We invited male and female individuals aged 40 to 65 y to undergo a standard nocturnal polysomnography at the Sleep Laboratory at the Heart Institute (InCor), School of Medicine, University of São Paulo. These volunteers were part of a large study investigating the effects of ET on cerebral function in patients with OSA. In the present study, in order to evaluate the effect of ET on diastolic function markers in OSA patients without other major diseases, we opted to exclude major confounders such as patients with body mass index (BMI) >40 kg/m², smoking or alcohol abuse (2 or more drinks/d), cardiopulmonary disease, chronic renal disease, diabetes mellitus, history of psychiatric disorders, shift workers, hypertension (resting BP >140/90 mmHg on more than one occasion), use of medicines that affect sleep or the vascular system, and any OSA treatment. Of note, some study subjects took part in our previous research dealing with the effects of exercise training on muscle metaboreflex control¹⁶ and cardiac autonomic modulation.¹⁷The sample size calculation for the present study was based on mean and standard deviation scores reported in a previous study.¹⁸ With 18 participants in this cross-sectional study design, we anticipated a priori that our sample size would be sufficient to detect a 10% difference in diastolic function markers between groups, assuming a power of 0.80 and significance level of 0.05 (2-sided). The Institutional Committee on Human Research (InCor-HCFMUSP) approved the study (0833/10), and all subjects provided written informed consent. Patients were randomly assigned to either the untrained or exercise-trained group on a one to one basis. Full nocturnal polysomnography, maximal exercise capacity, and echocardiographic measurements were performed at baseline and the end of the study.

All participants underwent overnight polysomnography (Embla N7000, Medcare Flaga, Reykjavik, Iceland) at the Sleep Laboratory of the Heart Institute (InCor), University of São Paulo Medical School. Polysomnography was performed using standardized techniques, with a standard staging system for sleep stages used as previously described.¹⁶ Obstructive hypopneas were defined as a reduction of at least 30% of the airflow amplitude captured by the nasal cannula, lasting at least 10 s, associated with oxyhemoglobin desaturation greater than 3% and/or arousal. The sum of apnea and hypopneas per hours of sleep determined AHI. Considering the consistent evidence suggesting that mild OSA is not associated with increased cardiovascular risk,¹⁹ we used a more conservative AHI cut-off of ≥15 events/hr of sleep.

Echocardiographic measurements were performed with commercially available ultrasound equipment (Vivid E9, GE Vingmed, Horton, Norway) using a 4 MHz multifrequency transducer, with patients in the left lateral decubitus position. Echocardiographic examination was performed as follows: 1) a comprehensive 2-dimensional echocardiogram, to exclude presence of structural heart disease; 2) measurements of LV diameters and wall thickness, with resultant ejection fraction (Teichholz) and LV mass²⁰ and mass index; 3) two-dimensional guided conventional and tissue Doppler echocardiography to evaluate diastolic function and estimate LV filling pressures. For mitral valve inflow tracings, the volume sample was positioned between the free edges of mitral valve leaflets, parallel to the flow obtained from color flow mapping. Three consecutive beats were selected for offline measurements at the end of expiration. Peak E (early) wave, A (late) wave, E wave deceleration time, and isovolumic relaxation time (IVRT) were measured, and the E/A ratio was calculated. IVRT (msec) was measured as the time between closing of the aortic valve and the earliest detection of mitral flow. Tissue Doppler tracings were obtained from the apical 4-chamber view with the sample volume positioned in the basal septal and lateral mitral annulus to assess early septal diastolic myocardial velocity (e') wave and late septal diastolic myocardial velocity (a') wave. The E/e' ratio was calculated for lateral and septal velocities and the average of these measurements was used for analysis. All measurements were performed according to American Society of Echocardiography and the European Association of Cardiovascular Imaging.²⁰Echocardiography was performed by a physician blinded to the group to which each participant was assigned. Intraobserver and interobserver variability of IVRT was measured in 10 individuals by 2 independent observers.

