ORIGINAL RESEARCH published: 27 June 2022 doi: 10.3389/fmed.2022.870906 Respiratory Polygraphy Patterns and Risk of Recurrent Cardiovascular Events in Patients With Acute Edited by: Marta Bueno, Coronary Syndrome University of Pittsburgh, United States Andrea Zapater 1,2† , Geoffroy Solelhac 3* †‡ , Alicia Sánchez-de-la-Torre 2,4 , Reviewed by: Esther Gracia-Lavedan 2,4 , Ivan David Benitez 2,4 , Gerard Torres 2,4 , Jordi De Batlle 2,4 , Sébastien Bailly, INSERM U1042 Laboratoire Hypoxie José Haba-Rubio 3 , Mathieu Berger 3 , Jorge Abad 2,5 , Joaquín Duran-Cantolla 2,6 , et Physiopathologies Amaia Urrutia 7 , Olga Mediano 2,8 , María José Masdeu 2,9 , Estrella Ordax-Carbajo 10 , Cardiovasculaires et Respiratoires Juan Fernando Masa 2,11 , Mónica De la Peña 12 , Mercé Mayos 2,13 , Ramon Coloma 14 , (HP2), France Josep María Montserrat 2,15 , Eusebi Chiner 16 , Olga Mínguez 4 , Lydia Pascual 4 , Beatriz F. Giraldo, Anunciación Cortijo 4 , Dolores Martínez 4 , Mireia Dalmases 2,4 , Chi-Hang Lee 17 , Institute for Bioengineering R. Doug McEvoy 18 , Ferran Barbé 2,3 , Raphael Heinzer 3,19 and of Catalonia (IBEC), Spain Manuel Sánchez-de-la-Torre1,2* ‡ and Spanish Sleep Network *Correspondence: 1 Precision Medicine in Chronic Diseases, Hospital Universitari Arnau de Vilanova-Santa Maria, IRB Lleida, Department Manuel Sánchez-de-la-Torre of Nursing and Physiotherapy, Faculty of Nursing and Physiotherapy, University of Lleida, Lleida, Spain, 2 Centro sanchezdelatorre@gmail.com de Investigación Biomédica en Red de Enfermedades Respiratorias (CIBERES), Madrid, Spain, 3 Center for Investigation Geoffroy Solelhac and Research in Sleep, Lausanne University Hospital and University of Lausanne, Lausanne, Switzerland, 4 Translation geoffroy.solelhac@chuv.ch Research in Respiratory Medicine, Hospital Universitari Arnau de Vilanova-Santa Maria, IRBLleida, Lleida, Spain, † These authors share first authorship 5 Respiratory Department, Hospital Universitari Germans Trias i Pujol, Barcelona, Spain, 6 Servicio de Investigación OSI, IIS ‡ These authors share senior Bioaraba, Hospital Universitario Araba, Vitoria-Gasteiz, Spain, 7 Servicio Neumologia, Hospital Universitario Cruces, Bizkaia, authorship Spain, 8 Respiratory Department, Hospital Universitario de Guadalajara, Guadalajara, Spain, 9 Respiratory and Sleep Department, Hospital Universitari Parc Taulí, Institut Investigació i Innovació Parc Taulí (I3PT), Universitat Autònoma Specialty section: de Barcelona, Barcelona, Spain, 10 Respiratory Department, Hospital Universitario de Burgos, Burgos, Spain, 11 Respiratory This article was submitted to Department, Hospital San Pedro Alcántara, Cáceres, Spain, 12 Clinic Analysis and Respiratory Services, Institut Pulmonary Medicine, de Investigació Sanitaria de Palma, Hospital Universitari Son Espases, Palma de Mallorca, Spain, 13 Sleep Unit, Department a section of the journal of Respiratory Medicine, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain, 14 Respiratory Department, Hospital Frontiers in Medicine General Universitario de Albacete, Albacete, Spain, 15 Respiratory Department, Hospital Clinic, Barcelona, Spain, 16 Respiratory Department, Hospital Universitari Sant Joan d’Alacant, Alicante, Spain, 17 Cardiovascular Research Institute, Received: 07 February 2022 National University of Singapore, Singapore, Singapore, 18 Adelaide Institute for Sleep Health, College of Medicine and Public Accepted: 30 May 2022 Health, Flinders University, Adelaide, SA, Australia, 19 Pulmonary Department, Lausanne University Hospital and University Published: 27 June 2022 of Lausanne, Lausanne, Switzerland Citation: Zapater A, Solelhac G, Sánchez-de-la-Torre A, Introduction: Obstructive sleep apnea (OSA) severity is based on the apnea-hypopnea Gracia-Lavedan E, Benitez ID, index (AHI). The AHI is a simplistic measure that is inadequate for capturing disease Torres G, De Batlle J, Haba-Rubio J, severity and its consequences in cardiovascular diseases (CVDs). Deleterious effects of Berger M, Abad J, Duran-Cantolla J, Urrutia A, Mediano O, Masdeu MJ, OSA have been suggested to influence the prognosis of specific endotypes of patients Ordax-Carbajo E, Masa JF, with acute coronary syndrome (ACS). We aim to identify respiratory polygraphy (RP) De la Peña M, Mayos M, Coloma R, Montserrat JM, Chiner E, Mínguez O, patterns that contribute to identifying the risk of recurrent cardiovascular events in Pascual L, Cortijo A, Martínez D, patients with ACS. Dalmases M, Lee C-H, McEvoy RD, Barbé F, Heinzer R and Methods: Post hoc analysis of the ISAACC study, including 723 patients admitted for a Sánchez-de-la-Torre M (2022) first ACS (NCT01335087) in which RP was performed. To identify specific RP patterns, Respiratory Polygraphy Patterns and Risk of Recurrent Cardiovascular a principal component analysis (PCA) was performed using six RP parameters: AHI, Events in Patients With Acute oxygen desaturation index, mean and minimum oxygen saturation (SaO2 ), average Coronary Syndrome. Front. Med. 9:870906. duration of events and percentage of time with SaO2 < 90%. An independent doi: 10.3389/fmed.2022.870906 HypnoLaus population-based cohort was used to validate the RP components. Frontiers in Medicine | www.frontiersin.org 1 June 2022 | Volume 9 | Article 870906 Zapater et al. OSA Patterns and Cardiovascular Risk Results: From the ISAACC study, PCA showed that two RP components accounted for 70% of the variance in the RP data. These components were validated in the HypnoLaus cohort, with two similar RP components that explained 71.3% of the variance in the RP data. The first component (component 1) was mainly characterized by low mean SaO2 and obstructive respiratory events with severe desaturation, and the second component (component 2) was characterized by high mean SaO2 and long-duration obstructive respiratory events without severe desaturation. In the ISAACC cohort, component 2 was associated with an increased risk of recurrent cardiovascular events in the third tertile with an adjusted