Sleep Breath (2013) l7:705-112DOI 10. I 007/s1 l32s-012-07 46-7

Effects of treatment with oral appliance on 24-h bloodpressure in patients with obstructive sleep apneaand hypertension: a randomized clinical trial

Ann Andr6n. Piir Hedberg. Marie-Louise Walker-Engstriim.Petra Wahl6n. Ake Tegelberg

Received: 15 May 2012 /Revised: 18 June 2012 /Accepted: 22 h:ne 2012 /Published online: 21 July 2012€; Springer-Verlag 20 12

AbstractB ac kground Continuous positive airway pressure treatmenthas been shown to lower blood pressure (BP) in patientswith obstructive sleep apnea (OSA). The aims of the present

pilot study were to evaluate the potential effects of oralappliance (OA) therapy on BP, to assess various outcomeBP n-reasures, and to inform sample size calculation.Methods Seventy-fwo patients with OSA and hyperten-sion were randomly assigned to intervention with eitheran OA with mandibular advancement (active group) or an

OA without advancement (control group). Before and

after 3 months of treatment, the patients underwent noc-turnal somnographic registration and 24-h ambulatory BPmonitoring.Results Among the various BP measures, the largest trendtoward effect of OA treatment was seen in 24-h mean sys-

tolic BP with a 1.8 mmHg stronger BP reduction in theactive group compared with controls. A stronger trend to-ward effect was seen in a subgroup with baseline ambula-tory daytime mean systolic BP >135/85 mmHg where the

A. Andr6n (F<)'4. TegelbergDepaftment of Stomatognathic Physiology,Vdstmanland County Hospital,729 81 ViisterAs, Swedene-mai1: ann.andren@bredband.net

P. Hedberg' M.-L. Walker-Engstrom. A. TegelbergCentre for Clinical Research, Uppsaia University,Vristmanland County Hospital,Viisterfls, Sweden

P. Hedberg'P Wahl6nDepartment of Clinical Physiology, Viistmanland County Hospital,Vzisteris, Sweden

A. TegelbergDeparlment of Stomatognathic Physiology, Faculty of Odontology,Mahrci University,Malmo. Sweden

mean systolic BP fell, on average, 2.6 mmHg. Addi-tional exclusion of patients with baseline apnea hypo-pnea index (AHD <15 gave a significant reduction inmean systolic BP of 4.4 mmHg (P:0.044) in the activegroup compared with controls.Conclusions ln patients with OSA and hyperlension, OAtreatment had a modest trend toward effect on reducing BP.

A stronger trend toward treatment effbct was seen afterexcluding patients with nomal baseline ambulatory BP.

Additional exclusion of patients with baseline AHI <15

showed a significant treatment effect. Data to inform sample

size for an adequately powered randomized study areprovided.

Keywords Hyperlension ' Mandibular advancement .

Obstructive sleep apnea . Oral appliance .RCT

Introduction

OSA is recognized as a serious health issue [1,2] and is the

most common sleep-related breathing disorder with a

reported prevalence of approximately 4 0% in males and

2 %o in females [3, 4]. During sleep, the breathing of an

OSA sufferer is characterized by repeated disruptions. The

sleep fragmentation and the accompanying hypoxemia leadto noctumal hyperlension l5l. Patients with OSA not onlyexperience dramatic fluctuations in BP during apneas and

hypopneas but their risk of daytime hypertension alsoincreases [6]. Hypertension has been shown to occur in a

substantial number of OSA patients, and there is a positivecorrelation between BP and apnea severity [6-8]. Dur-ing the last decade, there has been a steady increase inevidence linking OSA to long-term cardiovascular mor-bidity including hypertension, heaft failure, myocardialinfarction, and stroke [9-11].

6 Springe,

106 Sleep Breath (2013) l7:705-712

The efficacy of continuous positive airway pressure

(CPAP) treatment in reducing daytime sleepiness and low-ering systemic BP in patients with moderate to severe OSAhas been shown in a number of randomized controlledstudies 112 141. It has been demonstrated that therapeu-

tic CPAP reduces 24-h mean arterial BP by 3.3 mmHgand 24-h systolic BP by 3.4 mmHg compared withsubtherapeutic CPAP [3].

During the last two decades, there has been in-creased interest in the oral appliance (OA) as an alter-native treatment modality for OSA, showing beneficialeffects on the number of obstructive breathing events,

arterial oxygen saturation levels, and arousal frequency

[5 l7]. Results of non-controlled observational cohorlstudies [18-20] and of a small randomized controlledtrial [21] suggest that BP falls in patients with OSAtreated with an OA. To our knowledge, no previousrandomized controlled study has evaluated the effectof OA treatment on BP in patients with the combina-tion of sleep apnea and systemic hypertension. Thepresent study was designed to examine the suitableBP outcome measures and to provide data to underpinsample size calculation for a future well-powered eval-

uation of the effect of an OA on BP in patients withOSA and hypenension.

