Randomized controlled trial of an oral appliance (SomnoDent) for sleep‐disordered breathing and cardiac function in patients with heart failure

Abstract

In patients with heart failure (HF), sleep‐disordered breathing (SDB) is a common comorbidity and a risk factor for poor clinical outcomes. SDB can be ameliorated by continuous positive airway pressure; however, inadequate adherence remains a major cause of treatment failure. On the other hand, the efficacy of oral appliance (OA) has been proved in orthodontics and otolaryngology, although the efficacy of OA in patients with HF remains to be elucidated. This trial aims to determine the efficacy of OA for SDB in patients with HF. Patients with HF undergoing optimal medical therapy who were diagnosed as having SDB (apnea‐hypopnea index [AHI] ≥ 10 and percentage of central AHI per total AHI ≤ 70%) by using polysomnography (PSG) will be enrolled in the present study. Either patients with HF with reduced ejection fraction (left ventricular ejection fraction [LVEF] ≤ 50%) or HF with preserved ejection fraction (history of hospitalization because of acute decompensated HF) or plasma B‐type natriuretic peptide (BNP) level ≥ 100 pg/mL will be included. Each patient will be randomly assigned into active OA or sham OA. PSG, laboratory, and echocardiographic data will be obtained after 3 months of intervention. The main outcome measures are AHI, plasma BNP, and E/e’ determined with echocardiography. Furthermore, overnight urinary catecholamine, 6‐min walk distance, Epworth sleepiness scale, and health‐related quality of life will be assessed simultaneously.

This trial started on April 1, 2017, and the projected end date is March 31, 2019. This study was registered in the University Hospital Medical Information Network (UMIN000025731).

1. INTRODUCTION

Sleep‐disordered breathing (SDB) is a common comorbidity in a number of cardiovascular diseases. In patients with heart failure (HF), SDB is associated with worse prognosis and increased mortality.1, 2 Obstructive sleep apnea (OSA) is caused by upper airway obstruction, whereas central sleep apnea (CSA) is caused by instability of the respiratory center.2

Positive airway pressure therapy can improve prognosis by suppressing SDB2; however, sufficient therapy duration is occasionally difficult because of patient discomfort at the initial phase of treatment.3 Adaptive servo‐ventilation used to be applied for central‐dominant SDB in patients with left ventricular (LV) systolic dysfunction; however, a recent study indicated that it might increase mortality.4

On the other hand, the use of an oral appliance (OA) is a simple method to treat SDB.5 The efficacy of OA for OSA has been established in the fields of orthodontics and otolaryngology.6 Some studies suggested that the forward positioning of the mandible releases upper airway obstruction,7 which lowers the B‐type natriuretic peptide (BNP) level by suppressing sympathetic nerve activation and decreasing the alterations in intrathoracic pressure.8 However, limited data are currently available about its efficacy in HF patients with OSA. In addition, whether OA can successfully treat either OSA or CSA in patients with LV diastolic dysfunction remains to be elucidated. Thus, this prospective randomized study was designed to examine if treatment with OA improves SDB and cardiac function.

1.1. Purposes

This study aims to investigate (i) whether OA (SomnoDent; SomnoMed, Australia) can effectively suppress SDB in patients with HF, and (ii) whether SDB suppression with OA can improve cardiac function.

1.2. Hypothesis

OA can suppress SDB even in patients with HF and improve cardiac function.

2. METHODS

2.1. Study design

This study will be conducted in Juntendo University Hospital, Tokyo, Japan (Appendix S1, Supporting information). This is a single‐blinded randomized, sham‐controlled interventional trial with or without OA for SDB in patients with HF. SDB will be diagnosed in patients with HF by using polysomnography (PSG) at the hospital. OA will be set for each patient inside the dental clinic of Juntendo University Hospital, after checking whether oral hygiene and dental treatment are required. The supplied OA will be adjusted carefully by experienced dentists at the follow‐up visit (acclimatization phase for 2 to 4 weeks). Suppression of sleep apnea will be confirmed using a portable sleep breathing diagnostic device (WatchPAT) in 2 weeks. The target apnea‐hypopnea index (AHI)/oxygen desaturation index (ODI) is <10/hour or a > 50% reduction.9 Each patient will be randomized into the active OA group or the sham OA group (only the mandibular part with recording media) after a washout period of 1 week. The follow‐up PSG will be scheduled in 3 months. All measurements, including cardiac parameters such as plasma BNP level or E/e’ determined using echocardiography, will be obtained simultaneously. It should be noted that we plan to conduct a 3‐dimensional substudy of the upper airway on and off OA by using computed tomography10 (Appendix S2).

