Do Blood Transfusions Improve Outcomes Related to Mechanical Ventilation?\u003cxref rid=\"AFF1\"\u003e\u003csup\u003e*\u003c/sup\u003e\u003c/xref\u003e

$Page 1: Do Blood Transfusions Improve Outcomes Related to Mechanical Ventilation?\u003cxref rid=\"AFF1\"\u003e\u003csup\u003e*\u003c/sup\u003e\u003c/xref\u003e$
DOI 10.1378/chest.119.6.1850 2001;119;1850-1857 Chest

George Wells, John Marshall and Irwin Schweitzer Paul C. Hébert, Morris A. Blajchman, Deborah J. Cook, Elizabeth Yetisir,

Related to Mechanical Ventilation?Do Blood Transfusions Improve Outcomes

http://chestjournal.org/cgi/content/abstract/119/6/1850and services can be found online on the World Wide Web at: The online version of this article, along with updated information

). ISSN: 0012-3692. http://www.chestjournal.org/misc/reprints.shtml(of the copyright holder may be reproduced or distributed without the prior written permission Northbrook IL 60062. All rights reserved. No part of this article or PDFby the American College of Chest Physicians, 3300 Dundee Road,

2007Physicians. It has been published monthly since 1935. Copyright CHEST is the official journal of the American College of Chest

Copyright © 2001 by American College of Chest Physicians on June 11, 2008 chestjournal.orgDownloaded from

http://chestjournal.org/cgi/content/abstract/119/6/1850

http://www.chestjournal.org

Do Blood Transfusions Improve OutcomesRelated to Mechanical Ventilation?*

Paul C. Hebert, MD, MHSc; Morris A. Blajchman, MD;Deborah J. Cook, MD, FCCP, MSc(Epid); Elizabeth Yetisir, MSc;George Wells, MSc, PhD; John Marshall, MD; Irwin Schweitzer, MSc; and theTransfusion Requirements in Critical Care Investigators for the CanadianCritical Care Trials Group†

Background: Correcting the decrease in oxygen delivery from anemia using allogeneic RBCtransfusions has been hypothesized to help with increased oxygen demands during weaning frommechanical ventilation. However, it is also possible that transfusions hinder the process becauseRBCs may not be able to adequately increase oxygen delivery. In this study, we determinedwhether a liberal RBC transfusion strategy improved outcomes related to mechanical ventilation.Methods: Seven hundred thirteen patients receiving mechanical ventilation, representing asubgroup of patients from a larger trial, were randomized to either a restrictive transfusionstrategy, receiving allogeneic RBC transfusions at a hemoglobin concentration of 7.0 g/dL (andmaintained between 7.0 g/dL and to 9.0 g/dL), or to a liberal transfusion strategy, receiving RBCsat 10.0 g/dL (and maintained between 10.0 g/dL and 12.0 g/dL). The larger trial was designed toevaluate transfusion practice rather than weaning per se.Results: Baseline characteristics in the restrictive-strategy group (n 5 357) and the liberal-strategy group (n 5 356) were comparable. The average durations of mechanical ventilation were8.3 6 8.1 days and 8.3 6 8.1 days (95% confidence interval [CI] around difference, 2 0.79 to 1.68;p 5 0.48), while ventilator-free days were 17.5 6 10.9 days and 16.1 6 11.4 days (95% CI arounddifference, 2 3.07 to 0.21; p 5 0.09) in the restrictive-strategy group vs the liberal-strategygroup, respectively. Eighty-two percent of the patients in the restrictive-strategy group wereconsidered successfully weaned and extubated for at least 24 h, compared to 78% for theliberal-strategy group (p 5 0.19). The relative risk (RR) of extubation success in the restrictive-strategy group compared to the liberal-strategy group, adjusted for the confounding effects ofage, APACHE (acute physiology and chronic health evaluation) II score, and comorbid illness,was 1.07 (95% CI, 0.96 to 1.26; p 5 0.43). The adjusted RR of extubation success associated withrestrictive transfusion in the 219 patients who received mechanical ventilation for > 7 days was1.1 (95% CI, 0.84 to 1.45; p 5 0.47).Conclusion: In this study, there was no evidence that a liberal RBC transfusion strategy decreasedthe duration of mechanical ventilation in a heterogeneous population of critically ill patients.

(CHEST 2001; 119:1850–1857)

Key words: critical care; mechanical ventilation; oxygen delivery; RBC transfusion; transfusion trigger; weaning

Abbreviations: APACHE 5 acute physiology and chronic health evaluation; CI 5 confidence interval; RR 5 relativerisk; TRICC 5 Transfusion Requirements in Critical Care

A nemia is common in the critically ill and may bean important factor interfering with a patient’s

ability to wean from mechanical ventilation.1 Duringweaning, oxygen consumption may be increased,because respiratory muscles must overcome the

increased work of breathing imposed by the pres-ence of conditions affecting the lung and respiratorymuscles.2 The increased oxygen consumption mustbe accompanied by increased oxygen delivery to vitalorgans, including the heart,3,4 the respiratory mus-cles,5 and the splanchnic circulation.6 Improving

*From the Critical Care Programs (Dr. Hebert) and the ClinicalEpidemiology Unit (Dr. Wells, Mr. Schweitzer, Ms. Yetisir),University of Ottawa, Ottawa, Ontario; the University of Toronto(Dr. Marshall), Toronto, Ontario; and the Departments ofPathology (Dr. Blajchman) and Medicine and Epidemiology(Dr. Cook), McMaster University, Hamilton, Ontario, Canada.†A list of other study investigators is given in the Appendix.Drs. Hebert and Cook are Career Scientists of the OntarioMinistry of Health.

