Background: Stress-related biomarkers have the potential to provide objective measures of whether interventions directed at people with dementia (PWD) and their family caregivers (FCG) are successful. The use of such biomarkers has been limited by logistical barriers to sample collection.

Objective: Explore saliva concentration of steroid hormones in dementia care dyads during a music intervention.

Methods: Consecutive PWD attending a memory evaluation center and their FCG were allocated to either an intervention-with-music or a non-intervention control group. All were living at home. Stress biomarkers, salivary cortisol and dehydroepiandrosterone sulfate (DHEA-S) samples were collected by the PWD and their FCG, in the morning and evening, 5 days a week, for 8 consecutive weeks. Biomarker concentrations of the intervention and the control groups were compared at week 8, in an intention-to-treat approach with adjustment for baseline value.

Results: Twenty-four PWD in the intervention group and 10 in the control group, and their FCG were included in the analyses. The mean number of morning saliva collections was similar in the intervention and the control groups, ranging from 4.3 to 4.9 per participant weekly during the first 7 weeks, declining to 3.3 during week 8. Median log morning cortisol (pg/mL) among caregivers was lower in the intervention group than in the control group (8.09 vs. 8.57, P = 0.0133).

Conclusion: This study demonstrates that music intervention was associated with lower morning saliva cortisol concentrations for FCGs.