Measuring fractional exhaled nitric oxide concentration in asthma: NIOX MINO, NIOX VERO and NObreath

6.1 The Diagnostics Advisory Committee reviewed the evidence available on the clinical and cost effectiveness of measuring fractional exhaled nitric oxide (FeNO) to inform the diagnosis and management of asthma in children and adults. The Committee considered the report produced by the External Assessment Group and statements from patient experts on the Committee and from clinical specialists who acted as specialist Committee members on this assessment.

6.2 The Committee considered whether NIOX MINO, NIOX VERO and NObreath could be considered equivalent for the purpose of this assessment. It noted the review by the External Assessment Group, which indicated that, although some differences were observed in test results, there was generally a good correlation with results from other chemiluminescence devices. The Committee noted that there appeared to be poorer equivalence between devices in some circumstances, such as at higher FeNO levels, and that the direction of disagreement varied between studies and devices. However, the Committee acknowledged that there is no commonly accepted definition of clinically acceptable differences in FeNO measurements. The Committee concluded that, based on the available evidence, the 3 devices could, on balance, be considered to be broadly equivalent. The Committee also thought that standardisation of FeNO devices should be encouraged.

6.3 The Committee considered whether very young children would be able to perform the test. It heard from both manufacturers that the minimum recommended age for using FeNO monitoring devices is 5 years. The Committee also noted that the External Assessment Group's systematic review only included studies of children 5 years and older, in line with the review protocol. The Committee concluded that there was insufficient evidence to determine the suitability of FeNO testing for children younger than 5 years.

6.4 The Committee discussed the lack of gold standard for asthma diagnosis. It heard from the clinical specialists that there is no single clinical definition of asthma and that the diagnosis is based on multiple factors, including the presence of symptoms and evidence of airway obstruction. It heard that high FeNO levels in people with symptoms suggestive of asthma, such as coughing and wheezing, may indicate that the patient has eosinophilic asthma, which may be treated with inhaled corticosteroids. However, FeNO levels are not raised in all patients with asthma and, conversely, not all people with raised FeNO levels have asthma. The Committee further heard from a clinical specialist that, although it is generally accepted that FeNO levels correlate with eosinophilic airway inflammation, this effect may be stronger in some groups of patients than others and that there are other causes of inflammation. The Committee also heard that clinical practice varies and so the use of standard clinical practice as a reference standard is problematic. The Committee therefore concluded that the lack of gold standard for diagnosis, the complexity of diagnosis and the variation in clinical practice introduced an increased uncertainty in the assessment of the clinical validity of FeNO devices.

6.5 The Committee considered the diagnostic accuracy of FeNO testing. The Committee noted that the External Assessment Group had presented accuracy data for both children and adults against various reference standards and in different positions in the diagnostic pathway. It acknowledged that, because of the clinical heterogeneity of the data, a meta-analysis had not been performed and that estimates of specificity and sensitivity varied between studies. Overall, the Committee accepted the External Assessment Group's observation that the ranges of specificity were generally narrower than those for sensitivity, and that FeNO testing appeared to have a higher specificity than sensitivity. It heard from the clinical specialists that higher specificity would indicate that testing would be of greater use as a rule‑in test; that is, patients testing positive are assumed to have asthma and patients testing negative have further tests for asthma. The Committee considered that the absence of a meta-analysis of accuracy meant that there was a greater uncertainty about the accuracy of FeNO devices in this assessment. Nevertheless, it was satisfied that the specificity of the devices was acceptable if they are used in a rule‑in scenario.

6.6 The Committee considered FeNO cut-off values for guiding diagnosis of asthma in adults and children. The Committee heard from the External Assessment Group that cut-offs were generally not interchangeable between the FeNO devices, and that there appeared to be a high degree of heterogeneity between the studies. The Committee also heard from a clinical specialist that, in their experience, the devices generally produce different readings when used by the same patient. However, the differences appeared to be consistent. The Committee noted that the highest sum of sensitivity and specificity was obtained when the cut-offs ranged from 19–21 parts per billion (ppb) in children and the highest sum of sensitivity and specificity was wider for adults (20–47 ppb). The Committee further noted that a higher cut-off was needed to optimise the specificity of the devices; a cut-off between 47 ppb and 76 ppb resulted in specificity of 88–100% in adults and a cut-off range of 30 ppb to 50 ppb resulted in specificity of 92–100% in children. The Committee was informed by clinical experts that cut-off values in the higher range would be preferred to reduce the rate of indeterminate results, and that the test could be used to rule in a diagnosis of asthma in people whose test is positive. The Committee concluded that cut-off values in the higher ranges should generally be used, and that cut-off values should be lower in children than adults.

