Next Article in Journal
The Role of Neutrophil-to-Lymphocyte Ratio and Right Ventricular Dysfunction in Indonesian Patients with COVID-19: A Retrospective Cohort Study
Next Article in Special Issue
High-Flow Nasal Cannula Application After Extubation in Acute Respiratory Failure Patients
Previous Article in Journal
Factors Affecting Recurrence in 165 Patients with Serous Borderline Ovarian Tumours: The Pattern of Micro-Invasion Is Main Prognostic Factor
Previous Article in Special Issue
Bilateral Serratus Plane Block in a Critically Ill, Mechanically Ventilated Patient with Multiple Rib Fractures Due to Severe Thoracic Trauma: Case Report and Literature Review
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
JCM
Volume 14
Issue 6
10.3390/jcm14062049
jcm-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Systematic Review

Mortality in Critically Ill Patients with Liberal Versus Restrictive Transfusion Thresholds: A Systematic Review and Meta-Analysis of Randomized Controlled Trials with Trial Sequential Analysis

by
Daniel Arturo Jiménez Franco
1,2,
Camilo Andrés Pérez Velásquez
1,2,* and
David Rene Rodríguez Lima
2,3
1
Escuela de Medicina y Ciencias de la Salud, Universidad del Rosario, Bogotá 111221, Colombia
2
Critical and Intensive Care Medicine, Hospital Universitario Mayor-Mederi, Bogotá 111411, Colombia
3
Grupo de Investigación Clínica, Escuela de Medicina y Ciencias de la Salud, Universidad del Rosario, Bogotá 111711, Colombia
*
Author to whom correspondence should be addressed.
J. Clin. Med. 2025, 14(6), 2049; https://doi.org/10.3390/jcm14062049
Submission received: 26 February 2025 / Revised: 14 March 2025 / Accepted: 16 March 2025 / Published: 18 March 2025
(This article belongs to the Special Issue Clinical Advances in Critical Care Medicine)
Browse Figures
Versions Notes

Abstract

:
Background/Objectives: Anemia is common in critically ill patients, yet red blood cell (RBC) transfusion without active bleeding does not consistently improve outcomes and carries risks such as pulmonary injury, fluid overload, and increased costs. Optimal transfusion thresholds remain debated, with some guidelines recommending a restrictive target of 7 g/dL instead of a more liberal target of 9 g/dL. Methods: We conducted a systematic review and meta-analysis following PRISMA guidelines, searching PubMed, EMBASE, and LILACS from January 1995 to October 2024. Thirteen randomized controlled trials involving 13,705 critically ill adults were included, with 6855 assigned to liberal and 6850 to restrictive transfusion strategies. The risk of bias was assessed using the Cochrane Risk of Bias Tool 2, and the pooled effect sizes were estimated with a random-effects model. We registered the protocol in PROSPERO International Prospective Register of Systematic Reviews (CDR42024589225). Results: No statistically significant difference was observed in 30-day mortality between restrictive and liberal strategies (odds ratio [OR] 1.02; 95% confidence interval [CI], 0.83–1.25; I2 = 49%). Similarly, no significant differences emerged for the 90-day or 180-day mortality, hospital or intensive care unit (ICU) length of stay, dialysis requirement, or incidence of acute respiratory distress syndrome (ARDS). However, patients in the restrictive group received significantly fewer RBC units. The trial sequential analysis (TSA) indicated that the evidence accrued was insufficient to definitively confirm or exclude an effect on the 30-day mortality, as the required sample size was not reached. Conclusions: In conclusion, while our meta-analysis found no statistically significant difference in the short-term mortality between restrictive and liberal transfusion strategies, larger trials are needed to fully determine whether any clinically meaningful difference exists in critically ill populations.

Graphical Abstract

1. Introduction

Anemia is a common problem among critically ill patients and is associated with adverse outcomes [1]. The primary rationale for correcting anemia in this population is to enhance oxygen delivery to tissues and mitigate potential hypoxic injury by increasing the oxygen-carrying capacity [2]. Although anemia has been associated with a higher mortality risk in critically ill patients [3], the practice of administering red blood cell (RBC) transfusions—particularly in the absence of active hemorrhage—has not consistently demonstrated unambiguous clinical benefit [4]. Furthermore, evidence increasingly indicates that RBC transfusions introduce significant risks, including acute transfusional reactions like transfusion-related acute lung injury, volume overload, infectious complications, and higher healthcare costs [5,6].
Transfusion indications and targets for critically ill patients remain controversial [7]. Currently, thresholds for restrictive and liberal transfusion strategies are defined primarily by absolute hemoglobin values, with limited consideration of additional triggers such as systemic perfusion, the impact of anemia, and oxygenation parameters. Moreover, there is considerable heterogeneity in the hemoglobin cut-off used to distinguish restrictive from liberal transfusion strategies. Recent guidelines from the European Society of Intensive Care Medicine (ESICM) recommend a restrictive transfusion threshold (hemoglobin 7 g/dL) over a liberal threshold (hemoglobin 9 g/dL), based on the moderate certainty of evidence and the absence of reported harms in most clinical contexts when employing a restrictive strategy [8]. Nevertheless, the randomized clinical trials supporting these recommendations have methodological limitations that may affect the confidence in the findings, owing to the imprecision or inconsistency across studies [9].
These uncertainties have contributed to considerable variability in transfusion practices, particularly in subpopulations with heightened susceptibility to complications associated with anemia, such as patients with acute myocardial infarction, those undergoing mechanical ventilation weaning, individuals with septic shock, or those with acute brain injury [7,10]. Although several recent investigations have attempted to clarify the optimal transfusion threshold, persistent questions remain, underscoring the importance of more definitive, large-scale data in guiding clinical decisions.
Consequently, we conducted a systematic review and meta-analysis encompassing a general intensive care unit (ICU) population, supplemented by a trial sequential analysis (TSA). The objective was to ascertain whether the observed effect of restrictive versus liberal transfusion thresholds on mortality is sufficiently precise to remain unchanged with further research. By evaluating whether the cumulative evidence meets the required sample size and significance boundaries, the TSA helps to determine if future studies are likely to alter the current conclusions. Ultimately, this approach provides a clearer perspective on whether restrictive transfusion strategies can be confidently recommended for critically ill patients in a broad range of clinical scenarios.

