Translate this page into:
Factors associated with short and mid-term compliance to continuous positive airway pressure in obstructive sleep apnea
, Massongo Massongo1, Eric Walter Pefura-Yone1, Andre Noseda3
*Corresponding author: Virginie Poka-Mayap, Department of Internal Medicine and Subspecialties, Faculty of Medicine and Biomedical Sciences, The University of Yaoundé I, Yaoundé, Cameroon. pokavirginie2013@gmail.com
-
Received: ,
Accepted: ,
How to cite this article: Poka-Mayap V, Balkissou A, Massongo M, Pefura-Yone E, Noseda A. Factors associated with short and mid-term compliance to continuous positive airway pressure in obstructive sleep apnea. J Pan Afr Thorac Soc. doi: 10.25259/JPATS_6_2026
Abstract
Objectives:
Compliance with continuous positive airway pressure (CPAP) therapy determines the outcome of the management of obstructive sleep apnea (OSA) syndrome. The aim of the study was to examine a large panel of variables measured before CPAP therapy in OSA patients, in an attempt to identify factors associated with CPAP compliance.
Material and Methods:
In this retrospective cohort study, socio-demographic, psychometric, clinical, and polysomnographic (diagnostic and CPAP titration nights) variables were collected in moderate to severe OSA patients who started on CPAP therapy at home. Compliance was assessed at the 4th and 16th months of follow-up. Logistic regression was used to identify factors associated with poor compliance, according to different thresholds (4 h and 4 h + 70% of nights use, 5 h and 6 h).
Results:
After 4 months, patients (n = 261) used their machine a median (25th–75th percentiles) of 85% (57–97) of days, while the median daily effective use was 4.4 (2.6–6.2) h. Factors associated with poor compliance were non-Caucasian ethnicity (all thresholds), no improvement in slow wave sleep (SWS) percentage during CPAP titration (compliance thresholds of 4 h and 4 h + 70% of nights’ use), and subjective poor quality of sleep (compliance thresholds of 5 h and 6 h).
Conclusion:
CPAP compliance is associated to some extent with the baseline features. Non-Caucasian patients and patients with a lack of improvement in SWS at titration or with poor sleep quality are at risk of poor compliance and could benefit from an intensive follow-up aiming to enhance compliance.
Keywords
Compliance
Continuous positive airway pressure
Obstructive sleep apnea syndrome
Polysomnographic features
INTRODUCTION
Obstructive sleep apnea (OSA) syndrome is a condition with increasing prevalence and is responsible for adverse cardiovascular, metabolic, and neurocognitive complications.[1-4] Acting as a pneumatic splint, continuous positive airway pressure (CPAP) maintains upper airway permeability, limits nocturnal hypoxemia, micro-arousals, and therefore, significantly improves sleepiness, quality of life, and, to some extent, cardiovascular mortality.[2,3] CPAP therapy is effective for the management of moderate to severe OSA, and its initiation requires a diagnostic polysomnography (PSG) and a second PSG to titrate the effective pressure needed to improve breathing in all stages of sleep and body positions.[5,6]
Compliance determines the outcome in the management of any chronic disease. Thus, the benefits of CPAP therapy are proportional to its regular and sufficient use. The threshold defining good compliance with CPAP is not uniform, due to variations according to the objective of the treatment. When assessed from reading the built-in time counter of the CPAP machine, the mean daily use is between 5 and 6 h in most European studies.[7-11]
Identifying factors associated with CPAP compliance at the initiation of therapy, using baseline features of the patient, could enable optimizing the follow-up of patients at risk of poor compliance. It has been shown that the indicators of the severity of OSA obtained during the diagnostic PSG can facilitate the prediction of long-term use of CPAP.[9,11-14] In addition, the subjective perception after the CPAP titration night, as well as the objective improvement in polysomnographic variables, may influence subsequent compliance.[15,16] However, the association between the severity of OSA and CPAP compliance is still controversial in the literature because the degree of association between these variables is generally low, indicating a limited clinical relevance.[9,14,16]
In the present study, we retrospectively studied the compliance with CPAP in a large group of patients with OSA. Our first objective was to compare short-term compliance, assessed after 4 months of home CPAP therapy, with mid-term compliance assessed 1 year later (after 16 months). The second objective was to explore, among a wide range of socio-demographic, psychometric, clinical, and polysomnographic variables, those that were associated with subsequent compliance with domiciliary CPAP treatment.
MATERIAL AND METHODS
Study design
This monocentric and retrospective cohort study was approved by the Ethics Committee of the Brugmann University Hospital Center (UHC) in Brussels. Medical records of patients with OSA diagnosed at Brugmann UHC between April 2014 and January 2016 were included in our study. To ensure homogeneity in adherence measurements, only patients treated with the same CPAP device model (iSleep 20, Breas) were included, as adherence recording algorithms and pressure-monitoring systems differ across manufacturers and may affect comparability of usage data. All patients were followed for at least 16 months, during which the CPAP was equipped with a device including a pressure monitor able to record the effective use of the machine. Each patient was told to use CPAP every night, as many hours as possible, and was informed of the existence of a device able to evaluate the pattern of use at home.
