Introduction

Globally, it is estimated that nearly one billion adults have obstructive sleep apnea (OSA), with the highest prevalence in China, followed by the USA, Brazil, and India [1]. OSA is characterized by collapse of the upper airway during sleep, which leads to interruption in breathing, intermittent hypoxemia, and frequent arousals. It is estimated that 82% of men and 93% of women with moderate to severe sleep apnea are undiagnosed in the general population [2]. A simple screening tool such as the STOP questionnaire (snoring, tiredness, observed apnea, blood pressure) is useful to identify individuals at risk of sleep apnea before laboratory polysomnography or home sleep apnea testing [3, 4].

Since January 2020, the coronavirus disease 2019 (COVID-19) pandemic has significantly impacted global health and individual well-being. The severe acute respiratory syndrome coronavirus-2 virus is highly contagious and causes fever, cough, fatigue, respiratory distress, and death. Patients with OSA have a significantly increased risk of severe COVID-19, as well as hospitalization and mortality due to COVID-19 [5,6,7,8,9,10,11,12,13]. OSA has been hypothesized to increase COVID-19 severity and risk of death through proinflammatory pathways [13, 14]. However, the reported low levels of diagnosed OSA in an obese population with COVID-19 could reflect under-diagnosis of the disorder [15]. Since OSA is an emerging key risk factor for increased severity of COVID-19, screening for OSA may optimize the triage process in COVID-19 patients [5].

Previous studies on the association between diagnosed OSA and COVID-19 have so far examined hospitalized patients in a single country or retrospectively examined electronic medical records of COVID-19-hospitalized patients with OSA [5,6,7,8,9,10,11,12,13]. To the best of our knowledge, there has been no study on OSA, or at high risk of OSA, and COVID-19 in the general population. Although OSA increases the risk of severe COVID-19, the level of potential modulation has not yet been established. The objective of our study was to determine the associations between high risk of OSA, comorbidities, and increased risk of being afflicted by COVID-19, as well as the risk of hospitalization or intensive care unit (ICU) treatment in a large sample of participants from 14 different countries/regions. We hypothesized that there would be associations between those at high risk of OSA—as well as its comorbidities—with greater risk of being diagnosed with COVID-19 and increased risk of hospitalization or requiring ICU treatment.

Methods

Global survey during the COVID-19 pandemic

The research protocol and the final standardized survey questionnaire have been published previously [16]. All countries/regions obtained ethical approval or exemptions in keeping with national research governance and regulations. The cross-sectional survey data were collected online in each country/region in their native language (translated and then back translated) between May and August 2020. It was administered in 14 countries/regions (Austria, Brazil, Canada, China/Hong Kong, China/Jinlin, Finland, France, Germany, Italy, Japan, Norway, Sweden, UK, and USA).

The International COVID-19 Sleep Study (ICOSS) included sociodemographic variables such as age, gender, marital status, and comorbidities. Participants were asked questions about having been diagnosed with COVID-19, and whether they were hospitalized or treated in the ICU for COVID-19. To investigate sleep problems, standardized and validated questionnaires were included such as Basic Nordic Sleep Questionnaire and STOP questionnaire (snoring, tiredness, observed apnea, pressure) and Patient Health Questionnaire (PHQ)-4 for anxiety and depression [3, 17, 18]. The STOP and PHQ questionnaires have been validated for use in diverse geographical populations [4, 19, 20].

The survey was administered online by sharing a link on national newspapers, social media sites, and university/hospital webpages. Participants aged 18 years or older anonymously and voluntarily took part in the self-administered online survey. The most commonly used survey platforms for administration were Redcap and Qualtrics.

High risk of OSA (STOP questionnaire) and comorbidities

The STOP questionnaire is a screening questionnaire that consists of four yes/no questions on snoring, tiredness, observed apnea, and high blood pressure. Answering yes to at least two of these questions had been shown to be an effective screening tool for high-risk OSA. It has been validated by lab-polysomnography [3, 4]. All respondents included in the analysis completed the STOP questionnaire. In the ICOSS survey, instead of answering yes or no, participants provided ratings on 5-point Likert scales for snoring, tiredness, observed apnea (never, less than once per week, 1–2 days per week, 3–5 days per week, and daily or almost daily), and blood pressure (yes/no) [3]. The questions were as follows: (1) Do you snore loudly (louder than talking or loud enough to be heard through closed doors)? (2) Do you often feel tired, fatigued, or sleepy during daytime? (3) Has anyone observed you stop breathing or choking during your sleep? (4) Do you have or are you being treated for high blood pressure? The first three questions were dichotomized into 0 = less than three nights per week versus 1 = three or greater than three nights per week. The highest possible STOP score was four. STOP scores of two or greater were classified as high risk and a score of zero to one as low-risk of OSA [3]. Participants were asked (yes or no questions) whether they suffered from any comorbidities eliciting reports of physician-diagnosed diseases such as sleep apnea, cardiovascular disorder, hypertension, cerebrovascular disorder, diabetes, asthma, chronic obstructive lung diseases (COPD), depression, anxiety, insomnia, other sleep disorder, rheumatic disorder, autoimmune disease, chronic fatigue syndrome, attention deficit hyperactivity disorder, cancer, allergy, and other neurological disorder.

