Prevalence and Clinical Implications of Sleep Apnea in Hypertrophic Cardiomyopathy

Key Points

Question

What is the prevalence of sleep disordered breathing (SDB) in patients with hypertrophic cardiomyopathy (HCM) and its association with myocardial remodeling, subclinical injury, and functional impairment?

Findings

In this cohort study, 154 patients with hypertrophic cardiomyopathy underwent criterion-standard polysomnography. Previously undiagnosed SDB was identified in 59% of patients and was associated with significantly increased left ventricular mass, diastolic dysfunction, and elevated troponin-T levels.

Meaning

The findings suggest that SDB is a highly prevalent yet underdiagnosed comorbidity in HCM, underscoring the rationale for randomized clinical trials to treat SDB in patients with HCM.

Abstract

Importance

Sleep disordered breathing (SDB) is a well-established contributor to cardiovascular morbidity, mediated by intermittent hypoxemia, autonomic dysregulation, and endothelial dysfunction. Patients with hypertrophic cardiomyopathy (HCM) may be especially at risk for SDB, but the clinical impact of SDB in this population remains unclear.

Objective

To define the prevalence and subtypes of SDB in HCM and examine their association with echocardiographic parameters and cardiac biomarker expression.

Design, Setting, and Participants

A prospective cohort study was conducted between April 18, 2018, and January 15, 2024, at a single tertiary referral center specializing in HCM care. Adults with HCM (left ventricular wall thickness ≥15 mm or pathogenic variants) were recruited from an institutional registry. Patients diagnosed with SDB or current pregnancy were excluded. Patients underwent polysomnography, with comparative assessment of echocardiographic, electrocardiographic, and biomarker indices. Observers were blinded to polysomnographic results. Data analysis was performed from April 11, 2024, to July 25, 2024.

Exposures

SDB classified via polysomnography using apnea-hypopnea index thresholds, with subtypes including obstructive sleep apnea and central sleep apnea and with event severity, hypoxemia, and sleep architecture disruption quantified.

Main Outcomes and Measures

Echocardiographic indices, cardiac biomarker expression, functional status, apnea-hypopnea index, and overnight hypoxemia.

Results

Among 154 patients (median [IQR] age, 60 [48-68] years; 102 [66.2%] male), 91 (59.1%) were diagnosed with SDB. Those with SDB, compared with those without SDB, had higher left ventricular mass index (median [IQR], 128 [107-161] vs 109 [96-134] g/m²; P = .03), E/e′ ratio (median [IQR], 12.5 [10.0-15.0] vs 10.0 [8.3-14.5]; P = .04), and baseline troponin-T level (median [IQR], 0.013 [0.009-0.022] vs 0.011 [0.007-0.015] ng/mL [to convert to micrograms per liter, multiply by 1]; P = .04) and greater overnight troponin-T level increases (change in median [IQR], 0.0021 [−0.0029 to 0.0062] vs 0.0002 [−0.0022 to 0.0026] ng/mL; P = .02). New York Heart Association class II or III symptoms were more common in those with SDB (48 [52.7%] vs 17 [27.0%]; P = .005). Hypertension and diabetes were more prevalent among patients with SDB than without SDB (hypertension: 67 [73.6%] vs 36 [57.1%]; P = .03; diabetes: 14 [15.4%] vs 3 [4.8%]; P = .04), whereas rates of atrial fibrillation and prior myectomy did not differ significantly between groups.

Conclusions and Relevance

This study suggests that undiagnosed SDB is highly prevalent in patients with HCM and that SDB is associated with adverse myocardial remodeling, greater diastolic dysfunction, and elevated troponin-T levels, indicating subclinical myocardial injury. SDB may contribute to HCM pathophysiology and symptom burden, supporting the rationale for randomized clinical trials to determine the impact of treating SDB on symptoms and clinical outcomes in patients with HCM.

This cohort study examines the prevalence and clinical implications of sleep disordered breathing in patients with hypertrophic cardiomyopathy.

Introduction

Hypertrophic cardiomyopathy (HCM) affects 0.2% to 0.5% of the population and is characterized by left ventricular (LV) hypertrophy independent of loading conditions.^1,2,3 While the structural and functional abnormalities of HCM have been extensively studied, the role of comorbid conditions, such as sleep disordered breathing (SDB), in modulating disease progression and outcomes remains uncertain.

SDB, encompassing both obstructive sleep apnea (OSA) and central sleep apnea (CSA), is increasingly recognized as a cardiovascular risk factor in the general population due to its association with intermittent hypoxia, sympathetic overactivity, oxidative stress, and endothelial dysfunction.^4,5,6,7 In patients with HCM, these pathophysiological effects may exacerbate disease progression by increasing LV mass, impairing diastolic function, or promoting arrhythmogenesis. However, despite the established cardiovascular risks of SDB in the general population, its true prevalence and clinical significance in HCM remain poorly characterized.

