Health Technology Assessment (Oct 2015)
Measurement of exhaled nitric oxide concentration in asthma: a systematic review and economic evaluation of NIOX MINO, NIOX VERO and NObreath
Abstract
Background: High fractions of exhaled nitric oxide (FeNO) in the breath of patients with symptoms of asthma are correlated with high levels of eosinophils and indicate that a patient is likely to respond to inhaled corticosteroids. This may have a role in the diagnosis and management of asthma. Objective: To assess the diagnostic accuracy, clinical effectiveness and cost-effectiveness of the hand-held electrochemical devices NIOX MINO® (Aerocrine, Solna, Sweden), NIOX VERO® (Aerocrine) and NObreath® (Bedfont Scientific, Maidstone, UK) for the diagnosis and management of asthma. Data sources: Systematic searches were carried out between March 2013 and April 2013 from database inception. Databases searched included MEDLINE, EMBASE, the Cochrane Database of Systematic Reviews, the Database of Abstracts of Reviews of Effects, Science Citation Index Expanded and Conference Proceedings Citation Index – Science. Trial registers such as ClinicalTrials.gov and the metaRegister of Controlled Trials were also searched in March 2013. All searches were updated in September 2013. Review methods: A rapid review was conducted to assess the equivalence of hand-held and chemiluminescent FeNO monitors. Systematic reviews of diagnostic accuracy and management efficacy were conducted. A systematic review of economic analyses was also conducted and two de novo health economic models were developed. All three reviews were undertaken according to robust high-quality methodology. Results: The rapid review (27 studies) found varying levels of agreement between monitors (Bland–Altman 95% limits of agreement up to ±10 parts per billion), with better agreement at lower FeNO values. Correlation was good (generally r > 0.9). The diagnostic accuracy review identified 22 studies in adults (all ages) and four in children. No studies used NObreath or NIOX VERO and seven used NIOX MINO. Estimates of diagnostic accuracy varied widely. FeNO used in combination with another test altered diagnostic accuracy only slightly. High levels of heterogeneity precluded meta-analysis. Limited observations included that FeNO may be more reliable and useful as a rule-in than as a rule-out test; lower cut-off values in children and in smokers may be appropriate; and FeNO may be less reliable in the elderly. The management review identified five randomised controlled trials in adults, one in pregnant asthmatics and seven in children. Despite clinical heterogeneity, exacerbation rates were lower in all studies but not generally statistically significantly so. Effects on inhaled corticosteroid (ICS) use were inconsistent, possibly because of differences in management protocols, differential effectiveness in adults and children and differences in population severity. One UK diagnostic model and one management model were identified. Aerocrine also submitted diagnostic and management models. All had significant limitations including short time horizons and the selective use of efficacy evidence. The de novo diagnostic model suggested that the expected difference in quality-adjusted life-year (QALY) gains between diagnostic options is likely to be very small. Airway hyper-responsiveness by methacholine challenge test is expected to produce the greatest QALY gain but with an expected incremental cost-effectiveness ratio (ICER) compared with FeNO (NObreath) in combination with bronchodilator reversibility of £1.125M per QALY gained. All remaining options are expected to be dominated. The de novo management model indicates that the ICER of guidelines plus FeNO monitoring using NObreath compared with guidelines alone in children is expected to be approximately £45,200 per QALY gained. Within the adult subgroup, FeNO monitoring using NObreath compared with guidelines alone is expected to have an ICER of approximately £2100 per QALY gained. The results are particularly sensitive to assumptions regarding changes in ICS use over time, the number of nurse visits for FeNO monitoring and duration of effect. Conclusions: Limitations of the evidence base impose considerable uncertainty on all analyses. Equivalence of devices was assumed but not assured. Evidence for diagnosis is difficult to interpret in the context of inserting FeNO monitoring into a diagnostic pathway. Evidence for management is also inconclusive, but largely consistent with FeNO monitoring resulting in fewer exacerbations, with a small or zero reduction in ICS use in adults and a possible increased ICS use in children or patients with more severe asthma. It is unclear which specific management protocol is likely to be most effective. The economic analysis indicates that FeNO monitoring could have value in diagnostic and management settings. The diagnostic model indicates that FeNO monitoring plus bronchodilator reversibility dominates many other diagnostic tests. FeNO-guided management has the potential to be cost-effective, although this is largely dependent on the duration of effect. The conclusions drawn from both models require strong technical value judgements with respect to several aspects of the decision problem in which little or no empirical evidence exists. There are many potential directions for further work, including investigations into which management protocol is best and long-term follow-up in both diagnosis and management studies. Study registration: This study is registered as PROSPERO CRD42013004149. Funding: The National Institute for Health Research Health Technology Assessment programme.
Keywords
