Nicolle et al - 1
Manuscript # : Neurology/2005/102624
Sleep apnea in patients with myasthenia gravis
MW Nicolle MD FRCPC D.Phil.§, S Rask BSc§, WJ Koopman RN MScN§, CFP George MD FRCPC DABSM‡, J Adams MD FRCPC MSc† and S Wiebe MD FRCPC MSc¥
From the Division of Neurology, Department of Clinical Neurological Sciences§ (Dr. Nicolle, Ms. Rask, Ms. Koopman) and Division of Respirology, Department of Medicine‡ (Dr. George), University of Western Ontario, London, Ontario; the Pacific Parkinson’s Research Centre, Department of Medicine†, University of British Columbia, Vancouver, British Columbia (Dr. Adams) and the Division of Neurology, Department of Medicine¥, University of Calgary, Calgary, Alberta (Dr. Wiebe), Canada.
Corresponding author:
Michael W Nicolle MD FRCPC D. Phil.
Department of Clinical Neurological Sciences
London Health Sciences Centre
339 Windermere Rd, London, Ontario
Canada N6A 5A5
Telephone: (519) 663-3236
Fax: (519) 663-3328
Email:
Acknowledgments:
Ms. Sara Rask was awarded a Henry R. Viets Medical Student Research Fellowship award by the Myasthenia Gravis Foundation of America (2003) for this project.
Abstract
To assess the prevalence of obstructive sleep apnea (OSA) in myasthenia gravis (MG), we identified patients at risk of OSA using the multivariable apnea prediction index. OSA was diagnosed with polysomnography (PSG). The prevalence of OSA was 36%, compared to an expected prevalence of 15-20% in the general population. When including the presence of daytime sleepiness (OSA syndrome), the prevalence was 11%, compared to 3% in the general population.
Introduction
The prevalence of OSA on polysomnographic (PSG) studies in the general population is 15-20%, and 3% for OSA syndrome (OSA on PSG and excessive daytime sleepiness) (1). We performed a cross-sectional analysis to assess the prevalence of OSA in MG.
Methods
Random number generation was used to select 100 MG patients from approximately 400 patients followed in our clinic. The diagnosis of MG was made clinically and supported by positive AChR antibodies, Tensilon or electrodiagnostic testing. All patients gave informed consent (Ethics #09865E).
Demographic and basic anthropometric data including body mass index (BMI) were collected. Clinical data included: AChR antibody status, MG severity (MGFA Classification (2)) at first assessment, onset age, presence of thymoma, previous thymectomy, details of MG treatment within the previous 12 months, medical conditions potentially affecting cardiac or respiratory function and previous PSG assessments. At study onset age and MG severity were recorded, and a quantitative MG scale (QMG) assessment performed (2). Total (maximum 39) and bulbar subset (maximum 15) QMG score were recorded.
Assessments included the Epworth and Stanford Sleepiness Scale (ESS, SSS), Pittsburgh Sleep Quality index (PSQI), Apnea Symptom Frequency (ASF) and Functional Outcomes of Sleep Questionnaire (FOSQ®).
The multivariate apnea prediction (MAP) index was calculated (3). It incorporates data on apnea symptoms (snoring, snorting, loud gasping and cessation of breathing), BMI, age and gender to derive a value between 0 and 1. Patients were stratified into higher (MAP ≥0.5) and lower (MAP <0.5) risks of OSA. Patients with a MAP value of < 0.5 were not investigated with PSG. The ASF score was looked at as an independent predictor of OSA.
Patients with a MAP score of ≥0.5 were invited for PSG (Appendix (E) methods). Previous PSG had been done in 13 and was repeated in 4 with incomplete data. The OSA severity was graded mild (AHI 5-14), moderate (AHI 15-29) or severe (AHI >30) based on the apnea-hypopnea index (AHI) (1, 3). OSA syndrome was diagnosed with an AHI of >5 and ESS score ≥10.
Statistical analysis
We used means, proportions and 95%CI to describe and visually assess data. The strength of association between the main outcome variable (OSA) and predictor variables was assessed using Spearman’s Rho. Variables thought a priori to be clinically relevant between OSA-positive and OSA-negative patients were compared using unpaired t-tests for continuous data, and Chi-squared for categorical data. Significant variables in univariate analyses were used to construct a logistic regression multivariate model, with a dummy variable constructed for BMI (≤25 = 0, >25=1).
Results
Demographics of the study population were similar to MG patients in our database (n=487; data not shown). Most had relatively mild MG; 87% were asymptomatic or < MGFA II, and most were receiving combinations of AChEI, CST or azathioprine. However 25% were untreated in the previous 12 months (Table 1).
Of the 100 patients, 50 patients scored <0.5 on the MAP index and were not studied further (Figure 1). However, one had been diagnosed with OSA previously, and another was diagnosed subsequently. The remaining 48 patients were assumed not to have OSA.
