The Mediation Effect of Cotinine Between Depression, Inflammation, and Mortality: A National Cohort Study

doi:10.21203/rs.3.rs-5240323/v1

Download PDF

Research Article

The Mediation Effect of Cotinine Between Depression, Inflammation, and Mortality: A National Cohort Study

https://doi.org/10.21203/rs.3.rs-5240323/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Depression is a major mental health issue that affects millions globally. Inflammation is linked to depression, and smoking is prevalent among depressed individuals. Serum cotinine, a nicotine metabolite, may mediate the effects of smoking on inflammation and mortality in depression. This study aims to explore the mediating role of cotinine between depression, inflammation, and all-cause mortality.

Methods

This study utilized data from the National Health and Nutrition Examination Survey (NHANES) collected between 2005 and 2014. A total of 24,937 participants were included after excluding individuals with missing data on depressive symptoms, serum cotinine concentration, and mortality outcomes. Depressive symptoms were assessed using the Patient Health Questionnaire-9 (PHQ-9), and serum cotinine levels were measured by isotope-dilution HPLC-APCI MS/MS. Multivariable logistic regression was used to assess the association between depression and cotinine. Mediation analysis was conducted to evaluate the mediating role of cotinine in the relationship between depression and WBC count, as well as between WBC count and mortality. Cox proportional hazards models were employed to determine the associations between cotinine, WBC count, and all-cause mortality.

Results

The analysis showed that individuals with depression had significantly higher serum cotinine levels (95.615 ng/mL vs. 53.546 ng/mL, P < 0.001) and WBC counts (7.665 vs. 7.203, P < 0.001) compared to those without depression. Multivariable logistic regression revealed that serum cotinine was positively associated with depression (OR = 1.002, 95% CI: 1.002–1.002, P < 0.001). Mediation analysis indicated that cotinine significantly mediated the association between depression and WBC count, accounting for 37.01% of the total effect after adjusting for confounders. Higher cotinine levels were also associated with increased all-cause mortality (HR = 1.889, 95% CI: 1.691–2.111, P < 0.001). Furthermore, cotinine mediated the relationship between WBC count and mortality, contributing to 27.39% of the total effect after adjustment.

Conclusion

This study highlights cotinine's role as a dual mediator in the relationships between depression, inflammation, and all-cause mortality. These findings underscore the need for targeted interventions, including smoking cessation and inflammation management, to improve health outcomes in individuals with depression.

Depression is a common and serious mental illness that affects approximately 280 million people worldwide[1]. Depression not only significantly reduces the quality of life of sufferers, but also increases the risk of suicide, which kills about 700,000 people each year[2]. In recent years, a growing body of research has shown that the immune system and inflammatory response play a key role in the onset and development of depression[3, 4]. Higher levels of inflammatory markers and pro-inflammatory cytokines, such as C-reactive protein (CRP)[5], tumor necrosis factor-alpha (TNF-alpha)[6], interleukin-6 (IL-6)[7], and white blood cell counts (WBC)[8], are often detected in the serum of depressed patients. These inflammatory factors may be involved in the pathophysiologic process of depression by affecting neurotransmitter metabolism[9], neuroplasticity[1, 10], and neuroendocrine function[11]. At the same time, depressed patients are more likely to exhibit smoking behavior[12]. Some studies have shown that the prevalence of smoking is about 50% higher in depressed patients than in non-depressed patients[13]. Long-term exposure to tobacco smoke triggers a chronic inflammatory response in the body, leading to an increased release of pro-inflammatory cytokines and matrix metalloproteinases, further exacerbating the inflammatory state[14]. Serum concentrations of cotinine, a major metabolite of nicotine, can be used as an objective indicator for assessing an individual's level of exposure to tobacco smoke[15]. Although the association between smoking, inflammation, and depression has been extensively studied, the underlying mechanism of whether smoking influences the inflammatory response through cotinine, which in turn increases the risk of death in depressed patients, has not been fully explored. Understanding this mediating effect is important for unraveling the pathological mechanisms of depression and developing effective intervention strategies.

Study Design and Population

Data for this analysis are from the National Health and Nutrition Examination Survey (NHANES), conducted by the National Center for Health Statistics (NCHS), part of the Centers for Disease Control and Prevention (CDC). The sample is nationally representative of the civilian institutionalized population in the US. Informed consent was obtained in writing for all participants in NHANES, and study protocols were approved by the NCHS Ethics Review Board. We downloaded demographic and dietary, examination, laboratory, and questionnaire data from the NHANES website (https://www.cdc.gov/nchs/nhanes/index.htm) for five continuous NHANES cycles from 2005–2014. In total, 50,965 participants were originally interviewed; however, the final analysis included 24,937 participants after excluding those missing PHQ-9 scales(n = 24,769) serum cotinine concentration data (n = 1,235) and all-cause mortality data (n = 24) (Fig. 2).

