Outcome of noninvasive ventilation in children with neuromuscular disease

Noninvasive ventilation (NIV) is an effective symptomatic treatment for adults with neuromuscular respiratory failure.¹ The optimal timing of initiation of NIV and its impact on quality of life (QOL) in children are unknown.^2–4 In this study, we sought to establish the effect of institution of NIV on clinical outcome and QOL in a cohort of children with severe neuromuscular disorders.

Methods.

NIV was defined as the delivery of ventilatory support via a nasal or face mask interface without the need for an invasive artificial airway.

Inclusion criteria included all patients commencing nocturnal NIV for respiratory compromise associated with neuromuscular disorders at The Children's Hospital at Westmead during the period January 1, 1994, to January 1, 2004. Toddlers and children who had severe cognitive deficits rendering them unable to communicate were excluded from the QOL analysis, but included in the remainder of the study. Records were reviewed and data obtained from the year prior to institution of NIV and annually thereafter. The study had institutional ethics committee approval.

Data obtained included diagnosis, age at institution of NIV, nature of NIV, mortality, patient symptoms, growth centiles, presence of scoliosis, results of pulmonary function testing, overnight oximetry or polysomnography (PSG), number of hospitalizations, number of days in hospital and intensive care, and costs to the health service. Pulmonary function testing was performed annually in all children able to cooperate, in a seated or erect position, using a Sensormedics Vmax Spirometer.

Sleep-disordered breathing (SDB) was assessed using questionnaires and overnight study data. Details of symptoms were obtained from clinical records and questionnaires completed prior to PSG and before and after initiation of NIV. PSG data were obtained using the Compumedics S-series polysomnographic system. Data collected included frequency and duration of apneas and hypopneas, maximum transcutaneous CO₂ (TCO₂), minimum O₂ saturation, average O₂ desaturation, respiratory disturbance index (RDI), and RDI in REM sleep.⁵ The RDI was defined as the number of abnormal respiratory events (apneas and hypopneas) per hour of sleep. PSG was performed in the year prior to NIV and annually thereafter.

Direct costs associated with outpatient visits and hospital admissions prior to institution of NIV were calculated. These were obtained from the finance departments of the tertiary hospital, the patients' local hospitals, and the ambulance service. Additional costs after NIV included equipment and specialized nursing costs. Total costs were underestimated as visits to general practitioners, medications obtained externally, and indirect costs (such as loss of parental income) were not included.

The QOL study described in this paper was carried out using the Pediatric Quality of Life Inventory (PedsQL), developed by Dr. James W. Varni. Parents and children evaluated QOL in several different areas using both the PedsQL version 4.0 (PedsQL 4.0) generic and neuromuscular modules.^6,7 Different versions of the PedsQL were used in the following age groups; toddler, 0 to 2 years; young child, 3 to 7 years; child, 8 to 11 years; and teenager, 12 to 18 years. Scores were converted to a 0 to 100% scale, with higher scores indicating higher QOL. In order to establish the effect of NIV on QOL, parents completed two sets of questionnaires: one recalling what life had been like for their child prior to commencing NIV and one addressing their QOL after institution of NIV.

Statistical analysis was performed using SPSS for Windows version 11. The Wilcoxon nonparametric related samples test was used to compare means of categorical data including patients' symptoms, QOL, and health care costs before and after NIV. Repeated-measures analysis of variance assessed changes in PSG data over time. Differences in PSG data in patients with and without symptoms were compared using the Mann-Whitney U test for nonparametric unrelated samples. Intraclass correlation was used to measure correlation between child- and parent-reported QOL.

Results.

Seventeen patients were eligible for inclusion. Fourteen children (82%) were enrolled. Diagnoses included spinal muscular atrophy type 2 (six children), Duchenne muscular dystrophy (three children), congenital myopathy (one child), merosin-deficient muscular dystrophy (one child), Arts syndrome (one child), and myotonic dystrophy (one child). Three refused or had been transferred to adult units. Nine subjects were girls. Three patients commenced NIV between 1994 and 1999. Eleven patients commenced NIV between 2000 and 2004. The median age of patients at initiation of NIV was 7.7 years (range 1.5 to 16 years) and reflected the underlying disease (see table E-1 on the Neurology Web site at www.neurology.org). The median length of follow-up was 30 months (range 6 to 84 months).

Symptoms and QOL scores.

Nine patients had daytime somnolence prior to NIV. Four also had morning headaches. Following NIV, no children had daytime sleepiness (p = 0.003) and only one continued to have headaches (p = 0.046) (figure). Although several PSG indices also improved, no association was found between the abnormalities in PSG data and symptoms of SDB. Before NIV, seven patients had anorexia and four had failure to thrive (weight below the third centile for age). After NIV, only three patients had anorexia, but five had failure to thrive.

• Download figure
• Open in new tab
• Download powerpoint

Figure. Patients' symptoms before and after noninvasive ventilation (NIV).