hazard ratio (95% CI) of 2.44 (1.07 to 5.56; p-value = 0.03) compared to first tertile. For component 1, no significant association was found for the risk of recurrent cardiovascular events. Conclusion: A RP component, mainly characterized by intermittent hypoxemia, is associated with a high risk of recurrent cardiovascular events in patients without previous CVD who have suffered a first ACS. Keywords: acute coronary syndrome, cardiovascular disease, obstructive sleep apnea, precision medicine, respiratory polygraphy INTRODUCTION identified a specific endotype of patients with ACS where OSA was associated with an increased risk of recurrent cardiovascular Cardiovascular diseases (CVDs) are one of the leading causes events. The results of this post hoc study suggested that patients of death worldwide, and more than half are consequences of with OSA admitted by a first ACS without previous CVD present coronary heart disease, which is often manifested for the first an increased risk of recurrent cardiovascular events (9). ISAACC time as acute coronary syndrome (ACS). Ranging from unstable study and others intervention trials failed to find a positive effect angina to myocardial infarction, ACS is the main cause of of CPAP treatment in the reduction of cardiovascular events morbidity and mortality worldwide (1, 2). (8, 10, 11) on the risk of cardiovascular recurrence. Obstructive sleep apnea (OSA) is a prevalent and chronic Recent study (12) reported that CVD mortality was not disease that affects more than 20% of the adult population (3) and associated with the AHI when it was assessed as an independent is caused by the collapse of the upper airway during sleep, which predictor. In contrast, they found that the hypoxic burden, leads to transient asphyxia. These events lead to brain arousal, an estimation of the depth and duration of respiratory- intermittent hypoxemia, poor quality of life and metabolic related desaturation, is an alternative metric associated with disturbances. OSA is considered an important public health an increased risk of CVD mortality. Some authors argue that problem and is largely associated with cardiovascular outcomes the AHI may not be a real marker of clinical disease and (4, 5). In fact, basic, epidemiological and clinical research suggests encourage further investigation of underlying pathophysiological a role of OSA in the initiation and/or progression of several and clinical phenotypes that are not captured by this measure CVDs (6). Observational studies have suggested that OSA is (13). The AHI does not provide information about apnea depth an independent risk factor for cardiovascular morbidity and and length and is a simplistic measure that has many downsides mortality (7). However, there is a lack of a dose-response and could result in inadequate capture of the severity of the relationship between the severity of the disease [expressed by disease. Respiratory polygraphy (RP) allows the evaluation of a the apnea hypopnea index (AHI)] and cardiovascular risk, and large number of physiological signals associated with respiratory some studies did not find an association between OSA and disorders during sleep. Polygraphic recordings include at least cardiovascular risk (8, 9). In the ISAAC study, a randomized oronasal flow, thoracoabdominal movements, heart rate, and controlled trial to evaluate the effect of intervention with oxygen saturation. Dimensionality reduction technique allows continuous positive airway pressure (CPAP) treatment on the risk the transformation of data from a high-dimensional space of recurrence of cardiovascular event in patients with ACS, OSA into a low-dimensional space so that the low-dimensional was not associated with a worse cardiovascular risk prognosis representation retains some meaningful properties of the original in patients with ACS (8). A post hoc study from the ISAACC data. The dimensionality reduction analysis of the signals of RP study reported that the effect of OSA is heterogeneous and it was could contribute to the construction of models that would relate OSA to cardiovascular risk. Abbreviations: ACS, acute coronary syndrome; AHI, apnea-hypopnea index; Available evidence supports that the indicators of CI, confidence interval; CPAP, continuous positive airway pressure; CVD, cardiovascular risk in patients with OSA, mainly based on cardiovascular disease; HR, hazard ratio; IQR, interquartile rang; ODI, oxygen desaturation index; OSA, obstructive sleep apnea; PCA, principal component AHI, are insufficient. In the present study, we aimed to evaluate analysis; SaO2 , oxygen saturation. the RP patterns that could contribute to identifying the risk of Frontiers in Medicine | www.frontiersin.org 2 June 2022 | Volume 9 | Article 870906 Zapater et al. OSA Patterns and Cardiovascular Risk recurrent cardiovascular events in patients with a first ACS and in a general population aged 35–75 years who were selected at without previous CVD. random and were enrolled in the study. Procedures and Outcomes MATERIALS AND METHODS ISAACC Cohort Questionnaires to record demographic and anthropometric Study Design characteristics, medical history and usual pharmacological This is an ancillary study of the ISAACC (Impact of Sleep Apnea treatment were administered the day before the sleep study along syndrome in the evolution of ACS. Effect of intervention with with questionnaires associated with quality of life (EuroQol-5D CPAP) study, a multicenter, open-label, parallel, prospective, questionnaire) and the degree of daytime sleepiness (Epworth randomized, controlled trial (registered trial NCT01335087). Full Sleepiness Scale test). All patients were evaluated at baseline and details of the ISAACC study’s methods, including inclusion and 1 month, 3 months, 6 months, 12 months, 18 months, 24 months, exclusion criteria, have been previously reported (14). Briefly, 30 months, 36 months, and annually thereafter. All patients were eligible patients were ≥18 years old, had been admitted for ACS monitored