Prior studies have reported an increased burden of SDB in HCM, but most were limited by retrospective design, small sample sizes, and reliance on screening tools or home sleep testing rather than criterion-standard polysomnography (PSG).^8,9,10,11 Moreover, the associations of OSA and CSA with disease progression and adverse outcomes in HCM remain largely unknown.

This study aimed to systematically assess the prevalence and clinical implications of previously undiagnosed SDB in patients with HCM using criterion-standard overnight attended PSG. We further evaluated the associations between SDB and biomarkers of myocardial stress, echocardiographic parameters, and long-term cardiovascular outcomes.

Methods

Patient Population

This prospective cohort study received approval from the Mayo Clinic Institutional Review Board and was prospectively registered with ClinicalTrials.gov (ClinicalTrials.gov identifier: NCT03327623). Participants provided written informed consent. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.¹² Data analysis was performed from April 11, 2024, to July 25, 2024.

Eligible study participants were identified from the Mayo Clinic registry of adult patients with HCM, restricted to the tristate area (Minnesota, Iowa, and Wisconsin) to enhance study feasibility. From this eligible pool, patients were randomly sampled prior to recruitment to minimize selection bias. Diagnostic inclusion criteria included LV wall thickness of at least 15 mm assessed by echocardiography or cardiac magnetic resonance imaging, in the absence of other systemic or cardiac disease capable of explaining the degree of hypertrophy. Additionally, patients with a history of myectomy and first-degree relatives of individuals with HCM exhibiting myocardial segment thickness of at least 13 mm were included. Exclusion criteria included a prior diagnosis of SDB or current pregnancy. Study staff recruited participants between April 18, 2018, and January 15, 2024, by telephone call. Echocardiographic data from the clinical study closest to the time of participation were recorded. To more precisely evaluate the impact of comorbid SDB on echocardiographic parameters, the primary analysis focused on patients without a history of myectomy, with a supplementary analysis including all participants regardless of myectomy status.

PSG and Other Assessments

Study participants were admitted to the Clinical Trials and Research Unit, Mayo Clinic, Rochester, Minnesota, for 24 hours. Once informed consent was obtained, anthropometric and vital measurements were taken, followed by 12-lead electrocardiography (ECG). Baseline blood samples were obtained at 6 pm, followed by overnight PSG. To assess acute biochemical changes, postsleep blood samples were collected at 6 am and analyzed alongside baseline samples for high-sensitivity troponin-T level (Elecsys Troponin T Gen 5; Roche Diagnostics) and N-terminal pro–brain natriuretic peptide (NT-proBNP) level (Elecsys ProBNP II; Roche Diagnostics). Sleep studies were carried out and scored by a registered sleep technologist according to American Academy of Sleep Medicine guidelines and definitions.¹³ Measurement of respiratory effort using respiratory inductance plethysmography bands permitted delineation between OSA and CSA.

Episodes of at least 30% reduction in airflow lasting longer than 10 seconds, accompanied by an oxygen desaturation of at least 4%, are classified as hypopnea, whereas episodes of at least 90% airflow reduction lasting longer than 10 seconds are classified as apnea. The total number of these events was averaged over total sleep duration to calculate the apnea-hypopnea index (AHI). An AHI of 5 to 14 events per hour defines mild OSA; 15 to 29 events per hour, moderate OSA; and 30 or more events per hour, severe OSA. The diagnosis of CSA requires at least 50% of recorded respiratory events to result from diminished respiratory drive, indicating an absence of respiratory effort.¹³

SDB severity was further assessed by the minimum oxygen saturation during sleep and the cumulative hypoxic burden, quantified as the percentage of total sleep time with oxygen saturation less than 90%. Patients diagnosed with SDB received a summary report of their PSG findings and were advised to consult their primary physician for further evaluation.

Statistical Analysis

Continuous variables, predominantly nonnormally distributed, are presented as median and IQR, while categorical variables are reported as frequency and percentage. Between-group comparisons were performed using the Wilcoxon rank sum test for continuous variables and Pearson χ² test for categorical variables.

Given that male sex, elevated body mass index (BMI; calculated as weight in kilograms divided by height in meters squared), and older age are established risk factors for SDB, these variables were included as covariates in regression modeling.¹⁴ Multiple linear regression with stepwise selection was used to examine associations between SDB status, biomarker levels, and echocardiographic parameters, accounting for potential confounders, effect modification, and collinearity. Statistical analyses were conducted using BlueSky Statistics version 7.4 (BlueSky Statistics), with a 2-sided P < .05 considered statistically significant.

