Eighteen research papers were included in this review, and presented in 4 groups. Group 1 consists of 6 studies since 2000 that included PSG data in exclusively DMD cohorts (Table 1). One of the 4 studies involved portable monitoring (PM). Group 2 consists of 3 early studies by Smith et al. from 1985–1989. Group 3 consists of 5 studies whose cohorts had varying NMDs, including DMD. One of the 5 studies involved PM. Group 4 consists of 3 studies involving PM, 2 in exclusively DMD cohorts, and the other with a cohort of varying NMD. The final article reviewed and presented is a consensus statement by the American Thoracic Society (ATS) regarding sleep-disordered breathing in DMD.
Duchenne muscular dystrophy (DMD) literature with polysomnography (PSG) data.
Duchenne muscular dystrophy (DMD) literature with polysomnography (PSG) data.
Group 1: Recent Studies with Exclusively DMD Cohorts
Suresh and coworkers performed a 5-year retrospective analysis where DMD patients seen in a neuromuscular clinic were routinely offered an appointment to a sleep medicine clinic.1 Thirty-four DMD patients were included—33 male and 1 female who was identified as a “manifesting female patient.” Forced vital capacity (FVC) percent predicted median was 54%, indicating severe respiratory disease in this cohort. Spirometry was not predictive of either SDB or hypoventilation.
Eight of 34 had daytime symptoms (“headaches, lethargy, or daytime sleepiness”); and 22/34 had nighttime symptoms (“snoring, or sleep disturbance”). There was no significant association between symptoms (either daytime or nighttime) and SDB (either OSA or hypoventilation).
Thirty-two of 34 DMD patients went on to have PSG with transcutaneous capnography (TC-CO2), age range was 2–16, with median of 10. Fifteen of 32 had normal PSG; 10/32 were classified as OSA using total sleep time (TST) apnea-hypopnea index (AHI) criteria of ≥ 5. Seven of 32 had hypoventilation, defined as AHI ≥ 5, and TC–CO2 > 55 mm Hg, with no hypoventilation duration criteria specified.
Five patients were started on continuous positive airway pressure (CPAP) therapy, 4 on bilevel positive airway pressure (BPAP) therapy, and 3/5 CPAP users were eventually switched to BPAP as disease progressed. Compliance was not formally assessed, but all patients on positive airway pressure therapy reported nightly use.
In the opinion of this author, using the American Academy of Sleep Medicine (AASM) Scoring Manual and International Classification of Sleep Disorders, Third Edition (ICSD-3) pediatric criteria for both OSA and hypoventilation, would have most certainly increased the number of children diagnosed with OSA and/or hypoventilation. In addition, given inevitable respiratory worsening in DMD, initiation with BPAP rather than CPAP may avoid the need for device switching later.
Sawnani et al. performed a retrospective study of 110 male DMD patients on corticosteroid therapy, with all undergoing PSG with capnography using both end-tidal capnography (ETCO2) and TC-CO2. Age range was 5–18, with mean of 11.4. SDB and hypoventilation were scored using pediatric AASM Scoring Manual and ICSD-3 criteria.2
Mean FVC was 79.82%, and mean forced expiratory volume in 1 second (FEV1) was 78.05%, both indicating this cohort had less severe respiratory disease than seen in the Suresh cohort. Lower FVC, mean inspiratory pressure (MIP), mean expiatory pressure (MEP), and peak cough flow (PCF) were associated with hypoventilation, compared to OSA; but no spirometric thresholds for increased risk of either SDB or hypoventilation were provided.
OSA was diagnosed in 70/110, tended to be REM predominant, and was more common than sleep related hypoventilation (18/110) or central sleep apnea (37/110). OSA was strongly correlated with body mass index (BMI), which is of specific relevance for those on corticosteroid therapy to slow deterioration, as such therapy may promote weight gain.
Both Suresh and Sawnani showed that prolonged deterioration as defined by spirometry is not required for the development of hypoventilation in DMD. In Suresh, hypoventilation was interpreted as occurring at older ages compared to OSA, but significant overlap was clearly evident in the data; while in Sawnani, hypoventilation was interpreted as occurring at similar ages as OSA.
