To compare different non-surgical treatment modalities on treatment of obstructive sleep apnea: A systematic review and meta-analysis

Abstract

The study aimed to assess the effect of mandibular advancement device (MAD) in patients with obstructive sleep apnea for reduction in 24-h mean blood pressure, sleep quality, Apnea Hypopnea Index (AHI), and patient compliance, compared to continuous positive airway pressure (CPAP), other interventions, or no treatment. Three different databases such as PubMed, EMBASE, and CENTRAL were searched using different search terms till July 2021 as per the inclusion and exclusion criteria. After inclusion of studies, data extraction including risk of bias assessment was done. For each study, we used odds ratio, mean difference, and 95% confidence interval to assess and synthesize the outcomes. The quality of evidence was evaluated as per the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE). Twenty-one randomized controlled trials were included: 497 patients in the MAD group, 239 patients in the CPAP group, and 274 patients in the sham group. In MAD-CPAP comparison, the results favored CPAP in the reduction of AHI of 3.48 (1.76-5.19). However, unclear results were found for sleep quality measured as Epworth Sleepiness Scale (ESS), patient compliance, and 24-h mean blood pressure. In MAD-sham comparison, the results favored MAD in the reduction of AHI of − 8.39 (−10.90–−5.88] and ESS of − 0.91 (−1.70–−0.12) and favored sham in terms of patient compliance while, unclear results for 24-h mean blood pressure. The GRADE score indicated that the quality of evidence is very low, low, and moderate for different outcomes. CPAP in comparison to MAD and MAD in comparison to sham showed a significant AHI reduction. However, patient compliance and 24-h mean blood pressure were not significantly different in MAD-CPAP or MAD-sham. Quality of evidence is very low and low when MAD was compared with CPAP and sham, respectively, for AHI.

INTRODUCTION

Obstructive sleep apnea (OSA) is sleep-related breathing disorder. OSA if left untreated may lead to poor quality of life (QoL), increased chances of road traffic accidents, cardiovascular attack, endocrine, metabolic, urogenital, neurological other systemic disorders like hypertension.[1,2,3] The prevalence of OSA is 2%–4% in adult population.[4] Gold standard test to diagnose OSA is polysomnography through Apnea Hypopnea Index (AHI), which also indicates severity of OSA. Symptoms of OSA can be daytime and nighttime. Daytime symptoms include daytime sleepiness, morning headache, difficulty in concentration during daytime, awakening with dry mouth or sore throat, experiencing mood changes, and memory loss. Nighttime symptoms include apnea, gastroesophageal reflux disease, loud snoring during sleep, road traffic accidents, decreased libido, sexual dysfunction, and bruxism.[5,6,7] Treatment of OSA can be surgical or nonsurgical. Surgical management includes pharyngoplasty, uvulopalatopharyngoplasty, nasal surgery, tonsillectomy in adults, genioglossus advancement, maxillomandibular advancement, and adenoidectomy, but these are not acceptable to many patients[8] due to its invasiveness. Nonsurgical management includes continuous positive airway pressure (CPAP), oral appliances (OAs) like mandibular advancement device (MAD) and act as conservative treatment options.[9,10]

CPAP is used as therapeutic as well as diagnostic for OSA patient.[11,13,14] CPAP works to keep the airway open and therefore prevents airway collapse, improves quality of sleep, reduces mortality rate, reduces high blood pressure, and reduces sympathetic tone during daytime and nighttime.[15,16,17,18,19,20]

OA improves upper airway configuration and prevents airway collapse through alteration in positions of jaw and tongue.[21,22,23,24] OA for OSA patients can either be tongue retaining devices/MADs.[25] The mechanism of action is to protrude the lower jaw more anteriorly and pulls the genioglossus forward, which helps in forward movement of the tongue. This forward movement of the tongue creates more upper airway space, which reduces chances of snoring and improves symptoms of OSA.[25] These OAs are active and protrusion of the mandible can be titrated to various degree according to the need. These are named as MAD or mandibular-repositioning appliance.[26,27]

MADs are popular choice to patients as these are affordable, light weight, and easy to use than CPAP. CPAP is a complete assembly having mask which is attached to the patient's face.[28] This may not be easily acceptable by the patient. Therefore, MADs are recommended in mild-to-moderate obstructive sleep apnea cases.[28,29,30] This device also stated to eliminate compliance issues with CPAP and therefore may be a treatment of choice for CPAP-intolerant patients.[28]

Sham is nonactive MAD which can be given in upper or lower arch as placebo. Sham has similar design as active MAD or just in the form of a plate with no components attached to it, but unlike active MAD, these do not protrude mandible. Few studies showed no significant change in blood pressure and sleep quality between MAD and sham appliance.[1] Hence, to find the true treatment effect of these physical therapies and to avoid the effect due to possibility of regression to the mean or placebo effect, a comparison with sham is important.

