Background

Depression is one of the most common mental health disorders, affecting more than 280 million people globally [1]. A recent systematic review and meta-analysis revealed that the 12-month and lifetime prevalence rates of major depressive disorder in China were 1.6% and 1.8%, respectively, and the percentages had been increasing over time [2]. If the population in China is estimated to be 1.426 billion in 2023 [3], the 12-month prevalence of major depressive disorder may reach over 22.8 million of individuals. A longitudinal population study in Australia suggested that the severity of depression is a major predictor for suicidal ideation and suicidal attempt [4]. Based on a recent meta-analysis on 15 studies, the prevalence of suicidal attempt in a lifetime among individuals with major depressive disorder was 3.45 times higher than those without major depressive disorder [

Methods

Study design and population

This study employed a secondary data analysis using the baseline data from the China Kadoorie Biobank (CKB) study. The original CKB study was conducted between 2004 and 2008, which recruited over 0.5 million adults from 10 regions across mainland China. After providing informed written consent, each participant attended a face-to-face interview and a physical examination. A total of 512,681 adults aged between 30 and 79 years (without any major disability) with permanent residence were included in this baseline survey. A standardized electronic questionnaire was used to collect participant information including sociodemographic characteristics, lifestyle habits, exposure to passive smoking and domestic indoor air pollution, medical history, physical activity, and mental health status. The questionnaire used can be accessed via the official website of CKB (https://www.ckbiobank.org/study-resources/survey-data). Each participant’s resting blood pressure (BP) was measured using the A&D digital BP monitor (Model No.: UA-779). A body composition analyzer (Model No.: TBF-300GS) was used to measure body mass index (BMI), while a standing height measuring instrument was used to measure weight and height. BMI is calculated by using this formula: the participant’s weight in kilograms (kg) divided by the square of height (H) in meters (m), (BMI = kg/ H2) [24]. BMI at 28 kg / m2is recommended as the cut-off point for obesity for the Chinese people [25, 33].

Covariates

Adjustment for covariates was performed in this analysis, including sociodemographic characteristics (i.e., age, gender, marital status, education level and annual household income), lifestyle habits (i.e., smoking status, alcohol assumption, and physical activity), health status (i.e., BMI and blood pressure), stressful life events in the past two years, passive smoking, cookstove ventilation, and exposure to solid fuels used for heating. Smoking status was classified into four categories: 1) never smoke, 2) quitted, 3) occasional smoker, and 4) current smoker. Participants were classified as a “regular alcohol drinker” if they reported that they drank alcohol “usually at least once a week.” Otherwise, they were classified as a “non-regular drinker.” Physical activity was estimated as metabolic equivalent task hours per day spent on activities related to occupation, commuting, housework, and non-sedentary leisure-time activities. Exposure to stressful life events (Yes/No) was defined as the occurrence of common major life events in the past two years, such as death of a spouse, marital separation/divorce, traffic accident and major natural disaster. Exposure to passive smoking was assessed by self-report responses to the question related to frequency of secondhand smoking exposure. The variable was categorized into 4 levels (none, > 0 to 2 h/week, > 2 to 12 h/week, > 12 h/week). The cut-off points were conventionally selected based on the tertile points among those who had exposure to passive smoking, with the three exposure categories being anticipated to reflect low, middle and high levels of exposure to passive smoking.

Statistical analysis

All statistical analyses were conducted using the IBM SPSS 25.0 (IBM Corp., Armonk, NY). Data were summarized descriptively using statistics including means, standard deviations, frequencies and percentages. For continuous variables, skewness statistics and normality probability plots were used to assess normality. In this study, the outcome of interest was status of major depressive episode in the past year (Yes/No). The primary exposure of interest was duration of solid fuels used for cooking which was categorized into four levels. Specifically, those participants who had no previous exposure to solid fuels used for cooking or always used clean fuels were categorized as the reference group. The remaining participants were conventionally stratified into three tertiles to characterize low, middle and high levels of exposure with totally four levels for the exposure factor: (i) none, (ii) > 0 to 20 years, (iii) > 20 to ≤ 35 years, (iv) > 35 years. Likewise, the exposure to solid fuels used for heating was categorized into four levels: (i) none, (ii) > 0 to 8.2 years, (iii) > 8.2 to ≤ 13.5 years, (iv) > 13.5 years. The association between major depression in the past year and exposure to solid fuels used for cooking was examined by logistic regression analysis. Unadjusted and adjusted logistic regression analyses were conducted with adjustment for the covariates of sociodemographic characteristics and lifestyle habits, presence of stressful life events in the past two years, presence of cookstove ventilation, exposure to passive smoking, and level of exposure to solid fuels used for heating. As the time scope of the outcome of major depressive episode was the past 12 months from the time of survey, it was possible that some participants might have a major depressive episode prior to exposure to solid fuel usage. A sensitivity analysis was therefore conducted by excluding those participants who had no more than one year of solid fuel usage before the survey. All tests involved were 2-sided at 5% level of significance.

