Dietary and cardio-metabolic risk factors in patients with Obstructive Sleep Apnea: cross-sectional study

Introduction

Obstructive sleep apnea (OSA) is considered a public health problem (Lee et al., 2008) and is associated with an increased risk of diabetes, hypertension and cardiovascular diseases (CVD) (Pepperell, Davies & Stradling, 2002; Wang et al., 2013; Dong, Zhang & Qin, 2013). It has also been suggested that OSA may be independently associated with an increased risk of ischemic heart disease, stroke, arrhythmias and mortality (Lurie, 2011). OSA is characterized by repetitive partial or complete closure of the upper airway during sleep that results in hypoxemia and hypercapnia, is frequently associated with arousals, and leads to an increase in myocardial oxygen demand (Dorasamy, 2007). The sum of the number of apneas and the number of hypopneas per hour is described by the apnea hypopnea index (AHI). It classifies OSA patients into mild (5.0–14.9), moderate (15.0–29.9), and severe (≥30.0 events per hour) (Foster et al., 2009). A number of studies have identified obesity as an important risk factor for OSA (Lee et al., 2008; Gasa et al., 2011). As OSA is diagnosed in approximately 30% of obese patients, metabolic dysfunction often present in obese individuals, including hypertriglyceridemia, lower high-density lipoprotein cholesterol (HDL-C) or hypertension, may play an important role in the increased risk of cardiovascular morbidity and mortality associated with OSA (Gasa et al., 2011; Ahmed & Byrne, 2010).

The unique importance of lifestyle interventions in OSA outcomes has already been emphasized in previous studies (Tuomilehto et al., 2009; Chirinos et al., 2014). Different dietary strategies, based on a reduction in energy intake, have resulted in body weight loss, markedly improving metabolic parameters and ameliorating OSA severity (Gozal, 2015; Vasquez et al., 2008). It has been shown that patients suffering from OSA are characterized by either frequent cravings for carbohydrates or the consumption of greater amounts of total calories derived from protein and fat (saturated fatty acids, trans-fatty acids) (Vasquez et al., 2008).

There is limited literature highlighting diet as an important factor related to AHI, and thus, to severity of OSA. In routine practice of treatment in OSA patients, an adequate dietary intake could result in a step forward in supportive care in these patients. Therefore, we aimed to analyze the association between dietary intake and cardio-metabolic biomarkers with OSA severity in patients with risk of OSA.

Material and Methods

Results

Study baseline characteristics are presented in Table 1. Patients with normal AHI (no OSA) were younger than those with OSA. Mean BMI values in all AHI groups exceeded 30 kg/m². However, patients with OSA (mild, moderate, and severe) had higher BMI as compared to those with no OSA. A high proportion of smokers were observed in the group of patients with a mild OSA (AHI 5-14.9). Comparatively higher intakes of alcohol were observed in patients with moderate (57.28 g/day; 95% CI [55.83–58.73]) and severe (95.9 g/day; 95% CI [93.61–98.19]) OSA. According to the PSQI, approximately 60% of the studied individuals described their quality of sleep as poor. There were no differences in education, marital status, residence and financial status across AHI patient groups. Due to small sample sizes mean differences in smoking categories could not be estimated.

Cardio-metabolic markers, dietary intake and factors describing sleep architecture across categories of OSA severity are described in Table 2. The concentration of TG was higher in patients with severe OSA (2.10 mg/dl; 95% CI [2.05–2.15]) as compared to non-OSA subjects (1.50 mg/dl; 95% CI [1.45–1.55]), whereas HDL-C was higher in non-OSA (1.4 mmol/dl; 95% CI [1.38–1.42]) as compared to severe OSA patients (1.1 mmol/dl; 95% CI [1.08–1.12]). Patients from this group also had the highest values of systolic (131.30 mmHG; 95% CI [130.60–132.00]) and diastolic (79.70 mmHG; 95% CI [79.25–80.15]) blood pressure. Significant differences were also observed in the NREM phase of sleep (stage of sleep 1 and 2) across the analyzed groups of patients (non-OSA vs severe-OSA). The highest values of ODI were also observed in the severe OSA group.

