Postoperative anaemia and long-term survival following cardiac surgery: a multicentre, retrospective, observational study
Abstract
Introduction
Pre-operative anaemia is associated with adverse outcomes following cardiac surgery. However, the prognostic impact of postoperative anaemia is characterised incompletely. We examined whether anaemia at hospital discharge independently predicts long-term survival and characterised haemoglobin recovery trajectories during the first postoperative year.
Methods
This retrospective multicentre cohort study included 25,697 adult cardiac surgery patients from two Danish tertiary centres (2000–2024). Anaemia was classified using World Health Organization sex-specific criteria as mild (haemoglobin 110–129 g.l-1 for men; 110–119 g.l-1 for women) or moderate–severe (haemoglobin < 110 g.l-1 for both). The primary outcome was overall survival from surgery to death or censoring at end of follow-up. Cox proportional hazards regression assessed associations between anaemia at discharge and mortality, adjusting for age, sex, pre-operative haemoglobin and comorbidities. Restricted cubic splines examined non-linear relationships between haemoglobin and mortality.
Results
At discharge, 23,935 (93.1%) patients were anaemic (8384 (32.6%) mild anaemia and 15,551 (60.5%) moderate–severe anaemia). Moderate–severe anaemia was independently associated with worse survival (hazard ratio 1.11, 95%CI 1.01–1.21, p = 0.034) in adjusted analysis. Cubic spline analysis revealed a steep, non-linear relationship between pre-operative haemoglobin and mortality, whereas discharge haemoglobin provided limited prognostic information beyond baseline characteristics and surgical complexity. Haemoglobin recovery plateaued around 4 months after discharge, but the average haemoglobin level remained below baseline even after 1 year.
Discussion
Postoperative anaemia showed a weak independent association with long-term mortality, in contrast to the clear dose–response relationship observed for pre-operative anaemia. The near-universal prevalence of postoperative anaemia and incomplete haemoglobin recovery at 1 year support further trials of peri-operative anaemia management.