2. Materials and Methods

2.1. Protocol

This meta-analysis adhered to the guidelines outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [11]. The protocol was registered in International Prospective Register of Systematic Reviews (PROSPERO) in December 2024 (CDR42024589225).

2.2. Search Strategy and Data Extraction

A comprehensive search strategy was implemented via the PubMed, EMBASE, and Latin American and Caribbean Health Sciences Literature (LILACS) databases, covering the period from January 1995 to October 2024. Data extraction and eligibility assessment were standardized and carried out independently by two reviewers (J.F.D.A. and R.L.D.R.). In case of any disagreements, a third evaluator (P.V.C.A.) was included. Titles and abstracts were independently screened in duplicate using the Rayyan tool [12], and full-text versions of the relevant studies were subsequently retrieved. No language restrictions were applied. The complete search strategy, including the terms used, is detailed in the protocol registered in PROSPERO (registration number 589225).

2.3. Study Selection and Inclusion Criteria

The selection and inclusion of studies was based on the PICOS strategy, as described below:
  • Patients and Setting: We included studies with adult critical care patients diagnosed with anemia. Patients under 18 years of age, pregnant individuals, or those with limitations of therapeutic effort were excluded.
  • Interventions: Restrictive transfusion of RBCs with a target hemoglobin between 7 and 9 g/dL.
  • Comparison: Liberal transfusion of RBCs with a target hemoglobin greater than 9 g/dL.
  • Outcomes: Studies that evaluated mortality at 28 to 30 days as the primary outcome were included. If multiple data points were available within a study, all of the relevant data were considered.
  • Study Type: Only randomized clinical trials were included. Exclusions were made for studies with standardized transfusion protocols, such as preoperative optimization in major surgery, orthopedic surgery, or cardiovascular surgery.

2.4. Study Selection and Data Collection Process

For studies meeting the inclusion criteria, information was compiled into individual extraction sheets to facilitate comparison. Any disagreements regarding inclusion, study quality, data adequacy, or the final classification of included studies were resolved through author consensus.

2.5. Data Items

Primary data extraction was conducted independently by two evaluators, who recorded their findings using standardized reporting forms (J.F.D.A. and R.L.D.R.). In cases of disagreement, a third evaluator was consulted to achieve consensus (P.V.C.A.). Extracted data included the authors, title, year of publication, study population, inclusion and exclusion criteria, number of participants, restrictive transfusion goal, liberal transfusion goal, primary outcome, and secondary outcomes.

2.6. Risk of Bias in Individual Studies

The risk of bias assessment was performed using the Cochrane Risk of Bias Tool 2 (RoB 2), which focuses on the following five domains: randomization, intervention, missing outcome data, outcome measurement, and outcome reporting [13]. Accordingly, the risk of bias was categorized individually for each outcome as low, high, or unclear. The grade of evidence assessment was also performed according to the recommendations of the Grading of Recommendations Assessment, Development, and Evaluation working group using the GRADEpro software (v.3.0.; McMaster University).

2.7. Statistical Analysis

2.7.1. Analysis of Individual Studies and Summary Measures

The meta-analysis was conducted using the R statistical package (version 4.3.1) and the DerSimonian–Laird random-effects model. For binary outcomes, odds ratios were estimated to evaluate the association between categorical variables of interest, and the Mantel–Haenszel method was employed to determine the study weights. Continuous outcomes were assessed using the Hedges method, with the results reported as mean differences. A 95% prediction interval was also provided for the primary analysis. Statistical significance was set at p < 0.05.
Heterogeneity was assessed using the Chi2 test, with p > 0.01 indicating consistency across studies. In addition, Higgins’ I2 index was calculated to quantify the proportion of variability in effect sizes not attributable to chance. Heterogeneity was categorized as follows: I2 = 0–24.9% (none), I2 = 25.0–49.9% (low), I2 = 50.0–74.9% (moderate), and I2 > 75% (high).

2.7.2. Analysis of Risk of Bias Across Studies

Publication bias was assessed graphically using the funnel plot and Egger’s linear regression.

2.7.3. Subgroup Analysis and Trial Sequential Analysis

The robustness of the primary results was evaluated using TSA conducted in R Studio (v.4.3.1) with the RTSA statistical package. The estimated effect was assessed using the OR obtained from the forest plot, and the cumulative evidence was compared against trial sequential monitoring boundaries. For this analysis, a random-effects model was employed, assuming an 80% power, a 5% type I error, and a clinical significance threshold defined by an OR of 0.80. By applying the TSA, we determined whether the accrued evidence reached the required information size to draw firm conclusions or if additional studies would be necessary or futile to confirm the observed effect. Additionally, a subgroup analysis was conducted according to the specific intensive care populations examined (medical ICU, sepsis, cardiovascular, burns, and acute brain injury) to evaluate the effect of a liberal transfusion strategy compared with a restrictive transfusion approach on 30-day mortality outcomes.

