Preoperative Hemoglobin Level as a Predictor of Mortality After Aortic Valve Replacement

ObjectivesThe predictive value of preoperative hemoglobin (HB) level on the outcome of patients undergoing valve surgery is not well established. This study evaluated the predictive value of preoperative HB level on survival after aortic valve replacement (AVR).DesignThis was a retrospective analysis of prospectively collected data.SettingA single-center study performed in an educational hospital.ParticipantsAll consecutive patients (n = 1,808) who underwent AVR between January 1998 and December 2010.InterventionsAVR.Measurements and Main ResultsPatients were classified into 4 groups according to the preoperative HB level: very low (HB of <12 g/dL in men and <11 g/dL in women), low (HB of 12-13 g/dL in men and 11-12 g/dL in women), normal (HB of 13-14.5 g/dL in men and 12-13.5 g/dL in women), and high normal (HB of ≥14.5 g/dL in men and ≥13.5 g/dL in women). The mean follow-up duration was 5.58±3.5 years, and the median follow-up duration was 5.38 years. The mean preoperative HB was 14±1.6 g/dL for men and 13.0±2.1 g/dL for women. Early mortality (≤30 days) was 6.1% in the very-low-HB group, 5.4% in the low-HB group, 3.2% in the normal HB group, and 2.3% in the high-normal-HB group (p = 0.37). Late mortality (>30 days) was 26.1% in the very-low-HB group, 23.7% in the low-HB group, 17.1% in the normal-HB group, and 12.6% in the high-normal-HB group (p<0.0001). The multivariate logistic regression model did not identify low HB as an independent predictor for early mortality. Cox regression multivariate analysis revealed both HB level, as a continuous variable, (p = 0.006), and very-low-HB level (p<0.0001), as independent predictors of late mortality. Cox regression analyses, corrected for confounders, demonstrated that low-HB level is an independent predictor for higher overall mortality (hazard ratio = 2.00, CI 1.41-2.85, p≤0.0001).ConclusionsIn patients undergoing AVR, preoperative low-HB level is an independent risk factor for late mortality, but not for early mortality.