Methods

Patients

We aimed to include 70 patients with OSA and systemic

hypertension. The patients were consecutively recruitedfrom the Deparlment of Clinical Physiology at Vdstmanland

County Hospital, Vdsteris, Sweden, to where they had been

referred for an ambulatory somnographic recording. Patients

were eligible for the trial if they had verified OSA defined as

an apnea hypopnea index (AHI) >10, systemic hypertension

deflned as office systolic BP >140 mmHg or diastolicBP >90 mmHg at two separate occasions, and were not

currently being treated with an OA or CPAP. Patients

also had to possess a sufficient number of teeth for the

retention of an OA. Exclusion criteria included officesystolic BP >180 rnmHg or diastolic BP >110 mmHg,

body mass index (BMl) >35, atrial fibrillation, chronic ob-

structive lung disease, epilepsy, severe psychiatric disease,

maximal protrusion of the mandible <6 mm, and an inabiliffto speak or understand the Swedish language.

A11 patients gave written informed consent in accordance

with the Helsinki declaration and were also asked not toalter their medication during the 3-month study period.

Approval was obtained from the Uppsala University Ethics

Committee (Dnr 2005 :221 ).

A Springer

Study protocol

The study was conducted as a randomized controlledpilot trial. After inclusion. patients were assigned tointervention with either an OA treatment with mandibularadvancement >4 mm (active group) or an OA withoutmandibular advancement <0.5 mm (control group). Ran-

domization was made in blocks of four. Sequence allo-cation was determined by a random number generator. Atbaseline, all patients underwent an ambulatory noctumal

somnographic recording, 24-h ambulatory BP monitoring(ABPM), and completed a questionnaire conceming theirgeneral medical condition. Daytime sleepiness was assessed

with the Epworlh Sleepiness Scale (ESS) 122). After3 months of OA treatment, the patients underwent repeat-

ed ambulatory somnographic recording and 24-h ABPM.The outcome variables were the mean changes in the

different 24-h ABPM values from baseline to 3 months

of OA treatment.

Intervention

The treatment was performed with an OA. The active OAwith mandibular advancement (OAa) was custom-made and

of a monobloc design, as previously described by Tegelberg

et al. [23] (Fig. l). The OAa prohuded the mandible to 70-75 Yo of the patient's maximum mandibular protrusive ca-

pacity (>4 mm). The control OA (OAc) possessed the same

feature as the active device except for the lack of any

mandibular advancement (<0.5 mm). The patients were

informed that there were two types of devices to be evalu-

ated but not infotmed on which one of the devices they

would receive. Furtheq they were informed on their AHIvalue and office BP, but not ambulatory BP values at base-

line. A11 the treatment and oral measurements were per-

formed by the same dentist, and this person was not

involved in the outcome assessments and blinded to allbaseline ambulatory BP measures.

Sleep studies

Ambulatory somnographic recordings were made with a

validated porlable digital recording unit with sensors forthe registration of airflow, saturation, respiratory move-

ments of the chest, body position, and snoring sounds(Embletta PDS device; Medcare Flaga, Iceland) [2a]. The

recordings were undertaken in the patient's home, transmit-ted to a computer, and analyzed manually by one experi-

enced technician blinded to the intervention fype. At the 3-

month follow-up, the patients slept with the OA in situ

during registration.

Apnea was defined as a cessation of airflow for at least

10 s. Hypopnea was defined as a reduction in airflow of at

Sleep Breath (2013) 17:705-712 70'7

Patients consecutivelyscreened(n = 1623)

Office BP< 140/90 mmHg

(n = a66)

Office BP missing(n = 132)

Patients eligible for thestudy

(n = 137)

Allocated to active OA(n = 36)

Fig. 1 Study flow charl

least 50 o/o for 10 s or more with an accompanying desatura-

tion of at least 4 o/o. The events were considered to be obsfiuc-tive if they occulred in association with thoracic-abdominalbreathing efforts. The number of apneas and hypopneas per

hour of sleep was calculated to obtain the AHI. Mild, moder-

ate, and severe OSAwas def,med as AHI <15, 15-30, and >30,

respectively.

Ambulatory blood pressure monitoring (ABPM)

ABPM recordings were perforrned using a validated device(Oscar 2; Suntech Medical Instn.rments, Raleigh, NC, USA)

[25]. Office BP was recorded in both arns after 5-min rest ina lying position using a conventional sphygmomanometer

and an appropriate cuff size. The ABPM cuff was placed on

the non-dominant afln. However, if on conventional sphyg-

momanometric measurement the difference in systolic BP

between both arms was >10 mmHg, the arm with the high-est reading was chosen. Before starting the study, the reli-ability of BP values measured with the monitor was checkedagainst simultaneous measurements with a conventionalsphygmomanometer. Differences of <10 mmHg wereallowed. BP was continuously recorded every 20 min duringa 24-h period. The wake and sleep intervals were defined

based on diary entries. The data were recorded in the digitalmemory transferred to a computer, arrd analyzed by one

experienced technician blinded to the intervention type.

