Comparison of methacholine and mannitol challenges: importance of method of methacholine inhalation

Background Direct inhalation challenges (e.g. methacholine) are stated to be more sensitive and less specific for a diagnosis of asthma than are indirect challenges (e.g. exercise, non-isotonic aerosols, mannitol, etc.). However, data surrounding comparative sensitivity and specificity for methacholine compared to mannitol challenges are conflicting. When methacholine is inhaled by deep total lung capacity (TLC) inhalations, deep inhalation inhibition of bronchoconstriction leads to a marked loss of diagnostic sensitivity when compared to tidal breathing (TB) inhalation methods. We hypothesized that deep inhalation methacholine methods with resulting bronchoprotection may be the explanation for conflicting sensitivity/specificity data. Methods We reviewed 27 studies in which methacholine and mannitol challenges were performed in largely the same individuals. Methacholine was inhaled by dosimeter TLC methods in 13 studies and by tidal breathing in 14 studies. We compared the rates of positive methacholine (stratified by inhalation method) and mannitol challenges in both asthmatics and non-asthmatics. Results When methacholine was inhaled by TLC inhalations the prevalence of positive tests in asthmatics, 60.2% (548/910), was similar to mannitol, 58.9% (537/912). By contrast, when methacholine was inhaled by tidal breathing the prevalence of positive tests in asthmatics 83.1% (343/413) was more than double that of mannitol, 41.5% (146/351). In non-asthmatics, the two methacholine methods resulted in positive tests in 18.8% (142/756) and 16.2% (27/166) by TLC and TB inhalations respectively. This compares to an overall 8.3% (n = 76) positive rate for mannitol in 913 non-asthmatics. Conclusion These data support the hypothesis that the conflicting data comparing methacholine and mannitol sensitivity and specificity are due to the method of methacholine inhalation. Tidal breathing methacholine methods have a substantially greater sensitivity for a diagnosis of asthma than either TLC dosimeter methacholine challenge methods or mannitol challenge. Methacholine challenges should be performed by tidal breathing as per recent guideline recommendations. Methacholine (more sensitive) and mannitol (more specific) will thus have complementary diagnostic features.

generally normal, lung function. Stimuli used to measure AHR have been classified as direct and indirect [1]. Direct stimuli act directly on airway smooth muscle receptors; examples include methacholine acting on muscarinic receptors and histamine acting on H 1 receptors. Indirect stimuli act through one or more intermediate pathways most via mediators released from metachromatic inflammatory cells (mast cells, basophils); examples include exercise, eucapnic voluntary hyperpnea (EVH), non-isotonic aerosols, propranolol, adenosine monophosphate (AMP) and dry powder mannitol [2]. Direct AHR reflects airway smooth muscle function, perhaps modulated by inflammation, while indirect AHR reflects airway inflammation [1,2]. The consensus is that direct AHR is highly sensitive for current asthma whereas indirect AHR is highly specific while being relatively insensitive particularly for mild and/or well controlled asthma [2].
Dry powder mannitol (Aridol ® ) inhalation is an indirect challenge test [3] with several advantages. The advantages include the dose-response nature of the test (in contrast particularly to exercise and EVH), the lack of requirement for expensive and bulky equipment, and the fact that there is only a single method for administration of mannitol. In addition, we suspect that the mannitol challenge is less likely to be dose limited compared to other indirect challenges such as exercise, EVH, propranolol or AMP.
Studies comparing the diagnostic properties of the direct methacholine challenge and the indirect mannitol challenge have yielded conflicting results . Several studies show that the two challenges have unexpectedly comparable sensitivity for asthma [7,12,13,15] whereas other studies support the consensus that methacholine is more sensitive for a diagnosis of asthma [19,22,25,26,29]. A possible explanation is the observation from numerous studies that methacholine methods using a dosimeter with total lung capacity (TLC) inhalation (with a breath hold) demonstrate a marked loss of diagnostic sensitivity [30][31][32] due to deep inhalation bronchoprotection. This results in false negative challenges occurring in as many as 25% of overall methacholine tests and approaching 50% in asthmatics with mild AHR [33].
We hypothesized that deep inhalation methacholine methods with resulting bronchoprotection may be the explanation for conflicting sensitivity/specificity data. We have compared the diagnostic performance of the two challenges by examining studies where the two tests were performed in the same individuals (mostly) and where the methacholine inhalation method was clearly described.