6.7 The Committee considered the generalisability of the clinical evidence to the whole population both for diagnosing and managing asthma. The Committee acknowledged that the clinical evidence was heterogeneous in terms of clinical characteristics and results, and that studies were identified based on their similarity to UK practice and similarity to the subgroups of interest as defined in the protocol (that is, in people in whom the condition is difficult to diagnose, or the wider population of people presenting with symptoms of asthma). As such, no single study can be generalised to the whole population. The Committee therefore concluded that both the variation in UK clinical practice and the heterogeneity of studies included in the assessment would increase the uncertainty around the clinical benefits of measuring FeNO.

6.8 The Committee discussed the clinical evidence on the use of FeNO to guide asthma management in adults. The Committee noted that there was a statistically significant reduction in the meta-analysis for the composite outcome of any severity of exacerbations when FeNO measurement was used to guide management compared with treatment without FeNO measurement during the first 12 months of management. However, it acknowledged that the results from the individual studies were heterogeneous and that the meta-analysis did indicate a non-significant trend towards reduction in severe exacerbation rates. The Committee noted that the meta-analysis also showed a non-significant trend towards decreased inhaled corticosteroid use and considered that the effects of FeNO measurement on inhaled corticosteroid use were uncertain. It further noted that step-up and step-down protocols for inhaled corticosteroid use varied between studies and the effect of this on the outcomes was not known. The Committee concluded that FeNO‑guided management was likely to reduce exacerbation rates in adults, but the extent and duration of this effect was uncertain.

6.9 The Committee discussed the additional meta-analysis, performed by the External Assessment Group, which incorporated the more recent Honkoop et al. study (2013). It recognised that the addition of this study provided some further support that severe exacerbations could be reduced when FeNO measurement was added to asthma management. However, because the External Assessment Group had been unable to calculate accurate error rates for outcomes, the Committee was still uncertain whether this effect was statistically significant. The Committee therefore concluded that the Honkoop et al. study appeared to lend weight to the assumption that FeNO‑guided management was effective, but was unable to establish whether this was statistically significant.

6.10 The Committee then discussed the clinical results for FeNO‑guided management in children with asthma. As with the studies in adults, those in children appeared to have variations in design, including in the step-up and step-down protocols, medications and inclusion criteria. The Committee noted that all studies reported a decrease in exacerbations in the intervention arm, but only 1 reported a statistically significant reduction. It also noted that there was a greater uncertainty around whether inhaled corticosteroid use went up or down when FeNO measurement was added to asthma management. On balance, the Committee concluded that FeNO‑guided management in children is likely to result in reduced exacerbation rates, but that the extent and duration of this effect is still uncertain. The Committee further concluded that the effect of FeNO‑guided management on inhaled corticosteroid use is uncertain and recommended that further evidence is generated to establish its benefits.

6.11 The Committee discussed the role of measuring FeNO in diagnosing and managing asthma in children and adults. It heard from the clinical specialists that diagnosis of asthma needs consideration of many different elements, including symptoms, response to treatment and physiological testing. Given the complexities of diagnosis, the Committee considered that FeNO testing would not be able to replace current practice for diagnosis. The Committee concluded that measuring FeNO had not been shown to be able to reliably replace other tests and clinical observations, and therefore should be used as an add‑on to current clinical diagnosis and management in people with asthma.

6.12 The Committee discussed the additional benefits of measuring FeNO in the diagnosis and management of asthma. It heard from a patient expert that an accurate diagnosis of asthma can sometimes take many years, resulting in less than optimal treatment, which can have a direct impact on health. The Committee also heard from a patient expert that FeNO‑guided management could result in patients better understanding their own condition and disease progression, which could reduce hospitalisations and improve patient experience. The Committee also considered the effect that measuring FeNO could have on adherence to medication. The Committee was informed by a clinical specialist that approximately 30% of people do not take their medication as prescribed. The clinical specialist indicated that studies have shown that FeNO is a useful marker for medicine adherence and that FeNO devices could be a useful tool for doctors to improve concordance, by which the patient and clinicians make decisions together about the treatment. The Committee concluded that FeNO testing could potentially enable patients and doctors to improve treatment concordance in patients who are on medications for asthma.

6.13 The Committee considered the areas of uncertainty in the economic models produced by the External Assessment Group for diagnosing and managing asthma. The Committee noted that the diagnostic models were sensitive to: assumptions about the length of time needed to resolve misdiagnoses; assumptions about health losses incurred by patients who have false-negative results; the costs of asthma management; and the use of rule‑in and rule‑out diagnostic decision rules. The Committee considered the assumptions made in the management model, in which it noted that the results in both the children and adult subgroups were particularly sensitive to: assumptions about changes in inhaled corticosteroid use over time; the annual number of nurse visits for FeNO measurement; and the length of time FeNO measurement was assumed to impact on exacerbation rates and inhaled corticosteroid use. The Committee concluded that both models for diagnosis and management were subject to considerable uncertainty, and therefore the results should be interpreted with caution.

