Article Text

Download PDFPDF

Mandibular advancement device versus CPAP in lowering 24-hour blood pressure in patients with obstructive sleep apnoea and hypertension: the CRESCENT trial protocol
  1. Yi-Hui Ou1,
  2. Juliana Tereza Colpani2,
  3. Siew-Pang Chan1,3,
  4. Weiqiang Loke2,
  5. Crystal S Cheong4,
  6. William Kong5,
  7. Calvin W Chin6,
  8. Pipin Kojodjojo7,
  9. Philip Wong8,
  10. Peter Cistulli9,
  11. Chi-Hang Lee1,3
  1. 1Department of Medicine, National University of Singapore, Singapore
  2. 2Department of Endodontics, Operative Dentistry and Prosthodontics, National University of Singapore, Singapore
  3. 3National University Heart Centre, Singapore
  4. 4Department of Otolaryngology-Head and Neck Surgery, National University Hospital, Singapore
  5. 5National University of Singapore, Singapore
  6. 6Department of Cardiology, National Heart Centre Singapore, Singapore
  7. 7Cardiology, National University Heart Centre, Singapore
  8. 8Department of Medicine, Raffles Hospital, Singapore
  9. 9Charles Perkins Centre, The University of Sydney, Sydney, New South Wales, Australia
  1. Correspondence to Dr Chi-Hang Lee; mdclchr{at}


Introduction Although treatment of obstructive sleep apnoea (OSA) using continuous positive airway pressure (CPAP) reduces blood pressure (BP), adherence to CPAP is often suboptimal. A mandibular advancement device (MAD) is a guideline-endorsed alternative therapy for OSA. Still, there is limited evidence on the relative efficacy between MAD and CPAP on BP reduction. We evaluate whether treatment of moderate-to-severe OSA using MAD can improve BP and other health-related outcomes compared with CPAP.

Methods and analysis This is a randomised, controlled, non-inferiority trial conducted. We will recruit 220 Asians with a history of hypertension and high cardiovascular risk for an overnight polysomnography screening. Those with moderate-to-severe OSA (apnoea–hypopnoea index ≥15 events/hour) will be randomised to treatment with either MAD or CPAP in a 1:1 ratio. Stratified by age (60 vs <60 years old), body mass index (25 vs <25 kg/m2) and apnoea–hypopnoea index (30 vs <30 events/hour), an adaptive randomisation scheme with permuted blocks constructed in real-time is implemented to restrict imbalance. The overall study duration is 12 months. The primary endpoint is the 24-hour mean arterial BP difference between baseline and 6-month follow-up. The secondary endpoints include other measures of ambulatory BP monitoring, arrhythmia based on a 4-day electrocardiographic monitoring, biomarker and proteomic analysis, cardiovascular magnetic resonance-derived myocardial fibrosis and remodelling and quality-of-life questionnaires. Recruitment began in October 2019 and ended in December 2022. Comparison between MAD and CPAP will be performed using covariance (ANCOVA) analysis of the changes in 24-hour mean arterial BP while adjusting for the baseline 24-hour mean arterial BP. We will compare the 95% CIs around the treatment difference point estimate with the prespecified non-inferiority margin (1.5 mm Hg). If the upper limit of the 95% CI is <1.5 mm Hg and crosses 0, non-inferiority of the MAD relative to CPAP will be established.

Ethics and dissemination The Domain Specific Review Board-C, National Healthcare Group under approved the study protocol (NHG DSRB Ref: 2019/00359, approved on 28 August 2019). Study findings will be disseminated to various local, national, and international audiences through abstract presentations and publication in peer-reviewed journals.

Trial registration number NCT04119999.

  • sleep medicine
  • blood pressure
  • clinical trial
  • hypertension
  • adult cardiology
  • Singapore

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Strengths and limitations of this study

  • The primary endpoint is based on ambulatory blood pressure monitoring rather than office blood pressure.

  • Participants with a severe form of obstructive sleep apnoea and known hypertension will be included.

