Peptic ulcers with ChEIs, NSAIDs

Abstract

We read with great interest the Szilcz's study showing that the risk of peptic ulcer increased for the combination of cholinesterase inhibitors (ChEIs) and non-steroidal anti-inflammatory drugs (NSAIDs) more than for NSAIDs alone¹ in patients ≥65 years. Their work was a self-controlled study. Since approaches in pharmacovigilance and pharmacoepidemiology should be multisource,² we investigated this possible drug interaction (DI) using disproportionality analyses^{3, 4} in the global pharmacovigilance database Vigibase®.

All reports with ChEIs (N06DA following Anatomical Therapeutic Chemical (ATC) classification) and NSAIDs (M01AA butylpyrazolidines, M01AB acetic acid derivatives and related substances, M01AC oxicams, M01AE propionic and derivatives, M01AG fenamates, M01AH coxibs) registered as “suspected/interacting” in Vigibase® between 01/01/1994 and 31/12/2023 in adults (≥65 years) with known age and sex were included. Disproportionality analyses^{3, 4} were performed with cases being reports of “gastrointestinal ulcerations and perforations” (GUP) (HLGT according to Standardized MedDRA Queries classification, excluding anal, rectal, and esophagus ulcers) with the drug(s) of interest and non-cases all other reports with the same drug(s) of interest. Following this case non-case analysis,^{3, 4} results with ChEIs + NSAIDs were compared with NSAIDs alone. To minimize the potential reporting bias and increase the medical meaning, a sensitivity analyses was performed only including reports by physicians. Results are presented as reporting odds ratios (ROR),^{3, 4} a ratio similar in concept to the odds ratio in case–control studies with their 95% confidence interval. The research was performed and paper written according to the READUS-PV consensus statement for drug safety signal detection using Individual case safety reports in pharmacovigilance.^{5, 6}

Among the 7,054,411 reports registered in VigiBase® according to the criteria defined above, 31,494 were GUP with 283 including ChEIs alone (mainly donepezil 49.5%), 9060 NSAIDs alone (mainly propionic drugs like ibuprofen 33.0%) and 29 the combination ChEIs + NSAIDs. Patients were mainly women (57.9% for NSAIDs, 54.1% for ChEIs, 82.8% for combination). Most of them were ≥75 years old (60.0% for NSAIDs, 78.4% for ChEIs, 79.3% for combination).

Table 1 shows the number of GUP reports. Significant ROR values were found for ChEIs alone, NSAIDs alone and their combination for all reports (whatever the reporter) as well as for reports only coming from physicians. ROR values for the comparison ChEIs + NSAIDs vs NSAIDs alone was 3.24 (2.18–4.81) for all reports and 2.64 (1.56–4.46) for physicians.

Using a validated method for detecting risk signals,^3-6 our results are in line with the self-controlled study by Szilcz¹ with the sole exception of a significant ROR value for ChEIs alone. In fact, peptic ulcer is a known adverse drug reaction (ADR) with ChEIs.⁷ Our results allow to conclude to a potentiating synergistic drug interaction, since the ROR value of the combination (37.85) was significantly higher than the ROR sum for ChEIs (2.03) and NSAIDS (15.94) alone. This was true for all reports as well as for from those from physicians alone. It is interesting to underline that the present work also extends the Szilcz's data, since the direct comparison showed a reporting risk approximately three times higher with the combination ChEIs + NSAIDs than with NSAIDs alone. This was never quantified previously.

Strengths of this type of study on large pharmacovigilance databases are well known. We used a global database, allowing to extend the results of clinical trials in real life. Under-reporting, which is classic and compulsory in pharmacovigilance, did not represent a limitation, since our aim was to describe the reports in the global pharmacovigilance database and not to be exhaustive. The calculated risk (ROR values) does not relate to the true risk of occurrence but to the risk of reporting risk, as underlined several times in the text. Disproportionality analyses alone cannot measure incidence. Other limitations of the study include the lack of adjustment for potential health confounding variables (tobacco, alcohol, …), which was not possible as these informations are not systematically registered in Vigibase®.^{5, 6} However, as is usual in large pharmacovigilance studies, the very large number of cases justifies the current interpretation of the results. Finally, with more than 31,000 patients, our study, is one of the largest series ever published on GUP reports.

From a practical point of view, our study, which described a drug–drug interaction between NSAIDS and ChEIs, suggests that demented patients receiving ChEIs should avoid the association. However, if the prescription is compulsory, patients should be carefully monitored. Finally, the study confirms the importance of working on several databases and using several methods in pharmacoepidemiology to detect drug interactions or ADRs.

JLM designed the study, extracted the data from the database, performed the statistical analysis, analyzed the data, and wrote the paper.

JLM declares no conflict of interest.

No sponsor for this research work.

The work was performed during the university research time of the authors using the database, which is available without fees to the authors. There were no funding sources.

The work was performed using the database, Vigibase®, which is freely available to the author. None informed consent was required from patients since the database Vigibase® is anonymized. Institutional review board approval was not required because Vigibase® is an anonymized database open to pharmacovigilance centers. Patients' informed consent was not necessary since data from VigiBase® were deidentified. According to French law, review from an ethics committee is not required for such observational studies.