3. Results

3.1. Study Selection

The primary search identified 71 studies. A total of 54 articles were excluded due to duplicates or failure to meet the inclusion criteria. Additionally, three studies were excluded because the original database could not be acquired for data extraction. Finally, 13 randomized clinical trials were included for review and final analysis [14,15,16,17,18,19,20,21,22,23,24,25,26] (Figure 1).

3.2. Study Characteristics

A total of 13,705 patients were included in the meta-analysis, with 6855 in the liberal transfusion group and 6850 in the restrictive transfusion group. The general characteristics of the included studies are summarized in Table 1. Among the 13 studies, 5 were conducted in patients with cardiovascular conditions [16,18,21,23,25], 2 in those with septic shock [17,20], 3 in medical ICU patients [14,15,19], 1 in burn patients [22], and 2 in acute brain injury patients [24,26]. Of these, seven studies used a liberal hemoglobin target between 9 and 10 g/dL, while six studies used a target greater than 10 g/dL. The general characteristics of the included studies are summarized in Table 1.

3.3. Syntheses of Results

The primary 30-day mortality outcome was reported by 12 studies, and the meta-analysis showed no significant difference between groups (OR 1.02; 95% CI 0.83–1.25; p = 0.03, I2 = 49%). The 90-day mortality outcome was reported by three studies, and this analysis also revealed no significant differences between groups (OR 1.40; 95% CI 0.35–5.55; p = 0.03, I2 = 72%). The 180-day mortality outcome was assessed by three studies, showing no significant differences between groups (OR 0.93; 95% CI 0.19–4.69; p = 0.10, I2 = 57%) (Figure 2).
Hospital length of stay was reported in four studies and showed no significant differences between groups (Standardized Mean Difference [SMD] −0.05; 95% CI −0.24 to 0.14; p = 0.05; I2 = 62%). Similarly, the ICU length of stay was reported by three studies and revealed no significant differences between groups (SMD −0.04; 95% CI −0.36 to 0.27; p = 0.12; I2 = 53%) (Supplementary Material Figure S1).
Among the secondary outcomes assessed, the risk of requiring dialysis was reported by four studies, demonstrating no significant differences between groups (OR 0.86; 95% CI 0.35–2.11; p = 0.04; I2 = 63%). The incidence of ARDS, reported by three studies, also showed no significant differences (OR 0.65; 95% CI 0.07–6.12; p = 0.05; I2 = 66%) (Supplementary Material Figure S1). Finally, the number of RBC units transfused was reported by seven studies and was significantly lower in the restrictive transfusion strategy group (SMD −0.47; 95% CI −0.91 to −0.03; p < 0.01; I2 = 98%) (Figure 2).

3.4. Risk of Bias

Risk of bias was assessed using the RoB-2 tool (Figure 3), showing that most studies were classified as having a low overall risk of bias. However, all of the studies were deemed to have a high risk of bias in the performance and detection bias domains. Consequently, a sensitivity analysis was conducted to evaluate the individual contribution of each study to the overall effect estimate. The grade of evidence is shown in Supplementary Material Figure S2.

3.5. Additional Analysis

3.5.1. Sensitivity Analysis

A sensitivity analysis was performed to assess the individual contribution of each study to the overall effect estimate, and no single study demonstrated a statistically significant impact on the results (Supplementary Material Figure S3).

3.5.2. Trial Sequential Analysis

The TSA was performed to assess the robustness of the findings for the 30-day mortality outcome, using an OR of 1.02 derived from the primary meta-analysis. This analysis indicated that the results were not robust, as the cumulative Z-score remained in the nonsignificant region (TSA-adjusted OR 0.74–1.39; p = 0.8680) (Figure 4). Moreover, the total sample size analyzed represented only 41% of the required 31,906 participants, which would be needed to achieve a 90% statistical power for detecting a significant effect on 30-day mortality.

3.5.3. Subgroup Analysis

A subgroup analysis was performed according to the included populations—medical ICU, sepsis, and cardiovascular. The burns and acute brain injury subgroups were excluded, as only one study in each group assessed the 30-day mortality. In the subgroup analysis, no statistically significant difference in the risk of 30-day mortality was observed between the liberal and restrictive transfusion strategies (Supplementary Material Figure S4).

3.5.4. Publication Bias

No evidence of publication bias was detected based on visual inspection of the funnel plot, which showed a symmetrical distribution of the points (Supplementary Material Figure S5).