and followed up for a minimum of 1 year. At each to coronary care units or cardiology hospitalization ward at 15 visit, sociodemographic and anthropometric variables previously hospitals in Spain and had an Epworth Sleepiness Scale score of related to increased cardiovascular risk were recorded. Each ≤10 (i.e., patients without excessive daytime sleepiness or non- follow-up visit included assessments of the rate of a composite sleepy). Eligible patients underwent RP (Embletta, ResMed, Bella of cardiovascular events [cardiovascular death or non-fatal events Vista, NSW, Australia) during the first 24–72 h after admission to (acute myocardial infarction, non-fatal stroke, hospital admission evaluate the presence of OSA (14). For the current post hoc study, for heart failure and new hospitalizations for unstable angina or we analyzed the data from 723 ISAACC patients with a first ACS transient ischemic attack)]. and without CVD (Figure 1) and in whom a deleterious effect of For the present study, we explored RP parameters to identify OSA was observed (13). specific components, including the AHI, oxygen desaturation Acute coronary syndrome was defined as the acute index (ODI, >4%), mean and minimum oxygen saturation presentation of coronary disease with or without ST elevation (SaO2 ), average duration of events and percentage of time with infarction, unstable angina, or type 1 myocardial infarction. SaO2 < 90%. We explored the contribution of RP components to The ethics committee of each participating center approved the risk of recurrent cardiovascular events. the study (approval number in the coordinating center: 2010- The cardiovascular risk variables explored were age, sex, 852) and patients provided written informed consent. current cigarette smoking, current alcohol consumption, The RP components in the ISAACC cohort was externally obesity, hypertension, previous cerebrovascular disease, validated in the HypnoLaus population-based cohort (3), a diabetes mellitus, and dyslipidemia (low-density lipoprotein nested-study of CoLaus/PsyCoLaus (15, 16). This study was cholesterol ≥100 mg/dL or the use of lipid-lowering drugs). designed to assess the prevalence of sleep-disordered breathing Moreover, on the basis of cardiovascular risk prediction models FIGURE 1 | Flowchart of study. CPAP, continuous positive airway pressure; CVD, cardiovascular disease. Frontiers in Medicine | www.frontiersin.org 3 June 2022 | Volume 9 | Article 870906 Zapater et al. OSA Patterns and Cardiovascular Risk (17, 18), we explored the following other cardiovascular variables: A total of 11.5% of the patients had recurrent cardiovascular antihypertensive drugs, antiplatelet, and antithrombotic drugs, events during follow-up (Supplementary Table 2). heart rate, systolic blood pressure, initial serum creatinine, stents implanted, no in-hospital percutaneous coronary intervention HypnoLaus Cohort troponin and mean SaO2 . The external validation cohort was 2,168 individuals from the HypnoLaus population-based cohort. Finally, 1941 individuals HypnoLaus Cohort were selected without previous CVD and available data. The Participants underwent full polysomnography at home. Details of median age was 56.9 years, 47% were males, and the median the sampling and procedure methodologies have been described AHI was 9.8 events per hour (Supplementary Table 3). The elsewhere (3, 15). For the current study, we analyzed the data mean time of follow-up (or CV event) was 4.25 (standard from 1941 HypnoLaus patients. The ethics committee of the deviation: 1.77) years. University of Lausanne approved the CoLaus/PsyCoLaus cohort study and the HypnoLaus Sleep Cohort study. We obtained written informed consent from all participants. Identification of Respiratory Polygraphy Components Statistical Analysis Principal component analysis was performed using the six RP The main characteristics were described using medians (25th parameters: AHI, ODI, mean and minimum SaO2 , average percentile and 75th percentile) or percentages. The P value for duration of events and percentage of time with SaO2 <90%. trend was evaluated using Spearman’s rank correlation coefficient The first two components were selected to characterize the RP when data were continuous and non-normally distributed and patterns. In the ISAACC study, these two components, which the chi-square test for trend when they were categorical (19). contributed significantly to explaining the relationships among Principal component analysis (PCA) was used to avoid the six RP parameters (eigenvalues >1), accounted for 70% collinearity between the RP parameters and to identify relational of the variability of the RP data (Supplementary Table 4). patterns that reflect sleep apnea patterns. Six RP standardized The first component (Component 1) was highly correlated parameters previously mentioned were used in the PCA. with the AHI, ODI and percentage of time with SaO2 <90% RP parameters that were not normally distributed were log- and inversely correlated with the mean and minimum SaO2 . transformed and then included in the PCA (ODI and AHI). The Component 1 accounted for 42.2% of the variability of the magnitude of the eigenvalues > 1 (20) and variance explained RP data. The second component (Component 2) was highly by each component were used to determine the number of correlated with the AHI, ODI, mean SaO2 , and average duration principal components to select. The components are described of events and inversely correlated with the percentage of time by the correlations with each RP parameter included in the with SaO2 <90% (Figure 2A). Component 2 accounted for PCA. Selected components from the PCA were grouped into 26.6% of the variability of the RP data. Whereas the mean and tertiles and included in the model to explore the association with minimum SaO2 were negatively correlated in component 1, in the the primary composite endpoint. Cox proportional regression component 2 mean SaO2 was positively correlated. Additionally, models adjusted by the