05/06/12 haemodynamic Monitoring and Management
Haemodynamic instabilityHow to recognize it?
Bern, May 11-12 2012Jan Bakker
05/06/12 haemodynamic Monitoring and Management
Department of Intensive CareErasmus MC University Medical Center
• 45+ ICU beds in a 1300 bed hospital
• Trauma center, Only center allowed to do all transplantaPons in adults and children
• ECMO center (last 4 mo: 22 paPents)• AdmiUng 2800 paPents in ICU and 1000 in PACU
• 70% Mechanical VenPlaPon
• APACHE II: 20±9• Research
– CirculaPon: 10 PhD students– VenPlaPon: 4 PhD students– Ethics/EOL: 2 PhD students
05/06/12 haemodynamic Monitoring and Management
DefiniPon of haemodynamic instability
• Acute deterioraPon in organ funcPon due to a inadequate organ perfusion/oxygenaPon– Perfusion Pressure
• S/D/M Arterial Pressure
– Perfusion• Global (cardiac output)• Regional (organs)
05/06/12 haemodynamic Monitoring and Management
RegulaPon of Cardiac Output/Flow• Healthy 25 year old donates one kidney• CO measured in stable condiPons pre-‐OP• Following recovery again CO measured in same condiPons
The Cardiac Output post donaPon will be • LOWER -‐ SIMILAR -‐ HIGHER
compared to the pre-‐OP Cardiac Output
05/06/12 haemodynamic Monitoring and Management
tion of cardiac output and arterial pressure, the impor-tance of providing an overall conceptual frameworkwhen discussing cardiovascularphysiologywith students.
CARDIAC OUTPUT REGULATION
Let’s begin by asking a very simple question.
Which of the following changes in cardiac outputwould you expect to find seven days after surgicalreduction of kidney mass by 50% (removal of onekidney): an increase, a decrease, or no change?
When I asked this question at the ExperimentalBiology ’99 symposium, most of the audience (whichincluded primarily physiologists) answered either ‘‘nochange’’ or ‘‘an increase,’’ even though the correctanswer is a decrease in cardiac output. Why is cardiacoutput reduced by removal of a kidney when therehas been no obvious effect on cardiac pumping abilityor heart rate? This is not easy to comprehend if onethinks in a ’cardiocentric’ manner and focuses on thewell-known formula learned by all physiologists: car-diac output ! stroke volume " heart rate.
Cardiac Output is the Sum of Tissue and Organ
Blood Flows
If I had shown Fig. 1 before asking this question, Idoubt that anyone would have had difficulty answer-
ing it. The reason for the decrease in cardiac outputafter unilateral nephrectomy is that, except formomen-tary imbalances, cardiac output is equal to venousreturn, which is equal to the sum of the flows of all ofthe individual tissues and organs. Removing onekidney decreases blood flowing back to the heart by!10% (assuming that total flow to both kidneys is!20% of the cardiac output). The same effect wouldbe observed with amputation of an arm or a leg orwith removal of any other tissue from the body. Thisexample illustrates that cardiac output is determinednot only by the function of the heart but also by theperipheral circulation. Except when the heart isseverely weakened and unable to adequately pumpthe venous return, cardiac output (total tissue bloodflow) is determined mainly by the metabolic needs ofthe tissues and organs of the body, although intrinsicand neurohumoral mechanisms allow the heart toeffectively accommodate changes in venous return.
This conceptual framework is very helpful in explain-ing changes in cardiac output that occur duringexercise (when metabolic activity and blood flow toskeletal muscles are increased), after eating a largemeal (which increases metabolic activity and bloodflow in the gastrointestinal system), and in many otherphysiological conditions. In each of these circum-stances, cardiac pumping ability plays a relatively
FIG. 1 .
Relationship between cardiac output and per ipheral blood flow regula-
tion. GI, gastrointestinal.
A P S R E F R E S H E R C O U R S E R E P O R T
V O L U ME 22 : N U MBER 1 – ADVANCES IN PHYSIOLOGY EDUCATION – D ECEMBER 1999S175
05/06/12 haemodynamic Monitoring and Management
CO2
LungsInspiration
Expiration
Pulmonary circulationMicrocirculation
Organ
CO2 production
O2 consumptionCirculation
CO2 flow
O2 flow
SV HR RecruitDilate Vt FQCO2 VO2
O2
05/06/12 haemodynamic Monitoring and Management
Oxygen Transport and Delivery
• TO2 = Hb x SaO2 x CO x ∁• DO2 = microcirculatory perfusion
– ConvecPon» heart failure
– Diffusion» sepsis, hemodiluPon
05/06/12 haemodynamic Monitoring and Management
Before ECC 10min after start ECC
same place sublingual
Atasever et al. J Cardiothorac Vasc Anesth 2011;13(6):573-577
05/06/12 haemodynamic Monitoring and Management
CharacterisPcs of haemodynamic instability
• Increased HR• Increase in Respiratory Rate• Abnormal Blood Pressure
– context sensiPve!!!!