Statistical analysis

The analyses were performed using STATA/SE 14.2, and participants’ characteristics were summarized using mean (± standard deviation) scores or percentages (frequency counts). An independent sample t-test or chi-square was conducted to investigate potential differences in sociodemographics of participants with high risk of OSA vs low risk of OSA and OSA vs no-OSA. Multivariable logistic regression analyses, with 95% confidence intervals (CI), were conducted to examine the association between high risk of OSA and risk of COVID-19, hospitalization, and ICU treatment, adjusting for BMI, age, gender, marital status, education, ethnicity, financial suffering, confinement, and other comorbidities. We chose the variables for models based on clinical importance and their association with the dependent variable in the unadjusted analysis. Some analyses were stratified and weighted by countries. A P-value less than 0.05 was considered statistically significant (2-sided).

Results

A total of 26,539 participants responded to the survey, of whom 20,598 (65% F) provided complete data on both STOP questionnaire and the presence of comorbidities and were included in the analysis. The sample had a mean age of 41.5 ± 16.0 years and a BMI of 24.0 ± 5.0 kg/m2. Fifty-five percent were cohabiting/married, and 64% had bachelors or higher-level education. The most commonly reported comorbidities were hypertension (11%), cardiovascular diseases (CVD) (4%), diabetes (4%), OSA (4%), depression (8%), and insomnia (7%) (Table 1). In total, 9.5% had a STOP score two or greater and were deemed to be at high risk of OSA. Eight percent of participants reported scores in keeping with high risk of OSA but did not indicate having OSA. There was a significant difference in the sociodemographics in high-risk vs low-risk OSA groups and in OSA vs non-OSA (Table 1). Three percent (622/20,598) reported having had COVID-19, of whom 56% had been confirmed by laboratory testing.

Table 1 Sociodemographics characteristics of participants
Full size table

Characteristics and risk factors of participants with COVID-19

To examine the association between being at high risk of OSA and a reported diagnosis of COVID-19, we adjusted for participants’ characteristics and comorbidities. Participants with higher BMI (>35 kg/m2) (adjusted odds ratio (aOR) 1.91, 95% CI: 1.08, 3.40), Hispanics ethnicity (aOR 1.55, 95% CI: 1.05, 2.30), high risk of OSA (aOR 1.72, 95% CI: 1.20, 2.47), and diabetes (aOR 2.07, 95% CI: 1.23, 3.48) had higher odds of reporting a COVID-19 diagnosis (Table 2). The unadjusted analyses are shown in eTable 1.

Table 2 Characteristics and risk factors of participants with COVID-19 infection
Full size table

Risk factors associated with increased hospitalization and ICU treatment

In total, 622 participants reported having had COVID-19 infection, of whom 536 reported the disease severity level as follows: 42% reported having mild, 43% moderate, 12.5% severe, and 1.9% life-threatening COVID infection symptoms. Of 104 participants who were hospitalized, 86.5% were admitted to hospital ward, while 13.5% were treated in ICU. Of 431 participants who were not hospitalized, 77.3% were treated at home, while 22.7% required no treatment.

To examine if there was an association between high risk of OSA (based on STOP questionnaire) and increased hospitalization for COVID-19, we adjusted for BMI, age, gender, and comorbidities. This analysis was not weighted and stratified due to limited and inconsistent number of responses across countries/regions. High risk of OSA, being male, having diabetes, and having depression were associated with increased hospitalization and ICU treatment related to COVID-19. Participants at high risk of OSA (aOR 2.11, 95% CI: 1.10–4.01) and having depression (aOR: 2.33, 95% CI: 1.15–4.77) were at twofold increased odds of hospitalization and ICU treatment for COVID-19 versus low risk of OSA and no depression. Male sex (aOR: 2.82, 95% CI: 1.55–5.12) and diabetes (aOR: 3.93, 95% CI: 1.70–9.12) had three times higher odds of being hospitalized and receiving ICU care for COVID-19, compared with female and no diabetes (Table 3). The unadjusted analyses are shown in eTable 2. Figure 1 displays probability values for the significant variables.