Maximal exercise capacity was determined with a maximal progressive cardiopulmonary exercise test (SensorMedics - Vmax Analyzer Assembly, Encore 29S) on an electromagnetically braked cycle ergometer (Via Sprint 150P, Ergoline, Bitz, Germany), with work-rate constant increments (5 to 20 W/min) at 60 to 70 rpm until exhaustion as previously described.²¹ Peak oxygen uptake (peak VO₂) was defined as the maximum VO₂ attained at the end of the exercise period in which the subject could no longer maintain the cycle ergometer velocity at 60 rpm.

Supervised ET consisted of 3 weekly sessions lasting 50 mins in the first month and 60 mins from the second month onwards. Each session was distributed as follows: 5 mins of stretching; 40 mins (30 mins in the first month) of aerobic exercise (cycle ergometer), with intensity varying between the anaerobic threshold and the point of respiratory compensation (detected from the cardiorespiratory capacity test), which was measured by heart rate; 10 mins of resistance exercises; 5 mins of relaxation. The subjects in the untrained group were asked to control their non-participation in systematic physical activity programs during the control period.

Data are expressed as mean ± standard deviation (SD). The Kolmogorov–Smirnov test was used to assess the normality of the data distribution for each variable studied and Levene’s test was used to assess the homogeneity of variance for each variable studied. The categorical variable was expressed as absolute frequency. A χ² test was used to compare differences according to sex. Unpaired Student’s t tests were used to determine differences between groups (baseline and delta changes). Two-way analysis of variance with repeated measures was used to compare within and between group differences before and after intervention. In case of significance, Tukey’s post-hoc test was used. Pearson correlation analysis was used to examine the association between changes in echocardiographic parameters and changes in arousal index. P<0.05 was considered statistically significant. Statistical analysis were performed using STATISTICA 12 (StatSoft Inc., Tulsa, OK, United States) software. Interobserver and intraobserver variability were calculated as the difference between the 2 observations divided by the means of the observations and are expressed as percentages.

In the present study, 100% (72 sessions) of the post-intervention measurements were obtained by 11.65±0.86 months. The untrained group was paired with the exercise-trained group. Table 2 presents the physical, hemodynamic, and functional capacity post intervention measurements in both untrained and trained groups. There were no significant differences in physical characteristics after the intervention period. Regarding the hemodynamic parameters, there were also no significant differences in systolic BP, diastolic BP, or heart rate measurements in either group. There was a significant increase in peak VO₂ in the trained group (p=0.0001). The comparison between groups showed that the changes in peak VO₂ were greater in the trained group than those observed in the untrained group (p=0.0001).

Table 2

Parameters	Pre	Post	Change (∆)
Physical
BMI, kg/m2
Untrained	29±4	30±4	0.2±0.9
Trained	30±4	30±3	(-)0.1±1.2
Cardiovascular
SBP, mm Hg
Untrained	121±11	123±15	2.2±13.2
Trained	119±12	114±10	(-)5.4±15.7
DBP, mm Hg
Untrained	78±8	78±8	(-)0.6±10.8
Trained	78±6	77±6	(-)0.3±9.1
FC, beats/min
Untrained	65±10	66±10	0.9±12.7
Trained	68±8	66±7	(-)2.3±5.5
Functional capacity
Peak VO2, ml/kg/min
Untrained	26±6	25±7	(-)1.1±1.9
Trained	23±6	26±7*†	3.9±2.9††
Sleep Parameters
Total sleep time, min
Untrained	388±47	392±51	3.9±44
Trained	360±60	371±56	11±66
AHI, events/hr
Untrained	43±26	48±29	4.5±19
Trained	47±31	41±25	(-)5.7±13††
Arousal Index, events/hr
Untrained	31±16	33±19	1.5±8
Trained	33±17	26±14*	(-)6.1±13††
O2 desaturation, events
Untrained	32±26	44±30*	11±20
Trained	41±31	37±23†	(-)4.3±13††

