95% CI, Confidence interval; cmH ₂O, Centimetre of water; FEV ₁, Forced expiratory volume in 1 second; FVC, Forced vital capacity; GRADE, Grading of recommendations assessment development and evaluation; HFCWC, High frequency chest wall compression; IPV, Intrapulmonary percussive ventilation; IRR, Incidence rate ratio; MD, Mean difference; NMD, Neuromuscular disease; PRISMA, Preferred reporting items for systematic reviews and meta-analyses; QoL, Quality of life; RCT, Randomised controlled trial; RXO, Randomised cross-over study

Ineffective cough mechanisms can occur in patients with neuromuscular disease (NMD) as a result of inspiratory and expiratory muscle weakness, as well as impaired glottic function ^{1, 2} . In the long term, secretion retention leads to airway obstruction, inflammation, breathing difficulty, repeated acute respiratory tract infection, and consequently chronic lung disease and a predisposition to ventilatory failure ^{3– 6} . Maintaining clear airways is crucial in patients with NMD, because respiratory insufficiency is one of the main causes of death ⁴ .

For secretion clearance, secretion mobilisation techniques and assisted coughing techniques are recommended ⁷ . However, the standard secretion mobilisation techniques, such as postural drainage techniques, chest wall strapping, positive expiratory pressure and oscillatory positive expiratory pressure, are ineffective in very weak patients because they are effort-dependent ^{2, 8} , and these patients are generally unable to generate sufficient expiratory flow ³ . These techniques are also difficult to apply in cases of chest wall or spinal deformities, as well as osteoporotic ribs ⁹ . In these specific cases, other techniques are favoured, such as high frequency chest wall oscillations (HFCWO), high frequency chest wall compression (HFCWC) or intrapulmonary percussive ventilation (IPV) ⁹ . These methods have the benefit of working without the patient’s active participation, especially in patients with tracheostomy and/or bulbar failure, and/or intellectual impairment ³ .

Arcuri et al. conducted a systematic review of airway clearance and analysed patients with NMD. They reported that HFCWC does not improve the survival rate or the loss of FVC. In addition, HFCWC does not decrease the frequency of respiratory infections, and IPV is unsuccessful in enhancing peak expiratory flow ¹⁰ .

Even if there is a growing interest in the use of IPV and HFCWC, Arcuri et al. did not include every existing publication in their recent review ¹⁰ , probably because they included studies with a majority of adults and the severity of the chronic disease had to be identified with the National Hospice Organisation Criteria ¹¹ . It is thus of clinical relevance to make an up to date synthesis of the available evidence regarding the use of IPV and HFCWC ¹² .

We hypothesised that IPV and HFCWC might mobilise secretions, recruit obstructed areas of the lungs and prevent the negative consequences of muscle weakness related to neuromuscular disease. Hence, our main objective was to explore the effects of HFCWC and IPV, as compared with standard care or no treatment, on lung volume and capacity (as a result of secretions mobilisation), as well as quality of life (QoL) in patients with neuromuscular disease in acute or stable condition. We further assessed the effects of IPV and HFCWC on clinical value (arterial blood gases and the patient’s subjective respiratory perception of dyspnoea), complications and survival.

This systematic review followed the recommendations of the Cochrane Guidelines for systematic review of interventions as well as the recommendations of the updated Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement ¹³ . The protocol of this review was not published in a peer-review journal but it is available on PROSPERO, the registration number is CRD42017064703.

The electronic search identified 1588 records. After removing 398 duplicates, we screened 1182 titles and abstracts and 1058 records were excluded. In total, 124 full texts were screened, and 119 were excluded for the following reasons: wrong design 13, wrong intervention 4, wrong population 101, study in Hebrew 1. (PRISMA flow diagram ¹³ in Figure 1). Five studies were included according to the eligibility criteria ^{18– 22} . Three RCTs ^{18, 21, 22} studied HFCWC and one RCT ²⁰ and one RXO ¹⁹ investigated IPV.

The characteristics of the included studies are described in Table 1. We could not perform a meta-analysis due to the diversity of the reported outcomes, the limited number of studies per outcome assessments, missing data and heterogeneity among studies design. For all these reasons, the results are presented narratively.

Based on this systematic review, 104 patients in five randomised studies, including the diagnoses amyotrophic lateral sclerosis (two studies) or Duchenne muscular dystrophy (three studies, two of them with a mix of other neuromuscular diagnoses), we can report the following main findings: there is very low certainty of evidence that A) HFCWC is not superior to standard care for lung capacity, lung volume, antibiotic use, quality of life, and survival rate, B) HFCWC is superior to standard care for perception of dyspnoea, C) IPV is not superior to standard care for lung capacity, lung volume, and the risk of respiratory infection, and D) IPV is superior to standard care for the reduction of hospitalisation rate and number of days of antibiotic use.

