Introduction

Stroke is the primary source of death in China, and the burden has increased significantly between 1990 and 2010 resulting from an increasing the absolute number of people who suffered a stroke, the number of related deaths, and disability-adjusted life-years lost [1]. A recent review systematically explored the stroke prevalence and showed that the overall stroke prevalence was 5.0% in central China for those ≥ 40 years of age, with a higher incidence in men than women, and a significant increase in prevalence, especially among those aged 40–59 years [2]. Stroke, because of its suddenness and severity, places a heavy burden on individuals, families, and society. Therefore, the study of stroke risk factors and effective control may be a vital way to prevent stroke.

A systemic analysis of worldwide stroke burden and risk factors in 188 countries indicated that the risk factors for stroke include low physical activity, poor diet, smoking, high BMI, high SBP, high total cholesterol, and high fasting plasma glucose air pollution exposure [3]. Furthermore, previous studies indicated that getting more or less sleep than recommended and daytime napping increases stroke risk [4]. However, discussions on the risk of stroke have been inconsistent, with some research suggesting that shorter and prolonged sleep duration increases stroke risk [5], some showing that only longer or shorter sleep is associated with stroke [6], or neither was associated with stroke [7].

Insufficient sleep quantity and worse sleep quality are commonly seen in contemporary society and are related to many diseases, especially hypertension, depressive symptoms, type 2 diabetes (T2D), and stroke. A large community-based study showed that daytime napping with a high frequency (≥ 5 times/week) or long duration (> 30 min) increase the risk of stroke [8]. However, few studies have explored the effect of napping combined with night sleep duration on stroke. These studies mostly focused on the urban people and mainly investigated special groups, such as the elderly [9]. There are few studies that have paid attention to the rural areas, which make up over 40% of Chinese people [10]. Furthermore, there is no doubt that napping or night sleep duration is linked to the prevalence of stroke, but few studies have shown the relationship of 24 h of total 24-h sleep on stroke risk [11].

Exploring the connection of sleep duration and stroke may provide assistance in proposing appropriate sleep duration to reduce the incidence of stroke and provide effective sleep recommendations for individuals who aspire to a healthy lifestyle. Therefore, the objective of this study was to evaluate the relationship of napping, nocturnal sleep and 24-h sleep duration on stroke, as well as the combined effect of napping and nocturnal sleep on stroke. The study hypothesized that inappropriate sleep duration and stroke have a significant effect on a rural Chinese population.

Materials and methods

Study design and participants

Subjects were derived from a prospective cohort research study on chronic non-communicable diseases conducted in Henan Province, which included 39,259 adults (18–79 years old). Subjects were excluded if they were incapable of answering the questionnaire or reporting to the survey site because of a serious physical or mental illness. All participants fasted for at least 8 h before coming to the health screening center. Blood and urine sample collection, anthropometric measurements, clinical examination, and self-report questionnaires (general demographic characteristics, lifestyle, sleep status, personal history of illness, and family history of illness) were administered by trained physicians, nurses, and technicians. A baseline study was performed during July 2015 to September 2017 and the follow-up surveys were carried out every 3 years starting in 2018. The present study only included the data from the baseline survey. The survey details have been presented before [12]. In this survey, 37,341 participants were enrolled in the further analysis, after excluding individuals without nighttime sleep duration (n = 36), without daytime nap duration (n = 6), individuals who were shift workers (n = 1549), or suffered cancer (n = 327). This study complied with the principles which were summarized in the Helsinki Declaration. The Life Sciences Ethics Committee of Zhengzhou University endorsed this study, and every individual gave written informed consent.

Sleep duration

Sleep duration was collected using the Pittsburgh Sleep Quality Index (PSQI) [13]. The daytime nap time came from the question: “Have you napped during the middle of the day in the past 30 days?” The sleep duration at night was obtained from this question: “How many sleep hours did you actually get each night in the past 30 days?” Those who answered yes were then asked: “How many times a week do you nap (frequency) and how long (min) is the nap duration.” We used the following formula to calculate the average daily nap time: frequency (times/week) × duration (minutes/ per time)/7. The length of nighttime sleep plus the length of the nap gave the total length of 24 h of sleep. According to the existing epidemiological studies [14], nocturnal sleep duration was sorted by < 6 h, 6 ≤ nocturnal sleep duration < 7 h, 7 ≤ nocturnal sleep duration < 8 h (reference), 8 ≤ nocturnal sleep duration < 9 h, and nocturnal sleep duration ≥ 9 h. The daytime napping duration was categorized as 0 min (reference), 1 ≤ nap duration < 30 min, 30 ≤ nap duration < 60 min, 60 ≤ nap duration < 90 min, and nap duration ≥ 90 min. Furthermore, 24-h total sleep duration was classified into 6 groups: < 6 h, 6 ≤ total sleep duration < 7 h, 7 ≤ total sleep duration < 8 h, 8 ≤ total sleep duration < 9 h, 9 ≤ total sleep duration < 10 h, and total sleep duration ≥ 10 h.