05/06/12 haemodynamic Monitoring and Management
●●●●●●●●●● ●
●● ●
● ●● ●
● ●● ●
● ● ● ● ●● ●
AP
mm
Hg
Flow ml/min1 2 3 40 5 6 7
40
20
60
80
100
120
140
160
180
200
●
Normal
100
3.5
100 3.5Blood flowBlood pressure
Jan Bakker
●
●
● ●
05/06/12 haemodynamic Monitoring and Management
●
●●●●●●●●●● ●
●● ●
● ●● ●
● ●● ● ● ● ●
● ●
AP
mm
Hg
Flow ml/min1 2 3 40 5 6 7
40
20
60
80
100
120
140
160
180
200
●
Normal
100 3.5Blood flowBlood pressure
1.0
58 1.058
Jan Bakker
05/06/12 haemodynamic Monitoring and Management
●●●●●●●●●● ●
●● ●
● ●● ●
● ●● ●
● ● ● ● ●● ●
AP
mm
Hg
Flow ml/min1 2 3 40 5 6 7
40
20
60
80
100
120
140
160
180
200
●
Normal
100 3.5Blood flowBlood pressure
● 1.058●●
●●●●●●●
●●
●●
●●
●
● ●● ●
● ●● ●
● ● ●● ● ●
● Sympathetic stimulation
Jan Bakker
●
●
●●
05/06/12 haemodynamic Monitoring and Management
●●●●●
●●●● ●
●● ●
● ●● ●
● ●● ●
● ● ● ● ●● ●
AP
mm
Hg
Flow ml/min1 2 3 40 5 6 7
40
20
60
80
100
120
140
160
180
200
●
Normal
100 3.5Blood flowBlood pressure
1.058●●
●●●●●●●
●●
●●
●●
●
● ●● ●
● ●● ●
● ● ●● ● ●
● Sympathetic stimulation
●●
1.0
1.0
108
108
Jan Bakker
●
05/06/12 haemodynamic Monitoring and Management
●●●●●
●●●● ●
●● ●
● ●● ●
● ●● ●
● ● ● ● ●● ●
AP
mm
Hg
Flow ml/min1 2 3 40 5 6 7
40
20
60
80
100
120
140
160
180
200
●
Normal
100 3.5Blood flowBlood pressure
●●
●●●●●●●
●●
●●
●●
●
● ●● ●
● ●● ●
● ● ●● ● ●
● Sympathetic stimulation
● ●
1.0108
Jan Bakker
●●●●●●●●●●●●●●
●● ● ●
● ● ● ●● ● ● ● ●
●●●
●
05/06/12 haemodynamic Monitoring and Management
●●●●●
●●●● ●
●● ●
●● ●
● ●● ●
● ● ● ● ●● ●
AP
mm
Hg
Flow ml/min1 2 3 40 5 6 7
40
20
60
80
100
120
140
160
180
200
Normal
100 3.5Blood flowBlood pressure
●●
●●●●●●●
●●
●●
●●
●
● ●● ●
● ●● ●
● ● ●● ● ●
● Sympathetic stimulation
1.0108
Loss of sympathetic tone●
1.0 38
38
Jan Bakker
●
)LJXUH�&OLFN�KHUH�WR�GRZQORDG�KLJK�UHVROXWLRQ�LPDJH
05/06/12 haemodynamic Monitoring and Management
05/06/12 haemodynamic Monitoring and Management
• Young male• double femur #• HR: 120 /min• BP: 130/90• Normal consciousness
What do you want to know, what do you look for?