Efficacy of these treatments in OSA was assessed by measuring AHI, sleep quality, blood pressure, snoring events, nocturnal oxygenation, QoL, neurocognitive behavior, and patient compliance. Hence, the aim of the study was to assess the effect of MAD in comparison with CPAP and sham or no treatment for reduction in 24-h blood pressure, sleep quality, AHI, and patient compliance.

METHODS

This meta-analysis was registered in PROSPERO[31] (CRD42020131068) and followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) system.[32] The following databases were searched: Cochrane Central Register of Controlled Trials (CENTRAL); MEDLINE (from 1946 onwards); and EMBASE. Only English-language articles were included without any restriction on date of publication to compare MAD with CPAP, sham, or no appliance in patients with OSA. Details of PICO of meta-analysis were

1. Population (P) was OSA.

2. Intervention (I) were MAD or mandibular-repositioning device, mandibular protrusion.

3. Comparison (C) were CPAP, sham, occlusal splints, or no appliance.

4. Outcomes (O) were reduction in 24-h mean blood pressure, sleep quality (Epworth Sleepiness Scale [ESS]), AHI and patient compliance.

Various search terms including MeSH and Emtree were used as per attached Supplementary Table 1 for different databases. These terms were then combined with different Boolean operator like “AND” or “OR” or “NOT.” The authors (JV, PD, and BPS) have done search in these databases. Manual search of reference list of the included studies was also done by one author (PC). Duplicates were removed in EndNote (version 19), and all articles were exported from EndNote to Covidence[33] for screening of abstracts and full text. Abstract followed by full-text screening was done by the two different reviewers (JV and PD) independently and the third reviewer (BPS) resolved conflicts for screening done by the two reviewers. After full-text screening, data extraction including risk of bias assessment was done in Covidence.

Supplementary Table 1.

