Routine FEV₁ measurement is essential in diagnosis and monitoring of childhood asthma: myth or maxim?

What is the evidence that spirometry helps diagnose asthma?

Whilst many studies have described differences in spirometry between children with and without asthma, these results answer the question “How does spirometry differ between children with and without asthma?” However, in the clinic we are not asked to diagnose asthma in children who do and do not have asthma; instead, we see children who might have asthma. Thus, when considering an asthma diagnosis, the clinical question is “What is the precision of spirometry for diagnosing asthma in children who might have asthma?” To answer this question requires a study of children with possible asthma who have tests done and then have an asthma diagnosis made. In a perfect world, children with an initial asthma diagnosis would be reviewed after 1 year (or more) to identify those who have persistent asthma symptoms and therefore most likely have “asthma”. A recent ERS task force addressing asthma diagnosis reviewed more than 10 000 papers, from which only three papers described the precision of spirometry for diagnosing asthma [14–16]; the conclusions of the task force are described later in this review.

Which index of spirometry should be used?

A spirometer provides two spirometric indices that are recommended for diagnosing asthma: forced expiratory volume in 1 s (FEV₁), and the ratio of FEV₁ to forced vital capacity (FVC). A third index that can be used is BDR, which can be determined by comparing the change in FEV₁ 15 min after inhaling a short-acting β₂-agonist [17]. Some guidelines recommend peak expiratory flow (PEF) measurements twice a day for 2–4 weeks to identify PEF variability, which may be an index of variable airflow obstruction [18].

FEV₁/FVC is among the most commonly used spirometric indices [10, 12], whilst reduced FEV₁ can also be used as a standalone indicator of airflow obstruction [9, 12]. Two guidelines use reduced FEV₁ as the primary spirometric index, but this requires confirmation of airflow limitation with FEV₁/FVC ratio [1, 8]. Most guidelines recommend BDR if FEV₁ or FEV₁/FVC ratio are reduced [1, 8–10, 12], although the BTS/SIGN guideline suggests BDR could be done routinely [11].

PEF variability can be used as a secondary or tertiary test where there is diagnostic uncertainty, or as a primary test where resources are limited. In combination with other objective features, PEF can help to sustain asthma diagnosis [1, 10, 12]. The ICON paediatric asthma guideline states that PEF variability can be useful to support asthma diagnosis [9], whilst the NAEPP guidelines do not actively recommend using PEF reading for asthma diagnosis [8]. Duration of PEF monitoring ranges from 2 to 4 weeks: the ERS task force for diagnosis of asthma and GINA recommend 2 weeks of monitoring [1, 12], whilst NICE and BTS/SIGN guidelines advise 2–4 weeks [10, 11].

Which cut-off value in spirometry defines “abnormal”?

Thresholds that define “abnormal” spirometry are not consistent, and the different thresholds are not explained by some guidelines being older than others. Traditionally, the normal range is considered to be 80–120%, i.e. ±2 coefficients of variation (CV), but this assumes that children are like adults and have a CV of 10%. Children have an FEV₁ CV of 15% so the range of “normal” is closer to 70–130%, meaning that children with FEV₁ 70–80% are wrongly considered as being “abnormal” [19]. The NICE guideline defines “abnormal” as FEV₁/FVC ratio <70% of predicted or below the lower limit of normal (LLN), i.e. below the 5th centile [10]. This was found to have a sensitivity for diagnosed asthma of 52% and specificity of 73% [11]. Conversely, GINA recommends using the cut-off value for FEV₁/FVC ratio of below LLN, stating that this is approximately <0.90 for children [1]. The ERS task force for diagnosis of asthma uses either FEV₁/FVC ratio or FEV₁ <LLN or <80% [12], while the ICON paediatric asthma guideline has FEV₁ <80% [9]. Although the NAEPP guideline recommends spirometry as part of diagnosing asthma, it does not define abnormal spirometry [8]. Table 1 presents sensitivities and specificities and positive and negative predictive values for different FEV₁ and FEV₁/FVC cut-offs for asthma.

A BDR of at least 12% in FEV₁ is defined as a positive test by guidelines [1, 8–12]. Two guidelines also suggest a rise of >200 mL in FEV₁ is a positive test result [9, 12]. One guideline suggests either at least a 12% rise in absolute baseline FEV₁ or a rise of 10% from baseline FEV₁ % predicted [9]. A recent ERS/American Thoracic Society (ATS) technical standard states that 10% could be a significant BDR [22]. A 12% cut-off for BDR is associated with a sensitivity of 35–36% and specificity of 90–98% [23, 24].

Twice-daily PEF measurements are mainly recommended for 2 weeks or more, although the thresholds are different: >20% variability is mentioned in the NICE guideline [10], ≥12% in the ERS clinical practice guidelines for diagnosis of asthma [12] and >13% in the GINA report [1]. A 12.3% variability in PEF provides 50% sensitivity and 72% specificity [11].

