Time-dependent markers of comorbidity and prognosis in heart failure patients: transitions across the life course

Abstract

Heart failure (HF) disease carries a poor prognosis despite optimisation of cardiovascular (CVD) treatments. Non-CVD comorbid diseases are common and known to influence the HF clinical course. These comorbidities change in severity over time from new onset yet, only static measures of comorbidity have been included in prognosis. A major gap in the management and prognosis of HF is how non-CVD comorbidity severity and longitudinal change influences individual risk.

A systematic review (SR) and two phase observational study were conducted in the general HF population to test the hypothesis that increasing severity and change of non-CVD comorbidity would be associated with worse outcomes. The SR showed that the three most common non-CVD comorbidities included in prognosis were diabetes mellitus (DM), chronic obstructive pulmonary disease (COPD) and chronic kidney disease (CKD). Hospital admission outcome studies were limited and there were no studies on quality of life. With the exception of hospital based renal studies, comorbidity severity and change evidence was scarce.

The observational studies used a case-control study nested within the UK Clinical Practice Research Datalink database (2002- 2014), of 50,114 incident HF patients. Using risk set sampling, multiple controls were matched to cases on follow-up time. A framework for measuring recent comorbidity severity and change was devised using drug or physiological indicators for DM, COPD and CKD measured in two time-windows prior to the match date. Conditional logistic regression was used to estimate adjusted odds ratios for all-cause hospital admission and mortality.

The observational study findings were that all three comorbidities were common and associated with both outcomes. Severe and worsening comorbid disease was also common and independently and significantly associated with increased risk of hospital admission and mortality. These dynamic measures of non-CVD comorbidity significantly improved HF prognostic models which has important implications for HF management and prognosis.