Saskatoon studies
We began by identifying 46 unique individuals from four studies performed in our laboratory. We included the 20 subjects from the most recent study [29], 18 (of 20) additional subjects from a second study [26] and 8 (of 20) subjects from two allergen challenge studies [27,28]. For analysis we selected the first methacholine challenge performed in the four studies, the only mannitol study by the standard method [3] from 2 studies [26,29] and the pre-allergen mannitol challenge from the two allergen challenge studies [27,28]. The methacholine challenges were done with the two minute tidal breathing method [34] in three studies [26][27][28] and by the tidal breathing vibrating mesh nebulizer volumetric method (0.5 mL methacholine nebulized to completion, 1.5 to 2.5 min tidal breathing) [35] in one [29]. A normal result is a provocation concentration causing a fall in forced expired volume in 1 s (FEV 1 ) of 20% (PC 20 ) of > 16 mg/mL for the former method [34] and non-cumulative provocation dose causing a 20% FEV 1 fall (PD 20 ) of > 400 μg for the latter [35]. For analysis, PC 20 values were converted to PD 20 s based on the validated relationship that a PC 20 of 16 mg/mL equates to a post evaporation noncumulative PD 20 of 400 μg [35][36][37][38]. A normal (negative) mannitol result is a cumulative PD 15 > 635 mg [3]. Mannitol responsiveness was also assessed as the doseresponse slope (DRS) so that a value was available for all individuals. Fractional exhaled nitric oxide (FeNO) [39] was available for all individuals. Data were analyzed with a computerized statistics programme, (Statistix 9 Analytical Software, Tallahassee, FL, USA). All data were log transformed. Log methacholine PD 20 was compared to log mannitol DRS with linear regression and both log methacholine PD 20 and log mannitol DRS were regressed with log FeNO.

Other studies
Through a PubMed search, we identified 23 additional studies  that met the following criteria: 1. Mannitol testing was performed by the standardized protocol and results reported as the PD 15 [3]. 2. Methacholine challenges by various methods were done in the same subjects, with one exception where more subjects had methacholine tests than mannitol tests [25]. 3. The methacholine inhalation method was described. 4. The definitions of "asthma" and "non-asthma" were outlined.
When both methacholine TLC and methacholine TB studies were combined, the overall rate of a positive mannitol challenge in non-asthmatics was 8.3% or 76 of 913.

