Association and Risk Factors for Obstructive Sleep Apnea and Cardiovascular Diseases: A Systematic Review

Obstructive sleep apnea (OSA) is a serious, potentially life-threatening condition. Epidemiologic studies show that sleep apnea increases cardiovascular diseases risk factors including hypertension, obesity, and diabetes mellitus. OSA is also responsible for serious illnesses such as congestive heart failure, stroke, arrhythmias, and bronchial asthma. The aim of this systematic review is to evaluate evidence for the association between OSA and cardiovascular disease morbidities and identify risk factors for the conditions. In a review of 34 studies conducted in 28 countries with a sample of 37,599 people, several comorbidities were identified in patients with severe OSA—these were: heart disease, stroke, kidney disease, asthma, COPD, acute heart failure, chronic heart failure, hyperlipidemia, thyroid disease, cerebral infarct or embolism, myocardial infarction, and psychological comorbidities including stress and depression. Important risk factors contributing to OSA included: age > 35 years; BMI ≥ 25 kg/m2; alcoholism; higher Epworth sleepiness scale (ESS); mean apnea duration; oxygen desaturation index (ODI); and nocturnal oxygen desaturation (NOD). Severe OSA (AHI ≥ 30) was significantly associated with excessive daytime sleepiness and oxygen desaturation index. The risk of OSA and associated disease morbidities can be reduced by controlling overweight/obesity, alcoholism, smoking, hypertension, diabetes mellitus, and hyperlipidemia.


Introduction
Sleep disordered breathing, alternatively called obstructive sleep apnea (OSA), is one of the most important biological cause of chronic sleep fragmentation and sleep deficiency [1]. OSA syndrome represents a challenge for the healthcare system due to its high prevalence in the adult population [2,3]. It causes intermittent nocturnal hypoxemia, excessive daytime sleepiness, and persistent elevated sympathetic tone during the day [1]. In a population-based cluster sampling of Sao Paulo inhabitants in Brazil [3], the prevalence rate of OSA was 32.8% (95% confidence intervals (CI), 29.6% to 36.3%), one of the highest rates observed compared to other epidemiologic studies. OSA rates were higher in men than women (odds ratio (OR) = 4.1; 95% CI, 2.9 to 5.8; p < 0.001) and in obese individuals (OR = 10.5; 95% CI, 7.1 to 15.7; p < 0.001) compared with individuals of normal weight [3]. The prevalence studies indicate that sleep disturbances including chronic sleep deprivation and sleep fragmentation due to various environmental and biological factors lead to mood disorders, anxiety disorders, poor cognition, memory deficiency, and decreased performance in academia and the workplace [4][5][6].
OSA is a significant risk factor for major cardiovascular diseases, including arterial hypertension, ischemic heart disease, heart failure, rhythm/conduction disturbances, as well as cerebral stroke [7,8]. The underlying mechanisms explaining associations between OSA and cardiovascular disease are not entirely delineated. Important risk factors for OSA include obesity, craniofacial or oropharyngeal anatomic abnormalities, male sex, and

Search Guidelines
The primary search engines that were used to select articles included PubMed, EBSCOhost, Scopus, Cochrane Library, Google Scholar, MEDLINE, Alt HealthWatch, CINAHL Consumer Health Complete, and Health Source: Nursing/Academic Edition with full text. The studies were selected for the review based on inclusion criteria such as (1) English only, (2) human studies, (3) scholarly papers, (4) age groups 19 and older, (5) had cardiovascular diseases with or without comorbidities, and (6) had sleep apnea. The search was performed in April-September 2021. The time limit for the studies included in this review was from 2018-2021. The search terms used included: 'sleep apnea' OR 'obstructive sleep apnea' or 'apnea' OR 'hypertension' OR 'high blood pressure' OR 'cardiovascular diseases' OR 'risk factors' OR 'comorbidities'.