Where does spirometry sit in the diagnostic hierarchy?

Many guidelines have rather complex algorithms that describe the order in which tests should be performed, but spirometry is the initial recommended test in all algorithms. Importantly, all guidelines emphasise that symptoms are at the top of the diagnostic hierarchy. Guidelines differ in which test follows spirometry. In some guidelines, abnormal spirometry should be followed with BDR [1, 10, 12]. The BTS/SIGN guideline states that initial spirometry should be accompanied with BDR, without specifying whether spirometry is normal or abnormal [11]. Other tests recommended by guidelines, often if spirometry is normal, include exhaled nitric oxide fraction (F_ENO), PEF variability and exercise testing. Table 2 and figure 1 summarise the hierarchy of tests recommended by different guidelines.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

A summary of the place of spirometry in diagnostic algorithms from current guidelines. BTS/SIGN: British Thoracic Society/Scottish Intercollegiate Guidelines Network; NICE: National Institute for Health and Care Excellence; GINA: Global Initiative for Asthma; ERS: European Respiratory Society; FEV₁: forced expiratory volume in 1 s; FVC: forced vital capacity; LLN: lower limit of normal; F_ENO: exhaled nitric oxide fraction; BDR: bronchodilator response; PEF: peak expiratory flow; +: positive; −: negative. ^#: BTS/SIGN uses high/intermediate/low probabilities of asthma as part of the diagnostic algorithm; a patient in the high probability category has an increased probability to have asthma if spirometry shows airflow obstruction reversible with a bronchodilator; a child who is in the intermediate category can have asthma diagnosis confirmed if spirometry is abnormal and has significant responsiveness to a bronchodilator. ^¶: direct challenge tests: methacholine challenge test positive if provocative concentration causing a 20% fall in FEV₁ (PC₂₀) ≤8 mg·mL⁻¹; mannitol test positive if fall in FEV₁ ≥15% at cumulative dose of ≤635 mg; indirect challenge test: exercise challenge positive if 8–20% fall in FEV₁. ^§: indicated for children with severe disease or clinical suspicion of other conditions. ^ƒ: GINA guideline provides two options in case the patient is on controller treatment and requires asthma diagnosis confirmation: 1) if there are variable respiratory symptoms but spirometry is normal, the recommendation is to consider repeating spirometry either when withholding bronchodilator or when the patient is symptomatic; 2) if there are fewer respiratory symptoms and spirometry is normal, bronchodilator testing can be repeated when symptomatic or when temporarily stopping bronchodilator. ^##: if available, direct bronchial challenge test with methacholine or indirect bronchial challenge tests using either a treadmill, bike or both. ^¶¶: PEF variability would be an inferior choice to challenge tests and could be used if challenge tests are not available.

What about testing in preschool children?

Guidelines do not recommend that testing should be performed in children aged <5 years, and none suggest how clinical teams might interpret spirometry in this age group. Although spirometry can be performed in specialist centres in children as young as 3 years, few children aged <5 years can provide reliable spirometry in routine clinical practice. The section “Spirometry to monitor asthma” discusses practical challenges in getting spirometry measurements in children. The burden of respiratory symptoms, including asthma, is highest in those aged <5 years, and diagnosing asthma remains a purely clinical diagnosis in young children.

How do the guidelines perform in real life?

There are a number of challenges to applying guidelines to practice. For example, some children referred with possible asthma may already be on a preventer treatment, which may affect test results. The GINA guideline provides advice for diagnosing children in this context, e.g. withhold bronchodilator before testing, or stop preventer if the child's symptoms are well controlled [1].

A prospective observational study in Switzerland applied the GINA and NICE diagnostic algorithms to 514 children referred with possible asthma, of whom 70% were ultimately given an asthma diagnosis. The GINA algorithm had a sensitivity of 42% and specificity of 90% for asthma, and the NICE guideline had corresponding values of 69% and 67%. The authors recommend that an experienced health professional is involved when diagnosing asthma, since diagnostic algorithms are not adequately sensitive [21]. The low sensitivity is likely due to spirometry values being “normal” in most children with asthma.

A second study described the sensitivity and specificity of components of tests recommended by the NICE guideline, and also the diagnostic algorithm [20]. In a prospective birth cohort study, spirometry and other tests were performed in 630 children aged 13–16 years, of whom 34 were on regular inhaled corticosteroid (ICS) treatment. Only 10 children involved met NICE's criteria for abnormal spirometry (using FEV₁/FVC <70% predicted), and only two had asthma. BDR was found positive in 9% of children (54 out of 624), of whom 12 had asthma [20].