Evaluation of Depressive Symptoms

Depressive symptoms were assessed using the 9-item Patient Health Questionnaire (PHQ-9). This tool comprises nine questions, each scored from 0 to 3, with higher scores indicating more frequent symptoms. The total score ranges from 0 to 27. In this study, we defined participants with PHQ-9 scores of ≥ 10 as having clinically relevant depression (CRD), based on the criteria established by Kroenke et al. [16]and the classification from Manea et al.[17]

Measurement of Serum Cotinine Concentrations

Serum cotinine was measured using the isotope-dilution high-performance liquid chromatography-atmospheric pressure chemical ionization tandem mass spectrometry (ID HPLC-APCI MS/MS). The measurements were made manually between 2003 and 2012. From 2013 onwards, the whole procedure became automated, and the upgrading of laboratory equipment was changed to API 6500 from the previously used API 4000 mass spectrometer manufactured by SCIEX. However, the basic principle of the measurement technique remained unchanged. For analysis, serum cotinine concentrations were divided into quartiles to divide the groups into different ranges from the lowest to the highest levels of cotinine concentration. The study used quartiles to ensure that there is a sufficient distribution of participants across categories for which clearer comparisons may be allowed while mitigating the potential skewness arising with the use of continuous values.

Determination of Mortality Outcomes

Mortality status of the study population was ascertained by using the NHANES-linked mortality file available publicly and updated through December 31, 2015. Mortality status was matched with the National Death Index (NDI) using a probabilistic matching algorithm. Further, disease-specific deaths were identified according to the International Classification of Diseases, Tenth Revision (ICD-10). Heart diseases (ICD codes 054–064), malignant neoplasms (019–043), and other causes (010) were identified under disease-specific mortality categories.

Assessment of Covariates

NHANES staff had home interviews that could gather demographic information, including individual age, sex, race/ethnicity, educational background, marital status, family income, smoking history, alcoholic consumption, and concomitant diseases such as hypertension or diabetes. Anthropometric measurements were taken in the Mobile Examination Center (MEC): height, weight, waist circumference, and body mass index (BMI, kg/m²). Race/ethnicity was categorized as Mexican American, other Hispanic, non-Hispanic White, non-Hispanic Black, and other. The education level was categorized as less than high school, high school or equivalent, and college or above. Family income was categorized into low (< 1.30), middle (1.30–3.49), and high (≥ 3.50) income based on the poverty income ratio (PIR). Smoking status was determined by the question: "Have you smoked more than 100 cigarettes in your lifetime?" Based on participants' responses, "No" was categorized as "Never smokers", while "Yes" was categorized as "Ever smokers." Alcohol consumption was first dichotomized by the number of drinks a participant consumed ≥ 4 alcoholic drinks per day as "daily, " "occasional, " or "none." Serum specimens for participants during the period of 2005–2014 were collected as part of the NHANES laboratory examination component, and blood collection and analysis were performed under stringent procedures. At baseline, white blood cell count, lymphocyte percentage, and platelet count were measured.

Statistical analyses

All analyses in this study were conducted in accordance with CDC guidelines and utilized the recommended weighting scheme. Variance tests were used for normally distributed data, while the Kruskal-Wallis test was applied for skewed data. Categorical variables were tested using the Chi-square test, and Spearman correlation analysis was used to assess relationships between cotinine levels, white blood cell (WBC) count, and depression scores.

Generalized linear models (GLM) were used to evaluate the associations between depression, inflammatory markers, and cotinine levels. Multivariable logistic regression was employed to calculate the odds ratios (OR) and 95% confidence intervals (CI) for cotinine in relation to depression.In the mediation analysis, two models were constructed: the mediator model was used to evaluate the relationship between depression (exposure variable) and WBC count, with cotinine as the mediator; the outcome model included both the exposure and mediator variables to assess their effect on WBC count. The magnitude of the mediation effect was quantified by deriving the mediation percentage, calculated as the ratio of the indirect effect to the total effect. The significance of the mediation effect was tested using bootstrap resampling (n = 10,000).

Interaction analysis between cotinine and WBC count was conducted using linear regression models, including interaction terms and adjusting for covariates. Generalized additive models (GAM) were employed to detect nonlinear relationships between cotinine levels and all-cause mortality, separately analyzing the depression and non-depression groups. Cox proportional hazards models were used to assess associations between cotinine, WBC count, and all-cause mortality, adjusting for demographic and health factors, and reporting hazard ratios (HR) with 95% CI. The moderating effect of cotinine on the relationship between WBC count and all-cause mortality was evaluated using Cox models, including interaction terms.In the mediation analysis for cotinine as a mediator between WBC count (exposure variable) and all-cause mortality (outcome variable), two models were constructed: the mediator model assessed the relationship between cotinine and WBC count, while the outcome model used Cox regression to evaluate the effect of the exposure and mediator variables on all-cause mortality. The magnitude of the mediation effect was quantified as a mediation percentage, with significance tested using bootstrap resampling (n = 10,000).