Definition of stroke

Each participant was enrolled in the New Rural Cooperative Medical System (NRCMS), in which every individual has a uniquely identifiable ID card and health insurance card number for disease tracking purposes. The participant was defined as having had a stroke if there was a prior physician self-reported stroke diagnosis. The village physician then confirmed the NRCMS stroke history review and was identified by an endocrinologist, a cardiologist, an epidemiologist, and a physician according to standardization approved by the criteria of the World Health Organization [15, 16].

Covariates

In the current study, the information on demographic characteristics, lifestyles, and the history of the disease was acquired by standard questionnaires administered by well trained interviewers. The demographic characteristics included age, gender, per capita income per month (< 500, 500 ~ , and 1000 ~ RMB), marital status (married/living together, separated/widowed/unmarried), and education (elementary school or below, Junior High School, Senior High School or above). Lifestyles included a high-fat diet (≥ 75 g of meat per day), more vegetables and fruits intake (≥ 500-g vegetables and fruits per day), drinker (≥ 12 times in the past year), smoker (≥ 1 cigarettes per day for over 6 months), and physical activity (low, medium, and high levels) [17]. Body mass index (BMI) was calculated by dividing weight in kilograms by the square of height (kg/m2) [18].

Statistical analysis

Categorical and continuous variables are shown as proportions and mean ± standard deviation, respectively. Categorical and continuous and covariates between those suffering from and not suffering from stroke were respectively compared using the chi-square test and Student’s t-test. Restricted cubic curve uses three nodes (5th, 50th, and 95th centiles) to investigate the correlation of napping, night sleep, and total 24-h sleep duration with stroke [19]. To examine the influence of naps, night sleep, and 24-h total sleep duration on stroke, the binary logistic regression model was developed with nap duration/nighttime sleep duration/24-h total sleep duration as the independent variable. Three models were established: the first model adjusting for no variables. The second model adjusted for age and gender. The third model was further adjusted for education, per capita monthly income, marital status, fatty diet, greater vegetables and fruits intake, current smoker, current drinker, BMI, physical activity, nocturnal sleep duration, nap duration, family history of stroke, hypertension, and type 2 diabetes mellitus. Furthermore, analyses stratified by sex were conducted throughout the study to examine the effect of gender on the correlation.

All statistical analysis was performed using SPSS V21. Two-tailed P < 0.05 was regarded as statistically significant.

Results

Characteristics of participants

The subjects’ demographic characteristics and health behaviors are displayed in Table 1. The final sample of this study was 37,341 individuals (14,485 men), including 2600 individuals who were diagnosed as having a stroke. Compared with subjects without stroke, those with stroke tended to be older and had lower per capita monthly income, education level, physical activity level, and higher BMI. Additionally, participants with stroke tended to be unmarried/widowed/divorced, with longer night sleep duration, longer nap duration, and more extreme 24-h sleep duration than participants without stroke. The average 24-h sleep duration was 9.2 h and 8.8 h in men and women, respectively. Similarly, the average napping duration was 69.9 min and 58.2 min in men and women, respectively, revealing that men had longer napping duration and longer 24-h total sleep duration than women.

Table 1 The characteristics of participants according to stroke by gender
Full size table

Relationship of naps/night/24-h sleep duration and stroke

The ORs and 95% CI of stroke with naps moreover 90 min were 1.57 (1.39–1.76) in the unadjusted model compared with non-nappers. In model 3, the ORs (95%CI) for participants taking a nap longer than 90 min were 1.28 (1.14–1.45) for total, 1.37 (1.13–1.65) for men, and 1.23 (1.04–1.45) for women. Each 30 min increase in nap time was related to 6.1% (3.1–9.2%) for all participants, 7.7% (3.1–12.6%) for men, 5.0% (1.0–9.2%) for women in the model with fully adjusted (Table 2).

Table 2 Odds ratios of stroke according to napping duration
Full size table

Participants who had a short nighttime sleep (< 6 h/day) had an increased risk of stroke compared to subjects who slept 7–8 h/day for all and for men, but not for women. The ORs (95%CI) for less than 6 h at night were 1.26 (1.09–1.46) for all, 1.34 (1.06–1.69) for men, and 1.18 (0.97–1.43) for women, after adjusted fully confounding factors, which showed a slightly stronger relationship for men than for women. Prolonged night sleep duration (≥ 9 h/day) also raised the prevalence of stroke in those men who slept only 7–8 h (OR = 1.30, 95%CI: 1.06–1.59). In addition, the corresponding adjusted ORs for stroke in men were 1.15 (1.05–1.26) and 1.12 (0.99–1.26) for each hour of increase and decrease in nighttime sleep compared with the reference values (Table 3).