4. Discussion

In this meta-analysis, the use of restrictive transfusion strategies (7–9 g/dL) compared to liberal strategies (>9 g/dL) in critically ill patients did not show a significant reduction in the 30-day mortality (OR 1.02; 95% CI 0.83–1.25; p = 0.03; I2 = 49%). After performing the TSA, the cumulative Z-score line remained in the zone of nonsignificance and did not reach the required sample size to either refute or confirm the hypothesis. Consequently, these results are not conclusive, and additional studies are necessary to determine the magnitude of the effect. Nevertheless, while additional large-scale studies are required to conclusively exclude any small yet potentially significant advantages of restrictive transfusion strategies, the current evidence indicates that, if such an effect exists, its magnitude is likely too modest to have a meaningful clinical impact. Overall, our findings highlight the absence of harm associated with using restrictive transfusion strategies and underscore the potential to reduce costs and adverse events associated with the increased number of RBC units required under a liberal transfusion strategy. Furthermore, these findings indicate that a restrictive transfusion strategy is both safer and more cost-effective than a liberal approach, and that employing a liberal transfusion strategy based on higher hemoglobin levels does not offer sufficient clinical benefit to justify its continued use.
In critically ill patients, hypoxia reflects an imbalance between oxygen delivery and consumption across various organ systems, often arising from a confluence of factors, such as impaired pulmonary function, hemodynamic instability, and reduced oxygen-carrying capacity [27]. Prolonged periods of hypoxia, commonly referred to as the “oxygen debt”, have been linked to adverse outcomes, including organ failure and elevated mortality [28,29]. Anemia frequently exacerbates this condition by directly diminishing hemoglobin’s ability to transport oxygen to vital tissues, especially in individuals with compromised cardiovascular reserve [30]. Restrictive transfusion strategies, which typically employ hemoglobin thresholds of 7–9 g/dL, are often viewed to minimize exposure to allogeneic red blood cells (RBCs). By contrast, liberal strategies set higher targets, generally above 9 g/dL, on the premise that an increased oxygen-carrying capacity could improve the clinical outcomes and mitigate organ hypoxia [3]. Despite this intuitive rationale, evidence from multiple randomized controlled trials (RCTs) has challenged the necessity of liberal thresholds in the absence of acute hemorrhage [4,5,6]. Nevertheless, it is crucial to recognize that oxygen delivery in critical illness is multifactorial. While liberal transfusions might bolster hemoglobin levels, other determinants, such as cardiac output, microcirculatory function, ventilation status, and the patient’s underlying pathology, ultimately govern tissue oxygenation [12]. As such, the hypothesized benefits of higher hemoglobin targets may not always translate into demonstrable improvements in survival or organ function.
Most studies assessing the impact of transfusion strategies on mortality risk have used a restrictive transfusion threshold (7–9 g/dL) as the intervention group, requiring the demonstration of either benefit or no harm compared with more liberal thresholds [31]. However, previous meta-analyses [8,9,32], including our own, have not identified an increased risk of adverse outcomes with restrictive strategies, suggesting that the mortality risk likely does not increase, and that other safety- and cost-related outcomes may potentially be improved.
The body of evidence on this topic has grown considerably over the past 50 years, albeit with inconclusive results [33]. Traditional meta-analyses can evaluate the pooled effect size of an intervention on a specific outcome, but they are limited in determining whether the sample size and statistical power are sufficient to accept or reject the study hypothesis [34]. TSA is a frequentist method that, based on accepted type I and type II error thresholds, assesses whether meta-analysis results are robust enough to remain unaffected by subsequent studies [35]. This approach allows for the confirmation or rejection of the pooled effect estimate and clarifies whether the required sample size has been reached to draw definitive conclusions [36].
In our study, we prespecified an OR of 0.80 as the threshold for clinical significance for 30-day mortality. Assuming an 80% power and a 5% type I error rate, it was not possible to demonstrate the absence of an effect due to an inadequate sample size. These findings suggest that, if there is any benefit in terms of reducing the 30-day mortality in the restrictive transfusion group, it is likely too small to be clinically meaningful, and thus does not justify further studies using the same methodological design to confirm it. This approach suggests that additional criteria, beyond relying solely on hemoglobin levels, should be incorporated and evaluated in clinical studies to demonstrate the impact of correcting anemic hypoxia in critically ill patients without active bleeding [30].
This meta-analysis has several limitations. First, although we included all of the clinical trials that compared restrictive and liberal transfusion strategies in critically ill patients, the total number of patients and events did not reach the required sample size. Second, the initial heterogeneity among the studies was moderate; however, the subgroup analyses did not demonstrate any differences in effect, and the sensitivity analyses indicated that no single study substantially contributed to this heterogeneity. Still, there is likely clinical heterogeneity given the inclusion of a general ICU population, which may limit the external validity of the results, particularly for populations vulnerable to hypoxia.

5. Conclusions

In conclusion, our meta-analysis did not demonstrate a statistically significant difference in the 30-day mortality when comparing restrictive versus liberal transfusion strategies in a general ICU population (OR 1.02; 95% CI 0.83–1.25; p = 0.03; I2 = 49%). Moreover, the TSA suggests that these findings are neither robust nor conclusive, indicating the need for additional randomized studies are required to provide convincing outcome data in the selected patient population to achieve sufficient statistical power.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/jcm14062049/s1, Figure S1: Forest plots for A. renal replacement therapy, B. acute respiratory distress syndrome, C. ICU stay, and D. hospital stay; Figure S2: Grade of evidence assessment; Figure S3: Sensitivity analysis; Figure S4: Subgroup analysis 30-days mortality. Figure S5. Funnel Plot for 30-days mortality.

Author Contributions

Design: D.A.J.F., C.A.P.V. and D.R.R.L.; Literature review: D.A.J.F., C.A.P.V. and D.R.R.L.; Acquisition of data: D.A.J.F., C.A.P.V. and D.R.R.L.; Statistical analysis: D.A.J.F., C.A.P.V. and D.R.R.L.; Interpretation of data: D.A.J.F., C.A.P.V. and D.R.R.L.; Wrote the manuscript: D.A.J.F., C.A.P.V. and D.R.R.L. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding. Additionally, the author(s) received no financial support for the research, authorship, and/or publication of this manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Supporting reported results can be found in the Supplementary Materials.

