Since provocation testing usually continues until a positive test occurs or until approximately 200 CBU have been administered, we have presented only a single, 224 CBU curve for a negative test. The asthmatic patients of Casale et al were selected so that all had a positive test at this level or below. Studies on unselected adult asthmatic patients suggest that 100 percent sensitivity at this level is a reasonable expectation. Cockroft et al have shown similar results for histamine challenge. However, 100 percent sensitivity at this level has not been a universal finding, and we have chosen to use a sensitivity of 90 percent for the 224 CBU negative test, a more conservative approach.

These curves show 100 percent post-test probability of asthma for positive tests at low CBU, especially for the nonsmokers. This reflects the highly selective nature of the data base, with no false positives (100 percent specificity), a highly unlikely situation in a true clinical setting.

With a protocol similar to that of Casale et al,^{ }Balzano et al have recently shown an intrasubject between-day variability in PD20 of ±1.66 fold (95 percent confidence interval). Applying this to the dose units of the curves of Figure 1, as a rough approximation, a subject whose levels on a given day fell on a certain CBU curve, could, on another day, fall on a curve halfway between this curve and the next curve in either direction.

canadian discount pharmacy

The data of Hopp et al using methacholine challenge in children, and that of Popa et al using histamine challenge in adults, can by interpolation and recalculation, be presented as curves similar to those of Figure 1; these curves are shown in Figure 2. Hopps data have specificities of 98 and 99 percent at low CBU, rather than 100 percent as in the data of Casale. The curves for these low CBU show the profound effect that specificity has at low CBU at low pretest probability. Except for this area, the curves are similar to those of Figure 1. Popas data show curves similar to those of Casales nonsmokers for positive tests. However, since Popas data showed 100 percent sensitivity at histamine doses greater than 1.0 (for PD20), the negative curve for these doses shows a zero post-test probability regardless of the pretest probability.

*FIGURE 2. Curves similar to Figure 1 for the data of Hopp et al (left) and Popa et al (right). The solid lines represent positive tests, the single dashed line in each panel represents a negative test. For the data of Hopp, the numbers of the curves refer to cumulative breath units of methacholine; for the data of Popa, the numbers of the curves refer to concentrations of histamine.*

These comparisons emphasize how small changes in sensitivity and specificity can affect certain areas of the diagram. The curves shown in Figure 1, therefore, cannot be applied in a literal or exact quantitative sense to another population base or to a different methodology. The groups used to generate the data in Casales series were highly selected; the asthmatic patients were young (age 18 to 45) and easily indenti- fied as allergic. The nonasthmatic subjects were also young (18 to 35), the nonsmokers among them had never smoked, and the smokers were all current smokers. A group of older subjects, some exsmokers, some nonallergic asthmatics, would certainly not give the same sensitivities and specificities.

tadacip 20 mg

## Leave a Reply