Association of Kidney Function With Incident Heart Failure: An Analysis of the Women's Health Initiative

Abstract

Background: Studies have shown an association of chronic kidney disease with heart failure (HF); however, this association has not been adequately examined in postmenopausal women, who are at heightened risk of both chronic kidney disease and HF. Additionally, association with HF subtypes is not well characterized.

Methods and results: Incident HF was defined as first hospitalization for acute decompensated HF, obtained by self-reported outcomes followed by physician adjudication through review of hospital records. Chronic kidney disease was defined using estimated glomerular filtration rate (eGFR). Restricted cubic splines tested the association of eGFR with incident overall HF, and HF with reduced ejection fraction (HFrEF) and preserved EF (HFpEF). Cox proportional hazards regression models evaluated the multivariable-adjusted association of eGFR categories with incident HF and its subtypes. The primary analysis included 23 309 women with 11 814 eGFR ≥90, 10 191 eGFR between 60 and 89, 1048 eGFR between 45 and 59 and 256 eGFR <45 mL/min per 1.73 m². For overall HF, HFrEF and HFpEF, there was a stepwise increase in risk for incident HF with declining eGFR category. Associations were stronger for HFpEF (hazard ratio [HR], 2.80 [95% CI, 2.36-3.32]) than for HFrEF (HR, 2.18 [95% CI, 1.66-2.87]) for eGFR <45 as compared with eGFR ≥90. Heterogeneity of the HF subdistributions (HFpEF versus HFrEF) was significant (P=0.017).

Conclusions: Kidney dysfunction is associated with incident HF in postmenopausal women. Although lower eGFR is associated with both incident HFrEF and HFpEF, the association is stronger with HFpEF.

Registration: URL: https://clinicaltrials.gov; Unique Identifier: NCT00000611.

Keywords: HFpEF; chronic kidney disease; heart failure.

Figure 1. Consolidated Standards of Reporting Trials diagram for the current study.

The final cohort included 23 309 participants after applying exclusion criteria. *Exclusion criteria included no creatinine (n, 137 603), preexisting heart failure (n, 423), not part of the heart failure cohort (n, 459) or missing follow‐up (n, 14). eGFR indicates estimated glomerular filtration rate; HF, heart failure; and WHI, Women's Health Initiative.

Figure 2. Age‐adjusted incidence rates per 1000 person‐years for heart failure and its subtypes by eGFR level.

Incidence rates for each type of heart failure increased with declining eGFR groups. eGFR indicates estimated glomerular filtration rate; HF, heart failure; HFmEF, heart failure with mid‐range ejection fraction; HFpEF, heart failure with preserved ejection fraction; and HFrEF, heart failure with reduced ejection fraction.

Figure 3. Spline analysis for eGFR risk and heart failure (adjusted for age).

The y axis is log hazard ratio and x axis is eGFR. Estimated splines (blue lines) with their 95% CIs (red lines) are shown for (A) Any incident heart failure (P<0.001 for linearity); (B) Heart failure with preserved ejection fraction (P<0.001 for linearity); (C) Heart failure with reduced ejection fraction (P<0.001 for linearity); (D) Heart failure with midrange ejection fraction (P, 0.484 for linearity). eGFR indicates estimated glomerular filtration rate; HFmEF, heart failure with mid‐range ejection fraction; HFpEF, heart failure with preserved ejection fraction; and HFrEF, heart failure with reduced ejection fraction.

Figure 4. Forest plot of fully adjusted* hazard ratios for each 10‐unit decrease in eGFR (95% CIs) for overall heart failure stratified by selected risk factors.

*Adjusted for age, race, ethnicity, marital status, income, education, diabetes, hypertension, systolic blood pressure, diastolic blood pressure, atrial fibrillation, hysterectomy, history of coronary heart disease, body mass index, physical activity, smoking status, diet quality, coffee intake, alcohol intake, and ability to walk 1 block (note—risk factor excluded from model when examined). CHD indicates coronary heart disease; eGFR, estimated glomerular filtration rate; and HR, hazard ratio.