2.2. Study subjects, groups, and sample size

2.2.1. Subjects

Patients with symptoms or signs of HF from Juntendo University Hospital (Tokyo, Japan) will be recruited. The patients should have HF in the chronic phase (HF diagnosis and use of beta‐blockers for ≥3 months) and stabilized with optimal medical therapy (OMT). The inclusion and exclusion criteria are listed in Tables S1 and S2 of Appendix S3.

2.2.2. Groups

All patients will be randomly assigned to receive an active OA in addition to OMT for HF or sham OA with OMT, by using a random number table.

2.2.3. Sample size estimation

According to a previous study, the AHI reduction was −8.8 ± 11.9/h in patients with HF after 3 months of intervention with another OA.8 The natural change in AHI was +1.3 ± 3.2/h in patients with HF. To detect the effect size and set the α error to 0.05 and the β error to 0.20, we estimated that at least 13 patients are needed in each group. We will recruit at least 30 patients in total, accounting for two possible dropouts in each group.

2.3. Oral appliances

SomnoDent is a custom‐made two‐piece appliance to move the mandible forward into its appropriate position, to prevent the collapse of the oropharynx by the root of the tongue (Appendix S4, Picture S1). The screw mechanism on the upper splint enables a fine adjustment of mandibular advancement. A micro‐recorder chip is embedded in the lower splint so that the compliance can be monitored by the computer (DentiTrac system) (Appendix S4, Pictures S2 and S3). Sham OA includes only the lower splint with a micro‐recorder (ie, patients in the sham OA group will not wear an upper splint).

2.4. Measurements

In addition to the follow‐up PSG, the following measurements will be obtained at baseline (at the same time with randomization) and 3 months later: plasma BNP level, urinary catecholamine excretion, echocardiographic data, 6‐min walk distance, Epworth sleepiness scale (ESS), and health‐related quality of life (Appendix S5).

2.5. Intervention

First, all subjects will undergo an acclimatization phase (2‐4 weeks) when using active OA at home. The effect of OA on SDB will be confirmed using WatchPAT. The target will be one of the following: AHI obtained using WatchPAT (peripheral arterial tonometry apnea‐hypopnea index [pAHI]) <10 events/h, ≥50% reduction in pAHI, or ≥50% reduction in ODI.9 On the basis of the results of this study, OA will be adjusted by experienced dentists in Juntendo University Hospital. The WatchPAT assessment will be repeated until sufficient SDB suppression is confirmed.

Next, OA compliance will be checked using a monitoring device (DentiTrac) after the washout period (1 week). Noncompliant patients (<70% and < 4 hours/night) will be excluded from randomization and followed up for 3 months to obtain a natural change of primary or secondary endpoints. Compliant patients will be randomized into active OA or sham OA by using a random number table; the allocation will be concealed in an opaque envelope. Intervention with active OA or sham OA will be conducted for 3 months.

2.6. Follow‐up

After randomization, patients will make a hospital visit every month. This trial will be completed with the follow‐up PSG after 3 months.

2.7. Outcomes

The primary endpoint is AHI determined using PSG. The follow‐up PSG will be scheduled after 3 months of intervention with active OA or sham OA. The secondary endpoints include E/e’ determined using echocardiography (tissue Doppler image), plasma BNP level, compliance (intervention group only), New York Heart Association class, ESS score, quality of life score (Short Form 8 or Minnesota Living with Heart Failure Questionnaire), and 6‐min walk distance.

2.8. Progress and estimated study period

This trial started on April 1, 2017. The first study participant was enrolled on September 4, 2017. A total of eight patients have been enrolled at the date of writing. The projected end date of this trial is March 31, 2019.

2.9. Data collection methods

The WatchPAT will be attached at the same time as the baseline PSG to confirm the correlation of AHI obtained from these two devices. The OA will be set for each patient diagnosed as having SDB by using baseline PSG. The effect of OA will be assessed using WatchPAT at home after the acclimatization phase (2‐4 weeks). OA adjustments will be made according to the results of WatchPAT, as described earlier. Baseline data other than PSG will be collected on the same day as the randomization. Follow‐up PSG will be scheduled after 3 months of intervention. All measurements will be repeated at the same time as the follow‐up PSG.

