Preoperative anaemia and adverse CABG outcomes—“there is more to this than meets the eye”

Anaemia is defined as haemoglobin < 13 g/dl in men and < 12 g/dl in women by the World Health Organization. Preoperative anaemia is very common in patients presenting for coronary artery bypass grafting (CABG) occurring in 15 to 40% of patients [1]. Preoperative anaemia may be attributed to multiple reasons. The most frequent causes are chronic blood loss, commonly from the gastrointestinal tract and iron deficiency anaemia. Megaloblastic anaemia from folic acid or B12 deficiency, erythropoietin deficiency as seen in chronic kidney disease, haemolysis, and myelodysplastic syndrome are other causes of anaemia. The presence of chronic diseases, frailty, and advanced age also lead to anaemia [2]. Many reports have suggested that presence of anaemia increases the postoperative mortality of CABG patients [1, 3, 4]. Anaemic patients also have increased length of intensive care unit (ICU) stay and hospital stay and requirement of dialysis. Why should anaemia worsen the outcome of CABG and cardiac surgery? The traditional view was that low haemoglobin reduces oxygen transport and can potentially impair aerobic metabolism. However, increased blood flow due to the decreased blood viscosity, increased oxygen unloading to tissues, increased plasma volume, and redistribution of blood flow negate the harmful effect of low haemoglobin, unless the anaemia is very severe or develops acutely [5]. Anaemic patients are at increased need for perioperative red blood cell (RBC) transfusion [1]. Infective complications, myocardial infarction, stroke, renal failure, prolonged ventilation, atrial fibrillation, and mortality are increased after RBC transfusion [6]. These are often mediated by the proinflammatory effects of RBC transfusion. Stored blood has less deformable RBCs leading to capillary occlusion; the oxygen release from haemoglobin is also impaired in bank blood due to the leftward shift of the oxygen dissociation curve. This results in end-organ ischaemic damage [7]. Hence, it is often difficult to differentiate the independent contribution of anaemia and blood transfusion for adverse outcomes after CABG.

The reports on adverse outcomes of anaemia in CABG patients are largely based on Western population where the majority of cases are performed using cardiopulmonary bypass (CPB) (on-pump CABG). Haematocrit below 22% on CPB is associated with higher incidence of low cardiac output syndrome, sepsis, multiorgan failure, and mortality. This effect is accelerated by the proinflammatory effect of blood transfusion [7, 8]. Hence, their conclusions may not be relevant to Indian patients where the differences in the risk profile and the frequent use of off-pump CABG (OPCAB) technique for coronary revascularization can potentially lead to different outcomes. With this background, Shales et al., in a paper later in this issue, looked into the independent effect of anaemia on postoperative outcomes in Indian patients after OPCAB [9]. They found that anaemic patients had a higher mortality in univariate analysis, but on multivariate analysis it was not an independent factor for mortality after OPCAB. However, it was associated with an increased incidence of renal and respiratory insufficiency and an increased need for perioperative blood transfusion.

The incidence of anaemia in this study was 12.8% which is much lower than quoted in other studies. There are probably three reasons for this—the exclusion of mild anaemia from their analysis (haematocrit between 33 and 37% was excluded), the younger age of the study population (mean age < 60 years), and the low percentage of women in the study cohort (13.6%).