Acknowledgments

Hospital Universitario Mayor-Mederi.

Conflicts of Interest

The author(s) declare that they do not have conflicting interests.

Abbreviations

The following abbreviations are used in this manuscript: ARDS: acute respiratory distress syndrome; CI: confidence interval; ESICM: European Society of Intensive Care Medicine; GRADE: Grading of Recommendations Assessment, Development, and Evaluation software; ICU: intensive care unit; LILACS: Latin American and Caribbean Health Sciences Literature; OR: odds ratio; PICOS: Population, Intervention, Comparison, Outcomes, Study design; PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses; PROSPERO: International Prospective Register of Systematic Reviews; RBC: red blood cell; RoB 2: Cochrane Risk of Bias Tool 2; SMD: standardized mean difference; TSA: trial sequential analysis.

References

  1. Warner, M.A.; Hanson, A.C.; Frank, R.D.; Schulte, P.J.; Go, R.S.; Storlie, C.B.; Kor, D.J. Prevalence of and Recovery From Anemia Following Hospitalization for Critical Illness Among Adults. JAMA Netw. Open 2020, 3, e2017843. [Google Scholar] [CrossRef] [PubMed]
  2. McKenna, H.T.; Murray, A.J.; Martin, D.S. Human adaptation to hypoxia in critical illness. J. Appl. Physiol. 2020, 129, 656–663. [Google Scholar] [CrossRef]
  3. Lin, I.-H.; Liao, P.-Y.; Wong, L.-T.; Chan, M.-C.; Wu, C.-L.; Chao, W.-C. Anaemia in the first week may be associated with long-term mortality among critically ill patients: Propensity score-based analyses. BMC Emerg. Med. 2023, 23, 32. [Google Scholar] [CrossRef]
  4. Carson, J.L.; Stanworth, S.J.; Alexander, J.H.; Roubinian, N.; Fergusson, A.D.; Triulzi, D.J.; Goodman, S.G.; Rao, S.V.; Doree, C.; Hebert, P.C. Clinical trials evaluating red blood cell transfusion thresholds: An updated systematic review and with additional focus on patients with cardiovascular disease. Am. Heart J. 2018, 200, 96–101. [Google Scholar] [CrossRef] [PubMed]
  5. Blet, A.; McNeil, J.B.; Josse, J.; Cholley, B.; Cinotti, R.; Cotter, G.; Dauvergne, A.; Davison, B.; Duarte, K.; Duranteau, J.; et al. Association between in-ICU red blood cells transfusion and 1-year mortality in ICU survivors. Crit. Care 2022, 26, 307. [Google Scholar] [CrossRef]
  6. Trentino, K.M.; Farmer, S.L.; Swain, S.G.; Burrows, S.A.; Hofmann, A.; Ienco, R.; Pavey, W.; Daly, F.F.; Van Niekerk, A.; Webb, S.A.; et al. Increased hospital costs associated with red blood cell transfusion. Transfusion 2015, 55, 1082–1089. [Google Scholar] [CrossRef]
  7. Willems, S.A.; Kranenburg, F.J.; Le Cessie, S.; Marang- van de Mheen, P.J.; Kesecioglu, J.; van der Bom, J.G.; Arbous, M.S. Variation in red cell transfusion practice in the intensive care unit—An international survey. J. Crit. Care 2020, 55, 140–144. [Google Scholar] [CrossRef]
  8. Vlaar, A.P.; Oczkowski, S.; de Bruin, S.; Wijnberge, M.; Antonelli, M.; Aubron, C.; Aries, P.; Duranteau, J.; Juffermans, N.P.; Meier, J.; et al. Transfusion strategies in non-bleeding critically ill adults: A clinical practice guideline from the European Society of Intensive Care Medicine. Intensive Care Med. 2020, 46, 673–696. [Google Scholar] [CrossRef] [PubMed]
  9. Zhang, W.; Zheng, Y.; Yu, K.; Gu, J. Liberal Transfusion versus Restrictive Transfusion and Outcomes in Critically Ill Adults: A Meta-Analysis. Transfus. Med. Hemotherapy 2020, 48, 60–68. [Google Scholar] [CrossRef]
  10. Guglielmi, A.; Graziano, F.; Bogossian, E.G.; Turgeon, A.F.; Taccone, F.S.; Citerio, G.; The CENTER-TBI Participants and Investigators. Haemoglobin values, transfusion practices, and long-term outcomes in critically ill patients with traumatic brain injury: A secondary analysis of CENTER-TBI. Crit. Care 2024, 28, 199. [Google Scholar] [CrossRef]
  11. Page, M.J.; McKenzie, J.E.; Bossuyt, P.M.; Boutron, I.; Hoffmann, T.C.; Mulrow, C.D.; Shamseer, L.; Tetzlaff, J.M.; Akl, E.A.; Brennan, S.E.; et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ 2021, 372, n71. [Google Scholar] [CrossRef] [PubMed]
  12. Ouzzani, M.; Hammady, H.; Fedorowicz, Z.; Elmagarmid, A. Rayyan—A web and mobile app for systematic reviews. Syst. Rev. 2016, 5, 210. [Google Scholar] [CrossRef]
  13. Sterne, J.A.C.; Savović, J.; Page, M.J.; Elbers, R.G.; Blencowe, N.S.; Boutron, I.; Cates, C.J.; Cheng, H.Y.; Corbett, M.S.; Eldridge, S.M.; et al. RoB 2: A revised tool for assessing risk of bias in randomised trials. BMJ 2019, 366, l4898. [Google Scholar] [CrossRef] [PubMed]
  14. Hébert, P.C.; Wells, G.; Marshall, J.; Martin, C.; Tweeddale, M.; Pagliarello, G.; Blajchman, M. Transfusion Requirements in Critical Care: A Pilot Study. JAMA 1995, 273, 1439–1444. [Google Scholar] [CrossRef]
  15. Hébert, P.C.; Wells, G.; Blajchman, M.A.; Marshall, J.; Martin, C.; Pagliarello, G.; Tweeddale, M.; Schweitzer, I.; Yetisir, E. A Multicenter, Randomized, Controlled Clinical Trial of Transfusion Requirements in Critical Care. N. Engl. J. Med. 2024, 340, 409–417. [Google Scholar] [CrossRef]
  16. Hébert, P.; Yetisir, E.; Martin, C.; Blajchman, M.; Wells, G.; Marshall, J.; Tweeddale, M.; Pagliarello, G.; Schweitzer, I. Is a low transfusion threshold safe in critically ill patients with cardiovascular diseases? Crit. Care Med. 2001, 29, 227–234. [Google Scholar] [CrossRef] [PubMed]
