AIMS: Heart failure (HF) leads to an imbalance between heart and kidney function, resulting in poor outcomes. However, the prognostic significance of the estimated glomerular filtration rate (eGFR) trajectory in HF patients remains unclear. We analysed electronic health records (EHRs) of real-world HF patients, assessing eGFR trajectories and their impact on mortality. METHODS AND RESULTS: Retrospective clinical data of HF patients were processed using natural language processing. Chronic kidney disease (CKD) was evaluated, and eGFR trajectories were analysed using linear mixed-effects models. Cox proportional hazard models were used to evaluate the relationship between baseline variables and mortality, while joint modelling combined eGFR trends and mortality. The dataset comprised 1986 patients, with a mean age of 74.8 years (SD ± 11.7) and 58% male. At the time of HF diagnosis, 58% of patients were diagnosed with CKD, and 39% presented with heart failure with preserved ejection fraction (HFpEF). The median follow-up duration was 3.16 years, during which 399 patients (20%) died. Patients with CKD were significantly older and exhibited a higher prevalence of myocardial infarction (P = 0.048), coronary revascularization (P = 0.004), stroke (P <
0.001), atrial fibrillation (P <
0.001) and type 2 diabetes (P <
0.001). Mortality rates at 1 and 2 years were nearly twice as high in CKD patients compared with those without (P <
0.001). Notably, CKD was significantly less prevalent among survivors (55% vs. 71%, P <
0.001). Key predictors of mortality included older age, beta-blocker use, prior stroke, lower serum haemoglobin levels, and elevated potassium and NT-proBNP levels. Each 10 mL/min/1.73 m CONCLUSIONS: CKD is prevalent in a real-world HF population and is an independent predictor of mortality. The current eGFR value and the eGFR slope from the previous year have the potential to be used for assessing individual mortality risk in the clinical follow-up of HF patients.