cricoid pressure in chaos

Correspondence

Is excess intensive caremortality in the UnitedKingdom concealed by ICUmortality prediction models?

The comparative performance of inten-sive care units (ICUs) is measured withcasemix adjustment systems such asAPACHE, SAPS and MPM [1]. Thesesystems calculate predicted hospitalmortality based upon reason for ICUadmission, degree of physiologicalderangement, chronic health status, ageand medical intervention. Predicted hos-pital mortality is calculated using datacollected shortly before and after ICUadmission [1].

There are considerable limitations tocasemix adjustment systems [2, 3]. How-ever, if the standardised mortality ratio(SMR; observed mortality/predictedhospital mortality) is taken as an indicatorof the effectiveness of ICU treatment,outcome is not clearly worse in theUnited Kingdom compared with else-where [4–9]. If the average predictedhospital mortality of admissions, ratherthan the SMR, is used to compare ICUsthen large differences between countriesemerge.

In the North Thames region of theUnited Kingdom a group of ICUs con-tribute information to a database [10].The average predicted hospital mortalityby APACHE II [11] for 12 762 patientsfrom 15 ICUs in this database is 28.6%.Another British ICU database reports apredicted hospital mortality of 27.2%[4]. In other countries average predictedmortality is generally lower, for example19.8% [5], 18.8% and 15.1% in theUnited States [12]. Out of 37 ICUs in

the United States, four reported anaverage predicted hospital mortality ofmore than 25% [12] whereas only two of15 ICUs in North Thames had a pre-dicted mortality less than 25%. Datafrom over 13 000 ICU admissions inthe United Kingdom, eight other Euro-pean countries and North Americashowed that the British hospital mortal-ity for these patients was highest at32.4%, compared with a median of 21%for the other European countries and19.7% for North America [13]. Pre-liminary data from the European Con-sortium for Intensive Care Data (ECICD)using SAPS II to assess severity of illnessshow that intensive care patients in theUnited Kingdom are sicker than any ofthe eight other participating countries(ECICD abstract, not published). ACanadian study reported a predictedhospital mortality of 24.7% [6] and oneof 20% was calculated for a group ofBrazilian ICUs [7].

Patients already in hospital accountfor a high proportion of high-risk ICUadmissions. In our data, ICU admissionsfrom the ward are 21.7% of total admis-sions. These have a 52.9% hospital mor-tality (1466 deaths), compared with22.3% (1156 deaths) of those admittedfrom the operating theatre/recovery and30.2% (1081 deaths) from the accidentand emergency department. Of patientsadmitted to ICUs following externalcardiac massage or defibrillation, 42.9%(677 patients) came from the ward.These patients had a 79.5% mortality.In our own hospital 34.8% of ICUadmissions of patients who had been inhospital at least 24 h were following arespiratory or cardiac arrest on the ward

[14]. In 1996 there were 142 cardiacarrest calls to the wards following which33 patients (23%) were admitted to theICU.

Our research [14] and that of others[15] suggests that it is possible to identifyearly those ward patients likely to requireICU admission or suffer a cardiac arrest.Early recognition of these patients mayallow management to prevent deteriora-tion in physiological values or to preventarrest. Such intervention is likely toimprove outcome. The incidence ofcardiac arrest on the ward may thereforebe a useful indicator of the quality ofcare. Compared with the United King-dom, in some other countries a higherpercentage of resources is given to caringfor critically ill patients [16, 17] and ICUadmissions have a lower average pre-dicted mortality suggesting that patientsare likely to have access to appropriatecare earlier. If patients are identifiedearly and admitted to the ICU, in parti-cular before a respiratory or cardiacarrest, their predicted mortality will beless but, given appropriate ICU treat-ment, so will the observed mortality.There will be no difference in theSMRs and no indication of the improvedICU outcome.

The relative lack of critical careresources and the high predicted mor-tality of patients admitted to British ICUspoint to the possibility of an excess mor-tality compared with better resourcedmedical systems. Early identification ofcritically ill patients may help improvecare for these patients on the ward orfacilitate early admission to an appropri-ate high-dependency area or ICU. Thisis likely to decrease the number of deaths

Anaesthesia, 1998, 53, pages 89–101................................................................................................................................................................................................................................................

89Q 1998 Blackwell Science Ltd

All correspondence should be addressed to Dr M. Morgan, Editor of Anaesthesia, Department of Anaesthetics, Royal Postgraduate Medical School,Hammersmith Hospital, London W12 0HS, UK.

Letters (two copies) must be typewritten on one side of the paper only and double spaced with wide margins. Copy should be prepared in the usualstyle and format of the Correspondence section. Authors must follow the advice about references and other matters contained in the Notice toContributors to Anaesthesia printed at the back of each issue. The degree and diplomas of each author must be given in a covering letter personally signedby all the authors.

Correspondence presented in any other style or format may be the subject of considerable delay and may be returned to the author for revision. If theletter comments on a published article in Anaesthesia, please send three copies; otherwise two copies of your letter will suffice.

without altering the SMR. The use ofcasemix adjustment systems to compareICU performance will conceal ratherthan reveal this excess mortality.

D. R. GoldhillP. S. WithingtonThe Royal London Hospital,London E1 1BB

Acknowledgments

Funding to support the North ThamesICU audit system was provided bythe North Thames (East) RegionalHealth Authority. We are grateful tothe intensive care units who supplieddata and appreciate the time and effortthey contributed in order to collect thedata.

References1 Rowan K. Risk adjustment for

intensive care outcomes. In: GoldhillDR, Withington PS, eds. Textbook ofIntensive Care. Chapman Hall:London, 1997; 787–803.

2 Goldhill DR, Withington PS. Theeffect of casemix adjustment onmortality as predicted by APACHEII. Intensive Care Medicine 1996; 22:415–9.

3 Goldhill DR, Withington PS.Mortality predicted by APACHE II.The effect of changes in physiologicalvalues and post-ICU hospital mortality.Anaesthesia 1996; 51: 719–23.

4 Rowan KM, Kerr JH, Major E,McPherson K, Short A, Vessey MP.Intensive Care Society’s APACHE IIstudy in Britain and Ireland – II:Outcome comparisons of intensivecare units after adjustment for casemix by the American APACHE IImethod. British Medical Journal 1993;307: 977–81.

5 Knaus WA, Draper EA, Wagner DP,Zimmerman JE. An evaluation ofoutcome from intensive care in majormedical centers. Annals of InternalMedicine 1986; 104: 410–8.

6 Wong DT, Crofts SL, Gomez M,McGuire GP, Byrick RJ. Evaluationof predictive ability of APACHE IIsystem and hospital outcome inCanadian intensive care unit patients.

Critical Care Medicine 1995; 23:1177–83.

7 Bastos PG, Knaus WA, ZimmermanJE, Magalhaes A, Sun X, Wagner DP.The importance of technology forachieving superior outcomes fromintensive care. Intensive Care Medicine1996; 22: 664–9.

8 Knaus WA, Wagner DP, ZimmermanJE, Draper EA. Variations inmortality and length of stay inintensive care units. Annals of InternalMedicine 1993; 118: 753–61.

9 Moreno R, Morais P. Outcomeprediction in intensive care: results ofa prospective, multicentre, Portuguesestudy. Intensive Care Medicine 1997;23: 177–86.

10 Goldhill DR, Withington PS, BirchNJ. The North East Thames regionalintensive care audit system. TheoreticalSurgery 1993; 8: 1–5.

11 Knaus WA, Draper EA, Wagner DP,Zimmerman JE. APACHE II: aseverity of disease classificationsystem. Critical Care Medicine 1985;13: 818–29.