predictors mentioned in the previous the correlation of duration obstructive respiratory events was section were used. Hazard ratios (HRs) and 95% confidence higher in component 2. Table 2 shows the description of the intervals (95% CIs) were reported. The proportional hazard RP parameters by tertiles of each RP component identified and assumption was checked in all models. Discrimination was selected (components 1 and 2). assessed using the C-statistic (21). From the validation, the PCA was repeated in the HypnoLaus Statistical analyses were conducted using R, version 3.6 (22). cohort using the same six RP parameters explored in the Two-sided p-values were reported, and statistical significance ISAACC cohort, and the first two components explained 71.3% was set at 0.05. of the variance of the RP data (Supplementary Table 5). Both components presented a high degree of similarity to those identified in the ISAACC cohort (Figure 2B). The RP parameters RESULTS were described by tertiles of the selected component for the HypnoLaus cohort (Supplementary Table 6). Patient Characteristics ISAACC Cohort From a total of 1,851 patients with ACS enrolled in the ISAACC Respiratory Polygraphy Component cohort, we included 948 patients without CPAP treatment and Distribution without previous CVD. Finally, 723 patients with available data ISAACC patients were characterized using sociodemographic for the variables explored were included in the present study and clinical characteristics by tertiles of the two RP components (Figure 1). The median age of the patients was 58 years, and (Supplementary Table 7). Patients with high adherence 81.6% were males (Table 1). The patients were followed for a in component 1 had more comorbidities, as the third median of 3.33 (interquartile range: 3.72) years. Excluded patients tertile showed a greater proportion of patients with obesity, had sociodemographic and clinical characteristics similar to those hypertension and diabetes. Similar characteristics were shown by of the patients included in the analysis (Supplementary Table 1). tertiles of component 2. Frontiers in Medicine | www.frontiersin.org 4 June 2022 | Volume 9 | Article 870906 Zapater et al. OSA Patterns and Cardiovascular Risk TABLE 1 | Baseline characteristics of patients from the ISAACC cohort. TABLE 1 | (Continued) All No recurrent Recurrent P value All No recurrent Recurrent P value CVE CVE CVE CVE (N = 723) (N = 640) (N = 83) (N = 723) (N = 640) (N = 83) Medication Sex 0.203 Antihypertensive 293 (40.5%) 251 (39.2%) 42 (50.6%) 0.062 Female 133 (18.4%) 113 (17.7%) 20 (24.1%) drug Male 590 (81.6%) 527 (82.3%) 63 (75.9%) Lipid-lowering 188 (26.0%) 161 (25.2%) 27 (32.5%) 0.191 Anthropometric drug measures Antidiabetic oral 113 (15.6%) 87 (13.6%) 26 (31.3%) <0.001 Age, years 58.0 58.0 58.0 0.426 medication [51.0;66.0] [51.0;66.0] [52.5;66.5] Insulin 26 (3.60%) 21 (3.28%) 5 (6.02%) 0.207 Body mass 28.0 27.8 28.9 0.070 Antiplatelet and 69 (9.54%) 45 (7.03%) 24 (28.9%) <0.001 index, kg·m−2 [25.1;30.8] [25.0;30.5] [25.6;31.3] antithrombotic Waist-hip ratio 0.98 0.98 0.99 0.435 drugs [0.94;1.02] [0.94;1.02] [0.95;1.03] Cardiovascular Neck 40.0 40.0 40.0 0.662 variables circumference, [38.0;43.0] [38.0;43.0] [38.0;42.0] Heart rate, bpm 70.0 70.0 73.0 0.125 cm [63.5;80.0] [63.0;80.0] [65.0;80.5] Lifestyle risk Systolic blood 120 [110;133] 120 [110;133] 120 [110;130] 0.608 factors pressure, mm Hg Smoking 0.838 Creatinine, 0.85 0.84 0.88 0.131 Never 189 (26.1%) 166 (25.9%) 23 (27.7%) mg/dL [0.73;0.99] [0.73;0.98] [0.74;1.02] Former 166 (23.0%) 149 (23.3%) 17 (20.5%) Stents 652 (90.2%) 575 (89.8%) 77 (92.8%) 0.518 Current 368 (50.9%) 325 (50.8%) 43 (51.8%) implanted Drinking 0.168 Peak troponin, 0.070 No 534 (73.9%) 467 (73.0%) 67 (80.7%) quartiles Yes 189 (26.1%) 173 (27.0%) 16 (19.3%) Quartile 1 130 (18.0%) 107 (16.7%) 23 (27.7%) Quartile 2 191 (26.4%) 171 (26.7%) 20 (24.1%) Sleep parameters Quartile 3 203 (28.1%) 186 (29.1%) 17 (20.5%) AHI, events per 16.5 15.9 20.2 0.016 Quartile 4 199 (27.5%) 176 (27.5%) 23 (27.7%) h [7.00;33.3] [6.70;33.0] [11.1;37.1] Cardiovascular 0.818 event type ODI > 4%, per h 14.0 13.9 16.7 0.061 [5.25;29.9] [5.00;29.3] [8.75;34.9] Non-Q-wave 299 (41.4%) 264 (41.2%) 35 (42.2%) AMI Mean SaO2 ,% 93.1 93.1 93.3 0.785 Unstable angina 37 (5.12%) 32 (5.00%) 5 (6.02%) [92.0;94.4] [92.0;94.3] [92.0;94.5] Q-wave AMI 387 (53.5%) 344 (53.8%) 43 (51.8%) Minimum 86.0 86.0 86.0 0.351 SaO2 ,% [81.0;88.0] [81.0;88.0] [82.0;89.0] Data are presented as the n (%) or the median [25th percentile; 75th percentile]. Time with 1.00 1.05 0.60 0.512 CVE, cardiovascular event; AHI, apnea-hypopnea index; ODI, oxygen desaturation SaO2 < 90%, % [0.10;7.25] [0.10;7.35] [0.00;6.80] index; SaO2 , oxygen saturation; ACS, acute coronary syndrome; bpm, beats per minute; and AMI, acute myocardial infarction. Significant p values (p < 0.05) are Average 20.0 20.0 19.0 0.464 presented in bold. duration of [17.0;23.0] [17.0;23.0] [17.0;22.5] events, sec Epworth 5.00 5.00 5.00 0.152 sleepiness scale [3.00;7.00] [3.00;7.00] [3.00;7.00] In the same way, HypnoLaus subjects were described using Medical sociodemographic and clinical characteristics by tertiles of the history two RP components (Supplementary Table 8). Subjects with Hypertension 325 (45.0%) 278 (43.4%) 47 (56.6%) 0.031 high adherence in component 1 had more comorbidities, as Diabetes 141 (19.5%) 112 (17.5%) 29 (34.9%) <0.001 third tertile showed a greater proportion of subjects with mellitus hypertension, diabetes, dyslipidemia and a higher body mass Dyslipidemia 596 (82.4%) 530 (82.8%) 66 (79.5%) 0.556 index (BMI). In component 2, the third tertile showed a greater Cerebrovascular 16 (2.21%) 12 (1.88%) 4 (4.82%) 0.100 proportion of subjects with hypertension, a lower proportion disease of subjects with dyslipidemia and no differences regarding Chronic 35 (4.84%) 32 (5.00%) 3 (3.61%) 0.787 pneumopathy BMI and diabetes. Neurological 31 (4.29%) 26 (4.06%) 5 (6.02%) 0.387 disease Severity of 0.331 Relationship Between Respiratory