Sleep-disordered breathing (SDB) is characterised by recurrent apnoea and hypopnea events during sleep(Reference Peppard, Young and Palta1). An increasing amount of literature suggested that SDB was associated with adverse health outcomes, including hypertension(Reference Wang, Wang and Liu2), diabetes(Reference Li, Sotres-Alvarez and Gallo3), CVD(Reference Chan, Wang and Seet4), stroke(Reference Heinzer, Vat and Marques-Vidal5) and increased all-cause mortality(Reference Gottlieb and Punjabi6). Globally, approximately one billion adults aged between 30 and 69 years suffer from obstructive sleep apnoea, the most common type of SDB(Reference Benjafield, Ayas and Eastwood7). SDB has imposed a considerable health hazard and economic burden on populations worldwide(Reference Benjafield, Ayas and Eastwood7), and its prevalence is still increasing(Reference Peppard, Young and Barnet8).
Previous studies have reported that sleep restriction can increase food intake and total energy expenditure(Reference Capers, Fobian and Kaiser9). Likewise, a clinical trial suggested that sleeping for short hours can change the type of food intake(Reference Tasali, Wroblewski and Kahn10) and increase snack intake(Reference Nedeltcheva, Kilkus and Imperial11). Insomnia was associated with higher intakes of total energy, trans-fat, and Na and a lower intake of vegetables(Reference Cheng, Li and Winkelman12). At the same time, a literature review summarised that higher intakes of vegetables, fruits, fish, dairy products and nuts had benefits for sleep improvement(Reference St-Onge, Mikic and Pietrolungo13). These results suggested a potential bidirectional relation between diet and sleep. In recent years, there is a growing body of research examining the relationship between complete dietary patterns, such as the Mediterranean diet, with sleep disorders and insomnia(Reference Campanini, Guallar-Castillón and Rodríguez-Artalejo14,Reference Muscogiuri, Barrea and Aprano15) . Nevertheless, research on other sleep-related dietary patterns is still at the preliminary stage(Reference Zuraikat, Wood and Barragán16).
Dietary Approaches to Stop Hypertension (DASH) are thought to be the best example of how the nutrition-intensive diet model can prevent chronic diseases(Reference Monsivais, Rehm and Drewnowski17). Since the establishment of the DASH dietary pattern, trials have shown that the DASH diet can reduce the risk of CVD(Reference Steinberg, Bennett and Svetkey18), obesity, metabolic and other diseases, in addition to hypertension(Reference Soltani, Shirani and Chitsazi19). Several studies have recently reported an association between the DASH diet and sleep quality(Reference Daneshzad, Heshmati and Basirat20–Reference Liang, Beydoun and Hossain22). To our knowledge, however, there have been no studies exploring the association between the DASH diet and SDB. Thus, this study aimed to examine the association between the DASH diet and SDB in community-dwelling adults in Suzhou, Eastern China.
Methods
Study population
The Suzhou Food Consumption and Health Survey is a cross-sectional study conducted from August 2018 to September 2020. A multi-stage random sampling method was used to recruit potential participants. The methodology and survey process has been described elsewhere(Reference Wang, Liu and Zhang23). A total of 5058 adults participated in the survey and had FFQ data. Of these, we excluded the participants under the age of 18 years (n 249) and those with implausible dietary data (energy intake < 800 or > 6000 kcal for males, energy intake < 600 or > 4000 kcal for females) (n 459). In addition, participants with BMI < 14 kg/m2 or BMI > 45 kg/m2 (n 21) and abnormal sleep information (individuals with missing sleep data or sleep duration < 2 h/d or > 12 h/d) (n 390) were also removed. The final study population thus included 3939 participants (1869 males and 2070 females). The study protocol was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects were approved by the Institutional Review Board of the Suzhou Center for Disease Control and Prevention. Informed written consent was obtained from each participant.
Dietary intake assessment and calculation of Dietary Approaches to Stop Hypertension score
Dietary intake was assessed with the use of a FFQ derived from the China National Nutrition Surveys. It has been validated with weighted food records(Reference Zhao, Hasegawa and Chen24,Reference Li, He and Zhai25) . Spearman’s correlation coefficients between food intakes from the weighed food records and the FFQ ranged from 0·08 for nuts to 0·76 for rice and rice products among a nationally representative sample of 23 198 participants aged 15–75 years(Reference Li, He and Zhai25). For each of the seventy-four items, participants were asked if they had eaten the food in the past 12 months and, if so, how often they had eaten it on average. FFQ provides four frequency responses of ‘daily’, ‘weekly’, ‘monthly’ and ‘yearly’, asking for the corresponding consumption times and amount of consumption per time. Separate questions were asked about all the types of spices, such as oil and salt, those households used in cooking each month, and how much they ate. Total dietary energy intake and nutrients intake were calculated according to the Chinese Food Composition Table(Reference Yang26).