Hukins et al. prospectively assessed 19 severely affected male DMD patients, whose mean FEV1 was 28.6%, PSG included TC-CO2, spirometry, and arterial blood gas (ABG).3 All patients were ≥ 12 years of age, with mean age 18.6. Hypoventilation was assessed using hypoxic burden (TST with oxygen saturation < 90%) ≥ 2%, rather than capnography, with hypoventilation found in 11/19. More patients may have been identified with hypoventilation if it had been defined using a TC-CO2 cutoff > 50 mm Hg rather than hypoxic burden. Four of 19 had TST AHI ≥ 4 using adult duration criteria, rather than pediatric criteria. FEV1 < 40% was 90% sensitive, and 50% specific for hypoxic burden ≥ 2%. Neither FEV1 nor FVC correlated to AHI.
Nozoe et al. performed a retrospective analysis of 44 DMD patients, all of whom underwent PSG with ET-CO2 capnography.4 Age range was 7–19, with mean of 13.2. Seventeen of 44 were ≤ 12 years of age, and 27/44 were between 13–19. Obstructive apneas were scored using pediatric duration criteria requiring ≥ 2 missed breaths. Hypopneas and central apneas were also scored, but it is unclear whether AASM adult criteria or pediatric criteria were utilized, as this was not explicitly stated. Arbitrary criteria, not based on current ICSD-3 criteria, were used to identify those with OSA: AHI > 1.5 for children ages ≤ 12, and AHI ≥ 5 for children ages 13–19. Eight of 44 met criteria for OSA, but stratification based on criteria utilized was not provided. Hypoxic burden was also assessed, with 6/44 having oxygen saturation < 90% for > 5 minutes of REM sleep, and 4/44 having oxygen saturation < 90% for > 5 minutes of NREM sleep, but it is not clear whether these patient groups overlapped.
Hypoventilation was assessed with ET-CO2 capnography, using the criterion of > 50 mm Hg for > 10% of total sleep time; no patients met this threshold. This criterion differs from both AASM adult hypoventilation duration criterion of 10 minutes of total sleep time and the AASM pediatric hypoventilation duration criterion of > 25% total sleep time. Spirometric assessment of lung function was not provided.
Polat et al. retrospectively assessed 12 DMD patients, all of whom underwent diagnostic polysomnography without capnography, with sleep staged according to Rechtschaffen and Kales criteria.5,6 Mean age was 10.3. Respiratory events were scored according to AASM adult criteria. Seven of 12 had scoliosis, and 5/12 were wheelchair bound. OSA was reported in 2/35 patients, both were wheelchair bound; criteria for OSA were not specified. Spirometric assessment of lung function was not provided.
Group 2: Early Work in DMD Associated OSA
In 1988, Smith et al. performed 2 consecutive nights of PSG in 14 DMD patients, with first night used to acclimate and 2nd night PSG data presented.7 Mean TST AHI was 9.6 across all patients, with 83% of events across all patients occurring in REM sleep. Nine of 14 exhibited oxygen desaturations > 5% with mean oxygen saturation nadir of 74.2%, and 8/9 exhibiting oxygen desaturations > 10%. Almost all desaturations were confined to REM sleep. Five of 14 did not exhibit oxygen desaturations, with mean oxygen saturation nadir in these patients at 93.2%. Neither blood gases nor spirometry predicted who would have SDB and/or sleep hypoxemia, though lung volumes tended to be lower in the 9/14 with oxygen desaturation.
Smith et al. continued their work, this time evaluating 7 patients with DMD with normal ABG, using 3 consecutive nights of PSG. Supplemental oxygen was not used on the first night and then utilized on either the second or third night, with a flow rate of 2 liters per minute.8 When breathing room air, 6/7 had oxygen desaturations > 5%; for 5/6 the oxygen desaturations were exclusively in REM, and for 1/6 the oxygen desaturations occurred in both NREM and REM. With supplemental oxygen only 1/7 had oxygen desaturation > 2.5%. Supplemental oxygen increased hypopnea mean duration by 19%, with duration increase greatest during REM. Oxygen supplementation was reported as not affecting apneas or hypopneas frequency, but AHI data were not presented.
