Relationship between depressive disorders and biochemical indicators in adult men and women

There is widespread public health concern about depression, especially in developing countries [1]. It is expected that depression will contribute the most to the burden of disease by 2030 [2]. There are several diseases associated with depression, including suicide, obesity, hypertension and stroke, cardiovascular disease, and Alzheimer’s disease [2‐4]. The main known risk factors for depression are gender (most common in women), low education level, low income, smoking, nicotine dependence symptoms, alcohol consumption, body mass index (BMI), waist circumference, triglycerides, glucose, total cholesterol, blood urea nitrogen (BUN), genetic factors, etc. Several studies have reported a correlation between gender and body mass index and depression [5, 6]. But the correlation between depression and biochemically related indices is not clear, for example, triglyceride levels are strongly associated with depression [7], but not with depression [8]. Therefore, this study aimed to examine whether depressive disorders are significantly associated with biochemical indicators in men and women separately in a large cross-sectional study and to determine whether multiple biochemical markers can discriminate between depressed patients. It may be possible to gain a better understanding of depressive disorders in the US population based on our findings.

The sociodemographic characteristics of the study sample are shown in Table 1. Out of 49,693 participants in NHANES 2009-2018, 17,561 were included in the analysis. The study population included 6857 men and 7931 women in the control group, 1059 men and 1714 women in the depression group (Fig. 1). Age, race, income, education, marital status, BMI, waist circumference, and alcohol abuse were significantly associated with depression in men (Table 2). Drinking status was associated with depression significantly more in women than in men, and the rest was associated with depression significantly in men as well (Table 3).

Table 4 describes the relationship between biochemical indicators of depressive disorders in men. In the crude model, alanineamino transferase (ALT), alkaline phosphatase (AKP), bicarbonate, and chloride were all significantly linked to depression, yet neither model 1 nor model 2 exhibited these same associations. Albumin and total bilirubin (Tbil) levels were lower in the depressed group than in the control group in all models (p <  0.05). Gamma-glutamyl transferase (GGT), glucose, and triglycerides levels were higher in the depressed group than in the control group in all models (p <  0.05).

Table 5 shows the relationship between biochemical indicators of depression in women. Female models 1, and 2 were adjusted for all confounding factors in male models 1, and 2 plus alcohol consumption status. ALT, AKP, aspartate aminotransferase (AST), creatinine, GGT, glucose, iron, lactate dehydrogenase (LDH), Tbil, triglycerides, and globulin were significantly associated with depression in the crude model, but none of these indicators were significantly associated with depression in both model 1 and model 2. Although glucose and depression were strongly linked in models 1 and 2, sodium and depression were not connected even in the crude model. In contrast to the crude model and model 1, model 2 discovered a strong link between bicarbonate levels and depression. ALP, bicarbonate, and sodium levels were higher in the depressed group than in the control group in all models (p <  0.05).

The biochemical indicator variables were included in stepwise logistic regression analysis and the regression coefficients of these biochemical indicators were used to calculate logit equations for assessing depressed male and female patients. In male patients, the logarithm of odds = − 0.112 − 0.042 (Albumin) + 0.002 (GGT) + 0.046(Glucose) − 0.027 (Tbil) (Table 6). In female patients, the logarithm of odds = − 1.629 -0.018 (Albumin) + 0.002 (GGT) + 0.033(Glucose) + 0.003 (AKP) + 0.002(Creatinine) + 0.125(Triglycerides) (Table 7). Based on the results of the Hosmer-Lemeshow test (p = 0.059, Chi-square = 15.032), this computational model was evaluated in women. But this computational model wasn’t evaluated in men (p = 0.017, Chi-square = 18.631). The sensitivity, specificity, and area under the curve (AUC) of these biomarker combinations were calculated separately in depressed patients (Tables 8 and 9; Figs. 2 and 3). The combined AUC was 0.592 (95% CI: 0.57-0.61) in men, indicating that they were more effective than all single markers in identifying depressed patients.