DASH score was calculated using the method reported by Fung et al. (Reference Fung, Chiuve and McCullough27). The score was based on eight major ingredients including fruits, vegetables, nuts and legumes, dairy products, whole grains, Na, sweet beverages, and red/processed meats. Due to the low consumption of low-fat dairy products in the Chinese population, we replaced this ingredient with regular dairy products(Reference Xiao, Lin and Li28). For each ingredient, participants were classified based on the quintile categories of their intake. For fruits, vegetables, nuts and legumes, dairy products, and whole grains, scores of 1, 2, 3, 4 and 5 were assigned to those in the lowest quintiles, 2nd, 3rd, 4th and the highest quintiles, respectively. In contrast, for Na, sweet beverages and red/processed meats, the opposite scoring system was applied. Each person’s DASH score was calculated by adding up the scores for each of the eight components, which range from 8 to 40 points.
Definition of sleep-disordered breathing
SDB symptoms were obtained using apnoea symptom score, a subscale in the Multivariate Apnea Prediction Index (MAPI), which asked subjects if they have had at least 3 d of sleep in the past month with the following symptoms: loud snoring, breathing holding/pauses and snorting/gasping. Any of which present was assessed as SDB, as defined in our published article(Reference Wang, Liu and Zhang23).
Covariates
Age, sex, education, marital status, smoking, alcohol drinking and sleep duration were obtained using standard questionnaires administered by trained staff. A smoker was defined as a person who had averaged at least one cigarette a day in the previous 30 d. Drinkers were defined as those who had consumed alcohol on average more than once a month in the past year(Reference Wang, Liu and Zhang23). Insomnia was defined as having one or more of the following symptoms at least 3 d per week in the last 30 d: difficulty falling asleep, waking up too early and not being able to fall asleep again, waking up twice or more during the night, or taking sleeping pills at least 1 d in the last 30 d(Reference Wang, Liu and Zhang23). Physical activity was assessed by asking subjects about the times they usually had domestic, occupational, transport and leisure activities at different intensities (low, moderate and vigorous) per week over the past year. BMI was calculated as weight divided by the square of height (kg/m2).
Information on diabetes, hyperlipidaemia, stroke (ischaemic, haemorrhagic and other types) or other chronic diseases (CHD, chronic obstructive pulmonary disease, asthma, bone and joint disease, neck or waist disease, chronic digestive system disease, chronic urinary system disease, and malignant tumours) was obtained by questioning the participants whether they had been diagnosed by a qualified physician at a township or higher hospital. Hypertension was defined as blood pressure ≥ 140/90 mmHg, hospital diagnosis or antihypertensive medication use(Reference Wang, Liu and Zhang23,Reference Wang, Zheng and Zhang29) .
Statistical analyses
To compare the demographic characteristics of study participants across the DASH score quintiles and self-reported SDB, we applied the Kruskal–Wallis tests for continuous variables and the χ 2 test for categorical variables. To examine the association between the DASH dietary pattern and SDB, we used multivariable logistic regression in several models: (1) model 1 was adjusted for age (continuous) and sex; model 2 was adjusted as model 1 plus marital status (married or others), education (junior high school and below, senior high school, or university), physical activity (low: no moderate or vigorous activity; moderate: < 4 h/week moderate or < 2 h/week vigorous activity; high: ≥ 4 h/week moderate or ≥ 2 h/week vigorous activity)(Reference Chen, Koh and Yuan30), BMI (< 24 kg/m2 and ≥ 24 kg/m2), smoking (yes or no), alcohol drinking (yes or no); model 3 was adjusted as model 2 plus energy intake (continuous), hypertension (yes or no), diabetes (yes or no), hyperlipidaemia (yes or no), insomnia (yes or no) and sleep duration (< 7 h/d, 7–9 h/d and > 9 h/d); model 4 was further adjusted for vitamin D intake (continuous).
In addition, interaction and stratified analysis were conducted by age group (18–45 years, 45–60 years and ≥ 60 years), sex, BMI, smoking, alcohol drinking hypertension, diabetes and hyperlipidaemia. Sensitivity analyses were performed to determine the robustness of the primary findings. We restricted the risk association analyses to those participants without hypertension, diabetes, hyperlipidaemia, stroke and other chronic diseases.
SAS statistical software (version 9.4; SAS Institute, Inc.) and R (V.4.1.3, www. r-project.org) were used for statistical analyses. All tests were two-sided, and a P-value < 0·05 was considered statistically significant.