Search strategy for EMBASE, CENTRAL, PubMed

Search database	Search stratergy
EMBASE	(“sleep disordered breathing”/exp OR “apnea, sleep” OR “apnoea, sleep” OR “nocturnal apnea” OR “nocturnal apnoea” OR “obstructive sleep apnea” OR “obstructive sleep apnea hypopnea syndrome” OR “obstructive sleep apnea syndrome” OR “obstructive sleep apnoea” OR “obstructive sleep apnoea hypopnoea syndrome” OR “obstructive sleep apnoea syndrome” OR “obstructive sleep-disordered breathing” OR “sleep apnea” OR “sleep apnea syndrome” OR “sleep apnea syndromes” OR “sleep apnea, obstructive” OR “sleep apnoea” OR “sleep apnoea syndrome” OR “sleep apnoea syndromes” OR “sleep apnoea, obstructive” OR “sleep disordered breathing” OR “upper airway resistance syndrome”/exp OR “upper airway resistance syndrome” OR “apnea” OR “apnoea” OR “periodic apnea” OR “periodic apnoea” OR “obstructive airway disease”/exp OR “obstructive airway disease” OR “upper respiratory tract obstruction”/exp OR “upper respiratory tract obstruction” OR “airway obstruction”/exp OR “airway obstruction” OR “obstructive apnea hypopnea index”/exp OR “obstructive apnea hypopnea index” OR “sleep apnea hypopnea syndrome”/exp OR “sleep apnea hypopnea syndrome” OR “upper respiratory tract disease”/exp OR “upper respiratory tract disease” OR “obstructive apnea”/exp OR “obstructive apnea” OR “breathing disorder”/exp OR “breathing disorder” OR “apnea”/exp OR apnea) AND (“sleep apnea device”/exp OR “sleep apnea device” OR “mandibular advancement device”/exp OR “mandibular advancement device” OR “sleep apnea appliance”/exp OR “sleep apnea appliance”) AND (“positive end expiratory pressure”/exp OR “positive end expiratory pressure” OR “positive airway pressure mask”/exp OR “positive airway pressure mask” OR “bipap device”/exp OR “bipap device” OR “cpap device”/exp OR “cpap device” OR “tongue suspension device”/exp OR “tongue suspension device” OR “oral appliance”/exp OR “oral appliance” OR “oral appliance therapy”/exp OR “oral appliance therapy” OR “occlusal splint”/exp OR “occlusal splint”) AND (“randomized controlled trial”/exp OR “randomized controlled trial” OR “noninferiority trial”/exp OR “noninferiority trial” OR “controlled study”/exp OR “controlled study” OR “controlled clinical trial”/exp OR “controlled clinical trial” OR “superiority trial”/exp OR “superiority trial” OR “equivalence trial”/exp OR “equivalence trial” OR “double blind procedure”/exp OR “double blind procedure” OR “crossover procedure”/exp OR “crossover procedure” OR “single blind procedure”/exp OR “single blind procedure”)
CENTRAL	#1 (OSAHS):ti, ab, kw (Word variations have been searched) in Trials 208
	#2 (OSAS):ti, ab, kw (Word variations have been searched) in Trials 3525
	#3 (obstructive airway disease):ti, ab, kw (Word variations have been searched) in Trials 3486
	#4 (obstructive apnea):ti, ab, kw (Word variations have been searched) in Trials 5346
	#5 (obstructive sleep apnea):ti, ab, kw (Word variations have been searched) in Trials 5233
	#6 (obstructive sleep apnoea hypopnea syndrome):ti, ab, kw (Word variations have been searched) in Trials 891
	#7 (obstructive sleep apnoea hypopnea syndromes):ti, ab, kw (Word variations have been searched) in Trials 891
	#8 (obstructive sleep apnoea syndromes):ti, ab, kw (Word variations have been searched) in Trials 2180
	#9 (obstructive sleep apnoea syndrome):ti, ab, kw (Word variations have been searched) in Trials 2180
	#10 (obstructive sleep apnoea):ti, ab, kw (Word variations have been searched) in Trials 5233
	#11 (obstructive sleep apnoeas):ti, ab, kw (Word variations have been searched) in Trials 5233
	#12 (“sleep apnoea hypopnoea syndrome”):ti, ab, kw (Word variations have been searched) in Trials 325
	#13 (“sleep apnoea syndrome”):ti, ab, kw (Word variations have been searched) in Trials 2312
	#14 (“sleep apnoea syndromes”):ti, ab, kw (Word variations have been searched) in Trials 2312
	#15 (“sleep apnoea-hypopnea syndromes”):ti, ab, kw (Word variations have been searched) in Trials 0
	#16 (“sleep apnoea/hypopnea syndromes”):ti, ab, kw (Word variations have been searched) in Trials 325
	#17 MeSH descriptor: [Sleep Apnea, Obstructive] this term only 1844
	#18 MeSH descriptor: [] explode all trees 0
	#19 MeSH descriptor: [] explode all trees 0
	#20 MeSH descriptor: [Sleep Apnea Syndromes] this term only 1216
	#21 MeSH descriptor: [Sleep Apnea Syndromes] this term only 1216
	#22 #1 or # 2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16 962528
	#23 #17 or #18 or #19 or #20 or #21 in Trials 2450
	#24 #22 or #23 in Trials 950036
	#25 (mandibular advancement device):ti, ab, kw (Word variations have been searched) in Trials 212
PubMed	(((((((((Obstructive sleep apnea AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter]))) OR (Obstructive sleep apnea patients AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (obstructive sleep apnea syndrome AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (Obesity Hypoventilation Syndrome AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (Sleep Apnea, Obstructive AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (Sleep Apnea, Central AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (Sleep apnea AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter]))) AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter]))) AND ((((Mandibular advancement device AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter]))) OR (Mandibular advancement therapy AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (Mandibular repositioning device AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (Mandibular protrusion device AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter]))) AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) AND ((((((CPAP AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter]))) OR (continuous positive airway pressure AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (Oral appliance therapy AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (Occlusal splints AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (Sham appliance AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (No appliance AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter]))) AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) AND (((((Reduction in blood pressure AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter]))) OR (Sleep Quality AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (Sleep hygiene AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (AHI Index AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))) OR (Apnea Hypopnea Index
	AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter]))) AND ((randomizedcontrolledtrial[Filter]) AND (humans[Filter]) AND (english[Filter])))

Criteria for study selection

Inclusion criteria for the study were randomized controlled trial (RCT), cross-over trial (first period data were taken), following the above-mentioned PICO criteria, and published in English.

Exclusion criteria for the study were duplicate studies, studies with data errors, irrelevant outcome, case report, letter to editor, conference proceeding, systematic review, or meta-analysis.

