Obstructive sleep apnea (OSA) is a public health problem worldwide because of its high prevalence (24% in men, 9% in women aged 30–60) and important associated comorbidities, such as cardiovascular diseases, metabolic syndrome, accidents, and poor quality of life (QoL) [1,2]. Clinical presentations of OSA are varied and non-specific but include classic manifestations such as snoring, witnessed apneas, and excessive daytime sleepiness (EDS). However, OSA can be accompanied by symptoms of insomnia, depression, anxiety, stress, and nasal obstruction, all of which can severely affect sleep quality and functional outcomes [2-4].

The concepts of sleep quality and functional outcomes do not refer to adequate hours of sleep. Instead, they are constructs of aspects and domains that include sleep deprivation, the presence of sleep disorders, medication use, latency and performance for daytime functioning. In short, good sleep quality with adequate functional performance are indicators of a healthy life [5]. The Functional Outcomes Sleep Questionnaire (FOSQ) developed by Weaver et al. [6] is a widely used instrument to quantify the impact of sleepiness secondary to various sleep disorders; measures are designed to assess the impact of excessive sleepiness disorders on multiple activities. Some investigators have found that patients with severe OSA have reduced QoL, FOSQ, and sleep apnea quality of life index (SAQLI); but these findings have not been consistent in the presence of mild to moderate OSA, and, these reductions were primarily explained by sleepiness; furthermore, correlations between questionnaires are modest, indicating that they assess different QoL aspects [5,6].

Symptoms, QoL and functional outcomes, as well as their association with OSA could have variations due to race, and this aspect has not been reported in Mexico. This study aimed to describe the symptoms of OSA, including sleepiness, depression, and sleep quality, and to identify clinical factors associated with poor functional outcomes in OSA patients in southeastern Mexico.

This research was approved by the Centro de Especialidades Médicas del Sureste SA de CV (COFEPRIS 31050063/31050004) ethics committee (#CEM-2018-014), and all patients signed informed consent. A cross-sectional, descriptive study including all adult patients referred to our institution for diagnoses of OSA between October 2018 and July 2019. Briefly, our institute provides respiratory care to patients from public and private entities. It offers 1st- and 2nd-level attention in 7 states in southeastern Mexico (Campeche, Chiapas, Oaxaca, Quintana Roo, Tabasco, Veracruz and Yucatán). Exclusion criteria included incomplete information on the questionnaires and failure to achieve at least 4 hours of adequate sleep monitoring in the sleep study.

All patients referred during this period were examined in their homes using a type 3 monitor (Stardust II, Koninklijke Philips, Amsterdam, Holland) that included the following signals: respiratory flow by nasal pressure, snoring, body position, inspiratory effort by piezoelectric band, pulse oximetry, and heart rate. The sleep study was scored by one of the authors (JLCM) following the recommendations of the American Academy of Sleep Medicine, version 2.5 (2018). Hypopnea was defined as a 30% decrease in respiratory flow compared to baseline respiration with a desaturation ≥ 3% [7].

All patients answered questionnaires and clinical prediction tools to assess symptoms, comorbidities, and functional outcomes; these instruments are versions translated into Mexican Spanish, standardized, and widely used in our country, and some have a Mexican validation. The questionnaires and tools included were: 1) Berlin questionnaire: a clinical prediction tool for OSA based on nocturnal and daytime symptoms, and comorbidities. It has three categories and is considered positive when two are present [8]. 2) FOSQ: This instrument measures the impact of EDS has on patients’ daily functioning. It consists of five domains: activity, vigilance, intimacy/sexuality, general productivity, and social performance, rating the difficulty of performing a given activity on a 4-point scale. The cut-off point for determining significant functional impact is < 89.5 points. The lower the score, the worse the functional performance [9]. 3) Epworth sleepiness scale (ESS): This scale evaluates the presence of EDS. Patients are asked to rate their likelihood of falling asleep in 8 situations or daily activities with minimal stimulation. EDS is defined by a cutoff point > 10 [10]. 4) Nasal obstruction symptom evaluation (NOSE): This tool assesses the impact of nasal obstruction symptoms. The higher the score, the greater the obstruction. The cut-off point was > 15 [11]. 5) Pittsburg Sleep Quality Index (PSQI): An instrument designed to assess sleep quality. It consists of 19 questions in 7 clinical domains: quality, latency, duration, efficiency, sleep disorders, medication use and daytime dysfunction. Poor sleep quality is defined by a cut-off point > 5. The higher the score, the poorer the sleep quality [12]. 6) Insomnia severity index (ISI): This index measures the nature, severity, and impact of insomnia in adults and monitors response to treatment. It consists of 7 sections: sleep onset, sleep maintenance, problems with early morning awakening, sleep dissatisfaction, interference with daytime functioning due to sleep difficulties, perception of sleep problems by other people, and anxiety. The cutoff for defining abnormality is > 8 points [13]. 7) Beck depression inventory (BDI): This instrument measures the severity of depressive symptoms. It consists of 21 sections covering cognitive, affective, somatic, and autonomic symptoms of depression. The abnormaly level is > 10 points (this cut-off point defines abnormality in the Mexican population) [14]. 8) NoSAS: This is a clinical prediction tool that establishes the risk of OSA using the following variables: neck circumference, obesity, snoring, age, and sex. A result > 8 points were considered positive [15]. 9) Sleep apnea clinical score (SACS): This is another clinical prediction tool that can determine the risk of OSA. It is based on neck circumference but considers habitual snoring, witnessed apneas during sleep, and systemic arterial hypertension. Scores < 42, 42–48, and > 48 indicated low, intermediate, and high risk for OSA, respectively [16]. 10) STOP BANG: Also an instrument for clinical prediction used to establish the risk of OSA. It uses dichotomous questions to evaluate the presence of snoring, tiredness, witnessed apneas, systemic arterial hypertension, obesity, age > 50 years, neck circumference > 17 inches, and male gender. A result > 4 means a high risk for OSA [17].

