To determine whether the increased transcription of immune activation genes in high-lonely individuals might stem from reduced circulating levels of the anti-inflammatory glucocorticoid cortisol, we assayed salivary cortisol concentrations immediately upon waking, 30 minutes later, and at bedtime on three consecutive days during study years 1, 3 and 4. Mixed effect linear model analysis (Time × Loneliness factorial design) indicated no significant reduction in average daily salivary cortisol levels in high-lonely individuals (mean = 4.061 ± 0.417 μg/dl versus 4.541 ± 0.342 for low-lonely; Loneliness main effect, p = 0.2374). However, high-lonely individuals did show less diurnal decline in circulating cortisol levels than did low-lonely participants (Time × Isolation interaction, p = 0.0474; means in Table 2). This difference was driven by slightly higher evening levels of cortisol; change in salivary cortisol concentration from 0 to 0.5 h post-awakening did not differ significantly across groups (p = 0.8610), but 0-16 h and 0.5-16 h declines were both less pronounced in high-lonely individuals (p = 0.0260 and 0.0396, respectively). Blunted diurnal cortisol rhythm was also observed in analyses treating measurement occasion