Results

Patient Characteristics and PSG Results

Three hundred patients were selected from the registry using random sampling. Of 300 patients screened, 74 were ineligible due to a prior diagnosis of SDB and were instead invited to participate in an ancillary study. An additional 11 patients were excluded based on alternative diagnoses—9 with hypertensive cardiomyopathy and 2 with cardiac amyloidosis—leaving 215 eligible participants. Contact could not be established with 11 individuals despite 3 call attempts, and 48 declined study participation. Among the 156 enrolled participants, 1 was unable to tolerate PSG and withdrew from the study entirely, while another tested positive for COVID-19 on arrival and chose not to reschedule their participation. The final study cohort thus comprised 154 patients who successfully completed the study protocol.

Of 154 study participants (102 [66.2%] male; median [IQR] age, 60 [48-68] years), 91 (59.1%) were diagnosed with SDB. Among those with SDB, 38 (41.8%) had mild OSA, 22 (24.2%) had moderate OSA, 18 (19.8%) had severe OSA, and 13 (14.3%) had CSA, yielding an OSA to CSA ratio of 6 to 1. The prevalence of SDB was comparable between those with and without prior myectomy. Patients diagnosed with SDB, compared with those without SDB, had a higher prevalence of diabetes (14 [15.4%] vs 3 [4.8%]; P = .04) and hypertension (67 [73.6%] vs 36 [57.1%]; P = .03) and higher in-office systolic and diastolic blood pressures (systolic: median [IQR], 121 [113-129] vs 112 [106-121] mm Hg; P < .001; diastolic: median [IQR], 77 [68-86] vs 73 [68-78] mm Hg; P = .04). New York Heart Association class II or III symptoms were more common in those with SDB than in those without SDB (48 [52.7%] vs 17 [27.0%]; P = .005). Those with SDB were more likely to describe apneic episodes and to designate their sleep quality as poor (Table 1). A comparison of demographic characteristics including study participants (n = 154), those excluded due to a prior diagnosis of SDB (n = 74), and potentially eligible individuals who did not enroll (n = 59), due to either being unreachable or declining participation, is provided in eTable 1 in Supplement 1. Demographic characteristics were comparable between enrolled participants and those who declined participation or could not be contacted. In contrast, individuals with a known diagnosis of SDB were generally older, had a higher BMI, and were more often male, reflecting associations also observed within the enrolled study cohort.

Table 1. Clinical Characteristics.

Characteristic	No SDB (n = 63)	SDB (n = 91)	P value
Age, median (IQR), y	54.3 (41.3-63.5)	63.3 (54.5-69.6)	<.001a
Sex, No. (%)
Female	27 (42.9)	25 (27.5)	.047a
Male	36 (57.1)	66 (72.5)
BMI, median (IQR)	28.1 (25.4-32.0)	31.0 (28.7-35.8)	<.001a
Circumference, median (IQR), cm
Waist	96 (90-105)	107 (101-116)	<.001a
Hip	107 (102-115)	112 (104-118)	.13
Neck	38 (35-41)	41 (39-44)	<.001a
Waist to hip ratio, median (IQR)	0.92 (0.86-0.96)	0.97 (0.93-1.02)	<.001a
Non-Hispanic Whiteb	55 (87.3)	88 (96.7)	.90
Postmenopausal, No./total No. female (%)	17/27 (63.0)	18 (72.0)	.49
Hypertension, No. (%)	36 (57.1)	67 (73.6)	.03a
Diabetes, No. (%)	3 (4.8)	14 (15.4)	.04a
Sudden cardiac arrest, No. (%)	1 (1.6)	3 (3.3)	.51
Atrial fibrillation, No. (%)	20 (31.7)	38 (41.8)	.21
History of stroke or TIA, No. (%)	5 (8.1)	8 (8.8)	.87
Myocardial infarction, No. (%)	6 (9.5)	14 (15.4)	.29
Myectomy, No. (%)	25 (39.7)	37 (40.7)	.93
Implantable cardioverter-defibrillator, No. (%)	19 (30.2)	20 (22.0)	.25
β-Blocker, No. (%)	31 (55.4)	60 (68.2)	.12
Calcium channel blocker, No. (%)	9 (16.7)	23 (26.4)	.18
Current or former smoker, No. (%)	26 (41.3)	48 (52.7)	.16
Alcohol use
No. (%)	43 (69.4)	64 (70.3)	.90
Drinks, median (IQR), No./wk	2 (0-4)	2 (0-7)	.54
NYHA class, No. (%)
I	46 (73.0)	43 (47.3)	.01a
II	11 (17.5)	35 (38.5)
III	6 (9.5)	13 (14.3)
Pre-PSG questionnaire answer of yes, No. (%)
Do you snore?	38 (61.3)	62 (68.1)	.38
Do you stop breathing at night?	13 (21.0)	33 (36.3)	.04a
Is your sleep quality poor?	18 (29.0)	44 (48.4)	.02a
Are you sleepy in the daytime?	23 (37.1)	43 (47.3)	.21
In-office mean blood pressure, median (IQR), mm Hg
Systolic	112 (106-121)	121 (113-129)	<.001a
Diastolic	73 (68-78)	77 (68-86)	.04a
In-office heart rate, median (IQR), beats/min	62 (57-68)	63 (57-69)	.59

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); NYHA, New York Heart Association; PSG, polysomnography; SDB, sleep disordered breathing; TIA, transient ischemic attack.

