The rationale for perioperative red blood cell (RBC) transfusion is based on the observations that anemia is an independent risk factor for morbidity and mortality after cardiac surgery and that red blood cell transfusion would benefit a subset of patients presenting tissue hypoperfusion [1, 2]. However, transfusions have been associated with high rates of morbidity and mortality in critically ill patients, and there is increasing evidence for independent relationships between RBC transfusion and infectious complications, cardiac and respiratory morbidity, prolonged length of stay (LOS) and mortality after cardiac surgery [2–4].

The Transfusion Requirements After Cardiac Surgery (TRACS) study was a prospective, randomized, controlled trial that recently demonstrated the safety of a restrictive strategy of transfusion compared with a liberal strategy in patients undergoing elective cardiac surgery [2]. In this study, independent of the adopted transfusion strategy, the number of transfused RBC units was an independent risk factor for clinical complications and death at 30 days [2].

The primary objective of this study was to assess the relationship between RBC transfusion and hospital length of stay in a large, single reference center of cardiac surgery. Secondary objectives were to compare the characteristics of patients who received RBC transfusion with those who did not, to evaluate the relationship of the number of transfused RBC units with mortality and clinical complications, and to identify the predictive factors for a prolonged hospital LOS. We hypothesized that patients requiring early postoperative blood transfusion would have longer hospital length of stay.

Intraoperative characteristics for the RBC transfusion groups are compared in Table 1. As expected, a higher proportion of patients from the liberal strategy group received blood transfusion than did those from the restrictive strategy group (77% vs. 43%, p < 0.001). Most patients underwent coronary bypass artery grafting (CABG). Patients who were exposed to blood transfusion had lower initial hematocrit and hemoglobin levels compared to patients who did not receive RBC transfusions.

From 502 patients, 236 (47%) patients received transfusion at the operative room, 111 (22%) received at the first postoperative day (1PO), 74 (15%) at the second postoperative day (2PO) and 59 (12%) at the third postoperative day (3PO) (Additional file 2).

Analyzing postoperative clinical complications, patients who received RBC transfusion were more likely to present major events related to the composite endpoint (30-day all-cause mortality and cardiogenic shock, acute respiratory distress syndrome, or acute kidney injury requiring dialysis or hemofiltration, p < 0.001), in addition to cardiac (p < 0.001), respiratory (p < 0.001), neurologic (p = 0.014), infectious (p < 0.001), and inflammatory complications (p < 0.001) (Figure 1). We also found that patients from the high transfusion group presented a higher incidence of postoperative complications.

Length of stay analysis

The median hospital length of stay was 9 days (range, 1 to 67 days). Patients exposed to any RBC transfusion had longer median length of stay than patients in the non-transfusion group: 15 days [95% CI, 12–18] in high transfusion requirement group (> 3 units) vs. 10 days [95% CI, 9–11] in low transfusion requirement group (1–3 units) vs. 9 days [95% CI, 8–10] in non-transfusion group, p < 0. 001) (Figure 2).

Many covariates had statistically significant associations with hospital LOS. These include: female gender, older age, higher EuroSCORE, redo surgery, renal disease, lower initial hemoglobin and hematocrit, lower left ventricular ejection fraction, renal disease, valvular surgery or a combined procedure, longer cardiopulmonary bypass, lower initial hemoglobin and hematocrit, and exposure to RBC transfusion (Additional file 3 and Additional file 4).

All variables significantly associated with increased hospital length of stay were analyzed in a multivariate Cox proportional hazards model. In the multivariate model, the following factors were predictive of a prolonged hospital length of stay: age higher than 65 years (hazard ratio [HR], 1.36 [95% CI, 1.09-1.7]; p = 0.006), EuroSCORE higher than 3–5 (HR 1.46 [1.13-1.88]; p = 0.003, EuroSCORE higher than 5 (HR, 1.71 [95% CI, 1.27–2.3]; p < 0.001), valvular surgery (HR, 1.60 [95% CI, 1.28–1.99]; p < 0.001), combined procedure (HR, 1.73 [95% CI, 1.12–2.67]; p = 0.013), LVEF lower than 40% (HR, 1.63 [95% CI, 1.19-2.24]; p = 0.002), LVEF 40–59% (HR, 1.32 [95% CI, 1.07–1.64]; p = 0.001), and RBC transfusion of > 3 units (HR, 2 [95% CI, 1.44–2.79]; p < 0.001) (Table 2).