For the included studies, data extraction and risk of bias assessment were done in Covidence[33] by the two reviewers (JV and PD) independently and consensus was reached after. For any conflicts, the third reviewer (BPS) was consulted. Data extraction was done in data extraction form of all studies in five sections:

1. Identification details included sponsorship source, country, study setting, author, E-mail, and publication details

2. Methods included design of the study, aim of the study, duration of the study, ethical approval, key conclusions of the study, method of recruitment of patients, and null hypothesis

3. Population included inclusion criteria, exclusion criteria, any group difference, population description, total number randomized, withdrawals, and exclusion of patients

4. Intervention and comparison group included total number randomized, type of intervention and with device or appliance details, number of visits, duration of follow-up, and resource requirement

5. Outcome included AHI, ESS, 24-h mean blood pressure, and patient compliance.

Data analysis was finally filled in RevMan 5.4 software[34] for statistical analyses. For missing data, the corresponding author of studies was contacted.

Risk of bias assessment: It was done as per the Cochrane Handbook of Systematic Review[35] using RoB 1.0 (Risk of Bias 1.0) having following domains: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessor, incomplete outcome, selective outcome reporting, and other bias if any. Risk of bias assessment was selected for each domain as of low, high, or unclear risk of bias with supporting comments mentioned in the article.

A fixed-effects model was used for meta-analysis to generate forest plot using instructions as mentioned in the Cochrane Handbook of Systematic Reviews.[35] Heterogeneity was assessed using I² statistic; if I² value was >50%, it was considered substantial heterogeneity. Quality of evidence was assessed through the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) recommendation.[36,37]

RESULTS

Databases were searched till July 2021; out of 305 articles, 41 studies met the inclusion criteria after full-text screening. Twenty-one studies were included for meta-analysis because outcome values in 20 articles could not be synthesized quantitatively.[16,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56 The study filtering process is depicted in PRISMA flowchart [Supplementary Figure 1 ^{(308.8KB, tif)} ].

Summary of characteristic of the included studies is presented in Table 1. Out of 21 studies,[17,19,22,23,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73] 11 studies compared MAD and CPAP, while 10 studies compared MAD and sham.

Table 1.