Discussion
These data provide strong support for the hypothesis that tidal breathing direct methacholine challenge methods yield results that are substantially more sensitive for asthma than does the indirect mannitol challenge. By contrast, when methacholine is inhaled by TLC methods, the diagnostic sensitivity falls to a level similar to that seen with mannitol.
AHR improves with anti-inflammatory therapeutic strategies including allergen avoidance environmental control [49,50] and ICS [51][52][53]. In keeping with the above observations, indirect AHR (AMP [49][50][51][52]) shows greater improvement with these treatments than does direct methacholine AHR. Mannitol responsiveness improves greatly after ICS treatment [53] and can provide a useful predictive marker of a pending asthma exacerbation during ICS tapering [54]. Although direct AHR has been proposed to monitor and guide asthma treatment [55], indirect AHR may provide a particularly valuable tool as a guide to monitoring asthma control [56]. In fact, non-responsiveness to indirect challenge (e.g. AMP, mannitol) may be a goal for adequate asthma control with ICS [56]. This, of course, is consistent with a positive indirect AHR challenge (including mannitol) being insensitive for the diagnosis of well controlled asthma.
Deep inhalations to TLC produce potent bronchodilation and bronchoprotection, the latter greater than the former, in normal individuals but initially stated to not occur in asthmatics [57]. It had become apparent that this marked bronchoprotective   effect extends to mild asthmatics [30][31][32][33] and, in all likelihood may well extend to well controlled asthmatics.
Although not seen in all studies [58], eosinophilic airway inflammation impairs the bronchoprotective effect of deep inhalation [26,59,60]. Anti-inflammatory strategies, both allergen avoidance [61] and oral/inhaled corticosteroid [62], can restore or improve the deep inhalation bronchoprotection in asthmatics. In one study, lack of bronchoprotection (methacholine) and elevated levels FeNO as an indirect measure of airway inflammation were associated with indirect AHR to mannitol [26]. Collectively, these data suggest that airway inflammation (eosinophilic particularly), indirect AHR and loss of deep inhalation bronchoprotection will occur together in asthmatics. Conversely, deep inhalation bronchoprotection and low levels of airway inflammation will be associated with little if any indirect AHR [26]. Avoidance of TLC inhalations during methacholine inhalation will therefore result in many more positive direct challenge tests in mild (and possibly well controlled) asthmatics with no indirect AHR and minimal airway inflammation. This is confirmed by our current review.
Deep inhalation bronchoprotection during methacholine challenges is an important and underappreciated phenomenon [33]. This has been shown by three studies from our laboratory [30][31][32] and supported by studies from other laboratories [63,64]. This was first suggested in a study of 40 individuals [30] comparing the two methacholine methods outlined in the ATS document [40]. Follow up investigations demonstrated that asthmatics with negative TLC dosimeter methacholine tests had positive challenges when the identical dosimeter dose was administered with sub-maximal inhalations (approximately half TLC) [31] and that many asthmatics with positive tidal breathing methacholine challenges were negative when five TLC breaths were incorporated at equal intervals throughout the 2 min of tidal breathing [32]. These latter two studies provide convincing evidence of the bronchoprotective effect of deep TLC inhalations in many individuals with mild asthma. Our summary data from 55 asthmatic individuals with positive tidal breathing methacholine tests revealed that 13 (24%) had negative five TLC breath dosimeter methacholine tests [33]. This represents 50% of asthmatics with a tidal breathing PC 20 between 2 and 16 mg/mL (post evaporation non-cumulative PD 20 between 50 and 400 μg). This is exactly the range where a positive diagnostic methacholine challenge, done in individuals with symptoms suggestive of asthma and normal spirometry, is likely to fall. In this population, the TLC dosimeter methacholine method could, therefore, produce a false negative rate approaching 50% for individuals with asthma and mild AHR. For these reasons the recent methacholine guidelines have strongly suggested that methacholine challenges be performed with tidal breathing methods with a non-TLC dosimeter method as a second option [36]. By contrast, as anticipated by the above data, our recent study documented that removal of TLC inhalations from the mannitol challenge did not affect the result [29].
It is difficult to accurately comment on sensitivity and specificity of the different tests from the available references. A reasonable estimate of diagnostic sensitivity can be made by assessing the rate of positivity in subjects determined to have asthma. Based on this approach the tidal breathing methacholine test is about twice as sensitive for "asthma" as the mannitol test (83.1% and 41.5% respectively) in the studies assessed, whereas the sensitivities of TLC methacholine and mannitol tests were similar, at approximately 60% for both in the studies included. These data suggest that the loss of diagnostic sensitivity of the methacholine test when using a TLC dosimeter method is significant enough to make the sensitivity equivalent to an indirect challenge. It is even more difficult to comment accurately on specificity without a larger cohort of normal nonasthmatic individuals. The observation that there were fewer positive mannitol tests (about half ) compared to methacholine tests in non-asthmatics is consistent with the consensus that indirect challenges, including mannitol, are more specific for asthma [2,65]. The difficulties are further compounded both by the lack of an independent gold standard for the diagnosis of asthma and by the requirement for the symptoms under investigation to be clinically current, i.e. within the past few days [65,66].
We suspect that these results would translate to indirect challenges other than mannitol; these include AMP, propranolol, hypertonic saline, EVH and exercise (EIB). It is likely that all these indirect challenges would show minimal if any deep inhalation bronchoprotection. EVH and EIB are particularly important. It would, however, be difficult to design a study with and especially without deep inhalations for these two, especially for EVH.
Indirect challenges require a substantially larger dose of stimulus than direct challenges, up to or greater than three orders of magnitude mg for mg or mmol for mmol [65]. For example, the top doses for mannitol and methacholine are 635 (cumulative) and 0.4 mg (noncumulative) respectively. It is possible that mannitol might be more sensitive than many other indirect stimuli because the challenge is less likely to be "dose limited" [65]. There are physiologic limits on the "dose"  16:14 of stimulus that can be achieved with exercise or EVH, and, because of the large doses needed, a solubility limit on the doses that can be achieved with AMP or propranolol [65]. Mannitol, by contrast, is a dry powder inhalation and the dose is not limited by solubility. There is only one mannitol inhalation method [3]. However, the large number of different methacholine methods represents a difficulty when attempting to compare data. A conservative estimate is that there were at least 6 different TLC dosimeter methods and 4 different TB methods in the studies evaluated. The best case estimate is that these methods equated to a post-evaporation methacholine PD 20 range of only twofold (200-400 μg), however that is speculation without knowledge of the operating characteristics of the different nebulizers used.

Conclusion
The discordance between methacholine and mannitol comparisons can be explained by the method of methacholine inhalation. Tidal breathing methacholine tests are substantially more sensitive than mannitol tests for a diagnosis of asthma and equally more sensitive than TLC dosimeter methacholine methods. In order to preserve a high diagnostic sensitivity, methacholine challenges should be performed by tidal breathing [33,36,65], thus providing data that are complementary to the more specific mannitol challenge.