12 Zimmerman JE, Shortell SM, KnausWA, et al. Value and cost of teachinghospitals: a prospective, multicenterinception cohort study. Critical CareMedicine 1993; 21: 1432–42.

13 Le Gall J-R, Lemeshow S, Saulnier F.A new simplified acute physiologyscore (SAPS II) based on a European/North American multicenter study.JAMA 1993; 270: 2957–63.

14 Goldhill DR, White SA, Brohi K,Sumner A. Physiological values andinterventions in the 24 hours beforeadmission to the ICU from the ward.Clinical Intensive Care 1997; 8: 103.

15 Franklin C, Mathew J. Developingstrategies to prevent inhospital cardiacarrest: analyzing responses ofphysicians and nurses in the hoursbefore the event. Critical CareMedicine 1994; 22: 244–7.

16 Vincent JL. European attitudestowards ethical problems in intensivecare medicine: results of an ethicalquestionnaire. Intensive Care Medicine1990; 16: 256–64.

17 Bion J. Rationing intensive care:preventing critical illness is better,and cheaper, than cure. British MedicalJournal 1995; 310: 682–3.

Early death amongstanaesthetists: a statisticalhowler

Milner and Ziegler stated recently in aletter to Anaesthesia (1997; 52: 797–8)that ‘Ina reviewoftheobituariespublishedin the British Medical Journal, Wright andRoberts have shown the average age ofdeath . . . of anaesthetists . . . [to be] thelowestof all doctorsborn in theUK. [This]. . . gives sobering thought to the profes-sion.’Sober thoughtis indeedrequired,butbefore immediately jumping to the con-clusion that this difference is due to thestress under which anaesthetists work, it isnecessarytoaskwhether theconclusionsofWright and Roberts [1] are themselvesvalid.InfacttheEditoroftheBritishMedicalJournal subsequently admitted that thejournal had committed a ‘statisticalhowler’ and said that ‘many people wroteto point out the error’ [2] and publishedtwo letters [3, 4], explaining the well-known statistical problem, which is tech-nically the absenceof valid denominators.

The problem is easily stated. There isno doubt from the data of Wright andRoberts that obituaries of anaesthetistsreport a younger age at death than thatfor other doctors. However, this mightresult from anaesthetists also beingyounger in life than other doctors, becausebeing an anaesthetist is a marker ofbeing from a later birth cohort. Anaes-thetics is one of the younger medicalspecialities, having developed substan-tially in the past four or five decades andindeed its practitioners tend to be some-what younger (as shown in an analysis of128 417 doctors on the Medical Regis-ter [5]). This probably explains entirelythe results of Wright and Roberts.

Anaesthetists may be under more stressthan other doctors, which may be dele-terious for their health, but such a casewill not be helped by quoting datawhich are statistically flawed. Ironicallythe Editorial in the same issue ofAnaesthesia noted, ‘it is essential thatanyone using a statistical technique iscompletely familiar with it’ [6].

I. C. McManusProfessor of Psychology,Imperial College School of Science,Technology and Medicine,London W2 1PD

Correspondence Anaesthesia, 1998, 53, pages 89–101................................................................................................................................................................................................................................................

90 Q 1998 Blackwell Science Ltd

References1 Wright DJM, Roberts AP. Which

doctors die first? Analysis of the BMJobituary columns. British MedicalJournal 1996; 313: 1581–2.

2 Smith R. Editor’s choice: patients’consent and questions, and a statisticalhowler. British Medical Journal 1997;314: no page number.

3 Khaw K. Which doctors die first?Lower mean age of death in doctors ofIndian origin may reflect different agestructures. British Medical Journal 1997;314: 1132.

4 McManus IC. Which doctors die first?Recording the doctors’ sex might haveled authors to suspect theirconclusions. British Medical Journal1997; 314: 1132.

5 McManus IC. Anaesthetists areyounger than other doctors. BritishMedical Journal 1997; 315: 314.

6 Goodman NW. Meta-analysis.Anaesthesia 1997; 52: 723–5.

Self-administration of pre-operative analgesicsuppositories

I aim to tell all patients pre-operativelyabout my plans for postoperative anal-gesia, especially since an anaesthetist wasfound guilty of serious professional mis-conduct for inserting a suppository intoan anaesthetised patient without obtain-ing prior consent [1]. Occasionally Irealise after induction that I have omittedto mention the suppository. This pre-sents a dilemma: do I insert it anywaybecause I believe it provides good anal-gesia ( paternalistic attitude?) or sacrificethis quality of care in respecting thepatient’s autonomy?

Recently, after discussion with thenurses on the day care unit, I haveoffered day-case gynaecology patientspre-operative self-administration of anal-gesic suppositories. This has severaladvantages: the need to obtain specificconsent is avoided; the insertion of sup-positories is less likely to be overlookedin the rapid turn-around of patients inthe operating theatre; patients partici-pate in their own care; analgesia is pro-vided earlier (utilising any pre-emptiveeffect). The patients have readily acceptedthe practice of inserting their own

suppositories in spite of an anonymousquestionnaire finding that 50% of femaleday-case patients considered rectal drugadministration unacceptable [2]. The veryfew who balk at the idea of self-admin-istration are offered the choice of rectalanalgesia inserted under anaesthesia (forwhich consent is then obtained) or post-operative oral medication. The additionalwork for the ward nurses pre-operatively(in dispensing the suppositories andsupervising their administration) is com-pensated by better postoperative anal-gesia and more rapid discharge home.

This change in practice avoids insert-ing suppositories into the large majorityof unconscious patients and has beenwell accepted by all involved. Othersmight consider adopting it.

D. JolliffeLeicester Royal Infirmary,Leicester LE1 5WW

References1 Mitchell J. A fundamental problem of

consent. British Medical Journal 1995;310: 43–6.

2 Vyvyan HAL, Hanafiah Z. Patientsattitudes to rectal drug administration.Anaesthesia 1995; 50: 401–2.

Combined introducer forreinforced laryngeal maskairway

Since the introduction of the reinforcedlaryngeal mask airway (RLMA), manydevices have been described to facilitateits insertion. These include a variation ofthe standard Magill forceps [1], a metalstylet [2, 3], a small tracheal tube [4] andthe Bosworth introducer [5]. Disadvan-tages of these devices include trauma tothe larynx with the forceps and metalstylet, displacement of the mask onremoval of the stylet and reduced tactilefeedback with the Bosworth introducer.With a tracheal tube alone as anintroducer we find that it lacks therigidity necessary for the insertion ofthe RLMA, which has a tendency torotate along its longitudinal axis. Withthese considerations in mind we wish todescribe an introducer combining theadvantages of a metal stylet and trachealtube.

The tracheal tube used is essentially ofthe same length and diameter as describedby Asai et al. [4]. A metal stylet is thenpassed through the tracheal tube and thelength adjusted so that the distal enddoes not protrude through the grill of

Anaesthesia, 1998, 53, pages 89–101 Correspondence................................................................................................................................................................................................................................................


Figure 1

the mask (Figs 1 and 2). This arrange-ment allows the shaft of the RLMA tobe rigid enough to be precurved inorder to mirror the oropharynx, therebymaking insertion easy and atraumatic.Adequate lubrication of the stylet andtracheal tube allows the removal of thestylet without displacing the RLMA.This is easily achieved by holding thejunction of the tube and mask whenpulling back the stylet. We feel that thiscombined approach minimises the poten-tial disadvantages of the stylet and atracheal tube used alone as introducersfor the RLMA yet retaining their indi-vidual advantages.