This study was supported by the Medical Research Council ofCanada and an unrestricted grant from Bayer Inc.Manuscript received April 10, 2000; revision accepted November3, 2000.Correspondence to: Paul C. Hebert, MD, MHSc(Epid), Depart-ment of Medicine, Ottawa Hospital, General Site, 501 Smyth Rd,Box 201, Ottawa, Ontario K1H 8L6, Canada; e-mail: [email protected]

1850 Clinical Investigations in Critical Care



oxygen delivery in anemic patients using RBC trans-fusions is believed to help patients cope with theincreased oxygen demands during periods of venti-latory support and weaning.7,8 However, it is alsopossible that RBC transfusions may not improveoxygen delivery during weaning, but may hinder theprocess because of changes in RBC function re-ported to occur during storage. In addition, compli-cations such as pulmonary edema from volumeoverload9 or an increased rate of nosocomial infec-tions from transfusion-associated immune suppres-sion may directly prolong the length of time a patientreceives mechanical ventilation or decrease weaningsuccess.10

Although the use of allogeneic RBCs is common incritical care practice, there have been very few largestudies examining the consequences of anemia andRBC transfusions, before the recent publication ofthe results of the Transfusion Requirements in Crit-ical Care (TRICC) trial.9 In this randomized trial9comparing a restrictive transfusion strategy vs aliberal transfusion strategy, 713 of the patients en-rolled (85%) required mechanical ventilation, 357patients in the restrictive transfusion strategy armand 356 patients in the liberal transfusion strategyarm. An analysis of these patients receiving mechan-ical ventilation afforded an opportunity to examinethe effects of anemia and RBC transfusion on me-chanical ventilation outcomes.

Materials and Methods

Description of the TRICC Trial

The TRICC trial9 was a randomized, controlled trial thatenrolled 838 critically ill patients with hemoglobin concentrations# 9.0 g/dL within 72 h of ICU admission, and were consideredvolume resuscitated by the attending ICU staff. Patients withchronic anemia and acute severe blood loss, defined as a decreasein hemoglobin concentration . 30 g/L or a requirement for threeRBC units in 12 h, were excluded from the TRICC trial, as wellas this analysis. Physicians caring for patients allocated to therestrictive RBC transfusion strategy were instructed to transfuseone RBC unit when a patient’s hemoglobin concentration fell to, 7.0 g/dL, and to maintain the patient’s hemoglobin concentra-tion between 7.0 g/dL and 9.0 g/dL. In the liberal-strategy group,hemoglobin concentrations were maintained between 10.0 g/dLand 12.0 g/dL. RBC units were administered when a patient’shemoglobin values fell to , 10.0 g/dL in this group. Hemoglobinconcentrations were measured after each RBC unit was trans-fused in all study patients. The primary outcome in the study was30-day all-cause mortality. Secondary outcomes included othermortality rates and organ failure. Details of the TRICC protocoland overall results have been reported previously.9 The presentanalysis is limited to those patients who required mechanicalventilation through an endotracheal tube or tracheostomy regard-less of their duration of ventilation subsequent to being enrolledin the TRICC trial. Patients receiving noninvasive ventilationwere not considered. Because RBC transfusions might have a

greater effect on patients requiring a longer course of mechanicalventilation, a priori, we also decided to examine the subgroup ofpatients who received mechanical ventilation for . 7 days.

Outcome Measures

Outcomes from mechanical ventilation included a comparisonof (1) the proportion of patients considered successfully weanedand extubated (defined as not requiring mechanical ventilationfor at least 24 h); (2) the total duration of mechanical ventilationduring the 30-day study period; (3) the time to successfulextubation; and (4) ventilator-free days, defined as the number ofdays not requiring mechanical ventilation during 30 days ofobservation. A patient had to be free of mechanical ventilation forat least 24 h to have any day counted as being ventilator free.Patients were assigned zero ventilator-free days if they diedwithin the 30-day observation period.

Statistical Analysis

The analysis was conducted on an intention-to-treat basis. Thenumbers of mechanical ventilation days and ventilator-free dayswere compared using independent t tests. The proportion ofpatients who were considered to have been successfully weanedand extubated were compared using the Fisher’s Exact Test. Forthis analysis, we opted to evaluate successful weaning andextubation using similar definitions with two cutoff points. First,patients were considered successfully weaned and extubated ifalive and free from mechanical ventilation for at least 24 h; andsecond, if alive and free from mechanical ventilation for a 30-dayperiod. Using time to extubation success with both definitions,Kaplan-Meier survival curves were constructed and comparedusing log-rank tests. Cox proportional-hazards modeling was usedto adjust for differences in duration of ventilation. Covariateswere submitted for analysis in the survival model at p # 0.20. Asecond Cox model was constructed, forcing potential confound-ers, including patient age, APACHE (acute physiology andchronic health evaluation) II score, comorbid illness, includingischemic heart disease and other cardiac diagnoses, and treat-ment, into the model. The effect of hemoglobin concentrationswas forced into the model as a continuous variable, while theeffect of the number of RBC transfusions was examined in aseparate model. We performed these analyses in all survivingpatients and the subgroup of surviving patients who receivedmechanical ventilation for at least 7 days. Complications werecompared using the Fisher’s Exact Test. Lengths of ICU andhospital stay were analyzed using the Wilcoxon rank-sum test forindependent samples. Comparisons of primary outcomes wereconsidered statistically significant using an overall two-sided a of0.05. No adjustments were made for multiple comparisons.Absolute p values and 95% confidence intervals (CIs) were alsoreported.

