Sleep-Disordered Breathing in Adults with Precapillary Pulmonary Hypertension: Prevalence and Predictors of Nocturnal Hypoxemia

To evaluate the frequency of sleep-disordered breathing (SDB) and predictors of the presence of nocturnal desaturation in adults with pulmonary arterial hypertension and chronic thromboembolic pulmonary hypertension. Outpatients with a hemodynamic diagnosis of precapillary pulmonary hypertension who underwent portable polysomnography were evaluated. Diagnosis and severity of SDB were assessed using three well-established respiratory disturbance index (RDI) thresholds: 5.0/h, 15.0/h, and 30.0/h, while nocturnal hypoxemia was defined by the average oxygen saturation (SpO2) < 90%. Multiple linear regression analysis evaluated the potential relationships among explanatory variables with the dependent variable (average SpO2 values), with comparisons based on the standardized regression coefficient (β). The R-squared (R2; coefficient of determination) was used to evaluate the goodness-of-fit measure for the linear regression model. Thirty-six adults were evaluated (69.4% females). The majority of the participants (75.0%) had SDB (26 with obstructive sleep apnea [OSA] and one with central sleep apnea [CSA]); while 50% of them had nocturnal hypoxemia. In the linear regression model (R2 = 0.391), the mean pulmonary artery pressure [mPAP] (β − 0.668; p = 0.030) emerged as the only independent parameter of the average SpO2. Our study found that the majority of the participants had some type of SDB with a marked predominance of OSA over CSA, while half of them had nocturnal desaturation. Neither clinical and hemodynamic parameters nor the RDI was a predictor of nocturnal desaturation, except for mPAP measured during a right heart catheterization, which emerged as the only independent and significant predictor of average SpO2.


Introduction
Pulmonary hypertension (PH) is a serious hemodynamic condition that can co-develop in several diseases or can be idiopathic and is associated with high morbidity and mortality [1,2]. The presence of PH is usually confirmed by performing right heart catheterization (RHC), with a mean pulmonary artery pressure (mPAP) > 20 mmHg having been recently adopted as the diagnostic criterion [3]. According to the pathophysiological mechanism, PH is currently classified into 5 sub-groups: Group 1: Pulmonary arterial hypertension (PAH); Group 2: PH associated with left heart disease; Group 3: PH associated with pulmonary and/or hypoxemic diseases; Group 4: PH associated with pulmonary artery obstruction, and Group 5: PH due to an unknown or multifactorial mechanism [3]. PAH and chronic thromboembolic pulmonary hypertension (CTEPH) reflect the overall vascular responses aiming to overcome the underlying high resistance within the pulmonary circulation and maintain right ventricular cardiac output [3].
Obstructive sleep apnea (OSA) is the most frequent presentation of sleep-disordered breathing (SDB) and is characterized by the collapse of the upper airway during sleep, resulting in intermittent hypoxemia, increased respiratory effort, and sleep fragmentation [4]. There is growing evidence that OSA plays an important role in the pathogenesis of cardiovascular and metabolic diseases [5,6], and of increased overall and cardiovascular mortality [7,8], along with increased healthcare costs [9,10]. A Brazilian study conducted in the city of São Paulo that included 1,042 volunteers who were representative of the city population showed that 32.8% suffered from OSA [11]. More recent estimates suggest that nearly one billion people may be affected worldwide, with around 425 million individuals aged between 30 and 69 years suffering from moderate-to-severe OSA [12]. The gold standard for the diagnosis of OSA consists of an inlaboratory polysomnographic test (PSG); however, PSGs are not widely available for a large number of individuals with suspected OSA in countries with limited financial resources, such as Brazil [4].
Despite limited evidence to date, SDB is probably more prevalent in patients with PAH and CTEPH than in the general population [13][14][15][16][17][18][19]. Interestingly, patients with idiopathic PAH (Group 1) are usually women and are not necessarily obese [20], and such demographic characteristics fundamentally differ from the usual patients being referred to a sleep laboratory with suspected OSA [21]. Such discrepancies may therefore hamper the clinical recognition of PAH patients at risk of SDB, and reduce the screening of SDB among PAH patients [22]. Since SDB (including OSA) leads to intermittent nocturnal hypoxemia, increased intrathoracic pressure swings with associated hemodynamic changes, and increased sympathetic activity, the presence of SDB can further compromise cardiovascular function in PAH patients [23]. A previous study showed that the presence of nocturnal hypoxemia, but not the occurrence of SDB, was associated with poorer survival in patients with PAH [20]. In addition, routine treatment with continuous positive airway pressure (CPAP) among PH patients suffering from SDB resulted in a beneficial, albeit modest reduction in pulmonary arterial pressure levels [23][24][25].
Considering that studies evaluating the prevalence of SDB in adults with PAH and CTEPH are surprisingly scarce, in addition to the fact that, to date, there are no studies in Brazil evaluating the prevalence of SDB in these patients, our main objectives were to estimate the prevalence of SDB in adult outpatients with precapillary PH diagnosis (Groups 1 or 4) and to evaluate predictors of nocturnal hypoxemia in this population.

