Effects of telemonitoring follow-up, side effects, and other factors on CPAP adherence

Abstract

Study Objectives:

This study aimed to investigate the effect of telemonitoring compared with standard clinic visits on adherence to continuous positive airway pressure (CPAP) treatment after 6 months. In addition, the impact of other factors including CPAP side effects on treatment adherence were assessed.

Methods:

Consecutive patients (n = 217) who were prescribed CPAP treatment for obstructive sleep apnea were randomized to either telemonitoring or standard-care follow-up. All patients were followed up 6 months after treatment started. Clinical/anthropometric variables, socioeconomical and lifestyle factors, psychological distress, daily function, and personality traits along with CPAP side effects were assessed. Differences between groups were analyzed using 2-sample t-test, chi-square test, or Fisher’s exact test. Regression modeling was used to explore associations between dependent and independent variables.

Results:

There were no differences in CPAP adherence between telemonitoring and standard-care groups after 6 months (53.2% vs 48.7%; P = .54). CPAP side effects such as dry throat (odds ratio = 2.17; 95% confidence interval = 1.25–3.70), increased awakenings (2.50; 1.31–4.76), and exhaling problems (3.70; 1.25–10.1) were independently associated with low CPAP adherence, although these associations were weakened when adding smoking to the model. No other baseline or follow-up factors were associated with CPAP adherence at 6 months.

Conclusions:

We could not show that telemonitoring follow-up improved adherence levels. Dry throat, increased awakenings, exhaling problems, and smoking had negative effects on CPAP adherence. Preventing side effects and assessing smoking status is therefore of importance when wanting to improve CPAP adherence.

Clinical Trial Registration:

Registry: ClinicalTrials.gov; Name: Benefits of Telemedicine in CPAP Treatment; URL: https://clinicaltrials.gov/ct2/show/NCT03202602; Identifier: NCT03202602.

Citation:

Delijaj F, Lindberg E, Johnsson L, Kristiansson P, Tegelmo T, Theorell-Haglöw J. Effects of telemonitoring follow-up, side effects, and other factors on CPAP adherence. J Clin Sleep Med. 2023;19(10):1785–1795.

BRIEF SUMMARY

Current Knowledge/Study Rationale: The effect of telemonitoring on improving continuous positive airway pressure (CPAP) adherence in patients with obstructive sleep apnea has been mixed. The aim of this study was to assess the efficacy of telemonitoring in CPAP adherence in a large, nonselected patient cohort. We further aimed to explore what other factors, including side effects of CPAP treatment, affect CPAP adherence.

Study Impact: The findings of this study show that telemonitoring does not improve CPAP adherence, and that smoking and side effects of CPAP such as dry throat, increased awakenings, and exhaling problems have negative effects on CPAP adherence. Findings suggest that preventing side effects and assessing smoking status is of importance when wanting to improve CPAP adherence.

INTRODUCTION

Obstructive sleep apnea (OSA) is the most common sleep-related breathing disorder and is caused by total or partial obstruction of the upper airway, with subsequent desaturations and arousals that disturb sleep.¹ Untreated moderate to severe OSA is a risk factor for cardiovascular disease, traffic and occupational accidents, and premature death.^1–4 The most effective treatment for severe and moderate OSA is continuous positive airway pressure (CPAP).⁵ Adherence to treatment is crucial for treatment effect but is sometimes difficult to achieve, and several previous studies have indicated low CPAP adherence.^6,7 Several factors have an impact on CPAP adherence, including mask displacement, pressure intolerance, claustrophobia, and machine noise.^6,8 Socioeconomic conditions, including civil status (single), (low) educational level,⁹ and cigarette smoking,¹⁰ also have a negative effect on CPAP usage. In addition, anatomic characteristics of the upper airway, as well as social-cognitive factors may predict CPAP use.^11–13

Adherence during the first week of CPAP treatment is associated with long-term adherence,^14,15 and early troubleshooting and managing may therefore be of importance for improved adherence. A telemonitoring (TM) approach provides options for early follow-up. Different approaches to TM have been studied and have shown TM to be less expensive than standard care (SC)^16,17 and that TM saves nursing time.¹⁸ Nevertheless, there is an ambiguity regarding the effect of TM on improving CPAP adherence,^{15,17,19–25} and previous studies have either included few patients^15,17,21,24 or had a short follow-up time.²⁵

In this study, we aimed to study the efficacy of TM compared with standard clinic visits on adherence to CPAP treatment after 6 months in a large, nonselected patient cohort. We further aimed to explore what other factors, including which side effects of CPAP treatment, have an impact on adherence.