Ethical considerations
This retrospective study was conducted in accordance with the principles of the Declaration of Helsinki. Authorization to access and use anonymized patient medical records was obtained from the Brugmann University Hospital. As the study was based exclusively on fully anonymized retrospective data, formal Institutional Review Board approval and informed consent were not required according to institutional regulations.
Participants
During the study period, patients with an apnea-hypopnea index (AHI) > 20/h and a micro-arousal index (MAI) >30/h at diagnostic PSG were invited to receive a PPC test with titration of the effective pressure according to the recommendations of the National Institute of Health and Disability Insurance in Belgium. Patients who improved their breathing parameters and sleep quality were eligible for CPAP therapy. The CPAP machines were installed at home, and in patients whose compliance report showed a mean use at 4 months >3 h/h, there was an extension of the device availability for an additional 12 months.
Data collection
Demographic data, including age, sex, body mass index (BMI), and cervical circumference, were collected in the file of each subject. The following main comorbidities were also extracted: Stroke, heart failure, hypertension, diabetes mellitus, and gastroesophageal reflux disease (GERD).
During the study period, all patients completed a questionnaire on lifestyle, drinking habits, and sleep. Self-questionnaires of psychometric scales with classical semantic instructions for fatigue and sleepiness were administered to all participants on the first day of their stay in our sleep unit, before their first night of polysomnographic recording.
The Fatigue Severity Scale (FSS) is a self-reporting tool used to assess symptomatic intensity levels of 208 fatigue and its effect on daily functioning.[17] The FSS is a nine-item 7-point Likert-type scale. Scores are usually reported as “mean scores” (ranging from 1 to 7) obtained by dividing the total score (ranging from 7 to 63) by 9. For clinically significant relevant excessive daytime fatigue, a cut-off>5 on mean scores is often proposed.
The Epworth Sleepiness Scale (ESS) is the most widely used scale of subjective sleepiness and daytime sleep propensity. The ESS consists of eight items (described situations) arranged on a 4-point Likert scale ranging from 0 (“never doze”) to 3 (“high chance of dozing” during daytime). The summed score ranges from 0 to 24, and a score above 10 is commonly interpreted as clinically relevant for increased daytime sleepiness.[18]
The Pittsburgh Sleep Quality Index (PSQI) was used to assess the subjective sleep quality. The 19 items are grouped 226 into seven component scores, each weighed equally on a scale from 0 to 3. These components are subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbance, use of sleeping medication, and daytime dysfunction. The component scores are then summed to yield the global PSQI score. In validation studies, a global PSQI score >5 indicates that a subject is having severe difficulties in at least two areas or moderate difficulty in more than three areas.[19]
The Hospital Anxiety and Depression rating scale (HADRS) is a self-report rating scale of 14 items on a 4-point Likert scale (with a range from 0 to 3). It is the most commonly used psychometric tool in general internal medicine, designed to measure the intensity of anxiety and depression symptoms (seven items for each subscale, anxiety [HAD-A] and depression [HAD-D]) with respective scores ranging from 0 to 21]. The reliability and validity of the HADRS have been tested in a vast number of 246 studies. Scores ≥11 on each subscale have previously been considered as clinically relevant.[20]
Sleep study
A baseline PSG was performed for the diagnosis of OSA. Recordings were made by standard sleep recording leads, including a bilateral electro-oculogram, a 3-lead electroencephalogram (Fp2-A1, C4-A1, O2-A1), and a submental and bilateral anterior tibial electromyogram. Oral and nasal airflow were recorded by an oro-nasal cannula. Snoring was recorded through a microphone and respiratory efforts by thoracic and abdominal belts. Capillary oxygen saturation was monitored by photosensitive finger-oximetry. All PSG recordings were read by trained technicians unaware of the aims of the study, according to the American Association of Sleep Medicine (AASM) criteria.[21] All patients eligible for the CPAP therapy slept a second night with an auto-piloted CPAP for effective pressure titration.
PSG
Sleep staging was performed according to the criteria of Rechtschaffen and Kales.[21] Sleep onset latency (SOL) was defined as the time from lights-off to the first 30 s epoch of sleep; the sleep period time (SPT) was the time interval from sleep onset to final awakening. Total sleep time (TST) was defined as SPT minus the total duration of cumulated intra-sleep awakenings (wake time after sleep onset). The internal sleep efficiency index (SEI) expressed in percent was defined as the ratio between the TST and the total registration time. Non-rapid eye movement sleep included sleep stages N1, N2, and N3 (or slow wave sleep, SWS). An episode of apnea was defined as a reduction ≥90% in airflow for ≥10 s during sleep. A sleep hypopnea was defined as a reduction of ≥30% airflow amplitude accompanied by either a 3% or greater reduction in oxygen saturation or a micro-arousal. The oxygen desaturation index was defined as the number of desaturations ≥3% per hour of sleep. Micro-arousals were defined according to the AASM criteria, and the MAI represented the number of micro-arousals divided by TST.