Statistical analysis

Descriptive continuous variables were presented as means

and standard deviations, and categodcal variables as numb-ers and percentages. Mean changes in AHI and ESS frombaseline to follow-up were compared between groups using

covariance analysis adjusting for the baseline level of the

dependent variabie. The mean differences in BP changes

from baseline to follow-up between the allocated groups

were calculated with 95 oh corfidence intervals. As this

was a pilot study, we did not anticipate definitive answers

to the randomized comparisons and, thus, no formal statis-

tical hypothesis testing was perfofined. However, for ex-ploratory purposes the effects with the largest magnitudes

were statistically tested using unpaired I tests. In subgroup

analyses, we excluded patients with normal ambulatory BP

defined according to the European Hypertension Society as

daytime mean systolic BP 5135 mmHg and daytime mean

diastolic BP S85 mmHg [26]. Finally, we excluded patients

with baseline AHI S15 and normal ambulatory BP.

Data were analyzed on an intention-to-treat basis, includ-ing all the data obtained even if patients were known to have

not complied with the OA treatment. Missing follow-up data

were imputed by the "baseline observation carried forward"approach. This was based on our assumption that whenpatients withdrew from the study their BP values wouldretum to baseline levels.

Sample size calculations were made by using the ob-

seryed mean BP changes (and their standard deviations)from baseline to follow-up for the active and placebo treat-

ment groups, respectively. We set the statistical power to80 o/o and used a two-tailed test with the significance levelset at 5 o/o.

Potential determinants of 24-h mean systolic BP at

3 months follow-up (24-h meat systolic BP at baseline,

sex, age, body mass index, AHI reduction from baseline to

follow-up, and ESS reduction from baseline to follow-up)were evaluated by multiple linear regression. Subsequently,

an interaction term between treatment group and AHI re-

duction was included in the regression model to assess the

effect of treatment group on the association between AHIreduction and BP at follow-up.

6 Springer

708 Sleep Breath (2013) 17:105 712

For all comparisons, P <0.05 was considered statisticallysigrtificant. A11 statistical analyses were perfonned usingSTATA version 11 (StataCorp, College Station, TX, USA).

Results

A total of 1,623 patients were consecutively screened(Fig. 2). One thousand four hundred and eighty-six patients

were not eligible for the study because of an AHI <10 (n:888), office BP <140/90 mmHg (n:a6Q, or missing officeBP measurements (n:132). Among 137 eligible patients,

six were excluded because of severe hypertension or severe

psychiatric disease, and 59 declined to participate. Finally,72 patients were randomized into the study. In the subgroupanalysis, four patients were excluded from the active ffeat-ment group and two from the control group because ofnormal ambulatory BP (i.e., ABPM daytime mean BP<135/85 mmHg). One patient from the control group with-drew from the study and did not attend follow-up. Twopatients in the active group did not use their OA butattended follow-up and were analyzed as members of the

active group according to the intention-to-treat approach.The patients were predominantly men, middle-aged, and

overweight (Table l). The mean mandibular advancement inthe active and control groups was 6.4 mm (SD 1.2) and

0.0 mm (SD 0.2), respectively.

After 3 months of treatment, there was a significant mean

reduction of the AHI in the active compared with the controlgroup ( 15.4 vs. -3.8; P<0.001; Fig. 3). Normalization ofAHI (i.e., reaching an AHI <10) was significantly more com-rnon in the active than in the control group (78 o/ovs.25 o/o; P<0.001). The ESS scores improved significantly in the active

compared with the control group (-4.3 vs. -2.1; P<0.006).Table 2 shows the differences in change of various BP

measures from baseline to follow-up between the active and

Fig. 2 The oral appliance used were manufactured from heat-curedacrylic polymer in a monobloc design

6 Springer

Table I Baseline characteristics ofthe patients according to allocatedfeatment group

Active group Control groupn-36 n:36

Male, n (%)

Age (years)

BMI (kg/m'z)

Current srnoking. n (%)

Diabetes mellitus, n (%)

Hypnotic medication, n (%)

Antihypertensive medication,n (l%)

Offrce systolic BP (mmHg)

OtIce diastolic BP (nimHg)

ESS

AHI

AHI l0 t5,n(%)AHI l6-30, n (%)

AHI >30, z (%)

30 (83)

57+8

30+4

1 (te)1 (3)

3 (8)

e (2s)

149+ 1 8

90+'7

11+5.4

23+16

14 (39)

t7 (47)

s (14)

27 (7s)

59+9

29.I4

6 (t1)0 (0)