METHODS

Study design and patients

This randomized controlled study was conducted at the Centre for Sleep and Breathing at Uppsala University Hospital, Uppsala, Sweden. Figure 1 shows the study flow. All consecutive patients (n = 327) who were prescribed CPAP treatment for diagnosed OSA from August 2017 to September 2018 were invited to participate in the study. Of those invited, 217 patients agreed to participate. Patients were then randomized into 2 groups: TM or SC.

Figure 1. Study design.

AHI = apnea-hypopnea index, CPAP = continuous positive airway pressure, DS14 = Type-D Scale questionnaire, ESS = Epworth Sleepiness Scale, FOSQ-10 = Functional Outcomes of Sleep Questionnaire, HAD = Hospital Anxiety and Depression scale, ODI = oxygen desaturation index, OSA = obstructive sleep apnea, SC = standard care, SECI = side effects to CPAP treatment inventory, TM = telemonitoring.

Intervention

CPAP initiation was carried out in groups with up to 6 patients per group. All patients received treatment with the same type of CPAP equipment: AirSenseS10 (ResMed, San Diego, CA, USA) that also includes humidification within the machine. Patients in the TM group had their first follow-up by telephone after 3–5 days and patients in the SC group had their first follow-up as a visit after 2 weeks, following the SC routine. Apart from assessing CPAP adherence at contacts, any issue with the CPAP treatment was assessed and addressed. After this first follow-up phone contact/appointment, further contacts (telephone or visits) were scheduled as needed for each individual patient until CPAP treatment was considered well established, treatment goal (usually apnea-hypopnea index [AHI] <5 events/h and usage ≥4 h/night) were reached, and the patient was satisfied with the treatment. All patients also had the possibility of initiating contacts (telephone or visits) if and when they needed, regardless of other scheduled contacts. This could also be done at any time after the CPAP treatment initiation was considered complete. All phone calls and visits were carried out by experienced nurses. At all contacts with the CPAP clinic, nurses used open-ended questions to assess patient experience with the CPAP, mask, treatment as a whole, and possible side effects. Issues raised by the patient were further investigated in the nurse/patient contact and addressed as needed. Direct questions, for instance on specific side effects, could be raised by the nurse if judged to be needed in the contact with the patient. As different patients can have different questions they want to raise regarding their treatment, there were no standardized questionnaires used at each of the contacts with the clinic, apart from the questionnaires used in the current study. If needed, CPAP settings were changed and new appointments were made for evaluation. Masks were also changed if needed (eg, due to leakage problems) and in such cases new appointments were made for evaluation. CPAP initiation was considered complete when a sufficiently good treatment effect was achieved, as assessed by the CPAP nurse, and the patient was satisfied with the treatment. If CPAP treatment was not tolerated by the patient, the patient did not want to continue with the treatment, or treatment was not effective despite nursing interventions, patients were booked in to have an appointment with a physician. In total, 21 patients from the SC group and 22 patients from the TM group did not complete the study.

All patients were followed up 6 months after treatment started. Patients who discontinued the CPAP treatment before the end of the study filled in a questionnaire about side effects of CPAP (side effects to CPAP treatment inventory [SECI]²⁶; as described below).

Telemonitoring

Telemonitoring of CPAP treatment was carried out using AirView (ResMed, USA). AirView is a secure cloud-based system for managing and monitoring CPAP treatment when using ResMed CPAP devices. The CPAP AirSense S10 (ResMed, USA), which was used in the present study, has a built-in wireless module that uses the mobile phone service network and technology to send and receive data on CPAP usage, mask leakages, pressure, and residual AHI. This means that telemonitoring does not require an internet connection or special patient skills.

Measurements and assessments

Anthropometric and clinical variables

Baseline data on clinical and anthropometric variables, such as age, sex, and body mass index (BMI), were available for all patients and assessed at baseline before they commenced CPAP treatment or responded to questionnaires. In addition, information on OSA severity (AHI) and oxygen desaturations (oxygen desaturation index [ODI]) was collected from the nocturnal respiratory recording performed in the clinical evaluation prior to the diagnosis of OSA (Figure 1).