All statistical analyses were performed using R (version 4.3.2), with a significance threshold of two-sided P < 0.05. The "survey," "mediation," and "mgcv" packages were used for GLM, mediation analysis, and GAM, respectively.

Baseline characteristics of study participants

We divided participants into a non-depression group (score < 10, n = 22,664) and a depression group (score ≥ 10, n = 2,273) according to their depression scores. Baseline characteristics were compared between two groups (Table 1). Comparing the basic characteristics in the depression group, the percentage of females was significantly higher compared to that in the non-depression group (63.792% vs. 49.294%, P < 0.001). Furthermore, the values in the depression group were greater compared with the non-depression group for body mass index (BMI) (30.532 vs 28.709 kg/m², P < 0.001); pulse rate (75.512 vs 72.633 bpm, P < 0.001); WBC count (7.665 vs 7.203, P < 0.001). In the case of racial distribution, there was a significant difference between the two groups (P < 0.001). However, both groups were dominated by non-Hispanic Whites, but the depression group had a higher number of non-Hispanic Blacks: 22.217% versus 20.751%. The following co-morbid conditions were significantly higher in the depression group (P < 0.001): hypertension, diabetes, coronary heart disease, and stroke. Meanwhile, the depression group showed a significantly higher serum cotinine concentration compared with the non-depression group (95.615 vs. 53.546 ng/mL, P < 0.001), while lymphocyte percentage was lower in the depression group than in the others (29.890% vs. 30.462%, P < 0.001).

Table 1

Baseline Characteristics of Participants by Depression Status
Characteristic	Non-depression	Depression	P-value
Characteristic	< 10(n = 22664)	≥ 10(2273)	P-value
Age,years, M(SD)	47.468 ± 18.971	47.933 ± 17.152	0.128
Gender, n (%)			< 0.001
Male	11492 (50.706%)	823 (36.208%)
Female	11172 (49.294%)	1450 (63.792%)
Race, n (%)			< 0.001
Mexican American	3737 (16.489%)	377 (16.586%)
Other Hispanic	1942 (8.569%)	279 (12.275%)
Non-Hispanic White	10366 (45.738%)	974 (42.851%)
Non-Hispanic Black	4703 (20.751%)	505 (22.217%)
Other Race	1916 (8.454%)	138 (6.071%)
BMI,kg/m², M(SD)	28.709 ± 6.663	30.532 ± 8.236	< 0.001
Pulse,bpm,M(SD)	72.633 ± 12.057	75.512 ± 12.658	< 0.001
SBP, mmHg, M(SD)	123.616 ± 17.923	123.423 ± 19.145	0.263
DBP, mmHg,M(SD)	69.528 ± 12.797	70.382 ± 12.423	0.024
Cigarettes, n (%)	9522 (42.014%)	1290 (56.753%)	< 0.001
Alcohol, n (%)	16471 (72.675%)	1616 (71.095%)	0.108
Hypertension, n (%)	7369 (32.514%)	1042 (45.842%)	< 0.001
Diabetes, n (%)	2852 (12.584%)	484 (21.293%)	< 0.001
Coronary Heart Disease, n (%)	798 (3.521%)	135 (5.939%)	< 0.001
Stroke, n (%)	690 (3.044%)	158 (6.951%)	< 0.001
Cotinine, Serum,ng/mL,M(SD)	53.546 ± 123.773	95.615 ± 150.062	< 0.001
WBC count,10³cells/µL,M(SD)	7.203 ± 2.419	7.665 ± 2.389	< 0.001
Lymphocyte%, M(SD)	30.462 ± 8.624	29.890 ± 8.892	< 0.001
PLT count,10³cells/µL,M(SD)	250.118 ± 66.566	258.980 ± 71.723	< 0.001

Multivariable Analysis of Factors Associated with Depression

In the multivariable logistic regression analysis, serum cotinine (WBC) and lymphocyte percentage were significantly associated with the development of depression after adjustment for sex, age, race, and other confounders (Table 2). Serum cotinine showed a regular positive association with depression both in Model I (OR = 1.002, 95% CI: 1.002–1.003, P < 0.001) and in Model II (OR = 1.002, 95% CI: 1.002–1.002, P < 0.001). WBC count was also significantly associated with depression in Model I (OR = 1.064, 95% CI: 1.047–1.081, P < 0.001) and Model II (OR = 1.032, 95% CI: 1.016–1.048, P = 0.00008). Lymphocyte percentage showed a negative association with depression in Model I (OR = 0.989, 95% CI: 0.984–0.994, P = 0.00003) and Model II (OR = 0.994, 95% CI: 0.988–0.999, P = 0.02252). Nevertheless, platelet count was significantly associated with depression only in Model I (OR = 1.001, 95% CI: 1.000-1.002, P = 0.00191), whereas in Model II (P = 0.07229), it did not reach the statistical significance level.