05/06/12 haemodynamic Monitoring and Management
Haemodynamic instability: vasoconstricPon
05/06/12 haemodynamic Monitoring and Management
Clinical assessmentCRT -‐ SubjecPve Skin temperature
Pressure is applied for 5 sec - nail bed turns white
Champion et al. Crit Care Med 1981;9:672-‐676 -‐ Schriger et al. Ann Emerg Med 1988;17:932-‐935
normal ≤ 2 sec in children and young adults 4.5 sec in older patients
put your hand on the patient and assess temperature (normal, abnormal)
05/06/12 haemodynamic Monitoring and Management
Skin temperature to iden@fy clinical hypo-‐perfusion in the cri@cally ill
Kaplan et al. J Trauma 2001;50:620-‐628
– Cool vs warm skin• similar HR, BP, PAOP, Hb, FiO2, PaCO2 and PaO2
Cool
2.9 ± 1.2
7.32 ± 0.2
60 ± 4
4.7 ± 1.5
Cardiac Index
Arterial pH
SvO2
Lactate
Warm
4.3 ± 1.2 *
7.39 ± 0.07 *
68 ± 8 *
2.2 ± 1.6 *
05/06/12 haemodynamic Monitoring and Management
One of the women, the 23-year-old primipara Eva Rumpel, gave birth to a healthy child on 9 January 1843. The same night she developed a painfully swollen abdomen and became ill, feverish, and sweaty, with rapid pulse and severe thirst. The initiated treatment was bloodletting and clystering. The next evening she deter iorated, became delirious, with anxious breathing, a tense abdomen, cold extremities and r a p i d p u l s e , fi n a l l y l o s i n g consciousness . Again, bloodletting followed. At 4:30 a.m., 36 h after the onset of the first symptoms, she died. During autopsy, severe purulent endometritis, vaginal pus, pulmonary edema, and shock liver and shock spleen were found.
Fig. 1 Title page of Scherer’s 1843 book
the medical faculty in 1842, professor of organic chem-istry in 1847, and later professor of general, anorganic andpharmaceutical chemistry. His work especially concernedquantitative research on blood and urine in pathologicalconditions. In 1843 he published his book ‘Chemische undMikroskopische Untersuchungen zur Pathologie angestelltan den Kliniken des Julius-Hospitales zu Würzburg’(Chemical and microscopic investigations of pathologycarried out at the Julius Clinic at Würzburg) [15] (Fig. 1),in which he described 72 case reports, giving details onclinical course, diagnosis, and results obtained duringautopsy and analysis of body fluids. Scherer died on 17February 1869 [18].
The 1843 casesIn one chapter in his 1843 book entitled ’Untersuchungenvon krankhaften Stoffen bei der im Winter 1842–1843 inWürzburg und der Umgegend herrschenden Puerperal-Fieber-Epidemie’ (Investigations of pathological sub-
stances obtained during the epidemic of puerperal feverwhich occurred in the winter of 1842–1843 in and aroundWürzburg) Scherer described the cases of seven youngwomen who all died peripartum.
One of the women, the 23-year-old primipara EvaRumpel, gave birth to a healthy child on 9 January1843. The same night she developed a painfully swollenabdomen and became ill, feverish, and sweaty, withrapid pulse and severe thirst. The initiated treatment wasbloodletting and clystering. The next evening she deteri-orated, became delirious, with anxious breathing, a tenseabdomen, cold extremities and rapid pulse, finally losingconsciousness. Again, bloodletting followed. At 4:30a.m., 36 h after the onset of the first symptoms, she died.During autopsy, severe purulent endometritis, vaginal pus,pulmonary oedema, and shock liver and shock spleenwere found. The blood that was obtained directly from theheart was chemically analysed, in which lactic acid wasfound. Most likely this unfortunate woman had died froma fulminant septic shock caused by group A haemolyticstreptococci (Streptococcus pyogenes). Scherer diagnosedthis case as perimetritis with secondary peritonitis.
Another patient, the 28-year-old, 7 months pregnant(second pregnancy) Margaretha Glück, was, after beingicteric, nauseous, vomiting and complaining about epigas-tric pain for 8 days, admitted to the lying-in birth clinicon 6 February 1843. Four days later she was transferredto the hospital with severe nosebleeds and generalised ex-anthema or purpura. In the evening she suffered from se-vere gastric bleeding and epistaxis, showing rapid pulse,cold extremities and dizziness. The next morning, she wastransferred back to the birth clinic, where she gave birthto a premature child (30 weeks) and suffered from a severepost-partum fluxus. She was again transferred to the hospi-tal with the following symptoms: cold clammy skin, tachy-cardia, severe lochia and persistent exanthema or purpura,but without signs of an acute abdomen. During the night ofFebruary 11, she became aphasic and restless, followed bychills and profound sweating. On the morning of Febru-ary 13, she further deteriorated and bilirubinuria was de-tected. The next day she was comatose, finally developedrattling breathing and convulsions. Death occurred duringthe following night. Autopsy revealed a small intracerebralhaematoma, normal lungs without pulmonary oedema, as-cites and an anaemic, foul smelling uterus filled with puru-lent and decayed tissue and pus. Blood was also obtaineddirectly from the heart during autopsy and lactic acid wasfound.