Patients were classified into four groups according to the following apnea hypopnea index (AHI) cut-off points: 1) healthy, AHI < 5 events per hour; 2) mild OSA, AHI 5–14.9 events per hour; 3) moderate OSA, AHI 15–29.9; and 4) severe OSA, ≥ 30 events per hour.

Data were analyzed centrally. The variables were described according to their type and distribution, which was evaluated using the Kolmogorov-Smirnov test. A chi-square test or Fisher’s exact test was used to contrast categorical variables. For the continuous variables, comparisons were made according to their distribution with a Student’s t or Mann–Whitney U test. For comparisons of three or more groups, one-way analysis of variances with Tukey’s post-hoc adjustment were applied. Pearson’s linear correlation was performed to determine the association between continuous variables and the FOSQ questionnaire. Finally, univariate and multivariate logistic regression analyses were performed between questionnaires and clinical variables to determine the risk of poor functional outcomes. All statistical analyses were performed using the SPSS package, ver. 25.0 (IBM Corp., Armonk, NY, USA).

Based on this group of OSA patients living in southeastern Mexico, the following observations can be made: 1) symptoms measured by the NoSAS, SACS, and STOP-BANG scales were greater in patients with severe OSA; 2) AHI was not associated with depression, sleep quality or functional outcomes; 3) functional outcomes were inversely correlated with sleepiness, depression, sleep quality, insomnia, and the SACS, STOP-BANG, and NOSE scales; and 4) sleepiness and depression are independently associated with poor functional outcomes in patients with OSA.

Although questionnaires and prediction scales such as No-SAS, SACS, and STOP-BANG have been designed to measure the risk of OSA compared to the results of polysomnography, they are also clinical tools based on classic symptoms (snoring, witnessed apneas, fatigue) that include such comorbidities as systemic arterial hypertension and obesity. For this group of OSA patients, we found an association between these scales and the AHI, so higher scores may reflect also more severe forms of the disease.

EDS is a prominent symptom in OSA patients. The patients in our study had a mean score of 10.4 ± 6 on the ESS and above half met the cut-off point to define EDS. However, it is essential to note that no differences were found between the three groups, which is consistent with the report by Sharkey et al. [18] for a sample of 269 cases. As in our work, that study did not observe differences associated with AHI severity. However, EDS may be associated with depression, as the combined incidence of these two manifestations in OSA patients is higher than in the general population [19]. Previous work has shown that 17%–48% of OSA patients have mood changes and depression [20-22]. In the present study, 43% of patients met the criteria for depression (> 10 points on the Beck inventory). However, there was no association with the AHI, so it is unclear whether EDS is a clinical phenomenon or the result of overlapping psychosomatic factors that occur separately in both disorders [20]. Identifying these two affections has therapeutic implications, as depression can lead to weight gain due to sleep disturbances, lack of motivation to exercise, and inadequate consumption of calories. Furthermore, there is a risk of deregulation of the serotonergic system, which is affected by episodes of hypoxemia and sleep fragmentation [23].