Results
Of the 3939 participants, 2070 (52·6 %) participants were females, with a median age of 48 years (IQR 35–60) and mean BMI of 23·9 kg/m2 (sd 3·4), 968 individuals (24·6 %) had self-reported SDB. The baseline characteristics of the participants across the quintiles of the DASH score are shown in Table 1. Compared with the participants with lower DASH score quintiles, those with higher quintiles were more likely to be younger, female, more educated, non-smokers, non-drinkers, and with lower BMI and prevalence of hypertension (P < 0·001). As expected, achieving a higher DASH score was associated with lower intakes of Na, red/processed meats and sweetened beverages, but higher intakes of energy intake, vitamin D, potassium, fruits, vegetables, nuts and legumes, whole grains and dairy products (P < 0·001).
Table 1. Descriptive characteristics of the study participants by quintiles of DASH score (n 3939)
DASH, Dietary Approaches to Stop Hypertension; Q, quintile.
Data are presented as median (p25, p75) for continuous measures and n (%) for categorical measures. P values were determined by χ 2 tests for categorical and Kruskal–Wallis tests for continuous variables.
As shown in Table 2, those with SDB were more likely to be older, male, married, smokers, drinkers, less educated, have less physical activity, and have higher BMI and hypertension, diabetes, hyperlipidaemia, insomnia, and irrational sleep duration. Of note, those participants with self-reported SDB generally had a lower DASH score (P < 0·001).
Table 2. Descriptive characteristics of the study participants stratified by sleep-disordered breathing (n 3939)
DASH, Dietary Approaches to Stop Hypertension; Q, quintile.
Data are presented as median (p25, p75) for continuous measures and n (%) for categorical measures. P values were determined by χ 2 tests for categorical and Kruskal–Wallis tests for continuous variables.
A higher DASH score was significantly associated with decreased odds of SDB in the basic model adjusted for age and sex, and the model with further adjustment for BMI, education, marital status, physical activity, smoking, alcohol drinking, and in the model 3 with additional adjustment for energy intake, hypertension, diabetes, hyperlipidaemia, insomnia, and sleep duration. The inverse association remained significant in the fully adjusted model with additional adjustment for vitamin D intake. The multivariable-adjusted OR for SDB was 0·68 (95 % CI 0·52, 0·88, P for trend = 0·004) for those in the highest quintile of the DASH score compared with the lowest quintile (Table 3).
Table 3. Adjusted OR (95 % CI) for the association between DASH score and sleep-disordered breathing (n 3939)
DASH, Dietary Approaches to Stop Hypertension; Q, quintile.
Model 1: adjusted for age and sex.
Model 2: further adjusted for BMI, education, physical activity, marital status, smoking and alcohol drinking.
Model 3: additionally adjusted for energy intake, hypertension, diabetes, hyperlipidaemia, insomnia and sleep duration.
Model 4: additionally adjusted for vitamin D intake.
The associations between the DASH score and SDB were similar in subgroups stratified by age group, sex, BMI, smoking, alcohol drinking, hypertension, diabetes and hyperlipidaemia (all P for interaction > 0·05, Fig. 1). Moreover, a multivariable logistic regression analysis of the eight components of the DASH diet showed that vegetables, nuts and legumes, and dairy products were negatively associated with SDB (Table S1). In sensitivity analyses, the association with SDB remained significant when the study population was restricted to 2038 participants without hypertension, diabetes, hyperlipidaemia, stroke and other chronic diseases (multivariable-adjusted OR = 0·96; 95 % CI 0·93, 1·00; P < 0·05).
Fig. 1. DASH score in association with sleep-disordered breathing by strata in the study population (n 3939). Multivariable model: adjusted for age, sex, BMI, education, physical activity, marital status, smoking, alcohol drinking, energy intake, hypertension, hyperlipidaemia, diabetes, insomnia and sleep duration. DASH, Dietary Approaches to Stop Hypertension; Q, quintile.
Discussion
Principal findings
In this study, we observed that adherence to the DASH dietary pattern was associated with reduced odds of SDB. The association between DASH score and SDB appeared to be independent of potential confounding factors and was similar across subgroups stratified by age, sex, BMI, smoking, alcohol drinking, hypertension, diabetes and hyperlipidaemia. The association persisted when restricting the participants to those without chronic diseases. Inverse associations with SDB were also observed for the DASH components vegetables, nuts and legumes, and dairy products.