Results

Study Population

The TRICC trial randomized 838 patients (418 inthe restrictive allogeneic transfusion group and 420in the liberal-strategy group). In total, 713 patients(85%) required mechanical ventilation, 357 in therestrictive-strategy group and 356 in the liberal-strategy group. All patients in this analysis completedthe trial and were followed up for 30 days. Twopatients were unavailable for follow-up at 60 days.

CHEST / 119 / 6 / JUNE, 2001 1851



All baseline characteristics were equally balancedbetween the treatment groups among patients whorequired mechanical ventilation (p . 0.05; Table 1).Study participants had a mean APACHE II score of22, and . 95% were receiving mechanical ventila-tion at baseline.

Baseline characteristics were also comparable inthe subgroup of patients who required mechanicalventilation for . 7 days. Approximately 70% ofpatients in both groups were male (62% vs 66%;p 5 0.58), had an average age in the mid-50s(54.9 6 17 years vs 53.9 6 19 years; p 5 0.67), and anAPACHE II score . 20 (21.7 6 7 vs 22.3 6 7.9;p 5 0.59), respectively. In patients who received me-chanical ventilation . 7 days, the most commondiagnostic category in both groups was respiratorydisease (37% vs 31%; p 5 0.39), followed by cardio-vascular disease (13% vs 19%; p 5 0.20) and trauma(17% vs 24%; p 5 0.13), respectively. Comorbidillnesses were common in both the restrictive-strat-egy group and the liberal-strategy group (25% vs27%, respectively; p 5 0.76).

There were no study protocols for mechanical

ventilation during the trial; however, assist control,assist control with pressure control, and synchro-nized intermittent mandatory ventilation with pres-sure support were the primary modes of mechanicalventilation employed in all centers. Assessments ofweaning readiness, weaning methods, and extubationdecisions were made at the discretion of the ICUteam.

Successful Implementation of the Study

Hemoglobin concentrations averaged 8.4 6 0.62g/dL throughout the ICU stay in the restrictive-strategy group and 10.4 6 0.71 g/dL in the liberal-strategy group (p , 0.01). Once randomized, anaverage of 2.7 6 4.0 U of RBCs per patient wereadministered in the restrictive-strategy group, com-pared to 5.5 6 5.1 U of RBCs per patient in theliberal-strategy group (p , 0.01). Physicians main-tained hemoglobin concentrations within the pre-specified limits in . 97% of patients. Cointerven-tions potentially modifying oxygen delivery,including the use of vasoactive drugs, overall fluidbalance, and pulmonary artery catheter use, werecomparable in the two groups throughout the ICUstay (p . 0.05).

Outcome Measures

In the 713 patients receiving mechanical ventila-tion, the mean duration of mechanical ventilationwas 8.3 6 8.1 days in the restrictive-strategy groupand 8.3 6 8.1 days in the liberal-strategy group (95%CI for the difference between groups, 2 0.79 to1.68; p 5 0.48; Table 2). Ventilator-free days were17.5 6 10.9 days and 16.1 6 11.4 days in the restric-tive-strategy group vs the liberal-strategy group,respectively (95% CI for the difference betweengroups, 2 3.07 to 0.21; p 5 0.09). Eighty-two per-cent of the patients in the restrictive-strategy groupwere considered successfully weaned and extubatedfor at least 24 h, compared to 78% for the liberal-strategy group (p 5 0.19). The unadjusted relativerisk (RR) of successful extubation for the restrictivevs liberal strategies was 1.08 (95% CI, 0.92 to 1.27;p 5 0.35). The RR for being successfully weanedand extubated was not significantly different afteradjustment for the confounding influence of age,APACHE II score, and comorbid illness (RR, 1.07;95% CI, 0.91 to 1.26; p 5 0.43). Using 30 days as thethreshold for extubation success, both the unad-justed RR (1.12; 95% CI, 0.94 to 1.34; p 5 0.21) andadjusted RR (1.09; 95% CI, 0.92 to 1.30; p 5 0.33)were similar to each other and to the results whensuccessful weaning and extubation were measured at24 h.