  • The trial will only recruit Asians, who are known to have the restrictive craniofacial phenotype and may respond better to mandibular advancement devices.


Obstructive sleep apnoea (OSA) is a chronic sleep disorder characterised by repeated episodes of a complete or partial collapse of the upper airway during sleep. The resultant respiratory disturbance results in sleep fragmentation, negative intrathoracic pressure swings, hypoxaemia and sympathetic activation. One billion adults are estimated to suffer from OSA, with half having a moderate-to-severe disease.1 OSA is frequently associated with hypertension, with approximately half of the patients with OSA having coexisting hypertension.2 According to the 2017 American College of Cardiology/American Heart Association Hypertension Guidelines, OSA is an underlying cause of 25%–50% of secondary hypertension. While OSA is ubiquitous in obese individuals, the restrictive craniofacial phenotype among the Asian population makes them more susceptible to OSA.3 Unsurprisingly, many Asian patients with OSA have lower body mass indices than their Caucasian counterparts.

Treatment of OSA using continuous positive airway pressure (CPAP) modestly reduces blood pressure (BP). Yet, CPAP is often poorly accepted and tolerated by patients with OSA.4 A mandibular advancement device (MAD) is a guideline-endorsed alternative therapy for OSA in patients who do not accept or tolerate CPAP.5 There is limited evidence from short-term studies on the relative efficacy between MAD and CPAP on BP reduction. Randomised trials have compared MAD versus CPAP and concluded that both techniques have similar BP-lowering effects.6–8 However, these trials are limited by small sample sizes,5–7 the inclusion of patients without hypertension,5–7 the exclusion of patients with severe OSA5 6 and a short intervention period of 1 month–3 months.5–7

By repositioning the lower jaw into a more anterior position and thereby improving upper airway patency during sleep, we postulate that the MAD may be particularly effective in Asian patients in whom OSA is primarily a result of a restrictive craniofacial phenotype.9–12 In the CRESCENT (Cardiosleep Research Program on Obstructive Sleep Apnoea, Blood Pressure Control and Maladaptive Myocardial Remodelling—Non-inferiority Trial, NCT04119999), we aim to evaluate whether treatment of moderate-to-severe OSA using MAD can improve BP and other health-related endpoints when compared with CPAP. Given the high prevalence of OSA in Asians and the benefits of intensive BP lowering, including data on Chinese individuals,13 the CRESCENT trial will inform recommendations for healthcare policymakers in considering appropriate strategies for managing OSA in Asian societies.


Design and study population

The CRESCENT is a randomised, controlled, non-inferiority trial conducted in Singapore—an Asian country with a population of 5.6 million (Chinese 74%). The CRESCENT is a collaborative effort between two healthcare clusters in Singapore1: National University Health System (comprises National University Heart Centre Singapore (central coordinating institution), National University Centre for Oral Health, National University Hospital) and2 SingHealth (National Heart Centre Singapore). The CRESCENT was designed in 2018 to compare the BP-lowering effect and other health-related outcomes during the treatment of moderate-to-severe OSA with MAD versus CPAP. For the sake of simplicity, the term 'OSA' would be used to denote 'moderate-to-severe OSA' for the rest of this protocol.

The CRESCENT was funded by the National Medical Research Council of Singapore in March 2019 after anonymised external (international) and internal (Singapore) peer reviews. We recruited Chinese participants with a known history of physician-diagnosed hypertension and high cardiovascular risk for an overnight polysomnography screening. The detailed inclusion and exclusion criteria are shown in box 1. The participants with OSA will be treated with either MAD or CPAP under the relevant specialists according to this study protocol for 12 months. A team of clinical research coordinators provides coordination and administrative support.

Box 1

Inclusion and exclusion criteria

Inclusion criteria

  • Age of at least 40 years.

  • Chinese (based on the Identity Card or other Identity Document if the subject is a non-Singapore citizen or permanent resident).