Thirty-seven of the 50 patients with a MAP index of ≥ 0.5 completed PSG, 13 refused or PSG couldn’t be arranged (Figure 1). Of these 37, 34 had OSA; 10 mild, 9 moderate and 15 severe (Figure 1). Sleep-disordered breathing occurred mainly during NREM sleep. There were more obstructive hypopneic than apneic events (mean 77.5 SD 41 vs. 47 SD 70). These were concentrated in NREM sleep rather than REM (mean 104 SD 85 vs. 21 SD 22). The time in REM sleep was 15.6% SD 7.0, consistent with patients sleeping in non-familiar (i.e. laboratory) environments.
In total, 36 of 100 patients had OSA, 11 of whom had OSA syndrome (excessive daytime sleepiness and OSA). Compared to the general population, the prevalence of both was increased in MG patients (OSA 36%, 95%CI 27-45%, p < 0.001; OSA syndrome 11%, 95%CI 5-17%, p < 0.001). A multivariate analysis revealed that MG patients with OSA were more likely to be male (odds ratio [OR] 5.6, 95%CI 1.9-16.8, p < 0.002) and have a BMI >= 25 (OR 23.1; 95%CI 2.4-219.6, p < 0.006).
Compared to those not known to have OSA (n=48), MG patients with OSA (n = 36) were more likely to be male (75% vs. 29%, p<0.001), older (mean 66.1 vs. 54.6 years, p=0.001), have a greater BMI (34.1 vs. 26.1, p<0.001), be on corticosteroids within the previous 12 months (50% vs. 29%, p = 0.05) and to have apneic symptoms (ASF score 1.44 vs. 0.34; p<0.001)(Table 2). There were no significant differences in MG duration or severity (MGFA < III, QMG total or bulbar scores), frequency of cardiac or pulmonary disease, treatment except corticosteroids or of sleep questionnaire results (ESS, SSS or FOSQ scores)(Table 2).
Discussion
In the general population, the prevalence of OSA and OSA syndrome is approximately 15-20% and 3% (1), compared to at least 36% and 11% in our study. However, this likely underestimates the prevalence of OSA in MG for two reasons. First, given the high proportion (34 of 37) of patients with a MAP index ≥0.5 who had OSA, assuming a similar prevalence in the 13 high-risk patients not studied with PSG, the prevalence could be much higher. Second some patients with a MAP of <0.5 may have OSA, as this cut-off has an estimated sensitivity and specificity of 0.88 and 0.55 for OSA (3), highlighted by the two patients with a MAP <0.5 diagnosed with OSA outside of this study.
In many studies of sleep in MG, MG patients were included with other neurological disorders (4-10). Selection criteria were often unclear, potentially biasing for the selection of older more severe patients or those who reported daytime somnolence (6, 7, 9). Our prospective study is the first demonstrating that OSA, rather than just hypoventilation during sleep common in other neuromuscular disorders, is more common in MG.
This study was not designed to assess risk factors for OSA. Importantly, patients were selected for PSG based on the MAP index, which has weighted contributions from BMI, age, male gender and symptoms suggesting sleep apnea (3). Our finding that these were more common in those with OSA is not surprising. The ASF score alone is useful to predict OSA. Other possible risk factors for OSA were likely obscured by the selection bias of the MAP index. The association with recent CST use (50% in OSA+, 29% in OSA-) may be an indirect effect of CST-induced weight gain. However 25% of OSA+ MG patients were untreated, and 50% CST naive, suggesting that other MG-specific factors increased the risk of OSA. To determine this, PSG would need to be performed on MG patients without pre-selection using the MAP score.
Although the study was also not designed to study characteristics of OSA in MG, the majority of obstructive events were hypopneic and occurred in NREM sleep. This suggests that oropharyngeal weakness was more important than diaphragmatic weakness.
Our study suggests that inquiring about OSA symptoms and calculating the BMI will identify patients most at risk for OSA, important in the management of ‘fatigue’ in MG patients. Without considering a role for OSA, a history of ‘fatigue’ could lead to potentially deleterious increases in corticosteroids.
References
1.Young T, Peppard PE, Gottlieb DJ. Epidemiology of obstructive sleep apnea: a population health perspective. Am J Respir Crit Care Med 2002;165(9):1217-1239.
2.Jaretzki A, 3rd, Barohn RJ, Ernstoff RM, et al. Myasthenia gravis: recommendations for clinical research standards. Task Force of the Medical Scientific Advisory Board of the Myasthenia Gravis Foundation of America. Neurology 2000;55(1):16-23.
3.Maislin G, Pack AI, Kribbs NB, et al. A survey screen for prediction of apnea. Sleep 1995;18(3):158-166.