Values are expressed as means ± SD. BMI: body mass index; SBP: systolic blood pressure; DBP: diastolic blood pressure; HR: heart rate; Peak VO₂: peak oxygen consumption; AHI: apnea-hypopnea index. Pre and post-intervention data were analyzed using two-way ANOVA with repeated measurements followed by Tukey’s post-hoc test. Change (Δ) data were analyzed using Unpaired Student t test. *p<0.05 within groups; † p<0.05 between groups after intervention period; ††p<0.05 between groups in mean difference in change.

Concerning the sleep parameters, AHI slightly increased in the untrained group and decreased in the trained group. There was a significant increase in O₂ desaturation in the untrained group (p=0.03), whereas no change was found in the trained group. Arousal index slightly increased in the untrained group and significantly decreased in the trained group (p<0.05). The comparisons between groups showed that the absolute arousal index value was significantly improved in the trained group in the post-intervention period. The comparisons between groups showed that there were significant differences between the groups in delta changes in AHI (p=0.03), O₂ desaturation (p=0.001), and arousal index (p=0.02).

The echocardiographic measurements (Table 3) were unchanged in untrained and trained groups (see Supplemental Content - Table S2 for more details about echocardiographic measurements before and after the intervention). IVRT slightly increased in the untrained group and decreased in the trained group. The comparison between groups showed that absolute IVRT was significantly lower (p=0.01) in the trained group in the post-intervention period (Figure 1A) with a significant difference between groups in changes after intervention (Figure 1B). Further analysis showed a significant association between the changes in IVRT and changes in arousal index in the trained group (r = -0.54, p<0.05).

Table 3

Parameters	Pre	Post	Mean difference in change (∆)
Echocardiographic
E wave, cm/sec
Untrained	74±14	75±15	1.1±12.1
Trained	72±13	74±14	1.4±10.6
A wave, cm/sec
Untrained	66±17	68±17	2.2±8.5
Trained	68±14	69±12	1.0±9.0
E/A ratio
Untrained	1±0.3	1±0.3	0±0.3
Trained	1±0.2	1±0.3	0±0.2
IVRT, msec
Untrained	89±12	96±10	6.5±17
Trained	90±15	85±11†	(-)5.1±17††
DT, msec
Untrained	209±55	211±41	1.7±48.0
Trained	205±34	210±46	5.6±39.9
Septal e’, cm/sec
Untrained	9±2	9±3	(-)0.1±2.3
Trained	9±2	9±2	(-)0.1±1.9
Septal a’, cm/sec
Untrained	9±2	10±2	1.0±2.8
Trained	10±2	10±2	(-)0.3±2.1
Septal E/e’
Untrained	9±2	9±3	0.5±2.4
Trained	8±2	9±3	0.5±2.1
Lateral e’, cm/sec
Untrained	11±3	11±3	0.2±2.3
Trained	10±2	11±2	0.6±2.0
Lateral a’, cm/sec
Untrained	10±2	10±2	0.9±2.6
Trained	10±3	10±3	(-)0.1±2.1
Lateral E/e’
Untrained	7±2	7±2	0±1.5
Trained	7±2	7±1	0±2.2
E/e’ average
Untrained	8±2	8±2	0.2±1.5
Trained	8±2	8±2	0.1±1.3

Values are expressed as means ± SD. IVRT: isovolumetric relaxation time; DT: E wave deceleration time; †p<0.05 between groups after intervention period. Pre and post-intervention data were analyzed using two-way ANOVA with repeated measurements followed by Tukey’s post-hoc test. Change (Δ) data were analyzed using Unpaired Student’s t tests. ††p<0.05 between groups in mean difference in change.

Figure 1

Interobserver variability was 2±6% and intraobserver variability was 1±6% for IVRT.