Pulmonology · Medical Literature
Exploring the role of cold air in airway hyperresponsiveness and asthma diagnostic testing: a systematic review and meta-analysis.
Abstract
Asthma is a heterogeneous disease, making uniform diagnosis challenging, resulting in both under- and overdiagnosis. We conducted a systematic review and meta-analysis to summarise the data on diagnostic accuracy of a cold air bronchial challenge test (CACh) in exercise-induced bronchoconstriction and asthma. Databases (PubMed, Embase, Web of Science core collection, Cochrane Library and the Cumulative Index to Nursing and Allied Health Literature via EBSCO) were systematically screened for articles evaluating the use of CACh as a diagnostic tool in exercise-induced bronchoconstriction and asthma. A random-effects model was used to calculate the effect on the reduction in forced expiratory volume in 1 s (FEV1) by CACh in healthy controls compared to patients with asthma. Risk of bias was evaluated using the Quality Assessment of Diagnostic Accuracy Studies-Comparative. This study is registered with PROSPERO (www.crd.york.ac.uk/PROSPERO/ identifier number CRD42021290350). After screening and removal of duplicates, 39 studies were included. Aggregate sensitivity and specificity of a FEV1 decrease ≥10% after CACh were 0.808 and 0.959, respectively. The weighted mean difference in percentage FEV1 decrease after CACh for patients with asthma compared to healthy controls was -17.58. High heterogeneity (I2=99) was mainly explained by differences in reference tests to diagnose asthma. Application of the trim-and-fill method to account for publication bias imputed eight studies, resulting in a notable reduction in heterogeneity. The overall effect (a significant decrease in FEV1 in patients with asthma) remained statistically significant (-8.18, p<0.001). Our study shows the high sensitivity and specificity and possible clinical implications of a CACh in the diagnosis of asthma. Further research is needed to confirm this.
Clinical relevance
What is the diagnostic accuracy of the cold air bronchial challenge test (CACh) for detecting airway hyperresponsiveness in patients with exercise-induced bronchoconstriction (EIB) and asthma, and how does the magnitude of FEV1 reduction differ between asthmatic patients and healthy controls? __SNAPSHOT__ {"studyDesign":"Systematic review and meta-analysis of diagnostic test accuracy studies (PRISMA-DTA compliant; PROSPERO registered: CRD42021290350)","setting":"Multi-database systematic literature search (PubMed, Embase, Web of Science Core Collection, Cochrane Library, CINAHL via EBSCO); database inception to 15 July 2024; no restriction on country or institution type","sampleSize":"39 studies included after screening and deduplication; aggregate participant count not reported in abstract","population":"Patients with asthma and/or exercise-induced bronchoconstriction compared to healthy controls; age range varied across included studies; most studies predated 2000, with limited representation of EIB as a distinct entity","intervention":"Cold air bronchial challenge test (CACh): isocapnic hyperventilation of cold/subfreezing air with FEV1 measured post-challenge; protocol parameters (temperature, target ventilation, machine type, challenge duration) varied substantially across included studies","comparator":"Healthy controls (for meta-analysis of FEV1 magnitude); various reference standard tests used to establish asthma diagnosis across included studies (no gold standard available; high variability in reference tests)","primaryOutcome":"Diagnostic accuracy of CACh (sensitivity and specificity using ≥10% FEV1 decrease as a positive test threshold); weighted mean difference in percentage FEV1 decrease between asthmatic patients and healthy controls","followUp":"Not reported (cross-sectional diagnostic accuracy studies)","fundingSource":"Not reported in the abstract or accessible full-text sections"}
Key findings
Among 39 included studies, the aggregate (weighted mean) sensitivity of CACh using a ≥10% FEV1 decrease as the positivity threshold was 0.808 (range across studies: 0.22–1.00), and aggregate specificity was 0.959 (range: 0.38–1.00). Compared to cold air exercise challenge (CAEXCh; sensitivity 0.674, specificity 0.897) and room-temperature exercise challenge (RTEXCh; sensitivity 0.291, specificity 0.966), CACh demonstrated higher sensitivity while maintaining comparable specificity to CAEXCh. The meta-analysis estimated a weighted mean difference in percentage FEV1 decrease of −17.58% for asthmatic patients versus healthy controls. Critically, heterogeneity was extreme (I² = 99%), predominantly attributable to variability in the reference standard tests used to diagnose asthma across studies. Application of the trim-and-fill method imputed 8 potentially missing studies, which markedly reduced heterogeneity; the effect estimate diminished to −8.18% but remained statistically significant (p < 0.001). Confidence intervals for individual accuracy estimates and the meta-analytic pooled mean difference were not reported in available abstract sections. NNT and NNH are not calculable from the data reported. No specific adverse event data were reported.
