The validity and feasibility of saliva melatonin assessment in the elderly

Recent work in young and middle-aged subjects suggests that melatonin levels in saliva may represent a viable alternative to serum melatonin measurement. We hypothesized that it may be a valid measure of melatonin levels in older adults as well, but features unique to the elderly may limit its utility. To study this, subjects were admitted to an academic medical center where saliva and serum specimens were collected concurrently in dim light conditions during a 14-hr overnight study period and analyzed for melatonin levels with radioimmunoassays (RIAs). Eighty-five subjects over the age of 65 with a broad range of medical conditions participated in the study. Subjects with dementia, depression and anemia were excluded. We found that saliva volume was inadequate for analysis (<200 mu L) in 23.6% of specimens, with the majority of inadequate volume specimens occurring after midnight and inadequate specimens occurring more frequently in females than in males. The correlation coefficient for saliva melatonin and serum melatonin was r = 0.659 (Spearman, P < 0.001), and r = 0.466 for saliva dim light melatonin onset (DLMO) and serum DLMO. Saliva melatonin levels were 30.9% of serum melatonin levels, with a wide range of ratios noted between subjects. Overall melatonin levels influenced both the correlation and ratio of saliva melatonin to serum melatonin; higher correlations and lower ratios were noted when melatonin levels were high. Saliva specimens provide an economical and practical method for melatonin assessment, however, in older adults, issues such as hyposalivation and low melatonin levels limit the feasibility and validity, respectively, of saliva melatonin.