Issue Navigator

Volume 14 No. 07
Earn CME
Accepted Papers

Scientific Investigations

Positive Airway Pressure Therapy Is Challenging for Patients With Epilepsy

Véronique Latreille, PhD; Ellen J. Bubrick, MD; Milena Pavlova, MD
Department of Neurology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts


Study Objectives:

To investigate whether patients with epilepsy and comorbid obstructive sleep apnea (OSA) are more likely to be nonadherent to positive airway pressure (PAP) therapy than adults with OSA but without epilepsy.


This retrospective study included patients with epilepsy diagnosed with OSA and age-, sex-, and apnea-hypopnea index (AHI)-matched controls with OSA but without epilepsy who started PAP treatment between February 2014 and August 2017. Subjects' adherence to PAP therapy was continuously recorded electronically, and comparisons were made at 1 month, 3 months, and 1 year following PAP initiation. Predictors to poor adherence were also evaluated.


Patients with epilepsy (n = 23) were less adherent to PAP than controls (n = 23) during the first month of treatment (13% versus 78%, P = .03). During this first month, average PAP use was lower in patients with epilepsy (4.7 ± 2.1 hours) relative to controls (6.1 ± 1.2 hours, P = .03). Despite sustained PAP treatment, patients with epilepsy had a greater residual AHI and were five times more likely than controls to have residual apnea events above normal range at 3-month and 1-year follow-up. However, no clinical characteristics could significantly predict poor adherence in patients.


Patients with epilepsy are less likely to be adherent to PAP therapy during the first month of treatment, as compared to adults with OSA but no epilepsy. Moreover, PAP therapy could not sufficiently reduce AHI in up to 72% of patients. These findings highlight the need for careful monitoring of PAP treatment in patients with epilepsy, as untreated OSA may worsen seizure burden.


Latreille V, Bubrick EJ, Pavlova M. Positive airway pressure therapy is challenging for patients with epilepsy. J Clin Sleep Med. 2018;14(7):1153–1159.


Current Knowledge/Study Rationale: Obstructive sleep apnea (OSA) is common in patients with epilepsy, and treatment adherence issues are confronted daily. Positive airway pressure (PAP) adherence may be even more challenging for patients with epilepsy than those who are healthy, but this has not been examined systematically.

Study Impact: Not only are patients with epilepsy less adherent to PAP therapy, but they are also more likely to have a higher number of residual apneas following PAP treatment. This finding is of potential high relevance to clinical practice, especially given the relationship between seizures and central apneas, and possibly associated risk of sudden unexpected death in epilepsy.


Obstructive sleep apnea (OSA) is common in adults with epilepsy, affecting up to one-third of patients.1,2 By causing nocturnal episodes of intermittent hypoxemia and sleep fragmentation, occurrence of OSA in patients with epilepsy may increase seizure burden and morbidity.3 Positive airway pressure (PAP) is the gold-standard therapy for OSA. In addition to reducing the number of apnea events, PAP helps improve quality of life, daytime sleepiness, cardiovascular health, and cognition in adults with OSA.4

Yet, in the general population, adherence to therapy is highly variable, ranging from 29% to 83%, and constitutes a major limitation to PAP effectiveness.5 In patients with complex neurological conditions such as epilepsy―which is also associated with several comorbidities―it has been postulated that PAP adherence may be even poorer, although this has never been examined in detail. However, treating OSA in patients with epilepsy is of clinical importance, as prior reports have shown that treatment with PAP therapy may improve seizure control and reduce interictal epileptiform discharges during sleep.1,6

We conducted a retrospective study to examine whether patients with epilepsy and comorbid OSA are more likely to be nonadherent to PAP therapy as compared to adults with OSA but without epilepsy matched for age, sex, and OSA severity. We also examined possible predictors of poor PAP adherence.


Subject Selection and Clinical Characteristics

We retrospectively reviewed clinical and polysomnography (PSG) data from all adult individuals consecutively referred for evaluation of OSA at Brigham and Women's Faulkner Hospital's Sleep Center from February 2014 to August 2017 to identify patients with epilepsy in whom OSA had been diagnosed and for which PAP had been prescribed. The study was approved by the hospital's institutional review board.