^a Statistically significant difference (P < .05).

^b Race and ethnicity data were self-reported. To protect participant confidentiality and comply with reporting guidelines, only the majority racial and ethnic group is reported. All other groups, including individuals identifying as Native American, non-Hispanic Asian, non-Hispanic Black, Pacific Islander, other, or unknown, were collapsed into a single category and not displayed due to cell sizes fewer than 10.

The median (IQR) AHI for the mild, moderate, and severe OSA and CSA groups were 9.4 (7.3-11.6), 21.4 (16.3-25.2), 46.7 (37.5-56.1), and 40.1 (13.6-56.9) events per hour, respectively. The median (IQR) heart rate during sleep was higher in patients with SDB than in those without SDB (61 [56-66] vs 58 [54-63] beats/min; P = .02). The severity of sleep apnea, measured by AHI, correlated with the degree of overnight hypoxia, measured by median (IQR) nadir oxygen saturation as measured by pulse oximetry (SpO₂), which decreased to 80% (76%-84%) in those with severe OSA (Pearson correlation coefficient, −0.67; P < .001). Similarly, patients with severe OSA spent a median (IQR) 6.6% (3.9%-34.1%) of sleep with SpO₂ under 90%, compared with a median (IQR) of 0.0% (0.0%-0.1%) in those without SDB (Table 2).

Table 2. Polysomnographic Characteristics.

Characteristic	No SDB (n = 63)	SDB (n = 91)	P value
AHI, median (IQR), No. of events/h	2.0 (0.8-3.3)	15.8 (9.9-32.2)	NA
Mild OSA	NA	9.4 (7.3-11.6)	NA
Moderate OSA	NA	21.4 (16.3-25.2)	NA
Severe OSA	NA	46.7 (37.5-56.1)	NA
CSA	NA	40.1 (13.6-56.9)	NA
OAI, median (IQR), No. of events/h	0.0 (0.0-0.1)	0.3 (0.0-1.8)	<.001a
CAI, median (IQR), No. of events/h	0.2 (0.0-0.6)	1.8 (0.3-5.0)	<.001a
Hypopnea index, median (IQR), No. of events/h	1.0 (0.3-2.7)	11.5 (6.1-18.0)	<.001a
TST, median (IQR), min	355 (305-411)	346 (304-384)	.27
Sleep efficiency, median (IQR), %	78 (71-90)	78 (69-86)	.42
REM sleep, median (IQR), %	24 (21-30)	19 (14-25)	.01a
SpO2, median (IQR), %
Mean during sleep	96 (95-96)	94 (93-95)	<.001a
Mild OSA	NA	94 (93-95)	NA
Moderate OSA	NA	94 (93-94)	NA
Severe OSA	NA	94 (91-94)	NA
CSA	NA	95 (93-96)	NA
Minimum during sleep	90 (88-92)	85 (80-87)	<.001a
Mild OSA	NA	86 (84-88)	NA
Moderate OSA	NA	84 (80-85)	NA
Severe OSA	NA	80 (76-84)	NA
CSA	NA	86 (80-89)	NA
T90, median (IQR), % of sleep	0.0 (0.0-0.1)	1.6 (0.5-4.8)	<.001a
Mild OSA	NA	0.6 (0.3-1.9)	NA
Moderate OSA	NA	1.8 (1.2-4.1)	NA
Severe OSA	NA	6.6 (3.9-34.1)	NA
CSA	NA	2.2 (0.0-16.3)	NA
T90 ≥1% of sleep, No. (%)	2 (3.2)	56 (61.5)	<.001a
Mean heart rate during sleep, median (IQR), beats/min	58 (54-63)	61 (56-66)	.02a
Mild OSA	NA	61 (55-66)	NA
Moderate OSA	NA	60 (56-67)	NA
Severe OSA	NA	63 (59-66)	NA
CSA	NA	64 (53-67)	NA

Abbreviations: AHI, apnea-hypopnea index; CAI, central apnea index; CSA, central sleep apnea; NA, not applicable; OAI, obstructive apnea index; OSA, obstructive sleep apnea; REM, rapid eye movement; SDB, sleep disordered breathing; SpO₂, oxygen saturation as measured by pulse oximetry; T90, percentage of sleep with SpO₂ less than 90%; TST, total sleep time.

^a Statistically significant difference (P < .05).