  17. Bergamin, F.S.; Almeida, J.P.; Landoni, G.; Galas, F.R.B.G.; Fukushima, J.T.; Fominskiy, E.; Park, C.H.L.; Osawa, E.A.; Diz, M.P.E.; Oliveira, G.Q.; et al. Liberal Versus Restrictive Transfusion Strategy in Critically Ill Oncologic Patients: The Transfusion Requirements in Critically Ill Oncologic Patients Randomized Controlled Trial*. Crit. Care Med. 2017, 45, 766–773. [Google Scholar] [CrossRef] [PubMed]
  18. Carson, J.L.; Brooks, M.M.; Abbott, J.D.; Chaitman, B.; Kelsey, S.F.; Triulzi, D.J.; Srinivas, V.; Menegus, M.A.; Marroquin, O.C.; Rao, S.V.; et al. Liberal Versus Restrictive Transfusion Thresholds for Patients with Symptomatic Coronary Artery Disease. Am. Heart J. 2013, 165, 964–971.e1. [Google Scholar] [CrossRef]
  19. Walsh, T.S.; Boyd, J.A.; Watson, D.; Hope, D.; Lewis, S.; Krishan, A.; Forbes, J.F.; Ramsay, P.; Pearse, R.; Wallis, C.; et al. Restrictive Versus Liberal Transfusion Strategies for Older Mechanically Ventilated Critically Ill Patients. Crit. Care Med. 2013, 41, 2354–2363. [Google Scholar] [CrossRef]
  20. Holst, L.B.; Haase, N.; Wetterslev, J.; Wernerman, J.; Guttormsen, A.B.; Karlsson, S.; Johansson, P.I.; Åneman, A.; Vang, M.L.; Winding, R.; et al. Lower versus Higher Hemoglobin Threshold for Transfusion in Septic Shock. N. Engl. J. Med. 2014, 371, 1381–1391. [Google Scholar] [CrossRef]
  21. Mazer, C.D.; Whitlock, R.P.; Fergusson, D.A.; Hall, J.; Belley-Cote, E.; Connolly, K.; Khanykin, B.; Gregory, A.J.; de Médicis, É.; McGuinness, S.; et al. Restrictive or Liberal Red-Cell Transfusion for Cardiac Surgery. N. Engl. J. Med. 2024, 377, 2133–2144. [Google Scholar] [CrossRef] [PubMed]
  22. Palmieri, T.L.; Holmes, J.H., 4th; Arnoldo, B.; Peck, M.; Potenza, B.; Cochran, A.; King, B.T.; Dominic, W.; Cartotto, R.; Bhavsar, D.; et al. Transfusion Requirement in Burn Care Evaluation (TRIBE): A Multicenter Randomized Prospective Trial of Blood Transfusion in Major Burn Injury. Ann. Surg. 2017, 266, 595–602. [Google Scholar] [CrossRef]
  23. Ducrocq, G.; Gonzalez-Juanatey, J.R.; Puymirat, E.; Lemesle, G.; Cachanado, M.; Durand-Zaleski, I.; Arnaiz, J.A.; Martínez-Sellés, M.; Silvain, J.; Ariza-Solé, A.; et al. Effect of a Restrictive vs Liberal Blood Transfusion Strategy on Major Cardiovascular Events Among Patients With Acute Myocardial Infarction and Anemia: The REALITY Randomized Clinical Trial. JAMA 2021, 325, 552–560. [Google Scholar] [CrossRef] [PubMed]
  24. Turgeon, A.F.; Fergusson, D.A.; Clayton, L.; Patton, M.P.; Neveu, X.; Walsh, T.S.; Docherty, A.; Malbouisson, L.M.; Pili-Floury, S.; English, S.W.; et al. Liberal or Restrictive Transfusion Strategy in Patients with Traumatic Brain Injury. N. Engl. J. Med. 2024, 391, 722–735. [Google Scholar] [CrossRef] [PubMed]
  25. Carson, J.L.; Brooks, M.M.; Hébert, P.C.; Goodman, S.G.; Bertolet, M.; Glynn, S.A.; Chaitman, B.R.; Simon, T.; Lopes, R.D.; Goldsweig, A.M.; et al. Restrictive or Liberal Transfusion Strategy in Myocardial Infarction and Anemia. N. Engl. J. Med. 2023, 389, 2446–2456. [Google Scholar] [CrossRef]
  26. Taccone, F.S.; Rynkowski Bittencourt, C.; Møller, K.; Lormans, P.; Quintana-Díaz, M.; Caricato, A.; Cardoso Ferreira, M.A.; Badenes, R.; Kurtz, P.; Søndergaard, C.B.; et al. Restrictive vs Liberal Transfusion Strategy in Patients with Acute Brain Injury: The TRAIN Randomized Clinical Trial. JAMA 2024, 332, 1623–1633. [Google Scholar] [CrossRef]
  27. Siggaard-Andersen, O.; Fogh-Andersen, N.; Gothgen, I.H.; Larsen, V.H. Oxygen status of arterial and mixed venous blood. Crit. Care Med. 1995, 23, 1284–1293. [Google Scholar] [CrossRef]
  28. Shoemaker, W.C.; Appel, P.L.; Kram, H.B. Role of Oxygen Debt in the Development of Organ Failure Sepsis, and Death in High-Risk Surgical Patients. Chest 1992, 102, 208–215. [Google Scholar] [CrossRef]
  29. Perez-Garzon, M.; Poveda-Henao, C.; Bastidas-Goyes, A.; Robayo-Amortegui, H. Oxygen Debt as Predictor of Mortality and Multiple Organ Dysfunction Syndrome in Severe COVID-19 Patients: A Retrospective Study. J. Intensive Care Med. 2023, 39, 358–367. [Google Scholar] [CrossRef]
  30. Bunn, H. Oxygen Delivery in the Treatment of Anemia. N. Engl. J. Med. 2022, 387, 2362–2365. [Google Scholar] [CrossRef]
  31. Trentino, K.M.; Farmer, S.L.; Leahy, M.F.; Sanfilippo, F.M.; Isbister, J.P.; Mayberry, R.; Hofmann, A.; Shander, A.; French, C.; Murray, K. Systematic reviews and meta-analyses comparing mortality in restrictive and liberal haemoglobin thresholds for red cell transfusion: An overview of systematic reviews. BMC Med. 2020, 18, 154. [Google Scholar] [CrossRef] [PubMed]
  32. Braïk, R.; Jebali, S.; Blot, P.-L.; Egbeola, J.; James, A.; Constantin, J.-M. Liberal versus restrictive transfusion strategies in acute myocardial infarction: A systematic review and comparative frequentist and Bayesian meta-analysis of randomized controlled trials. Ann. Intensive Care 2024, 14, 150. [Google Scholar] [CrossRef] [PubMed]
  33. Yadav, S.; Hussein, G.; Liu, B.; Vojjala, N.; Warsame, M.; El Labban, M.; Rauf, I.; Hassan, M.; Zareen, T.; Usama, S.M.; et al. A Contemporary Review of Blood Transfusion in Critically Ill Patients. Medicina 2024, 60, 1247. [Google Scholar] [CrossRef] [PubMed]