Variable	Hazard ratio	p
	(95% CI)
Age (years)
<65	Reference
≥65	1.36 (1.09 – 1.7)	0.006
EuroSCORE
<3	Reference
3–5	1.46 (1.13 – 1.88)	0.003
>5	1.71 (1.27 – 2.3)	<0.001
Procedure
CABG	Reference
Valve	1.6 (1.28 – 1.99)	<0.001
CABG + valve	1.73 (1.12 – 2.67)	0.013
Left ventricular ejection fraction (%)
<40	1.63 (1.19 – 2.24)	0.002
40–59	1.32 (1.07 – 1.64)	0.010
≥60	Reference
Units of RBC transfused
None	Reference
Low-transfusion (1–3 units)	1.16 (0.94 – 1.42)	0.164
High-transfusion (>3 units)	2.00 (1.44 – 2.79)	<0.001

Abbreviations: CI, confidence interval; CABG, coronary artery bypass graft surgery; min, minutes; RBC; red blood cells.

In an adjusted model for age, EuroSCORE, type of surgical procedure and left ventricular ejection fraction, the exposure to RBC transfusion was associated with a prolonged hospital length of stay (Figure 3).

In this study of 502 patients undergoing elective cardiac surgery, we found that 60% received RBC within the first 72 postoperative hours. Patients requiring RBC transfusion were older, most of them were female, underwent redo and complex surgeries, had higher EuroSCORE, presented with lower levels of hemoglobin and hematocrit and had a higher incidence of renal disease. In a multivariable model adjusted for univariate associations, RBC transfusion was a predictor of longer hospital length of stay, and the strongest association was found in the high transfusion group. Independent of the strategy of transfusion, patients who received RBC transfusion in the first 72 postoperative hours had a higher incidence of complications, including renal failure, cardiogenic shock, acute respiratory distress syndrome, infectious, neurologic and inflammatory complications and longer hospital length of stay.

Significant variability in transfusion practice exists in cardiac surgery. A recent analysis [5] of blood practices across 798 US hospitals drew attention to this problem. In hospitals performing at least 100 on-pump CABG operations (82 446 cases), rates of blood transfusion ranged from 7.8% to 92.8% for RBCs, 0% to 97.5% for fresh-frozen plasma, and 0.4% to 90.4% for platelets. Multivariable analysis revealed that after adjustment for patient-level risk factors, hospital transfusion rates varied by geographic location (P = .007), academic status (P = .03), and hospital volume (P < .001).

We found patients who were female, older, underwent redo or complex surgeries, as well as those who had higher EuroSCORE, renal disease, or previous anemia, were more likely to receive RBC transfusion. This finding emphasizes the need to maximize efforts to improve perioperative care in these subgroups of patients in order to avert RBC transfusion adverse events, including increased hospital length of stay.

Rates of transfusion persist high after cardiac surgery despite studies showing that RBC transfusion does not result in better outcomes and even increases rates of complications after cardiac surgery [6]. Guidelines from the Society of Thoracic Surgeons and Society of Cardiovascular Anesthesiologists emphasize the lack of evidence on transfusion triggers after cardiac surgery [7]. Most transfusion indications occur in the first 72 hours after surgery, starting in the operating room, where usually the transfusion indication is due to hemodilution and based on triggers [2].