ECG and Biochemical Findings

Excluding patients with a paced rhythm (n = 17) or atrial fibrillation (AF) (n = 19), 118 ECGs were analyzed (ECGs from patients with paroxysmal AF or a device in situ but in sinus rhythm at the time of the study were included). Patients with SDB had higher resting heart rates (median [IQR], 65 [59-70] vs 60 [54-66] beats/min; P = .005), longer QRS duration (median [IQR], 108 [98-136] vs 102 [92-119] milliseconds; P = .04), longer QTc interval (median [IQR], 462 [441-478] vs 448 [425-471] milliseconds; P = .008), and leftward-deviated QRS axis (median [IQR], −10 [−30 to 22] vs 26 [−9 to 51] degrees; P < .001). Adjusting for between-group variance in age, sex, and BMI, differences in QRS axis remained significant (SDB presence effect size, −28.3 degrees [95% CI, −48.3 to −8.3 degrees]; adjusted P = .006).

Patients with SDB, compared with those without SDB, had greater baseline troponin-T levels (median [IQR], 0.013 [0.009-0.022] vs 0.011 [0.007-0.015] ng/mL [to convert to micrograms per liter, multiply by 1]; P = .04) and overnight troponin-T level increases (change in median [IQR], 0.0021 [−0.0029 to 0.0062] vs 0.0020 [−0.0022 to 0.0026] ng/mL; P = .02). Adjusting for age, sex, and BMI, SDB was associated with a morning troponin-T level increase of 0.0049 ng/mL (95% CI, 0.0012-0.0086 ng/mL; adjusted P = .01) compared with those without SDB. Baseline NT-proBNP levels in patients with SDB were similar to those in patients without SDB (median [IQR], 295 [165-653] vs 355 [163-667] pg/mL; P = .98), as were postsleep levels (median [IQR], 285 [150-590] vs 338 [144-656] pg/mL; P = .94), including after adjustment. ECG and biochemical test results are summarized in eTable 2 in Supplement 1.

Echocardiographic Characteristics

The median interval between standard-of-care ECG and study-related PSG was 10.7 months (IQR, 2.3-15.1 months). Rates of atrial fibrillation and prior myectomy did not differ significantly between groups. Of the 92 participants without history of myectomy, patients with SDB had a lower cardiac index (median [IQR], 2.7 [2.5-3.1] L/min/m² with SDB vs 3.1 [2.7-3.5] L/min/m² without SDB; P = .04) and stroke volume (median [IQR], 49 [40-52] mL with SDB vs 51 [48-56] mL without SDB; P = .04) and higher LV mass (indexed to body surface area) compared with those without SDB (median [IQR], 128 [107-161] vs 109 [96-134] g/m²; P = .03) (Table 3). After adjusting for age, sex, and BMI, differences in cardiac index (adjusted P = .03) and LV mass (adjusted P = .03) remained significant. Median LVEF was at least 65% in the non-SDB, OSA, and CSA groups. All groups had an E/e′ ratio of at least 10 (median [IQR], 12.5 [10.0-15.0] in those with SDB vs 10.0 [8.3-14.5] in those without SDB; P = .04), and the CSA group demonstrated a higher E/A ratio (≥1.5). Indexed LV mass was markedly higher in those with CSA compared with those with OSA (median [IQR], 176 [154-189] vs 116 [106-148] g/m²; P = .02). An additional analysis including all study patients found that the only significant difference was in the lateral mitral annular E/e′ ratio, which was higher in patients diagnosed with SDB (eTable 3 in Supplement 1).

Table 3. Echocardiographic Characteristics, Excluding Patients With a History of Myectomy.

Characteristic	No SDB (n = 38)	SDB (n = 54)	P value
Age, median (IQR), y	52.1 (41.1-63.1)	64.5 (57.6-69.7)	<.001a
Sex, No. (%)
Female	16 (42.1)	14 (25.9)	.10
Male	22 (57.9)	40 (74.1)
BMI, median (IQR)	28.9 (25.4-34.5)	31.0 (28.5-32.9)	.20
Body surface area, median (IQR), m2	2.1 (1.9-2.3)	2.1 (2.0-2.3)	.67
LVEF, median (IQR), %	67 (63-71)	66 (63-70)	.67
Maximal wall thickness, median (IQR), mm	17 (15-20)	18 (16-21)	.55
LVPWd, median (IQR), mm	11 (10-12)	12 (10-13)	.19
LVEDd, median (IQR), mm	48 (45-51)	50 (47-52)	.15
LV mass index, median (IQR), g/m2	109 (96-134)	128 (107-161)	.03a
LVOT peak velocity, median (IQR), cm/s	1.2 (1.0-1.6)	1.2 (1.0-1.8)	.80
LVOT gradient, median (IQR), mm Hg	12 (5-46)	14 (6-49)	.94
Obstructive HCM, No. (%)b	14 (36.8)	21 (38.9)	.93
Cardiac index, median (IQR), L/min/m2	3.1 (2.7-3.5)	2.7 (2.5-3.1)	.04a
Stroke volume, median (IQR), mL	51 (48-56)	49 (40-52)	.04a
E/A ratio, median (IQR)	1.2 (1.0-1.6)	1.1 (0.8-1.5)	.15
E/e′ ratio, median (IQR)
Medial	10.0 (8.3-14.5)	12.5 (10.0-15.0)	.04a
Lateral	7.0 (5.6-9.2)	9.5 (7.5-12.0)	.04a
RVSP, median (IQR), mm Hg	30 (24-36)	30 (27-35)	.36
LA volume index, median (IQR), mL/m2	37 (30-45)	39 (32-48)	.61
Mitral regurgitation ≥moderate, No. (%)	7 (18.4)	5 (9.3)	.24
SAM, No. (%)	16 (42.1)	23 (42.6)	.40

