Adding vitamin C to hydrocortisone lacks benefit in septic shock: a historical cohort study.

PURPOSE: Sepsis has high incidence and mortality rates, particularly in the intensive care unit (ICU). Corticosteroids may improve outcomes, and vitamin C may add benefit. We aimed to assess whether vitamin C and corticosteroids improved outcomes compared with corticosteroids alone. METHODS: This historical cohort study (11 December 2016 to 21 February 2018) was conducted in the ICU of a quaternary referral hospital. Patients with an ICU admission diagnosis of sepsis or septic shock who received vitamin C and hydrocortisone within 72 hr were compared with those who received only hydrocortisone. All patients received standard sepsis care including source control, antibiotics, and fluid resuscitation. Most patients received thiamine as standard ICU care. The primary outcome was hospital mortality. Secondary outcomes included ICU mortality, ventilator-free days, vasopressor-free days, dialysis use, and duration of ICU admission. RESULTS: One hundred and forty-four patients were included in the study. The mean (standard deviation [SD]) age was 64 (15) yr; 39% were female; and the mean (SD) Acute Physiology And Chronic Health Evaluation IV score was 89 (30). Eighty-eight patients did not receive vitamin C and 52 received vitamin C. There was no observed difference in hospital mortality between the non-vitamin C (36%) and vitamin C (39%) groups (adjusted odds ratio for hospital death, 0.52; 95% confidence interval, 0.20 to 1.34; P = 0.18). There were no statistically significant differences in any secondary outcomes. CONCLUSION: In this small observational study of ICU patients with septic shock, the addition of vitamin C to hydrocortisone therapy did significantly affect hospital mortality or other measures of mortality or organ dysfunction.

RéSUMé: OBJECTIF: Le sepsis comporte une incidence et des taux de mortalité élevés, particulièrement à l'unité de soins intensifs (USI). Les corticostéroïdes pourraient améliorer les pronostics, et la vitamine C pourrait être bénéfique. Notre objectif était d'évaluer si la vitamine C et les corticostéroïdes amélioraient les devenirs par rapport à un traitement de corticostéroïdes seulement. MéTHODE: Cette étude de cohorte historique (réalisée entre le 11 décembre 2016 et le 21 février 2018) a été réalisée à l'USI d'un hôpital quaternaire. Les patients ayant un diagnostic de sepsis ou de choc septique lors de leur admission à l'USI et ayant reçu de la vitamine C et de l'hydrocortisone dans les premières 72 heures ont été comparés à ceux n'ayant reçu que de l'hydrocortisone. Tous les patients ont reçu des soins standard pour le sepsis, soit un contrôle de la source de l'infection, un traitement antibiotique et une réanimation liquidienne. La plupart des patients ont reçu de la thiamine, un traitement standard à l'USI. Le critère d'évaluation principal était la mortalité hospitalière. Les critères d'évaluation secondaires comprenaient la mortalité à l'USI, les jours sans respirateur, les jours sans vasopresseurs, le recours à la dialyse et la durée de séjour à l'USI. RéSULTATS: Cent quarante-quatre patients ont été inclus dans notre étude. L'âge moyen (écart type [ÉT]) était de 64 (15) ans; 39 % étaient de sexe féminin; et le score APACHE IV moyen (ÉT) de 89 (30). Quatre-vingt-huit patients n'ont pas reçu de vitamine C et 52 en ont reçu. Aucune différence n'a été observée en matière de mortalité hospitalière entre les groupes sans vitamine C (36 %) ou avec vitamine C (39 %) (rapport de cotes ajusté pour la mortalité hospitalière, 0,52; intervalle de confiance 95 %, 0,20 à 1,34; P = 0,18). Il n'y a eu aucune différence statistiquement significative en ce qui touchait aux critères d'évaluation secondaires. CONCLUSION: Dans cette petite étude observationnelle portant sur des patients de l'USI en choc septique, l'ajout de vitamine C à un traitement d'hydrocortisone n'a pas eu d'impact significatif sur la mortalité hospitalière ou les autres mesures de mortalité ou d'atteintes organiques.

