prevention of postoperative nausea and vomiting – …

DEPARTMENT OF ANAESTHESIA AND INTENSIVE CARE MEDICINEHELSINKI UNIVERSITY HOSPITAL

UNIVERSITY OF HELSINKI, FINLAND

DEPARTMENT OF ANAESTHESIOLOGYUNIVERSITY OF OULU, FINLAND

PREVENTION OF

POSTOPERATIVE NAUSEA AND VOMITING

– STUDIES ON DIFFERENT ANTIEMETICS,

THEIR COMBINATIONS AND DOSING REGIMENS

Ritva Jokela

ACADEMIC DISSERTATION

To be presented by permission of the Medical Faculty of the University ofHelsinki for public discussion in Lecture Hall 3 of Biomedicum Helsinki,

Haartmaninkatu 8, 00290 Helsinki, on 26th September 2003,at 12 o’clock noon

Helsinki 2003

Supervised by

Professor Kari KorttilaDepartment of Anaesthesia and Intensive Care MedicineHelsinki University HospitalHelsinki, Finland

and by

Professor Seppo AlahuhtaUniversity of OuluOulu, Finland

Reviewed by

Docent Timo Ali-MelkkiläAdministrative CentreTurku University HospitalTurku, Finland

and by

Docent Pekka TarkkilaDepartment of Anaesthesia and Intensive Care MedicineHelsinki University HospitalHelsinki, Finland

Opponent

Docent Markku HynynenDepartments of Anesthesiology and Intensive Care MedicineHelsinki University Hospital, Jorvi HospitalEspoo, Finland

ISBN 952-91-6161-1 (paperback)ISBN 952-10-1300-1 (PDF)

YliopistopainoHelsinki 2003

http://ethesis.helsinki.fi/

To all those patients who have ever suffered frompostoperative nausea and vomiting

5

CONTENTS

ABSTRACT ............................................................................................................. 9

LIST OF ORIGINAL PUBLICATIONS ....................................................................... 11

ABBREVIATIONS .................................................................................................... 12

INTRODUCTION ..................................................................................................... 13

REVIEW OF THE LITERATURE ............................................................................... 15

INCIDENCE AND CLINICAL IMPLICATIONS OF POSTOPERATIVE NAUSEA

AND VOMITING ............................................................................................... 15

PHYSIOLOGY OF POSTOPERATIVE NAUSEA AND VOMITING ............................ 15

RISK FACTORS FOR POSTOPERATIVE NAUSEA AND VOMITING ........................ 19

Predictive scores for postoperative nausea and vomiting ........................... 19

Anaesthesia-related factors ............................................................................ 22

Premedication ............................................................................................ 22

Intravenous anaesthetics ........................................................................... 22

Anaesthetic gases ...................................................................................... 23

Intraoperative opioids ............................................................................... 25

Reversal of muscle relaxation ................................................................... 25

Postoperative pain and analgesics ........................................................... 25

Operation-related factors ............................................................................... 26

PREVENTION OF POSTOPERATIVE NAUSEA AND VOMITING ............................ 26

Anticholinergics ............................................................................................... 26

Dopamine receptor antagonists .................................................................... 27

Phenothiazines ........................................................................................... 27

Butyrophenones ......................................................................................... 27

Benzamides ................................................................................................ 28

Antihistamines ................................................................................................. 29

Serotonin receptor antagonists ..................................................................... 29

Ondansetron .............................................................................................. 29

Granisetron ................................................................................................ 30

Dolasetron .................................................................................................. 31

Tropisetron ................................................................................................. 31

Glucocorticoids ................................................................................................ 31

Combinations of antiemetics ......................................................................... 32

Double combinations ................................................................................ 33

Multimodal techniques ............................................................................. 33

6

Miscellaneous .................................................................................................. 34

Monitoring the depth of anaesthesia ...................................................... 34

Supplemental oxygen ................................................................................ 34

Clonidine .................................................................................................... 35

Ephedrine ................................................................................................... 35

P6 Acupressure ........................................................................................... 35

Future agents .................................................................................................. 36

TREATMENT OF ESTABLISHED POSTOPERATIVE NAUSEA AND VOMITING ...... 36

PONV – prophylaxis or treatment? ...................................................................... 36

Treatment of established postoperative nausea and vomiting ................... 37

AIMS OF THE STUDY ............................................................................................ 39

PATIENTS AND METHODS .................................................................................... 40

PATIENTS ............................................................................................................... 40

DESIGNS AND PROTOCOLS OF THE ORIGINAL STUDIES ..................................... 40

METHODS .............................................................................................................. 41

Premedication and monitoring ...................................................................... 41

Anaesthesia ..................................................................................................... 41

Postoperative symptoms ................................................................................. 44

Postoperative nausea and vomiting ......................................................... 44

Postoperative pain ..................................................................................... 44

Other postoperative assessments ............................................................. 45

Power analyses ................................................................................................ 45

Statistical methods .......................................................................................... 46

RESULTS ................................................................................................................. 47

PATIENTS ............................................................................................................... 47

OVERALL INCIDENCE OF POSTOPERATIVE NAUSEA AND VOMITING ............... 47

TROPISETRON COMPARED WITH ONDANSETRON, DROPERIDOL,

AND METOCLOPRAMIDE (AIMS 1 and 3) ...................................................... 48

Incidence of postoperative nausea and vomiting......................................... 48

Consumption of antiemetics .......................................................................... 49

Time to the first antiemetic dose ................................................................... 50

Pain and pain medication ............................................................................... 51

Adverse events ................................................................................................ 51

7

PROPOFOL COMPARED WITH SEVOFLURANE WITH AND WITHOUT

ONDANSETRON PROPHYLAXIS (AIM 2) ......................................................... 52


Consumption of antiemetics .......................................................................... 52

Time to the first antiemetic dose ................................................................... 53

Pain and pain medication ............................................................................... 53

Adverse events ................................................................................................ 53

ONDANSETRON, DEXAMETHASONE AND DROPERIDOL COMPARED TO

NO PROPHYLAXIS IN SINGLE, DOUBLE AND TRIPLE COMBINATIONS

(AIMS 3 AND 4) ................................................................................................ 53


PATIENT SATISFACTION (AIM 5) ........................................................................... 55

DISCUSSION .......................................................................................................... 56

METHODOLOGY ................................................................................................... 56

NAUSEA AND VOMITING, SEPARATE ENTITIES ................................................... 57

POSSIBLE IMPACT OF THE OPERATION SITE ON POSTOPERATIVE NAUSEA

AND VOMITING ............................................................................................... 58

POSSIBLE ASSOCIATION OF PAIN WITH POSTOPERATIVE NAUSEA AND

VOMITING ........................................................................................................ 58

5-HT3 RECEPTOR ANTAGONISTS, DROPERIDOL, DEXAMETHASONE AND

THEIR COMBINATIONS FOR PREVENTION OF POSTOPERATIVE NAUSEA

AND VOMITING ............................................................................................... 59

5-HT3 receptor antagonists ............................................................................. 59

Droperidol ....................................................................................................... 60

Metoclopramide .............................................................................................. 60

5-HT3 receptor antagonist, dexamethasone and droperidol alone and in

combinations .............................................................................................. 60

ORAL PREVENTION OF POSTOPERATIVE NAUSEA AND VOMITING .................. 63

PROPOFOL AND VOLATILE ANAESTHETICS FOR MAINTENANCE

OF ANAESTHESIA ............................................................................................ 63

CONCLUSIONS ....................................................................................................... 65

CLINICAL CONSIDERATIONS ................................................................................. 66

ACKNOWLEDGEMENTS ........................................................................................ 67

REFERENCES .......................................................................................................... 69

ORIGINAL PUBLICATIONS ..................................................................................... 79

9

ABSTRACT

Postoperative nausea and vomiting (PONV) are frequent complications having a medicalimpact and a considerable economic impact. For surgical patients, PONV is a most distress-ing adverse event, which they are reported to fear more often than postoperative pain.This study evaluated different antiemetics, their combinations and dosing regimens invarious surgical settings among patients at high risk for PONV.

The antiemetic efficacy of 5-HT3 receptor antagonists ondansetron and tropisetron,droperidol, dexamethasone and metoclopramide was investigated in altogether 5721 pa-tients. Two 5-HT3 receptor antagonists, ondansetron 8 mg and tropisetron 5 mg, bothgiven intravenously and combined with intravenous droperidol 0.75 mg, were comparedwhen given for prophylaxis after gynaecological laparoscopy. Furthermore, the two 5-HT3

receptor antagonists, ondansetron 16 mg, tropisetron 5 mg were compared withmetoclopramide 10 mg given orally for prophylaxis with premedication before thyroid orparathyroid surgery. The incidence of PONV in patients undergoing breast surgery aftersevoflurane anaesthesia with intravenous ondansetron 8 mg prophylaxis was comparedwith propofol anaesthesia; sevoflurane anaesthesia without antiemetic prophylaxis servedas a control. Tropisetron 5 mg was compared with droperidol 1.25 mg, both given intrave-nously at the beginning of laparoscopic cholecystectomy. Finally, ondansetron 4 mg, dex-amethasone 4 mg and droperidol 1.25 mg, all given intravenously, were compared assingle, double and triple combinations for prophylaxis after various types of surgery.

Ondansetron 8 mg and tropisetron 5 mg, both combined with droperidol 0.75 mg,resulted in the same incidence of PONV; 36% vs. 49%, respectively, P=0.28. The meantime to the first dose of rescue medication was shorter after tropisetron prophylaxis com-pared with ondansetron; 3 h 18 min vs. 6 h 25 min, P=0.007. Headache was significantlymore common after ondansetron prophylaxis than after tropisetron; 42% vs. 14%, P=0.004.There was no difference among the groups in regard to other side effects. Oral ondansetron16 mg, tropisetron 5 mg and metoclopramide 10 mg, resulted in the same incidence ofPONV, 68%, 58%, and 75%, respectively, during 0–24 h after surgery. During the earlyphase of recovery (0–2 h) the incidence of PONV was lower after tropisetron prophylaxisthan after ondansetron and metoclopramide; 15%, 32%, and 39%, respectively, P=0.051(tropisetron vs. ondansetron), P=0.004 (tropisetron vs. metoclopramide). The side effectprofile was similar after all three antiemetics. The patients were more satisfied with theprophylaxis with tropisetron than metoclopramide; 10.0 vs. 8.0 (numeric rating scale, NRS),P=0.001. The early (0–2 h) incidence of postoperative nausea was the same after propofolanaesthesia and sevoflurane with ondansetron 8 mg prophylaxis, but was higher aftersevoflurane anaesthesia; 8%, 7%, and 32%, respectively, P=0.002 (propofol vs.sevoflurane), P=0.001 (sevoflurane-ondansetron vs. sevoflurane). The incidence of vomit-ing did not differ after the three anaesthesia regimens during this phase. Sevoflurane withondansetron resulted in a lower incidence of PONV compared with propofol or sevoflurane;27%, 45%, and 60%, respectively, P=0.056 (sevoflurane-ondansetron vs. propofol),P<0.001 (sevoflurane-ondansetron vs. sevoflurane) during the 24-hour study period. Thesatisfaction of the patients was equal after all three anaesthesia types. Tropisetron 5 mgand droperidol 1.25 mg prevented postoperative nausea as efficaciously; 55% vs. 62%,respectively, ns., but tropisetron prevented vomiting more efficiently; 20% vs. 52%, re-spectively, P=0.001. The patients were drowsier after droperidol prophylaxis, the median

10

sedation score being 7 vs. 6, respectively (P=0.023). Otherwise the side effect profile ofthe two prophylaxes did not differ. The prophylaxis with droperidol 1.25 mg, ondansetron4 mg and dexamethasone 4 mg was associated with an early (0–2 h) reduction in theincidence of PONV with adjusted odds ratios (OR) of 0.48, 0.65 and 0.73, respectively. Thedouble and triple combinations resulted in a decrease in the incidence of PONV with anadjusted OR of 0.40 (ondansetron with dexamethasone), 0.31 (ondansetron withdroperidol), 0.46 (dexamethasone with droperidol) and 0.32 (ondansetron with dexam-ethasone and droperidol). All the single prophylaxis regimens resulted in an equal reduc-tion in the overall (0–24 h) incidence of PONV with an adjusted OR of 0.56 (ondansetron),0.56 (dexamethasone) and 0.57 (droperidol). The double and triple combinations wereassociated with a decrease with an adjusted OR of 0.34 (ondansetron with dexametha-sone), 0.37 (ondansetron with droperidol), 0.36 (dexamethasone with droperidol) and0.27 (ondansetron with dexamethasone and droperidol).

In conclusion, ondansetron and tropisetron are similar in the prevention of PONV. Theincidence of PONV is similar after propofol anaesthesia compared with sevoflurane anaes-thesia with ondansetron prophylaxis, while sevoflurane anaesthesia without antiemeticprophylaxis results in a significantly higher incidence of PONV. The anti-nausea effect oftropisetron resembles that of droperidol. The late anti-vomiting effect of tropisetron isstronger than that of droperidol. The antiemetic effect of ondansetron, droperidol anddexamethasone given as single prophylaxis is similar cutting the incidence of PONV byabout 10–15%. All the double combinations of the drugs are equally effective, decreasingfurther the incidence of PONV by 10%. The triple combination of the antiemetics reducesthe incidence of PONV additionally by 5%. Ondansetron, dexamethasone and droperidol,given in double or triple combinations, act with an additive effect without synergism.

Altogether, either of the 5-HT3 receptor antagonists, ondansetron or tropisetron, canbe used for prophylaxis. When a patient with a moderate individual risk of (20–40%)PONV is anaesthetized, either a volatile anaesthetic with antiemetic prophylaxis, or propofol,can be chosen for maintenance. Any of the studied antiemetics, 5-HT3 receptor antago-nist, dexamethasone and droperidol can be used for the prevention of PONV. When givenin combinations, all of the studied antiemetics act with additive effect without synergism.For high risk patients (≥ 40%) propofol anaesthesia with a combination of at least twoantiemetics prophylaxis is recommended.

11

LIST OF ORIGINAL PUBLICATIONS

I Jokela R, Koivuranta M. Tropisetron or droperidol in the prevention of post-operative nausea and vomiting. Acta Anaesthesiol Scand 1999; 43: 645–50.

II Koivuranta M, Ala-Kokko T, Jokela R, Ranta P. Comparison of ondansetronand tropisetron combined with droperidol for the prevention of emesis inwomen with a history of postoperative nausea and vomiting. Eur J Anaesthesiol1999; 16: 390–5.

III Jokela R, Koivuranta M, Kangas-Saarela T, Purhonen S, Alahuhta S. Oralondansetron, tropisetron, or metoclopramide in the prevention of postop-erative nausea and vomiting. A randomised, double-blind comparison in highrisk patients undergoing thyroid or parathyroid surgery. Acta AnaesthesiolScand 2002; 46: 519–24.

IV Jokela R, Kangas-Saarela T, Valanne J, Koivuranta M, Ranta P, Alahuhta S.Postoperative nausea and vomiting after sevoflurane with or withoutondansetron compared with propofol in female patients undergoing breastsurgery. Anesth Analg 2000; 91: 1062–5.

V Jokela R, Apfel CC, Kredel M, Bogacki I, Roewer N, Korttila K. Single, doubleand triple antiemetic prophylaxis with ondansetron, dexamethasone anddroperidol. (manuscript)

12

ABBREVIATIONS

5-HT3 5-hydroxytryptamine-subtype-3ASA American Society of AnesthesiologistsBMI body mass index (kg/m2)CI confidence intervalCIA confidence interval analysisCNS central nervous systemCO2 carbon dioxideCSF cerebrospinal fluidCTZ chemoreceptor trigger zoneD2 dopamine receptor subtype-2ECG electrocardiogramEPS extrapyramidal symptomsET end-tidalFDA United States Food and Drug AdministrationGABA γ-aminobutyric acidGI gastrointestinalHR heart rateIM intramuscularIV intravenousM muscarinicMAC minimum alveolar concentrationN numberNA not applicableNK neurokininNNT number needed to treatNRS numeric rating scaleN2O nitrous oxideNSAID nonsteroidal anti-inflammatory drugODT oral disintegrating tabletOR odds ratioP probabilityPACU postanaesthesia care unitPCA patient controlled analgesiaPONV postoperative nausea and vomitingPOV postoperative vomitingQ1 25% percentileQ3 75% percentileQTc corrected QT intervalSD standard deviationSpO2 oxygen saturationTD transdermalTIVA total intravenous anaesthesiaVAS visual analogue scaleVDS verbal descriptive scaleVRS verbal rating scale

13

INTRODUCTION

Despite the rapid progress in the field of modern anaesthesia, the incidence of postopera-tive nausea and vomiting (PONV) still remains 25–30% (Lerman 1992). The increase of daysurgery in the 1990’s has been challenged by the high incidence of PONV. One of themajor limiting factors in the early discharge of day surgery patients is the presence ofPONV (Gold et al. 1989) implying that economic consequences are involved.

Managing PONV incurs costs for day surgery units through personnel costs associatedwith the direct management of PONV, costs of drugs used to prevent and manage PONV,costs of supplies used in caring for patients suffering from PONV, and costs associatedwith the extra time spent by patients with PONV in the postanaesthesia care unit (PACU)(Carroll et al. 1994). When all these costs were calculated, it was estimated that a delay indischarge associated with PONV caused a $415 revenue loss in a U.S. outpatient surgicalcenter in 1994 (Carroll et al. 1994).

Several antiemetics have been investigated as prophylaxis of PONV. Traditionalantiemetics, droperidol and metoclopramide, have been studied in different surgical set-tings (Henzi et al. 1999; Henzi et al. 2000). Newer agents, 5-hydroxytryptamine3 (5-HT3)receptor antagonists were earlier shown to prevent effectively chemotherapy-induced nauseaand vomiting. During the past decade, 5-HT3 receptor antagonists have been investigatedextensively in the prevention of PONV (Tramèr et al. 1997). Dexamethasone, which simi-larly was first used to prevent chemotherapy-induced emesis, has been lately investigatedfor the prophylaxis of PONV (Henzi et al. 2000). The above agents in different combina-tions have been studied too. A most impressive result was obtained when propofol anaes-thesia was combined with droperidol, 5-HT3 receptor antagonist and dexamethasone pro-phylaxis (Scuderi et al. 2000).

The cost-effective management of PONV as an objective, it has been debated whetherto prevent PONV or treat the established symptom. Using decision analysis, Watcha andSmith (1994) found droperidol to be the most cost-effective antiemetic, followed byondansetron and metoclopramide. They concluded that prophylaxis was cost-effective, ifthe incidence of PONV exceeded 30% (Watcha and Smith 1994). In another study per-formed with a decision analysis model (Hill et al. 2000) droperidol 1.25 mg was more cost-effective than droperidol 0.625 mg, ondansetron 4 mg or a strategy of treating only theestablished symptom. The 1.25 mg dose of droperidol has also been shown to be more cost-effective than a prophylactic 12.5 mg dose of dolasetron or placebo (Frighetto et al. 1999).

For patients PONV is a most distressing symptom. In addition to the costs, the un-pleasantness of PONV has value for them. From the patients’ perspective, total avoidanceof PONV would be preferable. Indeed, the U.S. patients were willing to pay for a totallyeffective antiemetic, and those with a history of PONV were willing to pay even extra (Ganet al. 2001).

Antiemetic prophylaxis appears to be cost-effective when directed to patients who areat high risk for PONV (Watcha and Smith 1994). Risk factors of PONV have been definedduring the past ten years. According to a simplified risk score to predict PONV (Apfel et al.1999) the main risk factors for PONV are female gender, nonsmoking status, the history ofPONV or motion sickness, and the use of postoperative opioids.

The aim of the present study was to find a regimen to prevent PONV. The presentstudy compared the antiemetic efficacy of tropisetron, a 5-HT3 receptor antagonist, with

14

droperidol, and that of tropisetron with ondansetron, another 5-HT3 receptor antagonist,both in combination with droperidol. Also the antiemetic efficacy of the oral prophylaxiswith ondansetron, tropisetron and metoclopramide was compared. The incidence of PONVafter propofol anaesthesia was compared with sevoflurane anaesthesia with or withoutondansetron prophylaxis. Finally, the incidence of PONV after the single, double and tripleantiemetic prophylaxis with droperidol, ondansetron, and dexamethasone was compared.All the comparisons were carried out among patients at high risk for PONV.

15

REVIEW OF THE LITERATURE

INCIDENCE AND CLINICAL IMPLICATIONS OF POSTOPERATIVE NAUSEA ANDVOMITING

The overall incidence of PONV has been difficult to assess, as its aetiology is multifactorialand there is seldom only one trigger stimulus. During the “ether age” the incidence ofPONV was reported to be as high as 75–80% (Blumfield 1899). At present, the overallincidence of PONV after general anaesthesia has been estimated to range from 25% to30% (Lerman 1992; Watcha and White 1992). The incidence of intractable nausea andvomiting has been reported to be 0.18% (Gold et al. 1989).

Prolonged PONV may result in dehydration and electrolyte imbalance, and may exposethe patient to an increased risk of pulmonary aspiration of gastric contents, and may triggercardiorespiratory reflexes (Andrews 1992). Persistent vomiting can cause opening of thesutures, especially in the head and neck area, and venous hypertension (Vance et al. 1973).

Intractable PONV can increase medical costs by increasing nursing care time, and delay-ing discharge from the postanesthesia care unit (PACU) or leading to unanticipated hospitaladmission (Gold et al. 1989; Carroll et al. 1995). Among ambulatory surgery patients, PONValone or combined with pain is the most common anaesthetic reason for delayed dischargeand unanticipated hospital admission (Gold et al. 1989; Hedayati and Fear 1999; Junger etal. 2001; Wu et al. 2003). Nausea was reported to continue on average for 1.7 days andvomiting for 0.7 days among outpatients suffering from the symptoms (Carroll et al. 1995).

From the patients’ point of view, PONV appears to be one of the most undesirablepostoperative complications (van Wijk and Smalhout 1990). When healthy subjects wereasked which adverse events they wanted to avoid, they were ready to accept moderatepain, decreased alertness and extra payment to avoid PONV (Orkin 1992). Eberhart et al.(2002) reconstructed the Orkin approach among patients scheduled for surgery undergeneral anaesthesia, and confirmed that patients were ready to accept a certain degree ofpain, drowsiness and additional costs to avoid PONV. In another study, in which patientswere asked which adverse events are important to prevent they rated vomiting the mostundesirable symptom, and even pain was rated less undesirable (Macario et al. 1999).When the patients’ willingness to pay was assessed, they were ready to pay for a com-pletely effective antiemetic (Gan et al. 2001). Those with a history of PONV were the mostwilling to pay extra.

PHYSIOLOGY OF POSTOPERATIVE NAUSEA AND VOMITING

Vomiting is the forceful expulsion of gastrointestinal contents through the mouth. Retch-ing is the rhythmic action of the respiratory muscles preceding vomiting. Both retchingand vomiting are objective patient experiences. Nausea is a subjective experience whichmay or may not be associated with vomiting. (Andrews 1992)

The mechanism of the vomiting reflex was first described by Borison and Wang (1953)to be a complex act involving the respiratory, gastrointestinal, and abdominal musculaturecontrolled by the vomiting centre. The vomiting centre is situated in the lateral reticularformation close to the tractus solitarius in the brain stem (Borison 1989; Andrews et al.1990). The vomiting centre can be triggered by stimuli from the periphery: the orophar-

16

Table 1. Classes of different antiemetics, their sites of receptor antagonism, administration times for prophylaxis, doses for prophylaxisand treatment, half-lives, contraindications and adverse effects. EPS=extrapyramidal symptoms, GI=gastrointestinal, NA=not assessed.*Half-life of biological activity

Class Drug Site of Route Administration Dose for Dose for Max dose Half-life Contra- Adverse

receptor time for prophylaxis treatment per day indications effects

antagonism prophylaxis

Anticholinergics Atropine Acetylcholine IM, IV Premedication 0.01 mg/kg 2–3 h Urinary tract Dry mouth,

obstruction restlessness

Scopolamine IM, IV Premedication 0.2–0.65 mg Glaucoma Sedation, dry mouth,

TD 1.5 mg restlessness, central

cholinergic syndrome

Phenotiazines Chlorpromazine Dopamine2 IM, IV Premedication 25–50 mg 30 h Sedation, EPS,

PO 10–25 mg hypotension, restlessness

Promethazine IM, IV, Premedication 12.5–25 mg Sedation, EPS,

PO hypotension, restlessness

Perphenazine IM Premedication 2.5–5 mg 5 mg 10 h Pheochromo- Sedation, EPS

IV 1 mg cytoma hypotension, restlessness

PO 2–4 mg

Prochlorperazine IV Premedication 2.5–10 mg 40 mg Sedation, EPS,

IM, PO 5–10 mg hypotension, restlessness

5–10 mg

Butyrophenones Droperidol Dopamine2 IM, IV End of surgery 0.625–1.25 mg 0.625–1.25 mg 127 min Parkinson Sedation, EPS, QTc

disease prolongation, torsades de

pointes, neuroleptic

malignant syndrome

Haloperidol IM, IV End of surgery 0.5–4 mg 1 mg 12 h Parkinson Sedation, EPS, QTc

7 µg/kg disease prolongation

17

Class Drug Site of Route Administration Dose for Dose for Max dose Half-life Contra- Adverse

receptor time for prophylaxis treatment per day indications effects

antagonism prophylaxis

Benzamides Metoclopramide Dopamine2 IM, IV End of surgery 10–20 mg 10–20 mg 1–2 h Pheochromo- Sedation, EPS

GI motility↑ cytoma

Antihistamines Cyclizine Histamine IM, IV, Premedication 25–50 mg 20 h Sedation, dry mouth,

PO restlessness

Hydroxyzine IM Induction of 25–100 mg 14 h Porphyria Sedation, dry mouth,

anaesthesia restlessness

PO Premedication 25–50 mg

Diphenhydramine IM, IV Premedication 10–50 mg 300 mg Sedation, dry mouth,

PO mg/kg restlessness

Serotonin Ondansetron 5-HT3; IV End of surgery 4–8 mg 1–4 mg 24 mg 3 Liver disease Headache, dizziness, liver

receptor central PO 8–16 mg enzyme elevation,

antagonists peripheral arrhythmias

Granisetron IV Induction of 1 mg 1 mg 3–6 h Liver disease Headache, dizziness

PO anaesthesia

Tropisetron IV Induction of 2–5 mg 2 mg 5 mg 8–42 h Liver disease Headache, dizziness

PO anaesthesia

Dolasetron IV During 12.5 mg 12.5 mg 4-8 h Liver disease Headache, dizziness, QT

PO anaesthesia 50 mg interval prolongation

Glucocorticoids Betamethasone Unknown IM Induction of 12 mg NA 5.5 h Adrenal suppression,

anaesthesia hyperglycaemia, wound

healing prolongation

Dexamethasone IV Induction of 2.5–5 mg NA 3.5 h Adrenal suppression,

anaesthesia 36–72h* hyperglycaemia, wound

healing prolongation

18

ynx, the gastrointestinal tract, the mediastinum, renal pelvis, peritoneum and genitalia,and centrally from the CNS: cerebral cortex, labyrinthine, the visual centre, the vestibularportion of the eighth cranial nerve, and the chemoreceptor trigger zone (CTZ) located inthe area postrema at the base of the fourth ventricle (Borison and McCarthy 1983; Andrewset al. 1990). As the area postrema has no effective blood-brain barrier, CTZ can be directlystimulated by chemicals in the cerebrospinal fluid (CSF) and blood (Andrews 1992).

