Association between proton pump inhibitor use and incident depression among older adults with an exploratory analysis of suicide

The association between long-term proton pump inhibitor (PPI) use and incident depression remains uncertain among older adults. We evaluated the association between cumulative PPI use and incident depression. Suicide was assessed as an exploratory secondary outcome, and possible indirect associations involving vitamin B12 and magnesium deficiencies were examined in exploratory mediation analyses. We used Korea National Health Insurance Service-Senior Cohort of older adults from 2002 to 2019. Cumulative PPI use was assessed during the 4-year pre-index period from 2007 to 2010. Participants were followed from January 1, 2011, through December 31, 2019. Depression was set as primary outcome. Suicide was evaluated as an exploratory secondary outcome. Associations were estimated using multivariable Cox proportional hazards and Fine–Gray subdistribution hazard models, with death treated as a competing event. Mediation analyses were considered exploratory. Of the 553,362 participants, 50,872 developed depression and 57 died by suicide. Among PPI users, each 30-day increase in cumulative PPI use was associated with a higher incidence of depression (subdistribution hazard ratio [SHR], 1.020; 95% CI, 1.018–1.023). Participants with cumulative PPI use of ≥ 90 days had a higher incidence of depression than those with 1–29 days of use (SHR, 1.428; 95% CI, 1.383–1.475). Among PPI users, each 30-day increase in cumulative PPI use was associated with a higher suicide estimate (SHR, 1.067; 95% CI, 1.025–1.111), and cumulative use of ≥ 90 days was associated with a higher suicide estimate than use of 1–29 days (SHR, 3.767; 95% CI, 1.082–13.113). Exploratory mediation analyses suggested a possible indirect association involving vitamin B12 deficiency. Higher pre-index cumulative PPI use was associated with a higher incidence of depression. Suicide and mediation findings were exploratory and should not be interpreted causally.