Association between intensive care unit occupancy at discharge, afterhours discharges, and clinical outcomes: a historical cohort study.

PURPOSE: There is a paucity of evidence evaluating whether intensive care unit (ICU) discharge occupancy is associated with clinical outcomes. It is unknown whether increased discharge occupancy leads to greater afterhours discharges and downstream consequences. We explore the association between ICU discharge occupancy and afterhours discharges, 72-hr readmission, and 30-day mortality. METHODS: This single-centre, historical cohort study included all patients discharged from the Vancouver General Hospital ICU between 5 April 2010 and 13 September 2017. Data were obtained from the British Columbia Critical Care Database. Occupancy was defined as the number of ICU bed hours utilized divided by the available bed hours for that day. Any discharge between 22:00 and 6:59 was considered afterhours. Logistic regression models adjusting for important covariates were constructed. RESULTS: We included 8,862 ICU discharges representing 7,288 individual patients. There were 1,180 (13.3%) afterhours discharges, 408 (4.6%) 72-hr readmissions, and 574 (6.5%) 30-day post-discharge deaths. Greater discharge occupancy was associated with afterhours discharges (per 10% increase: adjusted odds ratio [aOR], 1.12; 95% confidence interval [CI], 1.03 to 1.20; P = 0.005). Discharge occupancy was not associated with 72-hr readmission (per 10% increase: aOR, 0.97; 95% CI, 0.87 to 1.09; P = 0.62) or 30-day mortality (per 10% increase: aOR, 1.05; 95% CI, 0.95 to 1.16; P = 0.32). Afterhours discharge was not associated with 72-hr readmission (aOR, 1.15; 95% CI, 0.86 to 1.54; P = 0.34) or 30-day mortality (aOR, 1.05; 95% CI, 0.82 to 1.36; P = 0.69). CONCLUSIONS: Greater ICU discharge occupancy was associated with a significant increase in afterhours discharges. Nevertheless, neither discharge occupancy nor afterhours discharge were associated with 72-hr readmission or 30-day mortality.

RéSUMé: OBJECTIF: Il n'existe que peu de données probantes évaluant si le taux d'occupation de l'unité de soins intensifs (USI) au moment du congé est associé aux devenirs cliniques. Nous ne savons pas si un taux d'occupation plus élevé au moment du congé entraîne davantage de congés pendant la nuit et si cette situation a des conséquences. Nous avons exploré l'association entre le taux d'occupation de l'USI au moment du congé et les congés donnés pendant la nuit, la réadmission dans les premières 72 h, et la mortalité à 30 jours. MéTHODE: Cette étude de cohorte historique et monocentrique a englobé tous les patients ayant reçu leur congé de l'USI de l'Hôpital général de Vancouver entre le 5 avril 2010 et le 13 septembre 2017. Les données ont été tirées de la Base de données des soins intensifs de Colombie-Britannique (British Columbia Critical Care Database). Le taux d'occupation était défini comme le nombre d'heures d'occupation de lit de l'USI utilisées divisé par le nombre d'heures d'occupation de lit disponibles pour ladite journée. Tout congé reçu entre 22 h et 6 h 59 était considéré comme survenant pendant la nuit. Des modèles de régression logistique ont été élaborés afin de tenir compte des covariables importantes. RéSULTATS: Nous avons inclus 8862 congés de l'USI, représentant 7288 patients individuels. Au total, il y a eu 1180 (13,3 %) congés donnés pendant la nuit, 408 (4,6 %) réadmissions dans les 72 h suivantes, et 574 (6,5 %) décès à 30 jours après le congé. Un taux d'occupation plus élevé au moment du congé était associé à des congés pendant la nuit (par augmentation de 10 % : rapport de cotes ajusté [RCA], 1,12; intervalle de confiance [IC] 95 %, 1,03 à 1,20; P = 0,005). Le taux d'occupation lors du congé n'a pas été associé à une réadmission dans les premières 72 h (par augmentation de 10 % : RCA, 0,97; IC 95 %, 0,87 à 1,09; P = 0,62) ou à une mortalité à 30 jours (par augmentation de 10 % : RCA, 1,05; IC 95 %, 0,95 à 1,16; P = 0,32). Les congés pendant la nuit n'ont pas été associés à une réadmission dans les 72 h suivantes (RCA, 1,15; IC 95 %, 0,86 à 1,54; P = 0,34) ou à une mortalité à 30 jours (RCA, 1,05; IC 95 %, 0,82 à 1,36; P = 0,69). CONCLUSION: Un taux d'occupation de l'USI plus élevé au moment du congé était associé à une augmentation significative des congés donnés pendant la nuit. Cependant, ni le taux d'occupation lors du congé, ni le congé donné pendant la nuit, n'étaient associés à une réadmission à 72 h ou une mortalité à 30 jours.

