Table 4Clinical evidence summary: Auto-CPAP versus fixed level CPAP for improving usage of continuous positive airway pressure machines in adults with OSAHS –severe OSAHS

Outcomes

No of Participants

(studies)

Follow up

Quality of the evidence (GNRADE)Relative effect (95 % CI)Anticipated absolute effects
Risk with ControlRisk difference with Auto-CPAP versus fixed CPAP (95% CI)

Machine usage (hours/night)

Median follow-up 6 weeks

1452

(31 studies)

⊕⊕⊝⊝

LOW1,5

due to risk of bias, indirectness

control group risk not available6

The mean machine usage (hours/night) in the intervention groups was

0.21 higher

(0.11 to 0.31 higher)

Number of participants who used CPAP therapy > 4 hours per night

Follow-up range 3 to 16 weeks

346

(2 studies)

⊕⊕⊝⊝

LOW1,5

due to risk of bias, indirectness

RR 1.06

(0.9 to 1.24)

Moderate
448 per 1000

27 more per 1000

(from 45 fewer to 108 more)

Symptoms (Epworth Sleepiness Scale)

Scale 0 to 24

Higher is worse

Median follow-up 6 weeks

1285

(25 studies)

⊕⊕⊝⊝

LOW1,5

due to risk of bias, indirectness

control group risk not available6

The mean symptoms (epworth sleepiness scale) in the intervention groups was

0.44 lower

(0.72 to 0.16 lower)

Withdrawals (parallel group trials/first arm crossover trials)

Median follow-up 6 weeks

1275

(13 studies)

⊕⊝⊝⊝

VERY LOW1,2,5

due to risk of bias, imprecision, indirectness

RR 0.91

(0.67 to 1.24)

Moderate
80 per 1000

7 fewer per 1000

(from 26 fewer to 19 more)

Quality of life (Functional Outcome of Sleep Questionnaire)

Scale from 5-20

Higher is better

follow-up range 4 to 104 weeks

352

(3 studies)

⊕⊝⊝⊝

VERY LOW1,,5

due to risk of bias, indirectness

control group risk not available6

The mean quality of life (functional outcome of sleep questionnaire) in the intervention groups was

0.12 higher

(0.21 lower to 0.46 higher)

Quality of life (Sleep Association Quality of Life Index)

Scale from 1-7

Higher is better

97

(2 studies)

⊕⊝⊝⊝

VERY LOW1, 5

due to risk of bias, , indirectness

control group risk not available6

The mean quality of life (sleep association quality of life index) in the intervention groups was

0.14 lower

(0.54 lower to 0.27 higher)

Quality of life (SF-36 questionnaire) - Physical functioning

Scale from 0-100

Higher is better

60

(3 studies)

⊕⊕⊝⊝

LOW1,2

due to risk of bias, imprecision

control group risk not available6

The mean quality of life (sf-36 questionnaire) - physical functioning in the intervention groups was

0.76 higher

(3.5 lower to 5.01 higher)

Quality of life (SF-36 questionnaire) - Role physical

Scale from 0-100

Higher is better

60

(2 studies)

⊕⊕⊝⊝

LOW1,2

due to risk of bias, imprecision

control group risk not available6

The mean quality of life (sf-36 questionnaire) - role physical in the intervention groups was

3.73 lower

(13.46 lower to 6.01 higher)

Quality of life (SF-36 questionnaire) - Bodily pain

Scale from 0-100

Higher is better

60

(2 studies)

⊕⊕⊝⊝

LOW1,2

due to risk of bias, imprecision

control group risk not available6

The mean quality of life (sf-36 questionnaire) - bodily pain in the intervention groups was

4.21 higher

(4.23 lower to 12.64 higher)

Quality of life (SF-36 questionnaire) - General health

Scale from 0-100

Higher is better

60

(2 studies)

⊕⊕⊝⊝

LOW1,2

due to risk of bias, imprecision

control group risk not available6

The mean quality of life (sf-36 questionnaire) - general health in the intervention groups was

2.49 higher

(4.99 lower to 9.97 higher)

Quality of life (SF-36 questionnaire) – Vitality

Scale from 0-100

Higher is better

298

(6 studies)

⊕⊝⊝⊝

VERY LOW1,2,5

due to risk of bias, imprecision, indirectness

control group risk not available6

The mean quality of life (sf-36 questionnaire) - vitality in the intervention groups was

1.32 higher

(1.25 lower to 3.88 higher)

Quality of life (SF-36 questionnaire) - Social functioning

Scale from 0-100

Higher is better

60

(2 studies)

⊕⊕⊝⊝

LOW1,2

due to risk of bias, imprecision

control group risk not available6

The mean quality of life (sf-36 questionnaire) - social functioning in the intervention groups was

3.31 higher

(4.29 lower to 10.92 higher)

Quality of life (SF-36 questionnaire) - Role emotional

Scale from 0-100

Higher is better

60

(3 studies)

⊕⊕⊝⊝

LOW1,2

due to risk of bias, imprecision

control group risk not available6

The mean quality of life (sf-36 questionnaire) - role emotional in the intervention groups was

0.7 higher

(4.19 lower to 5.59 higher)

Quality of life (SF-36 questionnaire) - Mental health

Scale from 0-100

Higher is better

60

(3 studies)

⊕⊕⊝⊝

LOW1,

due to risk of bias,

control group risk not available6

The mean quality of life (sf-36 questionnaire) - mental health in the intervention groups was

0.2 higher

(1.88 lower to 2.27 higher)

Apnoea Hypopnoea Index (events/hr)