We also examined outcomes in the 219 patients

Table 1—Baseline Characteristics of the 713 PatientsReceiving Mechanical Ventilation*

Characteristics

RestrictiveStrategy

(n 5 357)

LiberalStrategy

(n 5 356) p Value

Patient characteristicsMale gender 237 (66) 216 (61) 0.12Age, yr 57 6 18.1 59 6 18.4 0.26APACHE II score 22 6 7.1 22 6 7.8 0.44MOD score 7.4 6 3.5 7.6 6 3.6 0.49

Diagnostic categoryRespiratory 111 (31) 112 (31)Cardiovascular 59 (17) 75 (21)Trauma 73 (20) 70 (20)GI disease 49 (14) 51 (14)Sepsis 19 (5) 14 (4)Neurologic 22 (6) 11 (3)Other 24 (7) 23 (6) 0.37

Other conditions presentComorbid condition 104 (29) 125 (35) 0.5Ischemic heart disease 67 (19) 87 (24) 0.07

InterventionsPulmonary artery

catheter126 (35) 131 (37) 0.7

Mechanical ventilation 345 (97) 348 (98) 0.7Oxygen delivery

variablesHemoglobin, g/dL 8.2 6 0.68 8.2 6 0.69 0.64Total fluid intake, L 4.1 6 2.2 4.1 6 1.6 0.87Transfusions, No. 2.7 6 7.0 2.4 6 4.7 0.46Lactate, mmol/L 1.8 6 1.9 1.9 6 3.0 0.9Vasoactive drugs 188 (53) 192 (54) 0.73

*Data are presented as No. (%) or mean 6 SD. MOD 5 multipleorgan dysfunction.




who required mechanical ventilation for . 7 days(Table 3). In this subgroup, the mean duration ofmechanical ventilation was 16.3 6 7.8 days vs15.9 6 7.8 days, respectively, when comparing therestrictive-strategy group to the liberal-strategygroup (95% CI for the difference, 2 2.31 to 1.36;p 5 0.61). Ventilator-free days were 12.1 6 8.4 daysand 9.8 6 9.0 days in the restrictive-strategy group vsthe liberal-strategy group, respectively (95% CI forthe difference, 2 4.42 to 2 0.33; p 5 0.02). Othermechanical ventilation outcomes are described in

Table 4. The median time to successful extubationwas 17 days (interquartile range, 11 to 28 days) in therestrictive-strategy group and 20 days (interquartilerange, 12 to 30 days) in the liberal-strategy group(Fig 1). The unadjusted RRs of successful extubationfor the restrictive-strategy group vs the liberal-strat-egy group in patients requiring mechanical ventila-tion for . 7 days were 1.13 (95% CI, 0.86 to 1.49;p 5 0.37) using 24 h, and 1.30 (95% CI, 0.97 to 1.74;p 5 0.08) using 30 days of observation. The RR wasnot significantly different if adjusted for the con-

Table 2—Outcomes in the 713 Patients in the TRICC Trial Who Required Mechanical Ventilation*

OutcomesRestrictive Strategy

(n 5 357)Liberal Strategy

(n 5 356) p Value

MV outcomesReceiving MV, d

Length of ventilation 8.3 6 8.1 8.8 6 8.7 0.48Ventilator-free days† 17.5 6 10.9 16.1 6 11.4 0.09

Not receiving MV for at least 24 hPatients 292 (82) 277 (78) 0.19Time to wean, d 7.7 6 6.6 7.4 6 6.2 0.55

Not receiving MV for at least 30 dPatients 283 (79) 261 (73) 0.78Time to wean, d 8.8 6 7.3 8.5 6 6.9 0.63

Other outcomesMortality

30 d 76 (21.3) 94 (26.4) 0.1160 d 92 (26) 105 (30) 0.32ICU 58 (16) 67 (19) 0.38Hospital 90 (25) 112 (31) 0.07

Organ dysfunctionMOD score 8.7 6 4.7 9.3 6 4.4 0.11MOD score‡ 11.4 6 7.6 12.6 6 7.7 0.04Change in MOD score 1.1 6 4.3 1.5 6 4.2 0.29Change in MOD score‡ 3.8 6 7.2 4.7 6 7.4 0.09

Length of stay, dICU 12.2 6 11.1 12.7 6 11.8 0.54Hospital 36.1 6 19.5 36.5 6 19.2 0.8

*Intention-to-treat analysis. Data are presented as mean 6 SD or No. (%). MODS 5 multiple organ dysfunction syndrome; MV 5 mechanicalventilation.

†Patients who died were given zero ventilator-free days.‡Nonsurvivors are considered to have all organs failing on date of death.

Table 3—Outcomes in the 219 Patients Who Required Mechanical Ventilation > 1 Week*

OutcomesRestrictive Strategy

(n 5 116)Liberal Strategy

(n 5 103) p Value

Receiving mechanical ventilationLength of ventilation, d 16.4 6 7.9 16.9 6 8.5 0.7Ventilator-free days, No.† 13.6 6 7.9 13.1 6 8.5 0.64

Not receiving mechanical ventilation for at least 24 hWeaned 101 (87) 87 (84) 0.7Time to wean, d 14.1 6 6.7 13.3 6 7.3 0.44

Not receiving mechanical ventilation for at least 30 dWeaned 97 (84) 82 (80) 0.49Time to wean, d 15.8 6 6.5 15.5 6 6.7 0.78

*Intention-to-treat analysis. Data are presented as mean 6 SD or No. (%).†Patients who died were given zero ventilator-free days.

CHEST / 119 / 6 / JUNE, 2001 1853



founding influence of age, APACHE II score, andcomorbid illness using either 24 h (RR, 1.11; 95%CI, 0.84 to 1.45; p 5 0.47) or 30 days in the defini-tion of success (unadjusted RR, 1.30; 95% CI, 0.91 to1.85; p 5 0.08; adjusted RR, 1.26; 95% CI, 0.94 to1.70; p 5 0.13).