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); HCM, hypertrophic cardiomyopathy; LA, left atrial; LV, left ventricular; LVEDd, left ventricular end diastolic diameter; LVEF, left ventricular ejection fraction; LVOT, left ventricular outflow tract; LVPWd, left ventricular posterior wall thickness in diastole; RVSP, right ventricular systolic pressure; SAM, systolic motion of the mitral valve; SDB, sleep disordered breathing.

^a Statistically significant difference (P < .05).

^b Refers to a resting or provoked LVOT gradient of at least 30 mm Hg.

Characteristics Predictive of SDB in HCM

To identify variables most strongly associated with SDB in this cohort, multivariable logistic regression with stepwise selection was performed, using a threshold of P < .10 for inclusion in the final model. This found age, BMI, waist to hip ratio, and morning troponin-T level to be most predictive of SDB in this sample. In predicting the presence of SDB within this cohort, this model had an area under the curve of 0.82, sensitivity of 60%, specificity of 84%, positive predictive value of 0.74, and negative predictive value of 0.73 (Table 4).

Table 4. Univariable and Multivariable Regression Analyses Predictive of Sleep Disordered Breathing Presence.

Variable	Univariable		Predictive model
	OR (95% CI)	P value	OR (95% CI)	P value
Age, y
Median	1.06 (1.03-1.09)	<.001a	1.08 (1.04-1.12)	<.001a
≥50	5.12 (2.37-11.03)	<.001a	NA	NA
Male	1.98 (1.00-3.90)	.048a	NA	NA
BMI
Median	1.11 (1.04-1.18)	.002a	1.17 (1.09-1.27)	<.001a
≥30	4.48 (2.25-8.95)	<.001a	NA	NA
Waist to hip ratio per 0.1	3.12 (1.94-5.32)	<.001a	2.07 (1.15-3.87)	.02a
Neck circumference, cm	1.15 (1.06-1.26)	<.001a	NA	NA
NYHA class II or III	3.02 (1.51-6.03)	.002a	NA	NA
Hypertension	2.09 (1.06-4.15)	.03a	NA	NA
Diabetes	3.64 (1.00-13.23)	.049a	NA	NA
Atrial fibrillation	1.54 (0.79-3.03)	.21	NA	NA
Septal thickness, mm	0.99 (0.94-1.06)	.87	NA	NA
LVEF, %	1.03 (0.98-1.07)	.29	NA	NA
LVOT peak velocity, cm/s	1.17 (0.82-1.68)	.39	NA	NA
LA volume index, mL/m2	1.02 (0.99-1.04)	.22	NA	NA
Morning troponin-T level ≥0.02 ng/mL	3.21 (1.44-7.16)	.004a	NA	NA
QRS duration per ms	1.02 (1.01-1.03)	.001a	NA	NA
QTc ≥460 ms	1.02 (1.01-1.03)	.004a	NA	NA
Snoring	1.56 (0.81-3.02)	.18	NA	NA
Stop breathing at night	2.13 (1.04-4.52)	.04a	NA	NA
Poor sleep quality	2.30 (1.19-4.53)	.01a	NA	NA
Sleepy in daytime	1.41 (0.75-2.70)	.29	NA	NA

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); LA, left atrial; LVEF, left ventricular ejection fraction; LVOT, left ventricular outflow tract; NA, not applicable; NYHA, New York Heart Association; OR, odds ratio.

^a Statistically significant difference (P < .05).

Discussion

This study provides evidence that previously undiagnosed SDB is prevalent in patients with HCM, with nearly 60% of our cohort affected predominantly by OSA. While the high prevalence of SDB in HCM is consistent with prior reports primarily using overnight oximetry, our study advances existing knowledge by using criterion-standard attended PSG to establish definitive diagnoses. Second, while OSA was more prevalent overall, the relative proportion of CSA among those with SDB was high. Last, we observed significant associations between SDB and adverse cardiac remodeling, elevated troponin-T levels, and functional impairment, particularly in CSA, emphasizing the need for a nuanced understanding of SDB subtypes in this population.