Inclusion criteria were (1) diagnosis of epilepsy confirmed by board-certified epileptologists; (2) diagnosis of OSA (apneahypopnea index [AHI] ≥ 5 events/h, along with symptoms) per standard criteria7; and (3) initiation of PAP therapy. Patients with nonepileptic seizures or provoked epileptic seizures were excluded from the analysis.

As a comparison, we also included a group of control subjects without epilepsy in whom OSA was diagnosed and who underwent PAP treatment. Except for OSA and related cardiovascular risk factors (ie, hypertension and diabetes), these subjects were free of any neurological disorders. The control subjects were consecutively identified from the initial screening search to match patients with epilepsy on age, sex, and AHI at the time of diagnosis.

For each subject, we collected demographic and clinical data, including body mass index (BMI), smoking status, neck circumference, Mallampati classification (class 0–4), and cardiovascular risk factors. For patients with epilepsy, we also recorded epilepsy duration, seizure characteristics, number of antiepileptic drugs, and medication adherence.

Polysomnography Recordings

All subjects attended in-laboratory overnight PSG during which electroencephalography, electrooculography, submental and anterior tibialis electromyography, and electrocardiography were recorded continuously and simultaneously. Respiration was monitored continuously, using oronasal thermistor and nasal pressure, thoracic and abdominal belts, and finger pulse oximetry sensors. Leg movements were recorded using surface electrodes on anterior tibialis muscles. Sleep stages were visually scored per standard criteria.7 Baseline PSG variables included total sleep time, sleep latency and efficiency, wake after sleep onset, arousal index, number of awakenings, periodic limb movement in sleep index, nadir oxygen saturation (SaO2), and O2 desaturation index, as well as OSA-related variables (total AHI, rapid eye movement sleep AHI, non-rapid eye movement sleep AHI, supine AHI, obstructive apnea index, and central apnea index).

Assessment of PAP Adherence

Adherence data were collected at 1 month, 3 months, and 1 year following PAP initiation. At each endpoint, the following variables were examined: (1) PAP average usage, expressed in hours; (2) adherence rate (defined as ≥ 4 hours use, ≥ 70% of the time); (3) mean PAP pressure, expressed in cmH2O; (4) residual AHI; (5) number of cases with residual AHI ≥ 5 events/h; (6) residual obstructive apnea index (number of events/h of sleep); (7) residual central apnea index (number of events/h of sleep); and (8) average time in large leak per day, expressed in minutes. We also recorded total treatment duration, adverse effects, and whether subjects reported benefits of PAP therapy (ie, subjectively improved daytime sleepiness, snoring, and/or improved sleep quality) during follow-up visits with their treating physician. Seizure frequency, BMI, and Epworth Sleepiness Scale scores8 were examined at baseline (prior to PAP therapy) and at follow-up (at least 1 month after PAP initiation).

Statistical Analysis

Comparisons among patients with epilepsy and matched controls were performed using independent sample t test or nonparametric Mann-Whitney U test for continuous data, and Pearson chi-square test for categorical data. Because of the large number of comparisons, we adjusted for the false discovery rate using the Benjamini-Hochberg procedure on P values.9 Corrected P thresholds (p′) are reported for PAP adherence data. Binary logistic regression was performed to identify clinical predictors of poor PAP adherence at 3 months in patients with epilepsy. Statistical significance was set at P < .05.


A total of 2,120 adults underwent PSG for evaluation of OSA between February 2014 and August 2017 (Figure 1). Of that sample, epilepsy was diagnosed in 76, and 55 met OSA criteria. Twenty-eight patients with epilepsy were prescribed PAP; however, adherence data were unavailable in 5 patients (ie, no data card or follow-up care outside of Brigham and Women's Hospital). Thus, 46 subjects, including 23 patients with epilepsy and 23 matched controls, were included in the analysis.

Data flow diagram illustrating patients and controls selection process.