Basic characteristics of included studies

Study	Design	Severity of OSA	Outcome	Period	Intervention			Sample
					MAD	CPAP/sham	SHAM	MAD	CPAP	Sham
Banhiran 2018	Cross over	AHI >5	AHI ESS	12 weeks	AT-MAS (SomnoGuard AP; Tomed, Bensheim, Germany)	Transcend AUTO (Somnetics International, New Brighton, MN) with the pressure set at 5–15 cm H2O		43	43
Mehta 2001	Cross over	Mild to moderate OSA	AHI	6 weeks	MAS was custom made	-	Lower dental plate used as control	12	-	12
Venderveken 2008	Cross over	Mild OSA	AHI ESS	1, 4 months	Soft SR-Ivocap Elastomer (Ivoclar, Vivadent AG; Schaan, Liechten-stein) and provides full occlusal coverage of both dental arches	-	Thermoplastic MAD used in this study was the SomnoGuard plus, a development product designed by Tomed Dr. Toussaint GmbH, Germany	-	-	-
Dal-Fabbro 2014	Cross over	Mild to moderate OSA	AHI ESS 24 h ambulatory blood pressure	4 weeks	Mandibular advancement appliances (the BRD) were individually constructed and installed	CPAP device (REMstar Plus; Respironics Inc., Murrysville, PA, USA)	The lower arch of the same MAD was used	9	17	13
Uniken venema 2020	Longitudinal follow-up study		AHI ESS	3 months, 1, 2, 10 years	Thornton Adjustable Positioner type-1 (AirwayManagement Inc., Dallas, TX, USA)	Breas PV10 (Molnlycke, Sweaden)	-	14	17	-
Petri 2008	Three armed, parallel group design	Mild to moderate OSA	AHI ESS	4 weeks	Type of intervention and its details (type of device/material and method/technique and treatment used): The appliances were one-piece, custom-made acrylic dental devices, the acrylic covering only the molars and premolars. The appliance was secured to these teeth by four stainless steel Adams clasps in each jaw. The MAA advanced the mandible to the most protrusive position without discomfort with a 5-mm vertical opening in front	-	The appliances were one-piece, custom-made acrylic dentaldevices, the acrylic covering only the molars and premolars. The appliance was secured to these teeth by four stainless steel Adams clasps in each jaw	33	-	30
Gotsopolous 2004	Cross over	Mild to moderate OSA	AHI ESS	4 weeks	Design features of MAS: Separate upper and lower acrylic appliances anchored onto dental arches and covering the occlusal surfaces of all the teeth. Two acrylisc flanges situated bilaterally on the buccal surface of the lower appliance in the molar region. Two 10 mm “LEWA” screw devices to enable advancement of the slots; this permitsincremental protrusion of mandible	-	Control device (inactive oral appliance)	36	-	37
Nikolopolou 2017	Parallel study design	Mild to moderate OSA	AHI ESS	6 months	This oral appliance holds the mandible and tongue in a protruded position	REMstar Pro system was used (Respironics, Herrsching, Germany)	-	21	22	-
Andrean 2013	RCT	Mild to moderate	AHI ESS 24 h mean blood pressure	3 months	The active OA with mandibular advancement (OAa) was custom-made andof a monobloc design. The OAa protruded the mandible to 70%–75% of the patient’s maximum mandibular protrusive capacity (>4 mm)	-	The OAc possessed the same feature as the active device except for the lack of any mandibular advancement (<0.5 mm)	36	-	36
Bamagoos 2019	Parallel study design	AHI>10 events/h with MAS therapy (cutoff of 5 events/h)	AHI ESS Arousal index	6 weeks	MAD (5 mad positions 0%, 25%, 50%, 75%, 100% with advancement in 4–12 mm range)	CPAP	-	17	17
Blanco 2005	Cross over trial	AHI≥10/h	AHI ESS Compliance SF 36 FOSQ	3 months	Advanced group (Advanced mandible model OA)	-	Nonadvanced group (nonadvanced mandible model OA)	8	-	7
Deane 2009	Cross over trial	AHI≥10/h	AHI Compliance ESS Quality of sleep (arousal index) Subjective snoring frequency and intensity Side effects Patient satisfaction Appliance preference	1 month	MAS: Custom-made 2-piece device (Somnomed Ltd, Australia)	-	TSD: Nonadjustable silicon appliance constructed by injection molding (Aveo-TSD, Innovative Health Technologies, New Zealand)	11	-	11
Cantolla 2015	Randomized, placebo-controlled, double blinded, and crossover clinical trial	Mild-to-moderate OSA (≥5 AHI<30)	AHI Somnolence (ESS) Compliance Sleep characteristics Snoring	12 weeks	MAD The commercial device KlearwayTM (University of British Colum-bia, Vancouver, Canada)	-	Placebo device: The placebo device was the same KlearwayTM device but in centric occlusion and did not provoke mandibular advancement	39	-	38
El Solh 2011	Cross over trial	Confirmed diagnosis of OSA defined as AHI>5 by overnight polysomnography	AHI ESS	3 days	MAD: Mandibular advancement custom-made device	Combination therapy: Auto CPAP plus MAD-nasal CPAP mask and given aREMstar® CPAP machine (Respironics, Murrysville, PA)	-	10	10	-
Ferguson 1997	Cross over trial	Mild to moderate OSA	AHI ESS Compliance Efficacy Side effects Preference	4 months	AMP	nCPAP: Therapy (nCPAP) is a highly effective treatment for OSA, 2 but there can be substantial problems with patient acceptance and long term compliance	-	10	10	-
Gagnodoux 2009	Cross over trial	AHI between 10 and 60 events/h	AHI Sleepiness (ESS, OSLER) Compliance Home sleep study HRQOL Cognitive tests Side effects Preference	8 weeks	MAD: Adjustable bi-bloc acrylic oral appliance (AMCTM; Artech Me ?dical, Pantin, France)	CPAP: CPAP device (Sullivan S6EliteTM; Resmed, Bella Vista, NSW, Australia) equipped with a microprocessor and pressure monitor	-	30	29	-
Gagnodoux 2017	Cross over trial	AHI≥30	ESS Compliance AHI 24 h MAP RHI	2 months	MAD (AMO®, Orthosom, Beaucouzé, France). The MAD was custom made, consisting of an adjustable two piece acrylic OA (AMO®, Orthosom, Beaucouzé, France) with attachments of various sizes allowing mandibular advancement adjustment	-	The sham device consisted of the upper appliance only and did not advance the mandible	75	-	75
Goodoy 2017	Placebo controlled clinical trial	AHI≤5	AHI ESS PSQI FOSQ Beck anxiety and depression inventories Multiple sleep latency test PVT	1.5 year	MAD: BRD	-	Placebo: An open arch dental protection plate made of acetate with no effect on upper airway patency	15	-	15
Phillips 2013	Cross-over trial	AHI 10 events per h	24-h MAP 24-h ambulatory BP and central BP and arterial stiffness Neurolgical functional behavior and QOL using  FOSQ  SF-36  the ESS, and  The AusEd driving simulator (Austral-asian Sleep Trials Network, Australia  Side effects  Compliance  Treatment preference	1 month	MAD: Somnodent (SomnoMed Ltd., Sydney, Australia), a custom fitted and titratable two-piece device	CPAP (ResMed Autoset S8 (ResMed, Bella Vista, Australia)	-	56	52	-
Randerath 2002	Crossover trial	AHI of 5/h minimum and 30/h maximum, mild to moderate OSA	AHI Sleep quality (arousal index, respiratory induced arousals) Compliance	6 weeks	ISAD: ISAD (IST; Hinz; Herne, Germany) is an OA for the noninvasive treatment of sleep-disordered breathing	CPAP: (Max II, MAP, Martinsried, Germany; Somnotron, Wein-mann, Hamburg, Germany; and Vector, Hoffrichter, Schwerin, Germany)	-	12	8	-
Tan 2002	Crossover trial	Mild to moderate OSA (AHI between 10 and 49 events/h)	AHI ESS Patient preference	2 months	MAS	nCPAP	-	10	14	-