In this case we could think of a haemorrhagic shockand cerebral haemorrhage due to clotting disorderspossibly resulting from either acute fatty liver of preg-nancy/HELLP syndrome, idiopathic thrombocytopenicpurpura, thrombotic microangiopathy (TTP/HUS) orDIC. The case was most likely complicated by a sepsis(endometritis). Scherer himself diagnosed this case asseptic endometritis.
05/06/12 haemodynamic Monitoring and Management
DefiniPon of haemodynamic instability
• what windows do we have?– Brain
• consciousness
– Kidney• urine output
– Periphery• color, temperature
Vincent, Ince, Bakker (Submitted)
uration (SpO2). We measured capillary refilltime by applying firm pressure to the distalphalanx of the index finger for 5 secs andrecording the time for return of the normalcolor by using a conventional wristwatch. PFIand SpO2 were measured by using the Viridia/56S monitor (Philips Medical Systems). TheViridia system calculates the PFI as the ratiobetween the pulsatile component and the non-pulsatile component of the light reaching thelight-sensitive cell of the pulse oximetryprobe.
Group 2. The measurements included PFI,SpO2, ambient temperature, central tempera-ture, great toe temperature, finger tempera-ture, capillary refill time, and hemodynamicvariables including heart rate and mean arte-rial pressure. The central temperature wasmeasured by using either a pulmonary arterycatheter or a rectal probe. The peripheral tem-perature was measured on the ventral face ofthe great toe with a temperature probe (Phil-ips Medical Systems 21078A). The finger tem-perature was measured simultaneously withthe PFI measurement on the same finger byusing a similar probe. The central-to-toe tem-perature difference was calculated, and a dif-ference up to 7°C was considered normal (9).The doses of vasoactive drugs were recorded.Poor peripheral perfusion was defined as acapillary refill time !2 secs or a central-to-toetemperature difference !7°C.
Protocol
To evaluate the variation of the PFI inhealthy volunteers (group 1), measurementswere taken in the hospital restaurant before andafter their normal lunch after a 5- to 10-minrest. Volunteers were seated and instructed tokeep their hands still on the table to avoid mo-tion artifacts and to have the hands at the level ofthe heart. A questionnaire was used to collectinformation about history of smoking and vas-cular disease (diabetes, hypertension). In group2, two measurements were taken from each pa-tient. The first measurement was taken whenperipheral perfusion was abnormal; the secondmeasurement was taken when the peripheralperfusion profile had normalized. Patients withcentral hypothermia (core temperature "36°C)and limb ischemia attributable to vascular oc-clusion were excluded.
Statistical Analysis
Data are presented as mean # SD and me-dians with the 25th and 75th percentiles un-less otherwise indicated. Differences betweengroups or within groups were assessed by us-ing the Mann-Whitney test for nonparametricdata. Pearson’s correlation index was calcu-lated where applicable. We considered p " .05to be statistically significant. Statistical anal-yses were conducted with Statistical Packagefor the Social Sciences version 9.0 (SPSS, Chi-cago, IL).
Informed Consent
The Institutional Review Board waived theneed for written informed consent from thehealthy volunteers. Informed consent was ob-tained from the relatives of the patients.
RESULTS
Group 1. One hundred and eighthealthy volunteers were included, and atotal of 216 measurements were made.The distribution of age in the healthyvolunteers was normal: skewness, 0.06;median, 36 yrs (inner quartile range,
30–45 yrs). The distribution of PFI wasskewed (Fig. 1, Table 1).
Descriptive analysis showed no signifi-cant difference between variance and skew-ness of the measurements before and afterthe meal (Table 1). Also, no significant dif-ferences were found between smokers (n $26) and nonsmokers (n $ 82), as well asbetween volunteers with (n $ 11) or with-out (n $ 97) vascular disease (diabetes,hypertension). All volunteers had a normalcapillary refill time and arterial oxygen sat-uration (96% to 100%).
Group 2. A total of 74 measurementswere carried out in the 37 patients stud-ied. Descriptive statistics revealed a meanPFI of 2.2 # 0.22 with a median of 1.8(inner quartile range, 0.5–3.2). Table 2summarizes hemodynamic data duringabnormal peripheral perfusion and nor-mal peripheral perfusion, as well as themean doses of vasoactive drugs. No sig-nificant relationship between core tem-perature and PFI or core-to-toe tempera-ture difference was found. A significantexponential relationship between PFI andthe core-to-toe temperature differencewas found (R2 $ .52, p " .001; Fig. 2).