Lower is better

Median follow-up 6 weeks

1256

(26 studies)

⊕⊕⊝⊝

LOW1,5

due to risk of bias, indirectness

control group risk not available6

The mean apnoea hypopnoea index (events/hr) in the intervention groups was

0.48 higher

(0.16 to 0.8 higher)

Arousals (events/hr)

136

(4 studies)

⊕⊕⊝⊝

LOW1,3,5

due to risk of bias, indirectness

control group risk not available6

The mean arousals (events/hr) in the intervention groups was

0.66 lower

(2.9 lower to 1.58 higher)

Pressure of CPAP treatment (cm H2O)

Median follow-up 6 weeks

1171

(24 studies)

⊕⊝⊝⊝

VERY LOW1,4,5

due to risk of bias, inconsistency, indirectness

control group risk not available6

The mean pressure of CPAP treatment (cm h2o) in the intervention groups was

1.49 lower

(2.12 to 0.85 lower)

Systolic blood pressure

Follow-up 12 and 16 weeks

353

(2 studies)

⊕⊕⊝⊝

LOW1,

due to risk of bias,

Mean in control group was

132.8

The mean systolic blood pressure in the intervention groups was

1.87 higher

(1.08 lower to 4.82 higher)

Diastolic blood pressure

Follow-up 12 and 16 weeks

353

(2 studies)

⊕⊝⊝⊝

VERY LOW1,2,4

due to risk of bias, inconsistency, imprecision

Mean in control group was

77.9

The mean diastolic blood pressure in the intervention groups was

4.01 higher

(1.46 lower to 9.49 higher)

24 hour mean BP

530

(2 studies)

⊕⊕⊕⊕

HIGH

Mean in control group was

92.8

The mean 24 hour mean bp in the intervention groups was

0.59 higher

(1.05 lower to 2.22 higher)

24 hour systolic BP

530

(2 studies)

⊕⊕⊕⊕

HIGH

Mean in control group was

127.1

The mean 24 hour systolic bp in the intervention groups was

0.15 lower

(2.21 lower to 1.91 higher)

24 hour diastolic BP

530

(2 studies)

⊕⊕⊕⊕⊝

HIGH

Mean in control group was

75.9

The mean 24 hour diastolic bp in the intervention groups was

0.9 higher

(0.65 lower to 2.44 higher)

Tolerability outcomes - Intolerable treatment pressure

Follow-up 4 to 36 weeks

171

(1 study)

⊕⊝⊝⊝

MODERATE2

due to imprecision

RR 0.9 (0.66 to 1.23)Moderate
513 per 1000

51 fewer per 1000

(from 174 fewer to 118 more)

Tolerability outcomes - Mask Leak

Follow-up 4 to 36 weeks

171

(1 study)

⊕⊕⊝⊝

LOW2

due to imprecision

RR 1.11

(0.74 to 1.66)

Moderate
338 per 1000

37 more per 1000

(from 88 fewer to 223 more)

Tolerability outcomes - Dry mouth

Follow-up 4 to 36 weeks

171

(1 study)

⊕⊕⊕⊝

MODERATE2

due to imprecision

RR 0.82

(0.61 to 1.1)

Moderate
563 per 1000

101 fewer per 1000

(from 220 fewer to 56 more)

Tolerability outcomes - Stuffy nose

Follow-up 4 to 36 weeks

171

(1 study)

⊕⊕⊝⊝

LOW2

due to imprecision

RR 0.98

(0.63 to 1.54)

Moderate
313 per 1000

6 fewer per 1000

(from 116 fewer to 169 more)

Patient preference (auto-CPAP/not auto-CPAP)

1082

(14 studies)

⊕⊝⊝⊝

VERY LOW1,2,4,5

due to risk of bias, inconsistency, imprecision, indirectness

RR 0.99

(0.64 to 1.56)

Moderate
475 per 1000

5 fewer per 1000

(from 171 fewer to 266 more)

MortalityNo outcome reported
1

Downgraded by 1 increment if the majority of the evidence was at high risk of bias, and downgraded by 2 increments if the majority of the evidence was at very high risk of bias

2

Downgraded by one increment if the confidence interval crossed one MID and downgraded by two increments if the confidence interval crossed both MIDs. MID for machine usage (adherence)- 1 hour ; Established MIDs for SF-36 physical/mental- 2/3 ; FOSQ- 2 ; ESS −2.5; SAQLI – 2. GRADE default MID (0.5XSD) used for all other continuous outcomes.

3

Imprecision could not be assessed as control group SD not available

4

Downgraded by 1 or 2 increments for heterogeneity. Random effect analysis used. Subgroup analysis not conducted in Cochrane review.

5

Downgraded by 1 or 2 increments because the majority of the evidence included an indirect population (downgrade by one increment) or a very indirect population (downgrade by two increments). The population was deemed to be indirect when the outcome included evidence from studies with different severity OSAHS populations or when the study did not report the AHI of the population included.

6

Cochrane review used mean difference (SE) in the analysis, control group risk data not available.

From: Positive airway pressure therapy variants for OSAHS, OHS and COPD–OSAHS overlap syndrome

Cover of Positive airway pressure therapy variants for OSAHS, OHS and COPD–OSAHS overlap syndrome
Positive airway pressure therapy variants for OSAHS, OHS and COPD–OSAHS overlap syndrome: Obstructive sleep apnoea/hypopnoea syndrome and obesity hypoventilation syndrome in over 16s: Evidence review F.
NICE Guideline, No. 202.
National Guideline Centre (UK).
Copyright © NICE 2021.

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.