Repeatability of a rapid dosimetric method for methacholine challenge using a pocket turbine spirometer for FEV1 measurements

Reproducibility of non-specific bronchial challenge in adults: implications for design, analysis and interpretation of clinical and epidemiological studies

BACKGROUND

Poor reproducibility of an outcome measure reduces power and, in an independent variable, biases results. The intraclass correlation coefficient measures loss of power and degree of bias. Information is lacking on the intraclass correlation coefficient for bronchial responsiveness and factors affecting reproducibility.

METHODS

Papers containing information on reproducibility of bronchial responsiveness were identified using a Medline search and citations. Within and between person components of variance of PD20 or PC20 were expressed in doubling dose or concentration units, and the intraclass correlation coefficient calculated when not reported.

RESULTS

Results were extracted from 32 papers. Intraclass correlation coefficients were over 0.9 in short term studies of highly selected asthmatic patients, but larger and most long term studies had lower intraclass correlation coefficients, less than 0.5 in some cases, due to greater within person or lower between person variation. Reproducibility of dose or concentration-response slope was generally higher, but still less than that of forced expiratory volume in 1 second.

CONCLUSIONS

Information is available to calculate sample size for studies with bronchial responsiveness as the outcome, but results when bronchial responsiveness is an explanatory variable may be misleading.

Serum ECP and MPO, but not urinary LTE4, are associated with bronchial hyper-responsiveness

A random population-based sample of 131 subjects was used to assess the value of serum eosinophil cationic protein (ECP), serum myeloperioxidase (MPO), and urinary leukotriene E4 (LTE4) in predicting bronchial hyper-responsiveness measured by methacholine challenge. Special interest was focused on the history of aspirin intolerance and on smoking as contributing factors. The mean serum ECP and MPO were higher in hyper-reactive [provocational dose causing a 20% fall in forced expiratory volume in 1 sec. (PD20) < or = 6900 micrograms] than in non-hyper-reactive subjects (22.3 vs. 13.2 micrograms l-1, P < 0.001 and 377 vs. 278 micrograms l-1, P = 0.001, respectively). This was also seen in current smokers vs. never smokers (17.2 vs. 12.9 micrograms l-1, P = 0.03 and 372 vs. 286 micrograms l-1, P = 0.04, respectively). There were no differences in baseline urinary excretion of LTE4 between hyper-reactive and non-hyper-reactive subjects. During the 2 h after methacholine challenge, urinary LTE4 excretion increased from 53.8 and 69.0 ng mmol-1 creatinine in non-hyper-reactive subjects, but there was no change in hyper-reactive subjects (non-hyper-reactive vs. hyper-reactive, P = 0.06). The increase was greatest in subjects with aspirin intolerance causing urticaria or angioedema but not aggravation of asthma (from 58.5 to 87.2 ng mmol-1 creatinine), probably due to extrapulmonary leukotriene production. Our results indicate that serum ECP and MPO, but not urinary LTE4 (even in subjects with a history of aspirin intolerance), predict bronchial hyper-responsiveness to methacholine. The subject's smoking history must be taken into account when these parameters are considered.

Comparison of two methacholine challenge methods using Spira-2 or Mefar dosimeter

Two bronchial challenge protocols with breath-actuated dosimeters, Spira Elektro-2 and Mefar, with similar cumulative dose steps were compared in 28 patients with mild to moderate asthma. Methacholine challenges were performed after two different protocols at the same time of day in random order 3 or 4 days apart. The provocative dose of methacholine producing a 20% fall in forced expiratory volume in 1 second (PD20) was lower when determined by Spira than with the Mefar dosimeter (P < 0.05). Transition equations calculated by linear regression analysis were: PD20mefar = exp10[0.897 + 0.678(logPD20spira)] (P < 0.05; r = 0.62) and PD20spira = exp10[0.759 + 0.559 (log PD20mefar)] (P < 0.05; r = 0.62). The slopes were calculated by regressing the percentage fall in FEV1 on log10 (dose) and transformed as slope = 100/(regression coefficient + 10). The mean slope (95% CI) for Spira was 3.1 (2.6-3.7) and for Mefar 4.4 (3.6-5.1) (P < 0.005). Regression equations calculated by linear regression analysis were: slope(mefar) = 2.126 + 0.712 slope(spira) (P < 0.05; r = 0.51) and slope(spira) = 1.551 + 0.365 slope(mefar) (P < 0.05; r = 0.51). In conclusion, PD20 was smaller and the decline in FEV1/log(dose) curve steeper using the Spira compared with the Mefar protocol. The dose-response curves should be validated and transition equations calculated when bronchial reactivity to inhaled agents is compared, even while using apparently similar well-standardized dosimeter methods.