These two studies highlight the limitations of applying guideline algorithms to real-world settings: different results will arise when the same algorithm is used in different populations and also when different algorithms are applied to the same population. History and examination remain the most important part of diagnosing asthma; testing can help with decision making and spirometry is the first-line test.

What is the evidence that spirometry helps monitor asthma?

There is limited trial evidence supporting the use of longitudinal measurements of spirometry to guide asthma management. One randomised controlled trial (RCT) undertaken in primary care in Australia described the benefit of adding spirometry to routine (symptoms-based) care delivered every 3 months. The trial considered the impact of intervention on the quality of life, asthma attacks and symptoms over 1 year. There were no differences in any outcome between participants in the two arms of the trial [25]. Similar results were provided by a second RCT that assessed the benefit of adding PEF monitoring to usual asthma care over 12 weeks. The intervention group was asked to record twice-daily PEF readings and asked to increase their ICS treatment when PEF was <70% of their personal best and to start oral steroid treatment when PEF was <50% of their personal best. None of the study outcomes were different between participants in the two arms of the trial [26]. A third RCT sought to determine whether knowledge of PEF results improved adherence to ICS treatment among ethnic minority inner-city children over 6 weeks. Compared to peers who did not have PEF measures, those who received PEF values had better adherence to ICS treatment (28% versus 49% respectively) [27]. There is absence of evidence in the literature to say whether spirometry is or is not helpful in monitoring asthma. What evidence is available comes from relatively small studies with an open-label design.

How should change in spirometry be expressed?

Spirometry is dependent upon age and height and should be standardised and expressed as a percentage of predicted or z-score. While % predicted has traditionally been the preferred way to express spirometry, z-scores have increasingly found favour. Most recently, a “conditional change score” (Zc) has become the recommended way to express a change in spirometry. Zc expresses the change in z-score between paired FEV₁ measurements but also considers factors associated with change in spirometry independent of asthma, i.e. the age of the individual and interval between testing [22].

What cut-off value in change in spirometry is clinically significant?

Two guidelines, now >10 years old, recommend that asthma treatment should be increased when FEV₁ % predicted falls below 80% [8, 9]. As discussed earlier, a change of 12% or 10% in FEV₁ is considered to be clinically significant for BDR testing [22, 28], and this threshold is also assumed to be the same for a clinically significant change in paired measurements of FEV₁; it may not be valid to make this assumption.

A lot of evidence for what merits a clinically important change in FEV₁ comes from a study published in 1981 that used data from three studies and a total of 47 adults [29]. The authors concluded that week-to-week variation in FEV₁ was 12% for those with normal lung function and 23% for those with abnormal (obstructive) lung function [29]. One study used paired FEV₁ measurements made at an interval of 3 months in 1112 children and related this to subsequent asthma outcomes [30]. A 10% fall in FEV₁ % predicted was associated with a 28% increase in odds for an asthma attack in the following 3 months and a 21% increase in odds for loss of asthma control (among those who were initially well controlled) [30]. Change in FEV₁/FVC was not associated with a significant change in odds for future asthma outcome [30]. At the time of writing, no study or guideline has suggested what might be a significant fall in z-score or Zc to suggest treatment should be stepped up or down. Similarly, there is no evidence or guidance as to whether, for example, ICS treatment should be altered in light of changing spirometry or whether other treatments, such as long-acting β₂-agonists or leukotriene receptor antagonists, should be added or stopped.

What about other tests used in combination with spirometry?

At least nine clinical trials have used F_ENO to guide asthma management and three of these have used FEV₁ in their algorithms. Different cut-offs for FEV₁ were used during the studies: one used 80% [31], another one used >80%, 70–79% and <70% [32], whilst the third did not mention the thresholds [33]. All three of these trials found that guiding treatment by symptoms, FEV₁ and F_ENO was associated with reduced exacerbations. In contrast, none of the other studies that used only symptoms and F_ENO found the intervention was associated with reduced exacerbations. It is possible that the “success” of these three F_ENO trials was at least partly explained by the role of FEV₁ in the decisional algorithms, and this raises the possibility that change in both FEV₁ and F_ENO may have merit in monitoring asthma, compared to treatment guided by F_ENO and symptoms alone or FEV₁ and symptoms alone.

Can spirometry be performed in children in the community?

Spirometry requires apparatus and trained staff and these are not available in all clinical contexts, especially in community-based settings. In the UK, 46% of the nurses doing annual asthma reviews in primary care had no spirometry training beforehand, and none had training in doing spirometry in children [34]. One study describes the feasibility and acceptability of measuring spirometry in a community setting in the UK [35]. The researchers spoke with 27 nurses and parents of more than 600 children. The nurses involved had no prior expertise in paediatric asthma or spirometry, and had training prior to doing spirometry. Valid spirometry was obtained in 97% of children and 87% of these would be happy to repeat the tests. The test result changed asthma management in approximately one quarter of the children [35].