Medical and non-medical predictors of initiating long-term use of proton pump inhibitors: a nationwide cohort study of first-time users during a 10-year period

Abstract

Background: Studies of the increasing use of proton pump inhibitors (PPIs) have mainly focused on prevalent long-term use and associations with gastrointestinal morbidity and comedication. Little is known about non-medical characteristics of first-time users of PPI, and predictors of initiating long-term use of PPIs. Aims: To describe medical and non-medical characteristics of first-time PPI users during a 10-year period and to analyse predictors of initiation of long-term use (>60 defined daily doses (DDDs) within 6 months) of PPIs. Methods: A nationwide cohort study of first-time users of PPI. Data were collected from Danish national registers. Individuals redeeming their first prescription for a PPI (omeprazole, lansoprazole, pantoprazole, rabeprazole, esomeprazole) in 2001 and 2011 were identified. Redemption of more than 60 DDDs of PPI within 6 months defined long-term use. Logistic regression models were used to determine the associations between previous diagnoses, comedication and socio-economic characteristics and initiation of long-term use of PPIs in 2011. Results: From 2001 to 2011 incidence of first-time users increased with an incidence rate ratio of 1.53 and mean quantity of PPI redeemed at first prescription increased by 44.6%. In 2011 a total of 37.6% redeemed >60 DDDs within 6 months, and 96% of the long-term users did not have a diagnosis registered which indicated treatment. New onset long-term use was significantly associated with low income and low educational level when adjusting for other predisposing variables. Conclusions: Proton pump inhibitor treatment is increasingly initiated with larger quantities prescribed for indications that are unidentifiable from the registers. Morbidity and comedication seem to be the strongest predictors of new onset long-term use of PPIs. However, there is also an independent social gradient.