Criteria for OSA severity were defined as follows: mild, AHI between 5–14 events/h; moderate, AHI between 15–29 events/h; and severe, AHI ≥ 30 events/h. Alternative therapy to PAP included Provent (n = 1) and oral appliances (n = 1). AHI = apnea-hypopnea index, OSA = obstructive sleep apnea, PAP = positive airway pressure, PSG = polysomnography.


Figure 1

Data flow diagram illustrating patients and controls selection process.

(more ...)

Surprisingly, a relatively large proportion of patients with epilepsy and OSA (22%) refused to use PAP. Of these, most had mild OSA (75%), whereas two patients had severe OSA. Most of those who declined PAP were men (75%), and tended to be younger than patients with epilepsy and OSA who agreed to start PAP therapy (52.0 ± 15.0 versus 57.8 ± 12.8 years).

Patients with epilepsy did not differ from controls on most clinical characteristics, except for smoking status, where a higher incidence of active smokers was found in the patient group (Table 1). Most patients had left temporal lobe epilepsy, and seizures were well controlled in 19 patients (83%). One patient had a mild developmental delay, but none had severe epilepsy syndromes.

Clinical characteristics of patients with epilepsy and matched control subjects.


table icon
Table 1

Clinical characteristics of patients with epilepsy and matched control subjects.

(more ...)

No significant differences were observed between the groups for any sleep variables at baseline (Table 2).

Baseline PSG data in all participants.


table icon
Table 2

Baseline PSG data in all participants.

(more ...)

PAP Adherence and Response to Therapy

Most patients with epilepsy were on autoadjusting continuous PAP (n = 20), whereas three patients were on auto-adjusting bilevel PAP. In the control subjects, one was on autoadjusting bilevel PAP, whereas the remaining subjects were on autoadjusting continuous PAP. Patients with epilepsy were less adherent to PAP therapy than controls during the entire follow-up period, although the difference in adherence rates was statistically significant at 1 month only (Table 3). During this first month, patients with epilepsy used PAP on average 1.4 hours less than controls. Compared with controls, patients with epilepsy also tended to have higher levels of large air leak, especially within the first month and at 1-year follow-up, with values approximating one half hour per night on average. Moreover, a lower proportion of patients with epilepsy reported improvement of OSA symptoms with PAP therapy, whereas almost all control subjects reportedly benefitted from PAP at follow-up, although again this difference was not statistically significant.

PAP adherence data in all participants.


table icon
Table 3

PAP adherence data in all participants.

(more ...)

Despite PAP use, patients with epilepsy had a greater residual AHI and were about five times more likely than controls to have residual AHI ≥ 5 events/h at 3-month and 1-year follow-up. To examine whether PAP treatment was more efficient in reducing the number of residual apnea events over the follow-up periods in controls as compared to patients with epilepsy, we performed repeated measures analysis of variance (2 groups: epilepsy and controls × 3 endpoints: 1 month, 3 months, and 1 year) on residual AHI data. The analysis failed to show a statistically significant interaction between groups and time (endpoints), suggesting that the changes in residual AHI with PAP therapy did not differ between patients with epilepsy and controls over the follow-up periods. Yet a trend was observed for a larger decrease in residual AHI in control subjects relative to patients with epilepsy across the follow-up periods. In addition to the observed group effect (patients with epilepsy showing higher residual AHI), we found a significant effect of time (F2,72 = 5.80, P = .005), with post hoc comparisons indicating a significant decrease in residual AHI from 1-month (7.6 ± 8.5 events/h) to 1-year (6.0 ± 5.7 events/h) follow-up periods in both groups. Therefore, both groups showed overall an improved outcome after 1 year of PAP use.