The CTZ has high concentrations of encephalin, dopamine, and opioid receptors. Thearea postrema is rich in opioid, dopamine (D) and serotonin (5-hydroxytryptamine (5-HT))receptors. The nucleus of the solitary tract contains high concentrations of encephalin,histamine, muscarinic (M), and cholinergic receptors (Kovac 2000). Recently, substance Pand neurokinin-1 (NK1) receptors were found to be involved in the regulation of emesis(Maubach and Jones 1997). At present it is generally accepted that the key to controlPONV is the blocking of all the recognized receptors involved in emesis (Heffernan andRowbotham 2000). (Figure 1, Table 1 on the previous page).

Figure 1. Agonists and antagonists associated with nausea and vomiting.

19

RISK FACTORS FOR POSTOPERATIVE NAUSEA AND VOMITING

Predictive scores for postoperative nausea and vomiting

In order to direct the antiemetic prophylaxis to the patients at high risk, several attemptshave been made to identify the risk factors for PONV. During the 1990s, several indepen-dent models using logistic regression were created for this purpose (Palazzo and Evans1993; Koivuranta et al. 1997; Apfel et al. 1998; Sinclair et al. 1999) (Table 2).

Palazzo and Evans (1993) studied 147 patients undergoing orthopaedic surgery (Table2). The total incidence of vomiting was 27% in their study. They concluded that the historyof PONV, female gender, the history of motion sickness and interaction between gender,history of PONV and postoperative opioids are fixed risk factors. The relative effects of thefactors are summarized in an equation as follows:

Logit postoperative sickness = –5.03 + 2.24 (if postoperative opioids were used) + 3.97 (ifprevious PONV) + 2.4 (if female) + 0.78 (if history of motion sickness) – 3.2 (if female withprevious PONV).

Toner et al. (1996) further tested the above model in another hospital in 400 patientsundergoing various types of surgery. According to their study the accuracy of the equationin predicting the risk for PONV in a single patient is limited. However, the model waseffective in predicting the risk of PONV in groups of patients with various combinations ofrisk factors.

Koivuranta et al. (1997) surveyed 1107 inpatients undergoing various types of surgery(Table 2). The incidence of nausea was 52% and the incidence of vomiting 25% duringthe entire 24 h study period. According to the survey female gender, the history of PONV,long duration of surgery, the history of motion sickness, and nonsmoking are patient-related risk factors. The authors constructed the following risk score for PONV:

Score = 0.93 (if female) + 0.82 (if previous PONV) + 0.75 (if duration of surgery over 60min) + 0.61 (if nonsmoker) + 0.59 (if history of motion sickness).

The risk score of Apfel et al. (1998) was based on the data of 1137 patients undergoingear, nose and throat surgery (Table 2). Only postoperative vomiting (POV) was studied, notpostoperative nausea. The data of 1137 patients was split randomly into an evaluation setand a validation set, and the incidence of POV was 21% and 22% in the sets, respectively.According to the study age, gender, the history of motion sickness and/or POV, nonsmok-ing status and the duration of anaesthesia are independent risk factors. The risk of POVcan be estimated from the equation:

Risk (probability of POV) = 1 / (1 + e-z), where z = 1.28 (female gender) – 0.029 (age) – 0.74(smoking) + 0.63 (history of PONV or motion sickness) + 0.26 (duration) – 0.92.

The discrimination power of the score was then tested among 1091 patients undergoingvarious types of ophthalmologic and surgical operations under general anaesthesia withvolatile anaesthetics, and found to be accurate (Apfel et al. 1998).

Sinclair et al. (1999) studied 17,638 outpatients; they found the incidence of PONV tobe 4.6% in the PACU and 9.1% within 24 h of the surgery (Table 2). The patient popula-tion was divided into a model development set and a model validation set. The logisticregression model for assessing the risk of PONV age, gender, smoking status, history of

20

Table 2. Patient, anaesthesia and surgery related risk factors for postoperative nausea and/or vomiting in risk scores for predicting PONV.++ = strong correlation, + = some correlation, – = no correlation. *Oulu **Würzburg

Authors Study characteristics Factors in the score

Number of Types of Primary Total Female Non- History History BMI Age History of Use of Long Type of Type ofpatients surgery endpoint incidence (%) gender smoking of PONV of motion migraine post- duration anaesthesia surgery

of nausea sickness operative ofand/or opioids anaesthesia

vomiting

Palazzo& Evans 147 Orthopaedic Vomiting 27 ++ NA ++ + – – NA ++ – NA NA1993

Koivuranta 1107 Various Nausea 52 ++ ++ ++ ++ – – – ++ ++ – +et al. 1997 Vomiting 25 ++ – ++ – – – – – ++ – +

Apfel 1137 Ear, nose Vomiting 22 ++ ++ ++ ++ – ++ NA – ++ NA NAet al. 1998 and throat

Sinclair 17638 Various Nausea and 9 ++ ++ ++ ++ – ++ NA + ++ ++ ++et al.1999 vomiting

Apfel et al. 520 *+ Various Nausea and 56* ++ ++ ++ ++ – – NA ++ – NA –1999 2202** vomiting 31**

Stadler 671 Various Nausea 19 ++ ++ – – – – + + – ++ ++et al. 2003 Vomiting 10 ++ ++ – – – – – + – ++ –

PONV= postoperative nausea and vomiting, BMI = body mass index, NA = not applicable.

21

PONV, type and duration of anaesthesia, and type of surgery were included as indepen-dent factors (Sinclair et al. 1999).

The data of Koivuranta and Apfel scores were cross-validated afterwards by the twocentres (Apfel et al. 1999), and a simplified risk score was developed without losing dis-criminating power (Table 2). The simplified risk score included female gender, history ofmotion sickness or PONV, nonsmoking status, and the use of postoperative opioids. Ifnone, one, two, three, or four of the risk factors were present, the incidence of PONV was10%, 21%, 39%, 61% and 79%, respectively (Figure 2).

Figure 2. Flow chart for assessing the individual risk for postoperative nausea andvomiting according to Apfel et al. (1999) and suggested anaesthetic and antiemeticregimen.

When the first three scores (Palazzo and Evans 1993; Koivuranta et al. 1997; Apfel et al.1998) were evaluated, they all predicted PONV and POV with moderate accuracy, but thescore presented by Koivuranta was recommended for clinical use as it is easy to calculate(Eberhart et al. 2000). Compared with the scores presented by Palazzo & Evans and Sinclair,the score published by Koivuranta and the simplified Apfel score predicts PONV with betterdiscrimination power (Apfel et al. 2002). In another comparison with the Sinclair score (Pierreet al. 2002), the simplified Apfel score presented with favourable discriminating and calibra-tion properties for predicting the risk of PONV. However, when the simplified Apfel score andthe scores published by Palazzo & Evans, Koivuranta and Sinclair were evaluated, the agree-ment between them was found to be poor, and the authors claimed that this limited theusefulness of the scores (Thomas et al. 2002).

Recently, an epidemiologic study of 671 patients identified risk factors for postopera-tive nausea and vomiting as separate symptoms (Stadler et al. 2003) (Table 2). Accordingto this survey, female gender, nonsmoking status and general anaesthesia are risk factorsfor both symptoms, while history of migraine and the type of surgery, urological surgery

22

excluded, increase nausea only. In the study by Stadler and co-workers, the history ofPONV or motion sickness did not increase the risk of PONV.

Investigated separately, smoking was shown to reduce the risk of PONV (Chimbira andSweeney 2000). The effect of smoking in reducing the risk of PONV has been suggested toresult from the dopaminergic stimulus (Apfel et al. 1998) or the increased metabolism ofthe anaesthetic agents via induction of the hepatic P450 enzyme (Chimbira and Sweeney2000; Sweeney 2002). On the other hand, when the impact of obesity on the risk forPONV was studied separately, it was not found to be a risk factor (Kranke et al. 2001).Perioperative gastric emptying appeared not to be a predictor of early PONV in patientsundergoing laparoscopic cholecystectomy (Wattwil et al. 2002).

Anaesthesia-related factors

Premedication

Premedication with opioids stimulates CNS opioid receptors and, therefore, may increasethe incidence of PONV (Dundee et al. 1965). The incidence of PONV was shown to in-crease after intramuscular, intravenous, nasal, oral, and transmucosal administration ofopioids (Henderson et al. 1988; Nelson et al. 1989; Pandit and Kothary 1989).

The combination of anticholinergics with opioids was shown to decrease the inci-dence of PONV (Dundee et al. 1965). As scopolamine has a considerable sedative effect, ithas been used for premedication. Transdermal scopolamine in children before strabismussurgery decreased the incidence of POV (Horimoto et al. 1991). Compared with atropine,glycopyrrolate appears to possess no antiemetic effect (Salmenperä et al. 1992), probablybecause it does not cross the blood-brain barrier.

Used as premedication, benzodiazepines have been suggested to reduce the incidenceof PONV by decreasing the plasma levels of catecholamines (Jenkins and Lahay 1971), butthe finding has not been confirmed.

Clonidine, an alpha2 adrenergic agonist is an interesting option for premedicationbecause of its sedative property, and given as premedication before strabismus surgery itdecreases the incidence of PONV in children (Mikawa et al. 1995).

Altogether, appropriately chosen premedication may reduce the incidence of PONV byrelieving the anxiety and distress of patients.

Intravenous anaesthetics

Propofol is an IV anaesthetic used for the induction and maintenance of general anaesthe-sia, sedation for short and long operations, and in intensive care, and as an antiemetic atsubhypnotic doses (Rose and Watcha 1999). At a subhypnotic dose of 10 mg it is claimedto have a direct antiemetic effect when administered to treat established PONV in patientsafter general anaesthesia (Borgeat et al. 1992). Administered 20 hours after thyroidec-tomy at an infusion rate of 0.1 mg/kg/h propofol decreased the incidence of PONV incomparison with placebo (10% vs. 65%, respectively) (Ewalenko et al. 1996). A bolusdose of 10 mg followed by an infusion of 10 µg/kg was determined to reduce the inci-dence of PONV by 50% in patients after general anaesthesia in the PACU (Gan et al.1997). The mechanism of the antiemetic effect of propofol is unclear. Propofol does notantagonize D2 receptors, but may depress directly the CTZ, vagal nuclei and other centresinvolved with nausea and vomiting (Appadu et al. 1994). When nausea and vomiting wereinduced by the 5-HT releaser ipecacuanha in healthy volunteers, propofol reduced the

23

number and intensity of emetic events compared with placebo (Hammas et al. 1998). Inrats propofol was shown to reduce the level of serotonin in the area postrema and thecerebrospinal fluid by γ-aminobutyric acidA (GABAA)- mediated inhibition of the release of5-HT, and possibly to reduce directly the neuronal activity via GABAA in the area postrema(Cechetto et al. 2001).

When compared to thiopentone as an induction agent, propofol reduced the inci-dence of PONV (odds ratio (OR) 0.82, P=0.03) in a study on 4173 patients (Myles et al.1996). However, according to a meta-analysis, when used for induction of anaesthesia,the number needed to treat (NNT) of propofol was 9 and when used for maintenance, theNNT was 6 (Tramèr et al. 1997). For patients at high risk for PONV (20–60%), propofolappeared to have the best efficacy on early PONV with a NNT of 4.7. Consequently, high-risk patients benefited from the propofol maintenance. Intravenous induction of anaes-thesia with propofol has no effect on the incidence of PONV (Tramèr et al. 1997).

Because barbiturates are used only for inducing anaesthesia, and not for its mainte-nance, their emetogenic property is difficult to assess (Rose and Watcha 1999).Methohexithone as a sole anaesthetic resulted in a higher incidence of PONV comparedwith propofol after microlaryngeal surgery (Best and Traugott 1991). The use of the steroidhypnotic eltanolone resulted in a similar incidence of PONV compared with propofol andthiopentone in anaesthesia for termination of pregnancy (Kallela et al. 1994; Eriksson etal. 1995). The incidence of PONV was similar after the induction of anaesthesia withetomidate and propofol in the early period of recovery, but the incidence was lower afterpropofol induction in the late postoperative period (St Pierre et al. 2000). Ketamine, adissociative anaesthetic which has been used as an adjuvant in anaesthesia for its analge-sic properties, is associated with a dose-dependent increase in the incidence of PONV(Badrinath et al. 2000).

In conclusion, patients at high risk for PONV benefit from propofol maintenance, whilepropofol used only for induction of anaesthesia does not reduce the incidence of PONV.

Anaesthetic gases

The inhaled anaesthetics used earlier, i.e. ether and cyclopropane, appeared to be moreemetogenic than the halogenated anaesthetics, i.e. halothane, enflurane, isoflurane,sevoflurane and desflurane, of today (Bellville et al. 1960; Clarke 1984; Palazzo and Strunin1984). The emetic property of the old inhalational agents may result from the increase incatecholamines associated with their use (Jenkins and Lahay 1971). However, according toa recent study, inhaled anaesthetics appear to be the main cause of PONV in the earlyphase of postoperative recovery, but they seem not to be associated with the occurrenceof the delayed PONV (Apfel et al. 2002). In the study of Apfel et al. the incidence of PONVdid not differ after isoflurane, enflurane or sevoflurane anaesthesia. The incidence of emeticevents was also similar after sevoflurane and desflurane used in outpatient anaesthesia(33% vs. 38%, respectively) (Nathanson et al. 1995) and videoarthroscopy of the knee(8% vs.16%, respectively) (Naidu-Sjösvärd et al. 1998). However, sevoflurane resulted in alower incidence of PONV compared with isoflurane (9% vs. 24%) in a multicentre trialconducted among adult ambulatory patients (Philip et al. 1996). Isoflurane maintenance,on the contrary, resulted in a lower incidence of PONV compared with desflurane andsevoflurane (22% vs. 67% vs. 36%, respectively) after breast surgery (Karlsen et al. 2000).

The incidence of emetic events has been compared in several trials after differentinhaled anaesthetics and after propofol. In general, the incidence of PONV is lower afterpropofol maintenance than after maintenance with inhaled anaesthetics. A meta-analysis

24

showed a lower incidence of PONV after propofol maintenance compared with any inha-lational agent (Sneyd et al. 1998). In inpatients undergoing surgery under general anaes-thesia, Rowbotham et al. (1998) showed that total IV anaesthesia with propofol resultedin a decreased incidence of PONV compared with isoflurane, both in combination withremifentanil-infusion (complete response 68% vs. 50%, respectively). In a large trial(n=2010) total IV anaesthesia with propofol resulted in a reduced incidence of PONV com-pared with isoflurane during the first 24 h after surgery (31% vs. 40% among inpatients,17% vs. 29% among outpatients) (Visser et al. 2001). During hours 24 to 72 the inci-dence of emetic events did not differ between the two anaesthesia groups. Sevofluraneinduction followed by sevoflurane maintenance resulted in a higher incidence of POVcompared with propofol induction followed by propofol maintenance in patients under-going outpatient surgery; 33% vs. 10%, respectively (Fredman et al. 1995) and 32% vs.18%, respectively (Ræder et al. 1997). However, the occurrence of PONV was similar aftersevoflurane vs. propofol and sevoflurane vs. isoflurane anaesthesia (48% vs. 40% and51% vs. 50%, respectively) among patients scheduled for surgery under general anaes-thesia (Ebert et al. 1998). Desflurane resulted in a higher incidence of PONV comparedwith desflurane in combination with ondansetron or propofol maintenance (51% vs. 24%vs. 24%, respectively) among outpatients undergoing gynaecological laparoscopy (Erikssonand Korttila 1996). Desflurane and propofol maintenance, both in combination withondansetron 0.05 mg/kg and droperidol 20 µg/kg, led to a similar incidence of PONV(33% vs. 43%, respectively) after outpatient laparoscopic cholecystectomy (Ræderet al. 1998). In conclusion, propofol induction followed by propofol maintenance is war-ranted for high-risk patients. The emetic properties of inhalational agents do not differfrom each other.

For decades, nitrous oxide (N2O) has been considered an emetic agent and generallybelieved to increase the incidence of PONV. N2O can induce PONV by penetrating intoclosed spaces such as bowel (Eger and Saidman 1965; Giuffre and Gross 1986), by activat-ing the medullary dopaminergic system (Murakawa et al. 1994) and increasing cerebrospi-nal opioid peptides (Finck et al. 1995). In the mid-1990s, three meta-analyses or reviewsevaluated the evidence of N2O in regard to PONV (Divatia et al. 1996; Hartung 1996;Tramèr et al. 1996). According to Hartung, in 24 out of 27 studies the incidence of emeticevents was higher when N2O was omitted, but the analysis does not meet the criteria of asystematic review followed by a meta-analysis. Divatia et al. concluded that omission ofN2O reduced the odds of PONV by 37%. Tramèr et al. showed that omitting N2O de-creased the incidence of vomiting when the control incidence was high (NNT 5), anddecreased the incidence less when the control incidence was low (NNT 13). Nevertheless,the investigators concluded that omitting N2O had no effect on nausea or complete con-trol of emesis. Having several benefits, such as analgesic and hypnotic effects, predictabil-ity, fast onset and offset, and low cost (Ræder 1994), N2O has retained its position incommon anaesthesia practice. An anaesthesia regimen without nitrous oxide may be jus-tified for patients at high risk for PONV (Tramèr et al. 1996).

Altogether, volatile anaesthetics appear to be the leading cause of PONV in the earlyphase of recovery. The pro-emetic activity of different volatile anaesthetics is very similar.Omitting N2O reduces the incidence of vomiting but does not influence the incidence ofnausea.

25

Intraoperative opioids

The association between pain, administered opioids and PONV is complex. In the postop-erative period both pain (Andersen and Krohg 1976; Chia et al. 2002) and opioids havebeen reported to increase the incidence of PONV. In the logistic regression models of therisk factors for PONV (Palazzo and Evans 1993; Koivuranta et al. 1997; Apfel et al. 1998;Apfel et al. 1999) the use of postoperative opioids was included, while the use of intraop-erative opioids was not. In the survey by Stadler et al. (2003) patients with and withoutPONV received equal amounts of sufentanil during the anaesthesia. When the beta-adren-ergic blocking agent esmolol was compared with the ultrashort-acting opioid remifentanilas anaesthesia adjuvant, the incidence of PONV was significantly higher after remifentanilinfusion, but the patient satisfaction was equal after both treatments (Coloma et al. 2001).Remifentanil was compared with alfentanil, both in combination with propofol, and theincidence of PONV was equal in both groups (Dershwitz et al. 2002; Ozkose et al. 2002).In a comparison of fentanyl-propofol with ketamine-propofol in outpatient tubal ligations,there was no difference in PONV between the groups (Vallejo et al. 2002). According tothe data of today, therefore, the association of intraoperative opioids with PONV is notclear.

Reversal of muscle relaxation

The use of muscle relaxants has not been associated with a higher incidence of PONV. Theantagonism of muscle relaxation with neostigmine has been widely investigated in regardto PONV. The emetic effect is claimed to be prevented by using atropine concurrently withneostigmine (Salmenperä et al. 1992). The results of the trials concerning neostigmine inregard to PONV are contradictory. However, a systematic review by Tramèr and Fuchs-Buder (1999) showed a slight decrease in the incidence of PONV when the antagonismwith neostigmine was omitted, as well as some evidence of dose-responsiveness for neo-stigmine. A 2.5 mg dose of neostigmine appeared to be emetogenic. A similar findingconcerning the dose-responsiveness of neostigmine was published by Løvstadt et al. (2001).

Postoperative pain and analgesics

Postoperative pain, especially when its origin is visceral, was shown to be a considerablecause of PONV (Andersen and Krohg 1976; Chia et al. 2002). On the other hand, PONV isa well-known side effect of opioids, and the emetic centre has plenty of opioid receptors.In a simplified risk score of Apfel et al. (1999), the use of postoperative opioids has beenidentified to be an independent risk factor for PONV. In a recent survey, the growing doseof morphine administered after surgery increased the incidence of PONV, whereas postop-erative pain did not influence PONV (Stadler et al. 2003). At present, a multimodal ap-proach to treat postoperative pain is generally recommended to prevent postoperativeemetic events. The multimodal approach consists of the use of regional anaesthesia when-ever possible, and regular dosing of nonsteroidal analgesic inflammatory drugs (NSAID)and paracetamol (Jin and Chung 2001). There is evidence that NSAIDs and paracetamol,given concurrently with premedication, decrease the need for opioids in the postoperativeperiod (Montgomery et al. 1996). A multimodal analgesia with lidocaine gel on the steril-ization clips and ketoprofen IV lowered the pain and nausea scores, and lessened the needfor opioids and antiemetics in comparison to either of the analgesic regimens alone afteroutpatient tubal ligation (Eriksson et al. 1996).

26

There is little evidence in the literature that the emetic properties of opioids differ.However, postoperative pethidine was shown to lead to a higher incidence of PONV whencompared with morphine (Ezri et al. 2002). It is possible that the incidence of PONV andthe other side effects of opioids decrease, if they are administered orally instead of IV orIM. The oral administration of the slow-release opioid oxycodone was shown to reducethe total dose of opioid needed, and consequently the incidence of PONV compared withimmediate-release oxycodone within the first three postoperative days after outpatientanterior cruciate ligament surgery (Reuben et al. 1999). Furthermore, slow-releaseoxycodone given as premedication before ambulatory laparoscopic tubal ligation was re-ported to result in a lower incidence of PONV when compared with the traditional IVadministration of fentanyl postoperatively (Reuben et al. 2002).

Operation-related factors

There is evidence that surgery in the head, neck and ear, nose and throat, plastic,gynaecological and abdominal surgery are associated with a high incidence of PONV(Honkavaara et al. 1994; Sonner et al. 1997). It appears that the patient-related risk fac-tors are the most important ones to be considered (Apfel et al. 1999). However, in a recentepidemiological study the type of surgery, urological surgery excluded, appeared to influ-ence the incidence of nausea but not of vomiting (Stadler et al. 2003). In the study ofStadler et al. (2003) the highest incidence of PONV occurred among patients undergoinggynaecological (32%), abdominal (26%), maxillofacial (27%), plastic (25%), neurosurgi-cal (24%) and urological (19%) surgery. In certain cases, e.g. a patient with wired jaws,and after plastic or thyroid surgery, the consequences of emesis may be hazardous and/ordestroy the outcome of the surgery. Situations like this justify the employment of all avail-able preventive measures.