ACS Polygraphy Components and Recurrent Mild 653 (90.3%) 581 (90.8%) 72 (86.7%) Cardiovascular Risk Severe 70 (9.68%) 59 (9.22%) 11 (13.3%) In the present study, from the ISAACC study, we (Continued) explored specific RP parameters and components Frontiers in Medicine | www.frontiersin.org 5 June 2022 | Volume 9 | Article 870906 Zapater et al. OSA Patterns and Cardiovascular Risk FIGURE 2 | Correlations between respiratory polygraphy parameters and components derived from the PCA. (A) In the ISAACC cohort, these two components, component 1 and component 2, accounted for 43 and 26% of the total variance, respectively. (B) In HypnoLaus, these two components, component 1 and component 2, accounted for 49.2 and 22.1% of the total variance, respectively. AHI, apnea-hypopnea index; ODI, oxygen desaturation index; and SaO2 , oxygen saturation. Dashed lines represent –0.5 and 0.5 correlations. (Supplementary Table 9) in the subgroup of patients admitted risk of recurrent cardiovascular events (Supplementary by a first ACS and without previous CVD (n = 723). Table 10). In the subgroup of patients without previous ACS, OSA In the analysis of RP components and the risk of recurrent severity (measured by AHI) showed a trend toward statistical cardiovascular events, it was observed that component 2 showed association [HR (95% CI) of 1.62 (0.98–2.66)] with an increased a significant independent association with the risk of recurrent Frontiers in Medicine | www.frontiersin.org 6 June 2022 | Volume 9 | Article 870906 Zapater et al. OSA Patterns and Cardiovascular Risk TABLE 2 | Description of the respiratory polygraphy parameters by tertiles of the components derived from the PCA in the ISAACC cohort. Component 1 Component 2 1st tertile 2nd tertile 3rd tertile p for trend 1st tertile 2nd tertile 3rd tertile p for trend (N = 239) (N = 238) (N = 246) (N = 239) (N = 238) (N = 246) RP parameters AHI, events per h 6 [8.6] 18 [18] 35.2 [31.5] <0.001 5.3[11.7] 15.8 [17.2] 32.8 [26.3] <0.001 ODI > 4%, per h 4.2 [5.9] 15.1 [13.8] 32.9 [30.9] <0.001 4 [11.6] 13 [16.7] 27.6 [25.4] <0.001 Mean SaO2 ,% 94.3 [1.9] 93.2 [1.7] 91.8 [2.9] <0.001 92.2 [3.7] 93 [2] 94 [2] <0.001 Minimum SaO2 ,% 89 [4] 86 [4] 80 [8.8] <0.001 85 [8] 86 [6] 85.5 [6] 0.312 Average duration of events, sec 19 [7] 20 [6] 20 [7.8] <0.001 17 [6] 20 [6] 22 [8] <0.001 Time with SaO2 < 90%,% 0 [0.3] 0.8 [2.2] 13 [29.3] <0.001 3.1[33.4] 0.8 [5.1] 0.9 [3.8] <0.001 Data are presented as the median [IQR (interquartile range)]. PCA, principal component analysis; RP, respiratory polygraphy; AHI, apnea-hypopnea index; ODI, oxygen desaturation index; and SaO2 , oxygen saturation. The p value for trend was calculated using Spearman’s rank correlation coefficient. Significant p values (p < 0.05) are presented in bold. cardiovascular events in patients with the highest adherence (3rd management. In our study, we observed that the effect of OSA tertile) of component 2 with an adjusted HR (95% CI) of 2.44 severity, assessed only by the AHI, showed a trend toward a (1.07 to 5.56). For component 1, no significant contribution was statistical association (with an HR of 1.62) with the risk of found to the risk of recurrent cardiovascular events (Table 3 recurrent cardiovascular events in the group of patients without and Figure 3). The discriminatory power of the models showed previous CVD. When considering the RP components, we a C-statistic for the risk of recurrent cardiovascular events of observed an HR of 2.44 for component 2, indicating an increase 72.3% for patients without previous CVD. Similar results were in the risk of recurrent cardiovascular events. Based on the found exploring components from PCA as continuous [HR results, we suggest that the analysis of other parameters from (95%CI): for Component 2: 1.38 (0.99–1.94) and for Component RP could contribute to the determination of RP components 1: 0.82 (0.6–1.13)]. and their relationship with the development of recurrent In the HypnoLaus cohort the risk of recurrent cardiovascular cardiovascular events in patients who have suffered an ACS. event in patients with first ACS could not be assessed due to its The ISAACC study (8) and others (12) have failed to population-based design. demonstrate the relationship between the AHI and an increased risk of recurrent cardiovascular events. Previous studies have suggested that other components from sleep studies could DISCUSSION contribute to characterizing the underlying pathology of OSA In the present study, we explored the relationship between specific components of sleep-disordered breathing and the risk of TABLE 3 | Cox proportional hazard model for the primary composite endpoint in recurrent cardiovascular events in a cohort of patients with a first the ISAACC cohort. ACS and without previous CVD. Recently, it has been reported ALL (N = 723) that patients with OSA admitted by a first ACS without previous CVD present an increased risk of recurrent cardiovascular HR* (95% CI) p value p for trend events (13). The results of this study show an association Component 1 between a specific polygraphy component and the risk of 1st tertile 1 0.49 recurrent cardiovascular events. In the ISAACC population, we 2nd tertile 0.81 (0.41–1.59) 0.544 report that a polygraphy component mainly characterized by 3rd tertile 0.74 (0.31–1.73) 0.481 intermittent hypoxemia (component 2) increases the risk of Component 2 recurrent cardiovascular events in patients with ACS and without 1st tertile 1 0.03 previous CVD. Based on previously described cardiovascular risk 2nd tertile 1.58 (0.79–3.17) 0.194 variables, the C-statistic for the risk of recurrent cardiovascular 3rd tertile 2.44 (1.07–5.56) 0.034 events was 72% for both phenotypes of patients, similar to other studies (23). Remarkably, the RP components identified HR (95% CI), hazard ratio (95% confidence interval). All variables included in the model met the assumptions for proportional hazards. The p value for trend in the ISAACC cohort were also identified in the HypnoLaus was computed