The study aims to establish the association between medication classes and amyloid deposition in the brain.
Data from 824 cognitively normal subjects in the longitudinal APOE cohort study who underwent amyloid PET scans were analyzed. The centiloid score was compared with the total number of medications and the number of medications in various categories that the subject was taking at the time of the scan. The cohort was stratified by APOE ε4 allelic status (non-carriers, heterozygotes, and homozygotes) and gender. Spearman correlation coefficients were calculated to evaluate monotonic relationships. Thirteen medication classes were included as binary predictors in separate multivariable linear regression models.
The dataset included 577 females (70.02%) and 247 males (29.98%). The non-carriers, heterozygotes, and homozygotes comprised 429, 258, and 138 subjects, respectively. No significant correlation was observed between the total number of medications taken and centiloid scores across groups.
In the non-carriers, females showed a positive correlation with cardiovascular medications (p<0.01), while males showed positive correlation with the cardiovascular medications and ‘other supplements’ category (p<0.01).
In the heterozygous group, females showed positive correlations with cardiovascular medications, corticosteroids, anti-inflammatory, and opioids, and negative correlations with hypoglycemics, sex hormone treatment, and psychoactive medications (p<0.05). Males showed positive correlations with immunosuppressants, corticosteroids, and psychoactive medications.
Among homozygotes, females showed positive correlations with cardiovascular medications, opioids, immunosuppressants, and sex hormone treatment. Males showed positive correlation with hypoglycemic medications (p<0.01).
There is a consistent correlation between certain medication categories (cardiovascular medications and opioids) and amyloid deposition. Other classes, like immunosuppressants and psychoactive medications, showed variable correlations depending on genotype and gender. These correlations could be due to the underlying comorbidities rather than the medications, highlighting the need for further investigation.