Table 3 Risk factors of participants with increased hospitalization and ICU treatment
Full size table
Fig. 1
figure 1

Predicted probabilities of hospitalization/ICU related to COVID for gender (a), STOP (risk of OSA) (b), diabetes (c), and depression (d). Vertical lines show 95% confidence interval.

Full size image

Sleep problems and OSA (high risk; reported)

During the time of confinement of COVID-19 pandemic, we found that the prevalence of self-reported sleep problems such as poor sleep quality, sleep onset, sleep maintenance, early morning awakening problems, nightmares, hypnotic use, fatigue, and excessive sleepiness were significantly increased in participants with risk of OSA (high risk vs low risk) and for OSA vs no-OSA (Fig. 2).

Fig. 2
figure 2

Prevalence of sleep problems during COVID-19 pandemic for risk of OSA (a) and OSA (b). Vertical lines show standard error. High risk of OSA: STOP score 2 or greater. Low risk of OSA: STOP score less than 2. Chi-square analyses showed that there was a significant difference in the prevalence of risk of OSA (high vs low) and OSA (yes vs no) for all sleep problems during pandemic (P<0.001). OSA, obstructive sleep apnea; ICU, intensive care unit

Full size image

To examine the association between sleep problems during the pandemic and high risk of OSA as well as OSA, we adjusted for gender, age, BMI, marital status, ethnicity, education, presence of sleep problem before COVID-19 pandemic, duration of COVID-19 confinement, financial suffering, comorbidities, and the severity of the COVID-19 epidemic in each country. Analyses were weighted and stratified by countries/regions (eTable 3). There were a significant association between high risk of OSA and all sleep problems. For example, high risk of OSA was most strongly associated with largest increase for excessive sleepiness (aOR 2.42, 95% CI: 1.90–3.09). Additionally, there was an association between some sleep problems and OSA (eTable 3).

Discussion

We found that participants with high risk of OSA and diabetes had higher odds of becoming afflicted by COVID-19. Additionally, our study showed that participants with the following characteristics—high risk of OSA, being male, having diabetes, and having depression—were two to three times more likely to have been hospitalized or to require ICU treatment due to COVID-19. To the best of our knowledge, our study is the first in the literature focusing on hitherto undiagnosed OSA by using the STOP questionnaire and the risk of COVID-19, whereas previous studies examined COVID-19-hospitalized patients with physician-diagnosed OSA [5,6,7,8,9,10,11,12,13].

The common risk factors for poor outcomes of COVID-19 are older age, hypertension, and diabetes [21, 22], all of which are prevalent or associated with OSA. Clinical recognition of OSA is markedly underdiagnosed worldwide. Previous studies which reported OSA in COVID-19 patients recruited hospitalized patients, usually from one area, with no comparison to participants who had not been afflicted by COVID-19. A few retrospective studies, which examined the association between diagnosed OSA and COVID-19 in electronic medical databases, identified OSA as a risk factor for COVID-19 severity with higher ICU admission and mortality [5,6,7,8,9,10,11,12,13]. Screening patients for OSA is being recommended to aid in decisions for COVID-19 treatment [5, 7]. Our study is novel as we examined the risk of COVID-19 hospitalization and ICU treatment in individuals with high risk of OSA—using the STOP questionnaire, from a diverse and global general population of over 20,000 individuals.

The proinflammatory status of OSA may enhance the typical COVID-19 cytokine storm, thus worsening disease evolution [15]. Interestingly, we observed a significant association between high risk of OSA (STOP score 2 or higher) and increased COVID-19 hospitalization and ICU treatment. In contrast to the literature, we did not observe a significant association between physician-diagnosed OSA and worse outcomes in the general population. Many individuals at high risk of OSA are presumably undiagnosed and have not received treatment, whereas we hypothesized that those who reported physician-diagnosed OSA in our study might have received prehospital and/or in-hospital benefit of treatment such as continuous positive airway pressure (CPAP) possibly accounting for better outcomes. CPAP therapy decreases the underlying proinflammatory conditions which may help manage COVID-19 symptoms by reducing upper airway and systemic oxidative stress [7, 23]. COVID-19 patients with OSA treated with CPAP prescription appeared to have better outcomes in a study from electronic health records of the New England Health Care System [7].