  • Physician-diagnosed essential hypertension, on at least one medication for BP control.

  • High cardiovascular risk, as defined by one or more of the following: (a) diabetes mellitus, (b) stroke or transient ischaemic attack, (c) significant coronary artery disease (at least one stenosis of >50% in at least one major epicardial artery on invasive or CT coronary angiography, previous coronary revascularisation, previous myocardial infarction or abnormal stress test), (d) chronic kidney disease with an estimated glomerular filtration rate of less than 60 mL/min mL/min/1.73 m2, or (e) age of 75 years or older.

Exclusion criteria

  • Known obstructive sleep apnoea on treatment.

  • Cheyne-Stokes breathing or predominantly central sleep apnoea (>50% central events).

  • Known secondary hypertension: from renal (renal artery stenosis, chronic renal failure); endocrine (aldosterone excess, pheochromocytoma, Cushing’s syndrome, hyperthyroidism) or cardiac causes (aortic coarctation).

  • Contraindications to mandibular advancement device: less than six teeth in each arch; inability to advance the mandible and open the jaw widely, pre-existing temporomandibular joint problems, severe bruxism.

  • Limited life expectancy (<1 year).

  • Hypertensive crisis, acute coronary syndromes, or acute heart failure in the past 30 days.

The planned follow-up duration is 12 months. The primary endpoint is the 24-hour mean arterial BP difference between baseline and 6-month follow-up. The secondary endpoints include 24-hour, daytime, and night-time systolic and diastolic BP, pulse pressure, nocturnal BP dipping (more than 10% decrease in BP at night), percentage of participants with 24-hour systolic BP below 130 mm Hg and below 120 mm Hg at follow-ups, arrhythmia based on a 4-day electrocardiographic monitoring, biomarker and proteomic analysis, cardiovascular magnetic resonance (CMR)-derived myocardial fibrosis and remodelling and quality-of-life questionnaires. The Domain Specific Review Board-C, National Healthcare Group under approved the study protocol (NHG DSRB Ref: 2019/00359, approved on 28th August 2019). Recruitment began on 1 October 2019 and ended on 5 December 2022, giving a total recruitment period of 38 months. The recruitment progress is shown in figure 1.

Figure 1

Recruitment progress.

Patient and public involvement

There is no patient or public involvement in the design, recruitment, conduct, analysis or interpretation in the CRESCENT trial.


The overall hypothesis of the CRESCENT is that treatment of newly diagnosed moderate-to-severe OSA using MAD is non-inferior to CPAP in reducing 24-hour mean arterial BP and other health-related endpoints in Chinese (Asian) participants with known physician-diagnosed hypertension and high cardiovascular risk.

Sample size calculation

To detect the non-inferiority of MAD compared with CPAP with a desired statistical power of 90% and an anticipated attrition of 20%, a sample size of 220 participants with OSA is required with the assumption that there is no actual difference between the treatment means. Performed during the design stage in 2018, the calculation was based on a change in 24-hour mean arterial BP (ie, (systolic BP + 2 × diastolic BP)/3) of CPAP when used as an OSA treatment. The non-inferiority margin was determined with the two-step fixed-margin approach, in which previous studies comparing the active control with placebo would be used to derive a single fixed value for the margin.14 The value recommended in recent guidance from the Food and Drug Administration15 is the upper bound of the 95% CI around the treatment effect of a single placebo-controlled trial. Based on a reported result that CPAP could lower the 24-hour mean arterial BP by 3.3 mm Hg (95% CI: −5.3 to −1.3 mm Hg) with respect to sham CPAP,16 the upper bound of the CI (ie, 1.3 mm Hg) was rounded up, and the non-inferiority margin was set at 1.5 mm Hg.