PREVENTION OF POSTOPERATIVE NAUSEA AND VOMITING

Anticholinergics

The cerebral cortex and pons are rich with different cholinergic and muscarinic receptors(McCarthy and Peroutka 1988). Muscarinic subtype-3 and subtype-5 receptor antagonistsact as antiemetics (Kovac 2000). (Figure 1) Atropine and scopolamine cross the blood-brain barrier and help to prevent motion sickness and PONV. There is evidence that the useof glycopyrrolate, instead of atropine, results in a higher incidence of PONV (Salmenperäet al. 1992). Glycopyrrolate 10 µg/kg and neostigmine resulted in a higher incidence ofPOV compared with atropine 15 µg/kg (81% vs. 56%, respectively) also among childrenundergoing tonsillectomy with or without adenoidectomy (Chhibber et al. 1999). On theother hand, compared with placebo, glycopyrrolate given before spinal anaesthesia forcaesarean section, resulted in a lower incidence of PONV (Ure et al. 1999). By addinganticholinergic drugs like atropine and scopolamine to an opioid for premedication, theincidence of PONV can be decreased (Dundee et al. 1965). In a recent meta-analysis,transdermal scopolamine was shown to prevent PONV effectively (Kranke et al. 2002).However, the incidence of side effects was inexcusably high. The most often complainedof side effects are visual disturbances, sedation, dry mouth, dizziness and memory dys-function.

27

Dopamine receptor antagonists

Phenothiazines (e.g. promethazine, prochlorperazine), butyrophenones (droperidol, halo-peridol), and benzamides (metoclopramide) are strong dopamine-2 (D2) receptor antago-nists. (Figure 1)

Phenothiazines

Phenothiazines, i.e. promethazine, chlorpromazine, prochlorperazine, perphenazine,thiethylperazine, have been widely used for the prevention and treatment of postopera-tive emesis. The compounds consist of a tricyclic nucleus with an aliphatic (promethazine,chlorpromazine) or heterocyclic (prochlorperazine, perfenazine, thiethylperazine) ring at-tached to it. The antiemetic activity appears to be higher, and the sedative effect lower, ofthose phenothiazines with a heterocyclic ring than of those with an aliphatic ring (Dundeeet al. 1965). The phenothiazines block directly the D2 receptors in the CTZ, having alsosome antihistaminergic and anticholinergic action. They act as sedatives and counter theeffect of e.g. opioids in the CTZ (Howat 1960; Dundee et al. 1965; Wood 1979). The sideeffect profile of phenothiazines is much like that of the other dopamine receptor antago-nists. The incidence of extrapyramidal symptoms is higher with the heterocyclic phenothi-azines than the aliphatic ones (Howat 1960; Dundee et al. 1965; Dundee et al. 1975). Theneuroleptic malignant syndrome is a rare side effect, which has been reported with all thedopamine receptor antagonists (Kovac 2000).

The tricyclic nucleus of heterocyclic phenothiazines has a piperazine ring at the tenthposition of the nucleus. Prochlorperazine and perfenazine are the most widely used het-erocyclic phenothiazines, and they are claimed to exert antiemetic activity against emesiscaused by opioids (Dundee et al. 1975; Kovac 2000).

Butyrophenones

Butyrophenones, i.e. haloperidol and droperidol, are strong D2 receptor antagonists actingat the CTZ and area postrema (Loeser et al. 1979). Both are effective antiemetics for theprophylaxis and treatment of PONV (Kovac 2000). Droperidol has been studied widely indifferent anaesthetic settings, but among anaesthesiologists the experience of haloperidolmay be limited to its use in chronic pain. The mean elimination half life of droperidol is 127min (Lehmann et al. 1988). At large doses, 2.5 mg to 5 mg, the risk of potential sideeffects of droperidol, such as drowsiness, restlessness, abnormal movements, dizziness,vertigo, anxiety and extrapyramidal symptoms increases (Henzi et al. 2000). In this system-atic review, droperidol proved to be antiemetic, but this effect is short-lived. Its effect onnausea is more pronounced than its effect on vomiting. According to the systematic re-view (Henzi et al. 2000) the anti-nausea effect of droperidol is not dose-dependent, whileits anti-vomiting effect appears to be dose-dependent. There was dose-responsiveness foradverse events associated with droperidol (Henzi et al. 2000). Droperidol appears to pos-sess antiemetic activity in prophylaxis, even at small doses of 0.625 mg (Kreisler et al.2000). However, in a placebo-controlled multicentre trial on 2061 patients, a 1.25 mgdose of droperidol compared favourably with a dose of 0.625 mg, as well as with 4 mg ofondansetron, when one weighs the costs against the benefits gained from having a addi-tional PONV-free, side-effect-free patient as a result of successful prophylaxis (Hill et al.2000). At a dose of 20 µg/kg droperidol was shown to prevent postoperative emesis effec-tively among female patients but not among male patients, whose incidence of PONVis low anyway (Hechler et al. 2001). For nausea and vomiting caused by opioid–patient–

28

controlled–analgesia (PCA), droperidol seems to be the only effective antiemetic at a dose0.017–0.17 mg/mg of morphine (Tramèr and Walder 1999). When droperidol is used forthis purpose, there is no evidence of dose-responsiveness for its efficacy, but the risk ofadverse events is dose-dependent. Epidurally administered droperidol appeared to preventnausea, vomiting and pruritus, while IV administered droperidol prevented only nauseaand vomiting (Nakata et al. 2002). When 0.625 mg of droperidol was given for antiemeticprophylaxis before gynaecological dilatation and curettage in a placebo-controlled trial,body sway was significantly greater in the droperidol group at the time of achieving apost-anaesthesia discharge score of 9 (approximately 36 min after surgery), but no longerat the time of discharge home (approximately 92–102 min after surgery) (Song et al. 2002).In this trial the body sway velocity was assessed while the patient was standing first on afirm surface, then on a foam surface with eyes open and closed. The data is inconsistentfor optimal timing of the prophylactic dose (Henzi et al. 2000). Based on the pharmacoki-netics, the prophylactic dose of droperidol should probably be given at the end of anaes-thesia (Henzi et al. 2000).

After over 30 years of clinical use of droperidol in anaesthetic practice, a “black box”warning was placed in the labels by the United States Food and Drug Administration (FDA)in December 2001. In Europe, droperidol has been withdrawn in the United Kingdom andSpain. The FDA warning was based on the dose-dependency in cases of prolonging thecorrected QT interval (QTc) and torsades de pointes associated with the use of droperidol(Gan et al. 2002). The warning consists of a recommendation for a pre-treatment screen-ing by 12-lead electrocardiogram (ECG) and extended ECG monitoring for up to threehours (White 2002). The FDA decision has met strong criticism for being unjustified, as thearrhythmias and electrophysiological changes associated with droperidol appear to be ex-tremely rare (Ben-David et al. 2002; Gan et al. 2002). From November 1997 until the endof January 2002 altogether 273 individual safety reports concerning droperidol were pro-cessed; 120 of them dealt with cardiac or respiratory events (Dershwitz 2002). Of these120 cases 19 involved doses smaller than 10 mg. Moreover, most of the reports concern-ing anaesthetic doses contained confounding factors (Dershwitz 2002). In the anaestheticsetting, droperidol is typically given while the patients are being continuously monitoredby ECG (White 2002). Furthermore, prolongation of the QT interval has been reportedwhen 5-HT3 receptor antagonist dolasetron has been given, and arrhythmia has occurredwhen another drug of the same class, ondansetron, has been administered concurrentlywith metoclopramide (Baguley et al. 1997; Kantor 2002). A black box warning has never-theless not been added to the labels of these drugs. For the time being, the FDA is per-forming a pharmacokinetic and pharmacodynamic study to define the effects of 0.625mg, 2.5 mg and 5 mg doses of droperidol on the QTc interval (McCormick 2002). The FDAwill not make liberations concerning the recommendations on the use of droperidol, be-fore the results of the study are available. Nevertheless, the study carried out by the FDAhas been claimed to have a surrogate endpoint (Scuderi 2003). It was proposed that anacceptable risk be established for antiemetic use, based on the risk associated with un-treated postoperative vomiting (Scuderi 2003). Also the risk of adverse events associatedwith other commonly used antiemetics should be investigated (Scuderi 2003).

Benzamides

Metoclopramide and domperidone are benzamides which are strong D2 antagonists.Metoclopramide is the most effective and most widely used antiemetic in the class (Roseand Watcha 1999) acting both centrally in CTZ and in the area postrema, and in the

29

periphery, i.e. gastrointestinal tract (Harrington et al. 1983). It does not haveantihistaminergic properties (Kovac 2000). Metoclopramide increases lower oesophagealsphincter tone and promotes gastric motility from the distal oesophagus to the ileocaecalvalve (Rose and Watcha 1999). The side effect profile of metoclopramide is much like thatof the other dopamine receptor antagonists: extrapyramidal symptoms, sedation, dizzi-ness, drowsiness, but the side effects appear to be milder (Henzi et al. 1999). Having arelatively short duration of action, metoclopramide should be given at the end of surgery(Kovac 2000). In a meta-analysis a 10 mg dose of metoclopramide IV was the best docu-mented regimen (Henzi et al. 1999), but the dose appeared to be ineffective as an anti-emetic prophylaxis, and did not show an increased risk of adverse events either. However,the authors state that the doses used in anaesthetic practice are too low. In chemotherapy-induced nausea and vomiting, much higher doses of metoclopramide, 2 mg/kg, resultedin a lower incidence of emesis compared with placebo, with only minor side effects(Gralla et al. 1981).

Antihistamines

Antihistamines (dimenhydrinate, diphenhydramine, cyclizine, hydroxyzine) block acetyl-choline in the vestibular apparatus and histamine H1 receptors in the nucleus of the soli-tary tract. (Figure 1) They act mainly in the vomiting centre and vestibular pathways, actingonly slightly in the CTZ. Antihistamines are effective for treating vertigo and motion sick-ness (Dundee et al. 1975; McKenzie et al. 1981). For the prevention of chemotherapy-induced nausea and vomiting they have been usually combined with other antiemetics, asthey counteract the extrapyramidal effects of dopamine receptor antagonists (Rose andWatcha 1999). Their use may be justified when preventing PONV resulting from vestibularstimulation such as in middle ear surgery and in patients with motion sickness (Rose andWatcha 1999). Their main side effects are sedation and drowsiness, dry mouth, visiondisturbances, urinary retention and prolonged recovery (Kovac 2000). Their sedative effectprevents their use in the postoperative period, at least in outpatients.

In a recent meta-analysis, dimenhydrinate was shown to prevent PONV in the earlyphase of recovery with NNT 8 and in the whole investigated period with NNT 5 (Kranke etal. 2002), but the side effects were poorly documented.

Serotonin receptor antagonists

The 5-HT3 receptor is highly specific and selective for nausea and vomiting. The key todiscovering the serotonin receptor antagonists was the finding that compounds resem-bling metoclopramide structurally lacked dopamine receptor affinity but had antiemeticactivity. (Figure 1) These compounds appeared to be specific and selective 5-HT3 receptorantagonists. After they were shown to be effective in nausea and vomiting resulting fromchemotherapy and radiotherapy, they were introduced for anaesthetic practice. The firstagent of the class to be utilized was ondansetron in 1991, and later in the 1990s granisetron,tropisetron and dolasetron became available (Rose and Watcha 1999). The advantage ofthe drugs of this class is their favourable side effect profile with no adverse events ofdopamine, muscarinic and histaminergic receptor antagonists (Rose and Watcha 1999).

Ondansetron

Since ondansetron was launched, it has been investigated widely in different surgical set-tings, and by now it is the best documented 5-HT3 receptor antagonist regarding efficacy

30

on PONV. The bioavailability of the drug is approximately 60% when given orally to healthyvolunteers. The mean elimination half-life of ondansetron is 3 hours (Markham and Sorkin1993). In several meta-analyses, the antiemetic efficacy of ondansetron has been provento be superior to placebo and metoclopramide and equal to that of droperidol and dexam-ethasone (Tramèr et al. 1997; Figueredo and Canosa 1998; Domino et al. 1999). For theprevention of PONV, an 8 mg dose of ondansetron IV was not more effective than a 4 mgdose (Tramèr et al. 1997; Figueredo and Canosa 1998). In one of the systematic reviews,the anti-vomiting effect of ondansetron was more pronounced than its anti-nausea effect(Tramèr et al. 1997). As the anti-nausea effect was weaker with the 4 mg dose, the 8 mgdose was regarded as the optimal dose. For the treatment of established PONV, a 1 mgdose of ondansetron appeared to be as effective as higher doses (Tramèr et al. 1997).When cost-effectiveness is considered, the treatment of established PONV is more effec-tive than its prevention (Tramèr et al. 1999), but if the prevention is directed at the patientsat high risk for PONV, it is more effective than placebo (Hill et al. 2000). In the extensive Hillstudy, 1.25 mg of droperidol was more cost-effective than 0.625 mg of droperidol or 4 mgof ondansetron. All in all, the regimen of PONV prophylaxis for patients at high risk forPONV appears to be more cost-effective than treating established PONV (White and Watcha1999). Compared with dosing before the induction of anaesthesia, dosing near the end ofsurgery resulted in better control of PONV (Sun et al. 1997; Tang et al. 1998). In severalstudies, ondansetron appears to be more effective in controlling postoperative vomitingthan nausea (Tramèr and Walder 1999). The most hazardous adverse events of the agentappear to be allergic reactions (Rose and Watcha 1999). Other side effects associated withondansetron have been headache, transient increase in liver enzymes, profound hypoten-sion, dizziness, flushing at IV site, and constipation (Russell and Kenny 1992; Tramèr et al.1997). Cardiac arrhythmias (Baguley et al. 1997; Kantor 2002) and acute myocardial is-chemia (Bosek et al. 2000) have been reported in association with the use of ondansetron.According to recent findings, ondansetron seems to inhibit the analgesic effect of con-comitantly administered tramadol (De Witte et al. 2001; Arcioni et al. 2002). The mecha-nism of the phenomenon is suggested to be a counteraction at 5-HT3 receptors which areresponsible for the pain transmission at the spinal level, and tramadol acts by inhibitingthe reuptake and enhancing the release of norepinephrine and 5-HT (Arcioni et al. 2002).

Granisetron

Granisetron has a longer half-life than ondansetron, 3 to 6 hours in healthy volunteers.(Plosker and Goa 1991). Granisetron, as well as tropisetron, differ from ondansetron inbinding in 5-HT3 receptors. The receptor blockade induced by granisetron or tropisetroncan not be reversed even at high serotonin concentrations, while the effect of ondansetroncan (Newberry et al. 1993). It is unclear whether this difference is clinically relevant. The 40µg/kg IV dose of granisetron was shown to be the minimum effective dose (Fujii et al.1998). However, in another study 1 mg of granisetron appeared to be the minimum effec-tive dose in adults given immediately before the induction of anaesthesia (Wilson et al.1996). A considerable proportion of the trials published with granisetron have been con-ducted in one centre, which makes it difficult to evaluate the efficacy of the agent inantiemetic prophylaxis (Kranke et al. 2001). The high cost of a single dose of granisetronmakes it unsuitable for clinical practice, at least in the USA (Cieslak et al. 1996).

31

Dolasetron

After IV administration dolasetron is rapidly converted by the plasma enzyme carbonylreductase to hydrodolasetron, which is an active metabolite and appears to be responsiblefor the antiemetic effect (Rose and Watcha 1999). Dolasetron has an ultra-short elimina-tion half-life of 9 minutes, while that of hydrodolasetron ranges from 4 to 8 hours after IVadministration, and 5 to 10 hours after oral administration (Balfour and Goa 1997). TheP450 2D6 isoenzyme is involved in the metabolism of hydrodolasetron (Balfour and Goa1997). The bioavailability of an oral dose ranges from 70% to 89% (Balfour and Goa1997). Given at the start of surgery, the minimum effective dose of dolasetron is 50 mg,but 12.5 mg when given at the end of surgery (Diemunsch et al. 1997; Korttila et al.1997). In another study, the timing of the 12.5 mg prophylactic dose of dolasetron has noinfluence on its efficacy (Chen et al. 2001). Dosed orally 1 to 2 hours before surgery, theminimum effective dose of dolasetron is 50 mg (Diemunsch et al. 1998). For the treatmentof established PONV, a 12.5 mg dose of dolasetron was the minimum effective dose (Kovacet al. 1997). The treatment of established PONV with a single dose of dolasetron has beenconsidered to lead to a reduced need for health care resources (Kovac et al. 1999). Theside effect profile of dolasetron is similar to that of ondansetron: headache, dizziness,light-headed feeling. ECG analysis occasionally reveals transient prolongation in PR, QRSintervals and in QTc (Balfour and Goa 1997). These changes are visible at the 12.5, 25, 50and 100 mg doses, but at the 12.5 mg dose the changes in PR and QTc are comparablewith the placebo group (Kovac et al. 1997; Kantor 2002). The effective dose of dolasetronseems to be cheaper than the effective dose of ondansetron, at least in the USA (Rose andWatcha 1999).

Tropisetron

Tropisetron is metabolised in the liver by the cytochrome P450 2D6 system. There are inter-individual differences in the metabolising speed of the compound because of the polymor-phism of the enzyme system (Lee et al. 1993). The distribution ratio of the extensive andpoor metabolizers in Caucasian populations is 12:1, respectively (de Bruijn 1992). Themean elimination half-life ranges from 8 h in extensive metabolizers to between 30 and 42hours in poor metabolizers (Lee et al. 1993). The bioavailability of tropisetron is dose-dependent, ranging from 52–66% (Lee et al. 1993). In a dose-finding study, a 2 mg doseof tropisetron given before anaesthesia was the minimum effective dose for the preven-tion of PONV (Capouet et al. 1996). However, most of the studies on tropisetron haveemployed the 5 mg dose, which was shown effective for the prophylaxis of PONV afterbreast (Chan et al. 1998) and gynaecological surgery (Zomers et al. 1993; Alon et al.1996). For the treatment of established PONV, the 2 mg dose was the optimal effectivedose (Alon et al. 1998). In a placebo-controlled comparison with 1.25 mg of droperidol,given at the end of surgery, 5 mg of tropisetron prevented vomiting but not nausea, whiledroperidol failed to prevent any emetic events (Purhonen et al. 1997). The side effectprofile of tropisetron does not differ from the profile of other 5-HT3 receptor antagonists.Tropisetron is marketed in only a few countries at a price a little higher than that ofondansetron.

Glucocorticoids

After corticosteroids were found effective in chemotherapy-induced nausea and vomiting,they were used in anaesthesia as prophylaxis against PONV. The mechanism of the anti-

32

emetic action of corticosteroids is unknown, but prostaglandin antagonism has been pro-posed to be the mechanism (Rich et al. 1980), and the release of endorphins is anotherexplanation suggested for the action (Harris 1982). It has also been speculated that themechanism of action is related to the anti-inflammatory and membrane-stabilising effect(Kovac 2000). The glucocorticoids betamethasone and dexamethasone have been studiedfor the prophylaxis of PONV. The half-life of betamethasone is 300 min and that of dexam-ethasone 200 min, but the biological half-life of dexamethasone is 36–72 h (Schimmerand Parker 1995). Betamethasone 12 mg given before the start of surgery, compared withplacebo, resulted in a significant decrease of PONV and pain in ambulatory foot andhaemorrhoid operations (Aasbø et al. 1998). In the past few years, dexamethasone hasbeen investigated extensively in different surgical settings. In a dose-ranging study where10 mg, 5 mg, 2.5 mg and 1.25 mg doses of dexamethasone were compared with placebo,2.5 mg dexamethasone was the minimum effective dose for prophylaxis after majorgynaecological surgery (Liu et al. 1999). In another study comparing dexamethasone atthe same doses, a 5 mg dose appeared to be optimal for preventing PONV in patientsundergoing thyroidectomy (Wang et al. 2000). When the effect of the timing of 10 mg ofdexamethasone was studied in a placebo-controlled trial, the timing before the inductionof anaesthesia compared well with the timing at the end of anaesthesia in regard to pre-venting PONV. The authors suggested that the onset time of the antiemetic effect of dex-amethasone was about two hours (Wang et al. 2000), which makes its use as a rescueantiemetic unwarranted. In a quantitative systematic review, a single prophylactic dose ofdexamethasone, compared with placebo was antiemetic displaying a more apparent lateefficacy (Henzi et al. 2000). The systematic review evidenced an increased antiemetic ef-fect if dexamethasone was combined with a 5-HT3 receptor antagonist. Associated benefi-cial effects have been reported when glucocorticoids have been used for antiemetic pro-phylaxis: Considerable pain relief has been demonstrated with the preventive use of 12mg of betamethasone (Aasbø et al. 1998), and decreased time to home readiness with 4mg of dexamethasone without evidence of prolonged wound healing (Coloma et al. 2001).In another meta-analysis the authors recommend the routine use of a single dose of dex-amethasone during paediatric tonsillectomy for the reduction of PONV (Steward et al.2001). The side effect profile of a single dose of betamethasone and dexamethasoneappears to be comparable to placebo (Aasbø et al. 1998; Henzi et al. 2000). No problemswith wound healing, or disturbances of glucose metabolism have been reported after asingle prophylactic dose (Henzi et al. 2000).

Combinations of antiemetics

The multifactorial origin of nausea and vomiting gives grounds to presume that a combi-nation of antiemetics antagonising different receptors would provide better control ofPONV (Heffernan and Rowbotham 2000). At present, when antiemetics of different phar-macological classes are available, it is possible to combine them without producing addi-tive side effects via CNS (Kovac 2000). Indeed, combinations of drugs of different classeshave proven to be a method for preventing PONV in high-risk patients (Eberhart et al.2002). Today it is assumed that there is synergism among different antiemetics (Henzi etal. 2000). Dexamethasone is considered to potentiate the effect of other antiemetics bysensitising the pharmacological receptors (Sagar 1991) while it decreases levels of 5-hy-droxytryptamine in neural tissue (Young 1981) and prevents the release of serotonin in theGI-tract (Fredrikson et al. 1992).

33

Double combinations

Double combinations of different antiemetics have been investigated in various surgicalsettings, and as a rule a double combination of drugs is more effective than any drugalone. A combination of droperidol 2.5 mg and ondansetron 4 mg was shown to be moreeffective than either of the drugs alone after abdominal surgery (Pueyo et al. 1996).Ondansetron 4 mg combined with droperidol 1.25 mg increased the incidence of com-plete response, compared with droperidol 1.25 mg (92% vs. 78%, respectively) in womenundergoing tubal ligation (McKenzie et al. 1996). When ondansetron 4 mg was combinedwith cyclizine 50 mg, the combination resulted in a lower incidence of vomiting thanondansetron 4 mg after ambulatory gynaecological laparoscopic surgery (Ahmed et al.2000). Dexamethasone 8 mg combined with ondansetron 4 mg in patients undergoingmajor gynaecologic operations resulted in better control of PONV than either of theantiemetics alone (Lopez-Olaondo et al. 1996). The result was confirmed in a study, inwhich the incidence of late vomiting was 4% after the combination prophylaxis comparedwith 35% after ondansetron prophylaxis (Rajeeva et al. 1999). McKenzie et al. (1997)compared a combination of ondansetron 4 mg and dexamethasone 20 mg withondansetron 4 mg alone, both with propofol anaesthesia, and did not find a significantdifference between the regimens after major gynaecological surgery. Another comparisonfailed to show a decrease in the overall incidence of PONV after the combination of dex-amethasone 8 mg and ondansetron 4 mg compared with both drugs alone, probablybecause the study was underpowered, as stated by the authors (Thomas and Jones 2001).Granisetron 1 mg combined with dexamethasone 5 mg, and granisetron 1 mg alone,resulted in a lower incidence of PONV than droperidol 1.25 mg in patients undergoingabdominal, gynaecological, breast or otolaryngological surgery (Janknegt et al. 1999). Intheir study the incidence of nausea, however, was significantly lower after the combina-tion of granisetron and dexamethasone than after granisetron alone. The antiemetic effi-cacy of the combination of droperidol 0.625 mg and metoclopramide 10 mg was bettercompared with ondansetron 4 mg after laparoscopic cholecystectomy (Steinbrook et al.1996). Ondansetron 4 mg, droperidol 1.25 mg and dexamethasone 8 mg were comparedin three double combinations; ondansetron combined with droperidol or dexamethasonewas found to be more effective than droperidol with dexamethasone among women un-dergoing major gynaecological surgery (Sanchez-Ledesma et al. 2002).

Multimodal techniques

A multimodal approach seems to result in a promising outcome in the prevention of PONV.Among patients undergoing ambulatory laparoscopy, multimodal management (total IVanaesthesia with propofol and remifentanil, without N2O or neuromuscular blockade, withaggressive IV hydration, with triple prophylactic antiemetics with ondansetron 1 mg,droperidol 0.625 mg and dexamethasone 10 mg, and with ketorolac 30 mg) resulted in a2% incidence of PONV compared with a 24% incidence after standard anaesthesia withprophylaxis of ondansetron 4 mg, and a 41% incidence after standard anaesthesia with-out any prophylaxis (Scuderi et al. 2000). The incidence of PONV was 24% after a multidrugapproach with ondansetron 4 mg, dexamethasone 4 mg, droperidol 1.25 mg andmetoclopramide 10 mg, compared with 49% after prophylaxis with propofol-infusionduring the operation and the first four hours after the operation, and 70% in the controlgroup without any prevention among patients undergoing breast and abdominal surgery(Hammas et al. 2002). Eberhart et al. (2002) used a multimodal antiemetic approach (total

34

IV anaesthesia with propofol, 80% supplemental oxygen, dexamethasone 8 mg, haloperi-dol 10 µg/kg and tropisetron 2 mg) in a group of high-risk patients with a predicted risk of79–87%, and compared the outcome with a control group with a risk of 53–57%,anaesthetised in a standard fashion with propofol-desflurane. The incidence of PONV was7% after the multimodal approach, and 41% after the standard anaesthesia. The patientsatisfaction was significantly higher after the multimodal management than after the stan-dard anaesthesia when assessed by the patients’ willingness to pay. In a recent study inwhich the patients were anaesthetised with desflurane after propofol induction, the addi-tion of dolasetron 12.5 mg or ondansetron 4 mg to the antiemetic combination of droperidol0.625 mg and dexamethasone 4 mg, did not improve the outcome (Tang et al. 2003). Inthe light of current data, a multimodal approach with propofol anaesthesia and a combi-nation of antiemetics, i.e. dexamethasone, 5-HT3 receptor antagonist and droperidol, is astate-of-the-art protocol when anaesthetising patients with a predicted high risk of PONV.

