Eeva Pylkkö

Background: Several Alzheimer's disease (AD) clinical trials currently focus on drug repurposing. Machine learning (ML)-based pharmacoepidemiological approaches may help identify candidate drugs for trials. Objective: We applied an ML-based signal detection method to identify drugs associated with AD risk. Methods: Using Finnish nationwide register-based data, we defined drug exposure at both subgroup and drug substance levels according to the Anatomical Therapeutic Chemical classification. Three binary embedded variables were calculated for each drug based on the frequency of dispensations over three years, starting 10 years before the index date. An algorithm combining subsampling and lasso-logistic regression was applied, with 500 replications for each drug exposure variable. Drug variables with a selection percentage (SP) ≥ 70% across replications were considered as signals. The average beta describes the direction and strength of the association. Results: The study population comprised 43,926 AD cases and 43,926 matched controls. The mean age was 80.3 years in both groups, and 65% of participants were women. Four drug subgroups and three individual drugs were associated with a lower risk of AD, whereas 14 subgroups and 11 individual drugs were associated with a higher risk. Drugs associated with decreased AD risk belonged to cardiovascular drugs: antihypertensives affecting the renin-angiotensin and diuretics. Conclusions: The identified associations are consistent with established risk factors of AD and support the importance of effective hypertension management in AD prevention. Further studies are needed to evaluate whether antihypertensive drugs may exert neuroprotective effects and represent viable candidates for drug repurposing in AD.

Machine learning (ML)–based methods have been proposed as a potential approach for identifying candidate drugs to be repurposed as disease‐modifying treatments for Parkinson's disease (PD). We applied an ML‐based signal detection method to identify drugs associated with PD and evaluated the method's generalizability. An algorithm combining subsampling and lasso logistic regression was implemented in a case–control study of 12 257 PD cases and 81 103 matched controls identified in Finnish registers. Drug exposure was defined at the subgroup and drug substance levels of the Anatomical Therapeutic Chemical (ATC) classification, considering the frequency of dispensation over 2 years, starting 10 years before the index date (8‐year lag). Three subgroups and two individual drugs were associated with reduced PD risk. Inhalant anticholinergics, in particular tiotropium bromide, showed the most robust signal. Other signals included antimalarial drugs (aminoquinolines) and the antibiotic subgroup lincosamides. Several drugs were associated with increased PD risk, as expected. In addition to direct pharmacological effects, observed associations could be due to treatment of prodromal symptoms of PD, increased comorbidity in individuals later diagnosed with PD or a combination of these factors. These results support the feasibility of the approach. Associations of decreased PD risk observed should be further investigated in view of drug repurposing.