Table 2

Multivariable OR Analysis for Factors Associated with Depression
Exposure	Multivariable I		Multivariable II
Exposure	OR(95% CI)	P-value	OR(95% CI)	P-value
Cotinine, Serum,ng/mL,M(SD)	1.002 (1.002, 1.003)	< 0.001	1.002 (1.002, 1.002)	< 0.001
WBC count,10³cells/µL,M(SD)	1.064 (1.047, 1.081)	< 0.00001	1.032 (1.016, 1.048)	0.00008
Lymphocyte%, M(SD)	0.989 (0.984, 0.994)	0.00003	0.994 (0.988, 0.999)	0.02252
PLT count,10³cells/µL,M(SD)	1.001 (1.000, 1.002)	0.00191	1.001 (1.000, 1.001)	0.07229
Note: Multivariable I model adjusted for Gender, Age, and Race.Multivariable II model further adjusted for BMI, Hypertension, Diabetes, Coronary Heart Disease, Stroke, Cigarettes, and Alcohol.Results are presented as Odds Ratios (OR) with 95% Confidence Intervals (CI) and P-values.

Cotinine's Mediation Effect Between Depression and WBC

In the mediation effect analysis, the results revealed that cotinine plays a significant mediating role between depression scores and white blood cell (WBC) count. The unadjusted Model 1 (Fig. 3A) reported a direct effect of 0.133 (95% CI: 0.107–0.161, P < 0.001); an indirect effect of 0.038 (95% CI: 0.032–0.044, P < 0.001), while the total effect was 0.170 (95% CI: 0.143–0.198, P < 0.001). The mediation effect accounted for 22.11% of the total effect. In Model 2 (Fig. 3B), after adjusting for the confounding variables of gender, age, race, BMI, alcohol consumption, smoking status, hypertension, diabetes, coronary heart disease, and stroke, the direct effect became 0.053 (95% CI: 0.030–0.079, P < 0.001), while the indirect effect became 0.031(95% CI: 0.025–0.037, P < 0.001), and the total effect was 0.084 (95% CI: 0.060–0.111, P < 0.001). The contribution of the mediation effect to the total effect is 37.01%.

Interaction Analysis of Cotinine and WBC by Depression Status

Cotinine was positively correlated with WBC count in both the depression and non-depression groups(Table 3). In the unadjusted model, the β-value was 0.003 (95% CI: 0.003 to 0.003, P < 0.001) in the non-depressed group and 0.003 (95% CI: 0.003 to 0.004, P < 0.001) in the depressed group, with a P-value for interaction of 0.5522. After adjusting for gender, age, and race, the interaction effect remained non-significant (P > 0.3), indicating that depressive status did not significantly modify the relationship between cotinine and WBC count.

Table 3

Interaction Analysis of Cotinine and White Blood Cell Count by Depression Status
	Non-depression		Depression		P interaction
	β(95%CI)	P-Value	β(95%CI)	P-Value
Crude	0.003 (0.003, 0.003)	< 0.001	0.003 (0.003, 0.004)	< 0.001	0.5522
Model I	0.003 (0.003, 0.004)	< 0.001	0.004 (0.003, 0.004)	< 0.001	0.8367
Model I*	0.004 (0.003, 0.004)	< 0.001	0.003 (0.003, 0.004)	< 0.001	0.4403
Model II	0.003 (0.003, 0.004)	< 0.001	0.004 (0.003, 0.004)	< 0.001	0.3515
Note:Model I: Adjusted for gender, age, and race. Model I*: Adjusted for gender, age, race, and interaction terms for race. Model II: Adjusted for gender, age, race, BMI, hypertension, diabetes, coronary heart disease, stroke, cigarette smoking, and alcohol consumption.

The detection of nonlinear relationships

During a median follow-up of 112 months, a total of 3,118 participants (12.5%) out of 24,937 died, with 797 deaths (25.5%) attributed to suicide. Using a generalized additive model (GAM), after adjusting for factors such as age, gender, race, alcohol consumption, and smoking, the analysis revealed a significant nonlinear relationship between cotinine levels and all-cause mortality(Fig. 4A). In the non-depression group, the smoothing curve had 6.6392 degrees of freedom, χ²=126.847, P < 0.001; in the depression group, the degrees of freedom were 2.6218, χ²=28.18, P < 0.001. Further analysis of the relationship between cotinine levels and suicide mortality (Fig. 4B) showed a significant nonlinear association in the non-depression group (smoothing curve edf = 2.7915, χ²=25.1056, P < 0.001), where suicide mortality initially increased and then decreased with rising cotinine levels. However, in the depression group, the association between cotinine levels and suicide mortality was weaker and not statistically significant (smoothing curve edf = 1.828, χ²=1.0913, P = 0.6149).