In this cohort of patients with HCM, the prevalences of hallmark symptoms typically used to screen for SDB— such as self-reported snoring and excessive daytime sleepiness—were similar between those with and without SDB, illuminating a limitation of symptom-based screening in this population and underscoring the potential for underdiagnosis when relying on traditional indicators alone. Although CSA was less common, it was associated with the most adverse structural and biomarker profiles, raising important questions about its mechanistic role and whether this subset may represent a particularly high-risk phenotype that warrants further investigation.

Estimates of the prevalence of OSA in the general population vary considerably depending on study design and location, applied diagnostic thresholds, and population characteristics, such as age, sex, and BMI. In the United States, the American Association of Sleep Medicine’s 2012 diagnostic criteria are most commonly used.¹⁵ The high prevalence of SDB in our HCM cohort, 59%, is higher than most estimates from the general adult population, where the prevalence is 10% to 50%, with higher rates in older individuals, men, and obese individuals.^16,17,18,19 The high prevalence of SDB in our cohort aligns with reported rates of 50% to 80% among patients with heart failure, where SDB is recognized as a major comorbidity associated with worse outcomes, including higher rates of hospitalization and mortality, highlighting its clinical relevance in the management of heart failure.^20,21,22,23

Previous estimates of the burden of SDB in patients with HCM are also variable. The majority of prior studies of SDB in HCM have been either retrospective or based on findings derived from overnight oximetry testing, which detects significantly fewer variables of interest than criterion-standard, in-laboratory, attended PSG^{8,24,25,26,27} and, importantly, cannot differentiate between OSA and CSA. This inevitably leads to significant limitations when compared with PSG, as a number of key sleep-staging channels of data are not typically monitored, such as electroencephalography, electro-oculography, or electromyography, resulting in the use of recorded time rather than sleep time as the denominator by which total respiratory events are indexed per hour.²⁸ By prospectively screening a well-characterized cohort with HCM, we provide the most comprehensive assessment to date, reinforcing that SDB remains frequently undiagnosed in this population.

SDB could contribute to disease progression in HCM through multiple pathways. Frequent nocturnal desaturations and autonomic instability from OSA are well-documented drivers of sympathetic activation, which in turn can exacerbate LV hypertrophy and outflow tract obstruction.^29,30 Additionally, elevated troponin-T levels, both at baseline and following apneic sleep, suggest that subclinical myocardial injury may be a consequence of repetitive hypoxic stress and increased myocardial workload.³¹ In our study, patients with SDB, particularly those with CSA, demonstrated significantly greater LV mass and more advanced diastolic dysfunction, independent of hypertension. Notably, increased LV mass has been reported to be a stronger predictor of HCM-related mortality than maximal wall thickness.³² Impaired ventricular relaxation and elevated left atrial pressure in HCM may contribute to delayed effective circulation time between the lungs and central chemoreceptors, resulting in a mismatch between arterial carbon dioxide levels and chemoreceptor feedback.³³ This mismatch can destabilize ventilatory control—a phenomenon referred to as heightened loop gain—and contribute to disordered breathing patterns.³⁴ Further supporting this theory, animal studies have shown that elevated left atrial pressure is associated with reduced ventilatory responsiveness to carbon dioxide below the eupneic threshold, suggesting impaired chemoreflex sensitivity under conditions of chronic cardiac loading.³⁵ This pathophysiology may help explain the lower OSA to CSA ratio of 6:1 observed in our cohort, compared with the approximately 10:1 ratio reported in the general population.¹⁹ Similarly, rostral fluid shifts during sleep may promote pharyngeal fluid accumulation, increasing tissue pressure and contributing to upper airway narrowing and resistance. This physiological mechanism can reduce airway patency and increase susceptibility to obstructive events during sleep, thereby contributing to the development or worsening of OSA.³⁶ Supporting this mechanism, Solin et al³⁷ demonstrated a linear association between pulmonary capillary wedge pressure and the AHI.

The association between SDB and ECG abnormalities in this cohort and the well-established association between OSA and AF suggest that SDB could contribute to arrhythmic burden through both structural and electrophysiological mechanisms.^38,39,40,41 Several studies have linked SDB to changes in cardiac structure and function, including LV hypertrophy and QRS widening.^42,43,44,45 Large intrathoracic pressure swings during forceful inspiration against a closed airway increase right ventricular preload, LV afterload, and transmural pressure gradients, driving further hypertrophy.⁴⁶ Prolongation of the QT interval, leftward QRS axis deviation, and increased resting heart rate in affected individuals may signal abnormal ventricular filling and diastology and increased arrhythmic and sudden cardiac death vulnerability.⁴⁷

In our study, patients with higher AHI values experienced more severe overnight hypoxemia and were more likely to experience higher New York Heart Association class symptoms. Notably, these associations persisted despite similar measures of LV function, outflow tract gradient, septal thickness, right ventricular systolic pressure, mitral regurgitation, and outflow obstruction across groups. Collectively, these data highlight a substantial burden of previously unrecognized SDB in patients with HCM, associated with markers of physiological stress and functional impairment.