Table 3 Odds ratios of stroke according to night sleep duration
Full size table

Subjects with a long 24-h total sleep duration (≥ 10 h/day) increased their stroke risk (OR = 2.12, 95%CI: 1.76–2.55) compared to those with 7–8 h of sleep among total participants in the unadjusted model. After layering by gender, the long total sleep time was significantly correlated with stroke (OR = 1.40, 95%CI: 1.15–1.71) in men in model 3. Furthermore, subjects with short total sleep duration (< 6 h/day) showed a similar association with stroke. The ORs of 24-h total sleep duration per 1-h increase and decrease were 1.06 (1.02–1.10) and 1.09 (0.92–1.30) for stroke in men (Table 4).

Table 4 Odds ratios of stroke according to 24-h total sleep duration
Full size table

Additionally, restrictive cubic splines demonstrated that there was a significantly increased risk of stroke with increased daytime napping time. Additionally, a U-shaped trend was found for stroke risk and the duration of nighttime sleep. A similar association was also found in the 24-h total sleep duration and stroke (Fig. 1).

Fig. 1
figure 1

Associations between napping duration, night sleep duration, total sleep duration and stroke among total, men and women, when adjusted for age, gender (only for total participants), education levels, average income per month, marital status, more vegetables and fruits intake, high-fat diet, current smoker, current drinker, physical activity, body mass index, family history of stroke, hypertension and type 2 diabetes mellitus, napping duration, and night sleep duration

Full size image

Joint effects of naps and nocturnal sleep duration for stroke

To examine the combined effects of daytime nap and nighttime sleep duration, a heat map was used to visualize how the relationship of nighttime sleep duration with stroke risk changed as nap duration increased. In the graph, there is a higher risk of stroke the darker the color of the pattern. Figure 2 shows that with a prolonged nap duration and longer or shorter night sleep duration, the darker the squares, that is, the higher the risk of stroke. In addition, in participants’ nighttime sleep < 6 h, the adjusted OR tended to decrease as napping increased and the OR for stroke was the lowest in the napping duration > 90 min. To be specific, compared with non-nappers with 7–8 h of nighttime sleep duration, participants who had nocturnal sleep duration < 6 h and napping duration ≥ 30 min had a higher prevalence of stroke in men (OR: 1.85, 95%CI: 1.20–2.86 for men). Subjects with over 90-min napping duration and over 8-h nighttime sleep duration also appeared to have greater odds of stroke, especially in men (OR: 1.59, 95%CI: 1.13–2.25 for those with 8 ~ 9-h nocturnal sleep duration and OR: 1.69, 95%CI: 1.16–2.47 for those with over 9-h nocturnal sleep duration). Among women, similar trends were also observed, but were not statistically significant. However, participants with over 90-min napping duration and less than 6-h night sleep duration had a lower risk of stroke, the ORs (95%CI) was 1.21 (0.86–1.71) for all, 1.30 (0.78–2.17) for men, and 1.13 (0.70–1.80) for women, compared with participants who took 31–90 min of naps, although not statistically significant.

Fig. 2
figure 2

Combined effect between night sleep duration and napping duration on stroke among total, men, and women. The night sleep duration was divided into < 6, 6 ~ , 7 ~ , 8 ~ , and ≥ 9, and napping duration was separated into 0, 1 ~ , 30 ~ , 60 ~ , and ≥ 90. Take non-nappers with 7–8 h of night sleep duration as the reference. The data in each square represent the effect size of each outcome (odds ratios (95% CI) for stroke). Covariates: age, gender (only for total participants), education levels, average income per month, marital status, more vegetables and fruits intake, high-fat diet, current smoker, current drinker, physical activity, body mass index, family history of stroke, hypertension and type 2 diabetes mellitus, napping duration, and night sleep duration

Full size image

Discussion

In this study, it was shown that long naps (≥ 90 min) and longer (≥ 9 h) or shorter (< 6 h) nocturnal sleep duration, as well as extreme 24-h total sleep duration (≥ 10 h/day) increased the risk of stroke in people living in rural China. After stratification by gender, both men and women increased the risk of stroke when prolonging the napping duration, but long or short nocturnal sleep duration and 24-h total sleep duration were linked to higher odds of stroke only in men. Furthermore, there exists a combined effect between nap duration and nighttime sleep duration on stroke. The longer the nap time in people with adequate or excessive nighttime sleep (≥ 6 h), the greater the risk of stroke, the longer the nap time in people with short nighttime sleep (< 6 h), the lower the risk of stroke. The current study may help to better identify significant risk variables regarding sleep.