P. MainoM. PilkingtonM. PopatJohn Radcliffe Hospital,Oxford OX3 9DU

References1 Welsh BE. Use of a modified Magill’s

forceps to place a flexible laryngealmask. Anaesthesia 1995; 50: 1002.

2 Philpott B, Renwick M. Anintroducer for the flexible laryngealmask airway. Anaesthesia 1993; 48: 174.

3 Harris S, Perks D. Introducer forreinforced laryngeal mask airway.Anaesthesia 1997; 52: 603.

4 Asai T, Stacey M, Barclay K. Stylet for

reinforced laryngeal mask airway.Anaesthesia 1993; 48: 636.

5 Bosworth A, Jago HR. The Bosworthintroducer for use with the flexiblereinforced laryngeal mask airway.Anaesthesia 1997; 52: 281–2.

Cuff inflator for tracheal tubes

The ethos behind the work of Drs Abde-latti and Kamath (Anaesthesia 1997; 52:765–9) in developing a device for cuffinflation of tracheal tubes is to be com-mended in principle. The identificationof clinical need, the ingenious develop-ment of apparatus and the meticulous,accurate testing of new devices has beenthe cornerstone upon which much ofmodern-day anaesthetic practice hasbeen established. In their article DrsAbdelatti and Kamath highlight the dif-ficulties that a right-handed assistant mayhave in inflating a tracheal cuff with theirnondominant hand whilst applying cri-coid pressure with the other. In additionthey list the properties that the ‘ideal’cuff inflator device should possess, includ-ing ‘rapid and consistent inflation of thecuff to produce an air tight seal, mea-surement of the injected volume of airand safety precautions to prevent overinflation of the cuff.’ It has obviously

escaped the authors notice that connec-tion of the syringe to the tracheal cuffinflator before the rapid sequence induc-tion has been embarked on will elimi-nate the assistant’s need to carry out thisdextrous manoeuvre under stress. Fur-thermore the preconnected syringe willfulfil all the authors’ own criteria for an‘ideal’ tracheal cuff inflator device.

Is it really necessary to complicatewhat is essentially a simple procedureby the development of yet more obsoleteanaesthetic equipment? There is anexpression about mountains and mole-hills that springs to mind when con-sidering the efforts of Drs Abdelattiand Kamath.

J. FrancisBedford Hospital,Bedford MK42 5BB

A replyThank you for the opportunity to reply.We would like to reassure Dr Francisthat only after using/studying all differ-ent options available over the years, wasit decided that there was a need for anautomatic cuff-inflator. The recent sug-gestion that ‘cricoid pressure’ should beperformed as a ‘two-handed’ procedureby the anaesthetic assistant further sup-ports our contention [1]. We would alsolike to clarify that designing the inflatorwas only the first stage of developmentand the facility to measure/regulate cuffpressure is being incorporated presently,to fulfil all the criteria for an ideal deviceas outlined in our paper. When com-pleted, the device will be left connectedto the tracheal tube in the peri-operativeperiod, including during long-term ven-tilation, allowing continuous regulationof cuff pressure. In present-day anaes-thesia where all aspects of patient man-agement are routinely monitored, webelieve that we are obliged to devotemore attention to one of the essentialand sometimes crucial parts of anaes-thetic practice, namely tracheal intuba-tion [2].

M. O. AbdelattiB. S. K. KamathBarnet General Hospital,Barnet,Hertfordshire EN5 3DJ

Correspondence Anaesthesia, 1998, 53, pages 89–101.................................................................................


Figure 2

References1 Yentis SM. The effects of single-

handed and bimanual cricoid pressureon the view at laryngoscopy.Anaesthesia 1997; 52: 332–5.

2 Pipin LK, Short DM, Bowes JB. Longterm tracheal intubation practice in theUnited Kingdom. Anaesthesia 1983;38: 791–5.

Use of a tracheal tube andcapnograph for insertion of afeeding tube

A feeding tube can be inadvertentlyinserted into the tracheobronchial treeand may even penetrate the pleura, caus-ing pneumothorax [1, 2]. In dementedor unconscious patients, insertion of thetube into the trachea sometimes doesnot induce protective airway reflexesand thus the incident may not bedetected [1, 2]. Infusion of nutrientsthrough the tube in such a circumstancemay be lethal [3]. Therefore, it is crucialto detect inadvertent insertion of a feed-ing tube in the airway.

Previously, I suggested that the useof a capnograph would be one reliablemethod for detection [4]: after insertionof a feeding tube, a capnograph is con-nected to the proximal orifice of thefeeding tube; if carbon dioxide isdetected, the tube is likely to be in theairway. It is possible to minimise theincidence of migration of the tube intothe intrapleural space if the absence ofcarbon dioxide output is confirmed atthe point when the feeding tube hasbeen inserted to a depth of 25 cm. How-ever, this method requires removal ofthe guide wire from the feeding tube forthe measurement of carbon dioxide andre-insertion of the wire for final posi-tioning of the tube into the gastrointes-tinal tract. This re-insertion of the guidewire may damage the tube or may resultin the wire protruding from a side holeof the feeding tube [5].

I now suggest that the use of a trachealtube will solve this problem. A trachealtube is inserted into the oesophagus anda sampling catheter of a capnograph isconnected to the tracheal tube. Whenthe tube is correctly inserted into theoesophagus, there should be no increase

in carbon dioxide concentration. Afterthis confirmation, a feeding tube ispassed through the tracheal tube intothe oesophagus and advanced into thegastrointestinal tract. The tracheal tubeis then removed. Finally, the guide wireis removed, a capnograph is connectedto the feeding tube and the absence ofcarbon dioxide emission is confirmed.

T. AsaiKansai Medical University,Osaka 570, Japan

References1 McWey RE, Curry NS, Schabel SI,

Reines HD. Complications ofnasoenteric feeding tubes. AmericanJournal of Surgery 1988; 155: 253–7.

2 Scholten DJ, Wood TL, ThompsonDR. Pneumothorax from nasoentericfeeding tube insertion. A report of fivecases. American Surgeon 1986; 52:381–5.

3 Muthuswamy PP, Patel K, RajendranR. ‘Isocal pneumonia’ with respiratoryfailure. Chest 1982; 81: 390.

4 Asai T, Stacey M. Confirmation offeeding tube position; how aboutcapnography? Anaesthesia 1994; 49:451.

5 Grossman TW, Duncavage JA,Dennison B, Kay J, Toohill RJ.Complications associated with anarrow bore nasogastric tube. Annals ofOtology, Rhinology and Laryngology1984; 93: 460–3.

Testing regional anaesthesiabefore Caesarean section

The paper by Bourne et al. (Anaesthesia1997; 52: 896–903) on how the Britishobstetric anaesthetist tests regional anaes-thesia before Caesarean section is timelyand appropriate since regional anaesthe-sia obviously accounts for the majorityof anaesthetics for the procedure.

Their figure of 9/555 (essentiallyobstetric anaesthetists) who preferredgeneral anaesthesia for Caesarean sectionis encouraging since at the GuildfordObstetric Anaesthetists’ Association(OAA) meeting (April 1997) the inter-active voting system enabled the 330anaesthetists in the audience to indicate

their preferred technique for electiveCaesarean section: 9% preferred a gen-eral anaesthetic, 10% an epidural, 23% acombined spinal/epidural (CSE) and58% a one-shot spinal. These latter fig-ures should not be compared directlywith those provided by Bourne et al.since the composition of the audiencewas unknown, but was probably a mix-ture of specialist and nonspecialist anaes-thetists.