Miscellaneous

Monitoring the depth of anaesthesia

The dosing of sevoflurane by monitoring the depth of anaesthesia using the bispectral index(BIS®) of the electroencephalogram, compared with not monitoring, was shown todecrease the incidence of PONV in the phase II recovery room (16% vs. 40%, respectively),but not in the PACU among outpatients undergoing gynaecological laparoscopy (Nelskylä etal. 2001). Interestingly, in another study BIS®-monitoring decreased the incidence of post-operative nausea after desflurane but not after propofol anaesthesia in patients undergoinggynaecological surgery (Luginbühl et al. 2003). The impact of the titration of anaestheticsby monitoring the depth of anaesthesia on PONV warrants further investigation.

Supplemental oxygen

Supplemental oxygen at a concentration of 80% administered during colorectal surgeryand 2 h after surgery, compared to a concentration of 30%, was reported to reduce theincidence of PONV (Greif et al. 1999). The authors hypothesised that PONV would beincreased by the mild intestinal ischemia leading to a release of e.g. serotonin, and thesupplemental oxygen would prevent the ischemia in the intervention group. It should bepointed out, however, that in this trial the rescue therapy was not standardised, thus therescue antiemetic was given at the discretion of the anaesthetist. The risk factors for PONVwere not controlled between the groups either. The same investigators (Goll et al. 2001)showed that 80% oxygen administered as above was as effective as ondansetron 8 mgcombined with 30% oxygen in the prevention of PONV following gynaecologicallaparoscopy. According to these investigators, supplemental oxygen should not be a rou-tine anaesthetic regimen, and should not be used for premature infants, patients withchronic obstructive pulmonary disease or patients undergoing laser surgery, but should beconsidered in the case of patient at high risk for PONV (Greif and Sessler 2000). Theadvantage of the technique is its low cost and the high availability of oxygen. However, agroup in Finland reconstructed the approach of using supplemental oxygen for patientsundergoing gynaecological day-case laparoscopy, and did not find any difference between80% and 30% oxygen in regard to PONV (Purhonen et al. 2003). The Finnish group re-peated the study comparing 50% with 30% oxygen among women undergoing breast

35

surgery, and did not find any difference between the two regimens (Purhonen et al. 2003).Further studies are needed to evaluate the role of supplemental oxygen in the preventionof postoperative emesis.

Clonidine

The α2-adrenergic agonist clonidine at a dose of 4 µg/kg for premedication was shown toreduce the incidence of PONV, but the oral intake time was longer and recovery scorelower in comparison with clonidine 2 µg/kg, diazepam 0.4 mg/kg, or placebo in childrenundergoing strabismus surgery (Mikawa et al. 1995). Oral clonidine premedication 4 µg/kg was associated with a significant increase in the incidence of complete response (93%vs. 67%) compared with premedication with diazepam 0.4 mg/kg among children follow-ing strabismus surgery without any reported adverse effects (Handa and Fujii 2001).Clonidine 2 µg/kg given immediately before induction anaesthesia to women scheduledfor breast cancer surgery resulted in a significant decrease in the incidence of PONV com-pared with placebo with a NNT of 3 without any additive sedation or other clinically impor-tant side effects (Oddby-Muhrbeck et al. 2002). The mechanism of the antiemetic effect ofclonidine is unclear. Oddby-Muhrbeck considered the mechanism to be multifactorial, re-sulting from the reduction in anaesthetic need and in sympathetic outflow caused byclonidine, the analgesic effect of clonidine and its intrinsic effect on α2-adrenergic or otherreceptor types. Oral clonidine at a dose of 5 µg/kg given before surgery and at 12 and 24h after the initial dose, compared with placebo, was shown to reduce the cumulative PCAmorphine dose and the incidence of nausea and vomiting after major knee surgery (Parket al. 1996). As clonidine is an inexpensive drug and has a favourable side effect profilewhen dosed cautiously, it is an interesting addition to the antiemetic drug selection. De-layed recovery and drowsiness, as well as the risk of hypotension and bradycardia, mayprevent its use particularly in outpatient surgery.

Ephedrine

Ephedrine 0.5 mg/kg was first shown to have an antiemetic effect similar to that of droperidol0.04 mg/kg, both compared with placebo, during the first 90 min after surgery, but notduring the following 24 h (Rothenberg et al. 1991). In a placebo-controlled study, ephe-drine 0.5 mg/kg given 10 min before the end of surgery was shown to reduce the inci-dence of PONV from 46% to 18% during the first 3 h after abdominal hysterectomy, butthere was no difference between the groups during hours 3 to 24 (Hagemann et al. 2000).As there was no difference in blood pressure between the groups, and the patients werelying still during the first three hours, the preventive effect of ephedrine on motion sick-ness can not be the mechanism of action; ephedrine thus seems to have a specific anti-emetic effect on PONV (Hagemann et al. 2000). However, the antiemetic effect of ephe-drine seems to be of very limited duration, which makes it a less attractive option as anantiemetic.

P6 Acupressure

Acupressure is a non-pharmacological method, which has been employed for the prophy-laxis of PONV as well as for pregnancy-related and chemotherapy-induced nausea (Kovac2000). In a meta-analysis of non-pharmacological techniques, their use was found prom-ising in the prevention of PONV (Lee and Done 1999). In the trials included in the meta-analysis, acupressure was compared with placebo and prochlorperazine. In a placebo-

36

controlled comparison among children, acupressure was found as effective as droperidolin controlling early PONV (Wang and Kain 2002). When acustimulation performed withReliefband® was compared with ondansetron 4 mg, the acustimulation resulted in a lowerincidence of PONV (White et al. 2002). For the treatment of established PONV, acustimulationinduced by Reliefband® enhanced the antiemetic efficacy of ondansetron (Coloma et al.2002). Further studies are required to determine the role of P6 acupressure or acustimulationin the prevention and treatment of PONV.

Future agents

Substance P and NK1 receptors have turned out to be involved in the regulation of emesis.(Figure 1) With the use of radiolabelled substance P, neurokinin receptors, including NK1

receptors, have been found in the nucleus tractus solitarius and dorsal motor nucleus ofthe vagus nerve in rats and ferrets (Maubach and Jones 1997). The mechanism of theantiemetic action of NK1 receptor antagonists is suggested to result from their ability toblock the entrance of the emetic stimuli to the CNS (McLean et al. 1996; Grunberg 1997).Two NK1 receptor antagonists (GR205171, CP-122,721) are in process of clinical testing(Diemunsch et al. 1999; Gesztesi et al. 2000). The antiemetic activity of CP-122,721 ap-peared to be superior to that of ondansetron (Gesztesi et al. 2000). In the light of prelimi-nary clinical studies, the NK1 antagonists warrant further investigation as a new class ofantiemetics.

TREATMENT OF ESTABLISHED POSTOPERATIVE NAUSEA AND VOMITING

PONV – prophylaxis or treatment?

During the past few years there has been debate on whether PONV should be preventedor treated only when the symptoms turn up (Scuderi et al. 1999). The prophylaxis of PONVhas been opposed due to the limited efficacy and adverse effects of the antiemetic pro-phylaxis, as well as the cost-effectiveness of treating patients with the symptoms. Forthose who defend the treatment, the outcome measured with the incidence and frequencyof PONV is what they call the surrogate end point. From their point of view, the true endpoints that should be measured are unplanned hospital admissions and patient satisfac-tion (Fisher 1999). However, both parties seem to accept that patients at high risk of PONVbenefit from the prophylaxis, and the satisfaction among them is higher with the regimenof prevention (Scuderi et al. 1999; White and Watcha 1999). Using a decision analysismodel which includes costs, prophylaxis was shown to be cost-effective when the inci-dence of PONV was more than 30% (Watcha and Smith 1994). Altogether, for the abovereasons routine prophylaxis is not warranted for every patient but should be based on theassessed individual risk for PONV. Prophylaxis is not justified for patients with a lower than10% risk for PONV. For patients with a 10–30% risk, a single prophylaxis with e.g. droperidolshould be chosen, and for patients with an even higher risk a multimodal prophylaxis with5-HT3 receptor antagonist and dexamethasone should be used. When prophylaxis is used,the doses of the antiemetics chosen should be the minimal effective ones in order to avoidadverse effects. If a regimen of treatment of established PONV only is chosen, the symp-tom must be treated in an early phase.

37

Treatment of established postoperative nausea and vomiting

While immense efforts have been made to investigate the prevention of emetic events,there are only sparse trials dealing with established PONV. According to a meta-analysison the treatment of PONV (Kazemi-Kjellberg et al. 2001) 5-HT3 receptor antagonistsappear to have an effect on vomiting but their effect on nausea is more obscure. Themeta-analysis showed little evidence of the dose-responsiveness of the drugs: dolasetron12.5–100 mg, granisetron 0.1–3 mg, tropisetron 0.5–5 mg and ondansetron 1–8 mgseemed to control further emesis equally. There is even less evidence of the antiemeticeffect of the old antiemetics, droperidol and metoclopramide, in the treatment of PONV,so these should be further investigated.

If the patient has received antiemetic prophylaxis, the drug for the treatment of break-through PONV should be chosen from a class other than the one used for prevention. Theeffect of 5-HT3 receptor antagonists on the treatment of vomiting is confirmed, but theanti-nausea effect is not so clear (Kazemi-Kjellberg et al. 2001). If ondansetron 4 mg wasgiven for prophylaxis, repeat dosing of ondansetron for rescue treatment did not improvethe antiemetic effect compared with placebo among outpatients (Kovac et al. 1999). Theantiemetic effect of dexamethasone seems to start only after 2 h after its administration(Wang et al. 2000); thus the role of dexamethasone in the treatment of PONV is limited tothe most intractable symptom not responding to any other antiemetic agent. Although inthe literature there is not much evidence of the efficacy of droperidol on the treatment ofPONV, yet the agent has been used for this purpose for 30 years. Droperidol appears tohave an effect on PONV caused by opioids, and is the only antiemetic that has efficacyagainst nausea and vomiting associated with patient-controlled analgesia (Tramèr andWalder 1999). Metoclopramide 10 mg was found to have a placebo effect when used forprophylaxis (Henzi et al. 1999), but larger doses may have a role in treatment. (Figure 3)

Figure 3. Flow chart for the treatment of established postoperative nausea and vomit-ing when patients received no prophylaxis, droperidol, droperidol with dexametha-sone and droperidol with dexamethasone and 5-HT3 receptor antagonist for prophy-laxis.

38

In order to prevent further emetic events, the good nursing care of the patient is a keyissue. The patient should have stable hemodynamics and should be properly oxygenated.A >35% elevation in systolic blood pressure during the induction of anaesthesia was shownto be associated with an increased incidence of PONV (Pusch et al. 2002). Hypovolaemiacan cause nausea and vomiting, and should be treated aggressively. A multimodal paintreatment should be employed.

39

AIMS OF THE STUDY

The primary aim of the thesis was to find a regimen to prevent PONV. The effects ofdifferent antiemetics, their combinations and dosing regimens were studied among pa-tients at high risk for PONV undergoing elective surgery under general anaesthesia. Thespecific objectives of this thesis were:

1. To compare the prophylactic antiemetic property of two 5-HT3 receptor antago-nists, ondansetron and tropisetron dosed IV, both combined with droperidol (II),and orally in comparison with metoclopramide (III).

2. To test the hypothesis that the overall incidence of PONV after sevoflurane anaes-thesia with ondansetron prophylaxis is lower compared with propofol anaesthesia,sevoflurane anaesthesia without prophylaxis serving as a control (IV).

3. To compare the antiemetic property of 5-HT3 receptor antagonist tropisetron withdroperidol (I). Further, to test the hypothesis that the incidence of PONV after singleprophylaxis with 5-HT3 receptor antagonist ondansetron, dexamethasone ordroperidol is lower than after no prophylaxis, and that the incidence of PONV afterdouble combinations of the drugs is smaller than after any of the drugs given assingle prophylaxis. Finally, to test the hypothesis that the incidence of PONV after atriple combination of the antiemetics is lower than after any of the double combi-nations of the drugs (V).

4. To test the hypothesis that the interaction between ondansetron, dexamethasone,and droperidol is additive or synergistic when given in double and triple combina-tions to prevent PONV (V).

5. To find out whether the increased incidence of postoperative nausea and/or vomit-ing is associated with decreased patient satisfaction and increased distress (III, V).

40

PATIENTS AND METHODS

PATIENTS

The total number of patients in the present study was 567 in Studies I–IV and 5154 inStudy V. In Studies I–IV, 198 of the patients were scheduled for elective surgery in LaplandCentral Hospital, Rovaniemi, Finland, 343 in Oulu University Hospital, Oulu, Finland, and26 in Kuopio University Hospital, Kuopio, Finland (Study III). In Study V, 384 of the patientswere undergoing elective surgery in Women’s Hospital, Helsinki University Central Hospi-tal; the rest of the patients were scheduled for surgery in 28 hospitals in Austria (1 hospi-tal), Germany (20 hospitals), Great Britain (1 hospital), Finland (2 hospitals), Slovakia (1hospital), Spain (1 hospital), and Turkey (1 hospital).

Study I was based on 120 ASA physical status I–III patients who underwent laparoscopiccholecystectomy. The 88 ASA I–III patients in Study II were scheduled for gynaecologiclaparoscopy. Study III involved 179 ASA I–III patients undergoing thyroid or parathyroid sur-gery. Study IV consisted of 180 ASA I–III patients undergoing breast surgery. Study V wasbased on 5154 ASA I–III patients scheduled for various types of elective surgery. (Table 3)

All studies were approved by the ethics committees of the hospitals in question (Stud-ies I–IV Lapland Central Hospital, Rovaniemi, and Oulu University Hospital, Oulu, Study IIIKuopio University Hospital, Kuopio, and Study V Women’s Hospital, Helsinki UniversityHospital, Helsinki and University of Würzburg, Würzburg, Germany). All patients gavetheir oral and written informed consent before participation and randomization.

DESIGNS AND PROTOCOLS OF THE ORIGINAL STUDIES

A prospective, controlled and randomized study design was used in Studies I–V. Studies I,II, III, and V were performed in a double-blind fashion. The patients were randomizedusing a table of random numbers, and the codes were contained in a sealed envelope. Thesummary of the study design is presented in Table 3.

Study I. Altogether 120 female patients undergoing laparoscopic cholecystectomy wererandomly assigned to receive an IV injection of either tropisetron 5 mg or droperidol 1.25mg after the induction of anaesthesia. The primary endpoint was the incidence of PONV,and the second was the use of rescue medication among the study patients during 0–2 hand 2–24 h of follow-up. The patients were interviewed at 2 h and 24 h after surgery.

Study II. 88 female patients with a history of PONV scheduled for gynaecologicallaparoscopy were administered IV droperidol 0.75 mg after the induction of anaesthesia,and ondansetron 8 mg or tropisetron 5 mg at the end of surgery. The primary endpointwas the occurrence of PONV and the second was the use of antiemetics. The main out-comes were monitored for study periods of 0–2 h and 2–24 h and the patients wereinterviewed at 2 h and 24 h after surgery.

Study III. 179 female patients who underwent thyroid or parathyroid surgery wererandomly allocated to receive ondansetron 16 mg, tropisetron 5 mg, or metoclopramide10 mg orally 1 h before anaesthesia, together with premedication. The primary endpointwas the incidence of PONV and the second was the use of rescue antiemetics during thefollow-up of 0–2 h and 2–24 h. Interviews were conducted at 2 h and 24 h after surgery.

41

Study IV. 60 female patients scheduled for breast surgery received an IV anaesthesiawith propofol, while 120 patients were anaesthetized with sevoflurane. 60 patients anaes-thetized with sevoflurane received an IV injection of ondansetron 8 mg at the end ofsurgery. The primary endpoint was the incidence of PONV and the second was the use ofrescue antiemetic medication. The main outcomes were recorded during 0–2 h and 2–24h follow-up. The patients were interviewed at 2 h and 24 h after surgery.

Study V. 5154 patients with a risk of PONV of 40% or more according to Apfel et al.(1999) (two or more risk factors), undergoing various types of surgery were stratificallyrandomized into the study groups. They received, at an equal ratio, no prophylacticantiemetics or an IV injection of ondansetron 4 mg, dexamethasone 4 mg, droperidol 1.25mg, or a double or triple combination of the above drugs. Dexamethasone and droperidolwere administered within the induction of anaesthesia and ondansetron at the end ofsurgery. The primary endpoint was the incidence of PONV during the 0–2 h and 2–24 h offollow-up. The patients were interviewed at 2 h and 24 h after surgery.

METHODS

Premedication and monitoring

All the patients in Studies I–IV were premedicated orally with midazolam 7.5 mg about 1 hbefore anaesthesia. In Study V the patients were given premedication with benzodiaz-epine as customary in the department. The patients in Study III received the study drugaccording to their allocation concurrently with premedication. Monitoring included heartrate (HR), ECG, pulse oximetry (SpO2), end-tidal carbon dioxide (ET-CO2), end-tidal anaes-thetic concentration, and non-invasive blood pressure. Neuromuscular blockade was moni-tored using a train-of-four method.

Anaesthesia

In Studies I–IV anaesthesia was induced with propofol 1–3 mg/kg. Before the induction ofanaesthesia in Studies I–IV, the patients were given glycopyrrolate 0.2 mg and fentanyl2–4 µg/kg. Intubation was facilitated with rocuronium 0.5 mg/kg in Studies I–IV. In StudiesI–IV the patients were subjected to controlled mechanical ventilation via a tracheal tube withoxygen-enriched air. In Studies I-III the hypnosis was maintained with sevoflurane at an end-tidal concentration of 1–2.5% (Study I) and 1–2% (Studies II–III). In Study IV, for the mainte-nance of hypnosis, the patients received either propofol-infusion 3–10 mg/kg/ h or sevofluraneat an end-tidal concentration of 2%. In Studies I–IV further increments of fentanyl 1-2 µg/kgwere given during the operation as needed, judged on the basis of the haemodynamicmonitoring, i.e. if the blood pressure or pulse rose by 20% from the preceding value. Neuro-muscular reversal agents were not used in Studies I–IV. In Study I all patients had a nasogastrictube placed after the induction and removed before the extubation. In Study III, the functionof the vocal cords was examined using a laryngoscope directly after the extubation.

In Study V before the induction of anaesthesia with propofol 2–3 mg/kg the patientsreceived either a bolus of fentanyl 100–200 µg or an infusion of remifentanil 0.25 mg/kg/min according to stratification. The intubation was facilitated with rocuronium accordingto clinical needs. After the intubation, normocapnic mechanical ventilation was employedwith oxygen in nitrous oxide or oxygen enriched air according to the stratification. Mainte-nance of hypnosis was achieved with either propofol at a rate of 3–10 mg/kg/h or the

42

Table 3. Study designs and patient characteristics in Studies I-V. The primary endpoint of each study was PONV. (PONV=postoperativenausea and vomiting, IV=intravenously, Q1=25% percentile, Q3=75% percentile). NA=not assessed

I II III IV V

Number of patients 120 88 179 180 5154Study design Prospective, randomized, Prospective, randomized, Prospective, randomized, Prospective, randomized Prospective,

double-blind double-blind double-blind stratificallyrandomized,double-blind,

placebo-controlled

Study drugs and drug Tropisetron 5 mg, Ondansetron 8 mg, Ondansetron 16 mg, Propofol, sevoflurane, Ondansetron 4 mg,dosages droperidol 1.25 mg tropisetron 5 mg, tropisetron 5 mg, ondansetron 8 mg dexamethasone 4 mg,

droperidol 0.75 mg metoclopramide 10 mg droperidol 1.25 mg

Dosing regimen of study IV after induction of IV after induction of Orally 1 h before surgery IV at end of surgery IV after induction ofdrugs anaesthesia anaesthesia (droperidol), with premedication anaesthesia

at end of surgery (dexamethasone and(ondansetron and droperidol),

tropisetron) at end of surgery(ondansetron)

Type of surgery Laparoscopic Gynaecological Thyroid or parathyroid Breast surgery Various typescholecystectomy laparoscopy surgery of surgery

Gender; females/males 120 / – 88 / – 179 / – 180 / – 4200 / 954

Age (years); mean (±SD) 48 (±15) 39 (±8) 49 (±13) 51 (±13) 47 (±15)median [Q1, Q3] 49 [36; 61] 40 [33; 45] 49 [39; 57] 50 [43; 59] 46 [36; 58]

BMI (kg/m);mean (±SD) 27 (±4) 25 (±4) 26 (±5) 26 (±5) 26 (±5)median [Q1, Q3] 26 [24; 29] 24 [22; 27] 26 [23; 29] 26 [22; 29] 25 [22; 28]

43

I II III IV V

Number of risk factors forPONV

0 0 0 0 0 81 1 0 0 9 2222 18 2 20 31 12483 51 27 81 75 21124 50 59 78 65 1509

Risk for PONV (%)*;mean (±SD) 65 (±14) 73 (±10) 66 (±13) 61 (±17) 59 (±17)

Duration of anaesthesia(min); mean (±SD) 114 (±32) 93 (±54) 138 (±70) 111 (±66) 108 (±72)median [Q1, Q3] 109 [90; 125] 87 [44; 130] 137 [110, 177] 96 [56; 159] 95 [68; 135]

Duration of surgery (min);mean (±SD) 79 (±30) 64 (±46) 114 (±47) 79 (±61)median [Q1, Q3] 72 [58; 90] 54 [20; 99] 109 [79; 135] 62 [25; 120] NA

Intraoperative fentanylmg/kg; mean (±SD) 4.7 (±1.1) 4.4 (±1.7) 5.7 (±1.8) 4.5 (±1.6)median [Q1, Q3] 4.8 [3.9; 5.4] 4.0 [3.1; 5.3] 5.6 [4.5; 6.7] 4.3 [3.3; 5.3] NA

*According to a simplified risk score (Apfel et al. 1999).

44

volatile anaesthetics (sevoflurane, desflurane, or isoflurane) starting at 1 MAC accordingto the stratification. Analgesia during the anaesthesia was ensured with further boluses offentanyl 50–100 µg or remifentanil 0.1–0.5 µg/kg/min according to the stratification.

Postoperative symptoms

Postoperative nausea and vomiting

Assessment

In all studies the patients were observed for a period of at least two hours in the PACU,and for any episodes of PONV for the first 24 h postoperatively. The data were collected instructured questionnaires. In Studies I–IV a study nurse or one of the investigators inter-viewed the patients at two hours in the PACU and at 24 h on the ward. The number ofepisodes of nausea and vomiting, the request for and the time of rescue antiemetic treat-ment, and any adverse events were recorded at the time of the interviews. In Studies I–IIthe patients were asked to grade nausea as none, mild, moderate, or severe. In Study Vthe worst nausea in the PACU was assessed at two hours postoperatively in an interviewby an investigator or the study nurse. At 24 h another interview was conducted or awritten questionnaire including questions on nausea and emetic episodes was filled out bythe patients.

Rescue medication

In Study I the rescue antiemetic treatment was achieved with metoclopramide 10 mg IV. InStudies II–IV the further antiemetic medication was ensured with droperidol 0.75 mg IV. InStudy V the first emetic event was treated with ondansetron 4 mg IV, the second withdexamethasone 4 mg IV and the third with droperidol at a dose of 0.5–1.25 mg IV. Thetreatment of further emetic events was at the discretion of the investigator.

Postoperative pain

Assessment

In Studies I–IV a study nurse or one of the investigators interviewed the patients at twohours in the PACU and at 24 h on the ward. The intensity of postoperative pain, opioidpain medication, and any adverse events were recorded at the time of the interviews.To assess their average postoperative pain during the study period, the patients used an11-point numeric rating scale (NRS), 0 standing for no pain at all, and 10 for the worst painimaginable. The amount of oxycodone was recorded in milligrams. In Study V, worst painwas assessed at two hours postoperatively in an interview by an investigator or the studynurse, and at 24 h by an interview or written questionnaire including questions on pain.