Association of intensive care unit occupancy during admission and inpatient mortality: a retrospective cohort study.

PURPOSE: There is conflicting evidence regarding the influence of intensive care unit (ICU) occupancy at the time of admission on important patient outcomes such as mortality. The objective of this analysis was to characterize the association between ICU occupancy at the time of ICU admission and subsequent mortality. METHODS: This single-centre, retrospective cohort study included all patients admitted to the ICU at the Vancouver General Hospital between 4 January 2010 and 8 October 2017. Intensive care unit occupancy was defined as the number of ICU bed hours utilized in a day divided by the total amount of ICU bed hours available for that day. We constructed mixed-effects logistic regression models controlling for relevant covariates to assess the impact of admission occupancy quintiles on total inpatient (ICU and ward) and early (72-hr) ICU mortality. RESULTS: This analysis included 10,365 ICU admissions by 8,562 unique patients. Compared with ICU admissions in the median occupancy quintile, admissions in the highest and second highest occupancy quintile were associated with a significant increase in the odds of inpatient mortality (highest: odds ratio [OR], 1.33; 95% confidence interval [CI], 1.12 to 1.59; P value < 0.001; second highest: OR, 1.21; 95% CI, 1.02 to 1.44; P value < 0.03). No association between admission occupancy and 72-hr ICU mortality was observed. CONCLUSIONS: Admission to the ICU on days of high occupancy was associated with increased inpatient mortality, but not with increased 72-hr ICU mortality. Capacity strain on the ICU may result in significant negative consequences for patients, but further research is needed to fully characterize the complex effects of capacity strain.

Implementation of Neurocritical Care Is Associated With Improved Outcomes in Traumatic Brain Injury.

Background Traditionally, the delivery of dedicated neurocritical care (NCC) occurs in distinct NCC units and is associated with improved outcomes. Institution-specific logistical challenges pose barriers to the development of distinct NCC units; therefore, we developed a consultancy NCC service coupled with the implementation of invasive multimodal neuromonitoring, within a medical-surgical intensive care unit. Our objective was to evaluate the effect of a consultancy NCC program on neurologic outcomes in severe traumatic brain injury patients. METHODS: We conducted a single-center quasi-experimental uncontrolled pre- and post-NCC study in severe traumatic brain injury patients (Glasgow Coma Scale ≤8). The NCC program includes consultation with a neurointensivist and neurosurgeon and multimodal neuromonitoring. Demographic, injury severity metrics, neurophysiologic data, and therapeutic interventions were collected. Glasgow Outcome Scale (GOS) at 6 months was the primary outcome. Multivariable ordinal logistic regression was used to model the association between NCC implementation and GOS at 6 months. RESULTS: A total of 113 patients were identified: 76 pre-NCC and 37 post-NCC. Mean age was 39 years (standard deviation [SD], 2) and 87 of 113 (77%) patients were male. Median admission motor score was 3 (interquartile ratio, 1-4). Daily mean arterial pressure was higher (95 mmHg [SD, 10]) versus (88 mmHg [SD, 10], p<0.001) and daily mean core body temperature was lower (36.6°C [SD, 0.90]) versus (37.2°C [SD, 1.0], p=0.001) post-NCC compared with pre-NCC, respectively. Multivariable regression modelling revealed the NCC program was associated with a 2.5 increased odds (odds ratios, 2.5; 95% confidence interval, 1.1-5.3; p=0.022) of improved 6-month GOS. CONCLUSIONS: Implementation of a NCC program is associated with improved 6 month GOS in severe TBI patients.

