Sleep duration as a mediator of the association between caffeine intake and mood symptoms: An intensive longitudinal study of young adults with and without bipolar spectrum disorders

Psychosocial interventions for bipolar spectrum disorders (BSDs) often recommend limiting caffeine intake, yet few studies have examined whether caffeine intake differentially affects mood and whether sleep disruption is a key mechanism underlying these effects. The goals of this study were to investigate concurrent and prospective relationships between caffeine intake, sleep, and mood symptoms among individuals with and without BSD and test whether caffeine intake prospectively predicts mood symptoms via its impact on sleep duration. Participants with and without BSD completed a 20-day ecological momentary assessment protocol, reporting daily caffeine consumption and mood symptoms via smartphone, and wearing wrist actigraphs to objectively measure sleep. Results from multilevel models revealed that on days when individuals consumed more caffeine than usual, they reported lower same-day depressive symptoms and higher same-day hypomanic symptoms, even after accounting for sleep duration. Results from multilevel mediation models indicated that caffeine intake was associated with increased next-day depressive symptoms, and this effect was partially mediated by shorter sleep duration. Caffeine intake also predicted higher next-day hypomanic symptoms indirectly through shorter sleep duration, though the direct effect of caffeine intake on hypomanic symptoms was not significant - consistent with full mediation. Diagnostic status did not moderate any of our findings. These findings suggest that caffeine has dynamic, time-dependent effects on mood, providing short-term mood benefits while contributing to next-day mood disruption through its impact on sleep duration. There was no evidence that caffeine intake has more deleterious mood effects for individuals with BSD relative to those without BSD.