Screening Guidelines
The Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines (2009) [18] was followed to document the review process. Selected abstracts were assessed to make sure that the studies met the eligibility criteria for inclusion. Three researchers (AKM, ARB, and EAJ) independently screened the titles and retrieved the abstracts based on the inclusion criteria and exclusion criteria as mentioned in Table 1. Selected abstracts were reviewed to assess their eligibility for a full text review. Studies were included in the review if a consensus was achieved by all three researchers. Conflicts were controlled by having group discussions between all three of the researchers of this study. During the second phase of screening, articles without a full text and review articles were excluded. After discussion, more articles were excluded because they were not available in full text, were not inclusive of an adult population (19+), and were not directly related to hypertension and sleep apnea. Finally, all the researchers agreed upon 34 studies that were identified for further evaluation. The PRISMA flow chart (Figure 1) exhibits the search and the inclusion and the exclusion process for the systematic review.

Screening Guidelines
The Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines (2009) [18] was followed to document the review process. Selected abstracts were assessed to make sure that the studies met the eligibility criteria for inclusion. Three researchers (AKM, ARB, and EAJ) independently screened the titles and retrieved the abstracts based on the inclusion criteria and exclusion criteria as mentioned in Table 1. Selected abstracts were reviewed to assess their eligibility for a full text review. Studies were included in the review if a consensus was achieved by all three researchers. Conflicts were controlled by having group discussions between all three of the researchers of this study. During the second phase of screening, articles without a full text and review articles were excluded. After discussion, more articles were excluded because they were not available in full text, were not inclusive of an adult population (19+), and were not directly related to hypertension and sleep apnea. Finally, all the researchers agreed upon 34 studies that were identified for further evaluation. The PRISMA flow chart (Figure 1) exhibits the search and the inclusion and the exclusion process for the systematic review.

Quality Appraisal
Studies were appraised for quality by all three of the researchers separately based on a pre-determined rating scale. The rating scale from 0 to 4 was based on the following criteria: (1) study design: cross-sectional, case-control, or cohort study = 1, otherwise = 0; (2) sample size: large = 1, small = 0; (3) selection of samples: random selection or lack of bias = 1, nonrandom or convenience sample or presence of bias = 0; and (4) analyses: robust (such as multivariate analyses) = 1, otherwise = 0. We followed the similar rating scales published in an earlier study [19]. Based on the above-mentioned criteria, the three researchers rated each of the 34 articles independently from a range of 0 to 4 and categorized as low (a score of 0-1), moderate (a score of 2-3), or high (a score of 4). An average value of the three scores was presented as the final score, because there were no significant inter-observer variations in the assessment of the quality of the included articles (Table 2).

Quality Appraisal
Studies were appraised for quality by all three of the researchers separately based on a pre-determined rating scale. The rating scale from 0 to 4 was based on the following criteria: (1) study design: cross-sectional, case-control, or cohort study = 1, otherwise = 0; (2) sample size: large = 1, small = 0; (3) selection of samples: random selection or lack of bias = 1, non-random or convenience sample or presence of bias = 0; and (4) analyses: robust (such as multivariate analyses) = 1, otherwise = 0. We followed the similar rating scales published in an earlier study [19]. Based on the above-mentioned criteria, the three researchers rated each of the 34 articles independently from a range of 0 to 4 and categorized as low (a score of 0-1), moderate (a score of 2-3), or high (a score of 4). An average value of the three scores was presented as the final score, because there were no significant inter-observer variations in the assessment of the quality of the included articles (Table 2). There was an increased association of narcolepsy with anxiety, thyroid disease, hypertension, and hyperlipidemia. At the end of 9.9 years (SD 7.27 years), there was increased association of narcolepsy with peripheral neuropathy, glucose intolerance, and automobile-related trauma. OSA symptoms were significantly increased in the narcolepsy group as compared to controls (at diagnosis: OR = 69.25, 95% CI = 9.26 to 517.99, p < 0.0001; at end (after about 10 years) OR = 13.55, 95% CI = 5.08 to 36.14, p < 0.0001). Score = 4; high; study design, sample size, sample selection and data analyses, each scored 1.
Diseases 2021, 9,88    The patients were categorized into two groups: no-to-mild OSA and moderate-to-severe OSA, by using the respiratory disturbance index (RDI) and 3% oxygen desaturation index (ODI). BMI was correlated with RDI and 3% ODI (r = 0.56 and 0.43, respectively). Sleep apnea severity was associated with pulse rate variability but not with blood pressure variability.  The frequency of sleep-disordered breathing was 1.4 times higher in the cases than controls (52.5% vs. 37.5%, respectively; p = 0.18). Severity of sleep-disordered breathing was more than twice as common in women with gestational hypertension or pre-eclampsia compared to controls (35% vs. 15%, p = 0.039). Score = 2; moderate; study design = 1 robust analyses = 1; others = 0.