Methods
A cross-sectional study was carried out at the Instituto de Doenças do Tórax of the Universidade Federal do Rio de Janeiro from June 2021 to December 2021. Inclusion criteria were adult individuals (age ≥ 18 years) with a hemodynamic diagnosis of precapillary PH (Groups 1 or 4) who underwent a home sleep apnea test [HSAT] in a single night. Subjects were excluded for any of the following reasons: previously diagnosed OSA receiving CPAP treatment, technically inadequate HSAT (total recording time < 240 min), nocturnal use of supplemental oxygen, and any individual who did not sign the informed consent form. In addition, eligible patients who were clinically unstable (New York Heart Association class IV) or had been hospitalized in the month preceding the HSAT were also excluded.
Clinical, demographic, and anthropometric data were systematically collected by one researcher (M.S.M.) using a standardized form on the same day of the HSAT. Subject characteristics included clinical data and self-reported comorbidities (hypertension and excessive daytime sleepiness [EDS]). Body mass index (BMI) was calculated by dividing the weight in kilograms by the square of the height in meters (kg/m 2 ). Neck circumference (NC) was measured using a 150-cm tape measure, with all participants in the sitting position, with the upper edge of the measuring tape positioned below the laryngeal prominence and applied perpendicularly to the long axis of the neck. Epworth Sleepiness Scale (ESS), an 8-item questionnaire that assesses the subjective likelihood of falling asleep in various settings was completed: a score ≥ 11 (from 0 to 24 points) was considered indicative of the presence of EDS [26]. The ESS has already been validated for use in the Brazilian population [27].
The research project was conducted in accordance with the Declaration of Helsinki and was previously approved by the Ethics Committee of the Universidade Federal do Rio de Janeiro (CAAE: 44429421.5.0000.5257). All study participants gave written informed consent before study enrollment, and the anonymity of each participant was strictly preserved.

Right Heart Catheterization
A hemodynamic evaluation was performed on all study participants using a Swan-Ganz catheter. The hemodynamic variables collected and used in data analysis included pulmonary capillary wedge pressure (PCWP), mPAP, cardiac output, and pulmonary vascular resistance (PVR). The diagnosis of precapillary PH was based on the finding of mPAP > 20 mm Hg measured directly, in absence of elevated PCWP (PCWP ≤ 15 mm Hg), and with PVR ≥ 3 Wood units [3].

Sleep Test
All participants underwent an HSAT (ApneaLink TM AIR, ResMed, San Diego, California, USA) after careful instructions regarding the application of the portable monitor.
The HSAT consists of continuous monitoring of airflow through a nasal cannula, thoracic impedance belt, and pulse oximeter. All recording data were manually scored by a board-certified sleep physician (R.L.M.D.)-who was blinded to all clinical data collected by another researcher (M.S.M)-according to a previously published guideline by the American Academy of Sleep Medicine [28]. Obstructive apneas were classified from a drop of ≥ 90% from baseline in airflow with a persistent respiratory effort lasting at least 10 s. Hypopneas were classified from a drop of ≥ 30% from baseline in airflow for ≥ 10 s, which were associated with an oxygen desaturation of ≥ 4% [28]. The severity stratification of SDB was defined by three well-established respiratory disturbance index (RDI) thresholds, namely ≥ 5.0/h (any SDB), ≥ 15.0/h (moderate-to-severe SDB), or 30.0/h (severe SDB). In turn, each SDB was classified as OSA, i.e., the predominance of obstructive events over central ones, or as central sleep apnea (CSA) when there was a predominance of central respiratory events in relation to obstructive events [28]. Nocturnal hypoxemia was defined by average oxygen saturation (SpO 2 ) of less than 90% [20] obtained from continuous records of the pulse oximetry existing in the HSAT.