The Dietary Approaches to Stop Hypertension diet and sleep-disordered breathing
Our study found an inverse association between the DASH score and SDB. Similar results were reported by Liang et al. (Reference Liang, Beydoun and Hossain22), who observed the DASH diet score was inversely related to sleep-related daytime dysfunction among middle-aged and older individuals. A study from Southern Iran showed that mothers and their infants with higher DASH scores had fewer sleep disorders than those with lower DASH scores(Reference Karbasi, Azaryan and Zangooie31). Besides, the inverse association between the DASH diet and insomnia has been found in adolescent girls aged 12 to 18 years(Reference Pahlavani, Khayyatzadeh and Banazadeh21,Reference Rostami, Khayyatzadeh and Tavakoli32) . It seems that adherence to the DASH diet is associated wither better sleep quality in different populations, as reported in our study. In the final model, we furthermore adjusted for vitamin D intake levels, and the results were essentially not affected. Besides the DASH score, similar associations have been reported for other diet quality indices including the Mediterranean diet pattern(Reference Campanini, Guallar-Castillón and Rodríguez-Artalejo14,Reference Muscogiuri, Barrea and Aprano15,Reference Zuraikat, Makarem and Liao33) and the Dietary Inflammation Index (DII) with sleep problems(Reference Godos, Ferri and Caraci34).
The DASH diet supports a higher intake of foods considered healthy, including fruits, vegetables, nuts and legumes, whole grains, and dairy products, and lower consumption of red meat, carbohydrates, and fizzy drink, which have been shown to improve sleep(Reference St-Onge, Mikic and Pietrolungo13,Reference Zuraikat, Makarem and Liao33,Reference Godos, Grosso and Castellano35) . A randomised clinical trial showed that men with a Na-restricted diet could improve symptoms of SDB(Reference Fiori, Martinez and Gonçalves36). Reid et al. showed that lower whole-grain intake was associated with increased consumption of red meat and moderate to severe SDB(Reference Reid, Maras and Shea37). Of the eight components of the DASH diet in our study, vegetables, nuts and legumes, and dairy products were inversely associated with SDB, but other components were not associated. Therefore, the inverse association with SDB was mainly driven by vegetables, nuts and legumes, and dairy products in our study population.
Possible mechanisms
Although the exact mechanism is unclear, we hypothesised several potential explanations for the observed association. Reimund proposed a theory that sleep itself contained an antioxidant defence systems, meaning that excess free radicals can be reduced during sleep by a decreased rate of formation of free radicals and increased efficiency of endogenous antioxidant process(Reference Reimund38). Patients with SDB have increased oxidative stress in the body due to repeated ischaemia-reperfusion(Reference Kent, Ryan and McNicholas39) and meanwhile underutilisation of the antioxidant function. In addition, inflammation is considered relevant to the pathogenesis of obstructive sleep apnoea(Reference Unnikrishnan, Jun and Polotsky40). Fruits, vegetables, legumes and whole grains rich in the DASH diet have been shown with antioxidative(Reference Pirouzeh, Heidarzadeh-Esfahani and Morvaridzadeh41) and anti-inflammatory effects(Reference Esmaillzadeh, Kimiagar and Mehrabi42,Reference Soltani, Chitsazi and Salehi-Abargouei43) . Moreover, the tryptophan–hydroxytryptamine–melatonin system may be involved in the association between the DASH diet and SDB(Reference Scoditti, Tumolo and Garbarino44). Vegetables and fruits, whole grains, legumes, and dairy products are rich sources of tryptophan, serotonin, and or melatonin(Reference Zuraikat, Wood and Barragán16). It was reported that melatonin intake had a positive effect on improving oxidative stress and subsequent SDB(Reference Morvaridzadeh, Sadeghi and Agah45). Further studies are still warranted to clarify the underlying pathophysiological mechanisms.
Strengths and limitations
To the best of our knowledge, this is the first study to examine the association of the DASH dietary pattern with SDB. In addition, we performed stratification analyses, as well as sensitivity analyses to explore and confirm the robustness of our findings. However, the present study has also some limitations. First, due to the cross-sectional nature, it is impossible to conclude a causal relationship between the DASH score and SDB. Reverse causality (i.e. SDB-caused dietary changes) cannot be precluded. Second, self-report bias may have existed because both diet and SDB information were collected by questionnaires. Third, residual confounding may have existed due to unmeasured or unknown factors, including for example snack intake before sleep, body composition, and medications related to depression or sleep.
Conclusion
The present study provides novel evidence that adherence to the DASH diet is associated with a decreased likelihood of having SDB. Prospective studies are warranted to further confirm these findings.
Acknowledgements
The author would like to thank the Suzhou CDC for organizing the survey, as well as the local CDC and communities in each participated district for their support.
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
S. L. formulated the research questions, analysed the data, interpreted the findings and drafted the manuscript. C. W., S. T., Y. Z. and K. Z. proofread and revised the original manuscript. B. W. and H. Z. designed the study, supervised the work, critically revised the manuscript and approved the final manuscript.
The authors declare no conflicts of interest.