Questionnaires

Daytime sleepiness was assessed by the Epworth Sleepiness Scale (ESS)²⁷ at baseline and at 6 months. The ESS is a self-administered questionnaire that consists of 8 items with 4 levels (0–3). The ESS score can thus range from 0 to 24, where a higher ESS score indicates a higher degree of daytime sleepiness. A cutoff of 10 points is often used to indicate daytime sleepiness.²⁷

The impact of sleepiness on daily activities was assessed with the Functional Outcomes of Sleep Questionnaire (FOSQ-10)²⁸ at baseline and at 6-month follow-up. The FOSQ-10 is a self-reported questionnaire that consists of 10 items designed to assess the impact of sleepiness on vigilance, activity level, general productivity, social outcomes, and intimate and sexual relationships. Items are scored on a 4-point scale (no difficulty, little difficulty, moderate difficulty, and severe difficulty). An average score is calculated for each subscale, and the 5 subscales are totaled to produce a total score from 5 to 20 points, with higher scores indicating better functional status.

Type D personality (distressed personality) is characterized by the presence of negative affectivity (dysphoria, anxious apprehension, and irritability) and social inhibition (discomfort in social situations, reticence, and lack of social poise). To investigate whether those characteristics have an impact on CPAP adherence, type D personality was identified using the Type-D Scale questionnaire (DS-14)^26,29 at baseline. The DS-14 consists of 14 items that measure negative affectivity and social inhibition, with a range of 0 to 4 (0 = false, 1 = rather false, 2 = neutral, 3 = rather true, and 4 = true). The 7 items on each subscale (negative affectivity and social inhibition, respectively) were summarized. A score ≥10 on both negative affectivity and social inhibition indicated a type D personality.¹³

Side effects of CPAP treatment were assessed using the SECI.²⁶ All patients who completed the study were asked to fill in the SECI at the 6-month follow-up. In addition, patients who did not complete the study were asked to fill in the SECI at the time of study discontinuation. The SECI comprises 15 items assessing possible side effects of CPAP. These are as follows: (1) blocked up nose, (2) runny nose, (3) nose bleed, (4) dry throat, (5) irritated eyes, (6) irritated bowel, (7) transient deafness, (8) feeling uncomfortable because of wearing CPAP mask, (9) increased awakenings, (10) uncomfortable pressure from the mask, (11) mask leaks, (12) cold air, (13) disturbing noise, (14) exhaling problems, and (15) anxiety during treatment. Each item is assessed along 3 dimensions: frequency (how frequently does each side effect appear), magnitude (how great a problem does each side effect cause), and impact (to what extent does each side effect decrease the use of CPAP), and scored on a 5-point Likert scale, with higher scores indicating worse problems (Table 1).

Table 1.

An example of a side effect (1. Blocked-up nose) revealed by 3 subquestions (Frequency scale, Magnitude scale, and Impact scale) taken from the SECI.

1. Blocked-Up Nose	Answer option 1	Answer option 2	Answer option 3	Answer option 4	Answer option 5
How frequently does each side effect appear?	Never	Seldom	Sometimes	Often	Very often
	1	2	3	4	5
How great a problem does each side effect cause?	No problems	Small problems	Some problems	Many problems	Very many problems
	1	2	3	4	5
How does each side effect decrease the use of CPAP?	Not at all	A little	Moderately	Much	Very much
	1	2	3	4	5

Columns indicate answer options. CPAP = continuous positive airway pressure, SECI = side effects of CPAP inventory.

On each of the 45 five-point Likert items, higher scores indicate worse side effects. In the analyses, responses were trichotomized. For items regarding questions about frequency, scores of 1 and 2 were considered to represent “never,” 3 to represent “occasionally,” and 4 and 5 “almost always.” With regard to questions about magnitude and impact, scores of 1 and 2 were considered to represent “minor,” 3 to represent “moderate,” and 4 and 5 “major.”