by treating component variables coded as numbers in the model. population-based cohort. Significant p values (p < 0.05) are presented in bold. Currently, OSA severity is characterized by the AHI, which *Adjusted for age, sex, smoking, alcohol consumption, obesity, hypertension, stroke, diabetes, the severity of ACS, dyslipidemia, antihypertensive drugs, captures only one aspect of this heterogeneous disease. The antiplatelet and antithrombotic drugs, heart rate, systolic blood pressure, ability to dissect OSA heterogeneity using additional features is creatinine, stents implanted, cardiovascular event type, troponin peak, and necessary for incorporating personalized approaches in disease mean SaO2 . Frontiers in Medicine | www.frontiersin.org 7 June 2022 | Volume 9 | Article 870906 Zapater et al. OSA Patterns and Cardiovascular Risk FIGURE 3 | Hazard ratio for recurrent cardiovascular events by tertiles of respiratory polygraphy components in the ISAACC cohort. Dots are hazard ratios, and lines represent 95% confidence intervals. Hazard ratios were adjusted for age, sex, smoking, alcohol consumption, obesity, hypertension, stroke, diabetes, the severity of ACS, dyslipidemia, antihypertensive drugs, antiplatelet and antithrombotic drugs, heart rate, systolic blood pressure, creatinine, stents implanted, cardiovascular event type, troponin peak, and mean SaO2 . HR (95% CI), hazard ratio (95% confidence interval); ACS, acute coronary syndrome. *Significant p values (p < 0.05). and its association with the risk of CVD. Azarbarzin et al. (12) the identified RP components for recurrent cardiovascular events recently reported that CVD mortality was not associated with in this cohort. Since, the low number of available cardiovascular the AHI when it was assessed as an independent predictor. In events due to its population-based design. contrast, they found that the hypoxic burden, an estimation The existence of specific RP components that could be of the depth and duration of respiratory-related desaturation, associated with detrimental effects in specific phenotypes of is an alternative metric associated with an increased risk of patients with ACS invites us to reconsider the design strategies of CVD mortality. Nevertheless, the authors of that study admit new interventional studies to demonstrate the possible beneficial that this measure does not fully characterize the components effect of CPAP treatment on secondary CVD prevention. These of hypoxic stress because it does not distinguish short and preliminary results raise the possibility that, in studies such as deep desaturations from long and shallow desaturations. In fact, the ISAACC study (8), no beneficial effect was seen because of in our study, we found that the component associated with the inclusion of patients with different phenotypes of ACS where recurrent cardiovascular events was component 2. The highest OSA could have different effects. adherence to this component (3rd tertile), showing the highest The strengths of our study include its multicenter design risk of recurrent cardiovascular events, was characterized by with a large population of patients with diagnoses of ACS higher AHI, higher mean SaO2 and longer event duration. This and OSA. In this study, all participating centers from the tertile would suggest an intermittent hypoxemia component. ISAACC study used the same methodology, and the sleep In contrast, the lowest adherence to this component was study was performed with the same polygraph model. In characterized by lower AHI, ODI and mean SaO2 , which would addition, the RP components identified in the ISAACC cohort suggest a sustained hypoxemia component. were validated in the HypnoLaus population-based cohort. The results of the present study suggest that from available The study presents some limitations that deserve mention. variables from RP, it is possible to improve the evaluation First, to identify specific RP components, a PCA analysis was of the risk of recurrent cardiovascular events with an HR performed using six RP parameters, including the AHI, ODI, of 1.61 exclusively with AHI to 2.44 considering additional mean and minimum SaO2 , average duration of events and variables. The patients with the highest adherence to component percentage of time with SaO2 < 90%. The usefulness of other 2 (3rd tertile), which represents an intermittent hypoxemia variables such as the mean respiratory rate as a predictor component, strengthen the hypothesis that OSA exerts a of cardiovascular recurrence was previously reported. This risk in specific subgroups of patients with ACS. In contrast, variable was not available in the ISAACC study, and future the patients with the lowest adherence to this component, studies exploring the contribution of this specific variable as characterized by sustained hypoxemia, would have a lower risk a predictor of cardiovascular outcomes should be performed. of recurrence of a cardiovascular event. Notably, the same RP Second, the results of this study indicate the possibility of components were found in the HypnoLaus cohort suggesting that identifying RP components associated with different risks of these components are not specific to the ISAACC population. recurrent cardiovascular events in patients who have suffered a However, it was not possible to estimate the predictive nature of first ACS and without previous CVD. However, these exploratory Frontiers in Medicine | www.frontiersin.org 8 June 2022 | Volume 9 | Article 870906 Zapater et al. OSA Patterns and Cardiovascular Risk results should be validated in future independent cohorts. AUTHOR CONTRIBUTIONS Third, in our study, we determined RP components using RP parameters that do not include information about sleep AZ, GS, AS-d-l-T, EG-L, IB, FB, RH, and MS-d-l-T contributed phases. Some studies have suggested that rapid eye movement- to the study concept and design. GS, AS-d-l-T, GT, JD, JA, related respiratory disorders could be informative regarding JD-C, AU, OMe, MJM, EO-C, JFM, MDP, MM, RC, JMM, cardiovascular