Dose-response association of sleep quality with anxiety symptoms in Chinese rural population: the Henan rural cohort

The participants of the current study were included from the Henan Rural Cohort, which was registered in Chinese Clinical Trial Register (Registration number: ChiCTR-OOC-15006699) and has been previously described in detail [25, 26]. Briefly, villagers aged 18–79 years were recruited from July 2015 to September 2017 by a multistage cluster sampling method from the local general population. Firstly, five rural counties in Henan province (central, south, north, east, and west) were chosen through simple cluster sampling on the basis of the local sufficient population source, support of the masses and local leadership, and medical conditions. Secondly, one to three rural townships of each county were selected by the local Centre for Disease Control and Prevention. Thirdly, the residents who gave a written informed consent were included as the study sample from each village of the selected township. Finally, a total of 39,259 participants (15,490 men and 23,769 women) who signed informed consent were included.

For the current analysis, a total of 29,995 participants completed the evaluation of anxiety symptoms. Furthermore, participants were excluded if they had missing data on PSQI score (n = 269). Because of the impaired sleep quality in shift workers [27, 28] and cancer [29], the participants who had self-reported experience of night shift work (n = 1530) or a history of cancer (n = 285) were further excluded to minimize the confounding bias. Finally, a total of 27,911 subjects were included in the present study.

Ethics approval was provided by the Zhengzhou University Life Science Ethics Committee, and signed informed consent was obtained for each participant. In addition, permission to administer each of the questionnaires, measures, or scales in the current study was obtained from every participant.

Data collection was performed by well-trained investigators in a face-to-face interview adopting a structured questionnaire. Demographic variables of participants included gender, age (continuous variable), marital status (married/cohabitation, other), educational levels (primary school or below, junior high school and senior high school or above), smoking status (non-smoker, or current smoker), alcohol consumption (non-drinker, or current drinker), and personal and family history of diseases.

Physical activity levels were classified into three categories: light, moderate and vigorous referenced to the criterion of the International Physical Activity Questionnaire [30]. Additionally, the anthropometric measurement was conducted on the basis of a standard protocol [31]. Height and weight were measured with individuals wearing light clothes and barefoot to the nearest 0.1 kg and 0.1 cm. Body mass index (BMI) was computed by body weight in kilograms divided by square of height in meters.

Information on sleep was collected by PSQI [32], which consisted of 19 items. The scale that scores 0 to 21 has been widely used to evaluate sleep quality. A previous study reported that at least a cut-off PSQI score of 6 yields a sensitivity of 89.6% and a specificity of 86.5% [32]. Thus, participants with at least 6 PSQI score were considered as having a poor sleep quality in this study. Self-reported night sleep duration was obtained by asking the following question, “What time did you usually go to bed and wake up during the past month?” The sleep onset latency was assessed by the following question: “How long (in minutes) has it taken you to fall asleep each night during the past month?” The sleep initiation time was calculated as bed time plus sleep latency. The night sleep duration was computed on the basis of wake-up time and sleep initiation time [33].

The anxiety symptoms of participants were collected using the two-item generalized anxiety disorder scale (GAD-2) (feeling nervous, anxious, or on edge and not being able to stop or control worrying) yielding a sensitivity of 85% [34]. The scores of this scale ranged from 0 to 6. To examine the association between sleep quality and anxiety symptoms, this study dichotomized scores of the GAD-2 scale. Participants were defined as having anxiety symptoms if they scored ≥3 in the current study [35].

Means ± standard deviations (SD) and frequencies (percentages) were presented for continuous and categorical variables, respectively. Multivariable restricted cubic regression spline curves [36] with 3 knots (5th, 50th, and 95th) were fitted to observe the association between continuous PSQI score and anxiety symptoms. Furthermore, the PSQI score was dichotomized to examine the association between poor sleep quality and anxiety symptoms with good sleep quality as reference group by performing logistic regression models. In the fully adjusted model, the potential confounders were adjusted according to the previous studies [14, 37], including age, gender, physical activity, marital status, smoking status, drinking status, educational levels, average monthly income, BMI, night sleep duration and napping duration. Considering the gender-specific prevalence of sleep quality, or anxiety, we investigated the associations stratified by gender through the full analyses [22, 23, 38]. Additionally, stratified analyses were conducted to examine whether the association between poor sleep quality and anxiety symptoms was potentially modified by age, gender, marital status, smoking status, drinking status, average monthly income, physical activity, BMI, snoring, hypertension, and type 2 diabetes mellitus(T2DM). Considering potential bias resulting from exclusion of participants, we did several sensitivity analyses to identify the association between sleep quality and anxiety symptoms by including subjects with shift working or cancer. A two-tailed P value of less than 0.05 was determined the statistical significance in the current study. All analyses were run on SAS version 9.3 (SAS Institute) and R version 3.5.1.

Table 1 displays the demographic characteristics of participants by anxiety symptoms. A total of 27,911 participants were included in this study. The mean (SD) age was 55.96 (12.22) years; 16,743 (59.99%) subjects were women; the mean (SD) PSQI score was 3.79 (2.73); 6087 (21.81%) individuals were poor sleepers; 1557 (5.58%) subjects have anxiety symptom. Those with anxiety symptoms were more likely to have lower education and income, lower physical activity levels, and have poorer sleep quality. In addition, the difference of demographic characteristics of participants between missing and non-missing information on PSQI score was reanalyzed. The findings implied that there was no difference except age, educational levels, and napping duration between two groups (See supplementary Table 1 in additional file 1). Likewise, the minor differences were found in both men and women. Thus, the missing data might be random and not affect the robustness of the current research.

Figure 1 presents the association between continuous PSQI score and the prevalence of anxiety symptoms. There is an increased likelihood of anxiety symptoms with the elevated PSQI score for crude model in total population. After additional adjustment for age, gender, physical activity, marital status, smoking status, drinking status, educational levels, average monthly income, BMI, night sleep duration, and napping duration, the association appeared to be slightly enhanced and remained to be significant. Similar findings were available in men and women.

Table 2 reports results of sleep quality and anxiety symptoms, with less than 6 scores of PSQI as reference category. Compared to the reference group, poor sleep quality (PSQI ≥6) was associated with a higher possibility of anxiety symptoms [odd ratio (OR): 3.85, 95% confidence interval (CI): 3.42–4.33] in total populations, (OR: 4.60, 95%CI: 3.70–5.72) in men, and (OR: 3.56, 95%CI: 3.10–4.09) in women on multivariable analysis.

The results of the stratified analysis for anxiety symptoms are shown in Fig. 2. Anxiety symptoms was associated with poor sleep quality among aged 60 or older (OR: 4.41, 95% CI: 3.70–5.27), married (OR: 3.99, 95% CI: 3.52–4.52), smokers (OR: 4.73, 95% CI: 3.45–6.48), participants with light level of physical activity (OR: 5.55, 95% CI: 4.47–6.90), obesity (OR: 4.66, 95% CI: 3.42–6.33), snoring (OR: 4.60, 95% CI: 3.75–5.64), T2DM (OR: 4.83, 95% CI: 3.30–7.06) (Fig. 2).

In sensitivity analyses, similar results were observed when we included the participants with shift working (See supplementary Table 2 in additional file 1). Comparing participants with poor sleep, the ORs were 3.86 (95% CI: 3.44–4.33) for total, 4.82 (95% CI: 3.90–5.95) for men, and 3.49(95% CI: 3.04–4.00). When the participants with cancer were included, the results did not change materially (See supplementary Table 3 in additional file 1).