Although the residual apnea events tended to be more frequently obstructive than central in nature, both indexes were higher in patients with epilepsy than controls, particularly at 1-year follow-up. Only one subject from the epilepsy group, but none from the control group, reached the threshold criteria of central apnea index ≥ 5 events/h at the 1-month follow-up. However, this patient previously showed both obstructive and central sleep apnea during the baseline PSG, with indexes of 7.9 and 10.5 events/h respectively, and therefore does not meet the diagnostic criteria of treatment-emergent central sleep apnea.10

To test the hypothesis that treatment-emergent central sleep apnea is more frequently found in patients with epilepsy following PAP therapy than in control subjects, we further analyzed the frequency of emerging central events during PAP titration in the subset of subjects who underwent a baseline split-night PSG study (epilepsy n = 8; controls n = 8). Surprisingly, none of the patients with epilepsy met the criteria for treatment-emergent central sleep apnea (all had a central apnea index < 5 events/h), whereas one control subject was considered to have treatment-emergent central events (mean central apnea index of 5.7 events/h during titration).

Adverse Effects of PAP Therapy

The most common reported adverse effects in patients with epilepsy who were nonadherent to PAP therapy were mask discomfort (71%), mask leak (14%), and/or problems with pressure settings (14%). Among controls, adverse effects in subjects nonadherent to PAP therapy were quite similar and included mask discomfort (67%), mask leak (67%), and/or dry mouth (33%).

Medically Refractory Epilepsy: Adherence and Seizure Frequency Before and After PAP Therapy

Four patients had medically refractory seizures. Although these results have to be interpreted with caution given the small number of patients in this subgroup, patients with medically refractory seizures tended to have a lower PAP adherence rate (1 month 25%, 3 months 25%, 1 year zero) and a higher residual AHI (1 month 13.7 ± 4.9, 3 months 11.8 ± 5.8, 1 year 11.8 ± 6.0 events/h), as compared to patients with medically controlled epilepsy (PAP adherence 1 month 42%, 3 months 47%, 1 year 37%; residual AHI 1 month 8.1 ± 9.6, 3 months 7.5 ± 6.7, 1 year 7.4 ± 5.2 events/h).

After 1 year of PAP therapy, seizure frequency remained unchanged in 3 out of 4 patients (average frequency, 36.8 seizures/mo; range, 5.5–75 seizures/mo). In one case, initiation of PAP, along with replacement of the vagus nerve stimulation device, led to a decrease in seizure frequency (from three seizures a month to one seizure every few months).

Predictors of Poor PAP Adherence in Patients With Epilepsy

Binary logistic regression analysis failed to reveal any significant clinical predictors of nonadherence at 3-month follow-up in patients with epilepsy (n = 20; data not shown).


Our findings indicate that patients with OSA who have epilepsy struggle more with PAP treatment than their otherwise healthy counterparts. They use it less frequently and for shorter periods, especially within the first month of therapy. Moreover, our results demonstrate that even with consistent PAP use, patients with epilepsy are more likely to have residual respiratory events above normal range (AHI > 5 events/h); these include both obstructive and central apnea events. Patients with medically refractory seizures also tend to be less adherent to PAP therapy and have a higher residual AHI despite PAP use than patients with medically controlled seizures.

In the general population, adherence to PAP therapy is highly variable, ranging from 29% to 83%.5 Our results indicate that adherence to PAP therapy can also be greatly heterogeneous in the epilepsy population, varying from 13% to 50% within the first year of therapy. Yet overall, adherence rates of patients with epilepsy remain significantly lower as compared to patients with OSA without epilepsy. One previous study performed in patients with epilepsy reported an adherence rate of 68%, which is higher than what we found in our study, and approximates the adherence rate of our control subjects.11 However, the patients with epilepsy in the previous study had to be using PAP therapy for at least 6 months before entering the trial, and no controls without epilepsy were included as a comparison.11 As per our findings, it is likely that a significant proportion of patients with epilepsy could not be included in the study because of nonadherence to PAP therapy or discontinuation of PAP therapy within the first 6 months.

It is possible that the burden of chronic disease makes PAP use more challenging for patients with epilepsy, and as they often have other health-related issues to attend to, PAP might become less of a priority. However, most of our patients were adherent with their antiepileptic medications, so volitional nonadherence is not as likely to contribute in a major way. Other factors that may contribute include other concomitant sleep disorders, such as insomnia or occurrence of nocturnal seizures that may increase the number of awakenings, and thus reduce the likelihood of patients to keep their mask on throughout the night. Many patients with epilepsy also fear lack of sleep as a trigger for their seizures, and may be more likely to abandon PAP therapy sooner than others if they think it is interfering with their ability to sleep. Patients with epilepsy who are nonadherent to PAP therapy are at a higher risk of seizures than patients who are adherent,11 and therefore it is important to address these false beliefs and better educate patients about PAP therapy.