All-cause mortality Suicide Mortality

Figure 4 The Relationship Between Cotinine Levels and Predicted Mortality Risk Based on Depression Status

Relationships of cotinine concentration levels with mortality

As it is derived (Table 4) by the Cox proportional hazards regression model, increased levels of cotinine correspond to an increased rate of all-cause mortality, but more among those groups presenting higher levels of cotinine (Q3 and Q4). After adjusting for basic demographic variables (Model I), the risk of mortality increased by 28.1% in the third quartile (Q3) (HR = 1.281, 95% CI: 1.158–1.418, P < 0.001) and by 112.8% in the fourth quartile (Q4) (HR = 2.128, 95% CI: 1.918–2.361, P < 0.001). This association remained significant after further adjusting for health-related factors (Model II), with a 26.1% increase in mortality risk in Q3 (HR = 1.261, 95% CI: 1.139–1.397, P < 0.001) and an 88.9% increase in Q4 (HR = 1.889, 95% CI: 1.691–2.111, P < 0.001).

Table 4

Cotinine Levels and All-Cause Mortality Risk: Hazard Ratios Across Quartiles
Cotinine Quartile	Model I			Model II
Cotinine Quartile	HR	β(95%CI)	P-Value	HR	β(95%CI)	P-Value
Q1	1.0	-	-	1.0	-	-
Q2	0.954	(0.862, 1.054)	0.35380	0.969	(0.876, 1.072)	0.53939
Q3	1.281	(1.158, 1.418)	< 0.00001	1.261	(1.139, 1.397)	< 0.00001
Q4	2.128	(1.918, 2.361)	< 0.00001	1.889	(1.691, 2.111)	< 0.00001

Outcome Variable: Mortality,Exposure Variable: Cotinine quartiles.Adjust I Model: Adjusted for age, gender, and race.Adjust II Model: Adjusted for gender, age, race, BMI, cigarettes, alcohol, hypertension, diabetes, coronary heart disease, stroke, and categorical depression scores.Cox Model Time Variable: person-months of follow-up.

Relationships of WBC levels with mortality

The Cox proportional hazards regression analysis revealed a significant association between white blood cell (WBC) count quartiles and all-cause mortality (Table 5). In Model I (adjusted for gender, age, and race), the mortality risk in the Q4 group was significantly higher compared to the Q1 group, with an HR of 1.540 (95% CI: 1.394-1.700, P < 0.001), while the HR for the Q2 group was 0.900 (95% CI: 0.810-1.000, P = 0.05082). In Model II (adjusted for BMI, hypertension, diabetes, coronary heart disease, stroke, smoking, alcohol consumption, and depression scores), the Q4 group had an HR of 1.386 (95% CI: 1.253–1.534, P < 0.001), and the Q2 group had an HR of 0.892 (95% CI: 0.802–0.991, P = 0.03370).

Table 5

Association Between White Blood Cell Count Quartiles and All-Cause Mortality: Cox Proportional Hazards Analysis
Cotinine Quartile	Model I			Model II
Cotinine Quartile	HR	β(95%CI)	P-Value	HR	β(95%CI)	P-Value
Q1	1.0	-	-	1.0	-	-
Q2	0.900	(0.810, 1.000)	0.05082	0.892	(0.802, 0.991)	0.03370
Q3	1.084	(0.981, 1.198)	0.11341	1.019	(0.921, 1.127)	0.71638
Q4	1.540	(1.394, 1.700)	< 0.00001	1.386	(1.253, 1.534)	< 0.00001

Cotinine's modulating effect on the association between WBC Quartiles and all-cause mortality

The Cox proportional hazards regression analysis revealed a significant association between WBC count quartiles and all-cause mortality, with cotinine levels acting as a moderator of this association (Fig. 5). In the Adjust I model (adjusted for gender, age, and race), mortality risk in the Q4 group of WBC count was significantly elevated, with an overall hazard ratio (HR) of 1.340 (95% CI: 1.211–1.483, P < 0.001). In the cotinine Q2 and Q3 groups, the mortality risk in the WBC Q4 group increased by 51.9% (HR = 1.519, 95% CI: 1.239–1.862, P = 0.006) and 27.3% (HR = 1.273, 95% CI: 1.041–1.557, P = 0.01894), respectively. In contrast, in the cotinine Q4 group, the mortality risk in the WBC Q2 group was significantly lower compared to Q1 (HR = 0.760, 95% CI: 0.594–0.974, P = 0.02989), with an overall HR of 0.874 (95% CI: 0.786–0.971, P = 0.01256). In the Adjust II model (adjusted for BMI, hypertension, diabetes, coronary heart disease, stroke, smoking, alcohol consumption, and depression scores), the mortality risk in the WBC Q4 group remained significantly increased, with an overall HR of 1.250 (95% CI: 1.127–1.386, P = 0.00002). In the cotinine Q2 group, the WBC Q4 group had a mortality risk increase of 28.6% (HR = 1.286, 95% CI: 1.043–1.585, P = 0.01835). Additionally, the WBC Q2 group showed a significantly lower mortality risk in the cotinine Q4 group (HR = 0.739, 95% CI: 0.577–0.947, P = 0.01693), with an overall HR of 0.867 (95% CI: 0.779–0.964, P = 0.00813).