Statistical Analysis
Data analysis was conducted using the SPSS for Windows statistical software package (version 21.0; Chicago, IL, USA). Results were summarized as median and interquartile range (IQR) for continuous variables and as frequencies (n) and respective percentages (%) for categorical variables. Comparisons were performed using the chi-square test or Fisher's exact tests for categorical variables and the nonparametric Mann-Whitney or Kruskal-Wallis tests for continuous variables. The correlation among continuous variables was evaluated by Spearman's rank correlation coefficient (r). The presence of collinearity among the continuous variables was excluded if the variance inflation factor (VIF) > 10.0 with tolerance < 0.1. After excluding collinearity, continuous variables that were not normally distributed were logtransformed before being evaluated in the linear regression model. Multiple linear regression analysis evaluated the relationship among possible explanatory variables and the dependent variable (average SpO 2 ), with comparisons based on the standardized regression coefficient (β). The R-squared (R 2 ; coefficient of determination) was employed as a goodness-of-fit measure for the linear regression model. All two-tailed tests were considered statistically significant if the p value < 0.05.

Results
Thirty-six adult subjects were consecutively recruited and included in the final analysis (the flowchart of the study is illustrated in Fig. 1). Among the study participants, 31 (86.1%) were classified as Group 1, while the remaining 5 patients (13.9%) had Group 4. Of the 31 patients classified as Group 1 PAH, idiopathic (n = 21), congenital (n = 6), portal hypertension (n = 2), human immunodeficiency virus infection (n = 1), and connective tissue disease (n = 1) accounted for the putative etiologies. All 5 patients in Group 4 had CTEPH. The characteristics of the study participants are summarized in Table 1. Overall, 25 subjects were females (69.4%), with a median age: 49.0 years (IQR 38.2-57.0) and median BMI: 25.5 kg/m 2 (IQR 20.7-28.9). Half of the study participants (n = 18; 50%; Table 1) had nocturnal desaturation. Clinical and hemodynamic parameters were similar in patients with or without nocturnal desaturation. Moreover, RDI, apnea-hypopnea index (AHI), and oxygen desaturation index (ODI) values were also similar to those with or without nocturnal hypoxemia (p values: 0.504, 0.317, and 0.317; respectively). As expected, there was an excellent correlation between the RDI and the AHI (r 0.953; p < 0.001). Twenty-seven participants fulfilled the criteria for SDB: 26 suffered from OSA and one participant fulfilled the diagnostic criteria of CSA with no evidence of Cheyne-Stokes respiration (male, RDI: 20.7 events/h, and average SpO 2 : 89.0%). Overall, the severity of SDB was distributed as follows: mild (n = 19), moderate (n = 4), and severe (n = 4). As illustrated in the Fig. 2, the values of the average SpO 2 were not statistically different in no-SDB (n = 9) or SDB subjects (n = 27); p = 0.124. In addition, the average SpO 2 was similar across the three severity levels of SDB (p = 0.185; Fig. 2). Table 2 reports the Spearman rank-order correlation coefficients between oximetry data and RDI values. Baseline SpO 2 , nadir SpO 2 , and time spent with oxygen saturation under 90% (T90) were significantly correlated with the average SpO 2 : r 0.062 (p < 0.001); r 0.794 (p < 0.001), and r − 0.945 (p < 0.001); respectively. On the other hand, neither ODI nor RDI were statistically correlated with the average SpO 2 : p values: 0.087 and 0.164; respectively. Table 3 shows the findings from the multiple linear regression analysis containing nine possible relevant parameters regarding the dependent variable (average SpO 2 ). In the linear regression model (coefficient of determination [R 2 ] = 0.391), the mPAP (β − 0.668; 95% CI − 0.259 to − 0.015; p = 0.030) emerged as the only significant and independent predictor of the average SpO 2 values. Interestingly, the RDI did not significantly influence the SpO 2 measurements (β − 0.001; 95% CI − 0.149 to 0.049; p = 0.998); Table 3.

