than absolute accuracy, given that the usual application for peak flow involves an assessment of serial changes in peak flow in the same subject over time (Van As, 1982).
Interpretation of peak flow variability is best performed by visual inspection of a graph that plots peak flow over time. Computer-based algorithms for interpreting peak flow measures have been shown to be no better than the "eyeball" method for diagnosing work-related asthma (Perrin et al., 1992).
Bronchial hyperreactivity, a cardinal feature of asthma, is measured as follows. Patients perform initial spirometry, and the resultant FEV1 is defined as the baseline value. The patient then inhales saline, followed by increasing concentrations of an agonist such as methacholine or histamine, or a stimulus such as cold air or a distilled water aerosol. After each provocation, spirometry is repeated and FEV1 is determined for that dose. The test is stopped when a predetermined reduction in lung function is achieved or the maximum dose is administered. Exercise testing is also used occasionally to measure bronchial reactivity. Typically the test ends when a 20% reduction in FEV1 or a 35 percent reduction in specific airway conductance has occurred, and a provocative dose is calculated.
The challenge method and type of data recorded differ among laboratories (Chai et al., 1975; Chatham et al., 1982a). Some record the concentration of agonist at the endpoint, whereas others record the cumulative dose. The percent decrease in FEV1 (targeted as a positive end point) ranges from 10 to 40 percent, with a 20 percent decrease being most common. Popa and Singleton (1988) attempted to determine which of the current provocative doses for histamine optimally separates normal from asthmatic subjects. They concluded that new normative data for diagnostic provocation were needed because of a high misclassification rate using current methods. Inhalation challenge testing with methacholine, histamine, distilled water, exercise, etc., is commonly termed non-specific challenge testing to distinguish these from specific allergen challenge testing. However, each of these chemical and physical agents act on different bronchial receptors and therefore reflect different aspects of non-specific bronchial hyperreactivity. Furthermore, animal studies indicate that inherited hyperreactivity to these agents is transmitted at distinct genetic loci (Levitt et al., 1990).
Despite recognized limitations, testing for airways reactivity is widely used because it is a practical test which has great utility in clinical medicine and research. Asthmatics as a group develop reductions in lung function with provocation at much lower doses than do non-asthmatics. The considerable within-subject overlap in hyperreactivity to methacholine, histamine