Despite the similarly high prevalence of SDB in heart failure, treatment with positive airway pressure has not been shown to improve cardiovascular outcomes. Some predominantly observational studies have reported improvements in LV outflow tract obstruction,³⁰ symptom burden,^{48,49,50,51,52} and AF recurrence with continuous positive airway pressure (CPAP) therapy in patients with cardiovascular disease and SDB,^53,54 but large randomized clinical trials have yielded neutral or harmful findings. For instance, the Sleep Apnea Cardiovascular Endpoints (SAVE) trial found no significant reduction in major cardiovascular events with CPAP compared with usual care in patients with moderate to severe OSA,⁵⁵ and multiple meta-analyses have similarly failed to demonstrate a mortality benefit.^56,57 In CSA, adaptive servo-ventilation (ASV) resulted in increased mortality in patients with heart failure with reduced ejection fraction.⁵⁸ Subsequent studies, such as Adaptive Servo-Ventilation for Sleep-Disordered Breathing in Patients With Heart Failure With Reduced Ejection Fraction (ADVENT-HF),⁵² have found newer-iteration ASV—designed to control both CSA and OSA—to be safe, albeit without any significant impact on cardiovascular event rates. As such, while CPAP and ASV may improve sleep quality, daytime function, and quality of life, their utility in mitigating cardiovascular risk remains uncertain. Randomized clinical trials in patients with HCM are needed to determine the effectiveness of positive airway pressure or other interventions on disease progression and clinical outcomes. Given that our findings show that SDB in HCM was associated with adverse structural and biomarker profiles, identifying the presence and type of SDB in patients with HCM may aid prognostication, clinical monitoring, and recommendations regarding risk-factor modification, particularly for treatment of obesity and hypertension.^59,60,61,62

Strengths and Limitations

The present study has several strengths, including the prospective design and use of criterion-standard, attended, in-laboratory PSG, permitting the most accurate diagnosis of SDB and accurate differentiation between its subtypes. These data represent a large, single-site referral center experience with a well-defined group of patients with HCM.

There are several limitations of the present study. The study sample was drawn from a single center that is a tertiary referral center for patients with HCM, a potential source of selection and referral bias, which we sought to mitigate by limiting recruitment to residents of the tristate area (Minnesota, Wisconsin, and Iowa). All hypopneas were classified as obstructive, which reflects standard clinical practice but may limit the ability to distinguish central from obstructive hypopneas, potentially leading to underestimation of the true burden of CSA in patients with HCM.⁶³ Approximately 40% of the study cohort had undergone septal myectomy, reflecting the tertiary care setting and potentially limiting generalizability; however, the prevalence of SDB was similar regardless of myectomy status, and primary echocardiographic analyses were conducted excluding these patients, with supplementary analyses including all patients irrespective of myectomy status. Echocardiograms were obtained as part of standard clinical care rather than study protocol, resulting in variable timing relative to PSG and potential for temporal bias; however, the median (IQR) interval between echocardiography and PSG was 10.7 (2.3-15.1) months, providing reasonable proximity for comparative analysis. Screening for eligibility to participate in the study consisted of review of hospital and International Classification of Diseases diagnosis codes, medical records, and echocardiographic data for phenotypic data, but genotypic data were not readily available for most patients because of patient choice and significant out-of-pocket costs. The proposed predictive model requires validation in an external cohort. Also, most of the participants were of European descent, limiting generalizability of study findings to non-White populations.

Conclusions

This study highlights that SDB is highly prevalent and frequently undiagnosed in patients with HCM, representing a significant comorbidity with potentially important clinical implications. Given its association with myocardial injury, adverse cardiac remodeling, and functional impairment, systematic evaluation for SDB should be considered in patients with HCM exhibiting high-risk clinical features, such as comorbid hypertension, unexplained or progressive functional decline, or disproportionately severe diastolic dysfunction, including in the absence of traditional symptoms typically used to guide screening. Future studies aimed at improving cardiovascular health and quality of life in HCM should investigate the impact of SDB treatment on cardiac remodeling, arrhythmic risk, and long-term clinical outcomes. Given that randomized clinical trials in heart failure have shown no impact or increased mortality in patients with heart failure treated with positive airway pressure, randomized clinical trials are needed in patients with HCM and SDB to determine the effectiveness of such approaches before recommendations for treatment of SDB can be made, particularly in the absence of daytime sleepiness or impaired quality of life.^{64,65,66,67,68,69,70}

Supplement 1.

eTable 1. Demographic Characteristics of Participants, Excluded, and Nonenrolled Patients

eTable 2. ECG and Biochemical Characteristics (Sinus Rhythm/Non-Paced) No SDB vs Study Diagnosed SDB

eTable 3. Echocardiographic Characteristics (Myectomy Included)

Supplement 2.

Data Sharing Statement