2 (6)

17 (47)

154+16

90+9

l1+4.5

24+17

11 (31)

t6 (44)

e (25)

Data are presented as mean + SD unless otherwise stated

AHI apnea hlpopnea index, BP blood pressure, ES,S Epwofth Sleepi-ness Scale

the control groups. There were modest trends towardreductions in all BP measures from baseline to follow-up. The largest effect was seen in 24-h mean systolic BP

with a mean reduction of 1.8 mmHg (95 % CI -2.1 to

5.7; P:0.356) in favor of the active compared with the

control group. Sample size calculation revealed a need

for a minimum of 660 randomized patients to verify a

statistically significant difference in the 24-h mean sys-

tolic BP reduction of 1.8 mmHg between the active and

control groups.

In a subgroup analysis, we excluded patients with ambu-

latory daytime mean BP:135/85 mmHg. In this subgroup,

there were stronger trends toward treatment effects on all BP

measures (Table 3). h 24-h mean systolic BP, there was amean reduction of 2.6 mmHg (95 % CI -1.4 to 6.6; P:0.204) it the active group compared with controls. A min-imum of 308 randomized patients would have been required

to detect a difference of 2.6 mmHg inthe24-h mean systolicBP reduction between the active and control groups.

ln a subgroup with ambulatory daytime mean systolicBP >135/85 mmHg and AHI >15 at baseline (n:46),there was a statistically significant mean reduction in24-h mean systolic BP of 4.4 mmHg (95 % CI 0.1 to8.1; P:0.044) between the active and control groups.

A multiple linear regression analysis revealed thatAHI reduction from baseline to follow-up (beside base-

line BP leve1) was the only statistically significantmarker of 24-h mean BP level at follow-up (Table 4).

There was no interaction of treatment group on the

Sleep Breath (2013) 17:105-712 709

Active groupFig. 3 Change in the apneahypopnea index (AHI) frombaseline to follow-up accordingto treatmgnt group

baseline

association between AHI reduction and 24-htolic BP level at follow up (P:0.282).

Discussion

Control group

follow-up baseline follow-up

BP S135/85 mmHg. In this subgroup with manifesthyperlension, the 24-h mean systolic BP fell on average2.6 mmHg in the active compared with the conffol group. Ina subgroup with ambulatory daytime mean BP >135/85 mmHg and AHI >15 at baseline, we found a statisticallysignificant mean reduction o14.4 mmHg in 24-h mean systolicBP in favor of the active treatment group.

Systemic hypertension occurs in 28-57 % of patientswith OSA [8], and there is a dose response associationbetween BP and apnea severity 16, 1,271. In patients withan AHI >15, the risk of simultaneous hypertension wasaimost tripled [6]. As both these conditions, hypertensionand OSA, increase the risk of cardiovascular and

mean svs-

In the present study of OA treatment in patients withOSA, we found a modest trend toward effect ol reduc-tion on BP. Among the various BP measures in thisstudy, the 24-h meat systolic BP indicated the largestreduction (1.8 mmHg) in favor of the active group. Astronger OA effect trend on BP was seen after theexclusion of patients with an ambulatory daytime mean

Table 2 Effect oforal appliancetrealment on blood pressure inpatients with office blood pres-sure >140/90 mmHg at baseline

Data are presented as mean + SDunless otherwise stated

MAP meat arterial pressureuDifference in blood pressurechange from baseline to fo11ow-

up between the active and thecontrol groups

BP (mmHg) Active group n:36 Control group n=36

Baseline Follow-up Basel'ine Follow-up

Net difference betweengroups" (95 % CI)

24-h meansystolic

24-h meandiastolic

24-h MAP

Daytime meansystolic

Daytime meandiastolic

Daytime MAP

Night timemean systolic

Night timemean diastolic

Night time MAP

136.9+ 10.8

83.8 +8.7

102.0+9.0

143.6 + 10.5

89.6+8.8

1 07.6 + 8.5

123.9+ 13.9

73.1+10.'.7

89.9+ 1 1.4

I 34.6 + 10.4

8l .6+8.7

99.2 +8.8

141.3 + 10.5

87.1 +8.3

105.1+8.4

121 .4+13.3

71.0+il.1

87.8 + 1 1.5

139.3 +9.8

83.4+ 8.6

102.0+1 .9

145.4t9.4

88.4+9.2

107.5 + 8.2

127.4+12.'l

72.9+8.3

91.0+8.7

138.8+ 1 1 .0

82.2+9.1

99.9+10.2

144.9+10.9

87.L.9.5

105.0+ 1 1.2

126.4+tt 3

72.t+1 .9

89.9+8.5

1.8 (-2.1 to 5.7)

1.0 (-1.9 to 3.9)

0.7 ( 2.6to4.1)

1.8 ( 2.1 to 5.8)

1.2 ( 1.4 to 3.9)