As well as the variation in the recom-mended level of block for Caesareansection found in textbooks and the lackof description as to which sensation theparticular level of recommended blockrefers, there is a practical confusionabout the level of block recorded. Work-ing in a cephalad direction with thepatient instructed ‘tell me when youfeel me touch you’ is the level of blockthe first unblocked dermatome or thelast blocked dermatome: the patient’sresponse may indicate touch sensationis appreciated at the T4 dermatome, butis this a block to T4 or T5? This blockcould be correctly described as a ‘blockto (i.e. up to) T4’ or ‘the upper level ofblock is T5’. It was for this reason thatmy description [1] of the level of blockrequired for a pain-free Caesarean sec-tion was described as ‘up to andincluding T5’, i.e. the first unblockeddermatome was T4. It must also beremembered that my results wereobtained before the use of spinal andepidural opioids. Clinical experience andthe results of many studies indicate that apain-free Caesarean section is signifi-cantly more likely when fentanyl isadded to the local anaesthetic (particu-larly with epidural anaesthesia), althoughthe level of block assessed in the fentanyland nonfentanyl groups may be the same(i.e. the required level of block for apain-free section when fentanyl is addedis lower than that without fentanyl).

If assessment of the upper level of blockis confused, then this is even more so forthe lower level. While we all accept thatmissed segments may occur with epi-dural anaesthesia it is not true to say ‘it isclinically recognised that spinal anaes-thesia guarantees complete sensory lossbelow the most cephalad level’. This willdepend on how the block is assessed: ifthe block is assessed by cold or sharp pin



prick then there may well be significantother sensations below the most cepha-lad level.

Even with a spinal, many patients areaware of pulling sensations during Cae-sarean section (although some of thesesensations may travel by extraspinalroutes) and it is not uncommon to per-form a pain-free Caesarean section on apatient who can still move her toes. Aretired anaesthetist (E. H. (Ted) Morgan,Australia) drew my attention to the factthat with spinal cinchocaine, it was oftenvery difficult, or impossible, to operateon the sole of the foot (this is confirmedby several of my colleagues with experi-ence in Asia). Over the last few months Ihave checked the soles of my patients’feet during their Caesarean sections andhave found that over 90% can appreciatetouch on the soles of both feet! VASscores for intra-operative pain or dis-comfort were zero.

My final comment is one of conster-nation. After Bourne et al. discuss theimplications of the differential block(between cold, sharp pin prick andtouch) and describe how cold and pinprick may be misleading in assessing theadequacy of the block, I am mystifiedto note the authors’ comment ‘in ourdepartment we use ethyl chloride spray’!Apart from questions about what ethylchloride is assessing there are also prob-lems associated with environmentalpollution by an inflammable chlorinatedhydrocarbon.

I. F. RussellRoyal Hull Hospital,Hull Royal Infirmary,Hull HU3 2JZ

Reference1 Russell IF. Levels of anaesthesia and

intraoperative pain at Caesareansection. International Journal of ObstetricAnesthesia 1995; 4: 71–7.

Cricoid pressure in chaos

In their letter Brimacombe and Berry(Anaesthesia 1997; 52: 924–6) raise someinteresting but controversial issues. Thestudies that have shown cricoid pressureto be potentially harmful have all applied

excessive force. Forces of 40 N and overcan cause: airway obstruction [1, 2],difficulty in passing the tracheal tube[3] and oesophageal rupture [4]. Also,when excessive force is used the view atlaryngoscopy has to be improved withneck support [5] presumably because ofthe head flexing on the neck, but with30 N of cricoid pressure neck support isnot needed; indeed the view is betterthan without cricoid pressure [6]. It issurprising that these excessive forcescontinue to be used when 30 N hasbeen shown to prevent the regurgitationof oesophageal fluid with a pressure of40 mmHg in 10 cadavers [4]. Gastricpressures over 25 mmHg have neverbeen recorded while supine and at resteven during pregnancy; therefore 30 Nis probably more than enough to preventpassive regurgitation into the pharynx.After practising on weighing scales 3 kg(30 N) can be reproduced 50 times withinsafe limits: mean of 3.2 kg, range 2.5–3.7,SD 0.29 [6]. This force should thereforebe practised by all anaesthetic assistants.

Light cricoid pressure must be appliedto the awake patient during induction ofanaesthesia [7]; although 20 N of force istolerated by most people it is uncomfor-table [2] and it must not be exceeded incase retching and oesophageal ruptureoccurs [4]. If anaesthetic assistants prac-tised 20 N (2 kg) of force on weighingscales a range of forces would actually beapplied to patients and therefore 20 Nwould be exceeded 50% of the time. Asafer recommendation would be 10 N(1 kg) when awake increasing to 30 Nwith loss of consciousness [8]; this hasbeen the practice in our hospital for thepast 5 years and we have not seen regur-gitation with cricoid pressure.

Never mind cricoid pressure in chaos,what about the failed intubation drill?Brimacombe and Berry suggest that ifintubation is difficult, or if oxygen satu-ration on the pulse oximeter starts to fallduring attempts at intubation, then cri-coid pressure should be reduced, thenreleased, in order to improve the view atlaryngoscopy for another attempt at intu-bation. Releasing cricoid pressure gradu-ally under direct vision with suction inhand may be relatively safe as cricoidpressure could be increased again if regur-gitation occurred which may prevent

aspiration. However, if cricoid pressureis applied correctly, releasing it is unlikelyto improve the view at laryngoscopy andwill probably make it worse [6]. Thephilosophy of the failed intubation drillis oxygenation without aspiration and alsoto admit a failed intubation at an earlystage before desaturation starts to occur.Ventilation using a facemask and an oralairway should immediately follow thisdecision and continue until spontaneousventilation resumes. Cricoid pressureshould be maintained during mask ven-tilation as regurgitation may be pro-voked by a rise in oesophageal pressure[9]. Mask ventilation with cricoid pres-sure maintained was not difficult in 70%of cases of failed intubation in obstetriccases [10]. However, if ventilation is dif-ficult the applied cricoid pressure shouldthen be reduced by half (half of whateverforce they happen to be applying); if thisdoes not help a laryngeal mask is inserted.If the laryngeal mask fails to provide anairway it is removed and cricoid pressurecan then be released under direct vision,suction in hand, before a second attemptat laryngeal mask insertion, which canbe followed by re-application of cricoidpressure. Hopefully by this time thepatient is breathing and cricothyroidot-omy, which is the next step in the failedventilation drill, can be avoided.

R. G. VannerGloucestershire Royal Hospital,Gloucester GL1 3NN

References1 Allman KG. The effect of cricoid

pressure application on airwaypatency. Journal of Clinical Anesthesia1995; 7: 197–9.

2 Vanner RG. Tolerance of cricoidpressure by conscious volunteers.International Journal of ObstetricAnesthesia 1992; 1: 195–8.

3 Lawes EG, Duncan PW, Bland B,Gemmel L, Downing JW. Thecricoid yoke – a device for providingconsistent and reproducible cricoidpressure. British Journal of Anaesthesia1986; 58: 925–31.

4 Vanner RG, Pryle BJ. Regurgitationand oesophageal rupture with cricoidpressure: a cadaver study. Anaesthesia1992; 47: 732–5.



5 Yentis SM. The effects of single-handed and bimanual cricoid pressureon the view at laryngoscopy.Anaesthesia 1997; 52: 332–5.

6 Vanner RG, Clarke P, Moore WJ,Raftery S. The effect of cricoidpressure and neck support on theview at laryngoscopy. Anaesthesia1997; 52: 896–900.

7 Vanner RG, Pryle BJ, O’Dwyer JP,Reynolds F. Upper oesophagealsphincter pressure and the intravenousinduction of anaesthesia. Anaesthesia1992; 47: 371–5.