States of anxiety and depression were assessed by the Hospital Anxiety and Depression (HAD) scale.³⁰ The HAD scale was recorded both at baseline and at 6 months. The HAD questionnaire includes 14 questions on anxiety and depression (7 items each for the depression and anxiety subscales). Scoring for each item ranges from 0 to 3 (where 3 denotes the highest anxiety or depression level). A total subscale score ≥11 points out of a possible 21 points is considered to indicate symptoms of anxiety or depression.³¹

CPAP and adherence data

Data on CPAP usage as well as data on residual AHI, air leakage, and treatment pressure were downloaded from the CPAP devices using the ResScan software (ResMed, USA) in the SC group and through AirView (ResMed, USA) in the TM group.

Good CPAP adherence is usually defined as ≥ 4 hours/night for ≥ 5 nights/week (criterion A below).³² However, this criterion has been discussed in previous literature³³ and therefore we also included alternative adherence definitions (criteria B, C, and D).

Using the downloaded CPAP usage data (usage hours/night and number of nights used) adherence was defined as follows:

• Criterion A: CPAP usage ≥ 4 hours/night ≥ 5 nights/week

• Criterion B: CPAP usage ≥ 6 hours/night ≥ 5 nights/week

• Criterion C: CPAP usage ≥ 30 hours/week

• Criterion D: CPAP usage ≥ 42 hours/week

Ethical approval

This study was approved by the Ethical Review Board in Uppsala (2017/014-05/04). Patients who agreed to participate in the study gave their written informed consent during the visit. Participants from the TM group also gave written consent for the activation of a CPAP device for wireless data transmission via AirView (ResMed, USA).

Statistical analysis

Differences in continuous variables were analyzed using a 2-sample t test, whereas categorical variables were analyzed using the chi-square test or Fisher’s exact test. In the main analysis, the outcome was CPAP adherence according to criterion A and the main exposure was TM vs SC group. Criteria B, C, and D were used as alternative definitions.

In a secondary analysis we analyzed the impact of CPAP treatment side effects on adherence (criterion A). For this analysis, we included all patients including those who discontinued CPAP treatment.

In addition, a logistic regression analysis was performed concerning the impact of side effects on CPAP adherence based on the SECI questionnaire. Binary logistic regression modeling was used to explore associations between dependent and independent variables (ie, to determine whether some side effects were risk factors for reducing CPAP usage, indicating an increased risk of being nonadherent). The analysis was adjusted for age, sex, and smoking.

Probability values lower than 0.05 were considered statistically significant. All analyses were performed with the IBM Statistical Package for the Social Science (SPSS) version 22 (IBM Corporation, Armonk, NY).³⁴

RESULTS

Most patients were male, middle-aged, overweight, and had severe OSA. Age, BMI, sex, smoking, AHI, ODI, ESS, HAD, and FOSQ-10 at baseline did not differ between SC and TM groups (Table 2).

Table 2.

Baseline characteristics of all patients included in the study (including those who discontinued) at baseline and the standard-care group compared with the telemonitoring group.

	All Patients (n = 217)	Standard-Care Group (n = 82)	Telemonitoring Group (n = 92)	P for Difference
Age, y	56.2 ± 12.1	56.3 ± 11.8	56.2 ± 12.3	.94
BMI, kg/m2	32.5 ± 6.3	32.9 ± 6.7	32.1 ± 5.8	.37
Male, %	152 (70.0)	58 (71.1)	63 (68.9)	.73
Smoker, %	38 (17.5)	14 (17.0)	15 (16.3)	.58
AHI, events/h	31.9 ± 18.5	33.1 ± 20.4	30.6 ± 16.1	.33
ODI, events/h	29.4 ± 18.2	30.9 ± 19.9	27.7 ± 16.0	.19
ESS	9.7 ± 4.9	9.8 ± 4.8	9.6 ± 4.9	.85
HAD-A	12.5 ± 2.4	12.4 ± 2.3	12.5 ± 2.4	.96
HAD-D	8.3 ± 1.9	8.5 ± 1.7	8.2 ± 1.9	.52
FOSQ-10	14.6 ± 3.7	14.4 ± 3.6	14.8 ± 3.7	.43

Values are presented as mean ± SD or n (%). AHI = apnea-hypopnea index, BMI = body mass index, ESS = Epworth Sleepiness Scale, FOSQ-10 = Functional Outcomes of Sleep Questionnaire, HAD-A = Hospital Anxiety and Depression scale for anxiety, HAD-D = Hospital Anxiety and Depression scale for depression, ODI = oxygen desaturation index.