Does anaemia worsen the outcome after CABG? To answer this question, we ideally require randomized controlled trials (RCT), because they effectively negate the effect of known and unknown confounders. However, it is not always possible to do RCTs, especially in surgical cohorts. In retrospective studies, to show that anaemia is an independent risk factor for adverse outcomes, multiple logistic regression (MLR), Cox proportional hazards model, and propensity scoring can be used to adjust the confounders. Propensity scoring is the most validated tool as it leads to the matching of the covariates; however, it may lead to a considerable decline in the sample size. MLR analyses the effect of variables (both categorical and continuous variables) (independent variable) on a single dichotomous outcome (dependent variable) [10]. Since the MLR model controls and analyses the magnitude of each covariate’s influence on the outcome, selecting the appropriate variables for a model is important. Cox proportional hazards modelling is similar to MLR and is used for analysing survival data and provide a hazard ratio to quantify the effect of the variable on survival outcomes. The authors used MLR in this paper and the dependent variables included mortality, renal failure, and risk for blood transfusion. The authors found that anaemia was not a risk factor for mortality, but was a strong predictor for the need for blood transfusions and renal failure. The area under the curve (AUC) for receiver operating characteristics (ROC) was only 0.63 suggesting only moderate discrimination for anaemia with 30-day mortality. This result is in variance to most studies that examined the role of anaemia in CABG outcomes. Most studies show that preoperative anaemia is indeed a risk factor for operative mortality after risk adjustment.

What could be the reasons for this result? One reason could be the use of OPCAB for CABG. The small sample size is another reason. However, the methodology of the study can also be questioned. In this paper, few important variables, which have a direct impact on the postoperative outcomes (chronic kidney disease, dialysis-dependent renal failure, urgency of surgery), were left out from the univariate and multivariate analysis. Since the Society of Thoracic Surgeons (STS) score was low in both groups (anaemic and non-anaemic), one can only assume that very few such patients were present in the study cohort. The other possibility could be that since these variables were already included in the STS score, no independent analysis was carried out. Another important finding in this paper was that decreasing trend of haematocrit was not associated with increased mortality, which should happen if a direct linear relationship exists between the cause and the effect. This also points out that anaemia may not be directly related to the outcome. The reason could be the influence of perioperative blood transfusion. Even though anaemic patients are expected to receive blood transfusion, the high rate of transfusion (84.5%) in non-anaemic patients is a cause for concern and this effect could have influenced the outcomes significantly. Blood transfusion in OPCAB is shown to increase the inflammatory milieu similar to on-pump CABG patients and adversely affect the outcome [11].

How do we separate the effects of anaemia and perioperative blood transfusion on CABG outcomes? One way was to do MLR with blood transfusion as a confounding variable. Another way is by performing a mediation analysis. Mediation analysis aims to see whether the effect of the independent variable (anaemia) on the outcome (mortality) can be mediated by a change in the mediating variable (blood transfusion). In a recent study published in The Annals of Thoracic Surgery, the authors elegantly showed that mortality, ICU stay, and hospital stay were mediated to a large extent by the increased RBC transfusion while renal failure was mostly caused by anaemia alone [1]. Since the authors failed to do such an analysis, the study unfortunately falls short in many regards and remains a hypothesis-generating one without producing firm conclusions. The study cohort also received a low percentage of the left internal mammary artery to the left anterior descending artery (< 90%) which is a cause for concern.

Anaemia seems an easy target to correct and since iron deficiency anaemia is the commonest form, STS practice guideline states that “Assessment of anaemia and determination of its aetiology is appropriate in all patients undergoing cardiac surgery, and it is reasonable to treat with intravenous iron preparations, if time permits (Class IIA, Level B–R)” [12]. However, there are several logistic issues related to it and limited data on its usefulness, and not all anaemia is related to iron deficiency. Anaemia in cardiac surgery is a surrogate of aging, frailty, comorbidities, and chronic diseases. Correcting anaemia remains difficult and is often unsuccessful in these situations. A combined administration of parenteral iron and erythropoiesis-stimulating agents seems the most promising agent to correct preoperative anaemia [2]. An RCT—the Intravenous Iron for Treatment of Anaemia before Cardiac Surgery trial (NCT02632760)—is being conducted comparing efficacy, safety, and cost-effectiveness of preoperative parenteral iron therapy [13] and hopefully will shed some light to the usefulness of anaemia correction before CABG.

Author contribution

PKV has contributed towards conception, writing, and revising the manuscript.

Funding

None.

Data availability

Not applicable.

Declarations

Ethics approval

Not applicable.

Consent

Not applicable.

Competing interests

None.