8 Vanner RG. Mechanisms ofregurgitation and its prevention withcricoid pressure. International Journal ofObstetric Anesthesia 1993; 2: 207–15.

9 O’Mullane EJ. Vomiting andregurgitation during anaesthesia.Lancet 1954; 1: 1209–12.

10 Hawthorne L, Wilson R, Lyons G,Dresner M. Failed intubationrevisited: 17-yr experience in ateaching maternity unit. British Journalof Anaesthesia 1996; 76: 680–4.

The use of CO2 gap to monitorsplanchnic oxygenation

I read with interest the report of DrsShaw and Baddock on the use of con-tinuous gastric tonometry (Anaesthesia1997; 52: 709–10). This is an interestingearly report of new technology whichmerits further comment.

The use of intramucosal ÿarterialCO2 gap (P iCO2ÿPaCO2) may wellhold promise in identifying patientswith splanchnic ischaemia. It avoids theassumption that arterial and intramuco-sal bicarbonate concentrations are equal(a necessary assumption to allow calcu-lation of pHi) and therefore avoids thepotential error of a calculated low pHi inthe absence of splanchnic ischaemia, e.g.renal failure, diabetic ketoacidosis. How-ever the substitution of endtidal CO2

(PE 0CO2) for PaCO2 and the use ofP iCO2ÿPE 0CO2 gap may not be validas the relationship between PaCO2 andPE 0CO2 may well vary in critically illpatients. This point was mentioned byDrs Shaw and Baldock. However, theythen describe a case where PE 0CO2 felland P iCO2 was elevated in relation toocclusion of the inferior vena cava and

presumed fall in cardiac output. Duringthis period of vascular clamping therelationship between PE 0CO2 and PaCO2

would have been drastically altered. End-tidal CO2 and, although not measured,PaCO2 and tissue CO2 would almostcertainly have risen. It has been demon-strated that alterations in PaCO2 arereflected by similar alterations in P iCO2,although the CO2 gap remains constant[1]. Therefore the P iCO2ÿ PE 0CO2 gapwould have been increased by two mech-anisms in this case, neither of whichwould necessarily infer splanchnicischaemia. I do not dispute that thisoperative procedure and decreased car-diac output due to vascular clampingmay have caused splanchnic hypoper-fusion but the use of PE 0CO2ÿ P iCO2

gap under these circumstances isinappropriate.

My second reservation is the com-ment that in the 12 h prior to discon-tinuation of monitoring the CO2 gapnever rose above 3–4 kPa and that thisreflected the patient’s excellent condi-tion. Information from the accompany-ing illustration reveals that the pHi

90 min prior to this time was 7.19. Wehave recently shown that a CO2 gap of> 3 kPa has a sensitivity of 68% and aspecificity of 100% in recognising apHi< 7.32 [2], which is borne out bythe pHi displayed in the report by DrsShaw and Baldock. We would be con-cerned about such a high CO2 gapand low pHi as this has been shownmany times to be a poor prognosticfactor [3, 4].

In summary we feel that although thisis promising new technology the use ofthe P iCO2ÿ PE 0CO2 gap is a noisy para-meter due to the nonfixed relationshipbetween PE 0CO2 and PaCO2 and shouldbe used with caution. A possible refine-ment for the future would be to com-bine the digital output from the Tonocapwith an intra-arterial CO2 electrode thusproviding a real time P iCO2ÿ PaCO2 gapand pHi if desired. In addition the rangefor normal CO2 gap needs to be defined.

G. P. FindlayUniversity Hospital of Wales,Cardiff CF4 4XW

References1 Mas A, Gaigorri F, Joseph D, et al.

Effect of acute changes of PaCO2 inthe assessment of gastric perfusion bytonometry. Intensive Care Medicine1996; 22: S437.

2 Findlay GP, Kruger M, Smithies MN.The measurement of gastro-intestinalintra-mucosal carbon dioxide tensionby semi-continuous air tonometry.Clinical Intensive Care (in press).

3 Maynard N, Bihari D, Beale R, et al.Assessment of splanchnic oxygenationby gastric tonometry in patients withacute circulatory failure. Journal of theAmerican Medical Association 1993; 270:1203–10.

4 Doglio GR, Pusajo JF, Egurrola MA,et al. Gastric mucosal pH as aprognostic index of mortality incritically ill patients. Critical CareMedicine 1991; 19: 1037–40.

Prediction of outcome by pHi,pHa and CO2 gap

I read with interest the thought-provok-ing article by Gomersall and colleaguesregarding the prediction of outcome bygastric tonometry (Anaesthesia 1997; 52:619–23). In appraising this paper it isworthwhile considering the derivationof pHi and the similarity to the deriva-tion of pHa:

pHi� 6.1� log([HCOÿ3 arterial]/[0.225 ×P iCO2])

pHa� 6.1� log([HCOÿ3 arterial]/[0.225 ×PaCO2]),

where P iCO2 equals intramucosal CO2

tension. As can be seen from these twoequations the derivation of pHi can bereduced to

pHi� pHaÿ log P iCO2/PaCO2.

From this third equation it can be easilyappreciated that low pHi can be attri-buted to three mechanisms: (1) second-ary to low pHa due to mathematicalcoupling (low HCOÿ3 arterial or acuterises in PaCO2); (2) elevation of P iCO2

with respect to PaCO2; (3) a combinationof the above two mechanisms.

If one believes that low pHi as amarker of splanchnic ischaemia is super-ior in predicting poor outcome thanpHa, as has been shown by previousstudies [1, 2], then the only logical



conclusion is that mechanism 2 above isof importance. This has been elegantlydemonstrated in a theoretical model bySchlichtig et al. [3]. However, in thestudy by Gomersall et al., pHa wassuperior to pHi in predicting outcomeat 0 and 24 h, suggesting that low pHi

may be, in a large part, due to mech-anism 1 above. In addition the CO2

gap was poorer than both pHi andpHa at discriminating survivors fromnonsurvivors.

Of the three possible explanations ofthe findings of Gomersall and colleaguesthey focus on the hypothesis that theP iCO2ÿ PaCO2 gap is not a marker ofsplanchnic ischaemia. However, Fried-man et al. have shown that the CO2 gapwas significantly higher in patients withevidence of tissue hypoxia than thosewith no evidence of tissue hypoxia andthat pHi correlated better with P iCO2

than arterial bicarbonate or base deficit[4]. In addition we have recently shownthat a CO2 gap of $ 3 kPa has a sensi-tivity of 68% and a specificity of 100% inpredicting that pHi< 7.32 [5]. It may bethat the incidence of splanchnic ischae-mia in the patients studied by Gomersallet al. was lower than their calculated pHi

may suggest. An analysis of their admis-sion diagnosis reveals that many patientshave the potential to have a low pHi dueto mechanism 1 above. Asthma, chronicobstructive airways disease and pneu-monia may lead to respiratory acidosiswhich has been shown to produce lowpHi but unaltered P iCO2ÿPaCO2 gap[6]. Similarly, diabetic ketoacidosis maylead to low pHi due to low arterialbicarbonate. These patients, and the mis-cellaneous group, constitute over 50% ofthe study population thus potentiallyundermining the study.

In conclusion, the finding that pHa issuperior to pHi in predicting outcome isnot in agreement with much of thepublished literature. This may be dueto the heterogeneous study populationof small numbers and mathematicalcoupling leading to low pHi. Further-more the statement that P iCO2ÿPaCO2

gap is not a measure of splanchnicischaemia is not supported by the data,as no independent measures of splanch-nic oxygenation were performed. Finallythis paper does not give grounds for the

abandonment of the use of theP iCO2ÿ PaCO2 gap.

