Copyright 2016 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
In Reply We share the concern expressed by Dr Lozano, Iannuzzella et al, Zhao et al, and Drs Ness-Jensen and Fossmark that observational studies may suffer from residual confounding. In the original study,1 to ensure that the observed associations were robust, we performed multiple sensitivity analyses and replicated results in a separate cohort. These analyses included adjustment for many possible confounders, including concomitant medication use, hypertension, body mass index, and diabetes mellitus. We also tested associations using an active comparator group (patients using H2 antagonists), a new-user design, and propensity-score matching. In the latter analysis, matched cohorts were well balanced (Table 1, as requested by Zhao et al), and the study design obviates the concern about multicollinearity that was raised by Drs Ness-Jensen and Fossmark. All sensitivity analyses were consistent and showed an association between proton pump inhibitor (PPI) use and incident chronic kidney disease (CKD).
Lazarus B, Coresh J, Grams ME. Adverse Effects of Proton Pump Inhibitors in Chronic Kidney Disease—Reply. JAMA Intern Med. 2016;176(6):869-870. doi:10.1001/jamainternmed.2016.1863