Assessment of compliance
Readings from the pressure monitor at 4 and 16 months after initiation of the fixed CPAP therapy at home were retained for analysis. Each CPAP was equipped with a microprocessor, an analog-to-digital converter, a battery backup, a real-time clock, and a power-on reset circuit. The variable component of the pressure signal given by the pressure transducer is analyzed to determine if the patient is breathing in the mask. Up to 100 consecutive processing sessions can be recorded (a session being defined as the time between power on and power off). The pressure monitor yielded the following variables. First, the mean effective use was the total use time recorded by the monitor, divided by the number of days elapsed between the two readings. The percentage of the monitored days during which the machine had been effectively used was also available, as well as the mean effective use per effective day, defined as the cumulative time of effective use divided by the number of days the machine had been actually used. The mean delivered pressure was the arithmetic mean of the pressures delivered on all sessions, weighted by the duration of each session. The residual apnea and hypopnea index was the sum of all apneas and hypopneas identified by the monitor, divided by the cumulative time of effective use. For each subject, a printed report was obtained showing the pattern of CPAP use at home, including the time of starting with CPAP treatment, the number of episodes when the pressure was no longer applied during the night, the time of cessation of CPAP, and the presence or absence of a nap under CPAP. Particular care had been taken to change the clocks of the monitors on the dates of shift from winter to summer. Periods during which subjects made journeys involving clock changes were excluded from the analysis.
Statistical analysis
Data were analyzed using IBM Statistical Package for the Social Sciences (SPSS) Version 20 for Windows (SPSS Inc., Chicago, IL). Qualitative data were presented as counts and proportions, and quantitative variables as mean and standard deviation (SD) or median and 25th–75th percentiles. Chi-square test and Fisher’s exact test were used to compare proportions. Quantitative variables were compared by Student’s t-test or a non-parametric equivalent. The Spearman correlation test was used to study the relationship between the compliance variables (at the 4th and 16th months of follow-up) and continuous or ordinal variables. Associations involving categorical variables were assessed using appropriate statistical tests depending on the nature of the variables. Variables showing significant correlations and exhibiting skewed distributions were log-transformed (base 10) before inclusion in the linear regression analysis to improve normality and stabilize variance. The assumptions of linear regression were assessed before model fitting. Thresholds of good compliance being variable according to the objective, we tested different thresholds found in the literature at 4th and 16th months of follow-up to define the compliance: 4 h per night combined with 70% of nights of use, 4 h, 5 h, and 6 h. Logistic regression was used to investigate factors associated with different compliance thresholds. Variables included in the univariate analysis were selected based on their clinical relevance and previous literature on determinants of CPAP adherence. Variables with p < 0.10 in univariate analysis were subsequently entered into the multivariate models. A difference was considered significant for a value of p < 0.05.
RESULTS
During the study period, 458 patients with moderate to severe OSA were treated at home with CPAP, of whom 333 patients received an iSleep 20 (Breas) type device. Forty-eight patients were excluded due to missing data in the medical files. Seventeen and 7 patients had incomplete compliance reports at 4 and 16 months of follow-up, respectively. The proportion of missing compliance data was low (8.4%). A total of 261 patients were definitively included for data analysis [Figure 1]. An attrition analysis was performed comparing patients included in the final analysis (n = 261) with those with incomplete compliance data (n = 24). No significant differences were observed in baseline characteristics between the two groups (all p > 0.05)
Study population
Of the 261 patients definitively included in our analysis, 193 (73.9%) were men. The mean age (±SD) was 52 (±13) years. Mean values (±SD) of BMI and cervical circumference were 31.7 kg/m2 and 42.3 cm, respectively. As shown in Table 1, the main comorbidities in these patients were hypertension, GERD, and diabetes mellitus, respectively, in 44.8, 29.9, and 23.8% of participants. Women had a higher BMI and had more frequently a history of stroke than men. The median values of HAD-A, HAD-D, ESS, PSQI, and FSS were 7, 6, 10, 7, and 4.1, respectively.