risk associations with respiratory sleep disorders EC, OMí, LP, AC, DM, RM, and RH contributed to the data (24). The contribution of information about sleep phases acquisition. AZ, GS, AS-d-l-T, EG-L, IB, GT, FB, RH, and should be explored in further studies. Finally, the results from MS-d-l-T contributed to the data analysis and interpretation. the present study may not be extrapolated to a population All authors contributed to the drafting of the manuscript, different from that of patients admitted to the hospital for critically revised the manuscript for important intellectual ACS. This fact makes it necessary to specifically identify content and approved the final version. MS-d-l-T is the guarantor phenotypes in patients with OSA who are attended in different of the manuscript. clinical settings. As a conclusion, in patients who have suffered a first ACS and without previous CVD, there exists a specific RP component, mainly characterized by intermittent hypoxemia, FUNDING that is associated with a greater risk of recurrent cardiovascular Instituto de Salud Carlos III (ISCIII; PI10/02763, PI10/02745, events. These first exploratory results must be confirmed in PI18/00449, and PI19/00907), co-funded by FEDER, “Una future studies, which will evaluate the effect of RP components manera de hacer europa,” IRBLleida – Fundació Pifarré, of specific ACS populations and the possible beneficial effect CERCA Programme/Generalitat de Catalunya, SEPAR, of CPAP treatment for patients with a first ACS and without ResMed Ltd. (Australia), Esteve-Teijin (Spain), Oxigen Salud previous CVD, in whom a deleterious effect of this RP (Spain), Associació Lleidatana de Respiratori (ALLER), and component has been found. Sociedad Española de Sueño (SES). AZ is the recipient of a predoctoral fellowship “Ajuts 2021 de Promoció de la Recerca en Salut-9a edició” from IRBLleida/Diputació de DATA AVAILABILITY STATEMENT Lleida. JD acknowledges receiving financial support from ISCIII (Miguel Servet 2019: CP19/00108), co-funded by the The original contributions presented in the study are included European Social Fund (ESF), “Investing in your future.” in the article/Supplementary Material, further inquiries can be MS-d-l-T has received financial support from a “Ramón y directed to the corresponding authors. Cajal” grant (RYC2019-027831-I) from the “Ministerio de Ciencia e Innovación – Agencia Estatal de Investigación” co-funded by the European Social Fund (ESF)/“Investing ETHICS STATEMENT in your future.” FB received funding from from ResMed (an Australian company that develops products related The ethics committee of each participating center approved to sleep apnea), the Health Research Fund, the Spanish the study (approval number in the coordinating center: 2010- Ministry of Health, the Spanish Respiratory Society, the 852) and patients provided written informed consent. The Catalonian Cardiology Society, Esteve-Teijin (Spain), Oxigen patients/participants provided their written informed consent to Salud (Spain), and ALLER. The funders were not involved participate in this study. in the study design, collection, analysis, interpretation of data, the writing of this article, or the decision to submit it for publication. MEMBERS OF THE SPANISH SLEEP NETWORK Laura Abad, Aida Munþoz, Elisabet Zamora, Ignacio Vicente, ACKNOWLEDGMENTS Sandra Ingleìs, Carlos Egea, Jaime Marcos, Almudena Fernaìndez, Valentin Cabriada-Nuño, Sonia Castro, Leyre We thank the individuals who participated in this trial and their Serrano, Marina Floreìs, Anna Mas, Maricel Arboneìs, Silvia families, the clinical and research teams of the various sleep and Ortega, Alicia Martiìn, Jose Miguel Romaìn-Saìnchez, Ma cardiology departments, and the Spanish Sleep Network for their Isabel Valiente-Diaz, Ma Esther Viejo-Ayuso, Concepcioìn work on the ISAACC study. Rodriìguez-Garciìa, Laura Vigil, Enriqueta Ramiìrez, Mariìa Pinþar, Elisabet Martiìnez, Blanca Barriuso, Jaime Corral, Francisco Javier Goìmez de Terreros Caro, Antonia Barceloì, SUPPLEMENTARY MATERIAL Paloma Gimeìnez, Ana Ma Fortuna, Patricia Penñacoba, Abel Jesuìs Martiìnez-Garciìa, Sergio Garciìa-Castillo, Lara Navas, The Supplementary Material for this article can be found online Onintza Garmendia, Joseì Sancho, Salvador Perelloì, Gemma at: https://www.frontiersin.org/articles/10.3389/fmed.2022. Rubinoìs, and Rocío Gallego. 870906/full#supplementary-material Frontiers in Medicine | www.frontiersin.org 9 June 2022 | Volume 9 | Article 870906 Zapater et al. OSA Patterns and Cardiovascular Risk REFERENCES continuous positive airway pressure on patients with ACS and nonsleepy OSA: the ISAACC trial. Clin Cardiol. (2013) 36:495–501. doi: 10.1002/clc.22166 1. Kumar A, Cannon CP. Acute coronary syndromes: diagnosis and 15. Firmann M, Mayor V, Vidal PM, Bochud M, Pécoud A, Hayoz D, et al. The management, part I. Mayo Clin Proc. (2009) 84:917–38. doi: 10.4065/84.10.917 CoLaus study: a population-based study to investigate the epidemiology and 2. Florés M, De Batlle J, Sánchez-De-La-Torre A, Sánchez-De-La-Torre M, genetic determinants of cardiovascular risk factors and metabolic syndrome. Aldomá A, Worner F, et al. Central sleep apnoea is related to the severity BMC Cardiovasc Disord. (2008) 8:6. doi: 10.1186/1471-2261-8-6 and short-term prognosis of acute coronary syndrome. PLoS One. (2016) 16. Preisig M, Waeber G, Vollenweider P, Bovet P, Rothen S, Vandeleur C, et al. 11:e0167031. doi: 10.1371/journal.pone.0167031 The PsyCoLaus study: methodology and characteristics of the sample of a 3. Heinzer R, Vat S, Marques-Vidal P, Marti-Soler H, Andries D, Tobback N, population-based survey on psychiatric disorders and their association with et al. Prevalence of sleep-disordered breathing in the general population: the genetic and cardiovascular risk factors. BMC Psychiatry. (2009) 9:9. doi: 10. HypnoLaus study. Lancet Respir Med. (2015) 3:310–8. doi: 10.1016/S2213- 1186/1471-244X-9-9 2600(15)00043-0 17. Wilson PWF, D’Agostino R, Bhatt DL, Eagle