Given increased respiratory event duration with supplemental oxygen use, Smith et al. further explored the effect of oxygen on ventilation in DMD.9 In a study of 6 patients with advanced DMD and limited body movement, chest and abdomen respiratory inductance plethysmography belts across then were calibrated and quantified to produce an estimate of minute ventilation across 6–10 minute segments of wakefulness, N2 sleep, N3 sleep, and REM sleep. The authors argued that this method of analysis was more representative of ventilation status, as opposed to analysis of individual respiratory events, but basic PSG data such as TST AHI were not presented.
Minute ventilation during wakefulness in DMD was similar to normal control subjects, but fell below normal values in NREM, and fell even further in REM sleep. During sleep DMD patients were not able to maintain minute ventilation, despite normal daytime ventilation.
Compared to room air, supplemental oxygen at unspecified flow rates resulted in increased proportion of REM sleep occupied by apnea-hypopneas, but none of the following was presented: mean length of individual respiratory events, respiratory event breakdown (i.e., apneas versus hypopneas), and REM AHI. No oxygen desaturations were noted on supplemental oxygen.
In the patients with DMD, vis-à-vis control subjects, minute ventilation in wake, N2, N3, and REM was similar on both room air and oxygen. The authors concluded that though supplemental oxygen may prolong individual hypopneas, as demonstrated in their prior study, it had no effect on overall ventilation.
Group 3: Studies with Varying NMD, Including DMD
Five studies evaluated PAP ventilation and/or sleep in patients with an array of NMD, including a subset of patients with DMD. One study is presented in the section on PM. The remaining 4 studies are described below.
Guilleminault et al. prospectively evaluated 20 patients with NMD utilizing PSG with capnography (either ET-CO2 or TC-CO2), followed by Multiple Sleep Latency Test (MSLT); followed one month later by BPAP (bilevel positive airway pressure) titration and next-day MSLT.11 Approximately one year later, repeat PSG with BPAP followed by MSLT was performed. Three of 20 patients had DMD, other diseases included myotonic dystrophy, mitochondrial myopathy, glycogen storage disease, and posttraumatic injury. Across all patients, age range was 12–51, with mean age 32.7; DMD age range was 12–14. One of 3 DMD patients did not continue in the study past the first PSG, the remaining 19/20 patients accepted BPAP therapy in either spontaneous or timed mode.
The mean respiratory disturbance index (RDI) on the first PSG was 28.70, with all 20 patients having RDI > 5; with DMD RDI range 17–29. Prominent central apneas were noted across all patients, with mean central apnea percentage at 78% of all respiratory events. On the second and third PSG, the central apnea percentage was 100% for all patients. The percentage of NMD patients with central apnea was higher than reported in Sawnani et al., who found central apneas in only 37/110 DMD patients, and may indicate lack of effort belt sensitivity in detecting decreased chest/abdominal effort seen in NMD, i.e., pseudo-central apnea.
Mean sleep latency (SL) on first MSLT was 8.18 minutes, with SL in the 3 DMD patients ranging from 7–9.25 minutes. Mean SL on titration was 12.47 minutes, and after one year was 12.12 minutes. Though habitual sleep patterns were not assessed prior to MSLT, and BPAP compliance was not provided, PAP appeared to immediately improve sleepiness, with sustained response after one year of BPAP.
Ragette et al. prospectively evaluated 42 NMD patients, all underwent PSG with TC-CO2, sitting/supine spirometry measured by hand-held device, and ABG prior to PSG.12 Ten of 42 had DM, other diseases included: congenital muscular dystrophy, acid maltase deficiency, nemaline rod myopathy, and myo-tonic dystrophy. Mean age across all patients was 28.7, with age range 14–63. SDB, scored using adult AASM scoring criteria, was found in 33/40 patients, and 2/10 with DMD. Predictive thresholds for inspiratory vital capacity (IVC), were computed for 3 stages of SDB: SDB onset (TST RDI > 5, or REM RDI > 10), nocturnal hypoventilation (TC-CO2 > 50 mm Hg for > 50% of TST), and diurnal respiratory failure (PaCO2 > 45 mm Hg before PSG, and nocturnal hypoventilation on PSG). IVC thresholds of < 60%, < 40%, and < 25% have both high sensitivity and specificity in predicting SDB onset, nocturnal hypoventilation, and diurnal respiratory failure respectively (Table 2). Sitting/supine hand-held spirometry is a low cost screening tool, with IVC < 60% indicating predisposition to SDB. Though PSG with capnography is always advisable in those with DMD, it is especially useful when IVC is < 40%.