The independent effects of RBC transfusions andhemoglobin concentrations were also examined.Each additional transfusion was associated with anincreased duration of mechanical ventilation (RR,1.10; 95% CI, 1.14 to 1.06; p , 0.01), adjusting forthe effects of age, APACHE II score, and comorbidillnesses (Fig 2). Hemoglobin concentrations did notinfluence the duration of mechanical ventilation(RR, 0.99; 95% CI, 1.01 to 0.98; p 5 0.45). Compli-cations including pulmonary edema and ARDS wereincreased in patients in the liberal-strategy group(Table 4).

Discussion

In summary, we found no significant differences inthe duration of mechanical ventilation, in the num-ber of ventilator-free days, or in the time necessaryto successfully wean and extubate patients frommechanical ventilation among those receiving a re-strictive transfusion strategy vs a liberal transfusionstrategy. This was true for all patients receivingmechanical ventilation and in the subgroup whorequired mechanical ventilation for . 7 days. There-

fore, hemoglobin concentrations and RBC transfu-sions did not influence the duration of mechanicalventilation or other mechanical ventilation outcomesin this randomized trial.

The consequences of anemia on lung function,work of breathing, and weaning from mechanicalventilation have not been studied extensively.11,12

The two studies11,12 evaluating posttransfusion lungfunction in patients with chronic anemia have diver-gent conclusions, with one study finding worsenedgas exchange and the other finding improved gasexchange. Limited research has addressed the role ofRBC transfusions in treating anemic patients requir-ing ongoing ventilatory support or being weanedfrom mechanical ventilation. Two case series7,8 de-scribe weaning successes in a small group of anemicpatients who received RBC transfusions. In fivepatients with severe COPD who failed repeatedattempts to wean from mechanical ventilation,Schonhofer et al8 described successfully weaning allpatients after increasing their hemoglobin concen-trations to . 12.0 g/dL. In a second study of me-chanical ventilation,7 the same investigators mea-sured the effects of RBC transfusions on the work ofbreathing and other respiratory parameters in 10anemic patients with severe COPD and in 10 anemicpatients who required mechanical ventilation associ-ated with other diagnoses. They observed a decreasein work of breathing and minute ventilation aftertransfusion in the group of patients with severeCOPD. From this small study, it is unclear whetherthe work of breathing was improved because of

Figure 1. Time remaining on mechanical ventilation in 713patients. The time to successful extubation from mechanicalventilation is illustrated using Kaplan-Meier survival curves.Weaning success is defined as remaining free of mechanicalventilation, once extubated, during the 30 days of observation.The hatched line refers to the patients in the restrictive transfu-sion group; the solid line refers to the patients in the liberaltransfusion group. Survival curves were not statistically differentwhen compared using a log-rank test (p 5 0.21). The mediantime to extubation was 7 days (interquartile range, 2 to 18 days)in the restrictive group and 7 days (interquartile range, 3 to 23days) in the liberal group.

Figure 2. Time remaining on mechanical ventilation in the 283patients requiring mechanical ventilation for . 1 week. The timeto successful weaning from mechanical ventilation is illustratedusing Kaplan-Meier survival curves in patients who requiredmechanical ventilation for . 1 week. Weaning success is definedas remaining free of mechanical ventilation, once extubated,during the 30 days of observation. The hatched line refers to therestrictive group; the solid line refers to the liberal group.Survival curves were not statistically different when comparedusing a log-rank test (p 5 0.08).




higher cardiac output as a result of increased preloador because of increased oxygen delivery to therespiratory muscles.

Because anemia may result in a limitation inoxygen delivery, the left ventricle may not be able toincrease cardiac output during the weaning processand oxygen delivery to the respiratory muscles maynot keep up with the increased oxygen requirements.The weaning process may precipitate myocardialischemia because of the increased strain on leftventricular function. In this heterogeneous group ofpatients receiving mechanical ventilation, we did notidentify adverse effects of low hemoglobin values;however, we did not conduct systematic screeningfor adverse effects, such as ECG evidence of myo-cardial ischemia. In a study by Srivastava and col-leagues,13 among 83 patients with coronary arterydisease who received mechanical ventilation for amean of 4.6 days, 8 patients had ECG ischemiaduring weaning and 7 patients failed to be liberatedon the first day; ischemia was associated with a riskratio of weaning failure of 2.1 (95% CI, 1.4 to 3.1).However, we did observe a significant increase inrates of pulmonary edema in patients transfusedaccording to a liberal strategy. The increase ineffective circulating volume and subsequent pulmo-nary edema seen in many liberally transfused pa-tients may have offset any potential benefit fromincreased oxygen delivery.

