We evaluated consecutive patients referred for CR and hospitalized in the Day Treatment Cardiac Rehabilitation Department of the Upper-Silesian Medical Center (Katowice, Poland). The inclusion criteria for the study were a history of MI (up to 28 days before admission) and participation in the Polish Cardiac Society, the National Health Fund, and the Ministry of Health’s comprehensive Managed Care after Acute Myocardial Infarction (MC-AMI, KOS-ZAWAL) program [37]. The MC-AMI program provides early CR and one-year extensive cardiac care to patients diagnosed with MI, regardless of its type (ST-elevation MI or non-ST-elevation MI) and the treatment method in the acute period (procedural or conservative). According to the program’s rules, outpatient CR does not include patients with HF with significantly reduced LV ejection fraction (LVEF), especially below 35%, or in Functional Class IV in New York Heart Association (NYHA) and patients needing overnight care. In addition, patients with incomplete coronary revascularization and low coronary flow reserve by treadmill exercise stress test (EST) or with signs and symptoms of circulatory decompensation (i.e., pulmonary congestion or severe peripheral edema) were excluded from CR. They were referred for retreatment in the cardiology department. Moreover, the exclusion criteria for the study included the current treatment of SDB with positive airway pressure (PAP) or intraoral devices.

We used anthropometric and echocardiographic measurements and data concerning comorbidities from medical records collected upon admission to the rehabilitation department.

Then we assessed the presence and severity of SDB in a studied cohort with a portable polysomnography system (Alice NightOne, Philips Respironics) used as a Home Sleep Apnea Test (HSAT) [38]. The HSAT recordings were analyzed manually following the recommendations of the American Academy of Sleep Medicine [38]. We defined sleep apnea as a 90%–100% reduction in airflow through the airway for at least 10 s. Episodes of apnea with preserved breathing movements were classified as obstructive, while those without breathing movements were classified as central. We noted hypopnea when there was a decrease in the airflow by 30% that lasted $\ge$10 s and led to a $\ge$4% decrease in oxygen saturation of hemoglobin (SpO${}_2$). We assessed the severity of SDB using the respiratory event index (REI), defined as the frequency of apneas or hypopneas per hour of recording (events/h). The REI less than 5 we described as correct. We classified SDB with an REI of 5–14 events/h as mild, 15–30 events/h as moderate, and $>$30 events/h as severe.

Next, we estimated the patients’ maximum exercise capacity on a treadmill EST with electrocardiography and BP monitoring (according to the modified Bruce protocol) [39] and a six-minute walk test (6MWT) [40]. We presented exercise capacity as an estimated amount of oxygen consumed in METs, maximum heart rate (MHR), and its percentage adjusted to age (MHR%) obtained in EST and as a distance in meters (m) in a 6MWT. Then we assessed the effect of SDB on the results obtained. In the multivariate analysis, we verified the strength of the observed associations concerning age, anthropometric parameters, and LVEF.

The results were analyzed using MedCalc Version 22.006 (MedCalc Software Ltd., Ostend, Belgium). The quantitative parameters were presented as the median and interquartile range (IQR). Qualitative data were expressed as numbers and percentages. The normality of the distribution of the variables was assessed with the Kolmogorov-Smirnov test. Differences in the frequency and size of anthropometric and HSAT parameters, maximum exercise capacity parameters, LVEF, and coexisting diseases by SDB severity were calculated using the chi-square test with the Cochran–Armitage test for trend or the Kruskal–Wallis test with the Conover post-hoc analysis and the Jonckheere–Terpstra test for trend. Independent risk factors for impaired exercise capacity were determined using univariable and multivariable regression analysis, with candidate variables of age, male sex, body weight, BMI, LVEF, REI, average and maximal apnea episode duration, average and minimal SpO${}_2$ during sleep, and the percentage of total sleep time with oxyhemoglobin saturation below 90% (TST90). In multivariable regression analysis, according to the stepwise method, a variable was entered into the model if its associated significance level was 0.05 and removed from the model if its associated significance level was $>$0.10. A p-value less than 0.05 was considered statistically significant.