Discussion
In the present study of 36 clinically stable patients with precapillary PH, 75.0% were found to suffer from a degree of SDB, with the vast majority of SDB fulfilling the criteria of OSA. Furthermore, 50% of the subjects had evidence of nocturnal hypoxemia. Despite the high prevalence of OSA  in the cohort, most patients were classified as having mild OSA, but the prevalence of nocturnal hypoxemia was elevated, regardless of the presence of SDB. Moreover, mPAP values were the only significant and independent predictor of the average SpO 2 values, suggesting that the severity of PH (estimated by mPAP) can infer the presence of sleep hypoxemia. Our findings concur with those of previous studies and show that PH severity is associated with the degree of hypoxemia during sleep [15,29]. The frequency of SDB in patients with precapillary PH is generally high, with some studies reporting a high prevalence of both OSA and CSA [17,20,30], and another suggesting a marked predominance of OSA over CSA [31]. In one cross-sectional study that evaluated 52 subjects with Group 1 PAH, SDB was present in 71% (56% had OSA and 44% CSA) [30]. Similarly, among 38 outpatients with  idiopathic PAH (n = 23) or CTEPH (n = 15), a high SDB prevalence was found, with CSA being clearly predominant [17]. However, in another recent study that evaluated 71 patients with precapillary PH (41% males) and 35 matched controls, OSA was detected in 68% of the precapillary PH patients as compared to 5% among controls [31]. Moreover, only 1 PAH patient had CSA (AHI: 12 events/h) without Cheyne-Stokes respiration [31]. Similar to our findings, nocturnal hypoxemia was present in 48% of patients with PAH, and the presence of nocturnal hypoxemia was unrelated to the presence of OSA [31]. We should emphasize that sleep-related hypoxemia is a prevalent disturbance among precapillary PH patients with a widely variable percentage of subjects displaying sleeprelated hypoxemia depending on the oximetry-derived criterion used to assess nocturnal desaturation [32,33]. All such studies, albeit only a few in number, suggest that the presence of nocturnal hypoxemia occurs regardless of the presence of SDB, and when present is frequently associated with a worse prognosis in patients suffering from PAH [20,32,33]. In a small cohort of 13 PAH individuals, nocturnal hypoxemia was reported in 10/13 patients [32]. Another study found that 30 of 43 PAH patients exhibited nocturnal desaturation, with those patients with desaturations during sleep also suffering from higher mPAP values and lower cardiac index [34]. In a cohort of 63 individuals with PAH or CTEPH, 63% had nocturnal hypoxemia [35], while in another previous study that included 15 PAH patients, only 2 had sleep apnea, but 8 patients had sleep-related hypoxemia without the presence of sleep apnea [36].
Although hypoxemia during sleep is frequent in precapillary PH, the underlying mechanism still remains unclear. It is usually advanced that ventilation-perfusion mismatch and narrowing of the distal airways, increasing the pulmonary physiologic dead space underlie the higher frequency of hypoxemic events in PAH [36][37][38]. Of note, the presence of continuous nocturnal hypoxemia, rather than the presence of SDB-related intermittent hypoxemia has been implicated in a poor prognosis among PAH patients [20], possibly because chronic hypoxemia promotes hypoxic pulmonary vasoconstriction and pulmonary vascular remodeling [39]. Notwithstanding, pulse oximetry has been proposed as part of the routine assessment of patients with PAH to detect the presence of nocturnal hypoxemia [34]. However, oximetry alone may not be a sufficiently accurate diagnostic test for identifying SDB in precapillary PH patients [40] because of inadequate discriminatory power when compared to PSG [17].
Our study is not without limitations. First, it was conducted in a single institution, which may limit its generalizability and therefore requires external validation. Second, the sample size was relatively small, which could lead to type 2 beta errors, such that our findings must be considered with caution. Third, all study participants were evaluated with a sleep test on a single night and the possibility of night-to-night variability must be considered. Fourth, the diagnosis of OSA was based on the RDI obtained from an HSAT, which may underestimate the actual prevalence and severity of SDB. Finally, as the hemodynamic parameters were not evaluated in close temporal proximity or concurrently with the HSAT, inferences from the associations between these two sets of measures require caution. Our study had a cross-sectional design, such that extrapolation of the data to prediction of treatment outcomes is obviously outside the scope of the study. Therefore, future prospective and ideally multicenter studies should be forthcoming to assess the real impact of treating OSA and/or intermittent hypoxemia on the prognosis of patients with precapillary PH.

Conclusion
In summary, the majority of precapillary PH patients exhibited the presence of some type of SDB, and more particularly of OSA, the latter not being associated with nocturnal desaturation. The mPAP emerged as the only independent and significant predictor of the average SpO 2 assessed from the HSAT, suggesting that the nocturnal desaturation documented in these patients is likely influenced by the severity of precapillary PH than by the presence of SDB.