Intensity of continuous renal replacement therapy in acute kidney injury in the intensive care unit: a systematic review and meta-analysis.

BACKGROUND AND OBJECTIVE: Acute kidney injury is a common finding among patients in the intensive care unit (ICU) and is an independent predictor of mortality. The optimal intensity and timing of continuous renal replacement therapy (CRRT), in critically ill patients remain unclear. The purpose of this study was to conduct a systematic review and meta-analysis of all prospective randomized controlled trials (RCTs) to determine the effect of intensity of CRRT on the survival of patients with acute renal failure (ARF) in ICU setting. METHODS: Search strategy and data source. Electronic databases were searched on MEDLINE (through February 2010), ISIWeb of Science, and Cochrane Central Register of Controlled Trials (2010); Pub Med ''Related articles.'' Trial authors were also contacted for additional information. Study selection and data abstraction. All prospective clinical trials comparing the intensity of CRRT in adult patients with ARF and with explicit reporting of mortality were included. Three authors independently evaluated articles for eligibility and extracted data on study quality and outcomes. Meta-analysis used a random-effects model. RESULT: Of the 322 citations, 5 trials (n = 2402) were included in the meta-analysis, which met all the inclusion and exclusion criteria. Meta-analysis showed that in critically ill patients with acute kidney injury, the high-dose CRRT did not reduce mortality at 28 days. (risk ratio [RR], 0.88; 95% confidence interval [CI], 0.70-1.11; P = 0.28). CONCLUSION: In critically ill patients with acute kidney injury, the high-dose CRRT did not reduce mortality at 28 days.

The safety of synthetic colloid in critically ill patients with severe traumatic brain injuries.

PURPOSE: Although 4% albumin is associated with increased mortality in patients with traumatic brain injury (TBI), evidence concerning the safety of synthetic colloids is lacking. We aimed to determine if there is an association between synthetic colloids and mortality in patients with severe TBI. MATERIALS AND METHODS: A retrospective cohort study of patients with severe TBI was conducted. Data were collected on all intravenous fluids administered during the first 14 days of admission. Multivariable Cox proportional hazards regression was used to model the association between daily cumulative pentastarch quintiles and mortality. RESULTS: Patients receiving pentastarch had higher Acute Physiology and Chronic Health II scores (23.9 vs 21.6, P < .01), frequency of craniotomy (42.5% vs 21.6%, P = .02), longer duration of intensive care unit stay (12 vs 4 days, P < .01), and mechanical ventilation (10 vs 3 days, P < .01). On unadjusted Cox regression, patients in the highest quintile of cumulative pentastarch administration had a higher rate of mortality compared with those receiving no colloid (hazard ratio, 3.8; 95% confidence interval, 1.2-12.4; P = .03). However, this relationship did not persist in the final multivariable model (hazard ratio 1.0; 95% confidence interval, 0.25-4.1; P = .98). CONCLUSION: There was no association between cumulative exposure to pentastarch and mortality in patients with severe TBI.

A critical appraisal of the quality of critical care pharmacotherapy clinical practice guidelines and their strength of recommendations.