0.1 (-3.s to 3.7)

t .4 (-2.5 to s.3)

1.3 ( 1.5 to 4.1)

0.8 ( 2.s to 4.0)

A Springer

710 Sleep Breath (2013) 17:705-712

Table 3 Effect oforal appliancetreatmenl on blood pressure in

patients with office blood pres-

sure >140/90 mmHg and ambu-latory daytime n-rean bloodpressure >135/85 mmHg atbaseline

Data are presented as mean t SDunless otherwise stateduDifference in blood pressurechange from baseline to fo1low-up between the active and thecontrol groups

BP (mmHg) Active group n =32 Control group n:34

Baseline Follow-up Baseline Follow-up

Net difference betweengroupsu (95 % CI)

24-h meansystolic

24-h meandiastolic

24-h MAP

Daytime meansystolic

Daytime meandiastolic

Daytime MAP

Night timemean systolic

Night timemean diastolic

Night tims MAP

138.8+9.9 135.7+10.3

85.1+8.1 82.6+8.5

103.7+8.0 100.2+8.5

145.7 +9.1 142.7 +t0.1

9l .2+1 .8 88.3 +7.5

109.4+1 .2 106.4+7.7

t25.7 +13.5 122.2+13.7

74.3't10.6 71.4.}11.4

91 .3t1t.2 88.3+11.8

140.3+9.1 139.8+10.5

84.6+7.0 83.4+8.0

t03.2+6.4 100.9+9.6

146.3+8.8 146.0r10.3

89.7+7.7 88.3+8.5

108.6+6.7 106.1+10.6

128.2+12.4 127.0+1t.4

73.8+7 .4 72.8+7 .3

91.9I8.0 90.6+8.2

2.6 (-1 .4 to 6.6)

1.3 ( 1.8 to 4.3)

1.2 ( 2.3 to 4.7)

2.6 ( 1.4 to 6.7)

1.5 ( 1.4 to 4.3)

0.4 ( 3.4 to 4.3)

2.4 ( 1.4 to 6.2)

1.9 ( 0.8 to 4.6)

1 .5 (-1.7 to 4.7)

cerebrovascular diseases, the need for different treatmentmodalities is strengthened [6]. Antihyperlensive treatment

is one modality and epidemiological evidence suggests that

a long-tem difference of about 10 years is associated with20-25 % less cardiovascular disease and a 20 7o reduction inthe risk of stroke if the BP reduction is maintained over time

[28]. Other modalities, such as CPAP and OA, are intended

to reduce the upper airway obshuction in patients with OSA.

An OA is an effective treatment, especially in OSA of mildto moderate severity U 6, 29, 30]. For patients with moderate

to severe OSA, CPAP is the most effective treatment and is

recommended as the first treatment option [3 I ].Bazzano et al. described large variations in BP reduction

in a literature review of several RCTs of CPAP treatment

[14]. It is difficult to draw conclusions from the results in all

Table 4 Muitiple linear regression including possible determinants of24-h mean systolic blood pressure at 3 months follow-up in the studypopulation (n:72)

l, SE P value

of the studies due to different inclusion criteria and

methods. When restricted to studies that provided sham

CPAP as a control treatment, they found significantreductions for both systolic and diastolic BP (-3.1 and

2.2, rcspectively). However, a statistically significantmean net reduction in BP was seen more often among

studies with pafiicipants who had higher baseline BP

levels, higher BMI, and more severe OSA.What is the clinical relevance of the changes in BP?

Studies using conventional antihypefiensive agents in non-

OSA populations indicate that a reduction of 5 mmHg indiastolic blood pressure is associated with a 42 %o decrease

in stroke and a 74 0% decrease in coronary heart disease [32].Hypertension is traditionally diagnosed after office BP

measurements. However, there is an increasing use of ABPMin clinical practice. Recently, the British National Institute forHealth and Clinical Excellence recommended ABPM, rather

than office BP measurement, as the method by which to

confirm the diagnosis of hypertension [33]. Based on a mul-titude of repeated measurements, ABPM provides a more

accurate representation of a patient's overall BP load. There

are nurerous outcome studies showing that ABPM is a better

predictor of organ damage and cardiovascular events com-

pared with office BP readings [34]. Not surprisingly, our

findings indicated a stronger OA treatment trend toward effect

in patients with a higher BP load, i.e., in patients with ambu-

latory daytime mean BP >135/85 mmHg.

