Classic cardiovascular risk factors do not explain all the cardiovascular events that occur,1,2 and it is possible that other unknown cardiovascular risk factors may exist. In this sense, obstructive sleep apnoea (OSA) has been proposed as a new potential and prevalent cardiovascular risk factor.3,4

OSA is characterized by partial or complete episodes of collapse of the upper airway during sleep that cause intermittent hypoxia, hypercapnia, sleep fragmentation and exaggerated fluctuations in heart rhythm, blood pressure and intrathoracic pressure.5,6 In turn, these physiological disruptions can result in excessive daytime sleepiness, poor quality of life and cardiometabolic alterations.5,6

Several studies associated OSA to cardiovascular risk factors and support that OSA contributes to the pathogenesis of cardiovascular disease.3 Intermittent hypoxia and intrathoracic pressure swings have been associated with the initiation and progression of atherosclerosis.7,8 In contrast, some authors suggest that OSA may require other risk factors for the development of atherosclerotic plaques.9 Therefore, the impact of OSA on atherosclerosis and its interaction with other cardiovascular risk factors is not well known.

A higher prevalence of OSA has been described in individuals with cardiovascular conditions than in healthy adults10; furthermore OSA prevalence in cohorts with cardiovascular risk factors has triggered great interest.11

OSA and snoring have been associated with atherosclerotic plaque, especially in the carotid12,13; nevertheless, there are few studies with controversial data, and evidence is limited.

In this context, the current study aimed to describe the prevalence of OSA in a cohort of middle-aged subjects with mild–moderate cardiovascular risk and evaluate whether there is an association between polygraphy variables and the presence of atherosclerosis. Considering the controversial role of snoring on carotid but not femoral atherosclerosis, both territories were explored to assess potential differences.

The present study confirms a high prevalence of OSA in subjects with mild–moderate cardiovascular risk and shows an association between the presence of atherosclerotic disease and parameters of nocturnal oxygen saturation.

To the best of our knowledge, our study is the first to evaluate the prevalence of OSA in a well-characterized cohort of middle-aged subjects with mild–moderate cardiovascular risk. Our findings show that the overall prevalence of OSA is 72.6%, being the prevalence of mild OSA 35.7%, and the prevalence of moderate to severe OSA 36.9%. OSA prevalence in our study is higher than previously reported in the general population16–19; nevertheless, it is difficult to compare among studies because of different types of sleep studies and different criteria to score and define OSA.

To date, a wide range of OSA prevalence has been reported, even when the same AHI cut-off point was used to define OSA. We observe that in a study performed in Hispanic subjects, the total prevalence of OSA was 25.8%, of which 9.8% for moderate OSA and 3.9% for severe OSA.18 Peppard et al.17 reported an estimated prevalence of OSA of 36.6% and 20.2% in men and women, respectively. Considering that age range and BMI were similar to our cohort, the prevalence observed in our study is higher than that described by Peppard et al.17 for both sexes. However, our prevalence is lower than reported by Heinzer.20 Differences could be explained by the use of polysomnography and the percentage of symptomatic subjects included in their study. Therefore, compared with studies using respiratory polygraphy, our data show a higher prevalence of OSA than previously reported in the general middle-aged population and suggest the importance of assessing OSA in subjects with at least one cardiovascular risk factor, even if they are asymptomatic.

An association of OSA with the risk of hypertension and obesity was found, as previously reported.21 Moreover, we observed a low probability of being a smoker with moderate to severe OSA. This result could be in line with previous studies concluding that the association between smoking and OSA has not been clearly stablished.22 Moreover, it is important to note that in our cohort, 42% of smoking patients present tobacco as the only risk factor while non-smoking patients were included in the study for presenting other cardiovascular risk factors and therefore we cannot exclude a selection bias that could have influenced the results.

Atherosclerotic burden, total and femoral plaque were associated with CT90 and mean oxygen saturation but not with AHI, minimal oxygen saturation or ODI. Our results are in line with previous studies linking mean oxygen saturation and CT90 with subclinical atherosclerosis.12,23–27 Unlike other studies,13,28,29 we did not find a relationship between AHI and the presence of plaque or the atherosclerotic burden in any territory. Our results are consistent with previous studies describing an association between atherosclerotic disease and OSA-related hypoxaemia parameters without an association with the AHI26 and indicate, as previously described, that atherosclerosis could be more strongly associated with hypoxaemia than with episodes of obstruction (represented by AHI).12,23,24 CT90 has been described as the strongest predictor of subclinical atherosclerosis25 and several observational studies showed that CT90 and mean SaO2 better predict cardiovascular mortality than the AHI.30,31