Impaired CRF is a risk factor for morbidity, CV complications, and death in patients with CV diseases [1, 5]. On the other hand, among CAD patients, the risk of death from any cause decreases for every MET increase in maximum exercise capacity (hazard ratio 0.83, 95% confidence interval [CI] 0.76–0.91) [5]. Better physical performance is also independently associated with fewer risk factors for CAD and a lower risk of adverse CV events [41, 42]. Therefore, all agents influencing CRF, physical activity, and lifestyle changes in post-MI patients may be relevant as potential targets for medical intervention [43]. The intervention documented to improve CRF in this group is early CR [41, 42, 44, 45, 46]. Exercise improvement of $\ge$2 METs during CR is associated with a decrease in major adverse cardiac events, CV-related hospitalizations, and unplanned coronary angiography [45]. Moreover, CR may improve the quality of life and reduce anxiety and depression in patients after MI [41, 47].

The assessment of exercise capacity is an integral part of a patient’s examination upon admission to CR [48]. The cut-off point for the minimum CRF level considered normal and associated with lower CV risk is, for men and women, approximately nine and seven METs for 40-year-olds, eight and six METs for 50-year-olds, and seven and five METs for 60-year-olds [2]. In our cohort of 60-year-old men, CV response to exercise was preserved within a range of values associated with lower CV risk (i.e., 7.15 [6.10–9.50] METs). However, in patients with severe SDB, this value was outside the normal range (i.e., 6.80 [5.60–7.90] METs), which may indicate a higher risk in this subgroup. It was documented that, with the coexistence of CAD, the CV risk associated with OSA further increases [22, 23], and impaired exercise capacity in patients with OSA and CAD has a more serious prognostic value and is associated with an increased risk of all-cause mortality [12].

Some studies indicate that SDB, predominantly OSA, affects CRF [9, 18, 28, 29, 33, 49, 50, 51]. We also found a significant relationship between SDB severity and maximum exercise capacity in our cohort of post-MI patients. However, CFR also depends on other factors than SDB itself [6, 7, 8], although often identical to SDB risk factors [19, 20], such as sex, age, overweight/obesity, altered body composition (with increased visceral and neck adiposity), diabetes, hypertension, and congestive HF. It is known that maximum exercise capacity is approximately 2 METs lower in women than in men [6] and decreases by 0.1 MET for every year of aging [7]. Generalized loss of muscle mass and obesity also reduce physical performance [8]. In our group, age, and BMI strongly correlated with the severity of the SDB. The prevalence of diabetes also correlated with an increase in the SDB severity. The NYHA class and LVEF did not differ in groups depending on the SDB severity, but the peak exercise performance in METs and the distance in 6MWT was directly proportional to LVEF. Thus, independent risk factors for exercise capacity impairment assessed by multivariate regression analysis were female sex, older age, higher BMI, and decreased LVEF, but not SDB severity. Similar conclusions were drawn by other researchers [10, 12, 52, 53], stating that among patients with OSA, the influence of sex, obesity/BMI, and comorbidities on physical performance may outweigh that of SDB.

It can be assumed that the potential and independent effect of SDB on CRF may be more pronounced in middle-aged patients with severe SDB but without significant comorbidities. Under such conditions, SDB may be the only exercise limiting factor. For other CV diseases, such as AF, ventricular arrhythmias, or HA, the adverse effect of SDB is the most significant in middle-aged cohorts under 60–65 years old [54, 55, 56]. In younger patients, due to the short duration of the disease, the lack of deep nocturnal desaturation, and higher daily physical activity, SDB may not yet cause complications or lead to symptomatic decreases in CRF [11, 57]. Conversely, for the elderly, multiple comorbidities, including HF, lung diseases, and diabetes, as well as the effects of age, obesity, and physical deconditioning, may outweigh the impact of SDB [10, 50].

