Depressive symptoms status and nighttime sleep duration in relation to cognitive performance in Chinese elderly: Insights from a large cross-sectional inquiry.

Cognitive impairment was present in 28.0% of the study sample and depressive symptoms were present in 24.92% of participants.

• Data came from the 2019 HAELAS baseline survey of community-dwelling older Chinese adults aged ≥65 years (n=4145).
• Cognitive impairment was defined as MMSE score <24; 1160 participants (28.0%) met this threshold.
• Depressive symptoms were measured via the 15-item Geriatric Depression Scale (GDS-15); 1033 participants (24.92%) had depressive symptoms.

Each 1-point increase in GDS-15 score was independently associated with a 10.5% higher risk of cognitive impairment.

• Adjusted OR = 1.105 (95% CI: 1.079–1.131, P < 0.001).
• This association was assessed using multivariable logistic regression controlling for relevant covariates.
• The association was independent of nocturnal sleep duration.

The relationship between nocturnal sleep duration and cognitive function followed a U-shaped (non-linear) pattern with an inflection point at approximately 7 hours.

• Non-linear association was identified using restricted cubic spline (RCS) analysis (P < 0.001 for non-linearity).
• Short sleepers (<6 h/night) had a 27.4% higher risk of cognitive impairment compared to normal sleepers (6–7.9 h/night) (OR = 1.274, P < 0.05).
• Long sleepers (≥8 h/night) also showed a significant association with cognitive impairment.
• Normal sleep duration (6–7.9 h/night) was used as the reference category.

All joint combinations of depressive symptoms and sleep duration categories showed higher odds of cognitive impairment compared to the reference group of non-depressed normal sleepers.

• Six subgroups were formed by crossing depression status (depressed vs. non-depressed) with three sleep duration categories (<6 h, 6–7.9 h, ≥8 h/night).
• Reference group was non-depressed participants sleeping 6–7.9 h/night.
• The highest-risk group was Group F (depressed + ≥8 h/night), with OR = 2.62.
• All 6 non-reference subgroups had ORs greater than 1, indicating additive risk across combinations.

For participants sleeping ≤7 hours, extending sleep had a protective total effect on cognition, with 37.91% of this effect mediated indirectly through reduced depressive symptoms and 62.06% attributable to a direct protective effect.

• Mediation analysis used 4-way decomposition stratified by the sleep-cognition U-shaped inflection point of 7 hours.
• For the ≤7 h stratum, longer sleep duration reduced cognitive impairment risk both directly and indirectly via alleviation of depressive symptoms.
• The indirect pathway (through depressive symptoms) accounted for 37.91% of the total protective effect.
• The direct effect of sleep extension accounted for 62.06% of the total protective effect.

For participants sleeping more than 7 hours, longer sleep duration showed a harmful total effect on cognition, driven mainly by a direct effect, with depressive symptoms no longer serving as a meaningful mediator.

• This finding was identified in the >7 h stratum of the 4-way decomposition mediation analysis.
• In this stratum, the harmful effect of longer sleep on cognition was primarily direct rather than mediated through depressive symptoms.
• This contrasts with the ≤7 h stratum where depressive symptoms played a meaningful mediating role.
• The pattern suggests different mechanistic pathways operate below and above the 7-hour inflection point.

Depressive symptoms exerted a harmful effect on cognitive function primarily through a direct pathway, with sleep duration playing a negligible mediating role in both sleep strata.

• In 4-way decomposition analyses stratified by the 7-hour sleep inflection point, sleep duration did not meaningfully mediate the depression-cognition association in either the ≤7 h or >7 h strata.
• This indicates that the pathway from depressive symptoms to cognitive impairment is largely independent of sleep duration.
• This asymmetry in mediation contrasts with the finding that depressive symptoms do partially mediate the sleep-cognition relationship in the short sleep stratum.