Five patients experienced adverse effects related to NIV. Two children reported finding the mask initially uncomfortable but required no change in treatment. Three children developed skin breakdown over the nasal bridge, necessitating dressings or change to a full-face mask. One child experienced nausea, and one developed midface hypoplasia after 4 years of NIV.

One patient, a 16-year-old boy with Duchenne muscular dystrophy, developed acute respiratory failure during a respiratory infection and died despite full resuscitative efforts. PSG and echocardiography performed while on NIV 6 months prior to his death were within normal limits.

Eleven (79%) parents completed parent report forms from both QOL questionnaires before and after NIV. There was no significant deterioration in parent-reported QOL in any of the subscales after NIV (table E-2). Three children (ages 8 to 12) and four teenagers (ages 13 to 18) completed child report forms after NIV only. The PedsQL 4.0 generic total QOL for children ranged from 32% to 77.9%, and the generic total QOL for teenagers ranged from 32.1% to 79.1% Actual PedsQL 4.0 QOL child report scores are shown in table E-3. Published PedsQL QOL values for children with other chronic illness and children with spinal muscular atrophy (SMA) are given in tables E-2 and E-3 for comparison. Intraclass correlation between the child and adult reported QOL was 0.94, p = 0.002. The lowest PedsQL 4.0 QOL scores were consistently in the physical health domains for both parent- and child-reported QOL

Discussion.

Institution of NIV in children with neuromuscular disease resulted in a reduction in hospital and intensive care unit admissions and a 73% reduction in health care costs over the year after treatment, without adversely affecting the QOL of the patient or family. Improvements were also observed in patients' symptoms and PSG indices.

The reduction in hospital admissions was seen in all subgroups, including those children with normal PSG data prior to emergent ventilation. Hospitalizations may have been further decreased by earlier institution of NIV. This has not been well studied in children with neuromuscular disorders, the only previous such study having been skewed by inclusion of a number of children with cardiac disease.^8,9

Health service costs also decreased markedly after NIV. Cost reduction may be offset by prolonged survival, but this study was not designed to demonstrate survival advantage. Health professionals have become accountable for resource allocation, particularly in patients with progressive incurable disease. It could be postulated that indirect costs (i.e., parental loss of earnings) would also decrease after NIV. QOL scores that addressed family resources did not improve after NIV, however, suggesting that earning capacity remains curtailed in these families despite a reduction in their children's hospitalizations.

As the children in this study had progressive neuromuscular disorders, it could be expected that their QOL would deteriorate with increasing age. We found, however, that QOL remained stable after NIV despite disease progression, suggesting that NIV may prevent deterioration in QOL. These results must be interpreted with caution as the QOL data prior to NIV may have been affected by recall bias. Not surprisingly, lowest QOL scores were in physical health domains, but, despite their severe physical limitations, most children considered their emotional and social well-being to be good. The considerable pressures involved in caring for a ventilator-dependent child should not be underestimated.¹⁰ Despite this, parents reported limitation of family activities with no deterioration after NIV, suggesting that NIV was not additionally detrimental to the QOL of the family unit.

Patients' symptoms of headache and daytime somnolence improved significantly after NIV. Although several PSG indices also improved, no association was found between the abnormalities in PSG data and symptoms of SDB. This disparity has been identified by other authors and remains poorly understood.¹¹

Although NIV should reduce the work of breathing and metabolic demands, rates of failure to thrive remained high after NIV, possibly because of factors such as bulbar dysfunction. Malnutrition can cause reduced ventilatory drive, impaired immune function, and respiratory muscle wasting. Nutritional monitoring is therefore important in children with chronic neuromuscular respiratory failure.¹²

NIV should ideally be offered electively in order to avoid difficult decision making in an emergency situation. The frequency of emergent initiation of NIV in this and other studies suggests that current means of assessment or monitoring of respiratory function are inadequate.¹³ Children may be too young or too intellectually disabled to report symptoms of SDB or to cooperate with PFT or may present with severe hypoventilation in the absence of previous symptoms of SDB. Future research in this patient group should address use of pulmonary function assessment techniques that do not require patient cooperation (such as forced oscillation and nitrogen washout techniques).¹⁴

PSG is considered the gold standard for the diagnosis of SDB in children,¹⁵ but does not adequately identify children whose respiratory function worsens acutely with respiratory infections. This potential for acute decompensation is suggested by a history of frequent hospitalizations or PICU admissions. Such findings should be considered when assessing children for NIV, particularly those unable to perform PFT or report symptoms of SDB. Performing PSG (or at least overnight oximetry and transcutaneous capnography) at the time of admission for respiratory tract infections may identify children in whom infection precipitates respiratory insufficiency.¹⁶ Future research in this patient group should focus on reliable diagnosis of respiratory insufficiency and the optimal timing of institution of NIV.