Our results indicated a U-shaped effect of nocturnal sleep duration and stroke. A follow-up study discovered that the duration of sleep (< 6 h or > 8 h) independently raised the risk of stroke and that sleep quality and sleep duration had an interactive influence on stroke [20]. This study has shown similar results to our study. Furthermore, another study found that shorter and longer duration of nighttime sleep were both associated with stroke in older individuals among participants who reported poor physical health [21]. Other studies have come to similar conclusions as this study that both shorter nighttime sleep and longer nighttime sleep are linked to stroke risk [22, 23]. The present study also found that long naps were correlated with stroke. A study based on a large community reported that daytime napping duration with a high frequency (≥ 5 times/week) or long duration (> 30 min) increases stroke risk [8]. Similarly, a prospective cohort study concluded that napping habits were related to cardiovascular disorders risk [24]. In addition, the present study provided important evidence of the coeffect of prolonged napping and prolonged nocturnal sleep on stroke. Yang et al. also reported that prolonged nighttime sleep and prolonged napping were separately and jointly linked to the prevalence of the cardiovascular disease [25]. A prospective cohort study similarly concluded that sleep duration ≥ 9 h at night and naps > 90 min were jointly correlated with a higher risk of stroke [14]. This study found that nap may moderate the risk of stroke from short nighttime sleep. A previous study from 21 countries that included 116,632 study subjects found similar findings to ours, which found that daytime naps were related to cardiovascular events risk and death, except for those who slept less at night, suggesting that napping may be a compensatory mechanism when sleep deprivation occurs at night [26].

Although emerging research has established the link of sleep duration and the prevalence of stroke, a lot of conclusions remain controversial. Research from the China health and retirement longitudinal study of 4785 Chinese over the age of 65 found that short sleep duration per 24 h significantly increased the risk of stroke, which is in line with the current study. But, the research also reported that napping was not associated with stroke [27]. This could be because the study’s participants were older, so the connection between naps and stroke was not significant. A Mendelian randomization study found sleep duration is not significantly associated with stroke [28]. This may be due to residual confounding factors (obstructive sleep apnea or obesity) that may bias the results of the observational studies [29]. There is evidence that long night sleep rather than a short night sleep is linked to the risk of stroke [30, 31]. The source of this discrepancy may be due to differences in the study sample.

Several biological mechanisms related to sleep may explain the relationship between shorter or longer sleep duration with stroke. Potential risk factors for stroke are hypertension, diabetes, smoking, hyperlipidemia, and physical inactivity [32]. A study conducted in 1999 first showed that participants had a 24% decrease in insulin sensitivity after restricting sleep to 4 h per night for five nights [33]. Additional mechanisms of decreased insulin sensitivity after sleep interruption include increased sympathetic activity in the autonomic nervous system. A rapid increase in blood pressure and heart rate has been observed in the morning as well as after napping [34]. Similarly, plasma catecholamine concentration also increases rapidly after waking up [35]. The sympathetic nervous system has an important influence on the pathogenesis of hypertension [36]. Long sleep duration and daytime napping may lead to altered activity of the hypothalamic–pituitary–adrenal system, which increases the risk of adverse vascular events [37].

A novel finding of this study is that there is a joint effect of daytime nap time and nighttime sleep time on stroke. In addition, the relationship of 24-h total sleep duration on the stroke risk was also explored in this study, rarely reported in prior studies. A strength of the study is that it included participants who were recruited from rural areas which tended to suffer from inadequate medical resources. Furthermore, various potential confounding factors were adjusted for.

Several limitations exist in the current study. First, sleep conditions were self-reported by participants, with no objective measurement and a potential lack of accuracy. However, participants received adequate training on their sleep habits, and the study also included comprehensive information on sleep status, which may have led to more accurate results. Second, this was a cross-sectional study, and because stroke was self-reported by the study subjects, some of the study subjects may have changed their sleep duration due to stroke, that is, stroke may have increased or decreased the need for sleep or the opportunity to nap, so this study design could not establish causality. However, some cross-sectional surveys and prospective studies have obtained similar results [14, 29, 38]. Third, since we only counted the period after lunch as nap time, our study may have underestimated nap duration and total sleep duration. Fourth, although our study considered numerous possible confounding factors, it was not possible to exclude potential confounding factors completely such as depression and sleep apnea, which might modify or mediate the relationship. However, one large prospective cohort study adjusted for depression and obtained similar results to the current study [26].

In summary, the present findings suggest that night sleep, daytime napping, and total 24-h sleep duration are independently correlated with an elevated risk of stroke in individuals living in rural China. In addition, in the presence of adequate or excessive nighttime sleep, the risk of stroke increased with increasing nap time. However, when nighttime sleep is insufficient, proper napping may mitigate the risk of stroke. Our study highlights the importance of appropriate sleep duration as an important factor in adopting a healthier lifestyle, and people who want to maintain a healthy life should aim for an appropriate sleep duration.