Cotinine's Mediation Between WBC and Mortality

In the mediation effect analysis, the results showed that cotinine significantly mediated the relationship between WBC count and all-cause mortality. In the unadjusted Model 1 (Fig. 6A), the direct effect was − 34.591 (95% CI: -48.869–20.791, P < 0.001), the indirect effect was − 6.544 (95% CI: -9.393–3.847, P < 0.001), and the total effect was − 41.133 (95% CI: -55.474–27.473, P < 0.001), with the mediation effect accounting for 15.81% of the total effect. After adjusting for confounding factors such as gender, age, race, BMI, alcohol consumption, smoking, hypertension, diabetes, coronary heart disease, stroke, and depression, the direct effect in Model 2 (Fig. 6B) was − 84.981 (95% CI: -119.109–56.411, P < 0.001), the indirect effect was − 31.963 (95% CI: -40.662–24.109, P < 0.001), and the total effect was − 116.943 (95% CI: -155.289–87.580, P < 0.001), with the mediation effect accounting for 27.39% of the total effect.

In this study, we analyzed 24,937 participants and found that serum cotinine levels and WBC counts were significantly higher in individuals with depression compared to non-depression individuals, and both were independently associated with the occurrence of depression. Mediation effect analysis demonstrated that cotinine played a significant mediating role between depression scores and WBC counts, as well as between WBC counts and all-cause mortality. Furthermore, higher cotinine levels were strongly correlated with increased all-cause mortality, suggesting that smoking may contribute to the mortality risk in depressed individuals through inflammatory pathways.

Our findings indicate that, in depressed subjects, the level of cotinine in serum was significantly higher compared to the non-depression participants (95.615 ng/mL vs. 53.546 ng/mL, P < 0.001). Such findings are supported by research from various authors; for example, Han et al. have documented that the smoking rate among depressed individuals is more than twice that of the general population[18]. While smoking may serve as a way to deal with and reduce symptoms of depression, it can, in fact, worsen the symptomatology of depression and create a sort of vicious circle[19]. Multivariate logistic regression analysis confirmed this association, even after adjusting for various confounders, with serum cotinine levels remaining significantly associated with depression (OR = 1.002, 95% CI: 1.002–1.002, P < 0.001), suggesting that smoking may play a critical role in both the onset and maintenance of depression[20].

The WBC count in depressed individuals was significantly higher than in non-depressed individuals (7.665 vs. 7.203, P < 0.001), while the percentage of lymphocytes was lower (29.890% vs. 30.462%, P < 0.001). These results support the notion that depression is linked to inflammatory responses, as many studies have indicated that inflammatory markers such as C-reactive protein (CRP) and interleukin-6 (IL-6) are elevated in depressed individuals. Inflammation may contribute to depression through mechanisms involving neurotransmitter metabolism and hypothalamic-pituitary-adrenal (HPA) axis function[21].

Mediation analysis showed that cotinine significantly mediated the relationship between depression scores and WBC counts. After adjusting for confounding factors, the mediation effect of cotinine accounted for 37.01% of the total effect, suggesting that smoking may aggravate depressive symptoms by increasing inflammatory responses. Previous research has shown that smoking induces the production of pro-inflammatory cytokines, such as tumor necrosis factor-alpha (TNF-α) and interleukin-1 beta (IL-1β), which may serve as a bridge between smoking and inflammation in depression[22, 23]. Our interaction analysis found that cotinine and WBC counts were positively correlated in both depressed and non-depressed individuals, with no significant influence of depressive status on this relationship (interaction P-value > 0.3). This indicates that the inflammatory response triggered by smoking may operate through similar mechanisms in both groups.

Over a median follow-up period of 112 months, we found a significant association between high cotinine levels and all-cause mortality. The Cox proportional hazards model showed that the highest quartile of cotinine levels (Q4) had a significantly higher mortality risk compared to the lowest quartile (Q1), even after adjusting for confounders (HR = 1.889, 95% CI: 1.691–2.111, P < 0.001). This finding aligns with previous studies that have identified smoking as a major risk factor for increased all-cause mortality[24]. Additionally, using a generalized additive model, we found a nonlinear relationship between cotinine levels and all-cause mortality, suggesting that the impact of smoking on mortality risk may follow a complex dose-response relationship.High WBC counts were also associated with increased all-cause mortality. The highest quartile of WBC counts had a significantly higher mortality risk than the lowest quartile (HR = 1.250, 95% CI: 1.127–1.386, P = 0.00002). As an inflammatory marker, elevated WBC counts may indicate the presence of chronic inflammation, which is thought to play a crucial role in the development of chronic conditions such as atherosclerosis and diabetes, thereby increasing mortality risk[25, 26].In further mediation analysis, cotinine was found to be mediating the association of WBC count with all-cause mortality. Cotinine accounted for 27.39% of the total effect after adjustment for confounders, with a suggestion that smoking increases mortality risk by enhancing inflammatory responses. This result points out the central role of smoking in inflammation-related mortality risk.