4.Chokroverty S. Sleep-disordered breathing in neuromuscular disorders: a condition in search of recognition. Muscle Nerve 2001;24(4):451-455.
5.Labanowski M, Schmidt-Nowara W, Guilleminault C. Sleep and neuromuscular disease: frequency of sleep-disordered breathing in a neuromuscular disease clinic population. Neurology 1996;47(5):1173-1180.
6.Amino A, Shiozawa Z, Nagasaka T, et al. Sleep apnoea in well-controlled myasthenia gravis and the effect of thymectomy. J Neurol 1998;245(2):77-80.
7.Quera-Salva MA, Guilleminault C, Chevret S, et al. Breathing disorders during sleep in myasthenia gravis. Ann. Neurol. 1992;31(1):86-92.
8.Papazian O. Rapid eye movement sleep alterations in myasthenia gravis. Neurology 1976;26(4):311-316.
9.Manni R, Piccolo G, Sartori I, et al. Breathing during sleep in myasthenia gravis. Ital J Neurol Sci 1995;16(9):589-594.
10.Sonka K, Klinderova J, Kolejakova M, Simkova L, Nevsimalova S. MESAM4 evaluated nocturnal respiration disturbances in myasthenia gravis. Sb Lek 1996;97(1):97-102.
Table 1 - Baseline characteristics of study group
Variable* (n=100) / Value / St. Dev. / RangeMean Age (yrs) / 61.4 / 16.2 / 19-89
Gender (Male/Female) / 56/44
BMI / 29.98 / 6.4
Duration MG (yrs) / 9.15 / 8.4 / 0-39
AChR ab positive (%) / 85
Thymoma (%) / 10
Treatment in 12 months preceding study (%)
None / 25
AChEI / 48
CST / 39
Aza / 38
IVIg / 5
PLEx / 4
MG severity (MGFA classification) at time of study (%)
MM/PR/CSR* / 30
I / 25
IIa / 21
IIb / 11
IIIa / 1
IIIb / 10
IVa / 1
IVb / 1
V / 0
QMG at time of study
Total score (max 39) / 6.77 / 4.61
Bulbar subset score (max 15) / 2.21 / 2.07
* Minimal manifestations, pharmacological remission or complete remission
Table 2 - Comparison of OSA- and OSA+ groups
Age (mean, yrs) / 54.6 / 66.1 / -11.5 (-18.1, -4.9) / 0.001
Males (%) / 29 / 75 / -46 (-65, -27) / <0.001
BMI / 26.13 / 34.1 / -7.97 (-10.52, -5.43) / <0.001
MGFA Initial <III (%) / 63 / 75 / -12.5 (-32.2, 7.2) / 0.21
Comorbid cardiac (%) / 17 / 25 / -8.3 (-26.0, 9.3) / 0.36
Comorbid pulmonary (%) / 10 / 3 / 7.6 (-2.5, 17.8) / 0.14
Duration MG (mean, yrs) / 9.8 / 8.1 / 1.7 (-1.7, 5.1) / 0.32
QMG Total (mean score) / 6.7 / 6.5 / 0.2 (-1.89, 2.29) / 0.85
QMG Bulbar (mean score) / 2.1 / 2.4 / -0.3 (-1.24, 0.64) / 0.53
AChE < 12 months (%) / 52 / 39 / 13.2 (-8.1, 3.5) / 0.23
Steroids < 12 months (%) / 29 / 50 / -20.8 (-41.6, -0.04) / 0.05
Cumu Dose Steroids (mean, mg) / 1065 / 1787 / -722.2 (-1742.8, 298.4) / 0.16
Azathioprine < 12 months (%) / 31 / 42 / -10.4 (-31.2, 10.4) / 0.33
No treatment < 12 months (%) / 25 / 25 / 0 (-19.0, 19.0) / 1
ESS (0-24) / 6.3 / 7.6 / -1.3 (-3.27, 0.67) / 0.19
ESS >10 (%) / 20.8 / 30.6 / -9.7 (-2.8, 9.2) / 0.31
ASF (0-4)(mean, score) / 0.34 / 1.44 / -1.1 (-1.54, -0.66) / <0.001
SSS (0-7)(mean, score) / 2.3 / 2.3 / 0.0 (-0.57, 0.57) / 1
PSQI (0-21)(mean, score) / 8 / 7.4 / 0.6 (-1.30, 2.50) / 0.53
FOSQ (5-20)(mean, score) / 17.6 / 17.5 / 0.12 (-1.30, 2.50) / 0.82
*Continuous variables: T test for Independent groups. Proportions: Chi-square with continuity correction.
Figure legends
Figure 1 – Flow chart of study population MAP and PSG results. * PSG not performed due to subject refusal or inability to attend. † independently diagnosed with OSA before (n=1) or after (n=1) study.