Restrictive RBC transfusion strategy non-inferior to liberal approach in cardiac surgery

TRICS III – An International Multicenter Randomized Trial of Transfusion Triggers in Cardiac Surgery

News - Nov. 13, 2017


Anemia in the context of cardiac surgery is independently associated with adverse outcome. On the other hand, red blood cell (RBC) transfusion is associated with increased mortality. Few RCTs have been performed in this area, but based on these there has been a state of clinical uncertainty on what is the best strategy.

The TRICS III study addressed the question whether restrictive transfusion is as safe as (or as effective as) liberal transfusion in moderate and high risk cardiac surgical patients as defined by EuroSCORE I. The hypothesis tested was that a lower hemoglobin concentration for RBC transfusion (restrictive transfusion strategy) will be non-inferior to a liberal strategy in terms of mortality and vital organ function (heart, brain, kidney). To this end, patients of at least 18 years old, scheduled for cardiac surgery with CPB and a pre-operative additive EuroSCORE I ≥6 were randomized to either of the strategies. For the restrictive strategy, a trigger of<7.5 g/dL in the operating room, ICU and ward was used, while for the liberal strategy, a trigger of <9.5 g/dL was used in the operating room and ICU, and <8.5 g/dL in the ward. The primary outcome was a composite of all-cause mortality, myocardial infarction, new renal failure with dialysis and new focal neurological deficit. The primary analysis was a per protocol comparison, with 3% non-inferiority margin. During 2014-2017, 5243 patients were randomized in 74 sites in 19 countries, of whom 4860 were included in the per-protocol analysis (adherence >90%).

As can be expected in a large randomized study, baseline characteristics were very comparable, and they were representative of what may be anticipated in patients having cardiac surgery; they were at moderate to high risk.

Main results

  • The composite primary endpoint was seen in 276 out of 2430 patients (11.4%) in the restrictive threshold group, and in 303/2430 (12.5%) in the liberal strategy group.
  • The estimated risk difference between groups was: -1.11% (-2.93% to 0.72, P=0.0001).
  • The odd ratio of developing the primary outcome was 0.90 (95%CI: 0.76-1.07).
  • No differences were seen in the occurrence of any of the components of the primary outcome.
  • The peri-operative hemoglobin concentration was similar in both groups pre-operatively and intra-operatively, and separated from ICU admission onwards, with mean Hb concentration being higher in the liberal threshold group (at about 10 g/dL), as compared with about 9 g/dL in the restrictive threshold group.
  • In the restrictive group, 1271 persons (52.3%) had at least 1 unit transfused, as compared with 1765 persons (72.3%)(OR: 0.41, 95%CI: 0.37-0.47).
  • The number of transfusions was lower in the restrictive threshold group, at a median of 2 (IQR: 1-4), as compared with 3 (IQR: 2-5) in the liberal threshold group (OR: 0.85, 0.82-0.88).
  • Among multiple subgroup analyses, the only significant observation was that patients ≥75 years old showed a better outcome with the restrictive approach.
  • A variety of sensitivity analyses of other factors that might influence these findings revealed no such factors.


This study showed that a restrictive RBC transfusion strategy with 7.5 g/dL as trigger, reduces transfusion of allogenic RBSc and is non-inferior to a liberal strategy with 9.5 g/dL for mortality and major morbidity including myocardial infarction, stroke or new onset of renal failure with dialysis, in moderate to high-risk patients undergoing cardiac surgery.

During the press conference, dr. Frank W. Sellke (Brown Medical School and Rhode Island Hospital, Providence, RI) repeated that anemia during cardiac surgery is associated with increased mortality and rate of complications and that blood transfusions have been found to result in diminished short and long term survival after surgery, although the results of past studies have been inconsistent. He noted that projections suggest that there will be a lack of an adequate supply of blood in the future. He remarked that while the overall results demonstrated no significant difference in outcome between groups, a numerical benefit was seen in the restrictive group for the composite outcome and all components of the primary outcome except MI. That was not too unexpected if you look back to previous trials, but it was unknown thus far.

Sellke thought that the overall rate of transfusion seemed high, even in the restrictive group (>50%). Although he was very positive about the trial, Sellke had a few questions: he would be curious to know the 24 hour blood loss and what was the take back rate for bleeding in this study. Moreover, there was a 1.5 g/dL difference in Hb level between groups, but the final Hb was 9 g/dL in the restrictive group. Therefore, he wondered what the adherence to the protocol was, and could a higher threshold for transfusion be applied (6.5 g/dL)?

Moreover, he was curious about possible explanation of the counterintuitive finding that patients over 75 did better with a restrictive strategy. Lastly, these are very short-term results, and it would be equally important to know the long-term effects. During the Q&A, Mazer added that he was surprised as well about this age-effect. With several subanalyses they confirmed that indeed it was a differential effect. He thinks there might be several reasons, for instance preselection by surgeons to pick the ‘good elderly’. Alternatively, elderly patients may be more sensitive to the effects on inflammation or volume of transfusion. It is hypothesis-generating that this study did not find that a higher threshold should be applied to elderly patients, and the study clearly shows that a restrictive approach may be safe in these patients.


- Our reporting is based on the information provided at the AHA 2017 congress -

This study was published simultaneously at NEJM Watch webcast with dr. David Mazer on this subject

Facebook Comments


We’re glad to see you’re enjoying PACE-CME…
but how about a more personalized experience?

Register for free