OSA: Obstructive sleep apnea, CPAP: Continuous positive airway pressure, OA: Oral appliances, OAc: Control OA, OAa: Active OA, TSD: Tongue stabilizing device, ESS: Epworth Sleepiness Scale, SF 36: The short form-36, QOL: Quality of life, HRQOL: Health-related QOL, OSLER: Oxford sleep resistance, RHI: Reactive Hyperemia Index, MAP: Mean arterial pressure, PSQI: Pittsburgh Sleep Quality Index, PVT: Psychomotor vigilance test, BP: Blood pressure, FOSQ: Functional Outcomes of Sleep Questionnaire, MAS: Mandibular advancement splint, BRD: Brazilian dental appliance, MAD: Mandibular advancement device, AMP: Anterior mandibular positioner, nCPAP: Nasal continuous positive airway pressure, AHI: Apnea–Hypopnea Index, RCT: Randomized controlled trials, ISAD: Intra oral sleep apnea device, MAA: Mandibular adjustable appliance, AT: Appliance therapy

In most studies, the follow-up period was 6–12 weeks, but only one longitudinal study[60] has 10 years of follow-up. Two studies are same except the study by Aarab et al.,[47] which was 1-year follow-up of Aarab et al.[16] Hence, these two studies were merged for risk of bias assessment.

Summary of risk of bias assessment and risk of bias graph is shown in Figure 1. In sequence generation, 70% of studies showed low risk of bias and 30% showed unclear risk of bias. Methods of sequence generation were computer-generated random number,[23,38,42,48,52,53,54,56,59,62,65,68,69,73,74] block randomization,[16,22,43,44,47,60,71] and block of four.[40,67] Many studies did not mention method of sequence generation.[17,19,39,41,45,46,50,51,57,58,61,63,64,66,70,72,75]

Figure 1.

Risk of bias summary and risk of bias graph

In allocation concealment, 27% of studies showed low risk of bias, 27% showed unclear risk of bias, and 45% showed high risk of bias. Methods of allocation concealment used were sealed opaque envelope,[16,23,47,52,65,66,68] telephonic allocation,[69] flipping card,[50] sequence of arrival,[48] software,[45] and method not mentioned.[17,19,20,22,38,39,40,41,42,43,44,46,49,51,53,54,55,58,59,60,61,62,63,64,67,70,71,72,73]

In blinding of participants and personnel, 65%, 20%, and 15% of studies showed low, unclear, and high risk of bias, respectively. In blinding of outcome assessor, 65%, 25%, and 10% of studies showed low, unclear, and high risk of bias, respectively.

In incomplete outcome data, 67%, 27%, and 5% of studies showed low, unclear, and high risk of bias, respectively. In selective outcome reporting, 32% and 67% of studies showed low and unclear risk of bias, respectively. Few studies mentioned trial registry number and published studies followed trial document.[17,23,38,48,52,57,60,65,68,71,73]

The GRADE score was very low, low, and moderate for different outcomes and for comparison of MAD-CPAP and MAD-sham or no treatment as shown in Table 2. The main reasons of downgrading of quality were indirectness, high risk of bias, and imprecision.

Table 2.

