Differential effect of the diagnostic measures- CDRSB and clinical diagnosis on drug class associations with dementia

Epidemiological studies have shown an association between various medications and late life cognitive decline and dementia. This study evaluated some of the commonly prescribed drug classes and their association with progression to dementia, diagnosed using both Clinical Dementia Rating sum of boxes (CDRSB) scale and clinical diagnosis. We employed multivariate logistic regression on the longitudinal NACC data to identify drug classes associated with risk of progression to mild cognitive impairment (MCI) and dementia including antihypertensives, lipid lowering medication, NSAIDs, anticoagulants, antidepressants, antipsychotics, anxiolytics and antidiabetic drugs, generally and in men and women separately. Antihypertensives, Non-steroidal Anti-inflammatory Drugs (NSAIDs), anxiolytics and antidiabetic drugs were significantly associated with progression to MCI and/or dementia. Discrepancies were observed between clinical diagnosis and CDRSB analyses; NSAIDs reduced Healthy-to-MCI risk with clinical diagnosis as the outcome (odds ratio [OR]: OR:0.64, FDR p<0.001), but not for CDRSB. Additionally, some associations appear to be gender specific; antidiabetic drugs reduced MCI-to-Dementia risk for women specifically (OR:0.58, FDR p=0.006). In men, antihypertensives such as antiadrenergic agents (0.71, p: <2e-04), and Ca channel blockers (OR: 0.74, p: 0.0012) were associated with reduced risk of MCI-to-Dementia progression. This study demonstrates a need for consensus regarding the most appropriate diagnostic outcome measure to realise potential benefits/disadvantages associated with specific drugs. Additionally, gender differences associated with different drug classes warrants further investigation.