OBJECTIVE: Clinical practice guideline (CPG) quality assessment is important before applying their recommendations. Determining whether recommendation strength is consistent with supporting quality of evidence is also essential. We aimed to determine quality of critical care pharmacotherapy CPGs and to assess whether high quality evidence supports strong pharmacotherapy recommendations. METHODS: MEDLINE (1966-February 2008), EMBASE (1980-February 2008), National Guideline Clearinghouse (February 2008) and personal files were searched to identify CPGs. Four appraisers evaluated each guideline using the appraisal of guidelines, research and evaluation (AGREE) instrument. AGREE assesses 23 items in six domains that include scope/purpose, stakeholder involvement, rigor of development, clarity, applicability and editorial independence. Standardized domain scores (0-100%) were determined to decide whether to recommend a guideline for use. One appraiser extracted strong pharmacotherapy recommendations and supporting evidence quality. RESULTS: Twenty-four CPGs were included. Standardized domain scores were clarity [69% (95% confidence interval (CI) 62-76%)], scope/purpose [62% (95% CI 55-68%)], rigor of development [51% (95% CI 42-60%)], editorial independence [39% (95% CI 26-52%)], stakeholder involvement [32% (95% CI 26-37%)] and applicability [19% (95% CI 12-26%)]. The proportion of guidelines that could be strongly recommended, recommended with alterations and not recommended was 25, 37.5 and 37.5%, respectively. High quality evidence supported 36% of strong pharmacotherapy recommendations. CONCLUSION: Variation in AGREE domain scores explain why one-third of critical care pharmacotherapy CPGs cannot be recommended. Only one-third of strong pharmacotherapy recommendations were supported by high quality evidence. We recommend appraisal of guideline quality and the caliber of supporting evidence prior to applying recommendations.

Identifying missed opportunities to curtail antimicrobial therapy for presumed ventilator-associated pneumonia using the clinical pulmonary infection score.

BACKGROUND: Early discontinuation of antimicrobial therapy for ventilator-associated pneumonia can reduce the emergence of antimicrobial resistance, the occurrence of adverse drug events, and the cost of therapy. Evidence suggests that discontinuation of therapy by day 3 may be appropriate for patients with a clinical pulmonary infection score of 6 or less at baseline and on day 3. OBJECTIVES: To determine the proportion of patients eligible for antimicrobial discontinuation on day 3 and day 7 of therapy and to determine the proportion of eligible patients for whom antimicrobials were discontinued within these timeframes. METHODS: A 6-month observational study was conducted from October 3, 2005, to March 31, 2006, in a 27-bed medical-surgical tertiary care intensive care unit. Clinical pharmacists attended daily rounds and prospectively identified patients for inclusion in the study. A study pharmacist retrospectively calculated clinical pulmonary infection scores. Other data were obtained from the quality-improvement database and patient health records for the intensive care unit. RESULTS: Ninety-two patients were treated for ventilator-associated pneumonia during the study period, of whom 49 were included in the analysis. At day 3, 17 (35%) of the 49 patients were eligible for early discontinuation of antimicrobial therapy, but therapy was discontinued for only 2 (12%) of these 17 patients. At day 7, 10 (32%) of 31 patients were eligible for antimicrobial discontinuation, but therapy was discontinued for only 1 (10%) of these 10 patients. CONCLUSIONS: A significant opportunity exists at the authors' institution to develop and implement an antimicrobial discontinuation policy that uses the clinical pulmonary infection score to guide antimicrobial use for patients with ventilator-associated pneumonia.

Hyperglycemia in acutely ill emergency patients--cause or effect?