Our aim was to evaluate OA treatment for both of the

conditions, OSA and hyperlension. Up to the present, there

have only been a few published non-randomized controlledstudies which have evaluated this combination, and the

effect of OA treatment on BP in hypertensive patients is

not yet established. Gotsopoulos et al. performed a random-

ized controlled crossover trial in 61 patients to evaluate the

24-h mean systolic BP at baseline 0.651(1 mmHg increase)

Treatment group (active vs. control) 0.003

Age (1 year increase)

Sex (ma1e vs. female)

BMI (1 kg/m2 increase)

0.0s I <0.001

1.93 0.998

0.1 1 8 0.508

2.41s 0.92s

0.242 0.099

0.698 0.012

0.249 0.620

0.078

0.229

-0.406AHI reductionu (10 units increase) -1.799ESS reductionu (1 unit increase) 0.124

Dependent variable is 24-h mean systolic BP at 3 months follow-up.Adjusted R2 :0.590

AHI apneahypopnea index, BMI body mass index, ,BP blood pressure,

ESS Epworth Sleepiness ScaleuReduction flom baseline to 3 months of fol1ow-up

6 Springer

Sleep Breath (2013) 17:705 712

effect on BP of an OA with mandibular advancement com-pared with pn OAwithout advancement [21]. The inclusioncriteria included AHI >10 but not hypertension, although

39 o/o of the patients were on antihypertensive treatment.

Gotsopoulos et al. reported a non-significant reduction in24-h mean systolic BP of 1.5 mmHg in favor of active

treatment corresponding to our result of 1.8 mmHg. How-ever, they found a statistically significant reduction in24-h mean diastolic BP of 3.0 mmHg, which is consid-erably larger than our finding of 1.0 mmHg. Discrep-ancies between these study results may be explained bydifferences in characteristics of the study population. In

comparison with the present study, Gotsopoulos et al.

examined younger subjects where a minority were hy-pertensive as reflected by a lower average 24-h meanBP of 121 118 mmHg at baseline. An obvious advantage

of Gotsopoulos' study was the crossover design. requiringfewer subjects to reach statistically conclusive results.

When comparing the results in these kinds of studies withdifferent treatment modalities to reduce BP, there are com-plicating influencing factors such as the severity of OSA,daytime sleepiness, degree of hypertension, BMI, and the

use of antihyperlensive medication. In both OA and CPAP

treatment, one critical point for success is compliance, as the

patient needs to use the appliance or CPAP regularly [23].The majority of the patients in our study used their OAsregularly, which demonstrates a good acceptance of the

treatment modality. Compliance in the use of an OA was

comparable with other studies with similar follow-upperiods 130,351. In the present study, the patients'

self-repofied frequency of using their appliances was,

on average, 6.4 nights per week and they were used

during the whole sleeping time. In CPAP studies, the

use of the equipment was objectively recorded between

3.3 and 4.9 h per night [12, 13].

In earlier studies, AHI reduction was considered to be

due to the degree of mandibular advancement of the lowerjaw [36]. In a study by Kato el a1., different degrees ofmandibular advancement in the same patient gave an im-provement of20 oh in reduced number ofnocturnal desatu-

rations for each 2 mm of advancement and the therapeutic

effect appeared to be >4 mm advancement [37]. In the

present study, all the patients in the active group received a

standardized mandibular advancement >4 mm corresponding

to 70-75 % of the individual patient's maxin'tum mandibular

protrusive capacity.

In our study, the sleep-breathing disorder measured as

apnea frequency was significantly reduced by the mandibu-

lar advancement with an OA compared with no mandibular

advancement. However, 25 o/o of the patients with no man-

dibular advancement in their treatment (control group) were

responders to the treatment. The result might be due to

regression to the mean or an effect of the treatment. When

the lower jaw is held in a fixed position, even with less than

0.5 mm mandibular advancement, it may be enough in some

cases to stabilize the pharyngeal tissues and, thereby, keep

the airways open. In our findings, the main determinant ofBP reduction after 3 months of follow-up was the reductionin AHI irrespective of treatment group (Table 4). Thus, it isplausible to expect a BP reduction among the responders inthe control group leading to an attenuation ofthe differences

in BP outcomes between the treatment groups. This is a

limitation of the present study which is difficult to avoid.

Other limitations of this study are the relatively low grade ofhyperlension and the fact that only every fifth patient had a

severe OSA condition. These characteristics are other pos-

sible sources of attenuation in the difference in outcome

between the active and control groups. Finally, the non-

crossover design of the present study was a limitation to

reach statistically conclusive results.

It could be concluded that in patients with OSA and

hyperlension, OA treatment had a modest trend towardeffect on reducing BP. A minimum of 660 randomizedpatients would be required to verifiz a statistically significantdifference in the 24-h mean systolic BP reduction of1.8 mmHg found in the present study. Restricting the inclu-

sion to patients with ambulatory daytime mean BP >135/

85 mmHg, a minimum of 308 randomized patients would be

required. In a subgroup of patients (n:aO with ambulatory24-h mean systolic BP >135/85 mmHg and AHI >15 at

baseline, our sample size was enough to demonstrate a

significant treatment effect on BP.

