Depression is a leading cause of disability worldwide and disproportionately affects older adults (1). In China—one of the world’s most rapidly ageing societies—surveillance data indicate that the prevalence of depressive symptoms among older adults increased markedly between 2013 and 2020, underscoring the need for scalable, community-level strategies for early detection and prevention (2). Regional evidence from Guangdong Province likewise shows a substantial prevalence of depressive symptoms among community-dwelling older adults, and comparative syntheses of pre- and mid-pandemic data suggest meaningful temporal variation in estimates (3, 4).

Beyond psychosocial determinants, kidney health has emerged as a biologically plausible correlate of late-life mental health. Meta-analytic evidence indicates that depression is highly prevalent across all stages of chronic kidney disease (CKD) (5). Importantly, large-scale observational cohorts suggest a bidirectional relationship: kidney impairment is associated with an increased risk of subsequent depression requiring treatment, whereas depressive symptoms themselves may precede and predict incident CKD (6–8). Contemporary mechanistic and clinical reviews implicate a uremic milieu, microvascular injury, chronic inflammation, oxidative stress, and broader brain–kidney network dysfunction as converging pathways linking renal impairment to affective disturbances, particularly depression, in ageing populations (9, 10). Extending these established mechanisms, recent advances in kidney–brain crosstalk research highlight uremic toxin–induced neuroinflammation—including aryl hydrocarbon receptor (AhR)-related signaling—as a key driver of cerebral dysfunction, offering a biologically plausible explanation for its role in mood disorders (11, 12).

At the community level, routine biochemical indicators may reflect clinically relevant links between kidney function and mood; however, findings remain inconsistent across studies and populations, underscoring the need for further evaluation in older adults (13). In a nationwide analysis, higher blood urea nitrogen (BUN) levels were associated with an increased risk of depression, with the pattern varying by type 2 diabetes status (14). Conversely, a large ageing cohort reported that lower serum creatinine (Scr)—a biomarker influenced by both renal function and muscle mass—was linked to a higher risk of depression, suggesting possible nonlinearity and population heterogeneity (15, 16). The BUN-to-creatinine ratio (BUN/Scr) has also been related to cognitive function, with depressive symptoms acting as a mediator, implying that mood may lie along the pathway between renal biochemistry and neurocognitive outcomes (17). Additional evidence indicates sex-specific associations when renal function is represented by eGFR dynamics (18). Despite these advances, evidence among Chinese community-dwelling adults aged ≥ 65 years remains scarce and inconsistent, particularly regarding routinely measured biomarkers (Scr, BUN) and their ratio (BUN/Scr), and few studies in this context have formally examined potential dose–response nonlinearity.

Against this background, we analyzed a community-based cohort of adults aged ≥65 years in Guangzhou, China, to examine the associations of Scr, BUN, and BUN/Scr with depressive symptoms assessed using the Patient Health Questionnaire-9 (PHQ-9). Multivariable logistic regression models were employed to estimate these associations, and restricted cubic splines were used to characterize potential dose–response relationships. Subgroup analyses were conducted by sex, age group, body mass index (BMI) category, and histories of hypertension, diabetes, and dyslipidemia. As a sensitivity analysis, Scr was replaced with the creatinine-based estimated glomerular filtration rate (eGFR) to evaluate the robustness of findings across biomarker metrics. Given that Scr and BUN are low-cost and widely available laboratory tests routinely incorporated into health check-ups, elucidating their relationships with late-life depressive symptoms may help inform feasible strategies for risk stratification, early detection, and follow-up in community and primary care settings.

In this community-based sample of adults aged ≥65 years in Guangzhou (n = 1, 994), higher Scr and blood BUN levels were independently associated with increased odds of depressive symptoms after multivariable adjustment, whereas the BUN/Scr ratio showed no clear association. On the original measurement scales, the adjusted odds ratios (ORs) were 1.002 per 1 μmol/L increase in Scr and 1.064 per 1 mmol/L increase in BUN. Restricted cubic spline analyses demonstrated approximately linear associations across the observed ranges; uncertainty widened at the upper extremes where data were sparse, while confidence bands were narrower in regions with denser observations. In a sensitivity analysis substituting Scr with creatinine-based eGFR, higher eGFR was associated with lower odds of depressive symptoms (model 3: OR = 0.992; 95% CI = 0.985–0.999), consistent with the expected inverse relationship between Scr and eGFR, thereby supporting internal consistency.

These findings contribute community-based evidence from China to the broader literature linking kidney health with late-life depressive symptoms. Longitudinal cohort studies have demonstrated that chronic kidney disease (CKD) is associated with an increased risk of incident or treatment-requiring depression, while depressive symptoms themselves may precede and predict accelerated kidney function decline or new-onset CKD (6, 7, 24, 25). Moreover, synthesis studies indicate that depression is highly prevalent among individuals with CKD and is associated with poorer quality of life and greater healthcare utilization (5, 26, 27).

At the biomarker level, the positive association observed for BUN is consistent with nationwide evidence linking elevated urea nitrogen to higher depression risk, with some studies suggesting modification by type 2 diabetes—highlighting the role of metabolic context as a potential source of heterogeneity (14). In contrast, several ageing cohorts have reported an inverse association between serum creatinine and depression risk, often interpreted as reflecting lower muscle mass or nutritional deficiency rather than better renal function per se (15, 28). Variations in body composition, hydration status, dietary patterns, comorbidity profiles (notably diabetes and hypertension), laboratory calibration, and covariate adjustment strategies likely contribute to the divergent findings across studies. Additional evidence indicates that renal function indexed by eGFR may exhibit sex-specific associations with depressive symptoms, and that creatinine- and cystatin C–based estimates can differ substantially, adding further complexity to interpretation (16, 18). From a public health perspective, national and provincial data—including those from Guangdong—continue to document a substantial burden of late-life depressive symptoms, with temporal fluctuations observed around the COVID-19 period (4, 29–33). Utilizing routinely collected biochemical data from community health screenings could therefore support opportunistic depression screening or more targeted case identification in older populations.