All sleep study data, including PSG and WatchPAT results, will be analyzed at a sleep laboratory inside Juntendo University Hospital throughout the study. Trained nurses and technicians blinded to the assignments will perform blood collection, echocardiography, PSG measurements, and 6‐min walk test measurements. The ESS questionnaire will be completed by the patients.

2.10. Statistics

Data will be analyzed using SPSS version 23 (spss, IBM Corp.). Differences in the change in mean value between patients with and those without OA will be assessed using intention‐to‐treat analysis.

2.11. Monitoring

The study will be discontinued under the following conditions:

1. Withdrawal of consent by the participant;

2. Discontinuation of the present study;

3. Investigator's judgment that the study is suitable for discontinuation.

2.12. Ethics and dissemination

This study was approved by the independent ethics committee of Juntendo University Hospital. Written informed consent forms will be distributed to all participants before enrollment.

2.13. Discussion and anticipated results

The prevalence of OSA ranges from 3% to 7% in the general population.11 In contrast, it amounts to 36% in patients with HF,12 which implies that OSA is not merely a coexisting condition but an important confounder in the pathogenesis of HF. OSA has several adverse effects on the cardiovascular system, as follows: increased preload due to the exaggerated intrathoracic negative pressure during forced inspiration in the presence of upper airway obstruction, LV diastolic dysfunction caused by the shifting of the intraventricular septum due to right ventricle dilatation, and increased afterload during apnea. Continuous positive airway pressure (CPAP) can abolish OSA in patients with HF as well as improve LV systolic/diastolic function,13 and its long‐term efficacy on prognosis has been established. However, there are >30% dropouts or nonresponders,14 whose long‐term outcome is poor.15 The main reason has been attributed to the remaining CSA as a consequence of pulmonary congestion.

On the other hand, OSA increases in response to worsening of HF because the upper airway is easily collapsed by mucosal edema after systemic congestion. However, upper airway collapsibility may also play a role in the pathogenesis of CSA.16, 17 In other words, upper airway stabilization by OA may serve to alleviate CSA. A greater reduction of AHI will be expected in patients with predominant OSA, which is generally defined as the percentage of central AHI per total AHI ≤50%. However, we set ≤70% in this study and no upper limit of AHI to ensure the generalizability of the impact of OA in patients with HF.

Although a recent study showed that CPAP suppressed AHI, daytime blood pressure, and cardiac autonomic function more sufficiently than OA,18 this might have resulted from cardiac unloading by positive end‐expiratory pressure. It remains to be determined to what degree simple upper airway stabilization can alleviate either OSA or CSA in patients with HF. To our knowledge, only one single‐arm interventional study has been conducted to examine the efficacy of OA in patients with HF thus far.8 This is the first randomized controlled study (RCT) to systematically address the efficacy of OA for SDB, cardiac function, quality of life, and exercise capacity in patients with HF.

In this study, we will enroll patients with systolic and diastolic HF. As for systolic HF, we set a wider range of LVEF including HF with mid‐range ejection fraction to enroll more subjects to keep feasibility (LVEF 40%‐49%).19 On the other hand, changes in cardiac parameters are difficult to detect in patients with diastolic HF. Thus, we added two definitions regarding previous hospital admission and plasma BNP levels (>100 pg/mL).20, 21 Plasma BNP level will not be affected by HF medication because valsartan /sacubitril have not yet been approved in Japan.

One significant limitation is that only compliant patients will be included in the OA intervention group, which can be considered a selection bias. However, this is a pilot study aimed at investigating whether OA can effectively suppress SDB in patients with HF and its effect on cardiac function. The selection of study subjects should be focused on compliant patients first, and the effect of compliance on cardiac function and symptoms should be investigated in a future larger‐scale study. Another limitation is the small sample size. The sample size of 15 in each group is adequately powered for the primary endpoint: AHI determined using PSG. However, this is not true for the secondary endpoints. Further large‐scale RCTs are warranted to evaluate the effects of SDB treatment on cardiac function using OA.

Supporting information

File S1.

Matsumoto H, Kasai T, Suda S, et al. Randomized controlled trial of an oral appliance (SomnoDent) for sleep‐disordered breathing and cardiac function in patients with heart failure. Clin Cardiol. 2018;41:1009–1012. 10.1002/clc.23028