The association of proton pump inhibitors and hypomagnesemia in the community setting.

Evidence for the association between hypomagnesemia and proton pump inhibitors (PPIs), highlighted by the 2011 FDA Drug Safety Communication, rests mainly on studies in hospitalized patients. Our objectives were to determine the prevalence of hypomagnesemia and its association with PPIs in the community setting. We performed a retrospective cross-sectional analysis of a large health maintenance organization administrative database, including ambulatory patients with ≥1 serum magnesium concentrations between 2008 and 2011, the lowest referred to as "index magnesium." In cases with any (index magnesium ≤0.7 mmol/L) or severe (≤0.55 mmol/L) hypomagnesemia, we analyzed (vs. controls, >0.7 mmol/L) the association with PPI or H2 -blocker use during the 4-12 months preceding the index magnesium by logistic regression analysis, adjusting for confounders. Among 95,205 subjects, 5,696 (6.0%) had any hypomagnesemia, which was severe in 454 (0.5%), with twofold higher prevalences in those with established risk factors. PPI use during the 4 months preceding the index magnesium was more common in cases of any hypomagnesemia (adjusted OR = 1.66; 95% CI, 1.55-1.78) and severe hypomagnesemia (adjusted OR = 3.79; 2.99-4.82) than in controls without acid suppression. Hypomagnesemia remained significantly associated with PPI use when using H2 -blocker-users as reference (adjusted OR = 1.25 [P = 0.003] and 2.65 [P < 0.001] for any and severe hypomagnesemia, respectively). We conclude that hypomagnesemia is associated with PPI use in ambulatory patients.