A total of 283,170 participants were included in this secondary data analysis study. Among them, 2,171 participants were classified as having major depressive episode in the past year, and there were totally 91,611 participants without exposure to solid fuels used for cooking and 61,873 to 65,612 participants with different levels of exposure to solid fuels used for cooking. Such a sample size is adequate to detect an odds ratio of having major depressive episode of as small as 1.17 when comparing anyone of the exposure groups with the non-exposure group with over 80% power at 2-sided 5% level of significance.

Results

Characteristics of the study population

Amongst 283,170 participants who were included in the baseline survey of the CKB study, the average age was 51.4 (SD = 10.5) years, and 58.2% of them were female. About 68% of them used solid fuels for cooking, with a 27-year median. More than half of the study sample (67%) had at least some cookstove ventilation. Nearly 23% participants had exposure to passive smoking for more than 12 h per week. A total of 2,171 (0.8%) participants reported major depressive episode in the past year. Characteristics of the study population stratified by levels of exposure to solid fuels used for cooking are shown in Table 1.

Table 1 Characteristics of the study population by level of exposure to solid fuels used for cooking (N = 283,170)
Full size table

Association between household use of solid fuels for cooking and major depressive episode

Based on their duration of exposure to solid fuels used for cooking, participants were categorized into four levels: (i) none, (ii) > 0 to 20 years, (iii) > 20 to ≤ 35 years, (iv) > 35 years. Those participants who had no previous exposure to solid fuels used for cooking or always used clean fuels for cooking were categorized as the reference group (none exposure). The remaining participants were conventionally stratified into three tertiles to characterize low, middle and high levels of exposure. Unadjusted logistic regression analysis showed that an increased level of exposure to solid fuels used for cooking was associated with an increased odds of having a major depressive episode (unadjusted model in Table 2). After adjusting for sociodemographic characteristics, obesity and lifestyle habits, presence of stressful life events, presence of cookstove ventilation, passive smoking exposure, and level of exposure to solid fuels used for heating, the pattern of association between an increased odds of having a major depressive episode and an increased level of exposure was also noted. Participants who had exposure to solid fuels used for cooking for up to 20 years, more than 20 to 35 years, and more than 35 years were 1.09 (95% CI 0.94–1.27), 1.18 (95% CI: 1.01–1.38) and 1.19 (95% CI: 1.01–1.40) times greater odds of having a major depressive episode, respectively, compared with those who had no previous exposure to solid fuel used for cooking or always used clean fuels for cooking (adjusted model 1 in Table 2). A sensitivity analysis was conducted by excluding those participants who had no more than one year of solid fuel usage before the survey, the results were similar to the primary analysis one (adjusted model 2 in Table 2).

Table 2 Risk of major depression episode by level of exposure to solid fuels used for cooking
Full size table

Discussion

Approximately 46% of the population in China used solid fuels as a household energy source, leading to household air pollution; and the proportion was substantially higher in rural areas [13, 23]. In fact, the present study found that 68% of rural residents used solid fuels for cooking. To the best of our knowledge, this is the largest national study to explore the relationship between solid fuel use and depression in rural China. The results revealed an association between household use of solid fuels for cooking and major depression, particularly for those who had used solid fuels for more than 20 years, after controlling for potential confounding covariates, including sociodemographic characteristics, lifestyle habits, health status, presence of stressful life events, presence of cookstove ventilation, passive smoking exposure, and exposure to solid fuels used for heating. Although participants with longer exposure generally associated with an increased odds of having a major depressive episode, the odds ratio of the longest exposure group (> 35 years, OR = 1.19) was unexpectedly similar to the second longest exposure group (> 20 to 35 years, OR = 1.18). A possible explanation may be owing to the fact that people with longer exposure were more likely subject to a competing risk of death, which may diminish the strength of association, particularly in the longest exposure group.