| Characteristics | Overall n=261 | Men n=193 | Women n=68 | p-value |
|---|---|---|---|---|
| Mean age, years±SD | 52±13 | 51±13 | 57±14 | 0.36 |
| Mean BMI, Kg/m2±SD | 31.7±5.7 | 31.1±5.2 | 33.5±6.6 | 0.015 |
| Mean cervical circumference, cm±SD | 42.3±4.1 | 43.3±3.8 | 39.4±3.5 | 0.503 |
| Median HAD-A (IQR) | 7 (5–10) | 8±4 | 8±4 | 0.94 |
| Median HAD-D (IQR) | 6 (4–9) | 6±4 | 6±4 | 0.75 |
| Median ESS (IQR) | 10 (7–14) | 11±5 | 9±5 | 0.23 |
| Median PSQI (IQR) | 7 (5–10) | 7±3 | 7±3 | 0.87 |
| Median FSS (IQR) | 4.1 (2.4–5.5) | 4.9±6.9 | 3.9±7.8 | 0.96 |
| Hypertension, n (%) | 117 (44.8) | 78 (66.7) | 39 (33.3) | 0.023 |
| Diabetes mellitus, n (%) | 62 (23.8) | 47 (75.8) | 15 (24.2) | 0.74 |
| Heart failure, n (%) | 17 (6.5) | 12 (70.6) | 5 (29.4) | 0.77 |
| Heart rhythm disorders, n (%) | 20 (7.7) | 12 (60) | 8 (40) | 0.18 |
| Coronary syndrome, n (%) | 16 (6.1) | 13 (81.2) | 3 (18.8) | 0.77 |
| Stroke, n (%) | 14 (5.4) | 5 (35.7) | 9 (64.3) | 0.002 |
| GERD, n (%) | 78 (29.9) | 56 (71.8) | 22 (28.2) | 0.64 |
| Depression, n (%) | 51 (19.5) | 36 (70.6) | 15 (29.4) | 0.59 |
SD: Standard deviation, IQR: Interquartile range, BMI: Body mass index, HAD-A: Anxiety score, HAD-D: Depression score, PSQI: Pittsburgh Sleep Quality Index, ESS: Epworth sleepiness scale, FSS: Fatigue severity scale, GERD: Gastroesophageal reflux disease. Statistically significant values (p < 0.05) have been highlighted in bold.
Table 2 summarizes the polysomnographic data of the diagnostic night and the titration night of CPAP. There is a significant improvement with CPAP in almost all variables except for SOL, intra-sleep awakenings, and sleep efficiency.
| Parameters | Night 1 | Night 2 | p-value |
|---|---|---|---|
| TST (min) | 378 (325–421) | 364 (316.5–409) | 0.028 |
| SOL (min) | 24 (12–40) | 23.5 (11–39) | 0.38 |
| WASO (min) | 58 (35–96.8) | 57 (31–93) | 0.13 |
| SEI (%) | 86.3 (78.2–91.6) | 86.6 (77.9–92.2) | 0.67 |
| N1 (%) | 16 (10–26) | 11 (8–18) | <0.001 |
| N2 (%) | 52 (43–60) | 48 (38–56) | <0.001 |
| N3 (%) | 16 (9–23) | 22 (14–30) | <0.001 |
| REM (%) | 12 (7–16) | 15 (11–20) | <0.001 |
| CAI (h−1) | 0.31 (0–1.15) | 0,44 (0–1.56) | <0.001 |
| OAI (h−1) | 9.3 (3.35–18.25) | 0.43 (0–1.34) | <0.001 |
| HI (h−1) | 21 (16.46–29.83) | 3.18 (1.33–6.6) | <0.001 |
| AHI (h−1) | 34.6 (25.9–52.2) | 5.4 (2.5–10.1) | <0.001 |
| MAI (h−1) | 41.8 (34.1–51.5) | 16.5 (11.4–22.8) | <0.001 |
| Mean SaO2 (%) | 92 (90–94) | 94 (93–95) | <0.001 |
| Minimal SaO2 (%) | 83 (75–88) | 90 (87–92) | <0.001 |
| ODI (h−1) | 23.9 (14.8–44.2) | 3.9 (1.5–9) | <0.001 |
| Snoring (%) | 28 (11–46) | 4 (1–11) | <0.001 |
| PLM index (%) | 10.8 (1.7–27.8) | 3.4 (0.7–14.4) | <0.001 |
All variables are presented as median (25th–75th percentile). TST: Total sleep time, WASO: Wake time after sleep onset, REM: Rapid eye movement sleep, SEI: Sleep efficiency index ([TST/SPT]×100), AHI: Apnea-hypopnea index, SaO2: Arterial oxygen saturation, ODI: Oxygen desaturation index, MAI: Micro-arousal index, CAI: Central apnea index, OAI: Obstructive apnea index, HI: Hypopnea index, PLM: Periodic limbs movement, SOL: Sleep onset latency. Statistically significant values (p < 0.05) have been highlighted in bold.
Compliance
Different compliance variables during CPAP therapy are presented in terms of the median in Figure 2. At the 4th month of follow-up (M4), patients used their machine a median (25th–75th percentiles) of 85% (57–97) of days, and the compliance increased significantly at the 16th month of follow-up (M16). The pressure was applied in the mask during 4.4 (2.56–6.2) h per 24 h with a median effective use per effective day of 5.6 (4.25–6.73) h per 24h at M4. At M16, the latter variable increased significantly to 6.23 (5.22–7.3) h/24 h. As expected, the higher the compliance threshold, the more the prevalence of poor compliance increased [Table 3]. Furthermore, the proportion of subjects with poor compliance decreased significantly from M4 to M16 except when the compliance threshold was up to 6h.
| Compliance thresholds | 4th-month follow-up (%) | 16th-month follow-up (%) | p-value |
|---|---|---|---|
| 3 h | 30.3 | 14.9 | 0.01 |
| 4 h | 43.3 | 24.9 | 0.005 |
| 4 h+70% of nights use | 46.7 | 28.7 | 0.01 |
| 5 h | 58.6 | 43.6 | 0.032 |
| 6 h | 72.8 | 61.3 | 0.15 |
P< 0.05 was considered statistically significant, highlighted in bold.
