|| Checking for direct PDF access through Ovid
The course of patients following admission for acute heart failure (AHF) is of major importance to patients and healthcare providers. We examined predictors and associations of length of stay (LOS), 30-day post-discharge readmission and 90-day post-discharge mortality in 1990 patients enrolled in the PROTECT study.PROTECT was a randomized study that examined the effect of the adenosine blocker rolofylline in patients within 24 h of admission for AHF with mild to moderate renal impairment. Geographic-region-adjusted multivariable models showed that LOS was only partly explained by the severity of heart failure (HF), comorbidities (diabetes mellitus, renal impairment, ischaemic heart disease) and degree of metabolic dysfunction (cholesterol and albumin) at baseline (adjusted R2 0.27). Addition of in-hospital worsening heart failure (WHF) and changes in metabolic markers contributed significantly to prediction of LOS [R2 difference 0.050, 95% confidence interval (CI) 0.0282–0.072]. Thirty-day HF readmission was associated with more severe HF and previous HF admission but not with LOS (odds ratios 1.00, 95% CI 0.97–1.04). Death within 90 days after discharge was associated with older age, more severe HF, worse renal function, and lower sodium and bicarbonate at admission; LOS was a strong predictor of 90-day post-discharge mortality.In patients admitted for AHF, LOS is not well-predicted by traditional markers of disease severity, but strongly associated with the occurrence of in-hospital WHF. Longer LOS is a strong predictor of early mortality after discharge but not of readmission. These findings may help focus efforts to reduce LOS and post-discharge outcomes on patients' subgroups at increased risk.