Summary of finding table of MAD versus Sham or no treatment (above) and MAD versus CPAP (below) using GRADE approach

Patient or population: Individual with OSA
Intervention: MAD
Comparison: Sham or no treatment
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)
	Risk with sham	Risk with MAD
AHI	The mean AHI was 0	MD 8.39 lower (10.9 lower to 5.88 lower)	-	470 (9 RCTs)	⨁⨁◯◯ Lowa, b
24 h MAP	The mean 24 h MAP was 0	MD 0.11 higher (2.81 lower to 3.04 higher)	-	229 (3 RCTs)	⨁⨁⨁◯ Moderatec, d
ESS	The mean ESS was 0	MD 0.91 lower (1.7 lower to 0.12 lower)	-	497 (9 RCTs)	⨁⨁◯◯ Lowe, f
Compliance (h/night)	The mean compliance (h/night) was 0	MD 0.84 higher (0.32 higher to 1.36 higher)	-	190 (3 RCTs)	⨁⨁⨁◯ Moderateg, h

*The risk in the intervention group (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI), ^aDowngraded by one level for serious risk of bias as majority of studies are unclear risk of bias contributed data, ^bDowngraded by one level for serious Indirectness due to various designs of MAD like monoblock and twin-block and nCPAP or standard CPAP, ^cDowngraded by one level for serious Indirectness due to various designs of MAD like Monoblock and Twin-block and nCPAP or standard CPAP, ^dDowngraded by one level for serious Indirectness due to various designs of MAD like Monoblock and Twin-block and nCPAP or standard CPAP, ^eDowngraded by one level for serious risk of bias as majority of studies are unclear risk of bias contributed data, ^fDowngraded by one level for serious indirectness due to various designs of MAD like Monoblock and Twin-block and nCPAP or standard CPAP, ^gDowngraded by one level for serious risk of bias as majority of studies are unclear risk of bias contributed data, ^hDowngraded by one level for serious indirectness due to various designs of MAD like Monoblock and twin-block and nCPAP or standard CPAP. CI: Confidence interval, MD: Mean difference, OSA: Obstructive sleep apnea, MBP: Mean blood pressure, ESS: Epworth Sleepiness Scale, AHI: Apnea Hypopnea Index, GRADE: Grading of Recommendations, Assessment, Development and Evaluation, MAP: Mean arterial pressure, RCTs: Randomized controlled trials, MAD: Mandibular advancement device, nCPAP: Nasal continous positive airway pressure, CPAP: Continuous positive airway pressure

Forest plot comparing AHI between MAD and sham included nine studies having 237 patients in the MAD group and 233 patients in the sham group [Figure 2]. Out of nine studies, 5 studies[59,62,65,66,70] favored the MAD, but 4 studies[63,64,69,71] gave unclear result. Compared with sham, MAD significantly decreased AHI (weighted mean difference: 8.39, 95% confidence level [CI]: 10.90–5.88). Figure 2 also depicts comparison of AHI between MAD and CPAP. Two studies[17,58] favored CPAP and two studies[60,19] gave unclear results. Compared with MAD, CPAP significantly decreased AHI (weighted mean difference [WMD]: 7.77, 95% CI: 5.89–9.66).

Figure 2.

Forest plot interpretation of MAD-sham (above) and MAD-CPAP (below) for AHI. MAD: Mandibular advancement device, CPAP: Continuous positive airway pressure, AHI: Apnea Hypopnea Index

Forest plot comparing 24-h mean blood pressure between MAD and sham included 113 patients in the MAD group and 116 patients in sham [Figure 3]. Three studies included for this outcome and showed unclear result. Figure 3 also shows comparison between MAD and CAPA and found unclear result (WMD: 0.50, 95% CI: −3.41~2.41).

Figure 3.

Forest plot interpretation of MAD-sham (above) MAD-CPAP (below) for 24-h mean blood pressure. MAD: Mandibular advancement device, CPAP: Continuous positive airway pressure

Forest plot comparing ESS between MAD-sham included nine studies having 249 patients in the MAD and 248 in the CPAP groups [Figure 4]. Out of 9 studies, one study[73] favored MAD than sham, and the remaining 8 studies showed unclear result. In comparison to MAD and sham, MAD significantly reduces ESS (WMD: 0.91 CI: −1.70 ~ −0.12). Figure 4 also depicts forest plot of MAD-CPAP for ESS, in which 8 studies were included. For 3 months of follow-up, seven studies favored CPAP than MAD (WMD: 0.31, 95% CI: −0.38~1.01).

Figure 4.

Forest plot interpretation of MAD-sham (above) MAD-CPAP (below) comparison for ESS. MAD: Mandibular advancement device, CPAP: Continuous positive airway pressure, ESS: Epworth sleepiness scale

Forest plot comparing patient compliance between MAD and sham showed 100 patients for MAD and 902 patients for sham group [Figure 5]. All three studies showed unclear results. Compared with MAD, sham significantly showed better patient compliance (WMD: 0.84 CI: 0.32–1.36) Figure 5 also shows comparison between MAD and CPAP with unclear result (WMD: 0.24, 95% CI: −2.27~2.74).

Figure 5.