Results
A summary of the study design, the countries of the studies, major findings regarding the correlation between cardiovascular disease with sleep apnea, comorbidities, and the quality appraisal of the studies are presented in Table 2. The 34 studies included in the systematic review comprised of 37,599 participants. In this review, the studies represented 28 countries including Argentina, Australia, Belgium, Brazil, Bulgaria, Cameroon, China, Czech Republic, Finland, France, Germany, Greece, India, Iran, Italy, Ireland, Japan, Korea, Lebanon, New Zealand, Poland, Romania, Spain, Saudi Arabia, Slovakia, Taiwan, Turkey, and the United States. A multicenter study, conducted by Wächter et al., 2020 [51], represented 11 sleep centers in 8 European countries.

Sample Size
The sample sizes varied widely, mainly due to the differences in the study methodologies. However, a few notable studies involving large sample sizes were as follows: a cross-sectional study with 12,500 adults conducted by Paine et al. in New Zealand in 2019 [44]. In this study, the authors showed the highest prevalence of sleep disturbances among the indigenous Māori population. The associated comorbidities for sleep apnea included poorer mental health, high blood pressure, diabetes mellitus, heart disease, obesity, current smoking, and hazardous drinking. The second largest cross-sectional study (n = 5776 women) was conducted by Louis et al. in the United States [39], in which increasing age, obesity, and snoring were identified as predictors of sleep-disordered breathing.
Two other large studies, both being of prospective cohort design, included 3101 in a study in the United States [25] and 2535 in a Finnish study [35]. In the US study [25], OSA symptoms were significantly increased in patients with narcolepsy as compared to controls at diagnosis (OR = 69. 25, 9.26 to 517.99, p < 0.0001) at nearly 10 years of followup (OR = 13.55, 5.08 to 36.14, p < 0.0001). The Finnish study [35] identified a significant association of pre-existing diabetes and cardiovascular diseases with an increased risk for an expedited worsening of intermittent hypoxemia. Another cross-sectional study involving 2740 adults in Korea [50] identified the "high risk" individuals as being older aged and having a BMI ≥ 25 kg/m 2 for OSA. Several other studies that involved large samples included a cross-sectional study (n = 2208) in Japan addressing the association between hypertension and OSA [42], a multicenter study (n = 1877) in Europe looking for the effects of OSA severity and other comorbidities on sleep pattern [51], a prospective cohort study (n = 1825) in Korea determining the association between hypertension with OSA in patients with high levels of homocysteine in the blood [36], and a cross-sectional study (n = 1272) in Belgium determining the prevalence of hypertension with insomnia [32].

Study Design
Of the 34 studies, the majority were cross-sectional studies (n = 21, 62%), 10 (29%) were prospective cohort studies, and 3 (9%) were retrospective studies. Of the cross-sectional studies, 5 were performed in a hospital or a clinic setting, and 16 were community-based studies. Only one study by Seguro et al., in France [48] proposed a case-control design by comparing OSA severity measured by AHI scores among 100 cases with hypertension and 100 age-and gender-matched normotensive individuals as controls. In that study, an association was found between resistant hypertension and untreated obstructed sleep apnea.
Based on this review, several modifiable risk factors for OSA were identified. These include: overweight/obesity, alcoholism, smoking, hypertension, diabetes mellitus, and hyperlipidemia.

Bidirectional Association between OSA and Hypertension
The prevalence of hypertension was higher among subjects with OSA and with high homocysteine (Hcy) levels. After adjusting for confounding factors, subjects with OSA and high Hcy levels had a 1.86-fold risk of developing hypertension compared to those without OSA and high Hcy levels [36]. On the other hand, severe OSA syndrome was highly prevalent among hypertensive subjects [21]. Of the individuals with hypertension, 44% had an AHI of >15 events/h, and CPAP was recommended in 69 (41%) patients [23]. However, no causal association could be established based on the nature of the studies.