Hand-held spirometry inspiratory vital capacity (IVC) thresholds for sleep-disordered breathing (SDB) severity in Duchenne muscular dystrophy.
Hand-held spirometry inspiratory vital capacity (IVC) thresholds for sleep-disordered breathing (SDB) severity in Duchenne muscular dystrophy.
Diaphragm weakness was also assessed; defined as > 20% postural fall in IVC from sitting to supine, combined with development of either dyspnea, thoraco-abdominal paradox, or accessory respiratory muscle activity, it was found in 11/42, independent of degree of ventilator weakness.
In a subsequent study, the same German group prospectively evaluated daytime predictors of SDB exclusively in children and adolescents with a range of NMD.13 Forty-nine children, mean age 11.3, were assessed using non-validated questionnaires, ABG, and PSG with TC-CO2. Seven of 49 had DMD, other diseases included: congenital muscular dystrophy, spinal muscular atrophy, limb girdle muscular dystrophy, acid maltase deficiency, nemaline rod myopathy, centronuclear myopathy, and hereditary motor sensory neuropathy.
The 10-item questionnaire, answered on a 10-point Likert scale, assessed the following: (1) quality of sleep, (2) nocturnal breathing problems, (3) nocturnal sweating, (4) morning headaches, (5) appetite, (6) concentration, (7) mood, (8) daytime function and general well-being, (9) frequency of chest infections, and (10) dyspnea. Higher scores indicated more severe symptoms, and maximum score was 100. Total symptom score correlated with RDI, but not PaCO2 or spirometry.
Definitions of SDB, nocturnal hypoventilation, and diurnal respiratory failure were the same as in the prior study. Thirty-three of 49 had SDB, 2/7 with DMD having SDB. Six of 49 had nocturnal hypoventilation across NREM and REM; 4/49 had hypoventilation in REM only. PaCO2 threshold of > 40 mm Hg had 92% sensitivity and 72% specificity for nocturnal hypoventilation. The authors suggested that if PaCO2 is > 40 mm Hg, then PSG with capnography should be considered, even if IVC is > 40% predicted.
As in the prior study, IVC < 60% was determined to be the best predictor of SDB, while IVC < 40% was the best predictor of nocturnal hypoventilation. IVC threshold sensitivities and specificities for SDB and nocturnal hypoventilation were slightly higher in children/adolescents, compared to adolescents/adults (Table 2).
Ward et al. evaluated patients with a range of NMD for nocturnal hypoventilation using sleep studies with TC-CO2.14 Patients with daytime normocapnia on ABG and nocturnal hypercapnia on PSG, defined as peak TC-CO2 > 48.7 mm Hg (6.5 kPa) were randomized to either noninvasive ventilation (12 patients), or no ventilation (14 patients). Noninvasive ventilation (NIV) modality was not explicitly stated. DMD accounted for 2/12 patients in the treatment group, and 3/14 in the control group. The mean age in the treatment group was 21, with range 7–46. Mean age in the control group was 26, with range 11–51. No age ranges were provided for DMD patients. Other diseases included: Becker muscular dystrophy, congenital muscular dystrophy, hereditary sensory or motor neuropathy, limb girdle muscular dystrophy, and congenital kyphoscoliosis.
Both groups had sleep studies with capnography every 6 months for 2 years. Treatment of nocturnal hypoventilation, even in the absence of daytime hypercapnia, provided respiratory benefits for those with NMD. Across 5 sleep studies NIV group showed significant improvement in mean SaO2, and significant decrease in mean percentage of night spent at TC-CO2 ≥ 48.7 mm Hg. Of the 10 control patients who completed 24 months of evaluation, 9/10 eventually required NIV during the course of the study for a variety of reasons, including development of daytime hypercapnia, worsening symptoms of nocturnal hypoventilation, increasing TC-CO2, poor weight gain, or pneumonia.