Postoperative pain management

In Studies I–IV the postoperative pain relief was achieved with oxycodone 0.05 mg/kg inthe PACU and 0.1 mg/kg on the ward, and paracetamol 1 g given orally at 8 h intervals. InStudy V the patients received a dose of non-opioid analgesic during the anaesthesia. In

45

case the patient was stratified for remifentanil, he or she received a dose of piritramide,oxycodone, morphine or fentanyl before the end of surgery. In Study V the postoperativepain medication was provided with non-opioid analgesics given regularly. Further painrelief was obtained with opioids administered IV, IM, or with the PCA method, or using acontinuous epidural infusion technique with a mixture of local anaesthetics, opioids and0.9% sodium chloride.

Other postoperative assessments

In Studies I–II, to evaluate the adverse events of droperidol, postoperative sedation wasassessed after the first 2 h in the PACU by nurses on a four-point scale: 1=awake, 2=drowsy,3=asleep but easily arousable, and 4=fast asleep. The patients assessed their overall drowsi-ness during the whole study period using an 11-point scale (NRS), where 0 meant alertand awake, and 10 as drowsy as one could possibly feel.

In Study III the patients were asked to assess their satisfaction with the antiemeticprophylaxis using an 11-point scale (NRS) (0–10), where 0 meant the lowest possible and10 the highest possible satisfaction. In Study IV the patients were asked to evaluate theiranaesthesia experience using a scale of worse than expected, as expected, and better thanexpected. They were also asked about their willingness to have the same kind of anaesthe-sia in the future.

In Study V the Myles recovery score (1999) was filled in by the patients at 2 and 24 hafter the anaesthesia. The patients also filled out an anaesthesia satisfaction questionnaire24 h after anaesthesia. In the questionnaire they were asked to evaluate their distress withnausea and vomiting separately using a scale of not at all, a little, moderately, and verymuch.

Power analyses

The incidence of PONV was the primary endpoint that was used for power analyses inStudies I–V. Power analyses for Studies I–V were executed with a power of 0.80 anda significance level of 0.05. The sample sizes were calculated using a formula for categori-cal data provided by Pocock (1983): n (sample size per group) = [p1 × (100-p1) + p2 × (100-p2) / (p1-p2)

2] × k, where p1 is the incidence for the control group and p2 the incidence forthe intervention group, and k = constant (7.9, when 1–β = 0.8 and α = 0.05). For Study Ithe incidence of nausea after droperidol prophylaxis was assumed to be 50% and aftertropisetron prophylaxis 25%, resulting in a sample size of 55 patients per group. In StudyII the incidence of PONV was assumed to be 50% after tropisetron prophylaxis withdroperidol, and 20% after ondansetron with droperidol. The calculation resulted in a samplesize of 36 patients per group. In the design of Study III, the incidence of nausea wasexpected to be 70% after prophylaxis with metoclopramide and 45% after prophylaxiswith 5-HT3 receptor antagonists. The estimation resulted in a sample size of 58 per studygroup. In Study IV we estimated the incidence of nausea to be 60% after sevofluraneanaesthesia and 35% after ondansetron prophylaxis, resulting in a sample size of 59 pergroup.

In Study V the incidence of PONV after inhalational anaesthesia was expected tobe 60% and after propofol anaesthesia 40%. With a single antiemetic prophylaxis, theincidence was assumed to decrease to 26.67%, with a double combination to 17.78%,and with a triple combination to 11.85%. Based on these assumptions, to find out thesuperiority of a triple combination over a specific double combination after propofol

46

anaesthesia, 598 patients per group were calculated to be needed for detecting a differ-ence of 5.92% between the incidence of 17.78% of the double combination and 11.85%of the triple combination. The three double combinations were grouped and comparedwith the triple combination; thus 409 patients were needed per group. Consequently,23 × 409, i.e. 3272 patients were needed to investigate the effect of the three antiemeticson patients receiving propofol anaesthesia. As far as volatile anaesthesia was concerned,it was assumed to lead to a 3/2 higher incidence of PONV than propofol anaesthesia.We thus needed 249 patients per group to find out the difference between a doublecombination and the triple combination, i.e. between 26.67% and 17.78%. This summedup to 23 × 249=1992 patients and a total of 5264 patients.

Statistical methods

The statistics for all studies were analyzed using the Statistical Package for the Social Sci-ences (SPSS®) for Windows versions 6.1.3, 7.5, 8.0, 10.0 and 11.0.1 (SPSS Inc., Chicago,IL, USA). Confidence interval (CI) analyses for Studies I–IV were performed using the CIAprogram (Gardner and Altman 1989; Newcombe and Altman 2000). In Studies I–IV themain outcome was compared by calculating the 95% CI for the differences in proportionsand by calculating a P-value for the null hypothesis of no difference using the X2 or Fisher’sexact test. To compare the continuous data on the study groups, in Studies I and II we usedthe t-test assuming unequal variances, and in Studies III and IV one-way analysis of vari-ance. For detecting intergroup differences, a post-hoc analysis was performed in Studies III(Tukey’s adjustment) and IV (least significant difference adjustment). For comparisons ofnon-parametric data in Study III the Kruskal-Wallis and Mann-Whitney tests were em-ployed. In Study V the main outcome was compared between any two interventions byperforming a X2 test and calculating a corresponding risk ratio with a 95% CI. A multivari-ate analysis was performed by using a binary logistic regression analysis to investigatefurther the primary outcome.

47

RESULTS

PATIENTS

In Studies I and IV there were no exclusions. In Study II two of the consented 90 patientswere excluded for protocol violations. In Study III six of the consented 200 patients wereexcluded because of a re-operation during the study period and 15 for protocol violations.In Study V 5262 patients had consented to participate, of whom 63 were withdrawnbefore allocation into the study groups. Of the randomized 5199 patients, 45 were with-drawn due to incomplete outcome. Thus 5154 patients were left in the analysis.

OVERALL INCIDENCE OF POSTOPERATIVE NAUSEA AND VOMITING

The overall incidence of PONV after laparoscopic cholecystectomy was 70/120 (58%, 95%CI 50%–67%), after gynaecological laparoscopy 37/88 (42%, 95% CI 32%–52%), afterthyroid or parathyroid surgery 120/179 (67%, 95% CI 60%–74%), after breast surgery79/180 (44%, 95% CI 49%–63%) and after various types of surgery in the multicentrestudy 1724/5154 (33%, 95% CI 32%–35%) (Table 4, Figure 4).

Table 4. The overall incidences of nausea and vomiting (n (%)) in Studies I–V 0–2 h,2–24 h and 0–24 h after surgery.

Study I II III IV V Total

Type Laparo- Gynaeco- Thyroid or Breast Variousof surgery scopic logical parathyroid surgery types of

cholecys- laparo- surgery surgerytectomy scopy

Number ofpatients 120 88 179 180 5154 5721

0–2 hPONV-free 96 (80%) 77 (88%) 128 (72%) 151 (84%) 4333 (84%) 4785 (84%)PONV 24 (20%) 11 (12%) 51 (28%) 29 (16%) 821 (16%) 936 (16%)

2–24 hPONV-free 55 (46%) 57 (65%) 65 (36%) 111 (62%) 3719 (72%) 4007 (70%)PONV 65 (54%) 31 (35%) 114 (64%) 69 (38%) 1435 (28%) 1714 (30%)

0–24 hPONV-free 50 (42%) 51 (58%) 59 (33%) 101 (56%) 3430 (67%) 3691 (65%)PONV 70 (58%) 37 (42%) 120 (67%) 79 (44%) 1724 (33%) 2030 (35%)

48

TROPISETRON COMPARED WITH ONDANSETRON, DROPERIDOL,AND METOCLOPRAMIDE (AIMS 1 and 3)

Incidence of postoperative nausea and vomiting

After laparoscopic cholecystectomy, tropisetron 5 mg and droperidol 1.25 mg, both givenIV, resulted in the same incidence of PONV in the PACU (18% vs. 22%, respectively, ns).During the entire study period, the overall incidence of nausea was also the same afterboth prophylaxis regimens (55% vs. 62%, respectively, ns.), but the incidence of vomitingwas significantly lower after tropisetron (20% vs. 52%, respectively, P=0.01).

When combined with droperidol 0.75 mg IV, the prophylaxis with ondansetron 8 mgIV or tropisetron 5 mg IV resulted in the same incidence of PONV in the PACU (9% vs.16%, respectively, ns.) and during the entire study period (36% vs. 49%, respectively, ns.)among patients undergoing gynaecological laparoscopy.

Tropisetron 5 mg (T), dosed orally 1 h before thyroid or parathyroid surgery, resulted ina lower incidence of PONV in the PACU compared with ondansetron 16 mg (O) andmetoclopramide 10 mg (M) (15% vs. 32% vs. 39%, respectively, P=0.051 (O-T), P=0.004(M-T)). During the whole study period, the incidence of vomiting was significantly lowerafter the prophylaxis with tropisetron or ondansetron compared with metoclopramide (22%vs. 34% vs. 53%, respectively, P=0.001 (M-T), P=0.063 (M-O)). The incidence of nauseawas equal after each prophylaxis. (Figure 5)

Figure 4. The total incidences (%) and 95% confidence intervals (CI) of vomiting andPONV in Studies I–V.

49

Consumption of antiemetics

After laparoscopic cholecystectomy the use of antiemetics was the same in the PACU andduring the entire study period after administration of tropisetron 5 mg and droperidol1.25 mg. (Figure 6)

Figure 5. The incidence and 95% confidence intervals (CI) of vomiting and PONV inStudies I–IV, 0–2 h, 2–24 h and 0–24 h after surgery.

Figure 6. The proportions (%) and 95% confidence intervals (CI) of patients needingrescue antiemetic medication in Studies I–IV, 0–2 h, 2–24 h and 0–24 h after surgery.

50

Among patients undergoing gynaecological laparoscopy, tropisetron 5 mg andondansetron 8 mg, both combined with droperidol 0.75 mg, resulted in an equal use ofantiemetics in the PACU and during the whole study period. (Figure 6)

Among patients undergoing thyroid and parathyroid surgery, the use of antiemeticswas significantly lower after oral premedication with tropisetron 5 mg, compared withondansetron 16 mg and metoclopramide 10 mg in the PACU (13% vs. 30% vs. 36%,respectively, P= 0.045 (O-T), P= 0.006 (M-T)). During the entire study period the use ofrescue antiemetics was lower after tropisetron than after metoclopramide (50% vs. 71%,respectively, P=0.024) but did not differ after ondansetron prophylaxis (64%). (Figure 6)

Time to the first antiemetic dose

The mean time (95% CI) from the end of surgery to the first antiemetic dose was equalafter prophylaxis with tropisetron 5 mg and droperidol 1.25 mg among patients undergo-ing laparoscopic cholecystectomy (4 h 46 min (3 h 24 min–6 h 7 min) vs. 5 h 16 min (3 h38 min–6 h 57 min), respectively, ns.).

After gynaecological laparoscopy, the mean time (95% CI) to the first rescue anti-emetic dose was significantly longer after ondansetron 8 mg than after tropisetron 5 mg,both combined with droperidol 0.75 mg (6 h 25 min (4 h 25 min–5 h 4 min) vs. 3 h 18 min(1 h 57 min–4 h 16 min), respectively, P=0.007).

The mean time (95% CI) to the first rescue medication did not differ after the oralprophylaxis with tropisetron 5 mg, ondansetron 16 mg and metoclopramide 10 mg amongpatients undergoing thyroid or parathyroid surgery (5 h 47 min (4 h 8 min–7 h 26 min), 5h 14 min (3 h 38 min–6 h 51 min], 4 h 13 min, respectively (2 h 52 min; 5 h 33 min), ns).(Figure 7)

Figure 7. The mean times and 95% confidence intervals (CI) to the first antiemetic dosein Studies I–IV.

51

Pain and pain medication

In Studies I–III there were no differences among the groups in regard to average pain scoreduring the whole study period, or the doses of oxycodone in the PACU, or during theentire study period. (Table 5)

Table 5. Average pain scores (median [Q1, Q3]) and postoperative oxycodone consump-tion (mg/kg) (mean (±SD)) 0–2 h and 0–24 h after surgery in Studies I–IV. (Q1=25%percentile, Q3=75% percentile)

Average pain score Oxycodone0–2 h 0–24 h

Laparoscopic Tropisetron 5 mg 5 [3, 6] 0.15 (±0.10) 0.38 (±0.20)cholecystectomyStudy I Droperidol 1.25 mg 5 [4, 6] 0.14 (±0.08) 0.39 (±0.24)

Gynaecological Tropisetron 5 mg + 3 [2, 4] 0.15 (±0.10) 0.29 (±0.25)laparoscopy droperidol 0.75 mgStudy II Ondansetron 8 mg +

droperidol 0.75 mg 4 [2, 5] 0.19 (±0.12) 0.36 (±28)

Thyroid or Oral tropisetron 5 mg 5 [3, 5] 0.14 (±0.08) 0.35 (±0.22)parathyroid Oral ondansetron 16 mg 5 [3, 5] 0.13 (±0.07) 0.33 (±0.18)surgery Oral metoclopramide 10 mg 5 [3, 6] 0.13 (±0.07) 0.34 (±0.22)Study III

Breast surgery Propofol 3 [1, 4] 0.10 (±0.07) 0.16 (±0.14)Study IV Sevoflurane 3 [2, 4]* 0.11 (±0.11) 0.16 (±0.17)

Sevoflurane with 2 [1, 3]* 0.09 (±0.08) 0.22 (±0.22)ondansetron 8 mg

*P = 0.016.

Adverse events

The median of average sedation scores during the entire study period was significantlyhigher after prophylaxis with droperidol 1.25 mg compared with tropisetron 5 mg amongthe patients undergoing laparoscopic cholecystectomy. The median sedation scores didnot differ among the groups in Studies II and III. (Table 6)

52

Table 6. Sedation scores (median [Q1, Q3]) and the incidences of adverse events (%) inStudies I-IV. (Q1=25% percentile, Q3=75% percentile, NA=not assessed).

Sedation Headache Dizziness Pruritus Tremor, Visualscore restless- disturb-

ness ances

Laparoscopic Tropisetron 5 mg 6 [4; 8]* 37 28 13 7 NAcholecystectomyStudy I Droperidol 1.25 mg 7 [5; 8]* 27 32 15 15 NA

Gynaecological Tropisetron 5 mg + 6 [4; 8] 14† 23 12 12 NAlaparoscopy droperidol 0.75 mgStudy II Ondansetron 8 mg + 5 [4; 7] 42† 22 11 13 NA

droperidol 0.75 mg

Thyroid or Oral tropisetron 5 mg 6 [4; 7.25] 63 28 13 NA 7parathyroid Oral ondansetron 16 mg 7 [4.25; 8] 63 45 18 NA 13surgery Oral metoclopramide 10 mg 7 [5; 8] 54 36 10 NA 12Study III

Breast Propofol 4 [3; 6] 37 17 7 NA 5surgery Sevoflurane 4 [3; 7] 30 22 13 NA 5Study IV Sevoflurane with 4.5 [3; 6] 47 23 10 NA 2

ondansetron 8 mg

*P = 0.023, †P = 0.004.

In Studies I and III there were no differences in regard to the incidences of adverse eventsbetween the groups. Headache was significantly more common among patients receivingondansetron 8 mg combined with droperidol 0.75 mg compared with tropisetron 5 mgcombined with droperidol 0.75 mg after gynaecological laparoscopy in Study II. The inci-dences of other adverse events did not differ among the groups in Study II. (Table 6)

PROPOFOL COMPARED WITH SEVOFLURANE WITH AND WITHOUTONDANSETRON PROPHYLAXIS (AIM 2)


After breast surgery, sevoflurane anaesthesia (S) resulted in a higher incidence of nauseaonly in the PACU compared with propofol anaesthesia (P), or sevoflurane anaesthesia withondansetron 8 mg for prophylaxis (SO) (32% vs. 8% vs. 7%, respectively, P=0.002 (S-P),P=0.001 (S-SO)). The incidence of vomiting in the PACU was equally low in all groups. Theentire study period considered, the incidence of PONV was lower after sevoflurane anaes-thesia with ondansetron compared with propofol or sevoflurane alone (27% vs. 45% vs.60%, respectively, P=0.056 (SO-P), P<0.001 (SO-S)). (Figure 5)

Consumption of antiemetics

In the PACU, the use of rescue antiemetics was higher after sevoflurane anaesthesia thanafter propofol anaesthesia, or sevoflurane with ondansetron 8 mg for prophylaxis. Duringthe entire study period the use of rescue medication was higher after sevoflurane andpropofol anaesthesia than after sevoflurane combined with ondansetron 8 mg (48% vs.35% vs. 8%, respectively, P<0.001 (SO-S), P=0.001 (SO-P)). (Figure 6)

53

Time to the first antiemetic dose

The mean times (95% CI) to the first antiemetic dose were equal after propofol andsevoflurane anaesthesia with or without ondansetron 8 mg (7 h 26 min (5 h 3 min–9 h 48min) vs. 7 h 40 min (6 h 2 min; 11 h 17 min) vs. 6 h 15 min (4 h 31; 8 h 19 min),respectively, ns). (Figure 7)

Pain and pain medication

The median of the average pain scores was higher after sevoflurane anaesthesia comparedwith sevoflurane combined with ondansetron. The pain scores did not differ after propofolanaesthesia compared with either of the sevoflurane anaesthesias. The amounts ofoxycodone given per weight in the PACU and during the entire study period did not differbetween the different types of anaesthesias. (Table 5)

Adverse events

The median sedation scores were equal after all types of anaesthesia. The incidences ofadverse events did not differ after propofol or sevoflurane anaesthesia with or withoutondansetron. (Table 6)

ONDANSETRON, DEXAMETHASONE AND DROPERIDOL COMPARED TO NOPROPHYLAXIS IN SINGLE, DOUBLE AND TRIPLE COMBINATIONS (AIMS 3 AND 4)


During the first two hours after surgery, droperidol was the most effective of the threestudy drugs in the prevention of both nausea and vomiting (Table 7). Both ondansetronand dexamethasone prevented nausea better than vomiting, but the effect of ondansetronwas stronger. During the initial period of recovery the combination of ondansetron anddroperidol had the strongest efficacy on both nausea and vomiting compared with theother double combinations, and was equal to the efficacy of the triple combination.

During the later phase of recovery (2–24 h) dexamethasone had the strongest effecton nausea, and on nausea and retching or vomiting, but the effect on vomiting was equalto that of ondansetron. The anti-nausea and anti-vomiting effect of ondansetron wasmore evident than that of droperidol. Of the double combinations, the combination ofdexamethasone with droperidol was associated with the strongest reduction in retching orvomiting, and in nausea and retching or vomiting. The other two double combinations ofthe drugs were equally effective in the prevention of PONV. The triple combination was aseffective as the combination of dexamethasone with droperidol in decreasing retching orvomiting. The triple combination reduced nausea and nausea and retching or vomitingslightly better than the other two double combinations.

During the entire study period (0–24 h), all the three antiemetics were as effective inthe prevention of both nausea and vomiting. The combination of dexamethasone withdroperidol was slightly stronger than the other two double combinations in reducing retchingor vomiting, but did not differ from them in decreasing nausea. The other two doublecombinations were equal in preventing nausea and retching or vomiting. The combinationof dexamethasone and droperidol was almost as strong as the triple combination in reduc-ing retching or vomiting, otherwise the triple combination was the most effective in theprevention of nausea and vomiting.

54

Table 7. Frequencies and odds ratios of retching or vomiting and PONV after no prophylaxis and after single, double and triple prophylaxiswith ondansetron, dexamethasone and droperidol within 0–2 h, 2–24 h and 0–24 h after surgery in Study V. PONV = postoperative nauseaand vomiting.

OndansetronNo Ondansetron Dexametha- Droperidol Ondansetron Ondansetron Dexametha- with

antiemetic sone with with sone dexametha-dexametha- droperidol with sone and

sone droperidol droperidol

n = 5154 646 646 647 632 642 637 656 6480–2 h Retching or

vomiting Incidence n (%) 53 (8.2) 38 (5.9) 46 (7.1) 18 (2.8) 17 (2.6) 18 (2.8) 19 (2.9) 12 (1.9)Adjusted Relative odds 1.0 0.70 0.86 0.33 0.30 0.33 0.33 0.21

95% CI (ref) (0.45; 1.08) (0.57; 1.29) (0.19; 0.57) (0.17; 0.53) (0.19; 0.56) (0.20; 0.57) (0.11; 0.40)P 0.104 0.460 <0.001 <0.001 <0.001 <0.001 <0.001

PONV Incidence n (%) 170 (26.3) 121 (18.7) 134 (20.7) 92 (14.6) 81 (12.6) 63 (9.9) 93 (14.2) 67 (10.3)Adjusted Relative odds 1.0 0.65 0.73 0.48 0.40 0.31 0.46 0.32

95% CI (ref) (0.50; 0.84) (0.57; 0.95) (0.36; 0.63) (0.30; 0.54) (0.23; 0.42) (0.35; 0.61) (0.24; 0.44)P 0.001 0.018 <0.001 <0.001 <0.001 <0.001 <0.001

2–24 h Retching orvomiting Incidence n (%) 129 (20.0) 80 (12.4) 79 (12.2) 90 (14.2) 54 (8.4) 55 (8.6) 47 (7.2) 49 (7.6)

Adjusted Relative odds 1.0 0.57 0.56 0.67 0.37 0.38 0.31 0.3395% CI (ref) (0.42; 0.77) (0.41; 0.76) (0.50; 0.89) (0.26; 0.52) (0.27; 0.53) (0.22; 0.44) (0.23; 0.47)P <0.001 <0.001 0.007 <0.001 <0.001 <0.001 <0.001


95% CI (ref) (0.46; 0.72) (0.42; 0.66) (0.49; 0.77) (0.29; 0.46) (0.32; 0.52) (0.29; 0.47) (0.23; 0.39)P <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

0–24 h Retchingor vomiting Incidence n (%) 166 (25.7) 105 (16.3) 111 (17.2) 102 (16.1) 66 (10.3) 67 (10.5) 62 (9.5) 55 (8.5)

Adjusted Relative odds 1.0 0.56 0.60 0.56 0.33 0.34 0.30 0.2795% CI (ref) (0.43; 0.74) (0.46; 0.78) (0.42; 0.73) (0.24; 0.45) (0.25; 0.46) (0.22; 0.41) (0.19; 0.37)

P <0.001 <0.001 <0.001 <0.001 <0.001 <0.001P <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001


95% CI (ref) (0.45; 0.70) (0.45; 0.70) (0.45; 0.71) (0.27; 0.43) (0.29; 0.47) (0.29; 0.46) (0.21; 0.34)P <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

55

PATIENT SATISFACTION (AIM 5)

The satisfaction of patients was assessed in Studies III–V. After thyroid surgery in Study IIIthe patient satisfaction score was significantly (P=0.001) higher after oral tropisetron pro-phylaxis than after metoclopramide (Table 8). The patient satisfaction score did not differafter oral ondansetron prophylaxis compared with the other two prophylaxes. Dischargewas postponed because of PONV in one patient (2%) after ondansetron prophylaxis, threepatients (5%) after metoclopramide prophylaxis, compared with 0 patients after tropisetronprophylaxis.

After breast surgery in Study IV, the anaesthesia experience of the patients was equalafter each type of anaesthesia (Table 8). Also their willingness to have the same kind ofanaesthesia in the future did not differ after each type of anaesthesia.

In Study V the patients not receiving prophylaxis were significantly more distressedwith nausea and vomiting than the patients who had been given any single (P<0.001,P<0.001, respectively), any double (P<0.001, P<0.001, respectively) and the triple prophy-laxes (P<0.001, P<0.001, respectively). After any single prophylaxis they were more dis-tressed with both symptoms than after any double prophylaxis (P<0.001, P<0.001, respec-tively). There were no differences in patient distress after any double antiemetic prophy-laxes compared with the triple one.

Table 8. Patient satisfaction with the prevention of PONV and with the anaesthesiaexperience and willingness to have the same kind of anaesthesia in future.(PONV=postoperative nausea and vomiting, NA=not assessed, Q1=25% percentile,Q2=75% percentile).

Patient Willing tosatisfaction have the

with the Anaesthesia experience same kind ofprevention of anaesthesia

PONV in future

Score Worse As Better(0–10) than expected than

expected expected

Ondansetron 16 mg;median [Q1, Q3] 9 [8, 10] NA NA NA NA

Thyroid andparathyroid Tropisetron 5 mg;surgery median [Q1, Q3] 10 [8.25, 10] NA NA NA NAStudy III

Metoclopramide 10 mg;median [Q1, Q3] 8 [5, 10] NA NA NA NA

Propofol; n (%) NA 2 (3) 24 (40) 34 (57) 53 (88)Breastsurgery Sevoflurane; n (%) NA 6 (10) 28 (47) 26 (43) 53 (88)Study IV

Sevoflurane with NA 1 (2) 23 (38) 36 (60) 56 (93)ondansetron 8 mg; n (%)

56

DISCUSSION

METHODOLOGY

All the Studies (I–V) in this thesis have been carried out in at least two hospitals. Studies I,II and IV were performed in two hospitals and Study III in three hospitals in Finland. StudyV was a true multicentre study with 28 participating hospitals in Europe. The obviousreason for designing studies to be carried out in several hospitals is the power needed forthe study, i.e. sufficient sample size. When factors complicating recovery from surgery andanaesthesia are investigated, a relatively large sample is generally needed to achieve ad-equate power in the study (Apfel et al. 2002).