The Consolidated Standards of Reporting Trials diagram is shown in figure 2. Eligible participants with written informed consent will undergo polysomnography within the next 2 weeks. Overnight in-laboratory polysomnography will be performed for all participants using an American Academy of Sleep Medicine type I sleep diagnostic software (Embla RemLogic, Natus Medical, Canada). The same technologist with Registered Polysomnographic Technologist credentials prospectively scores each polysomnogram according to the American Academy of Sleep Medicine 2012 guidelines.17 The primary measure is the apnoea–hypopnoea index—the total number of apnoeas and hypopnoeas per hour of sleep. An apnoea is defined as a decrease of 90% or more in airflow from the baseline value for at least 10 s. Apnoeas are further classified as obstructive or central based on the presence or absence of respiratory effort-related chest wall movement. Hypopnoea is defined as a 30%–90% reduction in airflow from the baseline value lasting 10 s or more in conjunction with oxygen desaturation of at least 3%. OSA is defined as apnoea–hypopnoea index ≥15 events per hour. Measures of hypoxic burden will be performed. Those without OSA will be considered to have completed their participation.

Figure 2

Consolidated Standards of Reporting Trials diagram. CPAP, continuous positive airway pressure; MAD, mandibular advancement device; OSA, obstructive sleep apnoea; PSG, polysomnography


After the polysomnography screening, subjects with OSA will undergo randomisation. They are assigned to either MAD or CPAP based on a 1:1 allocation ratio. Stratified by age (≥60 vs <60 years old), body mass index (≥25 vs <25 kg/m2) and apnoea–hypopnoea index (≥30 vs <30 events/hour), an adaptive randomisation scheme with permuted blocks constructed in real-time is implemented to restrict imbalance throughout the trial. A web-based system is set up to ensure the allocation is adequately blinded. The unblinded statistician was tasked to oversee the randomisation and the creation and management of a computer-based system to facilitate checking and dissemination on a daily basis. After the random allocation was generated, an encrypted Microsoft Excel file was produced and sent to the study coordinator on request. The research coordinator will inform all the participants about the results of the polysomnography and the randomisation outcome by telephone call within 2 weeks after the polysomnography.

Baseline study visit and acclimatisation

All the randomised participants will have in-person, scheduled study baseline visits at the clinic with a clinical research coordinator. A subject’s medical history will be recorded at baseline and include a prespecified list of conditions (high BP, heart disease, diabetes mellitus and stroke), the number of years with each condition and current treatment (if any) the subject is taking for each condition. Medical history of heart disease or stroke in the subject’s family (parents or siblings) will be recorded at baseline. Both groups will undergo baseline evaluation of the endpoints: ambulatory BP monitoring, office BP, prolonged 4-day ECG monitoring, blood test for cardiovascular biomarkers and proteomics and cardiac magnetic resonance.

After the baseline visit, there is a 4-week acclimatisation period for the MAD and CPAP groups to optimise device setting and fitting. After the acclimatisation period, the treatment phase will last for 12 months. During this period, the patients will continue their usual healthcare appointment. Any changes in BP medication (type and dosage) will be captured during the study visits and verified using the electronic medical record whenever feasible.

Treatment groups

Treatment of the OSA, according to the results of randomisation, will begin within 2 months after the polysomnography and the baseline study visit. Crossover of the treatment group is not allowed.

MAD group (n=110)

The MAD is a custom-made, removable, two-piece, adjustable device (SomoDent Fusion, SomnoMed). The maximal protrusion/advancement tolerated by the patient will be recorded. Individual 3D dental intraoral impressions (3Shape TRIOS-4) for the manufacture of the device and bite recordings at a mandibular protrusive bite position of 70%–75% of maximum protrusion will be obtained.18–21 Once fitted with the device, participants will undergo acclimatisation, during which the device will be incrementally advanced until the maximum protrusion or maximum comfortable limit of advancement is achieved. The participants will be asked to use the device each night for as long as they can tolerate it. Both face-to-face and telephone dental support will be available during the adaptation. The advancements will be recorded in a diary, and the final position will be verified by the dentist, after which the device will be worn until the MAD treatment phase of the study ends. A follow-up sleep study will be performed to document the residual apnoea–hypopnoea index at the end of acclimatisation.