  34. De Cassai, A.; Tassone, M.; Geraldini, F.; Sergi, M.; Sella, N.; Boscolo, A.; Munari, M. Explanation of trial sequential analysis: Using a post-hoc analysis of meta-analyses published in Korean Journal of Anesthesiology. Korean J. Anesthesiol. 2021, 74, 383–393. [Google Scholar] [CrossRef]
  35. Riberholt, C.G.; Olsen, M.H.; Milan, J.B.; Hafliðadóttir, S.H.; Svanholm, J.H.; Pedersen, E.B.; Lew, C.C.H.; Asante, M.A.; Ribeiro, J.P.; Wagner, V.; et al. Major mistakes or errors in the use of trial sequential analysis in systematic reviews or meta-analyses—The METSA systematic review. BMC Med. Res. Methodol. 2024, 24, 196. [Google Scholar] [CrossRef]
  36. Claire, R.; Gluud, C.; Berlin, I.; Coleman, T.; Leonardi-Bee, J. Using Trial Sequential Analysis for estimating the sample sizes of further trials: Example using smoking cessation intervention. BMC Med. Res. Methodol. 2020, 20, 284. [Google Scholar] [CrossRef]
Jcm 14 02049 g001
Figure 1. Flow chart of the study selection. LILACS: Latin American and Caribbean Health Sciences Literature. ICU: intensive care unit. EMBASE: Excerpta Medica Database. PubMed: Public MEDLINE National Library of Medicine United States [11].
Figure 1. Flow chart of the study selection. LILACS: Latin American and Caribbean Health Sciences Literature. ICU: intensive care unit. EMBASE: Excerpta Medica Database. PubMed: Public MEDLINE National Library of Medicine United States [11].
Jcm 14 02049 g001
Jcm 14 02049 g002
Figure 2. Forest plot for the main outcomes comparing restrictive and liberal transfusion strategies. (A). 30-day mortality. (B). 90-day mortality. (C). 180-day mortality. (D). Units of red blood cells (RBCs) transfused, green color indicates a statistically significant result [14,15,16,17,18,19,20,21,22,23,24,25,26].
Figure 2. Forest plot for the main outcomes comparing restrictive and liberal transfusion strategies. (A). 30-day mortality. (B). 90-day mortality. (C). 180-day mortality. (D). Units of red blood cells (RBCs) transfused, green color indicates a statistically significant result [14,15,16,17,18,19,20,21,22,23,24,25,26].
Jcm 14 02049 g002
Jcm 14 02049 g003
Figure 3. Distribution of the risk of bias assessment.
Figure 3. Distribution of the risk of bias assessment.
Jcm 14 02049 g003
Jcm 14 02049 g004
Figure 4. Trial sequential analysis. AIS: achieved information size; HARIS: heterogeneity-adjusted required information size for a non-sequential meta-analysis; MVD: mean value difference; DL_HKSJ: DerSimonian–Laird with Hartung–Knapp–Sidik–Jonkman adjustment; esOF: Lan and DeMets version of O’Brien–Fleming boundaries; TSA: trial sequential analysis.
Figure 4. Trial sequential analysis. AIS: achieved information size; HARIS: heterogeneity-adjusted required information size for a non-sequential meta-analysis; MVD: mean value difference; DL_HKSJ: DerSimonian–Laird with Hartung–Knapp–Sidik–Jonkman adjustment; esOF: Lan and DeMets version of O’Brien–Fleming boundaries; TSA: trial sequential analysis.
Jcm 14 02049 g004
Table 1. General characteristics of the included studies. Data are presented as No. (%), No./Total No. (%), percentage, or mean (SD) unless otherwise indicated.
Table 1. General characteristics of the included studies. Data are presented as No. (%), No./Total No. (%), percentage, or mean (SD) unless otherwise indicated.
StudyYearSample SizeAge in Years, Mean ± SD, Median (IQR)Sex, Male/FemalePopulationRestrictive Thresholdsn=Liberal Thresholdsn=30-Day Mortality (Restrictive/Liberal) no (%).
Hébert et al. [14]19956958.6 ± 1533/36Medical ICU7–9 g/dL3310–12 g/dL368 (24%)/9 (25%)
Hébert et al. [15]199983857.1 ± 18.1524/314Medical ICU7–9 g/dL41810–12 g/dL42078 (18.7%)/98 (23%)
Hebert et al. [16]200135764.0 ± 14.1216/141Cardiovascular7–9 g/dL16010–12 g/dL19736 (23%)/45 (23%)
Bergamin et al. [17]201730061.6 ± 12.9154/146Septic shock<7 g/dL151<9 g/dL14967 (45%)/84 (55.6%)
Carson et al. [18]201311070.8 ± 12.855/55Cardiovascular<8 g/dL55<10 g/dL557 (13%)/1 (1.8%)
Walsh et al. [19]201310067 ± 760/40Medical ICU7–9 g/dL519–11 g/dL4912 (23.5%)/16 (32.7%)
Holst et al. [20]201499867 (57–73)531/467Septic shock≤7 g/dL502<9 g/dL436168 (31.6%)/175 (34.8%)
Mazer et al. [21]2017485672 ± 103139/1717Cardiovascular≤7.5 g/dL2427≤9.5 g/dL242974 (3%)/87 (3.6%)
Palmieri et al. [22]201734541 (30–55)273/72Burns7 g/dL168≥10 g/dL17716 (9.5%)/15 (8.5%)
Ducrocq et al. [23]202166678 (69–85)201/184Cardiovascular8–10 g/dL342>11 g/dL32419 (5.6%)/25 (7.7%)
Turgeon et al. [24]202473648.9 ± 18.8937/201Acute brain injury≤7 g/dL371<10 g/dL371NA
Carson et al. [25]2020350472.1 ± 11.61137/2367Cardiovascular7–8 g/dL17499–10 g/dL1755173 (9.9%)/146 (8.3%)
Taccone et al. [26]202482052 ± 16436/384Acute brain injury<7 g/dL423<9 g/dL39782 (20.7%)/94 (22.5%).
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Jiménez Franco, D.A.; Pérez Velásquez, C.A.; Rodríguez Lima, D.R. Mortality in Critically Ill Patients with Liberal Versus Restrictive Transfusion Thresholds: A Systematic Review and Meta-Analysis of Randomized Controlled Trials with Trial Sequential Analysis. J. Clin. Med. 2025, 14, 2049. https://doi.org/10.3390/jcm14062049