OBJECTIVES: To clarify the benefits, risks and timing of glucose control and intensive insulin therapy in several groups, specifically the neurologic, cardiac and septic populations of patients, commonly seen in the emergency department. METHODS: Electronic search of MEDLINE (1966-2005; once with PubMed and once with Ovid) and Embase (1980-2005) using the terms insulin and glucose combined with emergency medicine, intensive care, cardiology and emergency department. RESULTS: There is considerable controversy in the literature surrounding the use of strict glucose control in cardiac, neurologic and septic patients. Much of this literature is non-randomized, and the timing of therapy is poorly investigated. CONCLUSIONS: Hyperglycemia is associated with adverse outcomes in acutely ill neurologic, cardiac and septic patients, but it remains unclear whether this is a causative association. Glucose control and intensive insulin therapy may be useful in some patient subgroups; however, controlled trials of aggressive glycemic control have provided insufficient evidence to justify subjecting patients to the real risks of iatrogenic hypoglycemia. We recommend a cautious approach to the control of glucose levels in acutely ill emergency department patients, with a target glucose of below 8 to 9 mmol/L.

Severity of illness and risk of death associated with pulmonary artery catheter use.

OBJECTIVE: To examine the association between the use of the pulmonary artery catheter and mortality rate in critically ill patients with a higher vs. a lower severity of illness. DESIGN: Observational cohort study. SETTING: A tertiary care university teaching hospital from March 1988 to March 1998. PATIENTS: A total of 7,310 critically ill adult patients. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The main outcome measure was hospital mortality rate, controlled by multivariable logistic regression within four patient groups based on severity of illness. Cutoffs for severity of illness were chosen based on Acute Physiology and Chronic Health Evaluation (APACHE) II score 25th percentiles. Logistic regression analysis demonstrated no increased risk of death associated with exposure to the pulmonary artery catheter in the population as a whole. The associated odds ratio of hospital death for the entire cohort was 1.05 (95% confidence interval, 0.92-1.21). Subgroup analysis of severity of illness revealed the highest risk of death to be associated with the lowest APACHE II score quartile vs. a decreased associated mortality rate with the highest APACHE II score quartile after adjustment with multivariable logistic regression (APACHE II <18: odds ratio, 2.47, 95% confidence interval, 1.27-4.81; APACHE II 18-24: odds ratio, 1.64, 95% confidence interval, 1.24-2.17; APACHE II 25-31: odds ratio, 1.00, 95% confidence interval, 0.80-1.24; APACHE II >31: odds ratio, 0.80, 95% confidence interval, 0.64-1.00). CONCLUSIONS: The use of the pulmonary artery catheter may decrease mortality rate in the most severely ill while increasing it in a population with a lower severity of illness. These findings underscore the necessity of examining the effect of severity of illness in future randomized controlled trials.

Hypothermia in the management of traumatic brain injury. A systematic review and meta-analysis.

OBJECTIVE: Brain injury remains the leading cause of death in cases of trauma in North America and Europe. This article critically appraised and summarised all published and peer-reviewed, randomised, controlled trials of the use of hypothermia in traumatic brain injury. DESIGN: To be included, a study had to be a published, randomised, controlled trial of the use of hypothermia in the management of traumatic brain injury. Pooling of data and meta-analysis of results occurred. SETTING: Conducted at a tertiary level Canadian teaching hospital. PATIENTS AND PARTICIPANTS: Patients were combined from eight randomised, controlled trials to generate a population of 748 severely head-injured patients. MEASUREMENTS AND RESULTS: Eight studies provided data on the efficacy of hypothermia in the management of traumatic brain injury. The pooled odds ratio of mortality in the hypothermic group was 0.81 (95%CI =0.59-1.13, p=0.22). The OR of a poor neurological outcome (GOS 1,2 or 3) was 0.75 (95% CI=0.56-1.01, p=0.06). The odds ratio for pneumonia in the normothermic group was 0.42 (95%CI =0.25-0.70, p=0.001). CONCLUSIONS: Although meta-analysis suggests that iatrogenic hypothermia may confer a marginal benefit in neurological outcome, there does not appear to be clear evidence of lower mortality rates in unselected traumatic brain injury patients. Prolonged hypothermia may confer a benefit, particularly in patients with elevated intracranial pressure refractory to conventional manipulations. Conclusions regarding the use of hypothermia are controversial and not strongly supported by the available evidence.