6.14 The Committee discussed the results from the base-case analysis for the use of FeNO testing in diagnosis. It noted that, although the methacholine challenge test produced the most quality-adjusted life years (QALYs), the incremental cost-effectiveness ratio (ICER) was very high at £1.25 million per QALY gained when compared with FeNO testing plus bronchodilator reversibility. The Committee also noted that the difference in the health benefit for methacholine challenge test compared with FeNO testing plus bronchodilator reversibility was estimated to be very small (0.0005 QALYs) and that costs were considerably lower for FeNO testing plus bronchodilator reversibility. Given that FeNO testing plus bronchodilator reversibility dominated the other tests, the Committee concluded that FeNO testing plus bronchodilator reversibility testing in adults and children delivered equal or greater QALYs at a lower ICER than other tests. Moreover, the Committee noted that the use of FeNO testing in conjunction with existing tests is more cost-effective than when the existing tests are used alone. The Committee therefore recommended FeNO testing as an option to help diagnose asthma in adults and children who, after initial clinical examination, are considered to have an intermediate probability of having asthma and where FeNO testing is intended to be done in combination with other diagnostic options according to the British guideline on the management of asthma (2012).

6.15 The Committee then discussed the results from the base-case analysis for asthma management in children. It noted that, for base-case results, the External Assessment Group had used 2 studies, Szefler et al. (2008) and Pijnenburg et al. (2005), to inform the clinical effectiveness, and the ICER was £45,200 per QALY gained. The Committee then discussed the sensitivity analyses for management in children and noted that, when the analysis was based on Pijnenburg et al., a more favourable ICER of £19,000 per QALY gained was obtained. It also noted that, when it was assumed that FeNO measurement impacts for a shorter length of time on exacerbations and inhaled corticosteroid use, FeNO measurement plus management as recommended in the British guideline on the management of asthma (2012), hereafter referred to as the 'British guideline', dominated management by the British guideline use alone for up to 4 years. It also noted that the ICER was £7600 per QALY gained for 5 years and £27,700 per QALY gained for 10 years. The Committee assumed that shorter duration of up to 10 years, rather than lifetime duration, was a more reasonable assumption, given that children would not be expected to remain in the child-model for the rest of their lifetime. Moreover, the Committee heard from the External Assessment Group that the Szefler et al. study included patients with uncontrolled asthma, and the study protocol did not allow therapy to be stepped down on the basis of low FeNO levels. The Committee considered that, in clinical practice, it was unlikely that the assumption of higher inhaled corticosteroid use throughout the time horizon with FeNO measurement would be seen, and therefore preferred the sensitivity analysis based on the Pijnenburg et al. study over the base-case analysis. Considering the combined shorter duration of impact of FeNO measurement and analysis based on the Pijnenburg et al. study, the Committee concluded that the most plausible ICER for management in children was likely to be lower than £19,000 per QALY gained.

6.16 The Committee discussed the results from the base-case analysis for asthma management in adults. It noted that the ICER for FeNO measurement plus the British guideline compared with the British guideline alone was £2100 per QALY gained. The Committee considered that, although there were uncertainties relating to this estimate, this ICER was low and therefore the use of FeNO measurement for asthma management in adults was likely to be cost effective. The Committee therefore accepted the base-case results for adults.

6.17 The Committee then discussed its recommendation for the use of FeNO measurement in asthma management in children and adults. The Committee heard from clinical specialists that there was considerable uncertainty around the use of stepping-up and stepping-down protocols for inhaled corticosteroid use, and that the External Assessment Groups' review had not conclusively demonstrated that FeNO measurement would be effective and safe for guiding the stepping down of inhaled corticosteroids. The Committee expressed concerns about using FeNO measurement as a basis for stepping-down treatment and was not satisfied that the evidence was robust enough to show that the benefits out-weighed the potential harms of under treatment. The Committee therefore concluded that FeNO measurement should not be recommended to help with stepping down inhaled corticosteroid use in adults or children whose asthma is well managed. However, it considered FeNO measurement to be cost and clinically effective when used as an option to support symptomatic asthma management in people using inhaled corticosteroids.

6.18 The Committee discussed the evidence in the subgroups defined in the scope for this assessment, (pregnant women, older people, people who smoke and those who have been exposed to tobacco smoke). The Committee heard from the External Assessment Group that there is little robust evidence for most of these groups (study designs were generally of lower quality), which could lead to biased results. It noted that randomised controlled trial evidence shows that measuring FeNO is at least as useful during pregnancy as it is for the general population, but it appears likely that FeNO is a less reliable indicator of airway inflammation in older people. The Committee therefore concluded that, in view of the limited evidence, it was unable to provide any specific recommendations for the subgroups defined in the scope.