AMA Style

Jiménez Franco DA, Pérez Velásquez CA, Rodríguez Lima DR. Mortality in Critically Ill Patients with Liberal Versus Restrictive Transfusion Thresholds: A Systematic Review and Meta-Analysis of Randomized Controlled Trials with Trial Sequential Analysis. Journal of Clinical Medicine. 2025; 14(6):2049. https://doi.org/10.3390/jcm14062049

Chicago/Turabian Style

Jiménez Franco, Daniel Arturo, Camilo Andrés Pérez Velásquez, and David Rene Rodríguez Lima. 2025. "Mortality in Critically Ill Patients with Liberal Versus Restrictive Transfusion Thresholds: A Systematic Review and Meta-Analysis of Randomized Controlled Trials with Trial Sequential Analysis" Journal of Clinical Medicine 14, no. 6: 2049. https://doi.org/10.3390/jcm14062049

APA Style

Jiménez Franco, D. A., Pérez Velásquez, C. A., & Rodríguez Lima, D. R. (2025). Mortality in Critically Ill Patients with Liberal Versus Restrictive Transfusion Thresholds: A Systematic Review and Meta-Analysis of Randomized Controlled Trials with Trial Sequential Analysis. Journal of Clinical Medicine, 14(6), 2049. https://doi.org/10.3390/jcm14062049

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop