Domperidone and the Risk of Serious Adverse Events

This study will assess whether initiating domperidone at doses ≥30 mg/day compared to \<30 mg/day in patients with advanced chronic kidney disease (CKD) (estimated glomerular filtration rate \[eGFR\] \<45 mL/min/1.73 m² but not receiving dialysis) is associated with a higher 30-day risk of a composite outcome of all-cause hospitalization or all-cause emergency visits or all-cause mortality.

\*Summary\* Background: Domperidone is a dopamine antagonist used to relieve symptoms of upper gastrointestinal motility disorders, such as nausea, vomiting, and delayed gastric emptying. Its use in patients with chronic kidney disease (CKD) may increase the risk of adverse events. However, current prescribing guidelines and product monographs provide conflicting recommendations for dosing in this at-risk population, leading to uncertainty about the appropriate use of domperidone in patients with CKD. The investigators have recently developed a novel high-throughput computing approach to identify medications that may harm patients with CKD using healthcare databases from Ontario. High-throughput study findings suggest that older adults who are newly prescribed domperidone face an increased risk of adverse outcomes compared to a similar cohort of non-users. To confirm these signals, the investigators will conduct a population-based cohort study among older adults with advanced CKD across two Canadian provinces: Ontario and Alberta. Methods: The investigators will conduct a retrospective cohort study using large provincial administrative healthcare databases from Ontario and Alberta, Canada. The study will include older adults (≥66 years of age) who received a prescription for oral domperidone between 2008 and 2024. In Ontario, the sample will be accrued from January 1, 2008, through September 30, 2024. In Alberta, the accrual period will be determined based on data availability. The prescription date will be the index date (start of cohort follow-up). Patients with CKD initiating treatment with domperidone will be divided into two groups based on their dosage: high dose (≥ 30 mg/day) and low dose (\<30 mg/day). The investigators will use propensity score weighting to ensure groups are comparable across baseline characteristics. Patient outcomes will be assessed 30 days following the index date. Modified Poisson regression will be used to compute the risk ratio (95% Confidence interval (CI)), and binomial regression will be used to compute the risk difference (95% CI) using the weighted cohort, with the low-dose group as the referent. \*Literature Review\* Domperidone, a peripheral dopamine-2 receptor antagonist with gastrokinetic and antiemetic effects, was first marketed in Canada in 1985 as Motilium, which has since been replaced by 13 generic versions. It is commonly prescribed to alleviate symptoms linked to upper gastrointestinal motility disorders, including diabetic gastroparesis, gastroesophageal reflux, and nausea and vomiting related to treatment for Parkinson's disease and cancer. In March 2012, Health Canada issued a warning about the potential for serious side effects associated with domperidone, such as abnormal heart rhythms and sudden cardiac death, following evidence from two studies. Domperidone's elimination process is complex: two-thirds of the unchanged drug is excreted in feces and one-third in urine. In patients with reduced eGFR, the elimination of half-life of domperidone extends from 7 to 20 hours, prompting some prescribing guidelines to recommend a lower daily dose for those with impaired renal function. Current product monographs recommended a dose reduction for domperidone in patients with eGFR \<30 mL/min/1.73 m², recommending 10 mg once or twice daily. This recommendation is based on domperidone's large volume of distribution, which means it is unlikely to be significantly removed by dialysis. Similar dosing is advised for patients on intermittent hemodialysis or peritoneal dialysis. Despite these recommendations, domperidone is still prescribed at doses over 30 mg/day, which exceeds the recommended dose for older adults with CKD in Canada. Older adults, who are often underrepresented in clinical trials, are at increased risk for adverse drug reactions due to age-related declines in kidney function and the lack of clear prescribing dosing recommendations for patients with reduced kidney function. Population-based drug safety studies tend to focus on a single medication or a limited number of outcomes, potentially overlooking significant safety signals that could be relevant for these vulnerable patients. To address this, the investigators have developed a high-throughput computing approach to identify these signals more efficiently. This approach confirmed prior safety signals, such as the increased 30-day risk of encephalopathy with baclofen use in older patients with CKD, particularly at higher doses (≥20 mg/day) compared to lower doses (\<20 mg/day) and identified new ones that warrant further investigation in another jurisdiction. Notably, preliminary findings revealed that new users of domperidone have a higher 30-day relative risk of emergency room visits, hospitalizations, and mortality compared to non-users. To validate these findings, the investigators will conduct a population-based, new-user cohort study among patients with advanced kidney disease (eGFR \<45) using linked administrative healthcare databases in Ontario, Canada, from 2008 to 2024. Additionally, the investigators will replicate this analysis using linked administrative healthcare databases in Alberta. \*Research location\* This study will utilize data from linked provincial administrative healthcare databases in Ontario and Alberta. Data for Ontario residents will be sourced from Ontario's linked administrative healthcare databases, housed and managed by ICES (ices.on.ca). The databases offer secure, encrypted data at the individual level for Ontario residents, all of whom have universal access to hospital and physician services under a government-funded, single-payer healthcare system. The use of data in this study is authorized under section 45 of Ontario's Personal Health Information Protection Act, which does not require review by a research ethics board. The Ontario administrative datasets, which include the Canadian Institute for Health Information Discharge Abstract Database (CIHI-DAD), Ontario Drug Benefit Database (ODB), ICES Physician Database, Ontario Health Insurance Plan (OHIP) Database, National Ambulatory Care Reporting System (NACRS), Ontario Laboratories Information System (OLIS), Ontario Mental Health Reporting System (OMHRS), and the Registered Persons Database (RPDB), which are linked using unique encoded identifiers and analyzed at ICES. Trained personnel code hospital admissions and diagnoses using the 10th Revision of the International Classification of Diseases (ICD-10), relying solely on physician-documented diagnoses in patients' medical charts, without reviewing or interpreting symptoms or test results. The information required for analyzing the primary outcomes is available across specific databases within the ICES system: data on all-cause hospitalizations are available in CIHI-DAD, emergency visits in NACRS, and mortality in RPDB. \*Data Access in Alberta\* Data for each study in Alberta will be accessed in one of two ways, under the guidance of a Nephrologist and Professor of Medicine and Community Health Sciences at the University of Calgary. 1. Through the Alberta Kidney Disease Network (AKDN), Nephrologist in Alberta invited Dr. Garg to participate in prior drug safety studies that used this data source to generate new information on kidney diseases. This dataset ends in \~ 2021. 2. Through the Alberta SPOR SUPPORT Unit. Dr. Garg uses this approach with an active CIHR-funded published protocol. As a recommended research practice, the investigators publicly document the study description, design, and statistical analysis plan on clinicaltrials.gov before analyzing the study outcomes. \*Statistical analysis plan\* Software: All statistical analyses will be conducted using SAS software version 9.4 (SAS Institute, Cary, NC). Descriptive Statistics: Categorical variables will be reported as frequencies and proportions, while continuous variables will be presented as means with standard deviations (SD) or medians with interquartile ranges (IQR), as appropriate. Differences in baseline characteristics between the high-dose and low-dose groups will be examined using standardized differences, with differences ≥10% considered meaningful. As described, the investigators will use the inverse probability of treatment weighting on the propensity score to balance baseline characteristics between the high-dose and low-dose groups. Balancing comparator group: Inverse probability of treatment weighting on the propensity score will be used to balance characteristics between high dose (i.e., domperidone ≥30 mg/d) and low dose (i.e., domperidone \<30 mg/d) groups based on their baseline characteristics, including known indicators for domperidone use. The investigators will conduct multivariable logistic regression analyses to generate propensity scores using all baseline characteristics. Patients in the low dose group will be assigned weights based on the average treatment effect for the treated, calculated as (propensity score / \[1 - propensity score\]). In contrast, patients in the high dose group will receive a weight of 1. This method will produce a weighted pseudo-sample of patients in the referent group (i.e., low-dose: domperidone \<30 mg/day) with a similar distribution of measured characteristics as the high-dose (domperidone ≥30 mg/d) group. Baseline characteristics between groups will be compared using standardized differences in unweighted and weighted samples, with differences exceeding 10% to be considered meaningful. Regression analysis: To evaluate the primary outcome composite measure of all-cause hospitalization or all-cause emergency department visits, or all-cause mortality- the investigators will apply a modified Poisson regression analysis to estimate the weighted risk ratio (RR) with 95% confidence intervals (CIs) and a binomial regression to estimate the weighted risk difference (RD) with 95% CIs. Additional analysis: The investigators will conduct five additional analyses. 1. The investigators will conduct a pre-specified subgroup analysis of the primary outcome stratified by pre-2020 (which overlaps with high-throughput computing) (January 1, 2008, to March 01, 2020) and post-2020 (which does not overlap with high-throughput computing) (April 01, 2020, to September 30, 2024). Alberta data will be combined with Ontario data for the subgroup analyses. 2. The investigators will analyze the primary outcome stratified by kidney function (CKD stage using serum creatinine measurements) in patients with eGFR below 30 mL/min/1.73 m², 30-\<45 mL/min/1.73 m². 3. For each cohort, the investigators will compute the hazard ratio for all outcomes within 30 days, illustrating the effect of the intervention on each outcome over time. Although the difference between hazard ratios and relative risks may be minimal over a short follow-up period, the hazard ratio offers additional insight by accounting for time-to-event analysis. 4. The investigators will conduct an analysis using a negative control exposure, the investigators will repeat the regression analyses for the initial signals, assessing the risk ratios (RRs) and risk differences (RDs) for 30-day outcomes, beginning 90 days prior to cohort entry (i.e., 90 days before the prescription start date). 5. The investigators will perform E-value analyses to determine the minimum association strength an unmeasured confounder would need with the prescription drug and the outcome of interest (while adjusting for measured covariates) to eliminate the observed association. \*Combining Outcome Results from Ontario and Alberta\* The investigators will implement one of two approaches described below to combine outcome results from Alberta and Ontario, with a preference for privacy-preserving methods, while maintaining data privacy and regulatory compliance. 1. Privacy-preserving methods: The investigators plan to use a privacy-preserving Cox-based approach for estimating risk ratios for multisite studies where individual-level data cannot be shared due to privacy constraints. The proposed method requires only a single transfer of summary-level outputs from each province to the research team. It produces results identical to those from a corresponding modified Poisson regression with combined individual-level data. The method was developed by adapting the risk-set table approach for survival outcomes, assuming stratified province-specific baseline risks, and allowing for confounding mitigation strategies such as propensity score matching or weighting to be done individually in each province. Each province will independently calculate these summary tables using their individual-level data. The summary-level risk-set tables generated in Alberta will be shared in a single data transfer to the coordinating site's analyst where they will be used to estimate the combined risk ratios and 95% confidence intervals. This will enable a robust and consistent analysis while maintaining data privacy and adhering to regulatory compliance. To comply with privacy regulations for minimizing the chance of identification of any individual, in all manuscripts, numbers of individuals are suppressed in the case of 5 or fewer participants (reported as ≤5). All team members will sign any required data confidentiality and data use agreements. The Alberta SPOR SUPPORT Unit recently approved this approach in a published protocol that combines Alberta and Ontario data. 2. Meta-analytic techniques: The investigators will combine risk ratios and risk differences across regions, following the methodology employed in our previous work. The investigators will first examine region-level statistical heterogeneity using the I² statistic (where an I² \>50% would indicate substantial variability across regions). If substantial heterogeneity is found, the investigators will present the results of each region separately. Without substantial heterogeneity, the investigators will use meta-analytic techniques to combine the results across regions using a random effects model. Summary estimates and corresponding 95% CIs will be reported within the \<45 eGFR category using appropriate methods.