Hemorrhagic complications represent an important adverse prognostic factor in patients with acute coronary syndromes (ACS).1–3 Previous studies indicate that patients with major bleeding (MB) in this setting have up to a 20% of risk of death, myocardial infarction, or stroke during the first 30 days compared with 5% in those who do not develop MB during the first 30 days.1 This risk extends beyond the time of procedures, and even when the severity of the bleeding does not indicate that it is life-threatening.1 Mechanistically, the risk associated with MB is thought to be multifactorial, resulting from the interruption of effective antithrombotic drugs, the reduction of oxygen delivery to the myocardium as a result of hypoperfusion, platelet activation, and the potentially adverse effects of transfusion.4

Given the deleterious association between MB and outcome in ACS, the identification of risk factors able to predict hemorrhage is important and has been explored using clinical risk models. It is logical to explore laboratory factors as well as to provide information to complement bleeding risk scales. Red cell distribution width (RDW) is a quantitative measure of the variability in size of the circulating erythrocytes,5 and has been recently found to be strongly predictive of cardiovascular outcomes in multiple patient populations, including those with ACS.6–12

Most recently, increased RDW has also been linked to a higher rate of bleeding in patients with ACS.13,14 However, these studies assessed only the role of RDW for the prediction of in-hospital bleeding, and did not report the usefulness of RDW for the prediction of bleeding during follow-up.

Therefore, the objective of the present study was to assess the role of RDW values to predict risk of MB over the long-term follow up of patients with non–ST-segment elevation ACS (NSTEACS), and to evaluate whether RDW adds additional predictive value to a widely accepted model for predicting MB risk, the CRUSADE scale.15

The study population consisted of 293 patients with NSTEACS (Table 1). The median RDW value was 13.9% [13.2-14.9%]. Patients with higher RDW values were elderly, more frequently female, and had a greater prevalence of comorbidities including hypertension, diabetes mellitus, hyperlipemia, anemia, previous stroke, and renal dysfunction. These patients also had worse CRUSADE risk scale and were less commonly revascularized. As expected, patients with higher RDW had lower hemoglobin and mean corpuscular volume, while they had higher C-reactive protein levels. At discharge, patients with higher RDW values were less commonly on clopidogrel (Table 1 of supplementary Material).

During follow up, a total of 30 patients (2.3 per 104 patient year) had MB; 16 (53.3%) had type 3a, 8 (26.7%) type 3b, 2 (6.7%) type 3c, 2 (6.7%) type 4, and 2 (6.7%) type 5b. The location of MB was gastrointestinal in 12 (40%), intracranial in 4 (13.3%), cardiac tamponade in 4 (13.3%), pulmonary in 2 (6.7%), musculoskeletal in 2 (6.7%), urologic in 2 (6.7%), coronary artery bypass graft-related bleeding in 2 (6.7%), and puncture site in 2 (6.7%). Eight patients (27%) had MB during the index hospitalization, 14 patients (47%) had MB during the first 30 days, 22 (73%) during the first year, and 25 (83%) through the second year.

The distribution of characteristics and laboratory parameters according to the occurrence of MB is shown in Table 1. Notably, patients who experienced MB had higher RDW values (14.6% [13.4-16] vs 13.8% [13.1-14.7]; P=0.03) (Figure 1). Quartile analyses of RDW values examined as a function of MB rates revealed that there was an abrupt increase in MB at the fourth RDW quartile (RDW ≤ 13.2%=5.1 per 104 patient year, RDW from 13.3% to 13.9%, 5.8 per 104 patient year; RDW from 14.0% to 14.9%, 8.7 per 104 patient year, RDW>14.9%, 2.1 per 104 patient year) (Figure 2). In addition, Kaplan-Meier survival analysis showed that, compared with patients with RDW ≤14.9%, those with RDW values above the fourth quartile (RDW>14.9%) had a higher rate of MB (19.4% vs 7.5%, log rank test, P<.001). On multivariate Cox regression analysis (Table 2), RDW>14.9% (hazard ratio=2.41; 95%CI: 1.15–5.02; P=.02) was associated with a more than 2-fold increase in the risk for MB. In reclassification analysis, RDW>14.9% added significant information to the CRUSADE risk scale. The relative integrated discrimination improvement from the addition of RDW>14.9% was 5.2% (P=.01) whereas the net reclassification improvement was 10% (95%CI, 6-19%; P=.02). The probability of correctly predicting MB and non–MB events when RDW>14.9% was added to the CRUSADE risk scale was reflected in the percentage of both non–MB and MB events correctly reclassified (7% and 3%, respectively). Moreover, the addition of RDW>14.9% was associated with a nonsignificant improvement in the C-index from 0.63 (95%CI, 0.57-0.69) to 0.68 (95%CI, 0.62.-0.73); P=.15. Illustrating this finding, as detailed in Figure 3, patients with a CRUSADE risk scale <40 points (nonhigh risk for bleeding) and RDW ≤ 14.9% had the lowest MB rate (6.4%), while those with a CRUSADE risk scale >40 points (high risk for bleeding) and RDW >14.9% had the highest rate (26.7%; P<.001).

Figure 1.

Box plots of red cell distribution width values in patients with and without major bleeding. The bottom and top whiskers indicate the 5th and 95th percentile values, the lower and upper boundaries of the boxes represent 25th and 75th percentile values, and the horizontal line within the box, the median value.