There are a number of limitations to this study.The most important is that this study was designed toassess the overall effects of transfusion practices inthe critically ill, rather than evaluate the effects ofRBC transfusions on outcomes from mechanicalventilation. Consequently, decision algorithms forweaning and extubation were not used. However,

different approaches to weaning from mechanicalventilation have not generated conclusions aboutconsistently superior approaches,14,15 making it dif-ficult to estimate the effect of this potential con-founding variable. If such protocols had been shownto impact on the duration of ventilation and if theywere differentially applied in these two groups, ourconclusions would be different. Some randomizedtrials of different approaches to achieve safe andrapid extubation, including respiratory therapy-driven protocols16,17 and noninvasive ventilation,18,19

were not used in this study. Nevertheless, the lack ofa standard approach to weaning may increase thevariability in this trial and may have decreased ourability to detect meaningful differences between thetwo groups. Second, the majority of patients receiv-ing mechanical ventilation can be rapidly and safelyextubated14,20 and may easily tolerate low hemoglo-bin values during a relatively rapid process of liber-ation from mechanical ventilation. Therefore, thepotential benefit or harm associated with transfusionmay have been attenuated because we enrolled allpatients receiving mechanical ventilation regardlessof their duration of mechanical support; however, wedid not observe any benefit to a liberal transfusionstrategy among the subgroup of patients receivingmechanical ventilation for . 7 days. Another limita-tion to this study is that the analysis was based on asubgroup of patients from a larger randomized trial;therefore, all inferences should be interpreted cau-tiously, given that unknown prognostic variables maynot be completely balanced between the groups.However, baseline differences between these twogroups were similar, and multivariate analysis did notindicate any significant changes in the direction ormagnitude of the RR compared to the unadjusted

Table 4—Complications Throughout the Study in the 713 Patients Who Required Mechanical Ventilation*

Complications

RestrictiveStrategy

(n 5 357)

LiberalStrategy

(n 5 356) Difference

95% CI

p ValueLower Upper

Cardiac 51 (14) 80 (22) 8.2 2.3 13.9 , 0.01Angina† 5 (1) 9 (3) 1.1 — — 0.28Myocardial infarct† 2 (1) 12 (3) 2.8 — — , 0.01Cardiac arrest 29 (8) 31 (9) 0.6 2 3.5 4.7 0.78Pulmonary edema 18 (5) 38 (11) 5.6 2 0.1 7.3 , 0.01Pulmonary 102 (29) 114 (32) 3.5 2 3.3 10.2 0.33ARDS 31 (9) 48 (13) 4.8 0.2 9.4 0.04Hematologic 9 (3) 9 (3) 0.0 2 2.3 2.3 1GI 13 (4) 17 (5) 1.1 2 1.8 4.1 0.46Infectious 41 (11) 50 (14) 2.6 2 2.3 7.5 0.32Neurologic 24 (7) 32 (9) 2.3 2 1.7 6.2 0.27Shock 65 (18) 51 (14) 2 3.9 2 1.5 9.3 0.19Any 193 (54) 210 (59) 4.9 2 2.4 12.2 0.19

*Data are presented as No. (%) unless otherwise indicated.†We were unable to calculate 95% CI because of the small number of patients.

CHEST / 119 / 6 / JUNE, 2001 1855



analysis. Finally, a post hoc power calculation re-vealed that this analysis would have been able todetect relative differences of 25% in duration ofmechanical ventilation, 20% in ventilator-free days,and 10% for extubation success. Therefore, it ispossible that our sample size may have been toosmall to detect clinically important differences inmechanical ventilation outcomes.

In summary, this study does not support thehypothesis that RBC transfusions improve outcomesfrom mechanical ventilation. A restrictive transfusionpolicy maintaining hemoglobin concentrations be-tween 7.0 g/dL and 9.0 g/dL seems reasonable basedon this analysis until further studies measure theinteraction between hemoglobin values and differentmethods of weaning from mechanical ventilation.Careful observational studies will advance our un-derstanding of the relation between myocardial andrespiratory muscle function and hemoglobin values.Additional research should determine what RBCtransfusion strategy should be advocated for patientswho are difficult to liberate from mechanical venti-lation,21 whether hemoglobin concentrations influ-ence the success of unassisted breathing trials ofdifferent duration,22,23 and whether a liberal RBCtransfusion strategy is advantageous for patients re-ceiving mechanical ventilation with acute coronarysyndromes.

ACKNOWLEDGMENT: We thank Drs. Graeme Rocker, Dar-ren Heyland, Jacques Lacroix, Thomas Todd, and the membersof the Canadian Critical Care Trials group; the nurses and criticalcare teams who provided medical care; and Christine Piche forsecretarial support. We also thank Dr. Mark Pickett, Director ofResearch and Development at Bayer Inc., and Dr. Bert T. Aye,former Director of the Canadian Red Cross Society BloodServices.

Appendix

TRICC Trial Executive and Writing Committee

Paul C. Hebert, MD, Irwin Schweitzer, MSc, Ottawa Hospital,General Campus; George Wells, PhD, MSc, Guiseppe Pagliar-ello, MD, Ottawa Hospital, Civic Campus; Morris Blajchman,MD, McMaster University, Hamilton; John Marshall, MD, To-ronto Hospital, General Division; Claudio Martin, MD, MSc,Victoria Hospital, London; Martin Tweeddale, MD, PhD, Van-couver General Hospital.