CPAP group (n=110)

A commercially available automatic CPAP (AirSense 10, Resmed, Australia) will be used in all the participants. Before acclimatisation to the CPAP, the participants will be fitted with an appropriate mask interface. All the participants will receive CPAP education, including instruction on CPAP machine operation. The machines will be initially set in auto-CPAP mode for home use. The participants will then be asked to use the device each night for as long as they can tolerate it. Face-to-face and telephone support from a CPAP therapist will be available during this acclimatisation period, which will last for 1 month. At the end of the 6-month study period, adherence to CPAP will be measured objectively by downloading a card that contains the device’s time counter.22

Treatment adherence and treatment-related side effects

Adherence to the assigned treatment will be recorded at 1, 6 and 12 months. Data regarding CPAP use will be downloaded from a cloud-based telemedicine management platform (AirView, ResMed Corp, San Diego, California, USA), and adherence to the use of the MAD will be determined by an embedded compliance micro-recorder chip (DentiTrac, Braebon, Canada). Non-adherence to the assigned treatment would not exclude the participants from the analysis unless they chose to withdraw from the study. Side effects of treatment will be evaluated by self-reporting using Sleep apnoea Quality of Life Index (SAQLI) questionnaire in a clinical setting. Participants will report the treatment-related side effects at 6 and 12 months.

Ambulatory and office BP monitoring

Ambulatory BP monitoring

Ambulatory BP monitoring will be performed over 24 hours using a clinically approved oscillometric device (Welch Allyn ABPM 7100, Welch Allyn, Skaneateles Falls, New York, USA) programmed to record BP every 30 min. The BP cuff will be applied to the non-dominant arm. The participants will complete a sleep log and press an event marker to identify sleep and wake periods. Data will be included for analysis if the monitor records at least 21 readings over 24 hours.23 The ambulatory BP monitoring will be performed at baseline, 6-month, and 12-month follow-ups. Data from the baseline and 6-month follow-up will be used to determine the primary endpoint.

Office BP monitoring

Office BP is obtained using a digital sphygmomanometer and appropriate size cuff after the participants have been seated for 5 min in a quiet room. Three readings are consecutively performed at 2-min intervals, and the mean value is calculated. BP is measured from both arms, and the arm with the higher mean BP will be used. Office BP will be performed at baseline, 6-month, and 12-month follow-ups. All BP assessors are blinded to baseline characteristics and treatment allocation.

Prolonged ECG monitoring

Prolonged ECG monitoring will be performed using Spyder ECG, a CE marked, medical-grade, wearable single-lead ECG monitor utilising a smartphone to transmit ECG data to a cloud database wirelessly. It is approved by the Singapore Health Science Authority for remote and ambulatory ECG diagnostic purposes. In the CRESCENT, the Spyder ECG will be attached to the participants by trained personnel for 4 days to assess for arrhythmia events. Arrhythmia events are defined by the detection of any of the following findings: ventricular tachycardia, ventricular fibrillation, supraventricular tachycardia, sinus arrest (duration >3 s), supraventricular tachycardia (>4 beats, not including atrial fibrillation or flutter), atrial fibrillation (>30 ms), atrial flutter, atrioventricular block (second-degree 2:1, or third-degree atrioventricular block requiring advanced evaluation by the investigators) and ventricular tachycardia (>4 beats). Prolonged ECG monitoring using Spyder ECG is performed at baseline and 12-month follow-up.