AHI Severity and Cardiovascular Risk
Based on the ACC/AHA risk calculator, there was a gradual increase in the risk of cardiovascular events (coronary event or stroke) proportional to the increase in AHI severity categories: 8.7% for patients with normal AHI; and 19.7%, 27.8%, and 30.3% for patients with OSA of 5.1-14.9, 15-29.9, and ≥30, respectively [23]. The French study by Seguro et al. [48] is the only case-control study reviewed here. It is noteworthy that the authors reported a lower Epworth sleepiness scale (ESS) in hypertensive patients than in normotensive patients with severe OSA, which could be misleading. The authors, therefore, suggested estimating the OSA severity by polysomnography, even if the ESS score is low in patients with hypertension [48].

OSA and Comorbidities
The three most common comorbidities that were reported were: hypertension in 30 out of 34 (88%), followed by overweight/obesity in 17 (50%), and diabetes mellitus in 12 (35%) of the studies. Other important physical conditions that were associated with OSA included heart disease (six studies), stroke (three studies), kidney disease including renal insufficiency (three studies), asthma (two studies), COPD (two studies), acute heart failure (two studies), chronic heart failure (two studies), hyperlipidemia (two studies), thyroid disease (two studies), cerebral infarct or embolism (two studies), myocardial infarction (two studies), and psychological comorbidities including stress, anxiety, and depression (three studies). Figure 2 shows the adjusted OR and 95% CI of the comorbidities or risk factors for severe OSA.

Quality Appraisal
The mean ± SD score for the quality of the studies was 2.82 ± 1.06. Of the 34 studies, the majority (n = 20, 59%) scored moderate, 13 (39%) scored high, and one (3%) scored low [46]. Although the Iranian study [46] scored low, the study provided data on cardiovascular risk factors of OSA in a special population (locomotive drivers), which are pertinent to this review.

Discussion
This systematic review showed that the severity of OSA is directly associated with major risk factors for cardiovascular disease, including hypertension, overweight and obesity, diabetes mellitus, hyperlipidemia, and physical inactivity. The review also identified several risk factors for OSA, such as increasing age, BMI ≥ 25 kg/m 2 , alcoholism, smoking, higher ESS score, mean apnea duration, oxygen desaturation index, and nocturnal oxygen desaturation. Severe OSA was significantly associated with excessive daytime sleepiness and oxygen desaturation index.
Heart disease is the leading cause of death worldwide. Therefore, assessing the risk of its occurrence is a crucial step in predicting serious cardiac events. The characteristics of cardiovascular disease are identified by its primary risk factors. Those factors are diabetes mellitus, the level of lipids in the blood, coronary artery function, and kidney function [53]. A data mining model was developed with a clinical laboratory database using a naïve Bayes classifier to detect cardiovascular risk, and it was tested for its accuracy in predicting three levels of risk [53]. Our review results are aligned with the sophisticated data mining model as proposed in determining the risk of cardiovascular disease.
Women's risk assessment for cardiovascular disease is equally important. Traditional atherosclerotic cardiovascular disease (ASCVD) risk factors, such as hypertension, diabetes mellitus, hypercholesterolemia, and obesity, affect both sexes, but some may affect women differently and are considered to be more potent [54]. Although our primary focus is to identify the relationship between OSA and cardiovascular disease, our review described an Australian study that demonstrated the association of gestational hypertension

Quality Appraisal
The mean ± SD score for the quality of the studies was 2.82 ± 1.06. Of the 34 studies, the majority (n = 20, 59%) scored moderate, 13 (39%) scored high, and one (3%) scored low [46]. Although the Iranian study [46] scored low, the study provided data on cardiovascular risk factors of OSA in a special population (locomotive drivers), which are pertinent to this review.

