Background
As population ages, geriatric depression has become a public health concern. Depression in older people is related not only to poor health-related quality of life [
1], but also to morbidity, mortality, and even suicide [
2]. The Global Burden of Disease 2010 study identified that depressive disorder was a leading cause of years lived with disability (YLDs), and that major depressive disorder accounted for 8.2% of global YLDs [
3]. In China, the meta-analyses showed that the prevalence of depression ranges from ~ 20% to ~ 30% among older adults living in communities, with the pooled prevalence being ~ 23% [
4,
5]. However, depression has been both underdiagnosed and undertreated in primary care settings owing to common co-occurrence with other geriatric conditions in older adults [
6]. Therefore, it is particularly important to identify risk factors related to geriatric depression for possible intervention.
Chronic kidney disease or decreased renal function is increasingly common as people age [
7]. In US, chronic kidney disease, defined as an estimated glomerular filtration rate (eGFR) < 60 mL/min/1.73 m
2, affected nearly one-third of community-dwelling adults aged over 70 years [
8]. However, only a few population-based studies have so far examined the relationship between kidney function and depression in older adults, with mixed results. For instance, a cross-sectional study of Chinese older people aged 70–84 years indicated that eGFR < 60 mL/min/1.73 m
2 was associated with depressive symptoms [
9], whereas the Singapore Longitudinal Aging Study found no association between the serum creatinine-based eGFR and depressive symptoms (defined as the 15-item Geriatric Depression Scale [GDS-15] score ≥ 5) [
10], and the cross-sectional survey of people over 60 years of age with diabetes in US suggested that only very low eGFR (< 29 mL/min/1.73 m
2) was associated with an increased risk of depressive symptoms [
11]. However, the population-based longitudinal studies of the relationship between measures of kidney function and geriatric depressive symptoms in older adults are still lacking, especially among Chinese elderly people living in rural areas.
Therefore, we hypothesize that increased serum cystatin C or impaired kidney function is associated with an increased risk of geriatric depressive symptoms in older people. We sought to test this hypothesis in this population-based cohort study of older adults who were living in a rural community in China.
Results
At baseline, the mean age of the 1440 participants was 68.5 years (SD, 4.9), and 60% were women. Of them, 316 had geriatric depressive symptoms, which resulted in the overall prevalence of 21.9%, with the prevalence being higher in women than in men (24.8% vs. 17.7%,
p = 0.002). Compared with people without geriatric depressive symptoms, those with depressive symptoms were older and had a higher prevalence of chronic health conditions (e.g., hypertension, heart disease, stroke, COPD, cataract, arthritis, and ADL-disability) (Table
1). Furthermore, people with depressive symptoms had a higher level of LDL, serum cysC, and creatinine, but a lower level of HDL, eGFR
cysC, and MMSE score than those without depressive symptoms. The two groups had no significant difference in educational level, BMI, smoking, diabetes, nephritis, and tumor (Table
1).
Table 1
Baseline characteristics of study participants according to depressive symptoms at baseline
Female, n (%) | 864 (60.0) | 650 (75.2) | 214 (24.8) | 0.002 |
Age (years), mean (SD) | 68.5 (4.9) | 68.3 (4.9) | 69.4 (5.1) | 0.001 |
Education (years), mean (SD) | 3.9 (3.4) | 4.0 (3.4) | 3.7 (3.3) | 0.180 |
BMI (kg/m2), mean (SD) | 26.4 (8.1) | 26.5 (8.9) | 26.2 (4.2) | 0.599 |
Current smokinga, n (%) | 206 (14.3) | 162 (14.4) | 44 (13.9) | 0.970 |
Alcohol consumptiona, n (%) | 299 (20.8) | 247 (22.0) | 52 (16.5) | 0.018 |
Hypertension, n (%) | 768 (53.3) | 573 (51.0) | 195 (61.7) | 0.001 |
Diabetes, n (%) | 322 (22.4) | 246 (21.9) | 76 (24.1) | 0.415 |
Heart diseasea, n (%) | 486 (33.8) | 357 (31.8) | 129 (40.8) | 0.009 |
Strokea, n (%) | 124 (8.6) | 86 (7.7) | 38 (12.0) | 0.044 |
Cataract, n (%) | 200 (13.9) | 145 (12.9) | 55 (17.4) | 0.041 |
COPDa, n (%) | 199 (13.8) | 142 (12.6) | 57 (18.0) | 0.029 |
