Critical care trailblazers: transfusion thresholds in critical care – the TRICC trial 


Transfusion of blood was first reported in the early 19th century. In the early days of transfusion history, direct donor-to-recipient transfusion was commonly practiced. By 1914, the addition of citrate was found to prevent clotting, thus opening the door to the storage of collected blood. As World War I broke out in Europe, there was an increasing transfusion requirements for soldiers wounded in battle. Similar to other medical innovations that took root on the battlefield, type O blood mixed with citrate-glucose solution was stored for emergency use in army units. With the entry of the United States into the war in 1917, physicians became more familiar with transfusion practices. During the war, blood was collected from healthy soldiers, citrated and refrigerated, followed by transportation to the battlefront in sterilized milk bottles (1).

It took several years before Bernard Fantus established the first blood bank at the Cook County Hospital in Chicago in March 1937 (Fig. 1). In the following year, the number of transfusions saw a phenomenal rise, with a considerable decrease in the incidence of adverse reactions from 33% to 8% (1).  

Figure 1. Bernard Fantus who established the first blood bank at the Cook County Hospital (in the background) in March 1937

Early transfusion practices

Red cell transfusions were often administered based on clinician judgment, leading to wide variability in transfusion thresholds. In their treatise of 1942, titled “Anesthesia in Cases of Poor Surgical Risk: Some Suggestions for Decreasing the Risk”, Adam and Lundy (2) proposed that “When the concentration of hemoglobin is less than 8 to 10 grams per 100 cubic centimeters of whole blood, it is wise to give a blood transfusion before operation.” This suggestion appears to have been rapidly indoctrinated into anesthetic practice in that era. The widespread adoption of the 10/30 rule followed, with transfusion of red cells below a threshold hemoglobin level of 10 g/dl or a hematocrit of less than 30%. This approach was based on the physiological rationale that the hemoglobin level is the key determinant of oxygen delivery. However it failed to take into account compensatory mechanisms that are activated, including a rise in the cardiac output and increase in the oxygen extraction at the tissue level.   

Rising concerns with transfusion-transmitted infection

Transfusion practices became the subject of intense scrutiny from the 1980s with increasing concern regarding the transmission of infection, particularly with the human immunodeficiency virus (HIV) (3). It was soon realized that blood transfusions, once considered relatively safe, carry risks and may be potentially dangerous. In the meantime, the National Institutes of Health Consensus Conference in the United States of 1988 reported that there was no evidence to support a rigid threshold for transfusion of red cells (4). The term “transfusion trigger” was coined by Friedman et al. in 1981 comparing transfusion practices among 535,031 male and female surgical patients (5).  

Background to the Transfusion Requirements in Critical Care (TRICC) trial 

Besides the transmission of infection, the potential immunosuppressive (6) and microcirculatory complications of blood transfusion (7) were realized by the early 1990s. However, there was no credible evidence to support transfusion thresholds in critically ill patients, balancing the risks compared to possible benefits. Against this background, the Canadian Critical Care Trials Group planned a randomized controlled trial (RCT) to identify optimal transfusion thresholds among critically ill patients. They aimed to compare a liberal with a more restrictive transfusion strategy among normovolemic critically ill patients with anemia (8). 

The study population and design 

The TRICC trial enrolled patients from 22 tertiary and 3 community-level ICUs across Canada during a 3-year period between 1994-1997. The study patients had a hemoglobin level of 9.0 g/dl or less within 72 hours of ICU admission, and were expected to stay in the ICU for more than 24 hours. Patients with ongoing bleeding, chronic anemia, following elective cardiac surgery, and those in whom death seemed imminent were excluded. Patients were randomly assigned to a liberal or restrictive transfusion strategy, stratified by center and disease severity. Stratification by disease severity was based on an APACHE II score of ≤15 or >15.

The restrictive vs. liberal strategies

The hemoglobin level in the restrictive group was maintained between 7.0–9.0 g/dl; red cell transfusion was carried out if the hemoglobin level dropped below 7.0 g/dl. In the liberal group, the aim was to maintain a hemoglobin level of 10–12 g/dl, with transfusion of red cells if the level dropped below 10 g/dl. One unit of blood, between 240 to 340 ml, was transfused at a time, with hemoglobin levels measured after each unit. The overall care was left to clinician judgment. 