Discussion
This systematic review showed that the severity of OSA is directly associated with major risk factors for cardiovascular disease, including hypertension, overweight and obesity, diabetes mellitus, hyperlipidemia, and physical inactivity. The review also identified several risk factors for OSA, such as increasing age, BMI ≥ 25 kg/m 2 , alcoholism, smoking, higher ESS score, mean apnea duration, oxygen desaturation index, and nocturnal oxygen desaturation. Severe OSA was significantly associated with excessive daytime sleepiness and oxygen desaturation index.
Heart disease is the leading cause of death worldwide. Therefore, assessing the risk of its occurrence is a crucial step in predicting serious cardiac events. The characteristics of cardiovascular disease are identified by its primary risk factors. Those factors are diabetes mellitus, the level of lipids in the blood, coronary artery function, and kidney function [53]. A data mining model was developed with a clinical laboratory database using a naïve Bayes classifier to detect cardiovascular risk, and it was tested for its accuracy in predicting three levels of risk [53]. Our review results are aligned with the sophisticated data mining model as proposed in determining the risk of cardiovascular disease.
Women's risk assessment for cardiovascular disease is equally important. Traditional atherosclerotic cardiovascular disease (ASCVD) risk factors, such as hypertension, diabetes mellitus, hypercholesterolemia, and obesity, affect both sexes, but some may affect women differently and are considered to be more potent [54]. Although our primary focus is to identify the relationship between OSA and cardiovascular disease, our review described an Australian study that demonstrated the association of gestational hypertension and preeclampsia with OSA [52]. The study also showed that the frequency of sleep-disordered breathing was 1.4 times higher in women with gestational hypertension than controls without hypertension. The severity of sleep-disordered breathing among women with gestational hypertension or pre-eclampsia was more than twice as common compared to controls. These data should be helpful in identifying women at risk and controlling for the cardiovascular risk factors in women.
Moderate-to-severe obstructive sleep apnea syndrome is a common pathology in major depression. The identification of these different risk factors advances a new perspective for more effective screening of moderate to severe obstructive sleep apnea syndrome in major depression. Our study identified psychological comorbidities such as stress and depression, in addition to physical illnesses, in association with OSA. Indeed, in individuals with OSA, the prevalence of depressive affects may reach as high as 63% [55]. The coexistence of major depression and OSA may have a negative impact on quality of life. Further studies are needed to address the increasing morbidity of patients with coexistent OSA and depression.
In a large community-based prevalence study among 2400 adults in India [56], male gender, increasing age, obesity, and increased waist-hip ratio were identified as significant risk factors for OSA. Our study demonstrated the risk factors and prevalence for OSA in multiple countries, which are consistent with earlier findings. More importantly, we also identified a few modifiable risk factors for OSA, such as overweight/obesity, alcoholism, smoking, hypertension, diabetes mellitus, and hyperlipidemia. Given the fact that obesity is a major risk factor for OSA, and given the current global rise in obesity, the prevalence of OSA will increase in the future. More community-based educational interventions are needed for the reduction in the global epidemic of obesity.

Limitations
Due to the nature of the studies, no causal association could be made between OSA and the diseases. There were no clinical trials identified in this review. Although there were 10 prospective cohort studies reviewed, these studies did not report OSA or diseases as an outcome measurement. Future studies are needed to establish whether OSA can cause comorbidities and vice versa. Secondly, this systemic review included multiple structured searches using 10 sources. There is always a possibility that some of the important articles could have been missed. Again, the study did not aim to identify any relationships between specific traits of the population with OSA or any gender differences of OSA, although the study evaluated data obtained from multiple countries (e.g., European, Arab, Asian, American, Australasian) and in both genders. However, studies showed a higher severity of OSA in men [26]. Future studies are needed to establish relationships between OSA severity with traits and gender.

Strengths
This is a comprehensive systematic review comprising a large study population (n = 37,599) from 28 countries. The sources of bias in the literature selection were controlled using the PRISMA method of strict inclusion and exclusion criteria and involving an independent quality appraisal method using all three authors. The study results are generalizable to a wider population of multiple races and ethnicities.

Conclusions and Recommendations
This review provided substantial quantitative evidence to demonstrate the risk factors of OSA. Secondly, this study addressed an important public health problem of cardiovascular disease in association with OSA. The review also showed consistency of findings through multiple country-specific data and the strength of the association by providing evidence of rigorous quantitative analysis regarding the role of OSA in disease morbidity, both physically and psychologically. Reviewed studies consistently showed that severity of OSA, measured by the ACC/AHA risk calculator and by the AHI categories, determine the risk of several disease comorbidities including hypertension, gestational hypertension, diabetes, obesity, heart disease, stroke, asthma, and poor mental health. These results are in line with emerging evidence showing that patients with severe, untreated OSA have a higher incidence of cardiovascular events. Further studies are urgently needed to