We found a significant linear correla-tion between changes in PFI and changesin the core-to-toe temperature difference(R2 $ .52, p " .001; Fig. 3).
In all cases, a concordant change in
Figure 1. Frequency distribution of all 216 pe-ripheral perfusion index values in the normalvolunteers. Line represents normal distribution.
Table 1. Descriptive statistics of peripheral perfusion index (PFI) measurements in healthy volunteers(group 1)
PFIAll Measurements
(n $ 216)Before Meal(n $ 108)
After Meal(n $ 108) p Value
Mean 2.2 # 2.0 2.2 # 2.0 2.2 # 1.9 NSMedian (IQR) 1.4 (0.7–3.0) 1.4 (0.6–3.2) 1.5 (0.8–2.8) NSP5–95 0.3–6.0 0.3–6.0 0.4–6.3Skewness 1.61 # 2.44 1.59 # 2.42 1.64 # 2.42 NSVariance 3.84 4.12 3.59 NS
IQR, inner quartile range; P5–95, 5th and 95th percentiles; NS, not significant.
Table 2. Hemodynamics, variables of peripheral perfusion, and vasoactive medication during abnormaland normal peripheral perfusion in the 37 patients
T1 T2 p
Core temperature, °C 37.5 # 0.2 37.5 # 0.1 NSHeart rate, bpm 92 # 3 91 # 3 NSMean arterial pressure, mm Hg 81 # 3 78 # 2 NSSpO2, % 97 # 1 96 # 0 NSCore-to-toe temperature difference, °C 9.0 # 0.45 4.9 # 0.3 ".001Perfusion index 0.7 # 0.1 3.6 # 0.3 ".001Dopamine, %g/kg!min 2.5 # 0.83 1.4 # 0.44 NSDobutamine, %g/kg!min 2.9 # 1.1 2.1 # 0.5 NSNoradrenaline, %g/kg!min 0.09 # 0.03 0.05 # 0.02 NS
T1, condition of abnormal peripheral perfusion; T2, condition of normal peripheral perfusion;SpO2, arterial oxygen saturation; NS, not significant.
1211Crit Care Med 2002 Vol. 30, No. 6
haemodynamic Monitoring and ManagementCrit Care Med 2002;30:1210-‐1213
Use of a peripheral perfusion index derived from the pulseoximetry signal as a noninvasive indicator of perfusion
Alexandre Pinto Lima, MD; Peter Beelen, RN; Jan Bakker, MD, PhD
Early recognition of impairedorgan perfusion is importantto avoid tissue hypoxia thatultimately could lead to organ
failure. During circulatory shock, skinblood flow decreases to preserve vital or-gan perfusion. This results in the clinicalsigns of poor peripheral perfusion, suchas a cold, pale, clammy, and mottled skin(1). Indexes of peripheral perfusion thushave been used to identify inadequateperfusion in critically ill patients (2–4).Peripheral perfusion can be assessed from
clinical signs (1), from the central-to-toetemperature difference (2, 3, 5), or withtechniques such as laser Doppler and cap-illary microscopy (6). Recently, the pulseoximetry signal has been suggested toreflect changes in peripheral perfusion(7). In addition, the ratio between thepulsatile and nonpulsatile component ofthe pulse oximetry signal has been re-lated to peripheral perfusion (8). Becausea pulse oximeter is universally availablein the operating room and intensive careunit, this ratio could be used to monitorperfusion in these circumstances.
Although the manufacturer reportsthe lower and upper limit of normal to be0.3 and 10.0, respectively, the variation innormal subjects and the clinical applica-tion of this ratio as an index of peripheralperfusion in critically ill patients have notyet been studied. The objective of the
current study, therefore, was to assess thevariation of this perfusion index inhealthy adults and study the relationshipbetween the peripheral perfusion index(PFI) and clinical signs of poor peripheralperfusion in critically ill patients.
METHODS
Participants
The study was conducted at a university-affiliated teaching hospital. Group 1 consistedof 108 healthy adult volunteers (mean age, 30! 9 yrs). Group 2 consisted of 37 critically illpatients (mean age, 70 ! 13 yrs) admitted tothe medical/surgical intensive care unit.
Measurements
Group 1. The measurements included cap-illary refill time, PFI, and arterial oxygen sat-
From the Department of Intensive Care Gelre Hos-pital, Apeldoorn, The Netherlands.