In this large cross-sectional study, men in the depressed group had significantly higher levels of GGT, glucose, and triglycerides. Depressed men had lower albumin and total bilirubin levels than control men. As compared to the control group, women with depression had higher levels of AKP, Bicarbonate, and Sodium. For combined markers in men, the area under the curve was around 59.16%. The area under the curve for combined markers in women was 58.19%.

There has been a conflicting relationship between biochemical indicators and depression found in many studies to date. To provide some strong support, some large cross-sectional studies are lacking.

In the liver, kidney, and pancreas, GGT is primarily observed. Currently, GGT is the most sensitive enzyme indicator for liver diseases and is used to diagnose and monitor hepatobiliary diseases. Huang et al. reported that GGT was higher in NAFLD patients with depression than in patients without depression [12]. A positive association was also found between GGT levels and depression in men.

Kahn reported that fasting blood glucose levels were higher in depressed patients and these levels were significantly associated with depression scores [13]. Our findings are consistent with this. Loss of appetite is a common feature of depression, which can adversely affect blood glucose levels. In depressed patients, this difference may be caused by defects in glucose metabolism in brain regions such as the amygdala, prefrontal cortex, and hippocampus [14, 15].

We found that higher triglycerides was associated with depressive symptoms, and higher triglycerides subjects had higher levels of depression than normal subjects [16]. Triglycerides levels were significantly higher in the depressed group of men. In addition to lowering cholesterol and increasing triglycerides, interleukin-2 inhibits melatonin release, which reduces brain serotonin, resulting in depression and suicidal tendencies [17].

According to Pascoe MC, a low serum albumin level after stroke was associated with long-term depression symptoms in elderly Swedish patients [18]. We also found that serum albumin was negatively associated with depression.

Peng YF found correlations between BUN, fasting blood-glucose (FBG), TBil, and MDD in a Chinese Han population [8]. Bilirubin is an endogenous antioxidant, and low blood bilirubin levels are associated with seasonal depression, according to Shcherbinina MB [19]. Our results are further confirmed by this.

Recently, sodium was shown to modulate oxidative stress and inflammation, alter the autonomic nervous system, and cause innate and adaptive immune dysfunction [20]. .Many studies have shown that high sodium and chloride are directly associated with depression [21]. Women in the depressed group in our study tended to have higher levels of Sodium, which is more consistent with previous studies.

A measure of bone production, bone-specific alkaline phosphatase, was shown by Cizza G to be significantly greater in women with MDD than in controls [22]. Tissue non-specific alkaline phosphatase (TNAP), a globally expressed enzyme, is known for its activity in bone mineralization. Vitamin B6 molecules are calcified and transportable when this enzyme metabolizes phosphate compounds. Hypophosphatemia (HPP) is an uncommon metabolic disorder caused by hereditary loss-of-function mutations in the ALPL gene. In addition to decreased mineralization of bones and teeth, this systemic illness is also associated with anxiety disorders, seizures, and depression [23].

The study has some limitations. First, although though this study had a high sample size, the study group was only composed of Americans, and our findings might not be applicable to other nations due to variations in socio-demographic traits. Our findings also do not suggest a cause-and-effect link because this study was cross-sectional in nature. Finally, health status at the time of blood collection may affect biomarker results, and therefore the effect of certain disease information not obtained during the survey on biomarkers cannot be excluded. Despite these limitations of our results, the statistical results and findings in this study are robust due to the large-scale data.

In this study, a large sample size was analyzed, and a combined marker was constructed for both males and females, and the combined marker had a higher diagnostic value compared to the individual markers. Although similar studies have been conducted previously, the sample size was small, or the data collection was incomplete. This study provides a comprehensive analysis of 17,561 depressed patients from the NHANES database and provides some insight into the lack of laboratory indicators for depression diagnosis.