K, Pencina MJ, Smith 4. Nakashima H, Kurobe M, Minami K, Furudono S, Uchida Y, Amenomori SC, et al. An international model to predict recurrent cardiovascular K, et al. Effects of moderate-to-severe obstructive sleep apnea on the clinical disease. Am J Med. (2012) 125:695–703. doi: 10.1016/j.amjmed.2012.0 manifestations of plaque vulnerability and the progression of coronary 1.014 atherosclerosis in patients with acute coronary syndrome. Eur Hear J Acute 18. Eagle KA, Lim MJ, Dabbous OH, Pieper KS, Goldberg RJ, Van De Werf F, Cardiovasc Care. (2015) 4:75–84. doi: 10.1177/2048872614530865 et al. A validated prediction model for all forms of acute coronary syndrome 5. Mediano O, Cano-Pumarega I, Sánchez-de-la-Torre M, Alonso-Álvarez ML, estimating the risk of 6-month postdischarge death in an international registry. Troncoso MF, García-Río F, et al. Upcoming scenarios for the comprehensive J Am Med Assoc. (2004) 291:2727–33. doi: 10.1001/jama.291.22.2727 management of obstructive sleep apnea: an overview of the spanish sleep 19. Subirana I, Sanz H, Vila J. Building bivariate tables: the comparegroups network. Arch Bronconeumol. (2020) 56:35–41. doi: 10.1016/j.arbres.2019.05. package for R. J Stat Softw. (2014) 56:11. doi: 10.18637/jss.v05 017 7.i12 6. Sánchez-de-la-Torre M, Campos-Rodriguez F, Barbé F. Obstructive sleep 20. Kaiser HF. The application of electronic computers to factor analysis. apnoea and cardiovascular disease. Lancet Respir Med. (2013) 1:61–72. doi: Educational and psychological measurement. Stat Des Res. (1960) 20:141–51. 10.1016/S2213-2600(12)70051-6 21. Harrell FE, Lee KL, Mark DB. Multivariable prognostic models: issues in 7. Marin JM, Carrizo SJ, Vicente E, Agusti AGN. Long-term cardiovascular developing models, evaluating assumptions and adequacy, and measuring outcomes in men with obstructive sleep apnoea-hypopnoea with or without and reducing errors. Stat Med. (1996) 15:361–87. doi: 10.1002/(SICI)1097- treatment with continuous positive airway pressure: an observational study. 0258(19960229)15:43.0.CO;2-4 Lancet. (2005) 365:1046–53. doi: 10.1016/S0140-6736(05)74229-X 22. R Core Team. A Language and Environment for Statistical Computing. Vienna: 8. Sánchez-de-la-Torre M, Sánchez-de-la-Torre A, Bertran S, Abad J, Duran- R Foundation for Statistical Computing (2019). Cantolla J, Cabriada V, et al. Effect of obstructive sleep apnoea and its 23. Kumbhani DJ, Wells BJ, Lincoff AM, Jain A, Arrigain S, Yu C, et al. Predictive treatment with continuous positive airway pressure on the prevalence of models for short- and long-term adverse outcomes following discharge in a cardiovascular events in patients with acute coronary syndrome (ISAACC contemporary population with acute coronary syndromes. Am J Cardiovasc study): a randomised controlled trial. Lancet Respir Med. (2019) 8:359–67. Dis. (2013) 3:39–52. doi: 10.1016/s2213-2600(19)30271-1 24. Nisha Aurora R, Crainiceanu C, Gottlieb DJ, Kim JS, Punjabi NM. Obstructive 9. Zapater A, Sánchez-de-la-Torre M, Benítez ID, Targa A, Bertran S, Torres sleep apnea during REM sleep and cardiovascular disease. Am J Respir Crit G, et al. The effect of sleep apnea on cardiovascular events in different Care Med. (2018) 197:653–60. doi: 10.1164/rccm.201706-1112OC acute coronary syndrome phenotypes. Am J Respir Crit Care Med. (2020) 202:1698–706. doi: 10.1164/rccm.202004-1127OC Conflict of Interest: The authors declare that the research was conducted in the 10. Peker Y, Glantz H, Eulenburg C, Wegscheider K, Herlitz J, Thunström E. absence of any commercial or financial relationships that could be construed as a Effect of positive airway pressure on cardiovascular outcomes in coronary potential conflict of interest. artery disease patients with nonsleepy obstructive sleep apnea: the RICCADSA randomized controlled trial. Am J Respir Crit Care Med. (2016) 194:613–20. Publisher’s Note: All claims expressed in this article are solely those of the authors doi: 10.1164/rccm.201601-0088OC and do not necessarily represent those of their affiliated organizations, or those of 11. McEvoy RD, Antic NA, Heeley E, Luo Y, Ou Q, Zhang X, et al. CPAP for the publisher, the editors and the reviewers. Any product that may be evaluated in prevention of cardiovascular events in obstructive sleep Apnea. N Engl J Med. this article, or claim that may be made by its manufacturer, is not guaranteed or (2016) 375:919–31. doi: 10.1056/NEJMoa1606599 endorsed by the publisher. 12. Azarbarzin A, Sands SA, Stone KL, Taranto-Montemurro L, Messineo L, Terrill PI, et al. The hypoxic burden of sleep apnoea predicts cardiovascular Copyright © 2022 Zapater, Solelhac, Sánchez-de-la-Torre, Gracia-Lavedan, Benitez, disease-related mortality: the osteoporotic fractures in men study and the sleep Torres, De Batlle, Haba-Rubio, Berger, Abad, Duran-Cantolla, Urrutia, Mediano, heart health study. Eur Heart J. (2019) 40:1149–57. doi: 10.1093/eurheartj/ Masdeu, Ordax-Carbajo, Masa, De la Peña, Mayos, Coloma, Montserrat, Chiner, ehy624 Mínguez, Pascual, Cortijo, Martínez, Dalmases, Lee, McEvoy, Barbé, Heinzer and 13. Pevernagie DA, Gnidovec-Strazisar B, Grote L, Heinzer R, McNicholas WT, Sánchez-de-la-Torre. This is an open-access article distributed under the terms Penzel T, et al. On the rise and fall of the apnea-hypopnea index: a historical of the Creative Commons Attribution License (CC BY). The use, distribution or review and critical appraisal. J Sleep Res. (2020) 29:e13066. doi: 10.1111/jsr. reproduction in other forums is permitted, provided the original author(s) and the 13066 copyright owner(s) are credited and that the original publication in this journal 14. Esquinas C, Sánchez-de-la Torre M, Aldomá A, Florés M, Martínez M, Barceló is cited, in accordance with accepted academic practice. No use, distribution or A, et al. Spanish sleep network. rationale and methodology of the impact of reproduction is permitted which does not comply with these terms. Frontiers in Medicine | www.frontiersin.org 10 June 2022 | Volume 9 | Article 870906