Proton Pump Inhibitors and CKD

Proton pump inhibitor (PPI) use is extremely common worldwide, and the medication class is available without a prescription in many countries. PPIs work by altering the acidity of the stomach and are helpful in treating gastroesophageal reflux disease. However, these physiologic alterations may also lead to greater risk of nutritional deficiencies, bone fractures, and infections. Eight years ago, we found that PPI use was also linked to a slightly higher risk of CKD using observational data in two cohorts.¹ Multiple subsequent studies from different data sources—Veterans Affair data, population data from Stockholm, Sweden, among others—reported the same phenomenon.^2,3 In this issue of JASN, Pyne et al. expand the observational evidence by performing a post hoc analysis of a clinical trial suggesting consistent results: a small but significant increase in eGFR decline in the setting of PPI use.⁴

Although randomized clinical trials are the gold standard in establishing causal relationships between treatments and outcomes, trials are not always feasible or even ethical. In these cases, methods such as target trial emulation using observational studies can be helpful. Pyne et al. perform a variation on this approach with a post hoc analysis of the Cardiovascular Outcomes for People Using Anticoagulation Strategies clinical trial, a study randomizing participants with coronary artery disease or peripheral artery disease to rivaroxaban and aspirin, rivaroxaban alone, or aspirin alone. Among those patients who were not taking a PPI at study enrollment, PPI use was also randomized. The original trial reported no increase in risk of the safety outcomes of incident CKD, AKI, acute nephritis, or nephrotic syndrome among 17,598 participants randomized in the PPI study. However, the study ascertained incident CKD through case report forms that did not specify criteria for CKD nor abstract creatinine or eGFR from medical records. In the new post hoc analysis, the authors instead evaluated change in eGFR, leveraging eGFR ordered in routine care within the year before baseline and the open-label extension.

Among the 17,598 participants in the original analysis of kidney safety, 8991 had follow-up tests that enabled estimation of change in eGFR in the post hoc analysis. Mean eGFR at study initiation was 75 ml/min per 1.73 m² in both the placebo and PPI groups, and the rate of decline was −1.41 ml/min per 1.73 m² per year and −1.64 ml/min per 1.73 m² per year, respectively. In adjusted analysis, this translated to a faster eGFR decline by 0.27 ml/min per 1.73 m² per year within the group receiving PPIs (P < 0.001). Among the participants with eGFR ≥60 ml/min per 1.73 m², a composite kidney end point (follow-up eGFR <60 ml/min per 1.73 m², CKD on the case report form, death from kidney failure, or exclusion from follow-up because of eGFR <15 ml/min per 1.73 m²) was suggestive of a slightly higher risk associated with PPI use, although this was not statistically significant (odds ratio, 1.11; 95% confidence interval, 0.98 to 1.25; P = 0.09).

Acknowledging the limitations of short follow-up, potential selection bias invoked by requiring enrollment in the open-label extension, and the imprecision of a laboratory measure obtained for clinical purposes, the results of the study by Pyne et al. are generally supportive of previous observations: PPI use seems to confer a slight increase in the risk of CKD. Without more detailed data, the nature of this risk is left unexplained. PPIs have a consistent association with acute interstitial nephritis, which may present acutely or more indolently.⁵ While the rate of AKI did not differ significantly in those randomized to PPI versus placebo in this trial, AKI was ascertained using administrative codes, rather than changes in creatinine, and would not be expected to detect subclinical interstitial nephritis. PPIs also lead to magnesium depletion because magnesium absorption is dependent in part on low pH in the stomach, and hypomagnesemia may contribute to worsening kidney function through diverse processes.^6,7 Finally, it is worth noting that small changes in eGFR can be driven by changes in filtration or changes in creatinine generation, which is predominantly determined by muscle mass. Similar results could occur if PPI use were associated with changes in muscle mass.⁸ Future studies may benefit from examining changes associated with PPI administration for other GFR markers, such as cystatin C, in addition to creatinine.

For most patients, PPI use would be expected to be safe from a kidney perspective. As with all medications, however, optimal prescribing will weigh the balance of risks and expected benefits. An 8-week course of PPIs may be beneficial in managing symptoms and histological signs of gastroesophageal reflux disease, and most experts recommend chronic PPI use in patients with erosive esophagitis, Barrett's esophagus, esophageal ulcers, or peptic strictures. However, the benefits of PPI use in other settings may be small. For example, PPIs do not improve throat symptoms thought to relate to gastroesophageal reflux disease.⁹ The American Gastroenterology Association states, “Although PPIs are generally safe, patients should not use any medication when there is not a reasonable expectation of benefit based on scientific evidence or prior treatment response.”¹⁰ With some studies suggesting nearly one in four adults use PPIs,¹¹ it would appear there are many who have not gotten the message.

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/JSN/E687.

Funding

None.

Author Contributions

Writing – original draft: Morgan E. Grams.

Writing – review & editing: F. Perry Wilson.