Proton Pump Inhibitors and Interstitial Nephritis

Abstract

Abstract Background When the research described in this thesis was initiated, twenty years of published case reports strongly suggested an association between proton pump inhibitors (PPIs) and interstitial nephritis. Absolute risk estimates implied PPI-induced interstitial nephritis was a very rare adverse drug reaction (ADR), however there were no published estimates of relative risk. Additionally, the demographic characteristics of New Zealand PPI users, and the reporting rate of suspected cases of interstitial nephritis to the Centre for Adverse Reactions Monitoring (CARM), New Zealand’s ADR spontaneous reporting centre, had not yet been investigated. Objective The objectives of the research presented in this thesis were: • To estimate the relative risks of interstitial nephritis resulting in hospitalisation or death, among current and recent PPI users, compared to past users; • To estimate the absolute risks of interstitial nephritis in current, recent, and past PPI users; • To describe the demographic characteristics of PPI users in New Zealand; • To determine the number and proportion of relevant study cases not detected through data reported to CARM; • To ascertain whether hospitalised and fatal cases reported to CARM were identified using routinely collected data. Design A nested case-control study using routinely collected demographic, health, and prescription medicine dispensing data held by New Zealand’s Ministry of Health. The complementary descriptive studies utilised routinely collected demographic information, and CARM data.

Participants The study cohort consisted of all presumed new and restarting PPI users living in New Zealand who were dispensed a PPI from 1 May 2005 to 31 August 2009, who had no previous recorded diagnoses for renal disease. Cases were patients with a first diagnosis of interstitial nephritis, and were classified as either definite (biopsy proven) or probable (no biopsy confirmation). Ten controls, matched by year of birth and sex, were randomly selected for each case using risk set sampling. Participants were categorised as current, recent, or past users based on the last PPI dispensing before their index date (date of diagnosis of case). Past users were the reference group for all analyses. Main outcome measures Odds ratios and 95% confidence intervals were estimated with conditional logistic regression. Current, recent, and past person-years of use were calculated for the entire study cohort, and used to estimate crude and age-stratified absolute risks. The Poisson distribution was used to calculate 95% confidence intervals. Key results 72 cases (including 46 definite cases) of interstitial nephritis and 719 matched controls were identified. In the case-control analysis, current use of PPIs was associated with a statistically significant five-fold higher risk of interstitial nephritis compared to past use. The unadjusted matched odds ratio was 5.2 (95% CI 2.2 – 12.1) for definite cases and controls, and 4.8 (95% CI 2.4 – 9.6) for all cases and controls. Recent use was associated with a non-significant two-fold increase in risk. The unadjusted matched odds ratio was 2.4 (95% CI 0.7 – 8.7) for definite cases and controls, and 1.7 (95% CI 0.6 – 5.2) for all cases and controls. Current use of other medicines associated with interstitial nephritis, or diagnoses associated with increased risk of renal disease in general, did not confound these relationships. The crude absolute risks of interstitial nephritis were estimated at 12.0 (95% CI 9.0 – 15.6) per 100,000 current use person-years, 4.3 (95% CI 1.4 – 10.0) per 100,000 recent use person-years, and 1.7 (95% CI 0.9 – 2.9) per 100,000 past use person-years. Current users aged 60 and older had substantially higher absolute risks compared to younger users. PPIs were widely used in New Zealand during the study period, with approximately 14.7% of the usually resident population dispensed a PPI. Users were generally older; however over 10,000 infants under one year old were dispensed a PPI during the study period. The study found 76.7% of current and recent user cases were not reported to CARM. Further investigation using available CARM and hospital discharge data revealed hospitalised and fatal cases reported to CARM, which met the study’s case definition, were detected using routinely collected data. Conclusion The research presented in this thesis is the first to show a statistically significant increased risk of interstitial nephritis in current PPI users compared to past users. Bias and confounding are unlikely to explain this finding, suggesting this is likely to be a causal relationship. The implications of this research contribute to the debate about the appropriate use of these drugs. Considering the widespread use of PPIs, clinicians should exercise caution when starting patients on the treatment, and be vigilant for abnormal renal symptoms in users.