Forest plot interpretation for MAD-sham (above) MAD-CPAP (below) for patient compliance. Mandibular advancement device, CPAP: Continuous positive airway pressure

DISCUSSION

This meta-analysis compared MAD with CPAP, sham in obstructive sleep apnea patients for AHI, ESS, 24-h mean blood pressure, and patient compliance.

Most of the studies have at least one high risk of bias mainly in allocation concealment, blinding of participants, and outcome assessor. Many studies did not register prospectively in clinical trial registry which may lead to presenting selective outcome reporting for beneficial outcomes only. This leads to negative impact on certainty of evidence and future studies should focus on methodology for adequate allocation concealment, blinding, and trial registration.

Comparison of outcomes between mandibular advancement device and continuous positive airway pressure

In MAD-CPAP comparison, CPAP showed 3.48 times (ranged from 1.76 to 5.19 times) AHI reduction in comparison to MAD, but certainty of evidence is very low. CPAP is a device which creates pressure stent to open anatomical collapse of upper airway. It comes as an air pressure creating device with tube. The tube may be attached with nasal mask or face mask or simple nasal prongs. CPAP may show 24% better compliance (ranged from − 2.27 to 2.74) to MAD due to otorhinolaryngological reasons in nasal cavity and/or paranasal sinuses such as anatomical, physiological, or pathological. Other reasons of compliance with CPAP may be related to duration of use for effectiveness (>4 h use), side effects such as dermatitis, leakage from mask, claustrophobia, discomfort in nose, and rhinitis. Hence, patients who have mild-to-moderate OSA, not suitable for MAD treatment, or did not get improvement by MAD may take advise from sleep physician or dentist for CPAP. CPAP is also recommended in patients with controlled epilepsy, edentulous, or poor dentition. In case of poor compliance, low adherence due to side effect, or higher cost of CPAP, MAD is a treatment of choice, especially in short term. A meta-analysis by Li et al.[75] showed a similar finding with significantly decreased AHI by CPAP over MAD, but no significant difference in ESS. Another meta-analysis by Schwartz et al.[30] showed significantly decreased AHI in CPAP group in comparison to oral appliance. A study of Schwartz et al. also showed significantly lower compliance of CPAP in comparison to MAD. Studies showed that MADs have better compliance and QoL than CPAP, which leads to favorable side effects, increased usage time, and low rate of withdrawal.[68]

Comparison of outcomes between mandibular advancement device and sham

In MAD-sham comparison, results favored MAD in reduction of AHI, ESS and favored sham for patient compliance and unclear results for 24-h mean blood pressure. MAD is a jaw-repositioning device that repositions the jaw by forwardly protruding mandible and hyoid bone, thus preventing upper airway collapse by contracting genioglossus and increasing retroglossal distance. It has been determined by videoendoscopic and magnetic resonance imaging studies that MAD primarily increases the upper airway volume at velopharyngeal level.[56] The forward advancement of the mandible elevates the base of the tongue and stretches the soft palate, thus helping in improving the air patency. This systematic review also found various designs of MAD, which might influence outcome values.

Earlier data suggest that inactive OA (sham) played a role in the treatment of OSA and may help in lowering AHI levels. However, various RCTs have concluded better efficiency of MAD over sham appliance or placebo.[57,58,59,60,61,62,63]

This study suggests the use of CPAP or MAD in mild-to-moderate OSA. For severe apnea, CPAP is still treatment of choice. For noncompliant patients with CPAP, MAD is recommended and vice-versa if feasible. However, patient education including anatomical, physiological, and pathological condition should be considered for effective treatment.

The GRADE results showed low or very low quality of evidence due to indirectness, high risk of bias in allocation concealment, blinding of participants, outcome assessors, and inconsistency due to heterogeneity in studies. Reason of indirectness was due to various designs of CPAP or MAD was used in different studies.

Various databases were searched: PubMed, CENTRAL, and EMBASE to search the relevant studies related to this meta-analysis. The results of this review are applicable to mild-to-moderate OSA patients with no restriction of gender and age range from 24 to 55 years.

However, there are certain limitations to the present analysis, which are as follows: (1) the number of included studies is limited; (2) individual studies differed in exclusion/inclusion criteria; (3) the courses and detail of therapy were varied; (4) the severity of OSA in patients varied between studies; and (5) pooled data were analyzed, as individual patient data was not available, precluding more in-depth analyses.

CONCLUSION

Continuous positive air pressure significantly reduces AHI in obstructive sleep apnea patients, but quality of evidence is very low in comparison to MAD. Patient compliance and 24-h mean blood pressure were not significantly different when MAD was compared to sham.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Supplementary Figure 1

PRISMA flow diagram