Acknowledgments The project was financially supported by grants

from the County of Vtistmanland, Sweden, the Swedish Heart and

Lung Foundation, and the Swedish Dental Association.The authors wish to thank Yvonne Pettcrsson, dental assistant, for

her active role in the administration and rolling ofthe patients.

References

1. Somers VK, White DP, Amin R, Abraham WT, Costa F, Culebras

A, Daniels S, Floras JS, Hunt CE, Olson LJ, Pickering TG, Russell

R, Woo M, Young T (2008) Sleep apnea and cardiovasculardisease. Circulation 118(10):1080-11 I I

2. Shah NA, Yaggi HK, Concato J, Mohsenin V (2010) Obstructivesleep apnea as a risk factor for coronary events or cardiovasculardeath. Sleep Breathing 14(2):13 l-l 36

3. Bixler EO, Vgontas AN, Lin H-M, Have T! Rein J, Vela-BuenoA, Kales A (2001) Prevalence of sleep-disordered breathing in

women. Am J Respir Crit Care Med 163(3):608 613

4. Young ! Palta M, Dempsey J, Skatrud J, Weber S, Badr S (1993)

The occumence of sleep-disordered breathing among middle-agedadults. N Engl J Med 328(11):1230-1235

5. Davies RJ, Belt PJ, Roberts SJ, Ali NJ, Stradling JR (1993)

Arterial blood pressure responses to graded transient arousal fromsleep in normal hun.tans. J Appl Physiol 74(3):1 123 1130

A Sprirger

712 Sleep Breath (2013) 17:705-712

6. Peppard PE, Young T, Palta M, Skatrud J (2000) Prospective studyof the association between sleep-disordered breathing and hyper-tension. N Engl J Med 342(19):l378-1384

7. Lavie P, Herer P, Hoffstein V (2000) Obstructive sleep apnoeasyndrome as a risk factor for hypertension: population study. BMJ320(7233):479482

8. Nieto FJ, Young TB, Lind BK, Shahar E, Samet JM, Redline S,

D'Agostino RB, Newman AB, Lebowitz MD, Pickering TG (2000)Association ofsleep-disordered breathing, sleep apnea, and hyper-tension in a large community-based study. JAMA 283(14):1829-1 836

9. Yaggi HK, Concato J, Keman WN, Lichtman JH, Brass LM,Mohsenin V (2005) Obstructive sleep apnea as a risk factor forstroke and death. New Engl J Med 353(19):2034 2041

10. Marin JM, Carrizo SJ, Vicente E, Agusti AGN (2005) Long-termcardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positiveairway pressure: an observational study. Lancet 365(9464):10461 053

ll. Shahar E, Whitney C, Redline S, Lee E, Newman A, JavierNieto F,

O'Connor G, Boland L, Schwartz J, Samet J (2001) Sleep-disordered breathing and cardiovascular disease. Am J Resp CritCare Med 163(1):19 25

12. Faccenda J, Mackay ! Boon N, Douglas N (2001) Randomizedplacebo-controlled trial of continuous positive airway pressure onblood pressure in the sleep apnea-hypopnea syndrome. Am J Resp

Crit Care Med 163(2):344 34813. Pepperell JCI Ramdassingh-Dow S, Crosthwaite N, Mullins R,

Jenkinson C, Sfadling JR, Davies RJO (2002) Ambulatory bloodpressure after therapeutic and subtherapeutic nasal continuouspositive airway pressure for obstructive sleep apnoea: a rando-mised paraliel tria1. Lancet 3 59 (9 3 02):20 4-2 I 0

14. Bazzano LA,KhatZ, Reynolds K, He J (2007) Effect of noctumalnasal continuous positive airway pressure on blood pressure inobstructive sleep apnea. Hypertension 50(2):411 423

15. Mehta A, Qian J, Petocz P, Darendeliler MA, Cistulli P (2001) Arandomized, controlled study of a mandibular advancement splintfor obstmctive sleep apnea. Am J Resp Crit Care Med 163(6):1457 146l

16. Walker-Engstrcim M-L, Tegelberg A, Wilhelmsson B, Ringqvist I(2002) 4-year follou,-up of treatment with dental appliance oruvulopalatopharyngoplasty in pationts with obstructive sleep ap-nea. Chest 121(3):739 146

17. Aarab G, Lobbezoo F, Hamburger HL, Naeije M (2011) Oralappliance therapy versus nasal continuous positive airway pressure

in obstructive sleep apnea: a randomized, placebo-controlled trial.Respiration 8 1(5):41 1-419

18. Otsuka R, Ribeiro de Almeida F, Lowe A, Linden W, Ryan F(2006) The effect of oral appliance therapy on blood pressure inpatients with obstructive sleep apnea. Sleep Breathing 10(1):29-36

19. Yoshida K (2006) Effect on blood pressure of oral appliancetherapy for sleep apnea syndrome. Int J Prosthodont 19(1):61-66

