Reanalysis of a National Alzheimer's Coordinating Center cohort reveals that the association between documented glucosamine use and faster progression from mild cognitive impairment (MCI) to Alzheimer's disease is not glucosamine-specific. When glucosamine records were compared against other supplements recorded in identical EHR medication fields—including multivitamins, calcium, and vitamin D—similar risk-direction estimates emerged across all comparators. Additionally, glucosamine records showed no significant association with adjusted mortality among dementia patients, though the authors acknowledge limitations in death-record completeness.
This finding matters considerably because it exposes a structural problem in EHR-based supplement epidemiology: the act of recording any supplement may itself proxy for a distinct patient phenotype—perhaps more engaged, more medically complex, or further along a diagnostic pathway—rather than reflecting the supplement's pharmacological action. Glucosamine had attracted interest partly due to plausible anti-inflammatory mechanisms and prior observational signals suggesting cognitive benefits, making the original EHR finding alarming. This reanalysis suggests that signal was likely an artifact of documentation patterns rather than biological effect.
For adults currently taking glucosamine for joint health, this analysis is reassuring but not exonerative—confounding runs in both directions. Crucially, this is a preprint posted on medRxiv and has not yet undergone peer review, so conclusions remain provisional. Methodologically, the study is a valuable reminder that comparator supplement arms are essential controls in any EHR pharmacoepidemiology analysis. An incremental but important contribution to supplement research methodology.