CPAP adherence levels in the TM vs SC approach

CPAP adherence at the 6-month follow-up did not differ between the TM and SC groups for any of the definitions of adherence. Improvements in AHI, ESS, and global FOSQ-10 scores were seen in both groups at 6-month follow-up compared with baseline (Table 3).

Table 3.

Comparisons of CPAP adherence and other variables between the TM group and SC group at 6-month follow-up.

	SC Group (n = 82)	TM Group (n = 92)	P for Difference
CPAP adherence
Total CPAP usage per total number of days, min	270.6 ± 141.5	274.3 ± 137.0	.86
Total CPAP usage per usage days, min	335.4 ± 111.4	339.6 ± 107.9	.80
Total number of CPAP usage days/6 mo	141.5 ± 51.3	135.9 ± 50.4	.47
≥ 4 h/night ≥ 5 nights/wk	40 (48.7)	49 (53.2)	.54
≥ 6 h/night ≥ 5 nights/wk	28 (34.1)	36 (39.1)	.53
≥ 30 h/wk	46 (56.0)	57 (61.9)	.43
≥ 42 h/wk	27 (32.9)	26 (28.2)	.62
ΔAHI, 1 event/h	–31.1 ± 18.6	–28.5 ± 14.4	.65
ΔESS	–2.1 ± 0.2	–2.0 ± 0.8	.67
ΔFOSQ-10	2.0 ± 0.2	1.1 ± 0	.12
ΔHAD-A	0.4 ± 0.1	0.5 ± 0	.57
ΔHAD-D	0.1 ± 0.5	0.4 ± 0.3	.74

Values are presented as mean ± SD or n (%). CPAP = continuous positive airway pressure, SC = standard care, TM = telemonitoring, ΔAHI = difference in apnea-hypopnea index between follow-up (FU) and baseline (BL), ΔESS = difference in Epworth Sleepiness Scale between FU and BL, ΔFOSQ-10 = difference in Functional Outcome Sleep Questionnaire between FU and BL, ΔHAD-A = difference in Hospital Anxiety and Depression scale for anxiety between FU and BL, ΔHAD-D = difference in Hospital Anxiety and Depression scale for depression between FU and BL.

Side effects related to discontinuation of CPAP treatment

In total, 157 completed the SECI questionnaire (SC = 69, TM = 76, and 12 patients who discontinued treatment).

Patients who discontinued CPAP treatment scored significantly higher on 7 of the 15 side effects compared with patients who completed the study, as shown in Figure 2. Dry throat (SECI question [Q] 4; Q4), increased awakenings (Q9), uncomfortable pressure of the mask (Q10), cold air (Q12), disturbing noise (Q13), exhaling problems (Q14), and anxiety during treatment (Q15) were all common problems in patients who discontinued CPAP treatment (Figure 2). Of the patients who discontinued, a high percentage (25–50%) scored “almost always” on the frequency scale or “major” on the magnitude or impact scales compared with patients who completed the study (2.3–14%).

Figure 2. Comparison of data from the SECI on frequency (how frequently), magnitude (how great a problem), and impact scale (decreased use of CPAP) between patients who completed (n = 145) and discontinued (n = 12) treatment.

Values are presented as percentages. *P < .05. Presented side effects are selected based on significant differences between groups on the impact scale (P < .05). Not presented side effects showed nonsignificant differences between groups on the impact scale (P > .05). Questions about frequency, magnitude, and impact were trichotomized: scores of 1 and 2 represent never/minor, 3 represents occasionally/moderate, and 4 and 5 represent almost/always/major. CPAP = continuous positive airway pressure, Q = question, SECI = side effects of CPAP inventory.

Differences in side effects between adherent and nonadherent patients

Among patients that completed the study, the most common side effect for both adherent and nonadherent groups was a blocked-up nose. In addition, the nonadherent group scored significantly higher on the impact scale on the side effects of dry throat (Q4; P < .05), increased awakenings (Q9; P < .01), and exhaling problems (Q14; P < .05) (Figure 3).

Figure 3. Comparison of the most common side effects (presented side effects based on significant differences between groups on the impact scale; P < .05).