When the study is run in a multicentre mode, there are several problems that shouldbe avoided in order to maintain a high standard of research. Hospitals have different rou-tines and regimens in their anaesthesia practice. The doses of opioids used during theanaesthesia can vary, as well as the dosage of volatile and IV anaesthetics. The routine useof neuromuscular relaxation and its reversal may differ among the hospitals. The regimenof postoperative pain medication, the use of NSAIDs and opioids may also diverge. Suchproblems were minimized in Studies I-IV because there were no more than two investiga-tors in each hospital. The use of neuromuscular relaxants and reversal agents wasstandardised, as well as the use and the dosage of volatile and IV anaesthetics. The post-operative pain medication was also standardised. The doses of fentanyl used during ana-esthesia and the opioids used postoperatively were controlled among the groups. Further-more, the distribution of the study groups was controlled in each participating hospital.Study V was stratifically randomised for volatile anaesthetics or propofol, air or nitrousoxide, and fentanyl or remifentanil, and partly for light or deep anaesthesia according toBIS®. The use of anticholinergics, antibiotics, reversal agents and postoperative pain medi-cation was controlled in the groups. After closing the study, a plausibility check was per-formed in each centre. Thus, although performed in several centres, all the studies wereconducted in a controlled fashion.

Placebo was not used in Studies I–III. The type of surgery was laparoscopic in Studies Iand II, and thyroid or parathyroid surgery in Study III, both of them associated with a highincidence of PONV. According to an editorial (Gilbertson 1999) and the Declaration ofHelsinki (52nd WMA General Assembly, Edinburgh, Scotland, 2000) placebo may be usedfor comparison only if there is no effective treatment with which the study drug could becompared. Furthermore, in Studies I–III the patients’ mean individual risk (±SD) for PONVwas high (65% (±14), 73% (±10) and 66% (±13), respectively) calculated according toApfel et al. (1999). Hence, including a placebo group in Studies I–III was considered un-ethical. In Study IV the sevoflurane group served as a control, while the propofol groupand the sevoflurane group with ondansetron prophylaxis were to be compared with eachother. In Study V placebo was included. In Study IV the mean individual risk (±SD) was 61(±17) and in Study V it was 59 (±17). In Study IV breast surgery was included and in StudyV consisted of various types of surgery. The mean individual risk was thus smaller and thetype of surgery was not so uniformly associated with a high incidence of PONV, so acontrol group was justified in studies IV and V.

Nitrous oxide was omitted in Studies I–IV in order to avoid additional exposure to awell-known pro-emetic agent (Tramèr et al. 1996). Furthermore, the patients in Studies Iand II underwent laparoscopic surgery, where omitting N2O is justified because of its high

57

solubility and ability to expand cavities such as bowel (Eger and Saidman 1965). However,N2O was used in equal proportions of patients in the study groups in Study V. One of theaims of the large multicentre study was to compare the incidence of PONV after anaesthe-sia with or without N2O. Although N2O appears to be pro-emetic in high-risk patients,omitting N2O does not result in the total control of PONV. The required dose of opioidsand other anaesthetic agents may be higher when N2O is not used as an anaesthesiaadjuvant.

In a high-quality PONV study, the sample size should be large enough to provideadequate power, and the group comparability should be based on proven risk factors(Apfel et al. 2002). Nausea, vomiting and rescue medication should be recorded and re-ported separately, and they should be observed for 24 h after surgery (Apfel et al. 2002).The early (0–2 h) and late (2–24 h) recovery should be reported separately. For measuringnausea, a visual analogue scale (VAS) (0–10 or 0–100), an 11-point numeric rating scale(NRS) (0–10), or a verbal rating scale (VRS) (none, mild, moderate, severe) may be chosen(Boogaerts et al. 2000). Boogaerts et al. (2000) compared the VAS and a verbal descriptivescale (VDS) for assessing postoperative nausea, and concluded that the VAS is suitable forrecording the patient’s opinion about the intensity of nausea. A modification of VDS wasused in Studies I and II. The NRS has been widely used for assessing the intensity of pain asit correlates well with the conventional VAS (McQuay and Moore 1998). The NRS waschosen for assessing the intensity of pain in Studies I, II, III and IV, and the patient satisfac-tion in Study III in this thesis, as it is easy and quick to use. All the rating scales have certainlimitations, such as the transferral of sensation into another dimension (Boogaerts et al.2000).

NAUSEA AND VOMITING, SEPARATE ENTITIES

In this thesis Studies I–IV failed to identify the patients with vomiting and without nausea.The patients suffering from vomiting only tend to remain unidentified, as the proportionof these patients is so small. In Study I 27 (22.5%) out of 120 patients had only nausea,while 43 (35.8%) suffered from nausea and vomiting, 50 (41.7%) being free from thesymptoms. In Study II 37 patients out of 88 (42.0%) had only nausea, 12 (13.6%) hadnausea and vomiting, and 39 (44.3%) were without PONV. In Study III 56 out of 179patients (31.3%) had only nausea, and 64 (35.8%) both nausea and vomiting; 59 (33.0%)were free from the symptoms. In Study IV 32 out of 180 patients (17.8%) had nausea only,47 (26.1%) had both nausea and vomiting, and 101 (56.1%) had no symptoms at all. InStudy V 990 out of 5152 patients had only nausea (19.2%), 45 (0.9%) vomited withoutnausea, 687 (13.3%) suffered from both nausea and vomiting, and 3430 (66.6%) werefree from the complications. In Study V the incidence of vomiting was constantly less than50% of the incidence of PONV, which may indicate underreporting the incidence of vom-iting.

Nausea and vomiting are complications that generally occur together. Nausea is asubjective feeling and may be defined as a desire to vomit, or a feeling that vomiting isimminent (Andrews 1992). Nausea often occurs in waves. The intensity and duration ofnausea may vary in the course of recovery. Vomiting is an objective endpoint; it is a com-plex result of central autonomic reflexes and the motor system, which is followed by aforceful expulsion of GI contents through the mouth (Andrews 1992). Retching is an ex-pulsive movement of the stomach muscles without expulsion of GI contents. Retching isusually considered as vomiting (Tramèr 2001). The two endpoints are considered to be

58

two biologically different phenomena which should be reported and analysed separately(Tramèr 2001).

According to a recent epidemiologic study focusing on the difference in risk factorsfor PONV, about half of the patients with nausea suffer also from vomiting (Stadler et al.2003). The proportion of patients with vomiting and without nausea was 2% in theirsurvey. Female gender, nonsmoking status, and general anaesthesia seem to be risk fac-tors for both nausea and vomiting, while history of migraine and the type of surgery,urological surgery excluded, have an impact on nausea only (Stadler et al. 2003). Conse-quently, nausea and vomiting may have a different physiopathology. Interestingly,antiemetics prevent nausea and vomiting with a different profile of efficacy, e.g. droperidolprevents nausea better than vomiting, while 5-HT3 receptor antagonists prevent vomitingbetter than nausea (Tramèr 2001). Although there is no standard objective measure fornausea, it is not a less significant complication compared with vomiting. Intractable nau-sea lasting for hours may cause more suffering for an individual than a single vomit with-out any preceding nausea.

POSSIBLE IMPACT OF THE OPERATION SITE ON POSTOPERATIVE NAUSEA ANDVOMITING

In the present study, the overall incidence of PONV was 67% after thyroid or parathyroidsurgery (Study III) and 58% after laparoscopic cholecystectomy (Study I), while the inci-dence was 42% after gynaecologic laparoscopy (Study II), and 44% after breast surgery(Study IV). The incidences of PONV in the different studies should not be compared witheach other, e.g. because of the different antiemetic prophylaxis used in each. However, theanaesthesia regimen used in Studies I–III is quite similar: propofol induction, pain medica-tion with fentanyl, maintenance with sevoflurane and air, and omitting of neostigmine.Furthermore, 2/3 of the patients in Study IV were anaesthetized using the above regimenand 1/3 of the patients received propofol maintenance.

The models for the assessment of the individual risk for PONV differ among otherthings in proportion to the role of the type of surgery. The model of Sinclair et al. (1999)included the type of surgery. In a recent study (Stadler et al. 2003) investigating the differ-ence in risk factors for nausea and vomiting, the type of surgery, except for urologicalprocedures, appeared to influence postoperative nausea, but not vomiting. On the otherhand, in the models of Palazzo and Evans (1993), Koivuranta et al. (1997) and Apfel et al.(1998; 1999) the site of operation did not play a major role. However, the models showingno relation between the incidence of PONV and the type of surgery can be criticised fornot including an even proportion of patients scheduled for the different types of opera-tions.

POSSIBLE ASSOCIATION OF PAIN WITH POSTOPERATIVE NAUSEA AND VOMITING

In Studies I, II and III there were no differences in the average pain scores between thestudy groups. In Study IV the median of the average pain score was significantly higherafter sevoflurane anaesthesia than after sevoflurane with ondansetron prophylaxis. How-ever, the postoperative oxycodone consumption was similar in both study groups. As thedifference in the medians of average pain scores between the two groups is an occasionalfinding, it is probably a result of chance.

59

The possible association of pain with PONV remains unresolved. Only few studies havecome up with a correlation between postoperative pain and PONV (Andersen and Krohg1976; Chia et al. 2002). It is obvious that the association is extremely difficult to proveconclusively, as it tends to be demonstrated indirectly. Nevertheless, in order to minimisethe incidence of PONV, all the possible non-opioid methods are necessary in the preven-tion and treatment of postoperative pain. Further studies are needed to illuminate thepossible correlation between postoperative pain and PONV.

5-HT3 RECEPTOR ANTAGONISTS, DROPERIDOL, DEXAMETHASONE AND THEIRCOMBINATIONS FOR PREVENTION OF POSTOPERATIVE NAUSEA AND VOMITING

5-HT3 receptor antagonists

In Study II ondansetron was compared with tropisetron, both in combination with droperidol,as the prophylaxis was directed to the patients at high risk for PONV. In Study III ondansetronwas compared with tropisetron, both dosed orally with the premedication. In Study II theantiemetic effect of ondansetron and tropisetron appeared to be very similar, except thatthe mean time to the first emetic event was shorter after tropisetron than after ondansetron.The finding is contradictory with the elimination half-lives of the agents, and may be aresult of chance, as the sample size was rather small. In Study III tropisetron preventedPONV more efficaciously than ondansetron in the early phase of recovery, but in the latercourse of recovery the efficacy of the agents seemed to be the same. The better initialefficacy of tropisetron may be explained by the difference in the potency of the doses ofthe two agents. Droperidol given for rescue medication (30% vs. 13% after ondansetronand tropisetron prophylaxis, respectively) within the first two hours after surgery may inter-fere with the incidence of PONV during the later phase of recovery. The side effect profileof the two agents was very similar in both studies.

In the literature, there are only few studies comparing 5-HT3 receptor antagonists witheach other. Ondansetron was shown to be antiemetic and to have dose-responsiveness intwo systematic reviews followed by meta-analyses (Tramèr et al. 1997; Figueredo andCanosa 1998). The anti-vomiting effect of ondansetron is more evident than its anti-nau-sea effect (Tramèr et al. 1997). Other 5-HT3 receptor antagonists have not been investi-gated as extensively as ondansetron. Tropisetron was shown to possess an anti-vomitingeffect, but no anti-nausea effect in a placebo-controlled comparison with droperidol(Purhonen et al. 1997). The results in Study I were much the same: during the entire studyperiod tropisetron 5 mg prevented vomiting more effectively but not nausea, comparedwith droperidol 1.25 mg.

Altogether, according to the results of Studies II and III, the efficacy and side effectprofile of 5-HT3 receptor antagonists appear to be very similar. The choice of the 5-HT3

receptor antagonist used for clinical practice can be made based on economical reasons.In general, the dose for prophylaxis should be timed at the end of surgery. (Table 7)

The prophylactic dose of ondansetron varied in Studies II–V. In Study II and IVondansetron 8 mg IV was investigated, corresponding to ondansetron 16 mg given orallyin Study III, while ondansetron 4 mg was studied in Study V. The 8 mg dose IV and 16 mgorally were based on the meta-analysis by Tramèr et al. (1997), in which the anti-nauseaeffect of ondansetron 4 mg was weaker compared with 8 mg. For the large multicentretrial (Study V), ondansetron 4 mg was chosen, because in another systematic review(Figueredo and Canosa 1998) 4 mg was regarded as the optimal dose. The relative efficacy

60

of the 4 mg and 8 mg dose of ondansetron does not differ in Studies II-V; the use of thesmaller dose is thus suggested for prophylaxis.

Droperidol

The results of Study I show a similar overall anti-nausea effect after droperidol 1.25 mgcompared with tropisetron 5 mg for prophylaxis. The incidence of vomiting during theentire study period was higher after droperidol than after tropisetron. In a systematic re-view, droperidol was shown to have a more evident anti-nausea effect than an anti-vom-iting effect (Henzi et al. 2000). Moreover, the anti-nausea effect of droperidol was moreprofound during the early phase of recovery (0–6 h) than during the later course of recov-ery. The short-term antiemetic efficacy of droperidol supports the idea of repeating thelow prophylactic dose particularly when opioids are used for postoperative pain medica-tion (Tramèr and Walder 1999; Tramèr 2001). In the systematic review droperidol appearedto be the only antiemetic effective in preventing nausea and vomiting caused by opioidsadministered with PCA (Tramèr and Walder 1999). (Table 7)

The 1.25 mg dose of droperidol chosen for Study I corresponds to a dose of 20 µg/kg fora 60–kg patient. According to the systematic review of Henzi et al. (2000), the antiemeticeffect of droperidol has no dose-responsiveness, whereas the side effects of the agent, i.e.sedation and drowsiness, are dose-dependent. In the light of these results, a lower doseshould have been chosen for Study I, although the side effect profiles of the two groupsdid not differ.

Metoclopramide

Metoclopramide 10 mg was chosen for the comparison of oral ondansetron with tropisetronin Study III. The overall incidence of PONV was 75% after oral metoclopramide prophylaxisin patients undergoing thyroid or parathyroid surgery. As the mean individual risk forPONV was 68% (±12) in the metoclopramide group, the antiemetic efficacy appears to benegligible. This finding is in line with a systematic review in which metoclopramide 10 mgwas shown to have no significant anti-nausea effect, and the NNT for early and late vom-iting was 9.1 (95% CI 5.5–27) and 10 (6–41), respectively (Henzi et al. 1999). Unfortu-nately, Study III was started in 1998 when the above-mentioned systematic review was notavailable. In that review no clinically relevant adverse effects were associated with the 10mg dose of metoclopramide. It is possible that higher doses of metoclopramide have amore distinct antiemetic effect. In the light of the data in the systematic review, the doseof metoclopramide should have been higher in Study III.

5-HT3 receptor antagonist, dexamethasone and droperidol alone andin combinations

In Study V, during the early phase of recovery, the prophylaxis with droperidol was associ-ated with the strongest reduction in PONV compared with ondansetron and dexametha-sone. During the later course of recovery, the prophylaxis with dexamethasone was associ-ated with the strongest reduction of PONV. When the entire study period was considered,the antiemetic prophylaxis with ondansetron, dexamethasone and droperidol was associ-ated with an equal reduction of PONV. All the double combinations of the antiemeticswere equal in all phases of the recovery in preventing PONV. The triple combination re-sulted in an overall incidence of 22% in PONV.

61

All the double combinations appear to prevent PONV equally effectively. However,according to the results of Study V, all the antiemetics studied appear to act as indepen-dent drugs. Consequently, the studied antiemetics seem to have an additive effect withoutsynergism. Any of the combinations may thus be used for the prophylaxis. As the onsettime of dexamethasone seems to be around two hours, as suggested by Wang et al.(2000), it should be given for prophylaxis rather than as rescue medication. The combina-tion of droperidol with dexamethasone is justified, as droperidol is effective during theearly hours, and dexamethasone in the later course of recovery. On the other hand, thecombination of ondansetron and droperidol is warranted, as ondansetron has a pronouncedanti-vomiting effect, and droperidol has an evident anti-nausea effect. The combination ofondansetron and dexamethasone is another rational regimen (Henzi et al. 2000). Thetriple combination of antiemetics is justified for patients at high individual risk for PONV,and for patients for whom vomiting is hazardous. (Table 9)

62

Table 9. Clinical aspects in selecting antiemetics for the prevention and treatment of PONV. +++= excellent, ++=good, +=moderate, ±=pos-sible -=not recommended. (PONV=postoperative nausea and vomiting, EPS=extrapyramidal symptoms).

Optimal timeof Effect on Effective for

administrationAgent Prevention Treatment for prevention Special property Side effects

Early LateEarly Late phase of phase of Nausea Vomiting

recovery recovery

Effective for theprevention of opioid Sedation, EPS,

Droperidol +++ +++ – +++ +++ ± +++ + induced nausea QTc intervaland vomiting prolongation

More effective forfemales than

for males

Headache,5-HT3 receptor dizziness,antagonists +++ +++ – +++ ++ ++ + +++ liver enzyme

elevation,arrhythmias

AdrenalPain relief, suppression,

earlier recovery of hyperglycaemia,Glucocorticoids +++ – +++ – ± +++ ++ ++ anaesthesia, wound healing

stimulating action prolongation,disturbance ofdiurnal rhythm

Metoclopramide – ± – – ± ± ± ± Sedation,EPS

63

ORAL PREVENTION OF POSTOPERATIVE NAUSEA AND VOMITING

The antiemetic activity of any of the studied drugs in Study III was far from satisfying in theprevention of PONV after thyroid or parathyroid surgery. Oral prophylaxis for PONV hasnot been studied extensively. Indeed, it is an attractive idea to carry out the antiemeticprophylaxis using a regimen of oral premedication targeted at patients who are at high riskfor PONV. A practice like this would reduce the risk of ignoring the antiemetic prophylaxis.Considered economically, particularly 5-HT3 receptor antagonists are relatively costly, andoral drugs are as a rule cheaper compared with IV formulations. The results of the fewstudies on oral prophylaxis are nevertheless not very convincing.

The optimal timing of the prophylactic dose of ondansetron has been shown to benear the end of surgery rather than after the induction of anaesthesia (Sun et al. 1997;Tang et al. 1998). There are no published studies comparing the timing of the prophylacticdose for tropisetron. All in all, considering the optimal timing of the 5-HT3 receptor an-tagonists, oral premedication with them as traditional formulations is not a rational prac-tice. For ondansetron, an oral disintegrating tablet (ODT) with a freeze-dried formulationis available. Ondansetron ODT does not require water to be swallowed as it disintegratesin the mouth. Therefore it could be given to a patient right after anaesthesia or even forthe treatment of established PONV. There are only two published studies comparingondansetron ODT 8 mg with placebo. In the first study ondansetron ODT 8 mg was givenprior to discharge from the ambulatory surgery unit, but all the patients had received a 4mg IV dose at the induction of anaesthesia (Gan et al. 2002). Thus the study is not a truecomparison of ondansetron ODT with placebo. Ondansetron ODT 8 mg given twice a dayfor three days, starting on the first evening after operation, did not control PONV betterthan placebo among patients undergoing outpatient laparoscopic surgery (Thagaard et al.2003). Altogether, oral 5-HT3 receptor antagonists do not appear very promising for PONVprophylaxis.

The optimal timing of the prophylactic dose of dexamethasone was stated to be withinthe induction of anaesthesia rather than at the end of surgery (Wang et al. 2000). Consid-ering the two-hour onset time of the drug activity, the authors suggested, that the oraldosing of dexamethasone as premedication may be an alternative for IV administrationbefore the induction of anaesthesia.

For droperidol, only the IV formulation is available. Another butyrophenone, namelyhaloperidol, is manufactured in oral preparations. Although shown to possess antiemeticactivity (Kovac 2000), haloperidol has not been investigated nearly as extensively asdroperidol for the prophylaxis and treatment of PONV. If in future, studies on QTc prolon-gation associated with droperidol lead to the withdrawal of the drug, haloperidol may bea true alternative for perioperative use as an antiemetic, although QTc prolongation isassociated to some extent with haloperidol as well.

PROPOFOL AND VOLATILE ANAESTHETICS FOR MAINTENANCEOF ANAESTHESIA

The results of Study IV show a higher incidence of PONV in the PACU after sevofluraneanaesthesia than after propofol anaesthesia. Moreover, the difference in the incidences ofPONV after sevoflurane and propofol anaesthesia during the later phase of recovery wasnot distinct. When ondansetron prophylaxis was combined with sevoflurane anaesthesia,the incidence of PONV in the PACU was much the same as with propofol anaesthesia, andsmaller compared with propofol or sevoflurane anaesthesia during hours 2–24 after sur-

64

gery. Accordingly, in a recent study volatile anaesthetics appeared to be the leading causeof the PONV occurring in the early phase of recovery (Apfel et al. 2002). In this study thepro-emetic property of volatile anaesthetics exceeds all the other factors, including theindividual risk factors and antiemetics.

Propofol has been claimed to possess direct antiemetic property (Borgeat et al. 1992;Gan et al. 1997). When used only for induction of anaesthesia, propofol has no effect onthe incidence of PONV (Tramèr et al. 1997). In this meta-analysis, propofol was shown tolead to a smaller incidence of PONV compared with volatile anaesthetics, when used notonly for induction but also for the maintenance of anaesthesia. According to the meta-analysis, the favourable effect of propofol on PONV is short-term (Tramèr et al. 1997).When the studies with high control incidence for PONV (20–60%) were analysed, the NNTwas 4.7 (early, 0–6 h) and 4.9 (late, 6–48 h). All the studies included in the analysis, NNTfor any emetic event was 6.2 (early) and greater than 9 (late). According to the study ofApfel et al. (2002) the lower incidence of PONV after propofol anaesthesia in the earlyphase is associated with the omitting of volatile anaesthetics rather than with the anti-emetic property of propofol. Altogether, patients with a high individual risk for PONV and/or patients undergoing surgery associated with a high incidence of PONV, benefit frompropofol anaesthesia.

65

CONCLUSIONS

The following conclusions can be drawn from the studies:

1. Ondansetron and tropisetron are comparable in the prevention of PONV. Com-bined with droperidol, ondansetron had a longer duration of action than tropisetron.Dosed orally with the premedication, tropisetron had better efficacy during theearly phase of recovery. The true importance of the differences can be questioned,as they were inconsistent. Consequently, either of the 5-HT3 receptor antagonistscan be chosen for antiemetic prophylaxis.

2. The incidence of PONV is similar after propofol anaesthesia and after sevofluraneanaesthesia with ondansetron prophylaxis, while sevoflurane anaesthesia withoutantiemetic prophylaxis results in a significantly higher incidence of PONV. When apatient at a moderate (20–40%) individual risk of PONV is anaesthetized, eithervolatile anaesthetic with antiemetic prophylaxis, or propofol, can be chosen formaintenance. For high risk (≥40%) patients propofol anaesthesia with multimodalprophylaxis is recommended.

3. The anti-nausea effect of tropisetron is similar to that of with droperidol. The lateanti-vomiting effect of tropisetron is stronger than that of droperidol. The anti-emetic property during the initial phase of recovery (0–2 h) among the threeantiemetics appears to be, in sequence from the strongest to the weakest, droperidol-ondansetron-dexamethasone. In the later course of recovery (2–24 h) the antiemeticeffect is of the order, dexamethasone-ondansetron-droperidol. The entire 24-h studyperiod considered, the antiemetic effect of all three given as single prophylaxis iscomparable, lowering the incidence of PONV by approximately 10–15%. When theoverall outcome is examined, all the double combinations of the drugs are equallyeffective, decreasing further the incidence of PONV by 10%. The triple combina-tion of the antiemetics reduces the incidence of PONV additionally by 5%.

4. Ondansetron, dexamethasone and droperidol, given in double or triple combina-tions, act with an additive effect without synergism.

5. Increased incidence of either postoperative nausea or vomiting distressed the pa-tients and was associated with decreased patient satisfaction.

66

CLINICAL CONSIDERATIONS

1. For patients at a moderate risk (20–40%) for PONV, a standard anaesthesia regi-men may be chosen with single or double antiemetic prophylaxis with droperidol,dexamethasone or 5-HT3 receptor antagonists. For patients at high risk (>40%) forPONV, propofol anaesthesia is recommended with double or triple antiemetic pro-phylaxis with droperidol, dexamethasone or 5-HT3 receptor antagonists. (Figure 2)

2. Any of the three antiemetics, i.e. droperidol, dexamethasone, and 5-HT3 receptorantagonist, may be selected for single prophylaxis. Droperidol may be chosen ifopioids are used to treat postoperative pain. When droperidol is chosen for prophy-laxis, the repetition of the dose is recommended e.g. after 12 hours. Dexametha-sone is suitable for prophylaxis, as the onset time of the antiemetic action of thedrug is rather long. (Table 9)

3. Any double antiemetic combination with droperidol, dexamethasone and 5-HT3

receptor antagonist may be chosen for prophylaxis for patients at moderate or highrisk for PONV.