From a total of 6451 patients were screened, 418 patients were randomized to receive a restrictive transfusion strategy, and 420 to a liberal strategy. Baseline characteristics were well matched between the two groups. Respiratory and cardiac disease were the most common underlying cause for ICU admission. More than 80% of patients received mechanical ventilation. Vasoactive drugs were administered in 37% and the baseline hemoglobin level was 8.2±0.7 g/dl in both groups. 

Hemoglobin levels and the number of units transfused

The mean daily hemoglobin levels through the study period were significantly different between the two groups (restrictive vs. liberal: 8.5±0.7 vs. 10.7±0.7 g/dl). The liberal strategy led to significantly more units of red cell transfusion compared with the restrictive strategy (5.6±5.3 vs. 2.6±4.1units), equating to a 54% reduction in transfusions with the restrictive strategy. 


The all-cause mortality at 30-days post ICU admission, the primary outcome, was similar among both groups of patients. The mortality was 18.7% in the restrictive group compared with 23.3% in the liberal group (p=0.11). The hospital mortality was significantly lower in the restrictive group (22.2 percent vs. 28.1 percent, P=0.05). The 60-day and ICU mortality were comparable between the two groups. The number of patients with multiorgan failure was also similar between the two groups. 

Subgroup analysis

The authors analyzed subgroups of patients who were less than 55 years old and APACHE II score ≤20. In the both these subgroups, the mortality at 30 days post ICU admission was significantly lower with a restrictive compared with a liberal strategy on Kaplan-Meier analysis (p=0.02 for both subgroups). No difference in the primary outcome was observed between the two groups among patients with trauma, primary or secondary cardiac disease, severe infections, and septic shock. 

Adverse events

Cardiac adverse events, mainly acute pulmonary edema and acute myocardial infarction, were more common in the liberal group. 


The authors concluded that among normovolemic critically ill patients, a restrictive strategy with a transfusion threshold of 7.0 g/dl is at least as effective as a liberal strategy using a transfusion threshold of 9.0 gm/dl. A target hemoglobin level of 7–9 g/dl was comparable or even superior to a higher target level of 10.0–12.0 g/dl. In contrast to previous studies, a low transfusion trigger of 7 g/dl did not lead to adverse outcomes in patients with cardiac disease. The authors suggested that a transfusion threshold of 7.0 g/dl may be generalizable to most critically ill patients, except patients with acute coronary syndrome.  


The TRICC trial challenged the use of tradition-borne, arbitrary transfusion thresholds and provided strong evidence that injudicious administration of red cells might lead to adverse outcomes. It questioned the purported benefits of targeting oxygen delivery by raising hemoglobin levels in critically ill patients. The trial addressed appropriate clinical outcomes, with analysis by intention to treat. The mean APACHE II score was 21, with 80% patients on mechanical ventilation, and nearly 37% on vasoactive medication, all suggesting a fairly high severity of illness at baseline. Thus, the study is applicable to most real world ICUs. There was a substantial difference in the number of units transfused between the restrictive and liberal groups, with 33% patients in the restrictive group not receiving any transfusion.   


The study excluded a large number of patients. Nearly 87% of patients who were initially screened were excluded for various reasons; of the 3206 eligible subjects, only 838 were randomized (26.1%). A substantial number of patients were excluded either because the clinician did not agree to participate (598 patients) or the family refused consent (603 patients). Randomization was stratified according to an APACHE II score of  ≤15 or >15; however, inexplicably, subgroup analysis was based on an APACHE II score of ≤20 compared with >20. The study included only 838 patients against the calculated sample size of 1620 patients. Patients with traumatic brain injury were considered to be a population vulnerable to the harmful effects of anemia; perhaps a subgroup analysis may have been relevant. The study excluded elective cardiac surgical patients, leading to many cardiac surgical units holding on tenaciously to the time-honored hemoglobin threshold of 10 g/dl. The study was performed during a period when leucodepletion was not in common practice. The question remained whether transfusion of leucodepleted blood would make a difference in clinical outcomes.  

The aftermath and relevance in today’s world 

Historically, transfusion practices were largely based on tradition. Critically ill patients were considered to be especially prone to adverse outcomes related to anemia. A hemoglobin level of 10, corresponding to a hematocrit of 30% was considered to optimize oxygen delivery in critical illness. This dogmatic approach may have led to overzealous transfusion practices and resulted in innumerable adverse clinical outcomes. 