Analysis of nutrient intake showed a reduced energy intake in severe-OSA patients (1,546.1 kcal; 95% CI [1,525.87–1,566.33]) as compared to non-OSA patients (1,800.40 kcal; 95% CI [1,768.24–1,832.56]). Similarly, lower intake of carbohydrates and fats were observed in severe OSA as compared to non-OSA patients. Likewise, severe-OSA patients had significantly lower intake of fiber (17.1 g/day; 95% CI [16.81–17.39]) and Vitamin E (7.9 g/day; 95% CI [7.72–8.08]) as compared to non-OSA patients (21.6 g/day; 95% CI [21.15–22.05]) and 9.6 g/day; 95% CI [9.35–9.85], respectively. Similar results were also observed for magnesium. Intake of selected nutrients was also found to differ from guideline recommendations (Jarosz & Bułhak-Jachymczyk, 2008). Using the estimation of energy intake from the FFQ, we did not detect any systematic error in food intake assessment (Table S1).

Results from ordinal logistic regression models assessing association of lifestyle factors with OSA categories of severity showed that a higher intake of alcohol was significantly associated with increased odds of being in an OSA category of higher severity as compared to non-OSA (Odds Ratio (OR) 1.03; 95% CI [1.01–1.04]). Assessment of cardiometabolic parameters in relation to OSA categories of severity showed that patients with higher HDL-C levels had significantly lower odds (OR 0.01; 95% CI [0.0003–0.55]) of OSA severity. Association between sleep stages with OSA severity also showed significantly increased odds of high severity in patients with disturbed sleep stages. Furthermore, OSA severity was also related with obstructive sleep apnea in these patients (OR 5.17; 95% CI [2.8–9.55]). Finally, the investigation of nutrient intake in relation to OSA severity showed that higher intakes of dietary fiber was associated with decreased OSA severity (OR 0.84; 95% CI [0.71–0.98]), whereas higher intakes of Vitamin B1 was associated with increased odds (OR 1.01; 95% CI [1–1.03]) of OSA severity (Table 3).

Discussion

In this study we present the association between severity of OSA, defined according to AHI categories, with selected lifestyle, dietary and cardio-metabolic factors. Alcohol consumption and poor sleep profile were associated with increased OSA severity, whereas higher levels of HDL-C and intake of dietary fiber were associated with lower OSA severity.