*P < .05. Data from the SECI on frequency (how frequently), magnitude (how great a problem), and impact scale (decreased use of CPAP) between adherent (n = 88) and nonadherent (n = 57) patients. Questions about frequency, magnitude, and impact were trichotomized: scores 1 and 2 represent never/minor, 3 represents occasionally/moderate, and 4 and 5 represent almost/always/major. CPAP = continuous positive airway pressure, Q = question, SECI = side effects of CPAP inventory.

Factors related to CPAP adherence

When analyzing factors other than CPAP side effects in relation to CPAP adherence, no differences between the adherent and nonadherent patients were found regarding baseline age, BMI, sex, AHI, type D personality, education, employment, alcohol consumption, physical activity, or comorbidities (Table 4). There was, in addition, no difference between adherent or nonadherent patients in terms of mask leakage, mask pressure, and reduction in AHI at follow-up. Baseline total ESS score did not differ between adherent and nonadherent groups (P = .38), and both groups improved in ESS scores. However, at the 6-month follow-up, the adherent group had improved significantly more (P = .04). Also, for FOSQ-10, the baseline total score did not differ between adherent and nonadherent groups, whereas the adherent group improved significantly more than the nonadherent group after 6 months (P < .01). For HAD scores, no significant differences were seen after 6 months. Smoking was more common in the nonadherent group (P = .01). More patients who had never smoked were adherent (41 patients) as opposed to nonadherent (27 patients) (Table 4).

Table 4.

Comparison between adherent and nonadherent groups.

	Adherent (n = 89)	Nonadherent (n = 85)	P for Difference
Leakage (95%), L/min (FU)	14.8 (8.9)	15.2 (9.5)	.82
Pressure (95%), L/min (FU)	11.1 (1.7)	10.8 (1.9)	.40
ΔAHI	–32.7 ± 17.6	–28.5 ± 15.9	.18
ΔESS	–2.7 ± 1.2	–1.5 ± 0.1	.04
ΔFOSQ-10	2.2 ± 0.7	–1.4 ± 0.9	<.01
ΔHAD-A	–0.5 ± 0.4	0.6 ± 0.3	.43
ΔHAD-D	0 ± 0.1	0.2 ± 0.7	.69
Age,* y	57.9 ± 11.6	56.4 ± 10.7	.41
BMI,* kg/m2	32.0 ± 5.6	33.3 ± 6.0	.14
Male*	58 (65.2)	58 (68.2)	.25
Type D personality*	22 (27.8)	18 (27.7)	0.99
Physical activity,* min/wk	185.5 ± 200	161.0 ± 156	.47
Education*
Elementary school or lower	12 (13.4)	10 (11.7)	.78
Upper secondary school	31 (34.8)	23 (31.1)
At least 1 year of education after upper secondary school	12 (13.4)	9 (10.5)
Postsecondary school	30 (33.7)	32 (37.6)
Smoking status*
Never smoked	41 (46.0)	27 (31.7)	.01
Former smoker	26 (29.2)	14 (16.4)
Smoker	9 (10.1)	20 (23.5)
Employment*
Paid employment	48 (53.9)	42 (49.4)	.32
Student	1 (1.1)	1 (1.1)
Retired	27 (30.0)	18 (21.1)
Unemployed	1 (1.1)	1 (1.1)
Other	16 (17.9)	19 (22.3)
Alcohol* (standard glasses/wk)
No alcohol	20 (22.4)	15 (17.6)	.37
1–4	38 (42.6)	23 (27.0)
5–9	11 (12.3)	12 (14.1)
10–14	3 (3.4)	5 (5.9)
15+	2 (2.2)	0 (0.0)
Unknown	20 (22.4)	25 (29.4)
Comorbidity*
Asthma	14 (18.2)	11 (17.7)	.95
Allergy	16 (20.8)	14 (22.6)	.83
Airways problem	15 (19.5)	12 (19.4)	0.99
Hay fever	7 (9.1)	6 (9.7)	0.99
Chronic nasal congestion	20 (26.0)	14 (22.6)	.65
Myocardial infarction	5 (6.6)	4 (6.5)	0.99
Other heart disease	5 (6.6)	4 (6.5)	0.99
Hypertension	41 (53.9)	32 (51.6)	.78
Headache	23 (30.3)	19 (31.1)	.91
Impotence	13 (17.1)	7 (11.3)	.34
Joint problems	25 (32.9)	18 (29.0)	.62
Mental disorders	13 (17.1)	9 (14.5)	.68
Acid reflux	30 (40.0)	22 (35.5)	.59