Cardiovascular magnetic resonance

CMR (Siemens AERA 1.5T; Siemens AG, Healthcare Sector, Erlangen, Germany) will be performed on all participants who consented to the imaging substudy. Balanced steady-state free precision cine images of the vertical and horizontal long-axis planes and the sagittal left ventricular outflow tract view will be acquired (acquired voxel size 1.6×1.3×8.0 mm; 30 phases per cardiac cycle). Short-axis cines are obtained from the mitral valve annulus to the apex (acquired voxel size 1.6×1.3×8.0 mm; 30 phases per cardiac cycle). Myocardial fibrosis will be assessed using two approaches: late gadolinium-enhanced (LGE) imaging (focal replacement fibrosis) and myocardial T1 mapping (diffuse myocardial fibrosis). LGE imaging will be started 8 min after administering 0.1 mmol/kg of gadobutrol (Gadovist; Bayer Pharma AG, Germany). An inversion-recovery fast gradient echo sequence will be used. Myocardial T1 mapping is performed using the Modified Look-Locker inversion-recovery sequence (flip angle 35°; minimum TI 100 ms; TI increment of 80 ms). Native and postcontrast myocardial T1 maps (15 min after contrast administration) will be acquired using a heartbeat acquisition scheme of 5 (3)3 and 4 (1)3 (1)2, respectively. Deidentified CMR images will be analysed at the National Heart Research Institute Singapore CMR Core Laboratory using standardised protocols for measuring cardiac volumes, function and left ventricular mass using dedicated software (CV42; Circle cardiovascular Imaging, Calgary, Canada).24 25 According to the Society of CMR’s recommendations, LGE will be assessed qualitatively by two experienced readers.26 Extracellular volume fraction will be assessed as a mean of the basal and midventricular slices (including regions of non-ischaemic LGE) using the T1 mapping module in CVI42 (Circle Cardiovascular Imaging). Interstitial volume is defined as extracellular volume × myocardial volume, where myocardial volume (mL) is defined as myocardial mass (g)/1.05 g/mL. We will exclude regions of focal ischaemic LGE in the assessment of extracellular volume and interstitial volumes.

Cardiovascular biomarkers and proteomic analysis

Blood samples for measurement of high-sensitivity C reactive protein, high-sensitivity troponin T and N-terminal-pro hormone brain natriuretic peptide are drawn at the outpatient clinic. All samples are placed on ice and taken to the Cardiovascular Research Institute for analysis. Additional blood samples are collected for future proteomic assays and novel biomarker discovery. Blood samples are collected at baseline, 6 months and 12 months.

Proteomic analyses- whole blood samples collected in EDTA anticoagulant tubes were centrifuged at 1500 × g for 10 min at 4°C to separate the plasma fraction. All specimens are immediately aliquoted, frozen and stored in a dedicated −80 °C freezer. Protein biomarkers will be analysed in baseline plasma samples using the Olink Target Cardiovascular II, Cardiovascular III and Inflammation panels (Olink Proteomics, Uppsala, Sweden) which comprises 92 proteins each and uses the methodology based on the proximity extension assay.

Quality-of-life questionnaires

The following questionnaires will be administered through face-to-face interviews by an investigator before polysomnography at 6-month and 12-month follow-up1: Epworth Sleepiness Scale (ESS),2 Euroqol questionnaire (EQ5D),3 Functional Outcomes of Sleep Questionnaire (FOSQ),4 SAQLI and5 36-item Short Form Health Survey (SF-36). We will determine the changes in the quality-of-life based on the scores between the MAD and CPAP groups.

ESS is the most widely used subjective index to quantify sleepiness. The participants will score themselves, on a scale of 0–3, on how easily they would fall asleep in eight different situations, giving an overall score between 0 and 24; the higher the score the sleepier the individual. An ESS score of more than 10 is used to define excessive daytime sleepiness.