Practice implications
This is, to the authors' knowledge, the first systematic review and meta-analysis specifically evaluating CACh for the diagnosis of EIB and asthma. The aggregate diagnostic estimates suggest that CACh is highly specific (0.959) and moderately-to-highly sensitive (0.808), which would favor its use as a rule-in test for airway hyperresponsiveness in clinically suspected asthma. The specificity is consistent with the European Respiratory Society's characterization of indirect challenges as highly specific diagnostic tools. However, several factors substantially limit the strength of these conclusions. First, heterogeneity was extreme (I² = 99%), and the trim-and-fill analysis reduced the effect estimate from −17.58% to −8.18%, raising concern that the unadjusted figures may overestimate the magnitude of effect. Second, there is no validated gold standard for asthma diagnosis, and reference standards varied widely across included studies. Third, most included studies were conducted before 2000, predating current understanding of asthma heterogeneity and the distinct phenotype of EIB; accordingly, contemporary asthma subtypes and EIB-without-asthma patients are underrepresented. Fourth, CACh protocol parameters (air temperature, target ventilation, device type, challenge duration) were inconsistent across studies, limiting generalizability and reproducibility. Fifth, individual sensitivity values ranged from 0.22 to 1.00, indicating that test performance is highly dependent on protocol-specific factors. Findings should not be used to justify routine clinical adoption of CACh without standardized protocols. The test may be particularly informative in patients with suspected asthma or EIB in whom standard testing is inconclusive, but confirmation through adequately powered studies with standardized methodology is needed before practice changes are warranted. Only the abstract and select full-text excerpts were available for this summary; full methods, individual study-level data, and forest plots were not reviewed. __LIMITATIONS__ Extreme heterogeneity (I² = 99%) across included studies undermines the reliability of pooled estimates; the primary driver was variability in reference standards used to diagnose asthma.,No validated gold standard for asthma exists; high variability in reference tests across 39 studies makes sensitivity and specificity estimates population- and protocol-dependent.,Majority of included studies predate 2000, introducing potential spectrum bias given contemporary reclassification of asthma phenotypes and EIB as a distinct entity.,Wide per-study sensitivity range (0.22–1.00) and specificity range (0.38–1.00) indicate extreme methodological heterogeneity; pooled estimates may obscure clinically meaningful protocol-level differences.,CACh protocol standardization was absent: air temperature, target minute ventilation, challenge duration, and measurement device differed substantially across studies.,Trim-and-fill analysis imputed 8 missing studies and substantially reduced the pooled effect estimate, suggesting probable publication bias favoring studies with larger FEV1 decrements.,Aggregate participant counts were not reported, limiting the ability to assess statistical precision.,EIB as a distinct phenotype was infrequently evaluated, and findings may not apply to EIB without concurrent asthma.,Confidence intervals for pooled sensitivity, specificity, and mean difference were not available in the abstract; precision of the estimates cannot be fully assessed.,Absence of reported funding source introduces uncertainty about potential conflicts of interest.,Generalizability is limited by the predominantly pre-2000 study population; applicability to current clinical settings with modern spirometry standards is uncertain. __BOTTOM_LINE_CATEGORY__ promising_but_preliminary __BOTTOM_LINE__ CACh demonstrates high specificity (0.959) and moderate-to-high sensitivity (0.808) for airway hyperresponsiveness using a ≥10% FEV1 decrease threshold, and produces significantly greater FEV1 reductions in asthmatic patients than healthy controls (weighted mean difference −17.58%, attenuated to −8.18% after publication bias adjustment). However, extreme heterogeneity (I² = 99%), the absence of a standardized CACh protocol, and a predominantly pre-2000 evidence base preclude routine clinical adoption at this time. CACh should be considered a promising but investigational diagnostic adjunct requiring prospective validation with standardized protocols before integration into clinical asthma diagnostic pathways.