The rationale for implementing a restrictive transfusion strategy is based on analysis of studies reporting a lack of benefit and, at the same time, substantially increased costs and adverse effects associated with RBC transfusion. These adverse effects include acute hemolytic and nonhemolytic reactions, transmission of viral and bacterial diseases, transfusion-related acute lung injury, and transfusion-associated circulatory overload [8]. Immunosuppression has also been associated with transfusion and may explain the higher risk of infection and recurrence of neoplastic diseases observed in transfused patients [9].

In a retrospective analysis of 11, 963 patients who underwent isolated CABG surgery, Koch et al [10] showed that perioperative RBC transfusion was associated with a dose-dependent increased risk of postoperative cardiac complications, serious infection, renal failure, neurologic complications, overall morbidity, prolonged ventilator support, and in-hospital mortality. In a similar retrospective study, Murphy et al [11] showed that RBC transfusion was strongly associated with infection and with postoperative ischemic morbidity, hospital stay, increased early and late mortality, and hospital costs.

Recently, the TRACS study prospectively demonstrated the safety of a restrictive strategy of RBC transfusion in patients undergoing cardiac surgery [2]. Also, this study reported that the higher the number of transfused RBC, the higher was the number of clinical complications [2].

In our study, we found that patients exposed to RBC transfusion had more complications when compared to patients not exposed, a similar result described in previous reports [1–4, 6, 10]. The evaluation of mechanisms related to RBC transfusion that results in morbidity were not included in this study. However, our hypothesis it that these potential effects may underlie the independent association between blood transfusion and prolonged LOS. On univariate comparison, total length of stay was approximately 6 days longer in the population of patients who received more than 3 RBC units of transfusion and 1 day longer in patients receiving until 3 units as compared to non-exposed group. Multivariate analysis demonstrated a significant relationship between early postoperative transfusion (more than 3 units of RBC) and longer hospital stay.

De Cocker et al [12]. performed a retrospective analysis of 1566 patients undergoing cardiac surgery and identified age greater than 75 years, female gender, NYHA functional class greater than II, arrhythmias, mitral regurgitation, need for inotropic support or intra-aortic balloon pump, non-elective procedures and aortic surgery as predictive factors for a prolonged ICU stay [12]. However, in this study, blood transfusion was not evaluated as a potential predictor of increased hospital LOS [12].

Prolonged hospital LOS is being studied because it results in increased costs, and clinical complications as exposure to infectious agents. It constitutes a possible method to estimate complications after cardiac surgery [13]. Several models have been developed to predict prolonged stay after cardiac surgery. In our practice, the EuroSCORE is widely applied to predict both mortality and prolonged hospital stay [14, 15]. Our findings incorporate RBC transfusion higher than 3 units after cardiac surgery as a strong predictor of prolonged LOS. Also, age higher than 65 years, valvular surgery or a combined procedure, left ventricular ejection fraction lower than 60% were also predictors of increased hospital length of stay, as have been previously reported [15].

Limitations of our study include that it was performed in a single referral center for cardiac surgery, which could compromise the generalizability of our findings. Also, the associations we found between RBC transfusion and longer LOS do not explain causality due to the fact of RBC transfusion may only be a surrogate marker of perioperative complications or higher morbidity. In fact, older and sicker patients with more debilitating conditions were stratified into the transfusion group and it may lead to a misinterpretation. So, it is possible that other variables with impact LOS and not included in the results, may not have been present in the multivariable model. Moreover, the present study is a retrospective analysis from a previous randomized clinical trial with a dictated trigger. Retrospective studies are best done in a large cohort of consecutive unselected patients which could limit our interpretation.

The increased length of hospital stay related to the number of transfused RBC units supports a restrictive therapy in cardiac surgery. In addition, clinicians caring for patients after cardiac surgery should administer only 1 RBC unit at a time, because this may result in less exposure to risks but similar benefits.

Importantly, although the interpretation of our data must take account the restrictions of the substudy analysis, our results brings additional information regarding an independent association of early RBC transfusion with longer hospital stay.