Values are presented as mean ± SD or n (%). *Variables measured at baseline. BMI = body mass index, FU, follow-up, ΔAHI = difference in apnea-hypopnea index between FU and baseline (BL), ΔESS = difference in Epworth Sleepiness Scale between FU and BL, ΔFOSQ-10 = difference in Functional Outcome Sleep Questionnaire between FU and BL, ΔHAD-A = difference in Hospital Anxiety and Depression scale for anxiety between FU and BL, ΔHAD-D = difference in Hospital Anxiety and Depression scale for depression between FU and BL.

To assess if side effects were related to nonadherence after also adjusting for other factors, logistic regression modeling was performed adjusting for confounders including smoking, as the prevalence of nonadherence was higher in smokers.

Dry throat, increased awakenings, and exhaling problems were all significantly associated with nonadherence after adjusting for age and sex. Adjustment for smoking attenuated the associations, which no longer became significant (Table 5).

Table 5.

Logistic regression model to predict the risk of being nonadherent at the 6-month follow-up due to side effects during CPAP usage: dry throat, increased awakenings, and exhaling problems among the patient groups.

Side Effect with Major Impact	Model 1		Model 2
	OR (95% CI)	P	OR (95% CI)	P
Dry throat	2.17 (1.25–3.70)	<.01	1.88 (0.92–3.84)	.08
Increased awakenings	2.50 (1.31–4.76)	<.01	1.96 (0.97–4.00)	.06
Exhaling problems	3.70 (1.25–10.08)	.01	3.0 (1.00–10.00)	.06

Model 1 was adjusted for age and sex. Model 2 was adjusted for age, sex, and smoking status. CI = confidence interval, CPAP = continuous positive airway pressure, OR = odds ratio.

DISCUSSION

Within this randomized controlled study, we could not show a difference in CPAP adherence between TM and SC using several different adherence criteria. Among factors relating to adherence, we showed that the side effects of dry throat, increased awakenings, and exhaling problems are the most important factors relating to CPAP discontinuation and CPAP nonadherence, and that smoking status influenced these relationships.

An overall CPAP use of 4 hours per day or CPAP use of at least 4 hours per day at least 70% of nights has been established as a clinical benchmark for CPAP adherence.³⁵ The definition is arbitrary, however, and a general dose–response relationship between CPAP use and improvement in, for example, daytime sleepiness, accident risk, and blood pressure after 4 weeks of treatment has been reported.^32,36 Therefore, in the current study, we used several different criteria for CPAP adherence, although we used the criterion of 4 hours per day as the main definition. However, this did not change the results.

Previous studies have studied different TM approaches to promote CPAP therapy and to increase CPAP adherence, and the results have been mixed.^{15,17,19–22} Similar to our results, 2 previous studies with a patient population similar to our study did not show improvement in CPAP adherence using TM follow-up.^17,22

Fox et al¹⁵ and Sparrow et al²⁰ have, in contrast, shown increased CPAP adherence when using telehealth approaches. Fox et al found improvement in adherence in the TM arm by approximately 90 minutes compared with the SC arm after 3 months in a nonblinded, single-center, randomized controlled trial.¹⁵ The patients were contacted, as necessary, if having problems (mask leak, low residual OSA), and for information on side effects or other problems. If lack of motivation to use CPAP was present, patients were encouraged to use CPAP. In a larger study (n = 250), Sparrow et al used a telehealth intervention to improve CPAP adherence.²⁰ Here, an automated telephone-linked communication system was used in the TM arm to promote CPAP use with motivational interviewing, whereas the control arm received general health information. Results showed a 30% higher rate of CPAP adherence (> 4 h/night) in the intervention arm.