Depression was assessed using the Patient Health Questionnaire-9 (PHQ-9), which has been contemporarily validated in Chinese and older adult populations; however, potential age-related differences in reliability and measurement non-invariance warrant cautious interpretation (34–36). Endorsement of somatic symptoms, cognitive status, and cultural expression of mood may influence item responses and, consequently, the estimated associations. Single-time laboratory measurements are also susceptible to short-term biological variability (e.g., hydration status, intercurrent illness, diurnal fluctuations), which can attenuate associations and increase uncertainty. Repeated assessments of both symptoms and biomarkers would help quantify stability and reduce measurement error.

Multiple converging mechanisms may link renal status to affective outcomes in ageing, including a uremic milieu that alters neurotransmission, microvascular injury impairing cerebral perfusion, chronic inflammation and oxidative stress, and broader brain–kidney axis alterations (10, 37). Uremic neurotoxicity can disrupt the blood–brain barrier and compromise neural integrity, with partial reversibility observed after improved toxin clearance or renal replacement therapy (10). Experimental evidence further demonstrates that elevated urea can impair synaptic plasticity and induce depression-like behaviors through carbamylation-mediated suppression of mTOR signalling, while disruptions in urea transport may lead to brain urea accumulation and depressive phenotypes (37, 38). These pathways provide plausible biological explanations for the observed association with BUN and suggest that filtration-independent processes may contribute to mood disturbances in late life. Supporting this, meta-analytic evidence on anti-inflammatory interventions among older adults strengthens the hypothesis of an inflammation-mediated pathway linking renal dysfunction and depression (39).

This study has several methodological and practical strengths. It was conducted in a relatively large, community-based cohort of older Chinese adults, with standardized collection of fasting serum samples and validated assessment of depressive symptoms. The emphasis on routinely available renal biomarkers—Scr, BUN, and the BUN/Scr ratio—adds pragmatic relevance for primary care and community screening, where such indices are already part of routine health check-ups. The prespecified multivariable modelling strategy, which incorporated restricted cubic splines, subgroup analyses, and a substitution-based sensitivity analysis using eGFR, strengthened both the robustness and interpretability of the findings. Moreover, the direction and magnitude of associations remained generally consistent across alternative model specifications and population subgroups (age, sex, BMI, and hypertension), supporting the internal validity of our results and their potential value for risk stratification in late-life depression prevention.

Despite the strengths of this study, several limitations should be acknowledged. First, the cross-sectional design precludes causal inference and warrants caution when interpreting temporal relationships; therefore, the possibility of reverse causation cannot be ruled out. Moreover, our multivariate regression approach, while identifying independent associations, does not allow us to disentangle the potential complex pathways and intermediary mechanisms through which renal markers may influence depressive symptoms. Second, renal biomarkers and PHQ-9 scores were assessed at a single time point, which may introduce short-term biological variability. However, such non-differential measurement error typically biases associations toward the null rather than exaggerating them. Third, although we adjusted for multiple confounders, residual confounding may persist due to unmeasured factors such as dietary intake, hydration status, and socioeconomic conditions. Finally, the study sample was drawn from a single urban community, which may limit the generalizability of the findings, particularly to rural populations or other regions. Despite these limitations, this study provides valuable insights into the association between renal function markers and depressive symptoms, highlighting the potential relevance of these findings for older adults in China.

Clinically, although the observed effect sizes were modest, routine renal indices may serve as practical markers to flag older adults who could benefit from depression screening during community health check-ups where laboratory infrastructure is available (40). A feasible approach would be to integrate PHQ-9 assessment when Scr or BUN values fall outside age-appropriate reference ranges, with priority given to individuals exhibiting multimorbidity (e.g., diabetes, hypertension) or functional decline.

Future research should employ longitudinal designs incorporating repeated assessments of renal biomarkers (Scr, BUN, cystatin C, and uremic toxin panels) alongside inflammatory markers, while also including direct body-composition measures to disentangle the effects of muscle mass from filtration function. To move beyond association and elucidate the complex interplay among variables, advanced analytical frameworks are crucial. First, pathway analysis or structural equation modeling could be employed to delineate the direct and indirect effects of renal dysfunction on depressive symptoms, and to identify key mediating variables (e.g., inflammatory markers, nutritional status, or specific uremic toxins) along the hypothesized biological pathways (41). Second, applying modern causal inference methodologies—such as marginal structural models to account for time-varying confounding, or propensity score–based analyses—to longitudinal or intervention data would substantially strengthen the causal interpretability of findings regarding the impact of renal function on mental health (42, 43). Mechanistic studies are warranted to determine whether alterations in urea-related pathways—such as carbamylation burden or mTOR signalling activity—correspond to depressive trajectories and cognitive outcomes in ageing. From an analytical perspective, evaluating effect modification by diabetes status and sex, adopting flexible non-linear modelling approaches, and performing robustness analyses (e.g., excluding participants with acute kidney events) would further enhance causal inference and interpretability.

Among community-dwelling older adults in Guangzhou, higher Scr and BUN were independently associated with greater odds of depressive symptoms, whereas eGFR—modelled as a substitute for Scr—showed a modest inverse association. These findings highlight a clinically meaningful link between renal function and late-life depressive symptoms, underscoring the need for longitudinal studies to establish temporality and causality. Such evidence could ultimately inform scalable, laboratory-based risk stratification and screening approaches for depression prevention in ageing populations.