Address requests for reprints to: Jan Bakker, MD,PhD, Isala Clinics, Department of Intensive CareWeezenlanden, PO Box 10500, 8000 GM Zwolle, TheNetherlands. E-mail: [email protected]
Copyright © 2002 by Lippincott Williams & Wilkins
Objective: Peripheral perfusion in critically ill patients fre-quently is assessed by use of clinical signs. Recently, the pulseoximetry signal has been suggested to reflect changes in periph-eral perfusion. A peripheral perfusion index based on analysis ofthe pulse oximetry signal has been implemented in monitoringsystems as an index of peripheral perfusion. No data on thevariation of this index in the normal population are available, andclinical application of this variable in critically ill patients has notbeen reported. We therefore studied the variation of the peripheralperfusion index in healthy adults and related it to the central-to-toe temperature difference and capillary refill time in critically illpatients after changes in clinical signs of peripheral perfusion.
Design: Prospective study.Setting: University-affiliated teaching hospital.Patients: One hundred eight healthy adult volunteers and 37
adult critically ill patients.Interventions: None.Measurements and Main Results: Capillary refill time, periph-
eral perfusion index, and arterial oxygen saturation were mea-sured in healthy adults (group 1). Capillary refill time, peripheralperfusion index, arterial oxygen saturation, central-to-toe tem-perature difference, and hemodynamic variables were measuredin critically ill patients (group 2) during different peripheral per-fusion profiles. Poor peripheral perfusion was defined as a cap-illary refill time >2 secs and central-to-toe temperature differ-
ence >7°C. Peripheral perfusion index and arterial oxygensaturation were measured by using the Philips Medical SystemsViridia/56S monitor. In group 1, measurements were made beforeand after a meal. In group 2, two measurements were made, withthe second measurement taken when the peripheral perfusionprofile had changed. A total of 216 measurements were carriedout in group 1. The distribution of the peripheral perfusion indexwas skewed and values ranged from 0.3 to 10.0, median 1.4 (innerquartile range, 0.7–3.0). Seventy-four measurements were carriedout in group 2. A significant correlation between the peripheralperfusion index and the core-to-toe temperature difference wasfound (R2 ! .52; p < .001). A cutoff peripheral perfusion indexvalue of 1.4 (calculated by constructing a receiver operatingcharacteristic curve) best reflected the presence of poor periph-eral perfusion in critically ill patients. Changes in peripheralperfusion index and changes in core-to-toe temperature differ-ence correlated significantly (R2 ! .52, p < .001).
Conclusions: The peripheral perfusion index distribution in thenormal population is highly skewed. Changes in the peripheralperfusion index reflect changes in the core-to-toe temperaturedifference. Therefore, peripheral perfusion index measurementscan be used to monitor peripheral perfusion in critically ill pa-tients. (Crit Care Med 2002; 30:1210–1213)
KEY WORDS: skin temperature; peripheral perfusion; shock; he-modynamics; monitoring; central temperature
1210 Crit Care Med 2002 Vol. 30, No. 6
0.3 -‐ 10median: 1.4 (IQR: 0.7 -‐ 3.0)
PFI and core-to-toe temperature differ-ence was found. No significant relation-ship was found between mean arterialpressures, dose of vasoactive agents, andPFI or between changes in these variablesand changes in PFI.
In 16 patients, cardiac output wasmeasured. No significant relationshipwas found between changes in cardiacoutput and changes in either core-to-toetemperature difference or PFI.
We assessed the ability of the PFI toindicate an abnormal peripheral perfusion,as reflected by an abnormal core-to-toetemperature difference by constructinga receiver operating characteristic curve. APFI of 1.4 discriminated best between anormal and abnormal core-to-toe temper-ature difference in these critically ill pa-tients (area under the curve, 0.91; 95%confidence interval, 0.84–0.98). Table 3 re-ports the corresponding sensitivity, speci-ficity, and likelihood ratios.
DISCUSSION
We studied whether a perfusion indexcalculated from the pulse oximetry sig-nal, and available on-line in some moni-toring systems, can reflect clinical signsof decreased peripheral perfusion (capil-lary refill time and central-to-toe temper-ature difference) in critically ill patients.Because no data were available on normalvalues for this perfusion index, we alsostudied the variation of this variable inhealthy individuals. We show that a PFIof 1.4 can be used to detect abnormalperipheral perfusion in critically ill pa-tients, corresponding with the medianvalue found in the healthy volunteers. Inaddition, changes in this perfusion indexadequately reflect changes in clinicalsigns of peripheral perfusion and thuscan be used to assess effect of therapeuticinterventions on peripheral perfusion.