TRICC Investigators

Paul C. Hebert, MD, Ottawa Hospital, General Campus;Guiseppe Pagliarello, MD, Ottawa Hospital, Civic Campus; JohnMarshall, MD, Toronto Hospital, General Division; PatriciaHouston, MD, Toronto Hospital, Western Division; MartinTweeddale, MD, PhD, Vancouver General Hospital; RichardHall, MD, Queen Elizabeth II Health Sciences Centre, Halifax;David Mazur, MD, St. Michael’s Hospital, Toronto; ThomasStewart, MD, MSc, Wellesley Hospital, Toronto; Thomas Hillers,

MD, MSc, Hamilton General Hospital; Dean Sandham, MD,Foothills Hospital, Calgary; James A. Russell, MD, St. Paul’sHospital, Vancouver; Yoanna Skrobik, MD, Hopital Maison-neuve-Rosemont, Montreal; John Muscedere, Hotel Dieu-GraceHospital, Windsor; Claudio Martin, MD, MSc, Victoria Hospital,London; Sharon Peters, MD, Health Sciences Centre, St. John’s;David Fleiszer, MD, Montreal General Hospital; Alan Spanier,MD, Jewish General Hospital, Montreal; Ann Kirby, MD, SaintJoseph’s Hospital, London; Jaime Pinilla, MD, Royal UniversityHospital, Saskatoon; Mary van Wijngaarden, MD, University ofAlberta Hospital, Edmonton; Sheldon Magder, MD, Royal Vic-toria Hospital, Montreal; Gordon Wood, MD, Daren Heyland,MD, Kingston General Hospital; Navdeep Mehta, MD, Dr.Everett Chalmers Hospital, Fredericton; Michael Jacka, MD, St.John Regional Hospital; Sidney Viner, MD, Calgary GeneralHospital/Peter Lougheed Centre.

Data Monitoring Committee

Deborah Cook, MD, St. Joseph’s Hospital; Jack Hirsh, MD,Hamilton Health Sciences Centre; Richard Cook, PhD, Univer-sity of Waterloo; Thomas Todd, MD, Toronto General Hospital.

Data Management Committee

Paul C. Hebert, MD, Irwin Schweitzer, MSc, Elizabeth Yetisir,MSc, Ottawa Hospital, General Campus; George Wells, PhD,MSc, My-Linh Tran, Fiona Daigle-Campbell, BA, Anne Gray,Ottawa Hospital, Civic Campus.

Site Research Coordinators

Mustafa Seyidoglu, MD, Charlene Sexton, BScN, GhulamDostazadu, MD, Ottawa Hospital, General Campus; MerrileeLeowen, RN, Ottawa Hospital, Civic Campus; Debbie Williams,RN, Barbara Plumbstead, BScN, Vancouver General Hospital;Joan Kearney, RN, Gwen Williams, RN, Vivian Nedelcu, RN,Queen Elizabeth II Health Sciences Centre, Halifax; LindaPerkins, RN, Montreal General Hospital; Gail Sloane, BScN, St.Michael’s Hospital, Toronto; Violet Smirnios, RRT, Chanel McK-enna, RN, Eduardo Ng, MD, Toronto Hospital, Western Divi-sion; Marilyn Steinberg, RN, Debra Foster, RN, BSc, DeborahBaptiste, RN; Toronto Hospital, General Division; Joanne Ke-hoe, RPN, Linda McCarthy, RN, Dawn Gilliland, RT, BrianMartin, RRT, Victoria Hospital, London; Daisy Gibbons, RN,Health Science Centre, St. Johns; Deborah Jones, BScN, SoniaBertleff, BScN, Royal Victoria Hospital, Montreal; Diane Collins,RN, Jewish General Hospital, Montreal; Diana Schouten, BScN,Wellesley Hospital, Toronto; Lesley Crenshaw, RN, Linda Knox,RN, Judie Lasante, Foothills Hospital, Calgary; Mary-KatherineScott, RN, Saint Joseph’s Hospital, London; Judee Strickland,RN, Royal University Hospital, Saskatoon; Michelle Douglas,BScN, Karen Mulcahy, RN, Alana Drummond, RN, St. Paul’sHospital, Vancouver; Mario Racine, RN, Hopital Maisonneuve-Rosemont, Montreal; Melanie Amos, RN, Dr. Everett ChalmersHospital, Fredericton; Carol Gunderson, RN, Calgary GeneralHospital/Peter Lougheed Centre; Loree Morrison, RN, Hamil-ton General Hospital; Emmy Merkley, RN, Bev Armstrong, RN,University of Alberta Hospital, Edmonton; Ann Taite, RN,Kingston General Hospital; Karen Furlong, RN, St. John Re-gional Hospital; Carol Diemer, RN, Peggy Oldfield, RN, HotelDieu-Grace Hospital, Windsor.

References1 Corwin HL, Parsonnet KC, Gettinger A. RBC transfusion in

the ICU: is there a reason? Chest 1995; 108:767–771




2 Swinamer DL, Fedoruk LM, Jones RL, et al. Energy expen-diture associated with CPAP and T-piece spontaneous venti-latory trials: changes after prolonged mechanical ventilation.Chest 1989; 96:867–872

3 Nelson AH, Fleisher LA, Rosenbaum SH. Relationship be-tween postoperative anemia and cardiac morbidity in high-risk vascular patients in the ICUs. Crit Care Med 1993;21:860–866

4 Elia S, Liu P, Hilgenberg A, et al. Coronary hemodynamicsand myocardial metabolism during weaning from mechanicalventilation in cardiac surgical patients. Can J Anesth 1991;38:564–571

5 Jubran A, Tobin MJ. Pathophysiologic basis of acute respira-tory distress in patients who fail a trial of weaning frommechanical ventilation. Am J Respir Crit Care Med 1997;155:906–915

6 Mohsenifar Z, Hay A, Hay J, et al. Gastric intramural pH asa predictor of success or failure in weaning patients frommechanical ventilation. Ann Intern Med 1993; 119:794–798