Relationship between baseline variables measured before CPAP therapy initiation and compliance variables
The results of the correlation and linear regression analysis are shown in Table 4. The mean effective use at M4 was positively correlated with depression score (r = 0.153), central apnea number (r = 0.18), and improvement in SWS percentage during the CPAP titration night (r = 0.146).The mean use over the entire period at M4 was positively correlated with the number of central apnea (r = 0.174) and was the only association having a linear relationship.
| Variables | Correlation | Linear regression | |||
|---|---|---|---|---|---|
| r | p-value | Beta | t | p-value | |
| Percent of nights used during the first 4 months | |||||
| Age | 0.146 | 0.019 | −0.001 (−0.002–0) | −1.84 | 0.067 |
| CA | 0.18 | 0.004 | −0.024 (−0.048–0.001) | −1.91 | 0.058 |
| CA/CA+CO | 0.151 | 0.015 | −0.012 (−0.034–0.009) | −1.12 | 0.261 |
| Mean use on days of effective use at 4th month follow-up | |||||
| Cervical circumference | −0.123 | 0.048 | −0.007 (−0.015–0.001) | −1.79 | 0.074 |
| HAD-D | 0.153 | 0.024 | 0.086 (−0.008–0.179) | 1.81 | 0.072 |
| Δ N3 | 0.146 | 0.019 | −0.472 (−1.34–039) | −1.07 | 0.286 |
| CA | 0.141 | 0.023 | 0.072 (−0.003–0.146) | 1.89 | 0.06 |
| Mean use over the first 4-month period of follow-up | |||||
| CA | 0.174 | 0.005 | −0.08 (0.27–1.52) | 2.84 | 0.002 |
CA: Number of central apneas, CO: Number of obstructive apneas, Δ N3: Percent N3 night 2–Percent N3 night 1, HAD-D: Depression score. Statistically significant value (p < 0.05) has been highlighted in bold.
Factors associated with poor compliance
Factors associated with poor compliance according to the compliance threshold are presented in Table 5. The Non-Caucasian ethnicity is an independent factor associated with poor compliance regardless of the compliance threshold and the month of follow-up.
| Characteristics | 4th-month follow-up | 16th-month follow-up | ||||||
|---|---|---|---|---|---|---|---|---|
| Univariate | Multivariate | Univariate | Multivariate | |||||
| OR (CI 95%) | p-value | OR (CI 95%) | p-value | OR (CI 95%) | p-value | OR (CI 95%) | p-value | |
| 4 h+70% of nights use | ||||||||
| Non-Caucasian | 3.25 (1.88–6.62) | <0.001 | 3.42 (1.96–5.96) | <0.001 | 3.34 (1.68–6.63) | 0.001 | 3.45 (1.28–7.75) | 0.012 |
| Δ N3 <0 | 1.72 (1.01–2.94) | 0.046 | 1.93 (1.1–3.37) | 0.021 | - | - | - | - |
| PSQI >5 | - | 2.68 (0.94–7.65) | 0.065 | 2.8 (0.96–8.19) | 0.06 | |||
| 4 h | ||||||||
| Non-Caucasian | 2.98 (1.74–5.11) | <0.001 | 3.48 (1.82–6.68) | <0.001 | 3.24 (1.59–6.58) | 0.001 | - | - |
| ESS >10 | 1.64 (0.95–2.84) | 0.076 | 1.71 (0.95–3.07) | 0.073 | - | - | - | - |
| Δ N3 <0 | 1.97 (1.15–3.36) | 0.014 | 2.94 (1.57–5.48) | 0.001 | - | - | - | - |
| Interface | 1.86 (0.94–3.7) | 0.076 | 2.05 (0.86–4.89) | 0.105 | - | - | - | - |
| 5 h | ||||||||
| Non-Caucasian | 3.69 (2.02–6.72) | <0.001 | 2.67 (1.21–5.89) | 0.015 | 3.71 (1.89–7.79) | <0.001 | - | - |
| Depression | 1.79 (0.96–3.32) | 0.064 | 1.69 (0.76–3.74) | 0.192 | - | - | - | - |
| PSQI >5 | 2.1 (1.04–4.22) | 0.037 | 2.23 (1.10–4.50) | 0.025 | - | - | - | - |
| 6 h | ||||||||
| Non-Caucasian | 3.81 (1-83–7.89) | <0.001 | 4.47 (1.44–13.86) | 0.009 | 3.79 (1.75–8.2) | 0.001 | 5.53 (1.84–16.58) | 0.002 |
| Heart failure | - | - | - | - | 6.08 (0.75–49.14) | 0.09 | 4.56 (0.35–58.11) | 0.242 |
| Depression | 2.49 (1.31–4.72) | 0.005 | 1.81 (0.7–4.9) | 0.219 | 2.91 (11.4–6.11) | 0.004 | 3.78 (1.4–10.18) | 0.008 |
| FSS >5 | 1.89 (0.97–3.69) | 0.061 | 0.43 (0.18–1.05) | 0.065 | - | - | ||
| PSQI >5 | 2.1 (1.04–4.22) | 0.037 | 2.51 (1.12–5.61) | 0.025 | 2.08 (0.95–4.54) | 0.065 | 2.94 (1.18–7.33) | 0.021 |
| Interface | - | - | - | - | 0.44 (0.19–103) | 0.06 | 3.11 (0.93–10.34) | 0.065 |
OR: Odds ratio, 95%CI: 95% Confidence interval, Δ N3: Percent N3 night 2–Percent N3 night 1, PSQI: Pittsburgh Sleep Quality Index, ESS: Epworth Sleepiness Scale, FSS: Fatigue severity scale. Statistically significant values (p < 0.05) have been highlighted in bold.