Differences from protocol: We added the assessment of peak expiratory flow because we considered it a relevant information. In addition, we did not exclude a study if it did not contain lung function assessment, as we would have lost important articles responding to our secondary outcomes. We also decided to exclude non-randomised controlled trials.

This review has some limitations. First, it is possible that we missed unpublished studies, or studies that were only published in languages other than English and in journals not listed in the searched databases or the clinical trial registries. Because of the low number of studies, we were not able to evaluate the risk of a small study effect or a publication bias. Second, it is known that the reliability of the risk of bias tool is not very high and other authors might come to different conclusions regarding risk of bias of the include studies.

There are limitations regarding the included studies. First, the included studies were very small, had a high risk of bias, and it was not possible to perform a meta-analysis. Therefore, we only have very low confidence in our results, given the downgrading in the GRADE rating because of very serious risk of bias, serious inconsistency and very serious statistical imprecision ²⁵ . Second, our research question was the effects of HFCWC and IPV on lung volume and capacity as a result of secretions mobilisation. Airway clearance measurement trough sputum weight is inappropriate due to day-to day variability, as well as variability during the day and the fact that secretions can be swallowed ²⁶ . Therefore, our primary outcomes were lung volume and capacity as it is often used in the literature as an outcome for airway clearance and to assess the progression of the disease in patient with NMD. Many trials have reported that the use of lung function parameters is questionable because of the lack of observed changes ⁸ . Jones et al. (2006) suggest that lung function parameters are insensitive in assessing the acute effects of airway clearance techniques ²⁷ . In addition, van der Schans (2002) reports that lung function measurement does not appear to reflect differences in mucus transport or mucus expectoration ²⁸ . Thus, the ineffectiveness of the two interventions might be influenced by the inefficacy of lung function to assess airway clearance. The lack of significant results could be influenced by the inefficiency of lung function to assess airway clearance.

The results of studies using IPV and HFCWC for other pathologies show similarities and differences as compared with our results. Reychler et al. (2018) have performed a systematic review and compared IPV with other airway clearance techniques. They found a reduction in the duration of hospitalisation and have observed an improvement in gas exchange only during the exacerbation phase in patients with chronic obstructive airway disease. In patients with cystic fibrosis, no difference was observed in the static and dynamic lung volume ²⁹ . Cough is usually not impaired in patients with chronic obstructive airway disease and cystic fibrosis, thus potentially explaining the contrasting results.

In a recent publication, Chatwin et al. (2018) ³ highlight that airway clearance such as IPV and HFCWC depend on a normal cough to clear proximal airway. In patients with NMD, these interventions may be ineffective if not combined with cough augmentation technique or device. To our knowledge, no studies investigated the combination of a secretion mobilisation device such as IPV and HFCWC with cough augmentation technique or device.

Our review differs from previous ones in that we included only patients with NMD treated either with HFCWC or IPV. We could include additional studies ^{18, 20, 21} , but we still can only report very low certainty of evidence for or against the use of HFCWC or IPV for airway clearance. In one study, patients in the HFCWC groups showed substantial but statistically not significant fewer survival days compared to the standard care, which was in this case bilevel positive airway pressure. There are no other studies in any type of patients, to our knowledge, that reported increased mortality for patients treated with HFCWC. Therefore, we strongly believe that this decreased survival time in the mentioned study should not be overrated, but future studies should monitor mortality under HFCWC.

Future studies should include larger sample sizes in multi-centre trials involving international collaborations and should avoid risk of bias. The comparison treatment should avoid using a ‘no treatment’ group, and the intervention should be described precisely to facilitate comparison with other studies. We invite researchers to focus on the effects of combined treatments, such as secretion mobilisation interventions with cough augmentation technique, manual cough techniques or mechanical insufflation-exsufflation. We further encourage researchers to investigate more reliable, sensitive and patient-relevant outcome measures to assess the effects of airway clearance techniques.

In this systematic review we explored the effects of IPV and HFCWC, compared with standard care, and found no effects on lung volume and capacity, and QoL. HFCWC might decrease the perception of dyspnoea but shows no difference in the development of complications and survival. Treatment with IPV, compared with control treatment, appears to reduce the number of hospitalisation days and to lessen the need for antibiotics, but no difference was observed regarding the respiratory infection rate.

The certainty of evidence of these results is very low, and all studies presented high risk of bias. The implementation of these interventions in clinical practice should be further evaluated in clinical trials. We invite future studies to improve on these aspects, to explore the effects of combined treatments and to investigate more reliable, sensitive, and patient-relevant outcome measure.

This review was not supported financially.