Group 4: Portable Monitoring
Three studies to date have studied PM in DMD patients. PM has been categorized by the AASM into 4 types based primarily on channels utilized. Type 1 is an attended in-laboratory full PSG with ≥ 7 channels. Type 2 is an unattended full PSG with ≥ 7 channels. Type 3 utilizes a limited channel device with 4–7 channels. Type 4 also utilizes a limited channel device but with only 1–2 channels, and one of the channels is oximetry.
In 1993 Khan and Heckmatt performed the first assessment of DMD patients with portable monitoring, performing 2 consecutive nights of 8-channel, Type 2, unattended home polysomnography in 21 non-ambulatory DMD patients, with sleep scored according to Rechtschaffen and Kales criteria.10 The device channels utilized included a single channel right central derivation electroencephalogram, bilateral electro-oculogram channels, submental electromyogram, chest/abdominal effort belts, nasal/oral thermistor, and pulse oximetry. Vital capacity was also measured in sitting and supine positions. The number of patients meeting criteria for obstructive sleep apnea was not stated, but the mean apnea index was 7.5 across all recordings, with a range of 3.5–11.4. The relation between vital capacity and sleep-disordered breathing indices was not mentioned. Hypoxic burden was assessed, and the average percentage of total sleep time with oxygen saturation below 90% was 2.4%. There was no correlation between hypoxemia, and vital capacity, or fall in vital capacity when supine.
In Kirk et al., 10 boys with DMD, age range 9–21, were prospectively studied with 2 consecutive nights of PSG, and 1 night of home study with simultaneous recording from 2 different PM devices.15 The first PM was a Type 4 device with pulse oximetry, and snore monitoring, which used a proprietary oxygen saturation analysis algorithm to score respiratory events. The second PM was a Type 3 device including nasal/oral thermistor, pulse oximetry, electrocardiogram, chest/abdominal effort belts, and snoring monitors.
Most patients did not have end-stage respiratory disease, with FVC < 80% in 7/10, and < 50% in only 2/10. Though no significant differences in respiratory parameters were noted between the 2 nights of PSG, it was unclear whether presented PSG data were from first night recording, second night recording, or if it varied from patient to patient.
Eight of 10 had AHI > 1, 3/8 had an AHI > 3. In the 8 patients with AHI > 1 on PSG, the AHI range was 2.4–33.6. Type 3 devices identified significant SDB in 3/8, with an RDI range of 2.3–19.3; and Type 4 devices identified significant SDB in 8/8, with an RDI range of 3.0–39.0. Type 4 studies also identified 2 patients with normal AHI on PSG as having mild SDB. Though definitive conclusions are difficult to draw given small sample size, it appeared Type 4 studies may have a higher rate of false positivity in DMD, while Type 3 studies may have a higher rate of false negativity in DMD.
Though TC-CO2 was monitored on PSG, TST with TCCO2 > 50 mm was not presented, making it difficult to assess the utility of PM in detecting hypoventilation across the night in DMD. TC-CO2 maximum was > 50 mm Hg in 4/10, with 2/4 having moderate to severe OSA (AHI of 16.8 and 33.6), that was detected on both PM devices. Two of 4 had AHI < 1, normal Type 3 study, and mild SDB on Type 4 device, with RDI of 2.4 and 6.1.
Labanowski et al. prospectively assessed SDB in 60 NMD patients: 50 adults, 10 children.16 Eight of 60 had DMD, other diseases included: myasthenia gravis, myotonic dystrophy, inflammatory myopathies, unspecified myopathies, amyotrophic lateral sclerosis, and hereditary motor sensory neuropathy. All 60 patients had Type 3 PM studies performed, with the device consisting of nasal/oral thermistor, pulse oximetry, chest effort belts, electrocardiogram, and position sensor. The accuracy of the Type 3 study was not compared to in-lab PSG.
Most the studies were performed at home, but some were performed in a lab setting; the precise breakdown was not provided. The incidence of SDB was high in this cohort, with RDI ≥ 5 in 50/60; and ≥ 15 in 25/60; specific breakdown by disease was not provided. Strong correlation was not found between RDI and any of the following: spirometry, age, BMI, or Epworth Sleepiness Scale score. There was also no correlation of RDI with a 5-point disability classification [(1) climb stairs without support, (2) requires ambulatory assistance with stairs only, (3) requires assistance with any ambulation, (4) not ambulatory, capable of unassisted transfer, and (5) not ambulatory, requires transfer assistance].