G. P. FindlayUniversity Hospital of Wales,Cardiff CF4 4XW

References1 Maynard N, Bihari D, Beale R, et al.

Assessment of splanchnic oxygenationby gastric tonometry in patients withacute circulatory failure. Journal of theAmerican Medical Association 1993; 270:1203–10.

2 Doglio GR, Pusajo JF, Egurrola MA,et al. Gastric mucosal pH as aprognostic index of mortality incritically ill patients. Critical CareMedicine 1991; 19: 1037–40.

3 Schlichtig R, Mehta N, Gayowski TJ.Tissue-arterial PCO2 difference is abetter marker of ischemia thanintramural pH (pHi) or arterialpH-pHi difference. Journal of CriticalCare 1996; 11: 51–6.

4 Friedman G, Berlot G, Kahn RJ,Vincent JL. Combined measurementsof blood lactate concentrations andgastric intramucosal pH in patientswith severe sepsis. Critical Care Medicine1995; 23: 1184–93.

5 Findlay GP, Kruger M, Smithies MN.The measurement of gastro-intestinalintra-mucosal carbon dioxide tensionby semi-continuous air tonometry.Clinical Intensive Care (in press).

6 Mas A, Gaigorri F, Joseph D, et al.Effect of acute changes of PaCO2 inthe assessment of gastric perfusion bytonometry. Intensive Care Medicine1996; 22: S437.

A replyWe thank Dr Findley for his interest inour paper. In reply we would like tomake the following points. Schlichtig etal.’s mathematical model demonstratedthat P iCO2ÿ PaCO2 gap is independentof systemic acid–base state. They did notassess its importance in the causation oflow pHi [1]. In the complex setting ofcritically ill patients, the relative contri-butions of pHa, pHi and P iCO2 arealmost certainly not constant either inmagnitude or in relative importancewith regard to their contribution to

pHi. It is the clinical setting that weevaluated.

The superiority of pHa over pHi wasnot discussed and we do not believe thatour data allow Dr Findley to concludethat pHa was a superior predictor in ourpatients. Although not mentioned inour paper, there was no statistically norclinically significant difference betweenthe area under the receiver operatingcharacteristic (ROC) curves for pHa

and pHi. We chose not to discuss thisissue because our sample size was toosmall for such a comparison to be mean-ingful. Very large sample sizes arerequired for comparisons of the areaunder ROC curves if the comparisonis to have adequate power when the twoareas are similar [2]. Friedman et al. [3]only showed that P iCO2ÿ PaCO2 gapwas higher in septic patients withraised lactate levels. It must be pointedout that, although frequently raised insepsis, lactate is a poor indicator of tissuehypoxia [4, 5]. In addition we did notstate that P iCO2ÿ PaCO2 gap is not ameasure of splanchnic ischaemia butonly that that is the most likely explana-tion of our results. We are acutely awarethat we did not measure splanchnicoxygenation, for the simple reason thatthis would have been extremely difficultto do. Indeed we know of no humanstudies in which splanchnic oxygenationhas been measured by a validatedmethod.

As stated in our paper Vandermeer etal. have shown that pHi may be low inthe absence of tissue hypoxia in ananimal study [6].

We agree that a substantial number ofour patients may not have had splanch-nic ischaemia but disagree with Dr Find-ley regarding the effect of this on thesignificance of our study. We chose amiscellaneous group as it is only if a riskfactor is present in only some of thepatients studied that it is likely to be apredictor of outcome in that particulargroup – it might be expected to distin-guish those with and without mucosalischaemia and therefore those with worseand better outcome. If it is present in themajority of patients it would be lesslikely to be predictive.

This paper was not intended to givegrounds for abandoning investigation of



the P iCO2ÿPaCO2 gap, but we do notsee to what use it can be put until it isestablished that it has some clinical sig-nificance. Prior to acceptance of ourpaper only one other published studyhad examined the clinical significance ofP iCO2ÿ PaCO2 gap: the results of thisstudy are addressed in our paper. Sincethen the results of two small studies havebeen published. The first described thesignificance of P iCO2ÿPaCO2 gap in 20neonates being weaned from extracor-poreal circulation. The results indicatedthat in this highly specialised group ofpatients P iCO2ÿPaCO2 is of prognosticsignificance [7]. However, we were sur-prised that the authors of this paper onlychose to analyse the results of the singlereading taken during the weaning phasedespite taking regular measurementsfrom the time of initiation of extracor-poreal circulation. The second exam-ined the prognostic significance in 19young children with severe sepsis orseptic shock. In these patients P iCO2ÿPaCO2 gap was marginally higher (statis-tically significant) in nonsurvivors at24 h but not at 0, 12 or 48 h. Even at24 h the addition of P iCO2ÿPaCO2 gapto blood lactate and mean arterialpressure did not appreciably increasethe predictive power for death suggest-ing, as our data have, that its clinicalsignificance is limited [8]. In the era ofevidence-based medicine we feel it isinappropriate to abandon use of a newtechnique only when it is proved beyondall doubt to be of no value, rather webelieve that new techniques should beadopted only when there are data tosupport their use. At present we do notbelieve that there are good data to sup-port the use of P iCO2ÿ PaCO2 gap,except possibly in certain specialisedgroups.

C. D. GomersallG. M. JoyntPrince of Wales Hospital,Shatin, NT,Hong Kong,People’s Republic of China

References1 Schlichtig R, Mehta N, Gayowski TJP.

Tissue-arterial PCO2 difference is abetter marker of ischemia thanintramural pH (pHi) or arterial

pH-pHi difference. Journal of CriticalCare 1996; 11: 51–6.

2 Hanley JA, McNeil BJ. The meaningand use of the area under a receiveroperating characteristic (ROC) curve.Radiology 1982; 143: 29–36.

3 Friedman G, Berlot G, Kahn RJ,Vincent JL. Combined measurementsof blood lactate concentrations andgastric intramucosal pH in patientswith severe sepsis. Critical Care Medicine1995; 23: 1184–93.

4 Hotchkiss RS, Karl IE. Reevaluationof the role of cellular hypoxia andbioenergetic failure in sepsis. Journal ofthe American Medical Association 1992;267: 1503–10.

5 Gutierrez G, Wulf ME. Lactic acidosisin sepsis: a commentary. Intensive CareMedicine 1996; 22: 6–16.

6 VanderMeer TJ, Wang H, Fink MP.Endotoxemia causes ileal mucosalacidosis in the absence of mucosalhypoxia in a normodynamic porcinemodel of septic shock. Critical CareMedicine 1995; 23: 1217–26.

7 Duke T, Butt W, South M, Shann F,Royal Children’s Hospital ECMONursing Team. The DCO2 measuredby gastric tonometry predicts survivalin children receiving extracorporeallife support. Comparison with otherhemodynamic and biochemicalinformation. Chest 1997; 111: 174–9.

8 Duke TD, Butt W, South M.Predictors of mortality and multipleorgan failure in children with sepsis.Intensive Care Medicine 1997; 23:684–92.

Reshaping the Macintosh blade

We enjoyed the paper by Bucx et al. onthe biomechanics of a reshaped Macin-tosh blade (Anaesthesia 1997; 52: 662–7). This paper, as well as the previousones of these workers, greatly enhancesthe understanding of the biomechanicsof laryngoscopy. It is, however, not quitecomplete in references to blades similarto the one proposed by Bucx et al. asthey cite only that by Bizzarri and Giuf-frida [1].