There is a growing body of evidence that solid fuel use is associated with a high risk of depression [22, 23], which is consistent with the current findings. Individuals (N= 8637) with exposure to solid fuel combustion for over 4 years had 1.12 times greater odds of having depressive symptoms [23]. Supported by the following longitudinal survey (N= 7005) [22], individuals using solid fuels in cooking for more than 7 years had 1.36 times greater odds of depression risk than those who always used clean fuels.

This study, together with the aforementioned previous studies, provides evidence on the association between the exposure to solid fuels and the prevalence of depression. However, only limited evidence exists on the mechanisms linking the use of solid fuels for cooking with depression. The incomplete combustion of solid fuels generates various air pollutants including PM, carbon monoxide, sulfur oxides, and polycyclic aromatic hydrocarbons [34, 35]. One possible explanation may be that inhalation of air pollutants can trigger associated oxidative stress, cerebrovascular damage, neuroinflammation, and neurodegenerative pathology, which all might cause or exacerbate the risk of depression [36,37,38]. Animal experience revealed that PM might cause neurotoxicity by inducing microglia activation characterized by the release of TNFα, which damages the olfactory bulb and increases depression risk [39]. Moreover, studies indicated that PM causes elevated levels of cortisol [40], which has been related to the development of depression [41]. Furthermore, domestic cooking with solid fuels could increase the risk of chronic diseases, such as cancer and cardiorespiratory diseases [19, 42], which are strongly associated with depression [43, 44].

In rural China, solid fuels are reported to be the dominant cooking fuel, with biomass and coal accounting for 47.6% [45] and 13.5% [46], respectively. Our study gives valuable insights into the potential hazardous effects of using solid fuels for cooking on mental health. It indicates household solid fuels used for cooking is a critical public health issue and that policy makers must take responsibility to make the needed policy changes. It is necessary to encourage people to switch to cleaner fuels and technologies when cooking to reduce exposure to household air pollution. Moreover, in this study, depressive episode was more prevalent in those without cookstove ventilation. This result is in line with those of the previous studies [22, 47], showing that cooking ventilation may weaken the relationship of cooking with solid fuel and long duration cooking with depressive symptoms, suggesting that improvements in cooking ventilation should be strongly encouraged.

As a remark, although people with longer exposure to solid fuels used for cooking generally associated with an increased odds of having a major depression episode, the odds ratio of the longest exposure group (> 35 years, OR = 1.19) was unexpectedly similar to the second longest exposure group (> 20 to 35 years, OR = 1.18). A possible explanation may be owing to the fact that people with longer exposure were more likely subject to a competing risk of death, which may diminish the strength of association, particularly in the longest exposure group.

Despite the significance of the findings, there are several limitations in this study that may impact the generalisability of this study. First, the cross-sectional study design assesses both outcome of interest and exposure simultaneously. Therefore, it may not be able to establish a cause-and-effect relationship between household solid fuels used for cooking and depression. In addition, self-reported information is prone to recall bias when participants fail to accurately remember an event in the past. Nonetheless, the overestimation or underestimation of association between cause and effect may be resolved through a longitudinal cohort study in which an event may be observed first, followed by the effects. On the other hand, different cooking practices and chemical properties of fuel such as density, volatility and thermal capacity which could affect the indoor air pollution were not examined in the CKB study. Hence, this could result in imprecision of actual exposure to solid fuels used for cooking. Although this study had controlled for potential confounders (e.g., sociodemographic characteristics, obesity status and lifestyle habits, presence of stressful life events, presence of cookstove ventilation and passive smoking exposure), the results might be confounded by other unmeasured covariates. This is because our study was a secondary data analysis where the adjusted analysis was only able to be performed based on existing available variables.

Conclusion

This study demonstrates the significant association between the use of household solid fuels for cooking and the prevalence of depression in rural China; and the longer duration of exposure, the higher odds of having a depressive episode. Further studies are warranted to examine if there is a causal relationship between them. Nevertheless, reducing the use of solid fuels for cooking by promoting the use of clean energy should be encouraged.