At M4, the independent factors associated with poor compliance (odds ratio [95% confidence interval]) were the absence of improvement in SWS percentage during the CPAP titration night for a compliance threshold of 4 h (1.93 [1.1– 3.37]) as well as for a compliance threshold of 4 h combined with >70% of nights use (2.94 [1.57–5.48]) and the poor quality of sleep for a compliance threshold of 5 h (2.23 [1.10–4.50]) as well as for a compliance threshold of 6 h (2.51 [1.12–5.61]). At M16, patients with a history of depression are three times more likely to be non-observant when the compliance threshold is 5 h. The poor quality of sleep (2.94 [1.18–7.33]) is independently associated with poor compliance when the compliance threshold is 6 h.
DISCUSSION
The present retrospective study aimed to characterize compliance with CPAP therapy in patients with moderately severe to severe OSA. We found that compliance with CPAP is relatively good at the 4th month of follow-up and that all compliance variables improve 1 year later. A weak but statistically significant linear association was observed between compliance and the number of central apneas. Some factors are associated with poor compliance as the absence of improvement in SWS on the CPAP titration night (compared with the baseline night), and also with a poor quality of sleep. However, the results depended on the time of assessment (4 months or 16 months) and on the threshold chosen to define low compliance. The only factor associated with poor compliance, regardless of the compliance threshold, was non-Caucasian ethnicity.
In this study, involving compliance with CPAP, some methodological factors have to be discussed. First, the effective use of CPAP was obtained from the time during which the pressure necessary to control apnea was effectively applied in the mask. It has been shown that the use of a pressure monitor, compared with a simple time counter, avoids some overestimation due to periods during which the machine is running, while the mask is off.[9,22] Moreover,the use of a pressure monitor allows us to describe the mean effective use per day in terms of a product of two components, namely the percentage of days CPAP is effectively used and the mean effective use per effective day.[9] Second, we tested several compliance thresholds to define good compliance, namely >4 h, >4 h per night combined with >70% of nights of use, as recommended by ATS,[23] >5 h, and >6 h. This strategy of considering several thresholds is justified by the fact that several studies have shown that the compliance target may vary according to the objective. For example, using the machine >4 h is sufficient to control daytime sleepiness, using >5–6 h is required to control blood pressure in non-sleepy hypertensive apneics[24] and using >7 h is required to restore full neuropsychological performance.[25]
In the present study, the compliance was found to be reasonably good. Patients use the machine during the first 4 months, 85% (57–97)of days with a median effective use per effective day of 5.6 (4.25–6.73) h. These data compare favorably with the results of previous studies.[9,22,26] The median percentage of days used and the median effective use per effective day increased to 91% (P = 0.026) and 6.2 h (p < 0.001) at 16 months of domiciliary treatment. To the best of our knowledge, such an early improvement in compliance over time has not been reported previously. In a prospective study, Sucena et al. assessed compliance with CPAP over time in long-term (at least 5 years, up to 10 years) patients and reported that compliance did not change during the first 2 years but significantly and gradually improved from the 3rd year.[10]
Several studies attempt to determine predictive factors for subsequent compliance with CPAP. A positive correlation between compliance and the severity of OSA, reflected by the AHI[14,26-28] or the fragmentation of sleep, assessed by MAI[11] was found in some studies. However, the correlation coefficients were generally weak, with poor clinical relevance. In line with these findings, the correlations observed in our study were also modest (r ≈ 0.15–0.18) and should therefore be interpreted cautiously, as CPAP adherence is likely influenced by multiple interacting factors rather than a single determinant. Similarly, a correlation was reported with the degree of improvement in AHI between the baseline night and the CPAP titration night.[27] On the other hand, many authors found no correlation at all between the baseline features of the subjects and their subsequent compliance with CPAP.[22,26,29,30] In the present study, there was a linear relation between compliance at 4 months of CPAP therapy and the number of central apneas. Central apneas lead to dyssomnia, even insomnia, hence their potential link with low compliance.[29] Our main finding was that poor compliance was associated with non-Caucasian ethnicity as well, after 4 months and 1 year later, and for all the compliance thresholds selected. Our non-Caucasian patients presented with language and educational barriers, and poor compliance has been reported by other authors in groups of OSA patients with low educational levels and cultural barriers.[31,32] This association should be interpreted with caution, as unmeasured socioeconomic factors, access to care, cultural influences, or educational level may partly explain the observed relationship rather than ethnicity itself. In addition, we found that the poor compliance after 4 months was associated with a lack of improvement in SWS during the CPAP titration night, at least for the 4 h and 4 h combined with 70% of days used thresholds. These potential predictive factors have not been reported previously, except for a modest increase of 7% of the probability of persistent use of CPAP associated with a 10% increase in SWS during titration.[16] Similarly, we found that poor quality of sleep, as reflected by a PSQI >5, was associated with poor compliance for higher thresholds, namely 5 h (after 4 months of CPAP) and 6 h (both after 4 months and 16 months of CPAP). Little attention has been paid in the past to a possible link between insomnia and compliance with CPAP in OSA patients, except for a study on 73 subjects by Pieh et al., who reported that insomnia assessed through the Regensburg Insomnia Scale was linked to poor compliance.[29] Finally, the relevance of our finding that poor compliance was more frequent at 16 months in depressive patients seems limited, as it was significant only for the 6 h threshold.