Moreover, our results are in line with recent publications indicating that the assessment of the severity of OSA only with the AHI could be insufficient as it measures the frequency of respiratory events but does not include key aspects of the deleterious effect of OSA on the cardiovascular system such as the duration and severity of the gasometrical changes. In this sense, Azarbarzin et al. have recently shown that the “hypoxic burden”, a new measure of OSA severity related to oxygenation changes associated with the respiratory events, could be more relevant to assess the relationship between OSA and cardiovascular disease than the AHI.32

It is important to note that CT90 can characterize not only nocturnal hypoxaemia associated with obstructive events but also persistent hypoxaemia that could be related to other conditions such as chronic respiratory diseases. In our cohort, CT90 values increase progressively as the severity of OSA (established by the AHI) increases. This fact suggests that CT90 reflects hypoxaemia related to OSA and not due to other underlying respiratory diseases such as COPD. Moreover, subjects included in the study presented spirometry values within the normal range. The mean CT90 value in our cohort was 8%. Even though it is not a high value, an association between CT90 and markers of atherosclerotic disease25 and CV mortality33 have been described with values of lower magnitude.

Differences between arterial territories were observed. An association between nocturnal hypoxaemia and atherosclerosis in femoral arteries was reported, while, no association between carotid territory plaques and any polygraphy variables was found. Our results are consistent with studies showing an association between OSA and peripheral arterial disease34 and are in line with previous studies not showing an association between OSA and carotid atheromatosis; territory in which controversy on the effect of OSA exists.28,35,36

Unlike our study, authors usually do not evaluate the presence of atherosclerosis in different territories and therefore evidence on the chronology and the territory mostly affected in OSA is scare. In studies assessing carotid and femoral territories in general population of similar characteristics than our patients, it has been described that femoral atherosclerosis could be an early marker of subclinical atherosclerosis37 and therefore it is possible that OSA related-hypoxia may have a greater impact on the territory with earlier affectation.

Multiple factors intervene in the generation of atherosclerosis and their interaction with OSA related hypoxaemia probably could influence differently the atherosclerotic process in carotid and femoral territories. The risk factors for atherosclerosis in the two areas (carotid and femoral) are similar,38,39 although peripheral arterial disease seems to be more dependent on smoking habits38 than carotid disease. Therefore, with the results, we can hypothesize that hypoxia may have a greater synergistic effect to promote atherosclerosis when interacting with the risk factors for peripheral arterial disease such as tobacco. In contrast with previous data,35 our results suggest that vibration produced by snoring may not have an important role in the generation of carotid plaque.

The main strength of our study is that it was performed in a large and well-characterized cohort of subjects with mild–moderate cardiovascular risk in which both arterial territories were evaluated.

Some limitations should be mentioned. First, the cross-sectional design limits causal inferences. Second, history of cardiovascular risk factors was not assessed at the same time than the sleep study and it could make associations less robust and difficult to establish causality, therefore the results should be extrapolated with caution. Third, subjects included are of a specific age group and the results cannot be extrapolated to other age groups. Additionally, age range was slightly different according to sex, making difficult to compare both groups. Forth, OSA diagnosis was performed using a simplified method; therefore, OSA prevalence can be underestimated and subjects misclassified. In addition, this simplified method does not provide information related to sleep structure, sleep fragmentation and arousal index, factors that have been related to atherosclerosis

In conclusion, our study shows a high prevalence of OSA in patients with mild–moderate cardiovascular risk. Moreover, the study shows an association between the atherosclerotic burden, total and femoral plaque with CT90 and mean oxygen saturation, suggesting the importance of OSA-related hypoxaemia in the induction of atherosclerotic disease.

Our data encourage clinicians to perform an energetic intervention to diagnose and control OSA in patients with cardiovascular risk factors. Future randomized clinical trials performed in well-characterized subjects according to their atheromatous risk are needed to evaluate if treating OSA could be beneficial and establish if subjects with subclinical arterial disease could be a target population.

This work was supported by Diputació de Lleida, Instituto de Salud Carlos III [RETIC RD16/0009/0011]; Ministerio de Ciencia, Innovación y Universidades [IJC2018-037792-I]; Fondos recibidos por el ISCIII y fondos FEDER “Una manera de hacer Europa” [PI19/00907]; Oxigen Salud; Resmed; Jordi de Batlle acknowledges receiving financial support from Instituto de Salud Carlos III [ISCIII; Miguel Servet 2019: CP19/00108] and co-funded by European Regional Development Fund (ERDF)/European Social Fund (ESF), “Investing in your future”.

All authors declare that there is no conflict of interest