Among patients with acute respiratory distress syndrome (ARDS) due to COVID-19, undiagnosed SDB was independently associated with ARDS [10]. It is difficult to know the true prevalence of OSA among patients with COVID-19 as diagnostic sleep studies with polysomnography are not feasible at the time of COVID-19 infection. Since our study shows that participants at high risk of OSA are about two times more likely to need hospitalization or ICU care due to COVID-19, therefore, we recommend screening for OSA in order to enhance the COVID-19 triage process. The STOP questionnaire has been validated as an effective tool for OSA screening [3, 4], and its use could aid in predicting COVID-19 severity and the need for more intensive in-hospital treatment.

Consistent with the literature, we found that individuals with diabetes had an increased (three times) odds of being hospitalized for COVID-19. Two meta-analyses found that diabetes was the second most prevalent comorbidity in COVID-19 patients [24, 25]. Patients with diabetes are often treated with angiotensin converting enzyme (ACE) inhibitors, which result in upregulation of the protein ACE2 [26]. The SARS-CoV-2 virus uses the enzyme as its main receptor to gain entry into the host cells [27]. Individuals with high risk of diabetes are thus potentially more likely to experience COVID-19 severe symptoms due to the increased expression of ACE2.

We showed that males were at three-fold increased odds of being hospitalized for COVID infection. Our findings are consistent with the previous studies which showed an increased risk of hospitalization and mortality in males compared to females [21, 22, 28, 29].

There are high rates of depression in individuals with OSA in both community and clinical populations [30]. Up to 20% of those presenting with a diagnosed depressive syndrome may have OSA, and vice versa [31]. In our survey, we asked participants whether they have a physician-diagnosed depression. Importantly, we found that participants who reported depression had almost two-fold increased odds of hospitalization due to COVID-19. Our findings are supported by two recent studies. Wang et al. analyzed a US-wide database of electronic health records of 61 million patients. Patients with depression had over 7-folds increased risk for COVID-19 with a higher rate of death and hospitalization [32]. Also, Atkins et al. reported that in 269,070 older adults, a pre-existing diagnosis of depression emerged as an independent risk factor for COVID-19 hospitalization [21]. However, it is plausible that depression may be a consequence of COVID-19 as a result of having been ill or critically ill for a substantial period. Further research is needed to validate if COVID-19 hospitalization is due to pre-existing depression itself or secondary to other conditions.

Previous studies found age, hypertension, and COPD to be risk factors for mortality and hospitalization in COVID-19-hospitalized patients [21, 22, 28, 33], but we did not find an association for these variables. However, our study involved general population using an online platform. The participants in our study were younger, with fewer comorbidities, such as hypertension and COPD, than previous studies on COVID-19-hospitalized patients [5,6,7,8,9,10,11,12,13].

We found that all types of sleep problems such as poor sleep quality, sleep onset and maintenance problem, daytime fatigue, and sleepiness were associated with high risk vs low risk of OSA. OSA is characterized by intermittent complete and partial airway collapse, which results in sleep problems including frequent arousals, disruptive snoring, breathing pauses, and co-occurrence of insomnia [34]. More sleep problems were recorded in participants who were at high risk of OSA based on the STOP questionnaire, compared with those with OSA, which further implies that patients with OSA might have received treatment that was beneficial to their symptoms.

Limitations

There are several methodological limitations due to the nature of survey-based data collection. Potential inaccuracy from reported data means that we cannot be certain all participants answered each survey question to the best of their ability or knowledge. We conducted the survey using web applications, which only allowed individuals with access to the internet to participate, thus potentially limiting the generalizability of our results. Variations in sample sizes and frame of sampling posed another limitation and were relative to each different country. These variations were addressed using weighting in some analyses.

Another major limitation is that we used the STOP questionnaire to screen participants for high risk of OSA. Although the STOP questionnaire is a validated screening tool in the population with pulmonary diseases [35, 36], it has not been validated to use in patients with COVID-19. It is possible that an inflammatory state such as COVID-19 may produce enough upper airway inflammation to provoke snoring. Dyspnea associated with bilateral COVID-19 pneumonia could be misinterpreted as choking during sleep. These may affect the reliability of the STOP questionnaire in this setting. Hence, our results should be interpreted with caution. Nevertheless, our findings are important as we demonstrated the association between those at high risk of OSA and COVID-19 in the general public in a large dataset. Our results epitomize a large and diverse global population representing different ethnicities and a broad age range.

Conclusion

In summary, participants at high risk of OSA had higher odds of reporting having been diagnosed with COVID-19. Furthermore, participants at high risk of OSA were two times more likely to have been treated in a hospital or ICU. Being male, having diabetes, and having depression were also associated with increased risk of hospitalization and ICU treatment. Our study used the STOP questionnaire as a screening tool in the general population to identify individuals with a high risk of OSA. Identifying those at high risk of OSA by screening may enhance the COVID-19 triage process to optimize treatment.