Psychosocial stressors or presence of mood disorders such as depression and anxiety, which are very common in patients with epilepsy,12 may also put them at risk of being less adherent to PAP therapy. However, previous studies performed in adults with OSA but without neurological disorders have shown that the presence of depression or anxiety was not a major predictor of poor adherence to PAP therapy.13 Rather, PAP adherence was low if adults did not experience greater improvements in OSA symptoms with treatment,14 which was the case in our study. Indeed, as compared to controls, a lower proportion of patients with epilepsy reported to benefit from PAP therapy at follow-up visits, although the difference was not statistically significant after adjusting for multiple comparisons. Finally, one other contributor of poor adherence in our patients with epilepsy may be related to technological device factors such as air leak, which tended to be higher within the first month of treatment in patients with epilepsy, and it was previously shown to predict poor adherence to autoadjusting PAP therapy.15 Although larger studies will be needed to better identify predictors of poor PAP adherence in patients with epilepsy, it is possible that mask leaks may to some extent be related to facial morphological differences of patients with OSA and epilepsy relative to those without epilepsy. However, none of the patients included in our study had syndromes with distinct dysmorphic features.

To our knowledge, this is the first report demonstrating a higher number of residual apneas following PAP therapy in patients with epilepsy. Whether this might be reflecting treatment-emergent central sleep apnea is uncertain, and our additional analyses do not seem to support this hypothesis. It is possible that the residual apneas are due to other disturbances in the regulation of breathing specific to epilepsy, or even to some medications, such as antiepileptic drugs or benzodiazepines that may affect breathing and muscle atonia during sleep. However, a similar number of patients with epilepsy and controls had regular benzodiazepine use, and therefore it is unlikely that benzodiazepine intake could have explained our findings. Future prospective studies will help in determining the potential causes. Nevertheless, these findings are of potential high relevance to clinical practice, especially given the relationship between seizures and central apneas, and the possibly associated risk of sudden unexpected death in epilepsy.16,17

Some limitations of our study should be noted. It is a retrospective chart review and uses a patient population that is seen in a regular clinic, and as such, the main inclusion criteria are somewhat limited. However, we very specifically selected a population with available data and meticulously matched control subjects on parameters relevant for OSA (such as age, sex, BMI, neck circumference, number of cardiovascular risk factors, and OSA severity). Because of the small number of patients, it is not clear whether our results are fully generalizable to the epilepsy population. Moreover, the limited sample size precluded more precise analysis of PAP machine and pressure settings and their associated respiratory outcomes. Although some patients (n = 8) in our study have had a split-night study and pressure titrating, the standard of practice in the United States typically is to use auto-adjusting PAP or autoadjusting bilevel PAP. Currently, in the United States, many of the testing is insurance driven, and in the past several years most insurance providers have refused to pay for a titration test unless the patient has failed autoadjusting PAP. In some of the patients, pressure adjustments have been made based on other factors (subjective complaints of discomfort with pressure, or the high residual AHI). We could speculate that in some patients with epilepsy, there was an adjustment of pressure that led to a lower AHI and better adherence, although this cannot be sufficiently tested with our numbers.

Nevertheless, our study provides a good understanding of PAP adherence in a chronic neurology clinic population, and highlights the need for careful monitoring of PAP use in patients with epilepsy. It supports clinical decisions to perform repeated PAP titration studies in patients whom treatment-emergent central apneas or complex apneas are suspected, such as patients with high residual AHI despite adequate treatment. Further prospective studies are needed to explore the ways to improve PAP adherence and treat OSA more efficiently in patients with epilepsy.