Powell et al. [57] evaluated a group of military personnel whose members were approximately 41 years old and had moderate to severe OSA (an apnea-hypopnea index [AHI] of 32.7 events/h). In the group, the CRF and BP responses were correct, although the MHR trended toward a blunted response. Arora et al. [58] described male military personnel who were ten years older and mature in their careers. These individuals had a mean AHI of 29.0 events/h. Their oxygen consumption (VO${}_2$ peak) at peak exercise capacity was already reduced by 8% compared to the control group but remained in the supernormal range. De Carvalho et al. [28] found that in middle-aged women (approximately 55 years old), OSA (AHI $>$5 events/h) reveals its effect on exercise capacity only in the presence of obesity. Importantly, the impact of OSA and obesity on CRF worsening was synergistic (odds ratio 9.40, 95% CI 3.79–23.3 in obese women with OSA vs. 2.88, 95% CI 1.02–8.11 in obese women without OSA). In a group of 50-year-old individuals with moderate to severe OSA and good physical health, Beitler et al. [49] found that each 1-unit increase in log-transformed AHI was associated with a decrease in predicted peak oxygen uptake of 3.20% (95% CI 0.53–5.88). In this study, patients with OSA had the same BMI as the participants in the control group but were older. Mansukhani et al. [29] revealed a negative dose–response relationship between AHI and METs achieved at EST for a large cohort of obese men (mean BMI 32.4 kg/m${}^2$) aged approximately 56. As in our cohort, patients with SDB were older and more obese than those without SDB. However, after adjusting for confounding factors, AHI as a continuous variable remained negatively related to exercise capacity. Sonners et al. [18] showed 11.3% lower METs in 60-year-old patients with untreated OSA (AHI $>$ 5 events/h) referred to CR. The effect was significant after adjusting for sex, although the age and BMI differences appear not to be accounted for in the CRF pre-CR multivariable analysis. A study of elderly (around 70 years) hypertensives by Barbosa et al. [51] showed that moderate OSA (mean AHI 24.4 events/h) did not affect peak workloads. On the contrary, in a study by Vitacca et al. [33], conducted in patients over 60 years of age but with severe OSA (AHI 38.3 events/h), 29.2% of participants had mild exercise limitation during 6MWT (6MWT below 85% predicted and above the lower limit of normal) and severe up to 31.9% (6MWT below the lower limit of normal). However, in the group with severe OSA, the burden of nocturnal hypoxia (TST90 of 34%) and multimorbidity was high, which may have influenced the outcome. A study by Ben Saad et al. [50], carried out in patients with baseline severe OSA (AHI 62 events/h) treated with PAP, revealed the influence of age, anthropometric factors, and comorbidities on CRF. In the group successfully treated for OSA, the distance in 6MWT remained shorter compared to the control group (512 m vs. 585 m, percentage predicted of 83% vs. 100%) and rode on the presence of CV diseases, BMI, first-second-forced-expiratory-volume, and average SpO${}_2$ during sleep. It was also shorter in older patients.

Some authors point to chronotropic insufficiency and lower post-exercise heart rate recovery in OSA patients [27, 28, 51]. In addition, researchers describe dysregulation of BP response to exercise with an excessive peak or postexercise BP in SDB cohorts [27, 28, 30]. We refrained from analyzing the chronotropic competence of the sinus node and the exercise BP response because of the extensive use of pharmacotherapy in post-MI patients, including maximum tolerated doses of beta-blockers and ACE-Is, ARBs, or MRAs, which affect MHR and BP control. However, we observed the decrease in MHR proportional to SDB worsening, expressed as REI, mean, or maximal SpO${}_2$, and hypoxia burden described as TST90.

Regardless of the impact of SDB on CRF, OSA diagnosis in patients referred to CR may be relevant due to the increased likelihood of exercise-related complications and adverse events, such as AF, exercise-related ventricular arrhythmias, and nocturnal atrioventricular blocks [15, 21, 55, 56, 59]. In the era of home-based hybrid cardiac telerehabilitation [46], without direct medical supervision during exercise, the diagnosis of OSA as a risk factor for CV complications may change the patient’s qualification to the CR type. In addition, starting treatment of severe OSA with PAP before CR may contribute to the increase in the patient’s spontaneous physical activity [60], or CRF [18, 61], and an improvement of compliance, quality of life [36], and long-term CV outcome [62].

Women are poorly represented in the study group due to the lower incidence of CAD and MI among women compared to men. Another limitation is using EST and 6MWT to assess maximum exercise capacity instead of the gold standard, i.e., CTEP [63]. However, EST and 6MWT are the standard tests for qualification for the CR model routinely used in practice, and estimated CRF derived from the peak work rate is accepted as an expression of fitness by scientific societies [1, 48]. We used HSAT to diagnose the SDB’s presence and severity rather than level 1 polysomnography [38], as it is easy to use and possible to perform on an outpatient basis; it is also recommended as screening in CR [64]. Moreover, we repeated or rejected all inconclusive or technically inappropriate HSAT results to ensure the credibility of the research. We also did not have the opportunity to analyze respiratory parameters during exercise. However, the available data show that SDB does not limit ventilatory response and gas exchange during peak exercise, even in patients with high AHI and deep desaturation during sleep [65].