(0.04MB).

Figure 2.

Major bleeding complications according to red cell distribution width quartiles.

(0.07MB).

Figure 3.

Addition of red cell distribution width values to the CRUSADE risk score for risk stratification of major bleeding. Rates of major bleeding were significantly higher with rising combined scores. Red cell distribution width (−), ≤ 14.9%: red cell distribution width (+), >14.9%; CRUSADE (+), >40 points (high risk for bleeding); CRUSADE (–), <40 points (nonhigh risk for bleeding). Hazard ratio per category and adjusted by age (years) and left ventricular ejection fraction>50%. 95%CI, 95% confidence interval; HR, hazard ratio; RDW, red cell distribution width.

(0.08MB).

In the present study, we examined the role of RDW in the prediction of MB complications in a cohort of patients with NSTEACS. In this population, we demonstrated that most MB episodes occurred well after the index hospitalization. Additionally, patients with elevated RDW (> 14.9%) at the time of hospitalization were at higher risk for MB, with a more >2-fold increase in risk compared with those with RDW ≤ 14.9%. Moreover, we also showed that RDW adds useful information to the CRUSADE risk scale for predicting this important outcome. Lastly, the risk associated with an elevated RDW was presented early, and lasted throughout the follow up.

The RDW is routinely assessed as part of the complete blood count to gather information on the heterogeneity in the size of circulating erythrocytes. Computationally, RDW is the coefficient of variation of the mean corpuscular volume and therefore higher RDW values reflect greater heterogeneity in mean corpuscular volume (anisocytosis), which is usually caused by perturbation in erythrocyte maturation or degradation.5 The RDW is used as an auxiliary index to help in the diagnosis of different types of anemia, but has also been evaluated as a potential screening marker for colon cancer and celiac disease.21 Furthermore, increased RDW has also been related to nutritional deficiencies (eg, iron, folate, or vitamine B12)21,22 and the release of cytokines in response to inflammation.23 Thus, oxidative stress directly damages erythrocytes and leads to shortened erythrocyte survival; likewise, theses cytokines attenuate the activity of erythropoietin and cause production of ineffective red blood cells, leading to elevated RDW.24–26

As mentioned earlier, several studies have suggested that increased RDW is predictive of higher risk for adverse events in patients with a wide array of different cardiovascular disorders.6–12 However, there are few data about the relationship between RDW and bleeding complications in patients with ACS. In addition, the mechanism through which RDW is associated with bleeding remains largely unknown, but seems to be multifactorial. Given that elevated RDW is often seen in patients with extensive comorbidities,27 RDW may predict the occurrence of these diseases as well as that of other age-associated conditions. Accordingly, in the present study, patients with elevated RDW had an unfavorable baseline clinical profile that included older age, more frequent female sex, and a higher prevalence of comorbidities known to increase the risk of bleeding, such as hypertension, diabetes mellitus, anemia, previous stroke, and renal dysfunction. Moreover, these patients also had higher C-reactive protein levels, and consequently inflammation may play a role in bleeding risk. In fact, increased inflammation states have been linked to changes in hemostasis with a prothrombotic state predisposing to thromboembolism and coagulopathy with a tendency for bleeding.28,29 Additionally, other indirect mechanism such as erythropoietin deficiency may contribute to the increased risk of MB.27 Further research is needed to understand the pathways through which RDW is associated with bleeding.

In a recent study that included 513 patients with NSTEACS, Gonçalves et al13 showed that those with a RDW>15.7% had a 3-fold increased in-hospital MB compared with those with a RDW<15.7%. Similarly, Fatemi et al,14 also showed a 2.3-fold increased in-hospital MB risk by comparing the highest RDW quartile (> 14.3%) to the lowest quartile (< 12.7%) in a large population of 6689 patients who underwent percutaneous coronary intervention. Findings from our study not only confirm these previous results but also expand on them, given that we also show that RDW is a strong predictor for MB beyond the hospitalization period. Moreover, unlike these previous studies, we used a standardized MB definition based on the Bleeding Academic Research Consortium Definition criteria,19 which may be helpful for future comparisons. We also demonstrated that RDW measurement adds valuable clinical information about the potential increased likelihood of MB to the CRUSADE bleeding risk scale in patients with NSTEACS.

Interestingly, in this study enhanced predictions of RDW over the CRUSADE risk scale appeared to be the result of an increased percentage of correct reclassification of both events and nonevents, the net reclassification improvement being 10%. Another important finding of our study is that even in the setting of a CRUSADE risk scale <40 points (nonhigh risk for bleeding), the presence of RDW>14.9% confers a high risk for MB, suggesting that the addition of this parameter to bleeding risk scores may be valuable.

Although examined in a relatively small cohort, our proof-of-concept analysis showing RDW as a predictor for MB beyond the hospitalization period in patients with NSTEACS is novel. In light of these findings, and the results of previous studies that have firmly established the strong link between bleeding complications and subsequent adverse outcomes, we suggest that future research should assess the potential role of including RDW values into the bleeding risk scales in order to improve the stratification process of these patients, especially after hospital discharge.

Dr. Marín has received funding for research, consultancy and lecturing from Abbott, Boston Scientifics, Bayer, AstraZeneca, Daiichi-Sankyo, BMS/Pfizer and Boehringer Ingelheim. Dr. Januzzi reports grant support from Roche Diagnostics, Critical Diagnostics, Siemens and Singulex.