The main limitation of the present retrospective study is the existence of missing data, as expected by collecting data in the medical records of patients. The restriction to a single CPAP device model may, to some extent, limit our ability to generalize our results; on the other hand, this approach minimizes measurement variability related to device-specific adherence recording algorithms. Because adherence data extraction differed across CPAP devices, only patients treated with a single device model were included. Clinical data were not systematically available for patients using other devices, preventing comparison between included and excluded subjects. Device allocation was mainly driven by logistical availability, which likely reduced the risk of selection bias; however, residual selection bias cannot be completely excluded. Because several compliance thresholds were evaluated at two follow-up time points, multiple statistical tests were performed, which may increase the risk of type I error. However, the use of different thresholds was intended to explore the robustness of the observed associations across commonly used definitions of CPAP adherence. On the other hand, our study also has strengths. First, we studied a larger number of subjects than many previous studies involving CPAP compliance.[9,12,22,28,30] Second, all subjects were equipped with the same machine, including a pressure monitor able to assess the effective “mask on pressure.” Third, we did not restrict our analysis only to the correlation between the baseline and compliance variables. We also determine factors associated with poor compliance not only at a single threshold but also at several thresholds, potentially reflecting several targets of CPAP therapy.
CONCLUSION
This study, performed in a large group of 261 OSA patients, shows that compliance measured after 16 months of CPAP therapy at home significantly increases in comparison with that assessed early, after 4 months. We have found that poor compliance is associated with non-Caucasian ethnicity, probably as a consequence of low educational level, language, and cultural barriers. In addition, an absence of improvement in deep sleep at the titration night and a poor quality of sleep were also associated with poor compliance. The correlations between baseline features of patients and compliance with CPAP are not strong enough to allow a robust prediction of subsequent compliance. Thus, it remains mandatory to accurately assess compliance in all OSA patients with domiciliary CPAP. The associated factors we found, however, allow us to define target groups of patients, to whom an adapted intensive follow-up should be proposed to enhance compliance with CPAP.
Acknowledgments:
The authors thank the medical and technical staff of the Sleep Unit at Brugmann University Hospital for their support in data collection and patient management.
Ethical approval:
This retrospective study was conducted in accordance with the principles of the Declaration of Helsinki. Authorization to access and use anonymized medical records was first obtained from the Ethics Committee of the Brugmann University Hospital.
Declaration of patient consent:
Patient’s consent is not required as there are no patients in this study.
Conflicts of interest:
There are no conflicts of interest
Use of artificial intelligence (AI)-assisted technology for manuscript preparation:
The authors confirm that there was no use of artificial intelligence (AI)-assisted technology for assisting in the writing or editing of the manuscript and no images were manipulated using AI.
Financial support and sponsorship: Nil.
References
- The epidemiology of adult obstructive sleep apnea. Proc Am Thorac Soc. 2008;5:136-43.
- [CrossRef] [PubMed] [Google Scholar]
- Stroke and other cardiovascular events in patients with obstructive sleep apnea and the effect of continuous positive airway pressure. J Neurol. 2017;264:1247-53.
- [CrossRef] [PubMed] [Google Scholar]
- Sleep. 4: Sleepiness, cognitive function, and quality of life in obstructive sleep apnoea/hypopnoea syndrome. Thorax. 2004;59:618-22.
- [CrossRef] [PubMed] [Google Scholar]
- The association between obstructive sleep apnea and metabolic syndrome: A systematic review and meta-analysis. BMC Pulm Med. 2015;15:105.
- [CrossRef] [PubMed] [Google Scholar]
- Home treatment of obstructive sleep apnoea with continuous positive airway pressure applied through a nose-mask. Bull Eur Physiopathol Respir. 1984;20:49-54.
- [Google Scholar]
- Clinical guidelines for the manual titration of positive airway pressure in patients with obstructive sleep apnea. J Clin Sleep Med. 2008;4:157-71.