EQ5D is a standardised measure of health-related quality of life developed by the EuroQol Group to provide a simple, generic questionnaire for clinical population health surveys. EQ-5D assesses health status in terms of five dimensions of health and is considered a generic questionnaire. EQ5D allows participants to rate their health on five dimensions: (1) mobility, (2) self-care, (3) usual activities, (4) pain or discomfort and (5) anxiety or depression.27

FOSQ is a sleep-specific quality-of-life questionnaire. There are 30 questions grouped into five domains: activity, vigilance, sexual relationships, productivity and social. An average score across the domains and an overall score will be calculated, with high scores denoting better-perceived health.28

SAQLI is a disease-specific health-related quality-of-life questionnaire designed for adults with OSA that measures the outcome of treatment for OSA.29–31 The SAQLI includes four functional domains that measure daily functioning, social interactions, emotional functioning and symptoms from OSA. It completes with a fifth domain assessing the negative impacts of therapeutic interventions. The degree of impairment/difficulty is rated on a 7-point Likert scale,1–7 ranging from not at all1 to enormous.7

SF-36 is a measure of general quality-of-life containing 36 self-completed questions grouped into 1 of 8 domains: physical function, bodily pain, role limitations resulting from physical health problems, role limitations resulting from personal or emotional problems, emotional well-being, social functioning, energy/fatigue and general health perceptions. Higher scores denote greater well-being.32

Statistical analysis plan

The data will be presented with a mean±SD, median (IQR) and frequency (%). Exploratory analyses concerning the unadjusted comparison of MAD and CPAP groups will be performed with the Student’s t-test, Mann-Whitney U test, χ2 test and Fisher’s exact test.

The following confirmatory analysis concerning the comparison between MAD and CPAP is facilitated with the analysis of covariance (ANCOVA) of changes in 24-hour mean arterial BP while adjusting for the baseline mean arterial BP. We will compare the upper limit of the CI around the treatment difference point estimate with the prespecified difference (1.5 mm Hg). If the 95% CIs fall above the line of no difference, MAD is non-inferior and superior to CPAP. If the 95% CIs do not cross the non-inferiority margin, the MAD is non-inferior to CPAP. If the 95% CIs are below non-inferiority margin, MAD is inferior to CPAP. Potential covariates testing and adjustment for baseline covariates in the primary analysis will not be conducted even when there is a baseline imbalance in variables observed post hoc. However, sensitivity analyses, including such variables when there are large baseline imbalances, will be conducted to assess the robustness of the primary analysis. We will carefully ensure that the default ANCOVA model, estimated with ordinary least squares, provides a satisfactory fit to the data and offers a valid and meaningful interpretation. An alternative robust regression model estimated with the iteratively reweighted least squares is applied if outliers, influential observations and leveraged data are identified during analysis. Generalised linear models based on different underlying distributions (normal, binomial, gamma or inverse Gaussian) and link functions will be considered for other endpoints.

We will conduct prespecified subgroup analyses on the primary endpoint according to the following baseline subgroups: age (>60 vs ≤60 years old), gender (male vs female), body mass index (>25 vs ≤25 kg/m2), waist circumference (split by tertiles), apnoea–hypopnoea index (>30 vs ≤30 events per hour), oxygen desaturation index (>30 vs ≤30 events per hour), Epworth Sleepiness Scale (>10 vs ≤10), diabetes mellitus (yes vs no), significant coronary artery disease (yes vs no), number of BP medication (>2 vs ≤2) and device adherence (split by tertiles).

All statistical analyses will be performed according to the intention-to-treat principle, with the level of significance fixed at 5%. The 95% CIs are also routine reported if appropriate. The data will be exported to Stata MP V16 (Stata Corporation, Texas, USA) for further data management and analysis.

Follow-up schedules

All randomised participants will attend the MAD or CPAP clinics, respectively at 1, 2, 6 and 12 months. There are two optional visits between the second and sixth months for participants who need additional advice on the treatment devices or adjustments. During the study visits, the participants will be assessed for treatment-related side effects, changes in BP medications or hospitalisation since randomisation. The details of the follow-up schedule can be found in table 1.

Table 1

Follow-up schedule

Preliminary results

As of 30 November 2022, a total of 3479 have been assessed for eligibility and 321 participants were eligible and consented to join the trial. There are 15 participants who withdrew their consents before polysomnography, and 306 participants have validated polysomnography results. Complete baseline demographic and clinical characteristics are available for the 306 participants with validated polysomnography results as shown in table 2. OSA and non-OSA (including one central sleep apnoea) were diagnosed in 220 and 86 participants, respectively. The prevalence of OSA was 71.9%.

