Prognostic Implications of Declining Hemoglobin Content in Patients Hospitalized With Acute Coronary Syndromes

Background Contemporary definitions of bleeding endpoints are restricted mostly to clinically overt events. Whether hemoglobin drop per se, with or without overt bleeding, adversely affects the prognosis of patients with acute coronary syndrome (ACS) remains unclear. Objectives The aim of this study was to examine in the MATRIX (Minimizing Adverse Haemorrhagic Events by Transradial Access Site and Systemic Implementation of Angiox) trial the incidence, predictors, and prognostic implications of in-hospital hemoglobin drop in patients with ACS managed invasively stratified by the presence of in-hospital bleeding. Methods Patients were categorized by the presence and amount of in-hospital hemoglobin drop on the basis of baseline and nadir hemoglobin values and further stratified by the occurrence of adjudicated in-hospital bleeding. Hemoglobin drop was defined as minimal (<3 g/dl), minor (≥3 and <5 g/dl), or major (≥5 g/dl). Using multivariate Cox regression, we modeled the association between hemoglobin drop and mortality in patients with and without overt bleeding. Results Among 7,781 patients alive 24 h after randomization with available hemoglobin data, 6,504 patients (83.6%) had hemoglobin drop, of whom 5,756 (88.5%) did not have overt bleeding and 748 (11.5%) had overt bleeding. Among patients without overt bleeding, minor (hazard ratio [HR]: 2.37; 95% confidence interval [CI]: 1.32 to 4.24; p = 0.004) and major (HR: 2.58; 95% CI: 0.98 to 6.78; p = 0.054) hemoglobin drop were independently associated with higher 1-year mortality. Among patients with overt bleeding, the association of minor and major hemoglobin drop with 1-year mortality was directionally similar but had wider CIs (minor: HR: 3.53 [95% CI: 1.06 to 11.79]; major: HR: 13.32 [95% CI: 3.01 to 58.98]). Conclusions Among patients with ACS managed invasively, in-hospital hemoglobin drop ≥3 g/dl, even in the absence of overt bleeding, is common and is independently associated with increased risk for 1-year mortality.