- [CrossRef] [PubMed] [Google Scholar]
- Effective compliance during the first 3 months of continuous positive airway pressure. A European prospective study of 121 patients. Am J Respir Crit Care Med. 1999;160:1124-9.
- [CrossRef] [PubMed] [Google Scholar]
- Predictors of long-term compliance with continuous positive airway pressure. Thorax. 2010;65:829-32.
- [CrossRef] [PubMed] [Google Scholar]
- Compliance with nasal continuous positive airway pressure assessed with a pressure monitor: Pattern of use and influence of sleep habits. Respir Med. 2000;94:76-81.
- [CrossRef] [PubMed] [Google Scholar]
- Continuous positive airway pressure treatment for sleep apnoea: Compliance increases with time in continuing users. Eur Respir J. 2006;27:761-6.
- [CrossRef] [PubMed] [Google Scholar]
- Acceptance and long-term compliance with nCPAP in patients with obstructive sleep apnoea syndrome. Eur Respir J. 1996;9:939-44.
- [CrossRef] [PubMed] [Google Scholar]
- Determinants for adherence to continuous positive airway pressure therapy in obstructive sleep apnea. PLoS One. 2017;12:e0189614.
- [CrossRef] [PubMed] [Google Scholar]
- Long-term use of CPAP therapy for sleep apnea/hypopnea syndrome. Am J Respir Crit Care Med. 1999;159:1108-14.
- [CrossRef] [PubMed] [Google Scholar]
- Long-term compliance with CPAP therapy in obstructive sleep apnea patients and in snorers. Sleep. 1996;19(9 Suppl):S136-43.
- [CrossRef] [PubMed] [Google Scholar]
- A brief survey of patients' first impression after CPAP titration predicts future CPAP adherence: A pilot study. J Clin Sleep Med. 2013;9:199-205.
- [CrossRef] [PubMed] [Google Scholar]
- Polysomnographic predictors of persistent continuous positive airway pressure adherence in patients with moderate and severe obstructive sleep apnea. Kaohsiung J Med Sci. 2015;31:83-9.
- [CrossRef] [PubMed] [Google Scholar]
- The fatigue severity scale. Application to patients with multiple sclerosis and systemic lupus erythematosus. Arch Neurol. 1989;46:1121-3.
- [CrossRef] [PubMed] [Google Scholar]
- A new method for measuring daytime sleepiness: The Epworth sleepiness scale. Sleep. 1991;14:540-5.
- [CrossRef] [PubMed] [Google Scholar]
- The Pittsburgh sleep quality index: A new instrument for psychiatric practice and research. Psychiatry Res. 1989;28:193-213.
- [CrossRef] [PubMed] [Google Scholar]
- The hospital anxiety and depression scale. Acta Psychiatr Scand. 1983;67:361-70.
- [CrossRef] [PubMed] [Google Scholar]
- The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications United States: American Academy of Sleep Medicine; 2007.
- [Google Scholar]
- Compliance with CPAP therapy in patients with the sleep apnoea/hypopnoea syndrome. Thorax. 1994;49:263-6.
- [CrossRef] [PubMed] [Google Scholar]
- An official American Thoracic Society statement: Continuous positive airway pressure adherence tracking systems. The optimal monitoring strategies and outcome measures in adults. Am J Respir Crit Care Med. 2013;188:613-20.
- [CrossRef] [PubMed] [Google Scholar]
- Long-term effect of continuous positive airway pressure in hypertensive patients with sleep apnea. Am J Respir Crit Care Med. 2010;181:718-26.
- [CrossRef] [PubMed] [Google Scholar]
- Relationship between hours of CPAP use and achieving normal levels of sleepiness and daily functioning. Sleep. 2007;30:711-9.
- [CrossRef] [PubMed] [Google Scholar]
- Nasal CPAP: An objective evaluation of patient compliance. Am J Respir Crit Care Med. 1994;149:149-54.
- [CrossRef] [PubMed] [Google Scholar]
- Improving compliance to CPAP in sleep apnea syndrome: From coaching to telemedicine. Rev Mal Respir. 2012;29:7-10.
- [CrossRef] [PubMed] [Google Scholar]
- Adherence to continuous positive airway pressure in adults with an intellectual disability. Sleep Med. 2017;34:234-9.
- [CrossRef] [PubMed] [Google Scholar]
- Insomnia symptoms influence CPAP compliance. Sleep Breath. 2013;17:99-104.
- [CrossRef] [PubMed] [Google Scholar]
- Nasal CPAP therapy in obstructive sleep apnea syndrome: Patient compliance. Schweiz Med Wochenschr. 2000;130:1291-7.
- [Google Scholar]
- Race and residential socioeconomics as predictors of CPAP adherence. Sleep. 2011;34:1653-8.
- [CrossRef] [PubMed] [Google Scholar]
- Predictors of positive airway pressure therapy adherence in children: A prospective study. J Clin Sleep Med. 2012;8:279-86.
- [CrossRef] [PubMed] [Google Scholar]