‘Accessories for endotracheal anaes-thesia are without end. There is noliving anaesthetist who holds the distinc-tion of not having designed one ormore.’ [2]. Several curved blades with areduced flange are described [1, 3–9].Of these, we have tested the Bizzarri andGiuffrida [1], the Wiemers [8] and theRFQ-blades [9] (Fig. 1 below) in man-nequins [9, 10]. We found intubationtime and subjective handling with thevirtually flangeless Bizzarri–Giuffrida tobe significantly inferior compared withthe Macintosh blade [10]. Like Bucx et



Figure 1 The RFQ-blade is another example of a flange-reduced blade (RFQMedizintechnik, Tuttlingen, D-78532).

al., we observed inadequate mouth-opening with flangeless or flange-reducedblades as well as a tendency to lift thehead of the mannequin.

Maybe some head restraint and specialtraining are needed to make flange-reduced blades successful.

W. H. MaleckKlinikumLudwigshafen, Germany

K. P. KoetterLeopoldina-Hospital,Schweinfurt, Germany

References1 Bizzarri DV, Giuffrida JG. Improved

laryngoscope blade designed for easeof manipulation and reduction oftrauma. Anesthesia and Analgesia 1958;37: 231–2.

2 Sykes WS. The cheerful centenarian,or the founder of laryngoscopy. In:Essays on the first hundred years ofanaesthesia, Vol. II. Huntington:Robert E. Krieger, 1972; 95–111.

3 Bowen RA, Jackson I. A newlaryngoscope. Anaesthesia 1952; 7:254–6.

4 Callander CC, Thomas J.Modification of Macintoshlaryngoscope for difficult intubtion.Anaesthesia 1987; 42: 671–2.

5 Dance C. Blade for lateral intubation.Anesthesiology 1959; 20: 380–1.

6 Gabuya R, Orkin LR. Design andutility of a new curved laryngoscopeblade. Anesthesia and Analgesia 1959;38: 364–9.

7 Ibler M. Modification of Macintoshlaryngoscope blade. Anesthesiology1983; 58: 200.

8 Wiemers K. ModifizierterLaryngoskopspatel (FreiburgerModell). Zeitschrift fur PraktischeAnasthesia 1972; 7: 107–9.

9 Maleck WH, Kotter KP, Bauer M,Weber M, Herchet J, Petroianu GA.Geknickt oder aufrichtbar? FunfLaryngoskopspatel im Test.Anasthesiologie & Intensivmedizin 1996;37: 621–4.

10 Lenz M, Kotter K, Maleck W, PiperS, Triem J, Boldt J. Steglos odergekerbt? Vier Laryngoskopspatel imTest. Anasthesiologie, Intensivmedizin,

Notfallmedizin, Schmerztherapie 1997;32 (Suppl. 1): S181.

Non-rewireable plugs

The operating theatre environment hasbecome home to an increasing array ofelectrical equipment. The presence ofwires attached to an ever-growingnumber of plug sockets from whichflourish an abundance of infusiondevices, monitors and other equipmentmakes the operating theatre environ-ment a potentially hazardous place forboth patients and staff [1]. The two casesthat I report below involve problemswith non-rewireable electrical plugs.

The first incident occurred 20 mininto a surgical procedure in which themains supply failed to the anaestheticmachine. The machine was changedand connected to a different powersocket in an attempt to identify thesource of the problem. However, themains electric supply failed withinminutes and the Residual CurrentDevices [RCD] located outside theoperating theatre were tripped out. Itwas also noticed that immediately beforethe power supply failed, the display onone of the syringe drivers continued toshow ‘mains failure’. Following discon-nection of this particular syringe driverfrom the mains, the power supply to theanaesthetic machine and the rest of the

theatre remained intact. On closerinspection it was found that fluid hadingressed into the fuse holder of the plugof the syringe driver. The fluid had shortcircuited the fuse causing current to flowbetween live and the earth points on theplug which caused the RCD to trip.The 20-min delay before the circuitbreaker tripped after the faulty pumpwas switched on at the beginning of theoperation was most probably due to thetime it took to heat the fluid in the plug.The rise in temperature would have theeffect of lowering the electrical resis-tance of the fluid until a point wasreached where sufficient current flowedto trip the RCD.

In a separate incident a non-rewire-able plug had been dropped, the fuseholder had fallen out and the fuse incor-rectly re-inserted. The plug was theninserted into the electrical socket asshown in the figure. Because the fusewas incorrectly inserted, the electricalplate holding the socket provided theearth path and consequently currentflowed from the live fuse to earth. Thishad the effect of melting the fuse againstthe earth plate as shown in the figure.

Both of these incidents demonstratepotential problems with non-rewireableplugs. Whilst the second incident wasoriginally the fault of the person insert-ing the fuse, it would still not haveoccurred had the plug not been of thetype described above. Whilst these plugs



are a convenient means whereby fuses canbe changed easily and safely by nontrainedpersonnel, it has been demonstrated thattheir use can be associated with signifi-cant and potentially serious sequelae.

All plugs should conform to the Brit-ish Standards specification – BS 1363[specification for rewireable and non-rewireable plugs]. In some situations avariation of this specification, known asBS 1363/A [rough usage], is used to coverequipment which is used in more hostileconditions such as the operating theatre.According to a recent safety action notice[1, 2], a number of other incidents havebeen reported from outside this institu-tion concerning this particular type ofplug. The latest Safety Action Noticespecifies that non-rewireable plugsshould confirm to BS 1363/A specifica-tion, have a moulded one-piece bodyandelectrical connection pins should be solidand not of the folded sheet metal type.

It is now the policy of some manu-facturers to supply their syringe driverswith moulded non-rewireable plugs. Itmust be pointed out, however, that theseplugs, while being an improvement onthe nonmoulded counterparts, still havean access point for fluid to ingress. Therehave not been any reported incidentswith these plugs to date. It is currentlythe policy in this hospital to replace allnon-rewireable plugs, which do notmeet the recommendations of SafetyAction Notice [SAN] 97/10 [3], withwireable versions for safety reasons. Theminor inconvenience caused by havingto dismantle a plug, when changing thefuse, is a small price to pay if one is toavoid problems such as those describedabove. Although these incidents appearto be few and far between it is probablyfair to say that electrical safety problemsare generally under reported.

A. HindleHCI Medical Centre,Clydebank G81 4HX

References1 Safety Action Notice (SC) 96/48; 1996.

Detachable mains cords for use withmedical devices: safety considerations.

2 Safety Action Notice (SC) 96/51; 1996.Potential hazard for use of medicalelectrical equipment with non UKtype mains plugs.

3 Safety Action Notice (SC) 97/10 ; 1997.Suitability of electrical mains plugs forhigh use hospital applications.

Over reliance on pulse oximetry

We would like to report an incidentwhereby a pulse oximeter (HewlettPackard model number M11908) reada false high oxygen saturation which ledto a patient becoming markedly hypoxicthrough over reliance on this device.The lungs of a 41-year-old man suffer-ing from pneumocystis pneumonia werebeing ventilated on our ITU. Due toinadequate gas exchange, it was decidedto paralyse him with atracurium, where-upon his oxygen saturation by pulseoximetry rose from 92% to 100%.Over the following 90 min it was pos-sible to reduce the inspired oxygen from86% to 50% and all seemed to be goingwell, until his blood pressure fell drama-tically. It was noted that he had becomemarkedly cyanotic, although the pulseoximeter read 98% and showed a goodwaveform. Arterial blood gases revealeda PO2 of 4.09 kPa. This rapidly correctedonce inspired oxygen was increased andhe did not appear to suffer any adverseconsequences. The probe has beentested by our ‘in-house’ medical engin-eers, who could find no fault, and it wasreturned to Hewlett Packard for furtherassessment.