In comparison with our study, patients in the study by Fox et al were comparable in age, BMI, and sex distribution; however, they had a smaller cohort and did not include patients with cardiopulmonary or psychiatric disease. It is possible that patients in the study by Fox et al were somewhat healthier than those in our patient group and that this affected the results. In the current study we aimed to include a patient population as similar as possible to the population we see every day in the clinic and did not exclude patients based on other diseases. Sparrow et al had a telephone-linked communication system with motivational interviewing,²⁰ which could explain our contrasting results. The present study did not include motivational interviews for TM, but rather, all patients within the study received motivational support from the CPAP nurses within the clinic, based on the patient’s individual needs. It is possible that TM in itself does not increase CPAP adherence.²⁰ Structures for this may need to be built in to future TM systems for this approach to be fully effective. Although our study as well as some other studies did not show increased adherence with a TM approach, TM may have other benefits and be of importance as TM can reduce the number of clinic visits,¹⁷ can be more cost-effective,¹⁶ and can save time for health care staff.¹⁸

In the present study, we showed that the side effects most associated with cessation of and nonadherence to CPAP treatment were dry throat, increased awakenings, and exhaling problems. In addition, cessation of CPAP treatment was also associated with factors such as uncomfortable pressure from the mask, cold air, disturbing noise, and anxiety during treatment. We could, however, not show any differences between groups regarding other factors such as daytime sleepiness, impact of daytime sleepiness on daily activities, type D personality, education, or comorbidities. These results are consistent with several previous studies^8,19,37 showing nose blockage, dry mouth, and increased awakenings as the main reasons for nonadherence but no overrepresentation of type D personality in nonadherent patients.¹³ The side effects of CPAP remain important issues that need to be addressed properly within clinical practice.

In the current study, smoking attenuated the relationship between side effects and nonadherence. Smoking has previously been associated with severity of OSA³⁸ and is a risk factor for low CPAP adherence.¹⁰ It is therefore important to assess smoking status both prior to starting CPAP and with side effects during the time of starting CPAP. Humidification has been shown to decrease side effects and increase adherence,³⁹ and potentially this may be of even greater importance in smokers. Future studies need to assess this further.

Strengths and limitations

This study used a randomized controlled design and collected data in a real-life patient cohort of considerable size. In addition, there were objective data on adherence, treatment effect, and mask leakage. Side effects of CPAP were also assessed in detail and there was information on several important baseline factors. Nonetheless, we acknowledge that there are limitations to our study. First, as the structure of follow-up in our study was similar in both groups, although the TM group was followed up after 3–5 days and the SC group was followed up after 2 weeks, it is possible that we would have needed to also add interventions (eg, motivational interviews) or have yet another arm to test the effect of different combinations of TM and other inventions. However, this was not within the article’s scope. Second, it is most likely that there are subgroups of patients where TM is more beneficial, but the study cohort size did not allow for this type of analysis. Adherence data were not available to the patients in real time other than the fact that the device used in the study presented user time for the previous night upon turning the machine off, and also the possibility of accessing mean adherence data over longer periods of time via the patient menu in the device. The patients in the study were not explicitly informed about the possibility of accessing data on adherence over time in the device, but likely some of the patients will have accessed this, which could have affected adherence for these patients. Unfortunately, we have no information on which of the patients could have used this. From clinical experience, the possibility of accessing data from the patient menu is used with great variability within the OSA patient group. We have no reason to believe that the number of patients accessing the adherence data via the patient menu would differ significantly between the 2 groups.

CONCLUSIONS

In conclusion, we could not show that TM improved CPAP treatment adherence over 6 months. Side effects of CPAP treatment, such as dry throat, increased awakenings, and exhaling problems, had an independent impact on CPAP adherence and smoking had a negative impact on CPAP treatment adherence. Thus, preventing side effects and assessing smoking status is of importance when wanting to improve CPAP adherence.

DISCLOSURE STATEMENT

All authors have seen and approved the manuscript. This study was funded by a grant from the Selander Foundation (Uppsala University). F.D.’s research time within this project was supported by the Research and Development Centre, Sörmland County Council, Sweden. The authors report no conflicts of interest.

ABBREVIATIONS

AHI

apnea-hypopnea index

BMI

body mass index

CPAP

continuous positive airway pressure

ESS

Epworth Sleepiness Scale

FOSQ-10

Functional Outcomes of Sleep Questionnaire

HAD

Hospital Anxiety and Depression

ODI

oxygen desaturation index

OSA

obstructive sleep apnea

Q

question

SC

standard care

SECI

side effects to CPAP treatment inventory

TM

telemonitoring