Most of the patients suffering from OSA were self-reported poor sleepers (82%) according to the PSQI (Lau et al., 2013), which was also confirmed in our study. Consequently, poor sleep status in OSA patients is usually associated with fatigue and may result in a decrease in physical activity and a compensatory increase in caloric intake leading to weight gain (Stelmach-Mardas et al., 2016). In the current study, mean BMI values in all AHI groups exceeded 30 kg/m², although the mean energy intake did not meet the recommended value (Jarosz & Bułhak-Jachymczyk, 2008). Unfortunately, obesity is very often related to underreporting of food intake by patients, frequently observed in clinical practice and well described in the literature (Johnson, Goran & Poehlman, 1994; Tooze et al., 2004; Stice & Durant, 2014; Nadeem et al., 2014). Furthermore, our patients reported intake from the last 24-hours prior to clinic entry (single intake), so health status and psychological factors related to stress of medical examination in the upcoming day could have also influenced self-reported dietary intake. Nevertheless, 24-hr dietary recall is considered to be a more reliable and unbiased method than FFQ for dietary assessment in the dietary field (Eck, Klesges & Klesges, 1996). Here, we used FFQ frequency information to estimate energy intake using portion sizes as usual consumption amounts to assess whether 24-hr recall intake could have been affected by systematic error. From the estimation of energy intake, we did not detect any systematic error in food intake assessment. In the presented statistical model, we have shown selected lifestyle factors (alcohol consumption) and nutrients (dietary fiber) are associated with OSA severity. Smith et al. (2014) also showed a negative association between liking food rich in dietary fiber and OSA severity and a positive association with high fat food, which could be explained by high levels of ghrelin and leptin resistance. Regarding lipid profile parameters, we observed that the severe OSA group was characterized by lower concentrations of HDL-C. Unfortunately, even future treatment of OSA patients using continuous positive airway pressure (CPAP) may not positively influence the lipid profile, as might be expected (Goris, Meijer & Westerterp, 2001). As Xu et al. (2014) pointed out based on their meta-analysis, CPAP does not alter TG, LDL-C or HDL-C levels, suggesting that it may have no clinically important effect on lipid metabolism. Furthermore, the lipid profile studied in patients also partially reflects the dietary habits, which should be taken under consideration by health professionals to support OSA patients with dietary consulting. Additionally, decreasing HDL-C and increasing TG can be observed with increasing severity of obesity. It was suggested that the AHI can be considered as one of the main indicators of HDL dysfunction (Xu et al., 2014). Moreover, the formation of reactive oxygen species (ROS) is related to the resistive breathing causes respiratory muscle and furthermore may lead to damage of cardiac tissue (Zuo et al., 2012; Zuo et al., 2013). Various nutrient diets can affect ROS levels in OSA patients. Generally, diet can be assessed by using composite evidence-based scores that comprise multiple nutrients and food groups—for instance the Alternate Healthy Eating Index (AHEI) (Mattei et al., 2016). An association between higher AHEI (which reflects better diet quality, based on current evidence) and lower risk of diabetes, CVD, and Metabolic Syndrome has been confirmed in different studies (McCullough et al., 2002; Fung et al., 2007; Belin et al., 2011; Akbaraly et al., 2010; Chiuve et al., 2012). This can be explained by the inclusion of specific foods and nutrients, such as legumes (lower risk), sugar-sweetened beverages, and red or processed meats (higher risk) in the AHEI, which may play a role in the development of chronic diseases (Chiuve et al., 2012). Behavioural changes in food desire under sleep deprivation can also explained the positive association between B1 intake and OSA due to the more often selected food sources of this vitamin (Greer, Goldstein & Walker, 2013). According to data from The National Health and Nutrition Examination Survey III (NHANES III), the prevalence of hypertension can reach up to 42% in men at a BMI of >30 kg/m² and up to 38% in women (Brown et al., 2000). A 10 kg higher body weight is associated with a 3.0 mm Hg higher systolic and a 2.3 mm Hg higher diastolic blood pressure, leading to increased risk of CVD (Reisin, 1986). Furthermore, according to the American Heart Association, obesity can be associated with alterations in hemodynamics, as an increase in oxygen demand produced by excess adipose tissue (∼1.5 mL/kg per minute) requires an increase in cardiac output related to a dysregulation of peripheral vascular resistance (Poirier et al., 2006). In the present study, we have examined mostly obese patients, in which the differences in diastolic blood pressure were highly pronounced across the AHI groups.

The monitored number of desaturation events was also the highest in the group of patients with a high AHI value, which shows that ODI is a very good predictor of AHI. It has already been confirmed by Chung et al. (2012) that ODI can be used as a sensitive and specific tool in clinical practice to detect undiagnosed sleep-disordered breathing. Relief of apneas prevents deep NREM sleep. However, apneas are more commonly observed in stages I and II of NREM and in REM sleep (Chung et al., 2012). It has also been shown that age and sex differences play a role in breathing, especially when compared with breathing during REM sleep (Subramanian, Mattewal & Surani, 2013). However, in our study, sex and age distribution was similar across AHI groups, only showing differences close to significance in the R stage of sleep. Nevertheless, as suggested by Liu et. al Liu et al. (2011) and Subramanian, Mattewal & Surani (2013), patients with either a higher NREM-AHI than REM-AHI or a higher REM-AHI than NREM-AHI should be considered as a part of the spectrum of OSA, rather than a specific clinical entity.

Conclusions

The severity of OSA is related to high alcohol consumption, lower HDL-C and disturbed sleep. The significantly lower dietary fiber intake in patients with higher AHI is of high importance for dietary consultation in clinical practice, which may positively influence cardiometabolic outcomes.

Supplemental Information