The TRICC trial triggered a wind of change among clinicians, supporting lower hemoglobin levels under most clinical circumstances. The findings of this landmark study were bolstered by many clinical trials that were published later. The TRISS trial, using leucodepleted blood, compared transfusion thresholds of 7.0 vs. 9.0 g/dl in patients with septic shock and found no difference in the 90-day mortality (9). The TRACS trial evaluated patients who underwent cardiac surgery under cardiopulmonary bypass. A target hematocrit of ≥24% was compared with ≥30% in the liberal group (10). The composite primary outcome, including the 30-day all-cause mortality, cardiogenic shock, acute respiratory distress syndrome, and acute kidney injury requiring renal replacement therapy, was similar between the two groups. The TRICS III RCT further strengthened the evidence for a restrictive strategy among cardiac surgical patients using a transfusion trigger of 7.5 g/dl (11). 

The TRICC trial was carried out in an era when clinicians were sharply divided on red cell transfusion thresholds in critically ill patients. Although the study was planned to recruit 2,300 patients over 5 years, with just a year remaining, physicians and patients were increasingly reluctant to participate. The study was closed after randomizing 838 patients, well short of the calculated sample size. Paul Hébert and his co-investigators were surprised when their analysis revealed that a restrictive transfusion strategy seemed to be no worse, and even superior compared with a liberal strategy (12). Subsequent studies over the years largely corroborated their findings. This pioneering trial of more than two decades ago, comparing transfusion thresholds, remains a landmark in the field of critical care medicine.    


1. Meyer KA. The history of the Cook County Hospital blood bank. Q Bull Northwest Univ Evanst Ill Med Sch. 1949;23(3):318–20. 

2. ADAMS RC, LUNDY JS. Anesthesia in Cases of Poor Surgical Risk: Some Suggestions for Decreasing the Risk. Anesthesiology. 1942 Sep 1;3(5):603–7. 

3. Welch HG. Prudent Strategies for Elective Red Blood Cell Transfusion. Ann Intern Med. 1992 Mar 1;116(5):393. 

4. Perioperative Red Blood Cell Transfusion. JAMA. 1988 Nov 11;260(18):2700–3. 

5. Friedman BA, Burns TL, Schork MA. An analysis of blood transfusion of surgical patients by sex: a question for the transfusion trigger. Transfusion (Paris). 1980;20(2):179–88. 

6. Bordin JO, Heddle NM, Blajchman MA. Biologic effects of leukocytes present in transfused cellular blood products. Blood. 1994 Sep 15;84(6):1703–21. 

7. Baker CH, Wilmoth FR, Sutton ET. Reduced RBC versus plasma microvascular flow due to endotoxin. Circ Shock. 1986;20(2):127–39. 

8. Hébert PC, Martin C, Yetisir E. A Multicenter, Randomized, Controlled Clinical Trial of Transfusion Requirements in Critical Care. N Engl J Med. 1999; 

9. Holst LB, Haase N, Wetterslev J, Wernerman J, Guttormsen AB, Karlsson S, et al. Lower versus higher hemoglobin threshold for transfusion in septic shock. N Engl J Med. 2014 Oct 9;371(15):1381–91. 

10. Hajjar LA, Vincent JL, Galas FRBG, Nakamura RE, Silva CMP, Santos MH, et al. Transfusion Requirements After Cardiac Surgery: The TRACS Randomized Controlled Trial. JAMA. 2010 Oct 13;304(14):1559. 

11. Mazer CD, Whitlock RP, Fergusson DA, Hall J, Belley-Cote E, Connolly K, et al. Restrictive or Liberal Red-Cell Transfusion for Cardiac Surgery. N Engl J Med. 2017 Nov 30;377(22):2133–44. 

12. Transfusion Trigger [Internet]. 50 Landmark Papers every Intensivist Should Know. CRC Press; 2021 [cited 2023 Jan 3]. Available from:

1 thought on “Critical care trailblazers: transfusion thresholds in critical care – the TRICC trial ”

  1. very appropriately analysed Jose !! I think very few non critical care physicians now object to this practice of 7.0 Gms as the cut off. Thank you for this analysis.

Leave a Reply