INTRODUCTION
Poor sleep quality and inadequate sleep duration in adults have been associated with various adverse health outcomes, including diabetes, metabolic syndrome, cardiovascular diseases, and increased mortality. Additionally, insufficient sleep is associated with increased inflammation, diminished cognitive function, and a higher risk of frailty. These findings highlight the critical role of sleep in maintaining overall health and preventing disease progression [1-4].
A recent review has emphasized the need to explore the impact of dietary patterns and specific nutrients on sleep [5]. The Mediterranean diet (MD), widely regarded as one of the healthiest dietary patterns globally, emphasizes fruits, vegetables, legumes, olive oil, nuts, and whole grains, with moderate-to-high consumption of fish, and dairy products while limiting saturated fats, sweets, and processed meats [6,7]. This diet is associated with reducing overall mortality and offers numerous health benefits, including preventing obesity, type 2 diabetes, cardiovascular diseases, and some cancers [8,9]. Furthermore, adherence to the MD has been associated with improved cognitive function, mental health, and sleep quality [10]. Previous studies have also suggested that great er adherence to the MD is associated with adequate sleep duration and improvements in several sleep quality indicators [11,12].
However, most studies have primarily focused on Western populations, with limited research on Eastern populations, mainly Koreans, whose sleep habits may differ from those of typically studied groups. Therefore, in this study, we aimed to investigate the relationship between MD adherence and sleep quality in the Korean population. Additionally, all analyses were stratified based on sex, given that prior research on the relationship between sleep and diet has frequently reported sex-specific associations [13]. These analyses hold significant public health relevance, as sleep and diet are modifiable factors influencing the risk of chronic diseases and mortality.
METHODS
Study population
Data were derived from the Korean Genome and Epidemiology Study (KoGES), specifically from the KoGES_Ansan and Ansung cohort, which includes community-based participants from urban and rural areas. In this cohort, individuals aged 40–69 years were enrolled at baseline between 2001 and 2002. Participants were recruited through several methods, including on-site invitations, mailed letters, phone calls, media outreach, and community leader-led conferences. Those interested were asked to visit survey centers, where they underwent interviews, completed questionnaires administered by trained staff, and received physical examinations. Typical reasons for non-participation included outdated contact information, busy schedules, and non-responsiveness to outreach efforts. Therefore, of the initial 10,030 participants who completed the health interview survey, we excluded those with missing data on sleep duration (n=800), dietary components (n=1,067), and incomplete records on compounders (n=302). This resulted in a final analytical sample of 7,861 eligible participants (3,670 men and 4,191 women) (Figure 1).
Notably, all eligible participants provided written informed consent to participate. This study adhered to the 1975 Declaration of Helsinki’s ethical guidelines and was approved by Severance Hospital’s Institutional Review Board (IRB) (9-2024-1250).
Sleep assessment
Sleep duration was assessed by asking participants, “On average, how many hours do you typically sleep each night?” Sleep quality was evaluated using several indicators of poor sleep quality, including overall poor sleep quality, difficulty falling asleep, waking up during the night, early awakening with trouble falling back asleep, the need for daytime naps, not feeling rested in the morning, use of sleeping medications, snoring, and excessive daytime sleepiness (Epworth Sleepiness Scale score). The Epworth Sleepiness Scale is a self-administered questionnaire that measures daytime sleepiness by evaluating the likelihood of dozing off in eight common situations [14]. Each situation is rated from 0 to 3, with a total score ranging from 0 to 24. A score of >10 suggests excessive daytime sleepiness. The eight situations include sitting and reading, watching TV, sitting inactive in public, being a passenger in a car, lying down to rest, talking to someone, sitting quietly after lunch, and sitting in a car stopped in traffic. Therefore, based on these outcomes, we determined the total number of sleep quality indicators affecting each participant.
Assessment of dietary intake and MD adherence
For dietary assessment, a semi-quantitative food frequency questionnaire (FFQ) was used, which included 103 items and was administered as part of the KoGES study [15]. Participants reported the frequency and portion sizes of foods consumed over the past year. The FFQ data and a food composition database were utilized to estimate dietary intake, including total calories, macronutrients, and micronutrients.
Adherence to the MD was evaluated using the Korean version of the Mediterranean Diet Adherence Screener (K-MEDAS questionnaire), which had been developed and validated in a previous study [16]. The K-MEDAS questionnaire consists of 14 items, each scored as 0 or 1, resulting in a total score ranging from 0 to 14, with higher scores indicating greater adherence to the MD. The questionnaire evaluates dietary habits and the frequency of consumption of various foods, including perilla or olive oil, vegetables, fruits, red meat, butter and margarine, soft drinks, wine, beans and tofu, fish and seafood, sweets, nuts, poultry, and whole grains. Each food category is awarded 1 point if the specified criteria are met. Further details on the K-MEDAS questionnaire can be found in Supplementary Table 1. Based on tertiles, participants’ responses were classified into three adherence categories: low (score ≤6), moderate (score 7), and high adherence (score ≥8).
