Kidney function decline associated with proton pump inhibitors: results from the ELSA-Brasil cohort

Abstract

Objective: Investigate the longitudinal association of use and time of use of proton pump inhibitors (PPI) with incidence of chronic kidney disease (CKD) and kidney function change.

Methods: Prospective study with 13,909 participants from baseline (2008-2010) and second wave (2012-2014) of the ELSA-Brasil (mean interval between visits = 3.9 years (1.7-6.0)). Participants answered about use and time use of the PPI in the two weeks prior the interview. Renal function was assessed by glomerular filtration rate estimated by the Collaboration Equation for the Epidemiology of Chronic Kidney Disease. Values below 60ml/min/1.73 m² in wave 2 were considered incident CKD. Associations between PPI use and time of use at baseline and incident CKD and decline in renal function were estimated, respectively, by logistic regression and linear models with mixed effects, after adjusting for confounders.

Results: After adjustments, PPI users for more than six months had an increased risk of CKD compared to non-users. Compared to non-users, users PPIs for up to six months and above six months had greater decline in kidney function over time.

Conclusion: This cohort of adults and elderly, after a mean interval of 3.9 years, PPI use and initial duration were associated with kidney function change between visits.

Keywords: Chronic kidney disease; Kidney function; Pharmacoepidemiology; Pharmacovigilance; Proton pump inhibitors.

Fig. 1

Study population flowchart

Fig. 2

Longitudinal trajectories* of estimated glomerular filtration rate (eGFR) according to time of use of proton pump inhibitors (PPI). ELSA-Brasil (2008–2010 and 2012–2014). *As age was modeled as a random effect in the data analysis, this figure shows changes in estimated glomerular filtration rate over time (i.e., as individuals age). †Predicted numbers are dependent variable values based on estimated regression coefficients and a prediction of independent variable values after adjustments (age, sex, per capita household income, excessive alcohol consumption, obesity, cardiovascular disease, diabetes, hypertension, use of NSAIDs, BRAs, ACEs and interaction: time of PPI use × age)