Rapid imipenem/cilastatin desensitization for multidrug-resistant Acinetobacter pneumonia.

OBJECTIVE: To report a successful case of rapid imipenem desensitization in a critically ill patient with multidrug-resistant Acinetobacter baumannii ventilator-associated pnemonia (VAP). CASE SUMMARY: A 40-year-old white man who had a lengthy stay in the intensive care unit (ICU) following a motorcycle accident developed VAP caused by A. baumannii. treatment with imipenem was necessary due to the bacteria's resistance to all other antibiotics. However, this patient was diagnosed with an allergy to imipenem following exposure earlier in his hospitalization in addition to a positive penicillin skin test. Thus, we attempted rapid desensitization to imipenem using a continuous infusion protocol. The patient was desensitized within 4 hours and was successfully treated for 21 days with a continuous infusion of imipenem combined with daily amikacin. He experienced no adverse reaction during the desensitization process or the remainder of his treatment course. DISCUSSION: The protocol used in this case was modified from a previously reported case, and differed in the speed of desensitization and total daily dose. We assumed that a more gradual escalation of the dose in our modified protocol would prevent the occurrence of adverse events, thereby resulting in more rapid desensitization. Rapid desensitization was necessary in this patient due to the presence of a life-threatening infection. The lower total daily dose of imipenem was in response to impaired renal function. CONCLUSIONS: Therapeutic options for multidrug-resistant pneumonia in the ICU are significantly limited in the presence of imipenem allergy. An option of last resort is to desensitize the patient using a rapid administration protocol. Our modified rapid imipenem desensitization protocol was successful and allowed for effective treatment of life-threatening pneumonia.

Lack of agreement between thermodilution and fick cardiac output in critically ill patients.

STUDY OBJECTIVE: s: Individual comparison of cardiac output via intermittent thermodilution and Fick technique over a wide range of cardiac outputs. DESIGN: Prospective clinical investigation. SETTING: Multidisciplinary ICUs of two teaching hospitals in Vancouver, British Columbia. PARTICIPANTS: Eighteen critically ill patients who had pulmonary and systemic arterial catheters and in whom active support was being withdrawn. INTERVENTIONS: Measurement of thermodilution cardiac output and calculation of Fick cardiac output while support was withdrawn. Active support was withdrawn in a three-step process: removal of vasopressors followed by decrease in fraction of inspired oxygen to 0.21, and finally removal of mechanical ventilation. MEASUREMENTS AND RESULTS: Simultaneous Fick and thermodilution cardiac outputs were obtained over a wide range. Fick calculated cardiac outputs were obtained using the Fick equation with oxygen uptake (O(2)) being measured with indirect calorimetry. O(2) determinations were made using five measurements over 5 min, with the mean being used for subsequent analysis. Thermodilution cardiac outputs were determined by the mean of five measurements, with the first being discarded. Coefficient of variation was calculated for the O(2) and thermodilution cardiac outputs. One hundred thirty-six simultaneous cardiac outputs were obtained in 18 patients with a mean APACHE (acute physiology and chronic health evaluation) II score of 25.5. The range of cardiac outputs was 1.39 to 16.95 L/min. Linear regression analysis found a good correlation of the data sets, with an R of 0.85. Bias and precision calculations found a bias of - 0.17 L/min with the upper and lower limits of agreement being 2.96 L/min and - 3.30 L/min, respectively. In patients with high cardiac outputs (> 7 L/min), the bias was - 1.90 with the limits of agreement being 1.87 L/min and - 5.67 L/min. The coefficient of variation for O(2) was 4.6% and for thermodilution cardiac output was 7.75%. CONCLUSIONS: There was good consistency of each of the measurements with a low coefficient of variation. The bias for the whole group was small, but the limits of agreement extended into a clinically relevant area, resulting in a lack of agreement. In patients with high cardiac outputs, the Fick tended to consistently produce higher cardiac outputs compared to thermodilution, suggesting a systematic error.