During circulatory failure associatedwith hypovolemia and low cardiac out-put, redistribution of blood flow causedby increased vasoconstriction results indecreased perfusion of the skin (1).Therefore, in critically ill patients, skinperfusion frequently is used to assess ad-equacy of global blood flow. Clinical signsof poor skin perfusion consist of a cold,pale, clammy, and mottled skin. Recently,techniques have become available tomeasure perfusion of the skin. LaserDoppler flow measurements and capillarymicroscopy (6) can adequately quantifychanges in capillary blood flow but arenot readily available in the emergencydepartment or intensive care unit.
When blood supply to the skin de-creases, the temperature of the skin alsodecreases. Therefore, measurements of skintemperature have been used to indicate de-creases in skin blood flow as a marker ofvasoconstriction and poor oxygen delivery(3, 2). Also, peripheral skin temperaturehas been advocated as a marker of the se-verity of shock (4). In addition, becausevasoconstriction of the skin reduces bodyheat loss, the difference between the core
temperature and skin temperature may in-crease. The central-to-toe temperature dif-ference therefore has been used to diagnoseand treat patients with global blood flowabnormalities (3, 5). To have this parame-ter of peripheral perfusion available online,at least two temperature probes are neces-sary, and the skin temperature probeshould be carefully affixed. These require-ments may limit the use of these variablesin emergency situations and clinically un-stable patients.
Pulse oximetry is a monitoring tech-nique used in almost every trauma andcritically ill patient. Monitoring of pulseoximetry during surgery is mandatory inmany countries. The principle of thepulse oximetry is the difference in absor-bance of light with different wavelengths(660 and 940 nm) by oxygenated hemo-globin. Other tissues, such as connectivetissue, bone, and venous blood, also ab-sorb light and thus affect the resultingsignal. However, whereas the arterialcomponent of the signal is pulsatile, theabsorption of light by other tissues isfairly constant. So, to have a proper esti-mate of the arterial oxygen saturation ofthe hemoglobin, the pulse oximetry hasto distinguish the pulsatile componentfrom the nonpulsatile component, wherethe pulsatile component is used subse-quently to calculate the arterial oxygensaturation (10, 11). When the signal isweak, for example, during vasoconstric-tion, the pulse oximetry signal requiresamplification up to !109 (10). Althoughanalysis of the pulse oximeter waveformhas been used to assess the volume statusof patients during major surgery (7), theamplification necessary during a low sig-nal (vasoconstriction, hypovolemia)could limit its clinical application in crit-ically ill patients. The perfusion index,used in this study, is calculated as theratio between the pulsatile and the nonpul-satile component of the light reaching thedetector of the pulse oximeter. When pe-ripheral hypoperfusion exists, the pulsatile
Figure 2. Relationship between peripheral perfu-sion index (PFI) and core-to-toe temperature dif-ference in all 74 measurements in the 37 patientsstudied. Displayed is the best fit curve (logarith-mic) R2 " .52, p # .001. Reference lines are themedian PFI of healthy volunteers and the refer-ence for an abnormal core-to-toe temperaturedifference.
Figure 3. Relationship between changes in pe-ripheral perfusion index (PFI) and changes incore-to-toe temperature difference. Displayed arethe linear regression and the correlation coeffi-cient. dPerfusion index, PFI during abnormalperfusion $ PFI during normal perfusion; dCore-Toe temperature, temperature difference duringabnormal perfusion $ temperature differenceduring normal perfusion.
Table 3. Use of peripheral perfusion index (PFI) as a measure of abnormal core-to-toe (C-T) temper-ature difference, an abnormal capillary refill time, or either
C-T Refill Either
Sensitivity, % 81 84 86Specificity, % 86 88 100Likelyhood ratio after positive test 6.0 7.1 —Likelyhood ratio after negative test 0.19 0.18 0.14
Normal PFI was defined as a PFI ! 1.4. Definition of abnormal C-T temperature difference andabnormal capillary refill time: see text.
1212 Crit Care Med 2002 Vol. 30, No. 6
05/06/12
haemodynamic Monitoring and Management
‣ 50 criPcally ill paPents following iniPal resuscitaPon and stabilizaPon (within first 24h of admission)
‣ Abnormal peripheral circulaPon was defined as
• increase in capillary refill Pme (> 4.5 sec) or cool skin (subjecPve)
‣ Measurements: Forearm-‐Finger Skin temperature difference, Central-‐Toe temperature difference, Peripheral Perfusion Index, SOFA score
05/06/12
haemodynamic Monitoring and Management
Clinical significance
‣ Odds for increase in SOFA score during first 48h of admission are 7.4 Pmes higher (CI: 2-‐19, P<0.05) in paPents with abnormal peripheral perfusion
‣ Odds to have increased lactate levels following iniPal resuscitaPon are 4.6 Pmes higher (CI: 1.4-‐15, P<0.05) in paPents with abnormal peripheral perfusion
05/06/12 haemodynamic Monitoring and Management
• Young male• double femur #• HR: 120 /min• BP: 130/90• Normal consciousness
The pa@ent will die 20 minutes later
You can do only one or two things, what will you do?