The patients in this study presented high scores on the NOSE scale (33.3 ± 26.7), and more than 60% were above the cut-off point established as clinically significant. Once again, no association was found with AHI, but it has been suggested that NOSE scale may be related to other symptoms. For example, Ishii et al. [11], who studied 158 patients in an otorhinolaryngology service, found that NOSE scores > 10 points conferred an 88% risk of developing EDS.

Another component evaluated was insomnia, as its co-occurrence with OSA presents challenges for diagnosis and treatment. A meta-analysis by Zhang et al. [24] reported a 38% prevalence of insomnia among OSA patients, which is lower than that observed in our work, in which 74% scored > 8 on the ISI scale (mean 12.2 ± 7.4). This difference can be explained by heterogeneity in diagnosis and ethnic differences. Individuals with insomnia are characterized by hyper-alertness, while OSA patients typically present with EDS. Despite this paradox, these two entities share specific symptoms: fatigue, frequent awakenings, unrefreshing sleep, work/social dysfunction, and concern about sleep quality [25]. Some authors have proposed low awakening threshold and high loop gain as possible mechanisms that foster insomnia in patients with OSA [26].

The PSQI was initially conceived to evaluate patients with insomnia, but its usefulness has extended to other pathologies [27]. This questionnaire is usually carried out in sleep laboratories that serve as reference centers. Several authors have reported weak correlations with several polysomnography parameters, such as sleep latency, efficiency and the non-rapid eye movement sleep stage [28]. Lusic Kalcina et al. [28] evaluated 130 patients with OSA of varying severity. They reported a PSQI of 8.6 ± 3.9, which is very similar to that obtained with our patients (8.8 ± 4.7). In that series, as in the subjects in our work, no correlation was observed with the AHI. Tsai et al. [29] and Lau et al. [30] studied patients with EDS and OSA of different ethnicities. Their findings showed an inverse, significant correlation between the FOSQ-10 and PSQI (r = -0.26 and r = -0.42, respectively), as reported herein (r = -0.42). These results demonstrate the interrelation between poor sleep quality and functional outcomes in OSA patients.

The total mean score of the FOSQ questionnaire for our patients was 82.6 ± 25.4 points, higher than that reported for the sample in Weaver et al. [31] (68 ± 21.2). However, those authors excluded patients with depression and cardiorespiratory comorbidities, which may well explain this difference. Vidal et al. [32] examined a sample of patients with OSA and reported a total value score on the FOSQ similar to our results (88.7 ± 19.8). Interestingly, neither one of those studies found an association with OSA severity measured by AHI [31,32]. The lack of association between AHI and FOSQ was recently reported in Latin America. Thus, de Freitas et al. [33] reported this finding among 100 patients with OSA referred to a sleep medicine unit. This suggests that patients’ perceptions may underestimate their real functional status. However, we must remember that these subjective tools do not accurately detect patients with OSA due to the broad inter-individual variability of symptoms.

The description of functional outcomes by FOSQ questionnaire in subjects without OSA in our language and our region is scarce. To our knowledge, there is only one study by Guglielmi et al. [34] in a sleep unit in Spain, in which the FOSQ questionnaire was compared between 50 patients with OSA (AHI between 6.4 and 119.3 h-1) and 50 healthy subjects, finding a statistically significant difference between the two groups (77.5 vs. 107.1, respectively); this investigation excluded patients with comorbidities and healthy subjects did not have EDS (13 vs. 5.8 in the ESS), which may explain this difference.

Analyzing the association between EDS and functional outcomes, we found an inverse and statistically significant correlation between total ESS scores and FOSQ scale scores (r = -0.57). In this regard, Billings et al. [35], Gooneratne et al. [36], and Weaver et al. [31] also demonstrated an inverse correlation between the FOSQ and ESS tools (r = -0.41, r = -0.6, r = -0.47, respectively) in OSA patients. The consistency of these findings demonstrates a strong association between sleepiness and functional outcomes, despite the conceptual differences in the questions formulated in the two instruments (e.g., for the act of driving a car, the FOSQ inquires for the level of difficulty caused by sleepiness, while the ESS refers to the probability of falling asleep while driving).

The association between depression and poor functional outcomes in OSA patients has been previously reported. In their work, Billings et al. [35] described a statistically significant association between depressive symptoms and the FOSQ. Jackson et al. [21] reported that patients with OSA and depression have lower overall FOSQ scores than patients without depression. In this regard, our cases showed a statistically significant, inverse correlation between the FOSQ and BDI (r = -0.55, p = 0.01), similar to the findings of Kapella et al. [37] (r = -0.37, p = 0.001). Because patients with depression have altered emotional states that affect their productivity, sexual activity, and socialization, low FOSQ scores are expected to reflect this phenomenon. Therefore, this questionnaire makes it possible to obtain a broader picture of depression in OSA patients.