A rapid dosimetric methacholine challenge in asthma diagnostics: a clinical study of 230 patients with dyspnoea, wheezing or a cough of unknown cause

The rapid methacholine challenge test using a pocket turbine spirometer (Micro Spirometer) and the Spira Elektro 2 dosimeter was performed with 230 consecutive patients who had dyspnoea, wheezing or a prolonged cough of unknown cause. Patients with previous asthma diagnoses as well as those who had used inhaled steroids during the preceding 4 weeks were excluded. Seventy-eight patients (34%) were methacholine positive (PD20FEV1 < or = 6900 micrograms) 47 (60%) of whom had a final diagnosis of American Thoracic Society (ATS) criteria fulfilling bronchial asthma. One hundred and fifty-two patients (66%) were methacholine negative (PD20FEV1 > 6900 micrograms) 14 (9%) of whom had bronchial asthma according to clinical evaluation. Increased bronchial responsiveness was strongly associated with ATS criteria fulfilling asthma (P < 0.0001). When PD20FEV1 was used, 47 (77%) of the asthmatic patients were hyper-responsive (range 40-6900 micrograms) compared to 31 (18%) of the non-asthmatic patients (range 160-6900 micrograms). When using PD15FEV1, 51 (84%) of the asthmatic patients (range 28-6900 micrograms) and 52 (31%) of the non-asthmatic patients (range 100-6900 micrograms) were hyper-responsive. The level of bronchial responsiveness measured by both PD20FEV1 and PD15FEV1 differed significantly between asthmatic and non-asthmatic patients (P < 0.0001). Hyper-responsiveness was associated with an increased daily variation in peak expiratory flow (PEF) (P < 0.0001) and an increased number of blood eosinophils (P < 0.0001). Hyper-responsiveness was also associated with decreased levels of FEV1 and percentages of predicted FEV1 (P = 0.04 and P < 0.0001, respectively). Stepwise logistic regression analysis showed that the number of positive prick results (OR = 1.15, 95% CI 1.01-1.31), blood eosinophils (1.004, 1.00-1.01), level of FEV1 (0.56, 0.36-0.87) and current smoking (2.36, 1.00-5.59) were factors significantly associated with the probability of hyper-responsiveness. Age, gender, atopy, pets and a history of ex-smoking were not significantly associated with hyper-responsiveness, neither in univariate nor in multivariate analyses. The Bayesian analysis was used to investigate the diagnostic value of the rapid methacholine challenge test. A receiver operator characteristic curve demonstrated that PD20FEV1 separated asthmatic and non-asthmatic patients better than PD15FEV1. The best cutoff value of PD20FEV1 was 6000 micrograms, but the difference from 6900 micrograms was minimal. The best results of the test using a PD20FEV1 cutoff point of 6900 micrograms (PPV: 0.80, NPV: 0.79) were obtained when the pre-test probability was 0.48. The interval security of the test was established by a pre-test probability between 0.19 and 0.78. Maximal positive (0.34) and negative (0.31) final gains were achieved when pre-test probabilities were 0.33 and 0.65, respectively. The cutoff level of 150 micrograms gave 100% of specificity and predictive value of a positive test for clinical asthma diagnosis. The Bayesian analysis approach demonstrated that the test is useful in asthma diagnostics if not performed on patients with lowest or highest probabilities of asthma.