Covariates
Body mass index (BMI) was calculated by dividing weight (kg) by height (m) squared. Educational levels were categorized into primary, secondary, and university education. Smoking status was classified as never, former, or current smoker. Physical activity (PA) was measured using the Global Physical Activity Questionnaire, validated in Korean, covering work, transport, leisure, and sedentary behavior. PA was reported as minutes per week across these domains, and metabolic equivalent task (METs) · min/week was used for analysis. Socioeconomic factors included low household income (≤1 million won), employment status (unemployed or housewives), marital status (single, divorced, or separated), and residential area (urban vs. rural). Medical history, including cardiovascular disease (ischemic heart disease, stroke, and heart failure), diabetes, and cancer, was assessed through a self-reported questionnaire asking whether the participant had ever been diagnosed with a disease by a doctor. If a participant reported a medical history, this information was confirmed by the researcher during an interview.
Statistical analysis
Data are presented as the mean±standard deviation or as counts with percentages. Analysis of variance was used to analyze the demographic characteristics of study participants for continuous variables, whereas categorical variables were assessed using the chi-square test.
We employed 1:1 exact matching to minimize confounding by pairing subjects who have identical values for age. This method ensures that the low, moderate, and highly adherent groups are balanced in terms of key demographic characteristics, enabling more accurate and unbiased comparisons of outcomes. Postmatching demographics of the adherent groups were compared using the generalized estimating equation (GEE) method to account for the characteristics of the matched data.
Multinomial logistic regression models estimated the odds ratio (OR) and its 95% confidence interval (CI) using a generalized linear mixed model (GLMM) for sleep quality across tertiles of adherence categories. Sleep quality indicators were grouped into three categories: ≤1, 2–3, and ≥4, with those having ≤1 indicator as the reference group. Analyses were conducted separately for each of the nine sleep indicators using logistic regression models. The model was adjusted for age, education level, BMI, PA, current smokers, current drinkers, energy intake, household income, employment status, marital status, residential area, and a number of chronic diseases, selected based on clinical confounders. For all statistical tests, statistical significance was set at p<0.05, and all analyses were conducted using R version 4.3.0 (R Foundation for Statistical Computing, Vienna, Austria).
RESULTS
Characteristics of the study population
Supplementary Table 2 presents the characteristics of the 3,670 men and 4,191 women participants according to MD adherence before matching, stratified based on sex. In men and women, higher adherence to the MD was significantly associated with older age. Among men, higher adherence was associated with lower smoking rates and a reduced prevalence of cardiovascular disease but a higher prevalence of hypertension and diabetes. However, in women, higher adherence was also associated with an increased prevalence of diabetes and higher overall energy intake.
Table 1 presents the characteristics of participants according to MD adherence levels after age matching. Among men (n=1,485), higher adherence was significantly associated with lower smoking rates and an increased prevalence of diabetes. However, in women (n=2,676), higher adherence was associated with higher education levels, higher rates of being housewives, lower rates of being single, divorced, or separated, a higher prevalence of living in urban areas, and a slightly increased prevalence of diabetes. Notably, no significant differences in sleep duration were observed across the MD adherence groups for both men and women.
Table 2 presents the OR for the association between MD adherence and the number of poor sleep quality indicators in men and women. Among men, no significant association was observed between MD adherence and poor sleep quality. In contrast, higher MD adherence was significantly associated with lower odds of reporting multiple poor sleep quality indicators in women. Specifically, for women with 2–3 indicators, the OR for high adherence was 0.82 (95% CI: 0.69–0.97, p=0.024), and for women with ≥4 indicators, the OR was 0.83 (95% CI: 0.72–0.97, p=0.018). These results remained significant after adjusting for age, education, BMI, PA, smoking, alcohol consumption, energy intake, household income, employment status, marital status, residential area, and number of chronic diseases, highlighting a sex- and age-specific effect of the Mediterranean diet on sleep quality.
Table 3 shows the association between MD adherence and each indicator of poor sleep quality based on sex after adjusting for age, education, BMI, PA, smoking, alcohol consumption, energy intake, household income, employment status, marital status, residential area, and number of chronic diseases. In women, higher adherence to the MD was significantly associated with lower odds of not feeling rested in the morning (OR: 0.80, 95% CI: 0.66–0.97, p=0.024). However, no significant associations were observed in men.