Steffen et al. [38] demonstrated that patients with OSA and insomnia showed an association between functional outcomes and depression through statistically significant negative correlations between the FOSQ and Beck instruments (r = -0.67). In our study group, this correlation was also observed to take this direction (r = -0.56), indicating consistency with data obtained in other research. These findings may have therapeutic implications, as patients with OSA and insomnia may slightly improve their QoL after starting therapy compared with patients without insomnia.

Lam et al. [39] studied 172 OSA cases treated in a continuous positive airway pressure clinic. They reported a significant correlation between NOSE and FOSQ-10 (r = -0.46). The correlation between these two questionnaires in our study was similarly significant (r = -0.25), and the NOSE scale explained 25% of the variation in the FOSQ scores. Notably, the mean NOSE score in Lam et al.’s [39] cohort was 43 ± 26.9, while it was 33 ± 26.7 points in our patients. These differences may underlie the varying prevalences of rhinitis or associated pathologies in distinct populations.

Based on the results obtained, a univariate analysis was performed to establish the determinants of low functional outcomes (FOSQ < 89.5 points). We identified the following: age > 65 years, ESS > 10 points, BDI > 10 points, PSQI > 5 points, ISI > 8 points, and NOSE > 25 points. In multivariate logistic regression analysis and adjusting for gender, age, OSA severity, and comorbidities, we found that only ESS and Beck depression presence were sensitive factors for low FOSQ scores. This finding is interesting because there is evidence that symptoms overlap in OSA patients, such that EDS may be confused as a sign of depression. However, other authors report that the changes in mood and depression disappear when adjusting for covariates such as age, sex, BMI, and hypertension in patients with OSA [40]. The results of this analysis in our study group showed that confounding variables were not decisive in eliminating EDS and depression from the model as explanatory variables of poor functional performance in OAS patients.

Lau et al.’s case-control study [30] with 64 participants reported interesting findings regarding mood and daytime sleepiness. Using the profile of mood state (POMS) questionnaire, those authors extracted somatic from affective symptoms in patients with and without OSA. They found that excessive daytime sleepiness was associated with anxiety, depression, anger, and confusion, which, in turn, are associated with the affective subscales of the questionnaire. This finding suggests that altered moods can be separated from such symptoms as EDS, fatigue, and severity of OSA.

Our study has some limitations. First, there was no control group to compare the results of the questionnaires with non-OSA patients. Second, patients had respiratory and cardiovascular comorbidities, which may have emerged as confounding factors and influenced the scores of many of these questionnaires. Third, the patients were studied with a type 3 portable monitor, which limited the information on the micro and macro architecture of sleep and its possible association with the questionnaires by not including electroencephalogram, electrooculogram, and electromyogram (these variables define sleep states and arrousals); however, the monitor used is a validated and helpful tool for establishing obstructive sleep apnea diagnosis and measuring AHI compared to the gold standard polysomnography. Fourth, four of the ten clinical tools applied have been validated in Mexico (EES, FOSQ, PSQI, and BDI), including our main objectives, and although the remaining instruments have not yet been validated in Mexico, they have been translated, are frequently used and have extensive experience in its use in Mexico. Fifth, patients were referred for sleep studies because of specific symptoms such as drowsiness, for differential diagnoses of insomnia, to be screened for OSA, if they had metabolic syndrome or heart disease, because of those circumstances, there was a certain degree of heterogeneity in the study group.

To the best of our knowledge, this is the first study in Mexico to describe multiple clinical-functional domains in OSA patients systematically. A significant contribution and strength of our work are that it demonstrates the usefulness of questionnaires in diagnosing patients with OAS, depression, and EDS. The relationship between high scores on the Beck inventory and ESS instruments, variables associated with the risk of poor functional outcomes, often explained solely due to the fragmentation of sleep structure by respiratory events, highlights the intricate connection between moods and shared symptoms among the unique pathologies. Based on these results, the authors propose systematic use of depression and FOSQ questionnaires as a way to expand differential diagnoses and contribute to better decision-making in OSA patients.

In conclusion, depression and EDS were risk factors for impaired functional outcomes in OSA patients residing in south-eastern Mexico, whereas AHI was not.