7 Schonhofer B, Wenzel M, Geibel M, et al. Blood transfusionand lung function in chronically anemic patients with severechronic obstructive pulmonary disease. Crit Care Med 1998;26:1824–1828

8 Schonhofer B, Bohrer H, Kohler D. Blood transfusion facil-itating difficult weaning from the ventilator. Anesthesia 1998;53:181–184

9 Hebert PC, Wells G, Blajchman MA, et al. A multicenter,randomized, controlled clinical trial of transfusion require-ments in critical care. N Engl J Med 1999; 340:409–417

10 Bordin JO, Heddle NM, Blajchman MA. Biologic effects ofleukocytes present in transfused cellular blood products.Blood 1994; 84:1703–1721

11 Bacalo A, Kivity S, Heno N, et al. Blood transfusion and lungfunction in children with thalassemia major. Chest 1992;101:362–365

12 Santamaria F, Villa MP, Werner B, et al. The effect oftransfusion on pulmonary function in patients with thalasse-mia major. Pediatr Pulmonol 1994; 18:139–143

13 Srivastava S, Chatila W, Amoateng-Adjepong Y, et al. Myo-cardial ischemia and weaning failure in patients with coronary

artery disease: an update. Crit Care Med 1999; 27:2109–211214 Brochard L, Rauss A, Benito S, et al. Comparison of three

methods of gradual withdrawal from ventilatory supportduring weaning from mechanical ventilation. Am J RespirCrit Care Med 1994; 150:896–903

15 Herbertson MJ, Werner HA, Goddard CM, et al. Anti-tumornecrosis factor-a prevents decreased ventricular contractilityin endotoxemic pigs. Am J Respir Crit Care Med 1995;152:480–488

16 Ely EW, Baker AM, Dunagan DP, et al. Effect on theduration of mechanical ventilation of identifying patientscapable of breathing spontaneously. N Engl J Med 1996;335:1864–1869

17 Kollef MH, Shapiro SD, Silver P, et al. A randomized,controlled trial of protocol-directed versus physician-directedweaning from mechanical ventilation. Crit Care Med 1997;25:567–574

18 Nava S, Ambrosino N, Clini E, et al. Noninvasive mechanicalventilation in the weaning of patients with respiratory failurebecause of chronic obstructive pulmonary disease: a random-ized, controlled trial. Ann Intern Med 1998; 128:721–728

19 Girault C, Daudenthun I, Chevron V, et al. Noninvasiveventilation as a systematic extubation and weaning techniquein acute-on-chronic respiratory failure: a prospective, ran-domized controlled study. Am J Respir Crit Care Med 1999;160:86–92

20 Esteban A, Frutos F, Tobin MJ, et al. A comparison of fourmethods of weaning patients from mechanical ventilation.N Engl J Med 1995; 332:345–350

21 Manthous CA, Schmidt GA, Hall JB. Liberation from me-chanical ventilation: a decade of progress. Chest 1998; 114:886–901

22 Esteban A, Alia I, Gordo F, et al. Extubation outcome afterspontaneous breathing trials with T-tube or pressure supportventilation: Spanish Lung Failure Collaborative Group. Am JRespir Crit Care Med 1999; 156:459–464

23 Esteban A, Alia I, Tobin MJ, et al. Effect of spontaneousbreathing trial duration on outcome of attempts to discon-tinue mechanical ventilation: Spanish Lung Failure Collabo-rative Group. Am J Respir Crit Care Med 1999; 159:512–518

CHEST / 119 / 6 / JUNE, 2001 1857



DOI 10.1378/chest.119.6.1850 2001;119;1850-1857 Chest

George Wells, John Marshall and Irwin Schweitzer Paul C. Hébert, Morris A. Blajchman, Deborah J. Cook, Elizabeth Yetisir,

Ventilation?Do Blood Transfusions Improve Outcomes Related to Mechanical

This information is current as of June 11, 2008

& ServicesUpdated Information

http://chestjournal.org/cgi/content/full/119/6/1850high-resolution figures, can be found at: Updated information and services, including

References

http://chestjournal.org/cgi/content/full/119/6/1850#BIBLfor free at: This article cites 23 articles, 17 of which you can access

Citations

http://chestjournal.org/cgi/content/full/119/6/1850articles: This article has been cited by 8 HighWire-hosted

Permissions & Licensing

http://chestjournal.org/misc/reprints.shtml(figures, tables) or in its entirety can be found online at: Information about reproducing this article in parts

Reprints http://chestjournal.org/misc/reprints.shtml

Information about ordering reprints can be found online:

Email alerting service

online article. article sign up in the box at the top right corner of the Receive free email alerts when new articles cite this

Images in PowerPoint format

format. See any online article figure for directions. downloaded for teaching purposes in PowerPoint slide Figures that appear in CHEST articles can be


http://chestjournal.org/cgi/content/full/119/6/1850

http://chestjournal.org/cgi/content/full/119/6/1850#BIBL

http://chestjournal.org/cgi/content/full/119/6/1850

http://chestjournal.org/misc/reprints.shtml

http://chestjournal.org/misc/reprints.shtml


Do Blood Transfusions Improve Outcomes Related to Mechanical Ventilation?\u003cxref rid=\"AFF1\"\u003e\u003csup\u003e*\u003c/sup\u003e\u003c/xref\u003e

Documents