Association of habitual coffee consumption and kidney function: A prospective analysis in the Rotterdam Study

Background & aims: Population-based studies have suggested a protective effect of coffee against development of chronic kidney disease (CKD), possibly through coffee's anti-inflammatory and antioxidant compounds. Studies on coffee and kidney function decline in the general population are scarce. We studied associations of habitual coffee consumption with repeated assessments of estimated glomerular filtration rate (eGFR) and urinary albumin-to-creatinine ratio (ACR). Methods: We used data from 7,914 participants of the population-based Rotterdam Study. Baseline coffee consumption data (cups/day) were obtained from home interviews and validated food frequency questionnaires (1997–2008). Repeated assessments of eGFR (ml/min per 1.73 m2, 1997–2014) were calculated according to the creatinine-based CKD Epidemiology Collaboration equation of 2012. Repeated assessments of urinary albumin and creatinine were used to estimate ACR (mg/g, 2006–2014). Data were analyzed by applying linear mixed models, adjusted for sociodemographic, lifestyle and dietary factors, and cardiovascular disease risk factors. Predefined subgroup analyses were performed stratified by CKD risk factors. Results: Participants’ mean (SD) baseline age was 66 (10) years, 57% were women and median [IQR] coffee consumption was 3.0 [2.0, 5.0] cups/day. Those drinking more coffee were more likely to smoke, and to have type 2 diabetes (T2D) and obesity. Mean eGFR was 79 (15) ml/min per 1.73 m2. In the total study population, coffee was not associated with longitudinal eGFR during a median of 5.4 years of follow-up (β = 0.04 ml/min per 1.73 m2 per one cup/day [95% CI: −0.10,0.18]). However, among those aged >70 years, one additional coffee cup/day was associated with on average 0.84 (0.51,1.18) ml/min per 1.73 m2 higher longitudinal eGFR. Among obese participants this estimate was 0.32 (0.01,0.63). A protective trend was also observed among former smokers (0.17 [−0.03,0.39]) and those with T2D (0.42 [−0.05,0.88]). Coffee was not associated with longitudinal ACR (0.01 mg/ml [−0.01,0.02]). Conclusion: While coffee was not associated with eGFR and ACR in the total population, more coffee consumption was associated with higher longitudinal eGFR among those at higher risk for CKD, i.e., among those aged 70+ and obese participants. These findings require confirmation in other prospective cohort studies.