This study first reported that cotinine played a dual mediator role in the links between depression and WBC count and all-cause mortality. The present findings underlined the crucial role of smoking in the development of inflammatory status and further mortality risk in subjects with depression. Full-scale interventions towards depressed persons should cover smoking cessation and control of inflammation to improve long-term health outcomes.

Funding

This study is funded by the

STI 2030—Major Projects (No.2021ZD0200407 [to VV])
The National Natural Science Foundation of China (Grant No. T2250710686 [to VV])
Medical Research Council Senior Clinical Fellowship (Grant No. MR/W020408/1 [to VV])
The National Natural Science Foundation of China (grant: 81973157 and 82173646[to JZ])
The Natural Science Research Project in Minhang District, Shanghai City (2023MHZ046[to YL])
Training Program for High-Level Specialist Physicians under the Collaborative Health Service System of Education, Research, and Clinical Practice in Minhang District, Shanghai City(2024MZYS09)

Author Contribution

Y.L. and L.W. conceived and designed the study. Y.L. and J.X. conducted the data analysis, and J.X. prepared the first draft of the manuscript. J.Z. contributed to the interpretation of the data and provided critical revisions to the manuscript. V.V. supervised the project and provided significant input throughout the study. All authors reviewed and approved the final manuscript.

Kumar PM, Udaya Kumar V, Meenakshi S, Bahekar TN, Narapaka PK, Dhingra S, Murti K: Chap. 1 - Epidemiology and risk factors of mental disorders. In: A Review on Diverse Neurological Disorders. Edited by Bagchi D, Chaurasia RN, Ohia SE: Academic Press; 2024: 3–12.
Pranckeviciene E, Kasperiuniene J: Global Suicide Mortality Rates (2000–2019): Clustering, Themes, and Causes Analyzed through Machine Learning and Bibliographic Data. INT J ENV RES PUB HE 2024, 21(9).
Hassamal S: Chronic stress, neuroinflammation, and depression: an overview of pathophysiological mechanisms and emerging anti-inflammatories. FRONT PSYCHIATRY 2023, 14:1130989.
Chan KL, Poller WC, Swirski FK, Russo SJ: Central regulation of stress-evoked peripheral immune responses. NAT REV NEUROSCI 2023, 24(10):591–604.
Wang JM, Yang KD, Wu SY, Zou XG, Liao YS, Yang B, Xie BN, Huang Y, Li SJ, Ma HJ: Platelet Parameters, C-Reactive Protein, and Depression: An Association Study. INT J GEN MED 2022, 15:243–251.
Liu F, Yang Y, Fan XW, Zhang N, Wang S, Shi YJ, Hu WJ, Wang CX: Impacts of inflammatory cytokines on depression: a cohort study. BMC PSYCHIATRY 2024, 24(1):195.
Mao L, Ren X, Wang X, Tian F: Associations between Autoimmunity and Depression: Serum IL-6 and IL-17 Have Directly Impact on the HAMD Scores in Patients with First-Episode Depressive Disorder. J IMMUNOL RES 2022, 2022:6724881.
Lengvenyte A, Strumila R, Belzeaux R, Aouizerate B, Dubertret C, Haffen E, Llorca PM, Roux P, Polosan M, Schwan R et al: Associations of white blood cell and platelet counts with specific depressive symptom dimensions in patients with bipolar disorder: Analysis of data from the FACE-BD cohort. BRAIN BEHAV IMMUN 2023, 108:176–187.
Cui L, Li S, Wang S, Wu X, Liu Y, Yu W, Wang Y, Tang Y, Xia M, Li B: Major depressive disorder: hypothesis, mechanism, prevention and treatment. SIGNAL TRANSDUCT TAR 2024, 9(1):30.
Kadiyala S, Bhamidipati P, Malla RR: Neuroplasticity: Pathophysiology and Role in Major Depressive Disorder. Crit Rev Oncog 2024, 29(4):19–32.
Troubat R, Barone P, Leman S, Desmidt T, Cressant A, Atanasova B, Brizard B, El HW, Surget A, Belzung C et al: Neuroinflammation and depression: A review. EUR J NEUROSCI 2021, 53(1):151–171.
Wootton RE, Richmond RC, Stuijfzand BG, Lawn RB, Sallis HM, Taylor G, Hemani G, Jones HJ, Zammit S, Davey SG et al: Evidence for causal effects of lifetime smoking on risk for depression and schizophrenia: a Mendelian randomisation study. PSYCHOL MED 2020, 50(14):2435–2443.