Arthritisa, n (%) | 518 (36.0) | 384 (34.2) | 134 (42.4) | 0.015 |
Nephritisa, n (%) | 69 (4.8) | 52 (4.6) | 17 (5.4) | 0.763 |
Tumor, n (%) | 51 (3.5) | 37 (3.3) | 14 (4.4) | 0.333 |
cysC (mg/L), mean (SD) | 0.99 (0.29) | 0.97 (0.29) | 1.09 (0.26) | < 0.001 |
HDL (mmol/L), mean (SD) | 1.49 (0.44) | 1.50 (0.42) | 1.44 (0.49) | 0.023 |
LDL (mmol/L), mean (SD) | 2.69 (0.85) | 2.67 (0.85) | 2.78 (0.87) | 0.045 |
Creatinine (μmol/L), mean (SD) | 70.5 (21.8) | 69.7 (22.8) | 73.3 (17.7) | 0.010 |
eGFR (ml/min/1.73 m2), mean (SD) | 81.3 (30.6) | 84.2 (31.7) | 71.2 (23.4) | < 0.001 |
MMSE score, mean (SD) | 26.2 (4.5) | 26.4 (4.4) | 25.5 (4.8) | 0.002 |
ADL-disability, n (%) | 32 (2.2) | 12 (1.1) | 20 (6.3) | < 0.001 |
For the baseline cross-sectional relationship, compared to people with serum cysC level of ≤1.00 mg/l, those with levels of 1.01–1.25 and > 1.25 mg/l had an increased likelihood of having geriatric depressive symptoms, with the multiple-adjusted OR (95% CI) being 1.41 (1.01–1.97) and 3.20 (2.32–4.41), respectively (
p for linear trend < 0.001). Impaired kidney function (eGFR
cysC < 60 ml/min/1.73 m
2) was significantly associated with an increased likelihood of having depressive symptoms (multiple-adjusted OR = 2.95; 95% CI 2.22–3.92) (Table
2).
Table 2
Cross-sectional associations of serum cystatin C and impaired kidney function with prevalent geriatric depressive symptoms (n = 1440)
Cystatin C, mg/L |
≤ 1.00 | 768/119 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
1.01–1.25 | 360/79 | 1.43 (1.03–1.98) | 1.40 (1.004–1.95) | 1.41 (1.009–1.97) |
> 1.25 | 312/118 | 3.25 (2.39–4.42) | 3.30 (2.40–4.50) | 3.20 (2.32–4.41) |
P for trend | | < 0.001 | < 0.001 | < 0.001 |
Impaired kidney function |
No | 1070/176 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
Yes | 370/140 | 3.07 (2.34–4.03) | 3.03 (2.30–4.01) | 2.95 (2.22–3.92) |
At an average 3.6 years (SD, 0.62) of follow-up, 157 out of the 669 persons who had no depressive symptoms at baseline were ascertained to have incident geriatric depressive symptoms. People who developed incident depressive symptoms at the follow-up were more likely to be female and to have lower education than those who did not. Cox regression analysis suggested that having an elevated serum cysC level (> 1.25 vs. < 1.00 mg/L) was significantly associated with an increased HR of having incident depressive symptoms, even in model 3 when controlling for multiple potential confounders (HR = 2.16, 95% CI 1.43–3.27) (
p for linear trend < 0.001) (Table
3). Impaired kidney function at baseline was significantly associated with a 54% increased risk of incident geriatric depressive symptoms detected at the follow-up assessment (multiple-adjusted HR 1.54; 95% CI 1.03–2.30) (Table
3).
Table 3
Longitudinal associations of serum cystatin C and impaired kidney function with incident geriatric depressive symptoms (n = 669)
Cystatin C, mg/L |
≤ 1.00 | 388/78 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
1.01–1.25 | 175/40 | 1.11 (0.75–1.65) | 1.22 (0.82–1.81) | 1.21 (0.81–1.80) |
> 1.25 | 106/39 | 2.20 (1.48–3.27) | 2.23 (1.48–3.34) | 2.16 (1.43–3.27) |
p for linear trend | | 0.001 | 0.001 | 0.001 |
Impaired kidney function |
No | 544/122 | 1.00 (reference) | 1.00 (reference) | 1.00 (reference) |
Yes | 125/35 | 1.57 (1.06–2.32) | 1.57 (1.05–2.34) | 1.54 (1.03–2.30) |
Discussion
In this community-based cohort study, we found that geriatric depressive symptoms affected more than one-fifth of Chinese older adults who were living in a rural area. In addition, both the cross-sectional and longitudinal data suggested that higher levels of serum cysC were associated with an increased risk of having geriatric depressive symptoms in a dose-response manner, even after adjusting for multiple potential confounders, including sociodemographic factors, lifestyles, cardiovascular disorders, cognitive function, and physical disability. Finally, impaired kidney function, when assessed based on eGFRcysC, was associated with both prevalent and incident geriatric depressive symptoms, independent of multiple confounders. This study suggests that high serum cysC levels and impaired kidney function may be risk factors for depressive symptoms in geriatric populations.