Table 2

Baseline demographics of the 306 patients with valid polysomnography results

In a multivariate logistic regression analysis, age, sex, body mass index, neck circumference and waist–hip ratio were analysed, only age (p=0.008) and body mass index (p=0.003) were shown to independently predict OSA.


To the best of our knowledge, the CRESCENT is the largest randomised controlled trial to investigate the efficacy of MAD versus CPAP in reducing BP in participants with moderate-to-severe OSA and hypertension. The primary endpoint of the CRESCENT, change in 24-hour mean arterial BP, will be released in the last quarter of 2023; and the complete results will be available in mid-2024. The data generated by the CRESCENT trial will be made available as soon as possible on reasonable request, wherever legally and ethically possible.

While obesity is a well-known risk factor for OSA, the importance of craniofacial features in OSA has been increasingly recognised. The relative role of obesity and craniofacial dimensions vary by ethnic group. Our group conducted a population screening in Singapore and showed that the Chinese had a higher prevalence of OSA despite having a lower body mass index (23.3 kg/m2) than Indians (25.4 kg/m2) and Malays (26 kg/m2).33 A study comparing Chinese with Caucasians found that Caucasians had a higher body mass index (30.7 vs 28.4 kg/m2) and larger neck circumference (40.8 vs 39.1 cm). However, the Chinese had more craniofacial bony restrictions.11 This predisposition suggests that MAD may have a unique role in treating OSA in the Chinese population.

There are some limitations in the CRESCENT. The CRESCENT is conducted at a tertiary healthcare institution in Singapore. While all the specialists, including cardiologists, sleep physicians, otolaryngologists and dentists, are well-trained and experienced, the study’s findings may not be extrapolated to hospitals where the relevant expertise is unavailable. The trial will recruit Asians, who are known to have the restrictive craniofacial phenotype and may respond better to MAD. Hence, the findings may not be generalisable to other ethnicities.

Ethics and dissemination

The Domain Specific Review Board-C, National Healthcare Group under approved the study protocol (NHG DSRB Ref: 2019/00359, approved on 28 August 2019). Study findings will be disseminated to various local, national, and international audiences through abstract presentations and publication in peer-reviewed journals.

Ethics statements

Patient consent for publication



  • Contributors All the authors have made substantial contributions to the conception or design and drafting of the work. They have all approved the version to be published. Authors' specific contributions include: Y-HO and C-HL wrote the first draft. CWC wrote the section on cardiac magnetic resonance imaging. S-PC (a biostatistician) provided extensive inputs on sample size calculation, randomisation and statistical analysis plan. JTC and WL are leading the MAD arm. CSC is leading the CPAP arm. WK, CWC, PK and C-HL are the steering committee members. PC have provided inputs on the study design and execution through teleconference every 3 months. The National Medical Research Council Clinician Scientist Award was awarded to S-PC, WK, CWC, PK, PW and C-HL (the main applicant). The USyd-NUS Partnership Collaboration Award was awarded to PC (co-lead) and C-HL (co-lead).

  • Funding This study was supported by a Clinician Scientist Award from the National Medical Research Council of Singapore (Grant number: CSASI18may-0001) and the USyd-NUS Partnership Collaboration Award, a joint award from the National University of Singapore and the University of Sydney (Grant number: N/A).

  • Competing interests PW is a founder of WEB bio which develops the Spyder device. PC has an appointment as an endowed Academic Chair at the University of Sydney that was created from ResMed funding; he receives no personal fees, and this relationship is managed by an Oversight Committee of the University. Additionally, he has received research support from ResMed, SomnoMed, Zephyr Sleep Technologies, and Bayer, and is a consultant/adviser to Signifier Medical Technologies, SomnoMed, ResMed, Bayer and Sunrise Medical.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.