Park SK, Oh CM, Kim E, Ryoo JH, Jung JY: The longitudinal analysis for the association between smoking and the risk of depressive symptoms. BMC PSYCHIATRY 2024, 24(1):364.
Zhang W, Lin H, Zou M, Yuan Q, Huang Z, Pan X, Zhang W: Nicotine in Inflammatory Diseases: Anti-Inflammatory and Pro-Inflammatory Effects. FRONT IMMUNOL 2022, 13:826889.
Marques H, Cruz-Vicente P, Rosado T, Barroso M, Passarinha LA, Gallardo E: Recent Developments in the Determination of Biomarkers of Tobacco Smoke Exposure in Biological Specimens: A Review. INT J ENV RES PUB HE 2021, 18(4).
Kroenke K, Spitzer RL, Williams JB, Lowe B: The Patient Health Questionnaire Somatic, Anxiety, and Depressive Symptom Scales: a systematic review. GEN HOSP PSYCHIAT 2010, 32(4):345–359.
Manea L, Gilbody S, McMillan D: Optimal cut-off score for diagnosing depression with the Patient Health Questionnaire (PHQ-9): a meta-analysis. CAN MED ASSOC J 2012, 184(3):E191-E196.
Han B, Volkow ND, Blanco C, Tipperman D, Einstein EB, Compton WM: Trends in Prevalence of Cigarette Smoking Among US Adults With Major Depression or Substance Use Disorders, 2006–2019. JAMA-J AM MED ASSOC 2022, 327(16):1566–1576.
Wu AD, Gao M, Aveyard P, Taylor G: Smoking Cessation and Changes in Anxiety and Depression in Adults With and Without Psychiatric Disorders. JAMA NETW OPEN 2023, 6(5):e2316111.
Park SK, Oh CM, Kim E, Ryoo JH, Jung JY: The longitudinal analysis for the association between smoking and the risk of depressive symptoms. BMC PSYCHIATRY 2024, 24(1):364.
Zhang Y, Yang Y, Li H, Feng Q, Ge W, Xu X: Investigating the Potential Mechanisms and Therapeutic Targets of Inflammatory Cytokines in Post-stroke Depression. MOL NEUROBIOL 2024, 61(1):132–147.
Shyam PSB, Chaya SK, Kumar VS, Mahendra M, Jayaraj BS, Lokesh KS, Ganguly K, Mahesh PA: Inflammatory Biomarkers Interleukin 1 Beta (IL-1beta) and Tumour Necrosis Factor Alpha (TNF-alpha) Are Differentially Elevated in Tobacco Smoke Associated COPD and Biomass Smoke Associated COPD. TOXICS 2021, 9(4).
Goncalves RB, Coletta RD, Silverio KG, Benevides L, Casati MZ, Da SJ, Nociti FJ: Impact of smoking on inflammation: overview of molecular mechanisms. INFLAMM RES 2011, 60(5):409–424.
Thomson B, Emberson J, Lacey B, Lewington S, Peto R, Jemal A, Islami F: Association Between Smoking, Smoking Cessation, and Mortality by Race, Ethnicity, and Sex Among US Adults. JAMA NETW OPEN 2022, 5(10):e2231480.
Yang C, Yang Q, Xie Z, Peng X, Liu H, Xie C: Association of systemic immune-inflammation-index with all-cause and cause-specific mortality among type 2 diabetes: a cohort study base on population. ENDOCRINE 2024, 84(2):399–411.
Wang H, Nie H, Bu G, Tong X, Bai X: Systemic immune-inflammation index (SII) and the risk of all-cause, cardiovascular, and cardio-cerebrovascular mortality in the general population. EUR J MED RES 2023, 28(1):575.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

The Mediation Effect of Cotinine Between Depression, Inflammation, and Mortality: A National Cohort Study

Status:

Version 1

Abstract

Figures

Background

Methods

Study Design and Population

Evaluation of Depressive Symptoms

Measurement of Serum Cotinine Concentrations

Determination of Mortality Outcomes

Assessment of Covariates

Statistical analyses

Results

Baseline characteristics of study participants

Multivariable Analysis of Factors Associated with Depression

Cotinine's Mediation Effect Between Depression and WBC

Interaction Analysis of Cotinine and WBC by Depression Status

The detection of nonlinear relationships

All-cause mortality Suicide Mortality

Relationships of cotinine concentration levels with mortality

Relationships of WBC levels with mortality

Cotinine's modulating effect on the association between WBC Quartiles and all-cause mortality

Cotinine's Mediation Between WBC and Mortality

Discussion

Conclusion

Declarations

Author Contribution

References

Additional Declarations

Status:

Version 1