4. The selection of a 5-HT3 receptor antagonist used for prophylaxis can be madebased on economical considerations.

5. The doses of antiemetics used for prophylaxis should be the smallest effective onesin order to avoid possible side effects, which may be a problem particularly withdroperidol. The recommended dose of droperidol ranges from 0.5–0.75 mg IV andshould be administered at the end of surgery. The optimal dose of dexamethasoneranges from 2.5–5 mg IV, and the prophylactic dose should be timed at the induc-tion of anaesthesia. The optimal dose of ondansetron is 4 mg IV, and the dose isoptimally timed at the end of surgery.

6. Multimodal pain medication with NSAIDs, paracetamol and local anaesthetics isrecommended to reduce the need for postoperative opioids.

67

ACKNOWLEDGEMENTS

I wish to express my most sincere gratitude to all the following:Professor Kari Korttila, MD, PhD, FRCA, Department of Anaesthesia and Intensive Care

Medicine, University of Helsinki, for supervising this work and providing the facilities for carryingout Study V, giving me the opportunity to finish this thesis, and for his support in tight spots.Professor Seppo Alahuhta, MD, PhD, University of Oulu, the other supervisor of this work, forguiding Studies I–IV, and for his encouragement and constructive criticism of my thesis.

Professor Per Rosenberg, MD, PhD, Department of Anaesthesia and Intensive Care Medi-cine, University of Helsinki, for his supportive attitude to my work and for his most valuablecomments on the thesis.

The reviewers of the thesis, Docent Timo Ali-Melkkilä, MD, PhD, for his invaluable sugges-tions on the thesis and for focusing my attention on an essential detail in the work in a mostfriendly way, and Docent Pekka Tarkkila MD, PhD, for his constructive criticism of the thesis.

Docent Jukka Valanne, MD, PhD for suggesting to me the original idea of working on adoctoral thesis, for allowing me to carry out Studies I-IV in Lapland Central Hospital, and forhis kind support during the entire work. My co-workers Merja Koivuranta, MD, PhD for sug-gesting that I investigate the prevention of postoperative nausea and vomiting, Docent TuulaKangas-Saarela, MD, PhD, for her support during our co-operation, Docent Tero Ala-Kokko,MD, PhD, Sinikka Purhonen, MD, and Pirjo Ranta, MD, PhD for their encouraging co-opera-tion during the work. Aila Vuohelainen, MD, for her supportive attitude towards my workduring Study IV. Eija Ruoppa, RN, and Raili Valanne, RN, besides being my precious friends, formaking it easier for me to start the research in a peripheral hospital. Outi Törmänen, RN, forher active contribution to Study IV. Risto Bloigu, biostatistician in the Medical Informatics Groupof the Medical Faculty of Oulu University and Antti Nevanlinna, chief analyst in the IT Depart-ment of Helsinki University, for their readiness to offer continuous consultation help withproblems in statistics and SPSS.

Christian Apfel, MD, for the fruitful co-operation with Study V, Markus Kredel, MD, forbeing always helpful with the database of Study V. Anne Soikkeli, MD, Annika Åhlström, MD,Emilita Castejon, RN, and Riitta Drachman-Mertsalmi, RN, for their enthusiastic co-operationwith Study V.

Terttu Kaustia, MA, for linguistic editing of the thesis and of Studies III and V in a skilfuland prompt fashion.

Professor Arvi Yli-Hankala, MD, PhD, for being an exemplary scientist, and also for leavingme a peaceful office equipped with advanced hardware to work with. Elina Seitsonen, MD, forthe endless conversations about statistics as well as about life in general, and especially for herfriendship. Jouni Ahonen, MD, PhD for thought-provoking and interesting discussions on re-search and other topics. Anne Vakkuri, MD, PhD for always being there with an answer to allkinds of questions. Anna-Maija Korhonen, MD, for her friendship, co-operation, and admira-bly determined work with her own thesis. Päivi Laurila, MD, PhD, for sharing her recent expe-riences of preparing a doctoral thesis. Mika Jokinen, MD, PhD, for enriching discussions on theroad towards the same goal.

My colleagues in the anaesthesia departments in Lapland Central Hospital and Women’sHospital, HUCH, for flexibility and understanding the difficulties in scientific work. My col-leagues at the surgical and gynaecological departments in Lapland Central Hospital and at theDepartment of Gynaecology in Women’s Hospital, HUCH for the pleasant co-operation. Thestaff of the gynaecological ward in Women’s Hospital, HUCH for taking excellent care of allour patients and for the smooth co-operation. I want to express my appreciation to HelenaVänttinen, RN, and the entire staff of the operating theatre of Women’s Hospital, HUCH, forcreating an exceptionally inspiring work atmosphere.

Without my dear friends my life during the course of this work would have been muchduller. I thank my sister Pirkko M. Jokela, MSc, her husband Juha Kiljunen, MSc (Econ), mycousin Jarmo Kökkö, MA, her wife Riitta Korkeakivi, BA, my friends Hilkka Kyllönen, MSc,

68

Olli-Heikki Kyllönen, MSc, Hannu Paajanen, MD, Kaija Puhakka, MD, PhD, Seppo Suomela,MD, Katri Vuopala, MD, PhD, and Jorma Leskinen, MSc, for the countless stimulatingconversations over unforgettable dinners and splendid wines. I express my gratitude to myfriends from the North, Docent Aino-Liisa Oukka, MD, PhD, and Teuvo Määttä, MD, forsharing all the ups and downs and for helping me to keep the priorities right during thehectic working. I am grateful to my close relative Pirkko A. Jokela for her never-failingsupport and life-time friendship. I am indebted to my long-time friends Sirpa Koski-Tervo,MD, and Raija Valtonen, MD, for bringing cheer into my life and for sharing the fascinat-ing tradition of Sodankylä Midnight Sun Film Festivals. I owe my deep-felt thanks to myfriend Katri Selander, MD, PhD, for keeping me well informed about the latest achieve-ments of modern science on bones, and her husband Kevin Harris, MD, PhD, for editingthe English language of the manuscript of Study IV. I also thank Hannu Paajanen andJorma Leskinen for responding on-line to my endless questions concerning computers,particularly at the start of the work. I am grateful to Seppo Suomela also for the originalidea for the cover of this thesis.

I am deeply indebted to Docent Timo Raatikainen, MD, PhD, for his endless efforts tokeep me in one piece. Without his fantastic skill and patience, my struggle with my recur-rent fractures during the past few years would have been desperate.

I owe my loving gratitude to my parents, Mirja and Antti Jokela, my sister Liisa Jokelaand brother Jussi Jokela, and their families, for their continuous understanding and never-failing support. I am grateful to my dear godchild Aapo, 3 years, for just being there andbringing me moments of joy during the intensive work.

The financial support of Biomedicum Helsinki Foundation, Finnish Medical Founda-tion, HUS-EVO Committee, Maud Kuistila Foundation, Paulo Foundation and ResearchFoundation of Orion Corporation is gratefully appreciated.

Helsinki, September 2003

69

REFERENCES

Aasbø V, Ræder JC and Grøgaard B (1998). Betamethasone reduces postoperative pain and nausea afterambulatory surgery. Anesth Analg 87: 319–23.

Ahmed AB, Hobbs GJ and Curran JP (2000). Randomized, placebo-controlled trial of combination antiemeticprophylaxis for day-case gynaecological laparoscopic surgery. Br J Anaesth 85: 678–82.

Alon E, Buchser E, Herrera E et al. (1998). Tropisetron for treating established postoperative nausea andvomiting: a randomized, double-blind, placebo-controlled study. Anesth Analg 86: 617–23.

Alon E, Kocian R, Nett PC et al. (1996). Tropisetron for the prevention of postoperative nausea andvomiting in women undergoing gynecologic surgery. Anesth Analg 82: 338–41.

Andersen R and Krohg K (1976). Pain as a major cause of postoperative nausea. Can Anaesth Soc J 23:366–9.

Andrews PL (1992). Physiology of nausea and vomiting. Br J Anaesth 69: 2S–19S.Andrews PL, Davis CJ, Bingham S et al. (1990). The abdominal visceral innervation and the emetic reflex:

pathways, pharmacology, and plasticity. Can J Physiol Pharmacol 68: 325–45.Apfel CC, Greim CA, Haubitz I et al. (1998). A risk score to predict the probability of postoperative

vomiting in adults. Acta Anaesthesiol Scand 42: 495–501.Apfel CC, Greim CA, Haubitz I et al. (1998). The discriminating power of a risk score for postoperative

vomiting in adults undergoing various types of surgery. Acta Anaesthesiol Scand 42: 502–9.Apfel CC, Kranke P, Eberhart LH et al. (2002). Comparison of predictive models for postoperative nausea

and vomiting. Br J Anaesth 88: 234–40.Apfel CC, Kranke P, Katz MH et al. (2002). Volatile anaesthetics may be the main cause of early but not

delayed postoperative vomiting: a randomized controlled trial of factorial design. Br J Anaesth 88:659–68.

Apfel CC, Läärä E, Koivuranta M et al. (1999). A simplified risk score for predicting postoperative nauseaand vomiting: conclusions from cross-validations between two centers. Anesthesiology 91: 693–700.

Apfel CC, Roewer N and Korttila K (2002). How to study postoperative nausea and vomiting. ActaAnaesthesiol Scand 46: 921–8.

Appadu BL, Strange PG and Lambert DG (1994). Does propofol interact with D2 dopamine receptors?Anesth Analg 79: 1191–2.

Arcioni R, della Rocca M, Romano S et al. (2002). Ondansetron inhibits the analgesic effects of tramadol:a possible 5-HT(3) spinal receptor involvement in acute pain in humans. Anesth Analg 94: 1553–7,table of contents.

Badrinath S, Avramov MN, Shadrick M et al. (2000). The use of a ketamine-propofol combination duringmonitored anesthesia care. Anesth Analg 90: 858–62.

Baguley WA, Hay WT, Mackie KP et al. (1997). Cardiac dysrhythmias associated with the intravenousadministration of ondansetron and metoclopramide. Anesth Analg 84: 1380–1.

Balfour JA and Goa KL (1997). Dolasetron. A review of its pharmacology and therapeutic potential in themanagement of nausea and vomiting induced by chemotherapy, radiotherapy or surgery. Drugs 54:273–98.

Bellville JW, Bross IDJ and Howland WS (1960). Postoperative nausea and vomiting IV: factors related topostoperative nausea and vomiting. Anesthesiology 21: 186–193.

Ben-David B, Weber S and Chernus S (2002). Droperidol “box warning” warrants scrutiny. Anesthesiology97: 288.

Best N and Traugott F (1991). Comparative evaluation of propofol or methohexitone as the sole anaestheticagent for microlaryngeal surgery. Anaesth Intensive Care 19: 50–6.

Blumfield J (1899). The prevention of sickness after anaesthetics. Lancet 2: 833–835.Boogaerts JG, Vanacker E, Seidel L et al. (2000). Assessment of postoperative nausea using a visual

analogue scale. Acta Anaesthesiol Scand 44: 470–4.

70

Borgeat A, Wilder-Smith OH, Saiah M et al. (1992). Subhypnotic doses of propofol possess direct antiemeticproperties. Anesth Analg 74: 539–41.

Borison HL (1989). Area postrema: chemoreceptor circumventricular organ of the medulla oblongata.Prog Neurobiol 32: 351–90.

Borison HL and McCarthy LE (1983). Neuropharmacology of chemotherapy-induced emesis. Drugs 25Suppl 1: 8–17.

Borison HL and Wang SC (1953). Physiology and pharmacology of vomiting. Pharm Rev 5: 192–230.Bosek V, Hu P and Robinson LA (2000). Acute myocardial ischemia after administration of ondansetron

hydrochloride. Anesthesiology 92: 885–7.Capouet V, De Pauw C, Vernet B et al. (1996). Single dose i.v. tropisetron in the prevention of postoperative

nausea and vomiting after gynaecological surgery. Br J Anaesth 76: 54–60.Carroll NV, Miederhoff P, Cox FM et al. (1995). Postoperative nausea and vomiting after discharge from

outpatient surgery centers. Anesth Analg 80: 903–9.Carroll NV, Miederhoff PA, Cox FM et al. (1994). Costs incurred by outpatient surgical centers in managing

postoperative nausea and vomiting. J Clin Anesth 6: 364–9.Cechetto DF, Diab T, Gibson CJ et al. (2001). The effects of propofol in the area postrema of rats. Anesth

Analg 92: 934–42.Chan MT, Chui PT, Ho WS et al. (1998). Single-dose tropisetron for preventing postoperative nausea and

vomiting after breast surgery. Anesth Analg 87: 931–5.Chen X, Tang J, White PF et al. (2001). The effect of timing of dolasetron administration on its efficacy as

a prophylactic antiemetic in the ambulatory setting. Anesth Analg 93: 906–11.Chhibber AK, Lustik SJ, Thakur R et al. (1999). Effects of anticholinergics on postoperative vomiting,

recovery, and hospital stay in children undergoing tonsillectomy with or without adenoidectomy.Anesthesiology 90: 697–700.

Chia YY, Kuo MC, Liu K et al. (2002). Does postoperative pain induce emesis? Clin J Pain 18: 317–23.Chimbira W and Sweeney BP (2000). The effect of smoking on postoperative nausea and vomiting.

Anaesthesia 55: 540–4.Cieslak GD, Watcha MF, Phillips MB et al. (1996). The dose-response relation and cost-effectiveness of

granisetron for the prophylaxis of pediatric postoperative emesis. Anesthesiology 85: 1076–85.Clarke RS (1984). Nausea and vomiting. Br J Anaesth 56: 19–27.Coloma M, Chiu JW, White PF et al. (2001). The use of esmolol as an alternative to remifentanil during

desflurane anesthesia for fast-track outpatient gynecologic laparoscopic surgery. Anesth Analg 92:352–7.

Coloma M, Duffy LL, White PF et al. (2001). Dexamethasone facilitates discharge after outpatient anorectalsurgery. Anesth Analg 92: 85–8.

Coloma M, White PF, Ogunnaike BO et al. (2002). Comparison of acustimulation and ondansetron for thetreatment of established postoperative nausea and vomiting. Anesthesiology 97: 1387–92.

de Bruijn KM (1992). Tropisetron. A review of the clinical experience. Drugs 43 Suppl 3: 11–22.De Witte JL, Schoenmaekers B, Sessler DI et al. (2001). The analgesic efficacy of tramadol is impaired by

concurrent administration of ondansetron. Anesth Analg 92: 1319–21.Dershwitz M (2002). Droperidol: should the black box be light gray? J Clin Anesth 14: 598–603.Dershwitz M, Michalowski P, Chang Y et al. (2002). Postoperative nausea and vomiting after total

intravenous anesthesia with propofol and remifentanil or alfentanil: how important is the opioid?J Clin Anesth 14: 275–8.

Diemunsch P, Korttila K, Leeser J et al. (1998). Oral dolasetron mesylate for prevention of postoperativenausea and vomiting: a multicenter, double-blind, placebo-controlled study. The Oral DolasetronPONV Prevention Study Group. J Clin Anesth 10: 145–52.

Diemunsch P, Leeser J, Feiss P et al. (1997). Intravenous dolasetron mesilate ameliorates postoperativenausea and vomiting. Can J Anaesth 44: 173–81.

Diemunsch P, Schoeffler P, Bryssine B et al. (1999). Antiemetic activity of the NK1 receptor antagonistGR205171 in the treatment of established postoperative nausea and vomiting after majorgynaecological surgery. Br J Anaesth 82: 274–6.

71

Divatia JV, Vaidya JS, Badwe RA et al. (1996). Omission of nitrous oxide during anesthesia reduces theincidence of postoperative nausea and vomiting. A meta-analysis. Anesthesiology 85: 1055–62.

Domino KB, Anderson EA, Polissar NL et al. (1999). Comparative efficacy and safety of ondansetron,droperidol, and metoclopramide for preventing postoperative nausea and vomiting: a meta-analysis.Anesth Analg 88: 1370–9.

Dundee JW, Kirwan MJ and Clarke RS (1965). Anaesthesia and premedication as factors in postoperativevomiting. Acta Anaesthesiol Scand 9: 223–31.

Dundee JW, Loan WB and Morrison JD (1975). A comparison of the efficacy of cyclizine and perhenazinein reducing the emetic effects of morphine and pethidine. Br J Clin Pharmacol 2: 81–5.

Dundee JW, Love WJ, M. NR et al. (1965). Studies of drugs given before anaesthesia. VI: The phenotiazinederivatives. Br J Anaesth 37: 332–52.

Dundee JW, Nicholl RM, Clarke RS et al. (1965). Studies of drugs given before anaesthesia. VII: Pethidine-phenothiazine combinations. Br J Anaesth 37: 601–12.

Eberhart LH, Hogel J, Seeling W et al. (2000). Evaluation of three risk scores to predict postoperativenausea and vomiting. Acta Anaesthesiol Scand 44: 480–8.

Eberhart LH, Mauch M, Morin AM et al. (2002). Impact of a multimodal anti-emetic prophylaxis onpatient satisfaction in high-risk patients for postoperative nausea and vomiting. Anaesthesia 57:1022–7.

Eberhart LH, Morin AM, Wulf H et al. (2002). Patient preferences for immediate postoperative recovery.Br J Anaesth 89: 760–1.

Ebert TJ, Robinson BJ, Uhrich TD et al. (1998). Recovery from sevoflurane anesthesia: a comparison toisoflurane and propofol anesthesia. Anesthesiology 89: 1524–31.

Eger IE and Saidman LJ (1965). Hazards of nitrous oxide anesthesia in bowel obstruction and pneumothorax.Anesthesiology 26: 61–6.

Eriksson H, Haasio J and Korttila K (1995). Comparison of eltanolone and thiopental in anaesthesia fortermination of pregnancy. Acta Anaesthesiol Scand 39: 479–84.

Eriksson H and Korttila K (1996). Recovery profile after desflurane with or without ondansetron comparedwith propofol in patients undergoing outpatient gynecological laparoscopy. Anesth Analg 82:533–8.

Eriksson H, Tenhunen A and Korttila K (1996). Balanced analgesia improves recovery and outcome afteroutpatient tubal ligation. Acta Anaesthesiol Scand 40: 151–5.

Ewalenko P, Janny S, Dejonckheere M et al. (1996). Antiemetic effect of subhypnotic doses of propofolafter thyroidectomy. Br J Anaesth 77: 463–7.

Ezri T, Lurie S, Stein A et al. (2002). Postoperative nausea and vomiting: comparison of the effect ofpostoperative meperidine or morphine in gynecologic surgery patients. J Clin Anesth 14: 262–6.

Figueredo ED and Canosa LG (1998). Ondansetron in the prophylaxis of postoperative vomiting: a meta-analysis. J Clin Anesth 10: 211–21.

Finck AD, Samaniego E and Ngai SH (1995). Nitrous oxide selectively releases Met5-enkephalinand Met5-enkephalin-Arg6-Phe7 into canine third ventricular cerebrospinal fluid. Anesth Analg 80:664–70.

Fisher DM (1999). Surrogate outcomes: meaningful not! Anesthesiology 90: 355–6.Fredman B, Nathanson MH, Smith I et al. (1995). Sevoflurane for outpatient anesthesia: a comparison

with propofol. Anesth Analg 81: 823–8.Fredrikson M, Hursti T, Furst CJ et al. (1992). Nausea in cancer chemotherapy is inversely related to

urinary cortisol excretion. Br J Cancer 65: 779–80.Frighetto L, Loewen PS, Dolman J et al. (1999). Cost-effectiveness of prophylactic dolasetron or droperidol

vs rescue therapy in the prevention of PONV in ambulatory gynecologic surgery. Can J Anaesth 46:536–43.

Fujii Y, Toyooka H and Tanaka H (1998). Granisetron in the prevention of nausea and vomiting aftermiddle-ear surgery: a dose-ranging study. Br J Anaesth 80: 764–6.

Gan T, Sloan F, Dear Gde L et al. (2001). How much are patients willing to pay to avoid postoperativenausea and vomiting? Anesth Analg 92: 393–400.

72

Gan TJ, Franiak R and Reeves J (2002). Ondansetron orally disintegrating tablet versus placebo forthe prevention of postdischarge nausea and vomiting after ambulatory surgery. Anesth Analg 94:1199–200, table of contents.

Gan TJ, Glass PS, Howell ST et al. (1997). Determination of plasma concentrations of propofol associatedwith 50% reduction in postoperative nausea. Anesthesiology 87: 779–84.

Gan TJ, White PF, Scuderi PE et al. (2002). FDA “black box” warning regarding use of droperidol forpostoperative nausea and vomiting: is it justified? Anesthesiology 97: 287.

Gardner MJ and Altman DG (1989). Calculating confidence intervals for proportions and their differences.Statistics with confidence: confidence intervals and statistical guidelines. Gardner MJ and AltmanDG (Ed). London, BMJ: 29–30.

Gesztesi Z, Scuderi PE, White PF et al. (2000). Substance P (Neurokinin-1) antagonist prevents postoperativevomiting after abdominal hysterectomy procedures. Anesthesiology 93: 931–7.

Gilbertson AA (1999). Ethical review of research. Br J Anaesth 82: 6–7.Giuffre M and Gross JB (1986). The effects of nitrous oxide on postoperative bowel motility. Anesthesiology

65: 699–700.Gold BS, Kitz DS, Lecky JH et al. (1989). Unanticipated admission to the hospital following ambulatory

surgery. Jama 262: 3008–10.Goll V, Akca O, Greif R et al. (2001). Ondansetron is no more effective than supplemental intraoperative

oxygen for prevention of postoperative nausea and vomiting. Anesth Analg 92: 112–7.Gralla RJ, Itri LM, Pisko SE et al. (1981). Antiemetic efficacy of high-dose metoclopramide: randomized

trials with placebo and prochlorperazine in patients with chemotherapy-induced nausea and vomiting.N Engl J Med 305: 905–9.

Greif R, Laciny S, Rapf B et al. (1999). Supplemental oxygen reduces the incidence of postoperativenausea and vomiting. Anesthesiology 91: 1246–52.

Greif R and Sessler DI (2000). Supplemental oxygen for the prevention of nausea and vomiting.Anesthesiology 93: 585.

Grunberg SM (1997). Innovative approaches in the treatment of emesis. Support Care Cancer 5: 9–11.Hagemann E, Halvorsen A, Holgersen O et al. (2000). Intramuscular ephedrine reduces emesis during the

first three hours after abdominal hysterectomy. Acta Anaesthesiol Scand 44: 107–11.Hammas B, Hvarfner A, Thorn SE et al. (1998). Effects of propofol on ipecacuanha-induced nausea and

vomiting. Acta Anaesthesiol Scand 42: 447–51.Hammas B, Thorn SE and Wattwil M (2002). Superior prolonged antiemetic prophylaxis with a four-drug

multimodal regimen - comparison with propofol or placebo. Acta Anaesthesiol Scand 46: 232–7.Handa F and Fujii Y (2001). The efficacy of oral clonidine premedication in the prevention of postoperative

vomiting in children following strabismus surgery. Paediatr Anaesth 11: 71–4.Harrington RA, Hamilton CW, Brogden RN et al. (1983). Metoclopramide. An updated review of its

pharmacological properties and clinical use. Drugs 25: 451–94.Harris AL (1982). Cytotoxic-therapy-induced vomiting is mediated via enkephalin pathways. Lancet

1: 714–6.Hartung J (1996). Twenty-four of twenty-seven studies show a greater incidence of emesis associated

with nitrous oxide than with alternative anesthetics. Anesth Analg 83: 114–6.Hechler A, Neumann S, Jehmlich M et al. (2001). A small dose of droperidol decreases postoperative

nausea and vomiting in adults but cannot improve an already excellent patient satisfaction. ActaAnaesthesiol Scand 45: 501–6.

Hedayati B and Fear S (1999). Hospital admission after day-case gynaecological laparoscopy. Br J Anaesth83: 776–9.

Heffernan AM and Rowbotham DJ (2000). Postoperative nausea and vomiting – time for balancedantiemesis? Br J Anaesth 85: 675–7.

Henderson JM, Brodsky DA, Fisher DM et al. (1988). Pre-induction of anesthesia in pediatric patients withnasally administered sufentanil. Anesthesiology 68: 671–5.

Henzi I, Sonderegger J and Tramèr MR (2000). Efficacy, dose-response, and adverse effects of droperidolfor prevention of postoperative nausea and vomiting. Can J Anaesth 47: 537–51.

73

Henzi I, Walder B and Tramèr MR (1999). Metoclopramide in the prevention of postoperative nausea andvomiting: a quantitative systematic review of randomized, placebo-controlled studies. Br J Anaesth83: 761–71.

Henzi I, Walder B and Tramèr MR (2000). Dexamethasone for the prevention of postoperative nausea andvomiting: a quantitative systematic review. Anesth Analg 90: 186–94.

Hill RP, Lubarsky DA, Phillips-Bute B et al. (2000). Cost-effectiveness of prophylactic antiemetic therapywith ondansetron, droperidol, or placebo. Anesthesiology 92: 958–67.

Honkavaara P, Saarnivaara L and Klemola UM (1994). Prevention of nausea and vomiting with transdermalhyoscine in adults after middle ear surgery during general anaesthesia. Br J Anaesth 73: 763–6.