Intensive Care MedDOI 10.1007/s00134-007-0679-y E D I T O R I A L
Jan BakkerTim C. Jansen Don’t take vitals, take a lactate
Received: 19 March 2007Accepted: 4 April 2007
© Springer-Verlag 2007
This editorial refers to the article available at:http://dx.doi.org/10.1007/s00134-007-0680-5.
J. Bakker (!) · T. C. JansenErasmus MC University Medical Centre, Department of IntensiveCare, Room Hs320,P.O. Box 2040, 3000 CA Rotterdam, The Netherlandse-mail: [email protected]
The resident internal medicine called from the EmergencyDepartment (ED). “Can you please come and see mypatient, I think he is becoming septic and needs admissionto intensive care”. In the ED we found a confused olderpatient with an oxygen mask who was clearly dyspnoeic,the urinary catheter was filled with a dark brown fluid, thecollecting bag was empty. The resident reported that headmitted the patient 4 h earlier as he suspected pneumonia.On admission the patient was hypoxic but this clearlyimproved with the supplemental oxygen. The resident wasstill waiting for all the laboratory results and the chestX-ray. However, now that the patient had developedhypotension he thought the patient was clearly at riskand intensive care admission was required. When weasked why he had not called us earlier, he replied that heintended to admit the patient to the general ward as he washaemodynamically stable and oxygenation had improvedon supplemental oxygen so intensive care admissionwas not required. When reviewing the blood samplethat was drawn 30 min following presentation, besideshypoxaemia, an increased lactate level of 4.6 mmol/l waspresent. The resident pointed out that hyperlactataemia insepsis is not related to tissue hypoxia but rather is a markerof increased aerobic metabolism. Therefore he thought
there was no need to react to this hyperlactataemia. Wheredid this resident go wrong?
Increased blood lactate levels in critically ill patientsare generally associated with increased morbidity andmortality [1, 2]. Even haemodynamically stable patientswith raised lactate levels, a condition referred to ascompensated shock, are at increased risk of dying [3, 4].This not only applies to patients admitted to the intensivecare unit; also early in the course of illness, increasedblood lactate levels are related to increased morbidity andmortality. In a study published in this issue of IntensiveCare Medicine, Howell et al. [5] evaluated the prognosticvalue of one single venous lactate measurement shortlyafter admission to the ED in patients with clinicallysuspected infection. Their study is a follow-up on a pre-liminary report [6], where they did not take into accountpossible confounding factors such as co-morbidities andvital signs. In the current prospective observational cohortstudy (n = 1,287), the authors constructed a multivariatemodel, controlling for age, blood pressure, presenceof malignancy, platelet count and blood urea nitrogenlevel. They showed that venous lactate predicted 28-day in-hospital mortality. The predictive power of thelactate level was independent of blood pressure and co-variates. In patients with normal blood pressure, increasedblood lactate levels (>4.0 mmol/l) were associated witha ten times higher mortality rate than normal lactatelevels (mortality 26.5%). Others have reported simi-lar results in other patient populations. Lavery et al. [7]measured venous lactate within 10 min following admis-sion to the ED in 375 trauma patients. This study showedthat an increased lactate level (> 2.0 mmol/l) was a betterpredictor of morbidity and mortality than physiologicaltriage criteria (composed of heart rate, blood pressure,Glasgow coma scale and respiratory rate). Rivers et al. [8]also showed that traditional physiological variables did notadequately determine septic patients at risk of increased
05/06/12 haemodynamic Monitoring and Management
Lactate on admission: 13 mmol/L
• Young male• double femur #• HR: 120 /min• BP: 130/90• Normal consciousness
05/06/12 haemodynamic Monitoring and Management
Causes of hemodynamic instabilityPipes -‐ Pump -‐ Volume
WEIL, M. H., & SHUBIN, H. (1971). Advances in experimental medicine and biology, 23(0), 13–23.
ConclusionsNever deny your gut feeling
05/06/12 haemodynamic Monitoring and Management
• Hemodynamic instability refers to a condiPon usually involving inadequate perfusion but at the end its about the Pipes, the Pump and the Volume
• Global parameters maybe misleading• Only 2 real windows in clinical judgement• Don’t forget the signs of the periphery• Never forget lactate