All authors have seen and approved the manuscript. Dr. Latreille receives support from the Canadian Institutes of Health Research. Dr. Pavlova receives research support from Biomobie Inc and Lundbeck Inc. The authors report no conflicts of interest.



apnea-hypopnea index


body mass index


non-rapid eye movement


obstructive sleep apnea


positive airway pressure




oxygen saturation



Lin Z, Si Q, Xiaoyi Z. Obstructive sleep apnoea in patients with epilepsy: a meta-analysis. Sleep Breath. 2017;21(2):263–270. [PubMed]


Sivathamboo S, Perucca P, Velakoulis D, et al. Sleep disordered breathing in epilepsy: Epidemiology, mechanisms, and treatment. Sleep. 2018;41(4). [PubMed]


Devinsky O, Ehrenberg B, Barthlen GM, Abramson HS, Luciano D. Epilepsy and sleep apnea syndrome. Neurology. 1994;44(11):2060–2064. [PubMed]


Cao MT, Sternbach JM, Guilleminault C. Continuous positive airway pressure therapy in obstructive sleep apnea: benefits and alternatives. Expert Rev Respir Med. 2017;11(4):259–272. [PubMed]


Weaver TE, Grunstein RR. Adherence to continuous positive airway pressure therapy: The challenge to effective treatment. Proc Am Thorac Soc. 2008;5(2):173–178. [PubMed Central][PubMed]


Pornsriniyom D, Shinlapawittayatorn K, Fong J, Andrews ND, Foldvary-Schaefer N. Continuous positive airway pressure therapy for obstructive sleep apnea reduces interictal epileptiform discharges in adults with epilepsy. Epilepsy Behav. 2014;37:171–174. [PubMed]


Iber C, Ancoli-Israel S, Chesson AL Jr, Quan SF; for the American Academy of Sleep Medicine. The AASM Manual for the Scoring of Sleep and Associated Events: Rules, Terminology and Technical Specifications. 1st ed. Westchester, IL: American Academy of Sleep Medicine; 2007.


Johns MW. A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep. 1991;14(6):540–545. [PubMed]


Benjamini Y, Hochberg Y. Controlling for the false discovery rate: A practical and powerful approach to multiple testing. J R Stat Soc Ser B (Methodol). 1995;57(1):289–300


American Academy of Sleep Medicine. International Classification of Sleep Disorders. 3rd ed. Darien, IL: American Academy of Sleep Medicine; 2014.


Vendrame M, Auerbach S, Loddenkemper T, Kothare S, Montouris G. Effect of continuous positive airway pressure treatment on seizure control in patients with obstructive sleep apnea and epilepsy. Epilepsia. 2011;52(11):e168–e171. [PubMed]


Rai D, Kerr MP, McManus S, Jordanova V, Lewis G, Brugha TS. Epilepsy and psychiatric comorbidity: a nationally representative population-based study. Epilepsia. 2012;53(6):1095–1103. [PubMed]


Sawyer AM, Gooneratne N, Marcus CL, Ofer D, Richards KC, Weaver TE. A systematic review of CPAP adherence across age groups: Clinical and empiric insights for developing CPAP adherence interventions. Sleep Med Rev. 2011;15(6):343–356. [PubMed Central][PubMed]


Wells RD, Freedland KE, Carney RM, Duntley SP, Stepanski EJ. Adherence, reports of benefits, and depression among patients treated with continuous positive airway pressure. Psychosom Med. 2007;69(5):449–454. [PubMed]


Valentin A, Subramanian S, Quan SF, Berry RB, Parthasarathy S. Air leak is associated with poor adherence to autoPAP therapy. Sleep. 2011;34(6):801–806. [PubMed Central][PubMed]


Latreille V, Abdennadher M, Dworetzky BA, et al. Nocturnal seizures are associated with more severe hypoxemia and increased risk of postictal generalized suppression. Epilepsia. 2017;58(9):e127–e131. [PubMed Central][PubMed]


Pavlova M, Singh K, Abdennadher M, et al. Comparison of cardiorespiratory and EEG abnormalities with seizures in adults and children. Epilepsy Behav. 2013;29(3):537–541. [PubMed]