Horimoto Y, Tomie H, Hanzawa K et al. (1991). Scopolamine patch reduces postoperative emesis inpaediatric patients following strabismus surgery. Can J Anaesth 38: 441–4.

Howat DDC (1960). Anti-emetic drugs in anaesthesia. A double blind trial of two phenothiazine derivatives.Anaesthesia 15: 289–97.

Janknegt R, Pinckaers JW, Rohof MH et al. (1999). Double-blind comparative study of droperidol, granisetronand granisetron plus dexamethasone as prophylactic anti-emetic therapy in patients undergoingabdominal, gynaecological, breast or otolaryngological surgery. Anaesthesia 54: 1059–68.

Jenkins LC and Lahay D (1971). Central mechanisms of vomiting related to catecholamine response:anaesthetic implication. Can Anaesth Soc J 18: 434–41.

Jin F and Chung F (2001). Multimodal analgesia for postoperative pain control. J Clin Anesth 13: 524–39.Junger A, Klasen J, Benson M et al. (2001). Factors determining length of stay of surgical day-case

patients. Eur J Anaesthesiol 18: 314–21.Kallela H, Haasio J and Korttila K (1994). Comparison of eltanolone and propofol in anesthesia for

termination of pregnancy. Anesth Analg 79: 512–6.Kantor GS (2002). Arrhythmia risk of antiemetic agents. Anesthesiology 97: 286.Karlsen KL, Persson E, Wennberg E et al. (2000). Anaesthesia, recovery and postoperative nausea and

vomiting after breast surgery. A comparison between desflurane, sevoflurane and isofluraneanaesthesia. Acta Anaesthesiol Scand 44: 489–93.

Kazemi-Kjellberg F, Henzi I and Tramèr MR (2001). Treatment of established postoperative nausea andvomiting: a quantitative systematic review. BMC Anesthesiol 1: 2.

Koivuranta M, Läärä E, Snåre L et al. (1997). A survey of postoperative nausea and vomiting. Anaesthesia52: 443–9.

Korttila K, Clergue F, Leeser J et al. (1997). Intravenous dolasetron and ondansetron in prevention ofpostoperative nausea and vomiting: a multicenter, double-blind, placebo-controlled study. ActaAnaesthesiol Scand 41: 914–22.

Kovac A, Mingus M, Sung YF et al. (1999). Reduced resource utilization in patients treated for postoperativenausea and vomiting with dolasetron mesylate. MCPR44 Study Group. J Clin Anesth 11: 235–41.

Kovac AL (2000). Prevention and treatment of postoperative nausea and vomiting. Drugs 59: 213–43.Kovac AL, O’Connor TA, Pearman MH et al. (1999). Efficacy of repeat intravenous dosing of ondansetron

in controlling postoperative nausea and vomiting: a randomized, double-blind, placebo-controlledmulticenter trial. J Clin Anesth 11: 453–9.

Kovac AL, Scuderi PE, Boerner TF et al. (1997). Treatment of postoperative nausea and vomiting withsingle intravenous doses of dolasetron mesylate: a multicenter trial. Dolasetron Mesylate PONVTreatment Study Group. Anesth Analg 85: 546–52.

Kranke P, Apefel CC, Papenfuss T et al. (2001). An increased body mass index is no risk factor forpostoperative nausea and vomiting. A systematic review and results of original data. Acta AnaesthesiolScand 45: 160–6.

Kranke P, Apfel CC, Eberhart LH et al. (2001). The influence of a dominating centre on a quantitativesystematic review of granisetron for preventing postoperative nausea and vomiting. Acta AnaesthesiolScand 45: 659–70.

Kranke P, Morin AM, Roewer N et al. (2002). Dimenhydrinate for prophylaxis of postoperative nauseaand vomiting: a meta-analysis of randomized controlled trials. Acta Anaesthesiol Scand 46: 238–44.

74

Kranke P, Morin AM, Roewer N et al. (2002). The efficacy and safety of transdermal scopolamine for theprevention of postoperative nausea and vomiting: a quantitative systematic review. Anesth Analg95: 133–43, table of contents.

Kreisler NS, Spiekermann BF, Ascari CM et al. (2000). Small-dose droperidol effectively reduces nausea ina general surgical adult patient population. Anesth Analg 91: 1256–61.

Lee A and Done ML (1999). The use of nonpharmacologic techniques to prevent postoperative nauseaand vomiting: a meta-analysis. Anesth Analg 88: 1362–9.

Lee CR, Plosker GL and McTavish D (1993). Tropisetron. A review of its pharmacodynamic andpharmacokinetic properties, and therapeutic potential as an antiemetic. Drugs 46: 925–43.

Lehmann KA, Van Peer A, Ikonomakis M et al. (1988). Pharmacokinetics of droperidol in surgical patientsunder different conditions of anaesthesia. Br J Anaesth 61: 297–301.

Lerman J (1992). Surgical and patient factors involved in postoperative nausea and vomiting. Br J Anaesth69: 24S–32S.

Liu K, Hsu CC and Chia YY (1999). The effect of dose of dexamethasone for antiemesis after majorgynecological surgery. Anesth Analg 89: 1316–8.

Loeser EA, Bennett G, Stanley TH et al. (1979). Comparison of droperidol, haloperidol and prochlorperazineas postoperative anti-emetics. Can Anaesth Soc J 26: 125–7.

Lopez-Olaondo L, Carrascosa F, Pueyo FJ et al. (1996). Combination of ondansetron and dexamethasonein the prophylaxis of postoperative nausea and vomiting. Br J Anaesth 76: 835–40.

Løvstad RZ, Thagaard KS, Berner NS et al. (2001). Neostigmine 50 microg kg(-1) with glycopyrrolateincreases postoperative nausea in women after laparoscopic gynaecological surgery. Acta AnaesthesiolScand 45: 495–500.

Luginbühl M, Wüthrich S, Petersen-Felix S et al. (2003). Different benefit of bispectal index (BIS) indesflurane and propofol anesthesia. Acta Anaesthesiol Scand 47: 165–73.

Macario A, Weinger M, Carney S et al. (1999). Which clinical anesthesia outcomes are important toavoid? The perspective of patients. Anesth Analg 89: 652–8.

Markham A and Sorkin EM (1993). Ondansetron. An update of its therapeutic use in chemotherapy-induced and postoperative nausea and vomiting. Drugs 45: 931–52.

Maubach KA and Jones RS (1997). Electrophysiological characterisation of tachykinin receptors in the ratnucleus of the solitary tract and dorsal motor nucleus of the vagus in vitro. Br J Pharmacol 122:1151–9.

McCarthy BG and Peroutka SJ (1988). Differentiation of muscarinic cholinergic receptor subtypes in humancortex and pons: implications for anti-motion sickness therapy. Aviat Space Environ Med 59: 63–6.

McCormick C (2002). FDA responds, continues watch. APSF Newsletter 17: 17.McKenzie R, Riley TJ, Tantisira B et al. (1997). Effect of propofol for induction and ondansetron with or

without dexamethasone for the prevention of nausea and vomiting after major gynecologic surgery.J Clin Anesth 9: 15–20.

McKenzie R, Uy NT, Riley TJ et al. (1996). Droperidol/ondansetron combination controls nausea andvomiting after tubal banding. Anesth Analg 83: 1218–22.

McKenzie R, Wadhwa RK, Uy NT et al. (1981). Antiemetic effectiveness of intramuscular hydroxyzinecompared with intramuscular droperidol. Anesth Analg 60: 783–8.

McLean S, Ganong A, Seymour PA et al. (1996). Characterization of CP-122,721; a nonpeptide antagonistof the neurokinin NK1 receptor. J Pharmacol Exp Ther 277: 900–8.

McQuay H and Moore R (1998). Pain measurement, study design, and validity. An evidence-based resourcefor pain relief. McQuay H and Moore R (Ed). Oxford, Oxford University Press: 14–18.

Mikawa K, Nishina K, Maekawa N et al. (1995). Oral clonidine premedication reduces vomiting in childrenafter strabismus surgery. Can J Anaesth 42: 977–81.

Montgomery JE, Sutherland CJ, Kestin IG et al. (1996). Morphine consumption in patients receiving rectalparacetamol and diclofenac alone and in combination. Br J Anaesth 77: 445–7.

Murakawa M, Shingu K and Mori K (1994). Effects of nitrous oxide on the brain catecholamines in rats.Ann Acad Med Singapore 23: 452–5.

Myles PS, Hendrata M, Bennett AM et al. (1996). Postoperative nausea and vomiting. Propofol orthiopentone: does choice of induction agent affect outcome? Anaesth Intensive Care 24: 355–9.

75

Myles PS, Hunt JO, Nightingale CE et al. (1999). Development and psychometric testing of a quality ofrecovery score after general anesthesia and surgery in adults. Anesth Analg 88: 83–90.

Naidu-Sjösvärd K, Sjöberg F and Gupta A (1998). Anaesthesia for videoarthroscopy of the knee.A comparison between desflurane and sevoflurane. Acta Anaesthesiol Scand 42: 464–71.

Nakata K, Mammoto T, Kita T et al. (2002). Continuous epidural, not intravenous, droperidol inhibitspruritus, nausea, and vomiting during epidural morphine analgesia. J Clin Anesth 14: 121–5.

Nathanson MH, Fredman B, Smith I et al. (1995). Sevoflurane versus desflurane for outpatient anesthesia:a comparison of maintenance and recovery profiles. Anesth Analg 81: 1186–90.

Nelskylä KA, Yli-Hankala AM, Puro PH et al. (2001). Sevoflurane titration using bispectral index decreasespostoperative vomiting in phase II recovery after ambulatory surgery. Anesth Analg 93: 1165–9.

Nelson PS, Streisand JB, Mulder SM et al. (1989). Comparison of oral transmucosal fentanyl citrate andan oral solution of meperidine, diazepam, and atropine for premedication in children. Anesthesiology70: 616–21.

Newberry NR, Watkins CJ, Sprosen TS et al. (1993). BRL 46470 potently antagonizes neural responsesactivated by 5-HT3 receptors. Neuropharmacology 32: 729–35.

Newcombe RG and Altman DG (2000). Proportion and their differences. Statistics with confidence. AltmanDG, Machin D, Bryant TN et al (Ed). Bristol, BMJ Books: 48–50.

Oddby-Muhrbeck E, Eksborg S, Bergendahl HT et al. (2002). Effects of clonidine on postoperative nauseaand vomiting in breast cancer surgery. Anesthesiology 96: 1109–14.

Orkin F (1992). What do patients want? – Preferences for immediate postoperative recovery. AnesthAnalg 74: S225.

Ozkose Z, Yalcin Cok O, Tuncer B et al. (2002). Comparison of hemodynamics, recovery profile, and earlypostoperative pain control and costs of remifentanil versus alfentanil-based total intravenous anesthesia(TIVA). J Clin Anesth 14: 161–8.

Palazzo M and Evans R (1993). Logistic regression analysis of fixed patient factors for postoperativesickness: a model for risk assessment. Br J Anaesth 70: 135–40.

Palazzo MG and Strunin L (1984). Anaesthesia and emesis. I: Etiology. Can Anaesth Soc J 31: 178–87.Pandit SK and Kothary SP (1989). Intravenous narcotics for premedication in outpatient anaesthesia. Acta

Anaesthesiol Scand 33: 353–8.Park J, Forrest J, Kolesar R et al. (1996). Oral clonidine reduces postoperative PCA morphine requirements.

Can J Anaesth 43: 900–6.Philip BK, Kallar SK, Bogetz MS et al. (1996). A multicenter comparison of maintenance and recovery

with sevoflurane or isoflurane for adult ambulatory anesthesia. The Sevoflurane MulticenterAmbulatory Group. Anesth Analg 83: 314–9.

Pierre S, Benais H and Pouymayou J (2002). Apfel’s simplified score may favourably predict the risk ofpostoperative nausea and vomiting. Can J Anaesth 49: 237–42.

Plosker GL and Goa KL (1991). Granisetron. A review of its pharmacological properties and therapeuticuse as an antiemetic. Drugs 42: 805–24.

Pocock S (1983). The size of a clinical trial. Clinical trials: a practical approach. Pocock S (Ed). Chichester,John Wiley & Sons: 123–127.

Pueyo FJ, Carrascosa F, Lopez L et al. (1996). Combination of ondansetron and droperidol in the prophylaxisof postoperative nausea and vomiting. Anesth Analg 83: 117–22.

Purhonen S, Kauko M, Koski EM et al. (1997). Comparison of tropisetron, droperidol, and saline inthe prevention of postoperative nausea and vomiting after gynecologic surgery. Anesth Analg 84:662–7.

Purhonen S, Niskanen M, Wustefeld M et al. (2003). Supplemental oxygen for prevention of nausea andvomiting after breast surgery. Br J Anaesth 91: 284–287.

Purhonen S, Turunen M, Ruohoaho UM et al. (2003). Supplemental oxygen does not reduce the incidenceof postoperative nausea and vomiting after ambulatory gynecologic laparoscopy. Anesth Analg 96:91–6, table of contents.

Pusch F, Berger A, Wildling E et al. (2002). The effects of systolic arterial blood pressure variations onpostoperative nausea and vomiting. Anesth Analg 94: 1652–5, table of contents.

76

Ræder J, Gupta A and Pedersen FM (1997). Recovery characteristics of sevoflurane- or propofol-basedanaesthesia for day-care surgery. Acta Anaesthesiol Scand 41: 988–94.

Ræder JC (1994). Total intravenous anaesthesia – free from nitrous oxide, free from problems? ActaAnaesthesiol Scand 38: 769–70.

Ræder JC, Mjåland O, Aasbø V et al. (1998). Desflurane versus propofol maintenance for outpatientlaparoscopic cholecystectomy. Acta Anaesthesiol Scand 42: 106–10.

Rajeeva V, Bhardwaj N, Batra YK et al. (1999). Comparison of ondansetron with ondansetron anddexamethasone in prevention of PONV in diagnostic laparoscopy. Can J Anaesth 46: 40–4.

Reuben SS, Connelly NR and Maciolek H (1999). Postoperative analgesia with controlled-release oxycodonefor outpatient anterior cruciate ligament surgery. Anesth Analg 88: 1286–91.

Reuben SS, Steinberg RB, Maciolek H et al. (2002). Preoperative administration of controlled-releaseoxycodone for the management of pain after ambulatory laparoscopic tubal ligation surgery. J ClinAnesth 14: 223–7.

Rich WM, Abdulhayoglu G and DiSaia PJ (1980). Methylprednisolone as an antiemetic during cancerchemotherapy – a pilot study. Gynecol Oncol 9: 193–8.

Rose JB and Watcha MF (1999). Postoperative nausea and vomiting in paediatric patients. Br J Anaesth83: 104–17.

Rothenberg DM, Parnass SM, Litwack K et al. (1991). Efficacy of ephedrine in the prevention ofpostoperative nausea and vomiting. Anesth Analg 72: 58–61.

Rowbotham DJ, Peacock JE, Jones RM et al. (1998). Comparison of remifentanil in combination withisoflurane or propofol for short-stay surgical procedures. Br J Anaesth 80: 752–5.

Russell D and Kenny GN (1992). 5-HT3 antagonists in postoperative nausea and vomiting. Br J Anaesth69: 63S–68S.

Sagar SM (1991). The current role of anti-emetic drugs in oncology: a recent revolution in patient symptomcontrol. Cancer Treat Rev 18: 95–135.

Salmenperä M, Kuoppamäki R and Salmenperä A (1992). Do anticholinergic agents affect the occurrenceof postanaesthetic nausea? Acta Anaesthesiol Scand 36: 445–8.

Sanchez-Ledesma MJ, Lopez-Olaondo L, Pueyo FJ et al. (2002). A comparison of three antiemeticcombinations for the prevention of postoperative nausea and vomiting. Anesth Analg 95: 1590–5,table of contents.

Schimmer BP and Parker KL (1995). Adrenocorticotropic hormone; adrenocortical steroids and their syntheticanalogs; inhibitors of the synthesis and actions of adrenocortical hormones. Goodman & Gilman’sThe pharmacological basis of therapeutics. Hardman JG, Limbird LE, Molinoff PB et al (Ed). NewYork, McGraw-Hill: 1465–6.

Scuderi PE (2003). Droperidol: many questions, few answers. Anesthesiology 98: 289–90.Scuderi PE, James RL, Harris L et al. (1999). Antiemetic prophylaxis does not improve outcomes after

outpatient surgery when compared to symptomatic treatment. Anesthesiology 90: 360–71.Scuderi PE, James RL, Harris L et al. (2000). Multimodal antiemetic management prevents early postoperative

vomiting after outpatient laparoscopy. Anesth Analg 91: 1408–14.Sinclair DR, Chung F and Mezei G (1999). Can postoperative nausea and vomiting be predicted?

Anesthesiology 91: 109–18.Sneyd JR, Carr A, Byrom WD et al. (1998). A meta-analysis of nausea and vomiting following maintenance

of anaesthesia with propofol or inhalational agents. Eur J Anaesthesiol 15: 433–45.Song D, Chung F, Yogendran S et al. (2002). Evaluation of postural stability after low-dose droperidol

in outpatients undergoing gynaecological dilatation and curettage procedure. Br J Anaesth 88:819–23.

Sonner JM, Hynson JM, Clark O et al. (1997). Nausea and vomiting following thyroid and parathyroidsurgery. J Clin Anesth 9: 398–402.

St Pierre M, Dunkel M, Rutherford A et al. (2000). Does etomidate increase postoperative nausea?A double-blind controlled comparison of etomidate in lipid emulsion with propofol for balancedanaesthesia. Eur J Anaesthesiol 17: 634–41.

Stadler M, Bardiau F, Seidel L et al. (2003). Difference in risk factors for postoperative nausea and vomiting.Anesthesiology 98: 46–52.

77

Steinbrook RA, Freiberger D, Gosnell JL et al. (1996). Prophylactic antiemetics for laparoscopiccholecystectomy: ondansetron versus droperidol plus metoclopramide. Anesth Analg 83: 1081–3.

Steward DL, Welge JA and Myer CM (2001). Do steroids reduce morbidity of tonsillectomy? Meta-analysisof randomized trials. Laryngoscope 111: 1712–8.

Sun R, Klein KW and White PF (1997). The effect of timing of ondansetron administration in outpatientsundergoing otolaryngologic surgery. Anesth Analg 84: 331–6.

Sweeney BP (2002). Why does smoking protect against PONV? Br J Anaesth 89: 810–3.Tang J, Chen X, White PF et al. (2003). Antiemetic prophylaxis for office-based surgery: are the 5-HT3

receptor antagonists beneficial? Anesthesiology 98: 293–8.Tang J, Wang B, White PF et al. (1998). The effect of timing of ondansetron administration on its efficacy,

cost-effectiveness, and cost-benefit as a prophylactic antiemetic in the ambulatory setting. AnesthAnalg 86: 274–82.

Thagaard KS, Steine S and Ræder J (2003). Ondansetron disintegrating tablets of 8 mg twice a day for 3days did not reduce the incidence of nausea or vomiting after laparoscopic surgery. Eur J Anaesthesiol20: 153–7.

Thomas R and Jones N (2001). Prospective randomized, double-blind comparative study of dexamethasone,ondansetron, and ondansetron plus dexamethasone as prophylactic antiemetic therapy in patientsundergoing day-case gynaecological surgery. Br J Anaesth 87: 588–92.

Thomas R, Jones NA and Strike P (2002). The value of risks scores for predicting postoperative nausea andvomiting when used to compare patient group in a randomised controlled trial. Anaesthesia 57:1119–28.

Toner CC, Broomhead CJ, Littlejohn IH et al. (1996). Prediction of postoperative nausea and vomitingusing a logistic regression model. Br J Anaesth 76: 347–51.

Tramèr M, Moore A and McQuay H (1996). Omitting nitrous oxide in general anaesthesia: meta-analysisof intraoperative awareness and postoperative emesis in randomized controlled trials. Br J Anaesth76: 186–93.

Tramèr M, Moore A and McQuay H (1997). Propofol anaesthesia and postoperative nausea and vomiting:quantitative systematic review of randomized controlled studies. Br J Anaesth 78: 247–55.

Tramèr MR (2001). A rational approach to the control of postoperative nausea and vomiting: evidencefrom systematic reviews. Part I. Efficacy and harm of antiemetic interventions, and methodologicalissues. Acta Anaesthesiol Scand 45: 4–13.

Tramèr MR (2001). A rational approach to the control of postoperative nausea and vomiting: evidencefrom systematic reviews. Part II. Recommendations for prevention and treatment, and research agenda.Acta Anaesthesiol Scand 45: 14–9.

Tramèr MR and Fuchs-Buder T (1999). Omitting antagonism of neuromuscular block: effect on postoperativenausea and vomiting and risk of residual paralysis. A systematic review. Br J Anaesth 82: 379–86.

Tramèr MR, Phillips C, Reynolds DJ et al. (1999). Cost-effectiveness of ondansetron for postoperativenausea and vomiting. Anaesthesia 54: 226–34.

Tramèr MR, Reynolds DJ, Moore RA et al. (1997). Efficacy, dose-response, and safety of ondansetron inprevention of postoperative nausea and vomiting: a quantitative systematic review of randomizedplacebo-controlled trials. Anesthesiology 87: 1277–89.

Tramèr MR and Walder B (1999). Efficacy and adverse effects of prophylactic antiemetics during patient-controlled analgesia therapy: a quantitative systematic review. Anesth Analg 88: 1354–61.

Ure D, James KS, McNeill M et al. (1999). Glycopyrrolate reduces nausea during spinal anaesthesia forcaesarean section without affecting neonatal outcome. Br J Anaesth 82: 277–9.

Vallejo MC, Romeo RC, Davis DJ et al. (2002). Propofol-ketamine versus propofol-fentanyl for outpatientlaparoscopy: comparison of postoperative nausea, emesis, analgesia, and recovery. J Clin Anesth 14:426–31.

van Wijk MG and Smalhout B (1990). A postoperative analysis of the patient’s view of anaesthesia in aNetherlands’ teaching hospital. Anaesthesia 45: 679–82.

Vance JP, Neill RS and Norris W (1973). The incidence and aetiology of post-operative nausea and vomitingin a plastic surgical unit. Br J Plast Surg 26: 336–9.

78

Visser K, Hassink EA, Bonsel GJ et al. (2001). Randomized controlled trial of total intravenous anesthesiawith propofol versus inhalation anesthesia with isoflurane-nitrous oxide: postoperative nausea withvomiting and economic analysis. Anesthesiology 95: 616–26.

Wang JJ, Ho ST, Lee SC et al. (2000). The use of dexamethasone for preventing postoperative nausea andvomiting in females undergoing thyroidectomy: a dose-ranging study. Anesth Analg 91: 1404–7.

Wang JJ, Ho ST, Tzeng JI et al. (2000). The effect of timing of dexamethasone administration onits efficacy as a prophylactic antiemetic for postoperative nausea and vomiting. Anesth Analg 91:136–9.

Wang SM and Kain ZN (2002). P6 acupoint injections are as effective as droperidol in controlling earlypostoperative nausea and vomiting in children. Anesthesiology 97: 359–66.

Watcha MF and Smith I (1994). Cost-effectiveness analysis of antiemetic therapy for ambulatory surgery.J Clin Anesth 6: 370–7.

Watcha MF and White PF (1992). Postoperative nausea and vomiting. Its etiology, treatment, and prevention.Anesthesiology 77: 162–84.

Wattwil M, Thorn SE, Lövqvist A et al. (2002). Perioperative gastric emptying is not a predictor of earlypostoperative nausea and vomiting in patients undergoing laparoscopic cholecystectomy. AnesthAnalg 95: 476–9, table of contents.

White PF (2002). Droperidol: a cost-effective antiemetic for over thirty years. Anesth Analg 95: 789–90.White PF, Issioui T, Hu J et al. (2002). Comparative Efficacy of Acustimulation (ReliefBand(R)) versus

Ondansetron (Zofran(R)) in Combination with Droperidol for Preventing Nausea and Vomiting.Anesthesiology 97: 1075–81.

White PF and Watcha MF (1999). Postoperative nausea and vomiting: prophylaxis versus treatment.Anesth Analg 89: 1337–9.

Wilson AJ, Diemunsch P, Lindeque BG et al. (1996). Single-dose i.v. granisetron in the prevention ofpostoperative nausea and vomiting. Br J Anaesth 76: 515–8.

Wood CD (1979). Antimotion sickness and antiemetic drugs. Drugs 17: 471–9.Wu CL, Berenholtz SM, Pronovost PJ et al. (2002). Systematic review and analysis of postdischarge symptoms

after outpatient surgery. Anesthesiology 96: 994–1003.Young SN (1981). Mechanism of decline in rat brain 5-hydroxytryptamine after induction of liver tryptophan

pyrrolase by hydrocortisone: roles of tryptophan catabolism and kynurenine synthesis. Br J Pharmacol74: 695–700.

Zomers PJ, Langenberg CJ and de Bruijn KM (1993). Tropisetron for postoperative nausea and vomiting inpatients after gynaecological surgery. Br J Anaesth 71: 677–80.

prevention of postoperative nausea and vomiting – …

Documents