The relationship between renal function and depression among older adults has rarely been explored so far in the general population settings. The cross-sectional data from the Maastricht Study of older adults indicated that albuminuria was associated with incidence of depressive symptoms, whereas the reduced eGFR based on creatinine and cystatin C (eGFR
Cr-cysC) was not associated with minor or major depressive episodes [
24]. The follow-up data from the US Health, Aging and Body Composition Study of community-dwelling older adults suggested that an elevated serum cysC level was associated with a 2-fold increased risk of depression, but impaired renal function, assessed based on eGFR
Cr-cysC < 60 ml/min/1.73 m
2, was not related to depression [
25]. Of note, some of the previous studies that do not show an association between reduced eGFR and depressive symptoms in older adults have used either serum creatinine or a combination of serum creatinine and cysC to assess eGFR [
10,
24,
25]. Because creatinine excretion is dependent on age, muscle mass, and nutritional status, especially among older adults, serum cysC has been considered to be more accurate and more sensitive in assessing renal function [
26,
27]. Indeed, we did not find any association between serum creatinine and depressive symptoms (data not shown). When estimating the eGFR based on serum cysC, we found that reduced eGFR
cysC or impaired kidney function was associated both cross-sectionally and longitudinally with an increased risk of having geriatric depressive symptoms among Chinese older adults. This is consistent with a cross-sectional study among Chinese older people (age 70–84 years), in which impaired kidney function was associated with a 1.71-fold increased likelihood of having depressive symptoms (GDS-15 score ≥ 5) [
9].
The biological and pathological mechanisms linking serum cysC and impaired kidney function with geriatric depressive symptoms are not fully understood, but multiple pathways are supposed to be involved. First, impaired kidney function or reduced eGFR
cysC due to glomerular small vessel disease has been correlated with subclinical cerebral microvascular disease in older adults [
28,
29]. Further, cerebral microvascular dysfunction and diseases in brain regions involved in mood regulation were associated with late-life depression (“vascular depression”) [
30,
31]. Thus, impaired kidney function or reduced eGFR might be linked with late-life depressive symptoms through cerebral small vessel disease. Second, serum cysC, as a cysteine protease inhibitor, affects the migration of neutrophils and involves the inflammatory process [
32,
33], which may impair function of the brain-serotonin system and stimulate the activation of the hypothalamus-pituitary-adrenal axis to cause depressive symptoms through inflammation pathway [
34]. In addition, the population-based studies of older adults have linked inflammatory markers (e.g., C-reactive protein, IL-6, and TNF-α receptor 1) with elevated serum cysC or impaired kidney function defined by eGFR
cysC [
35,
36]. The meta-analysis of population-based studies also supported an association of several inflammatory markers with depression in older adults [
37], suggesting the involvement of inflammation in the development of depression. Thus, inflammatory mechanisms may mediate the relationship between serum cysC and depressive symptoms. Of note, our data showed that the association of high serum cysC and impaired kidney function with depressive symptoms was present independent of the inflammation-related diseases such as atherosclerotic disorders, arthritis, and cognitive and functional impairment, suggesting that additional pathways may be involved in linking high serum cysC with geriatric depressive symptoms.
Our cohort study involved both cross-sectional and longitudinal data of older adults who were living in a rural area in China. Moreover, we collected comprehensive data (e.g., demographics, lifestyles, health history, and cognitive and physical functioning) following a standard approach. Thus, we were able to control for a broad range of potential confounding factors. However, our study also has limitations. First, we used GDS-15 to assess geriatric depressive symptoms instead of a clinical diagnosis of depression, although the GDS-15 cut-off score ≥ 5, as a widely screening instrument for major depressive disorders in older adults, did show high sensitivity and specificity in geriatric populations [
10,
21,
22]. Second, we used the cysC-based eGFR as an approximation of glomerular filtration rate instead of a direct measurement, although eGFR has been widely used to define kidney function. Finally, the study population was derived from a single rural community in Eastern China, where people had relatively low education and low socioeconomic position. Thus, caution is needed when generalizing our results to other populations.
Acknowledgements
The authors would like to thank all the study participants for their contribution to the Confucius Hometown Aging Project.