ATS Consensus Statement
In response to DMD patient insurance coverage denials purportedly based on lack of literature, the ATS published a consensus statement regarding respiratory care of DMD patients.17,18 ATS recommended annual PSG with continuous CO2 monitoring. If PSG is not available then overnight pulse oximetry with CO2 monitoring is suggested, though it is less sensitive than PSG for detecting sleep-disordered breathing without associated oxygen desaturation or CO2 retention. Morning capillary, rather than arterial, blood gas may be used to demonstrate overnight CO2 retention, but it is less sensitive than capnography. PSG or nocturnal oximetry with capnography was recommended as part of scoliosis repair surgery pre-operative assessment as it will identify those with SDB requiring PAP before surgery, and such patients may be placed on PAP following post-surgical extubation.
Regarding PAP therapy, serial evaluation is recommended as pressure requirements may change as the disease progresses. Also regular PAP follow-up to assess mask fit was suggested, as it allows opportunity to asses for PAP use associated complications, such as eye irritation from pressure leaks, skin breakdown from over tightening of straps, and gastric distention from air swallowing. Oxygen supplementation without ventilator support was discouraged, as hypoxemia is almost always a manifestation of hypoventilation.
CONCLUSION AND RECOMMENDATIONS
A “PSG for all” policy may be the most effective screening plan in those with DMD as the absence of typical SDB associated daytime or nighttime symptoms is not sufficient screen for SDB in DMD; and prolonged deterioration, as indicated by daytime spirometry, is not necessary to develop nocturnal hypoventilation. Though consistent spirometric thresholds predictive of SDB or hypoventilation are not available, IVC appears to be both sensitive and specific for SDB at < 60%, and for nocturnal hypoventilation at < 40%.
PSG with capnography, scored with AASM age appropriate criteria, is the gold standard in assessing SDB and hypoventilation in DMD. Using adult criteria for SDB and hypoventilation in children with DMD may result in underestimation of SDB burden. Obstructive apneas and hypopneas may predominate early in DMD associated SDB, but as the disease progresses individual respiratory events may become less circumscribed as flow limitation becomes more persistent, thereby making capnography critical in assessing SDB burden. TC-CO2 capnography, rather than ET-CO2, may provide the most accurate assessment of possible nocturnal hypercapnia, given limitation of ET-CO2 with poor respiratory effort.
This author disagrees with ATS consensus statement suggesting that overnight pulse oximetry with CO2 monitoring is a reasonable substitute for PSG; as this screening modality has limited ability to detect sleep-disordered breathing not associated with either oxygen desaturation or hypercapnia. The utility of PM screening for SDB or hypoventilation in DMD, or other NMD, is not clear from the available literature. To date no studies have assessed the utility of PM with capnography in NMD. The 2007 AASM Task force guidelines indicate that though PM is an option for patients with limited mobility, as is often the case for those with NMD, PM is not appropriate for the diagnosis of OSA in patients with NMD, as limited data exist regarding the diagnostic accuracy of in this population.19 A single study of 53 adults compared PSG to Type 3 PM with ET-CO2 monitoring, electrocardiogram, pulse oximetry, and respiratory effort belts. Though this study included no patients with NMD, results from Type 3 PM with ET-CO2 were comparable to PSG.20
Though serial PSG is suggested by ATS, no specific guidelines are provided regarding when to order repeat testing. Given potential logistical difficulties associated with repeated in-lab PSG, this author feels repeat PSG is warranted if the patient has clinically deteriorated vis-à-vis: declining muscle strength; development of new daytime/nighttime symptoms typical of SDB; declining respiratory status, as measured by spirometry; or increasing residual AHI on positive airway pressure device compliance download.
This author agrees with the ATS statement regarding use of supplemental oxygen, given that supplemental oxygen may actually increase hypopnea length in DMD. Regarding the choice of positive airway pressure device modalities to treat SDB and/ or hypoventilation noted on PSG, initiation of BPAP rather than CPAP may prevent the need for device switching to BPAP as disease inevitably progresses, though CPAP may adequately treat SDB and/or hypoventilation early in the disease course.