20. Andr6n A, Sjoquist M, Tegelberg A IzOOO; Effects on bloodpressure after treatment of obstructive sleep apnoea with a man-dibular advancement appliance-a three-year follow-up. J OralRehabil 36(10):719 12s

21 . Gotsopoulos H, Kelly JJ, Cistulli PA (2004) Oral appliance therapyreduces blood pressure in obstructive sleep apnea: a randomized,controlled trial. Sleep 27(5):934 941

22. Johns MW (1991) A new method for measuring daytime sleepi-ness: the Epworlh sleepiness scale. Sleep l4(6):540 545

23. Tegelberg A, Wilhelmsson B, Walker-Engstrom ML, Ringqvist M,Andersson L, Krekmanov L, Ringqvist I (1999) Effects and

adverse events of a dental appliance for treatment of obstructivesleep apnoea. Swed Dent J 23(.4):111 126

24. Ng SSS, Chan T-O, To K-W, Ngai J, Tung A, Ko FWS, Hui DSC(2010) Validation ofEmbletta porlable diagnostic system for iden-tiflzing patients with suspected obstructive sleep apnoea syndrome(OSAS). Respirology \5(2):336 342

25. Jones SC, Bilous M, Winship S, Finn B Goodwin J (200a) Vali-dation of the OSCAR 2 oscillometric 24-hour ambulatory bloodpressure monitor according to the Intemational Protocol for thevalidation of blood pressure measuring devices. Blood Press Monit9(4):219 223

26. O'Brien E, Asmar R, Beilin L, lmai Y, Mallion JM, Mancia G,Mengden ! Myers M, Padfield P, Palatini P, Parati G, Pickering T,

Redon J, Staessen J, Stergiou G, Verdecchia P, European Society

of Hypertension Working Group on Blood Pressure (2003)European Society of Hypertension recommendations for con-ventional, ambulatory and home blood pressure measurement.

J Hypefiens 21(5):821-84827. Grote L, Ploch ! Heitmann J, Knaack L, Penzel ! Peter JH (1999)

Sleep-related breathing disorder is an independent risk factor forsystemic hypertension. Am J Resp Crit Care Med 160(6):18751 882

28. MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J,

Abbott R, Godwin J, Dyer A, Stamler J (1990) Blood pressure,

stroke, and coronary heaft disease. Part 1, prolonged differences inblood pressure: prospective observational studies corrected for the

regression dilution bias. Lancet 335(8692):165 71429. Lim J, Lasserson TJ, Fleetham J, Wright J (2004) Oral appliances

for obstructive sleep apnoea. Cochrane Database Syst Rev (4):

cD00443530. Hoekema A, Stegenga B, De Bont LG (2004) Efficacy and co-

morbidity of oral appliances in the treatment of obstructive sleepapnea-hypopnea: a systematic review. Crit Rev Oral Biol Med 15

(3):137 1ss31. Franklin K, Rehnqvist N, Axelsson S (2007) Obstructive sleep

apnoea syndrome-repofi of a joint Nordic project. A systematic

review. Swedish Council on Health Technology Assessment,Stockholm

32. Collins R, Peto R, MacMahon S, Hebert P, Fiebach NH, EberleinKA, Godwin J, Qizilbash N, Taylor JO, Hennekens CH (1990)Blood pressure, stroke, and coronary heart disease. Part 2, short-

term reductions in blood pressure: overview of randomised drugtrials in their epidemiological context. Lancet 335(8693):827-838

33. Krause ! Lovibond K, Caulfield M, McCormack ! Williams B(2011) Management of hyperlension: summary of NICE guidance.

BMJ 343:d489134. Verdecchia P, Angeli F,Mazzolta G, Gentile G, Reboldi G (2009)

Home blood pressure measurements will not replace 24-hour am-

bulatory blood pressure monitoring. Hypertension 54(2): 1 88-1 95

35. Wilhelmsson B, Tegelberg A, Walker-Engstrom ML, Ringqvist M,Andersson L, Krekmanov L, Ringqvist f (1999) A prospectiverandomized study of a dental appliance compared with uvulopala-topharyngoplasty in the treatment of obstructive sleep apnoea.

Acta Otolaryngol 1 1 9(4):503-50936. Walker-Engstrom ML, Ringqvist I, Vestling O, Wilhelmsson B,

Tegelberg A (2003) A prospective randomized study comparingtwo different degrees of mandibular advancement with a dental

appliance in treatment of severe obstructive sleep apnea. Sleep

Breathing 7(3):119 130

37. Kato J, lsono S, Tanaka A, Watanabe ! Araki D, Tanzawa H,Nishino T (2000) Dose-dependent effects ofmandibular advance-

ment on pharyngeal mechanics and nocturnal oxygenation inpatients with sleep-disordered breathing. Chest l11 (4):1065-1072

A Springet