We could find only one report ofpulse oximetry giving a false high read-ing, in hypoxic children using an adultprobe; interestingly it was the samemodel Hewlett Packard probe as in ourcase. The incident we report occurred inan adult patient using the standard adultprobe. We would like to stress the samemessage, that electronic monitoringequipment is merely an aid, which isliable to error and should not replacevigilant clinical observation.

S. PickardS. FayekBirmingham Heartlands Hospital,Birmingham B9 5SS

A replyThe Hewlett-Packard M1190A AdultReusable SpO2 Sensor which wasreported as giving an incorrect reading

was returned to Hewlett-Packard asdefective. The incoming inspectionrevealed damage to the sensor housing,which could cause an incorrect readingdue to ambient or other light sourcesshining on the sensor. The sensor wasreplaced to the customer’s satisfactionand no further problems have beenreported. Hewlett Packard is continu-ously endeavouring to improve thestrength and reliability of its reusablesensors, but this case illustrates theimportance of regular inspection of allpulse oximetry sensors for signs of phy-sical damage.

B. WoodSupport quality Manager,Hewlett Packard,Bracknell,Berkshire RG12 1HN

The ‘non-empty’ bottlerevisited

I have recently been handed a boxful ofused isoflurane bottles by our operatingtheatre staff. It was noticed that a sig-nificant volume of fluid was left in thebottle when the vaporiser had been filledby the key filler method and I was askedhow we should dispose of these con-taminated bottles. Our pharmacy wasunable to help. The basic problem isthat the key filler system for isofluraneempties the bottle, but the filler tubedoes not empty, so that when the filler isdisconnected from the vaporiser, a fewmillilitres of fluid is returned to thebottle leading to wastage of drug andthus NHS money and possible pollution.Further enquiry among members of theSouth Thames (East) Region Anaes-thetic Specialty Committee about theirmanagement of this problem demon-strated a wide variation in practice inspite of its having been discussed incorrespondence in Anaesthesia in 1992and 1992 [1–3].

Over a period of 2 months in our mainoperating theatre unit, 250 ml of isoflur-ane or one bottle BNF price £97.50 isbeing wasted in this way and being sentoff for glass disposal. The quantity ofunused drug which must be wastedglobally given an annual production inexcess of 800 tonnes does not bear



thinking about. It seems to have simila-rities with Coleman’s mustard who weresaid to have made their money from themustard left on the side of the plate. Iapproached the manufacturer (Abbott);individual members of the companywere extremely interested and sympa-thetic, but a corporate solution was notforthcoming. As this correspondencecommenced, they increased the size oftheir bottles, so the proportion of iso-flurane wasted was reduced, but there isstill no plan to change the filler system toone similar to that for sevoflurane whichempties completely. We need some sortof solution to prevent the currentwastage; now.

We could copy the brewery industryand call it ullage and then demand thatAbbott, like a brewery, put this amountextra into the bottle and take back theincompletely emptied bottles. Not sur-prisingly, they have rejected this idea.The filler could be modified. Unfortu-nately, as I understand it, the BritishStandards and European CommitteeCEN have completed the standards forthe isoflurane filler and did not addressthe problem of the dregs, but standardsare not required for patented devices. Itseems unlikely that it will be economic-ally worthwhile for a manufacturer todesign a new filler. We probably have toaccept that somehow one bottle has tobe emptied into the next. I do not acceptthat this practice would make qualitycontrol difficult to enforce [3] becauseit is universal practice to mix differentbatches in one vaporiser already. Pro-vided that the bottles are kept in a lockeddrug cupboard, substitution of the con-tents should not be a problem, as openedbottles are frequently returned to thedrug cupboard following topping-up ofthe vaporisers at the beginning of a listalready. Dr Robin Loveday has sug-gested a filler with bottle ends atboth ends so one bottle could beemptied safely into the next (personalcommunication). This would seem to bethe best solution.

Meanwhile, to avoid waste, can wesolve this problem?

A. FergusonQueen Elizabeth the Queen MotherHospital,Margate, Kent CT9 4AN

References1 Wittmann PH, Wittmann FW,

Connor T, Connor J. The ‘Nonempty’empty Bottle. Anaesthesia 1992; 47:721–2.

2 Martin LVH. The ‘non-empty’ bottle.Anaesthesia 1993; 48: 94.

3 Connor T, Connor J, Wittmann FW.The ‘non-empty’ bottle. Anaesthesia1993; 48: 647.

Attenuation of pain oninjection of propofol – anunexpected benefit of co-induction with midazolam

Drs Driver et al. (Anaesthesia 1997;52: 698–700) once more highlightthe facilitation of insertion of thelaryngeal mask airway where a tech-nique of co-induction employing mid-azolam (0.04 mg.kgÿ1) 3 min prior toalfentanil and propofol is advocated. Iagree that ease of insertion is muchimproved and have used co-inductionroutinely since their first publication onthis subject [1].

Propofol frequently causes pain oninjection and to date the widespreadresponse to this problem is to mix 20–40 mg of lignocaine with 20 ml of theemulsion. In doing so, changes in thenature of the micelles result [2] and thepractitioner assumes end-user responsi-bility from the manufacturer. I soonbecame aware that the prior administra-tion of midazolam attenuated the painon injection produced by propofol. Thereason for this is unclear, but is notsimply due to amnesia of the event, as0.04 mg.kgÿ1 does not typically causesedation and patients only report slightdiscomfort (if at all) when asked. As aconsequence I have not found it neces-sary to add lignocaine to propofol for thepast 16 months, as pain on injectionwhen using midazolam for co-inductionis no longer a clinically significant prob-lem. Drs Driver et al. do not mentionthis incidental observation, but appearto have used unmodified propofol ineach of their studies.

J. W. MackenzieRoyal Berkshire Hospital,Reading RG1 5AN

References1 Driver I, Wiltshire S, Mills P,

Lillywhite N, Howard-Griffin R.Midazolam before induction improvesconditions for laryngeal maskinsertion. British Journal of Anaesthesia1995; 75: 664.

2 Lilley EMM, Isert PR, Carasso ML,Kennedy RA. The effect of theaddition of lignocaine on propofolemulsion stability. Anaesthesia 1996;51: 815–8.

Use of heparin in arterial lines

We read with interest the article by Heapet al. (Anaesthesia 1997; 52: 640–5) onthe effect of heparin in arterial lines oncoagulation studies performed on bloodsampled from such lines. While this isencouraging in demonstrating little clin-ical effect, provided at least 4.5 ml ofblood is withdrawn first, we questionwhether heparin is necessary at all.Stimulated by an article in the nursingliterature [1] we omitted heparin from10 consecutive arterial lines in the ITUand could find no observable difference.Following this, we have not put heparinin our arterial lines for the last 3 yearsand have found no reason to revert toadding heparin. Consequently the ques-tion of validity of coagulation studiesdoes not arise.

M. BlackmoreR. MaundrillN. G. LaviesWorthing Hospital,Worthing,West Sussex BN11 2DH

Reference1 Gamby A, Bennett J. A feasibility

study of the use of non-heparinised0.9% sodium chloride for transducedarterial and venous lines. Intensive andCritical Care Nursing 1995; 11:148–50.

Anaesthesia, anaesthetics andanaesthesiology

With regard to spelling: as long ago as1755 Dr Johnson in his famous diction-ary remarked: ‘AE or Æ. A diphthong of



very frequent use in the Latin language,which seems not properly to have anyplace in the English . . .’ This diphthongis an affectation and should be quietlyput to rest, as was that in ‘aether’, a formstill occasionally used when I was astudent.

R. I. BodmanGlenbrook,Cork, Ireland

Footnote: This correspondence is now closedEditor.



cricoid pressure in chaos

Documents