DISCUSSION
Our study revealed differences in the association between MD adherence and sleep quality between men and women. After age matching, no significant associations were found between MD adherence and poor sleep quality in men. However, higher adherence to the MD in women was significantly associated with lower odds of reporting multiple poor sleep quality indicators for those with 2–3 and those with ≥4 indicators. Specifically, women with higher adherence were less likely to report not feeling rested in the morning. These results suggest that MD adherence may positively impact sleep quality, particularly in women.
Our findings are consistent with previous research showing that greater adherence to the MD is associated with improved sleep quality [17]. For instance, Godos et al. [18] found that each point increase in the MD score was associated with a 10% higher likelihood of better sleep quality. The balanced composition of carbohydrates, fats, and proteins in the MD and its high content of vitamins and polyphenols, primarily from foods like fruits, vegetables, fish, nuts, cereals, and olive oils, may explain its beneficial effects on sleep [19]. Conversely, diets high in saturated fats, sugary beverages, and processed meats, which are less common in the MD, have been associated with poorer sleep quality and an increased prevalence of insomnia symptoms [20,21]. Additionally, macronutrient imbalances, especially excessive intake of refined carbohydrates and fats, may negatively impact sleep; however, the evidence remains inconsistent [22].
Notably, several mechanisms associated with the MD may explain its connection to improved sleep quality. Rich in plant-based foods and healthy fats like olive oil and omega-3-rich seafood, the MD provides essential nutrients that support brain health and neurotransmission, such as docosahexaenoic acid and eicosapentaenoic acid, which may enhance sleep [23]. Foods typically found in the MD, like olives and grapes, are natural sources of melatonin, a hormone that regulates circadian rhythms and improves sleep patterns [24]. Additionally, the high content of antioxidants in the MD, including polyphenols and monounsaturated and polyunsaturated fats, helps reduce inflammation and oxidative stress, which can further improve sleep quality [25]. Furthermore, the emphasis on fresh, unprocessed ingredients in the MD may also promote regular sleep patterns and reduce stress, contributing to better overall sleep outcomes [26]
Notably, in our study, we found that higher adherence to the MD was associated with better sleep quality indicators in women but not men. The exact reasons for this sex difference are unclear; however, current research shows mixed findings on whether diet impacts sleep more strongly in women or men. Furthermore, some studies suggest that men may be more affected by poor diet regarding insomnia or weight gain from sleep deprivation [27,28], whereas others show that women experience a more vital link between diet and sleep quality [5,29]. Consistent with our results, a recent longitudinal study found that changes in fruit and vegetable consumption were associated with changes in sleep quality only in women [30]. The observed association between MD adherence and sleep quality in women, but not in men, could be explained by several factors. Research shows that men and women have distinct sleep patterns, with women typically experiencing longer sleep durations but also more frequent sleep disturbances, such as fragmented sleep and difficulty falling asleep [31].
Hormonal fluctuations are also significant in sleep patterns in women, with studies showing that sleep patterns can vary during different life stages, such as the menstrual cycle and menopause [32,33]. These hormonal and physiological differences may lead to more significant variability in sleep patterns in women, potentially making them more sensitive to external factors, like diet, which could explain the stronger association observed in women.
This study has some limitations. First, self-reported sleep duration and quality may be subject to recall and social desirability biases. Second, the cross-sectional design limits our ability to establish a causal relationship between MD adherence and sleep quality, and reverse causation cannot be entirely excluded. It is well-established that insufficient sleep can negatively impact dietary choices; however, there is also growing evidence that diet can influence sleep quality [34], suggesting that future studies should consider both factors simultaneously. Additionally, variables such as shift work or jet lag, which can significantly affect sleep patterns, were not accounted for. These limitations underscore the need for further research, including clinical trials, to better understand the complex relationship between diet and sleep.
However, despite these limitations, our study has several strengths. To our knowledge, this is the first study to examine the relationship between dietary patterns and sleep quality in a Korean population, specifically analyzing sex differences and providing new insights into the role of diet in sleep health for men and women. The study produced robust results even after adjusting for various potential confounders. Furthermore, the use of propensity score matching, particularly for age, strengthened the validity of our findings by minimizing potential selection bias. These strengths enhance the reliability and significance of our study’s conclusions.
In this study, we found that MD adherence did not affect sleep quality in men. However, it was associated with significantly lower odds of poor sleep quality indicators in women, particularly in reducing morning fatigue. These findings suggest that adherence to the MD may be a practical approach to improving sleep quality in women, potentially reducing the risk of related health issues. Therefore, promoting the MD is a valuable intervention in clinical settings, particularly for women experiencing poor sleep quality.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement1
