The reproducibility of adenosine monophosphate bronchial challenges in mild, steroid-naive asthmatics

Abstract

WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT

• Repeated adenosine monophosphate (AMP) challenges are used to assess drug effects in asthma clinical trials, but may be prone to tachyphylaxis when repeated at short intervals.

• Possible tachyphylaxis at 12- and 24-h intervals has not been studied.

WHAT THIS STUDY ADDS

• Clinically relevant tachyphylaxis after repeated AMP challenges does not occur when repeated at 12- and 24-h intervals.

• AMP challenges at these intervals can be used to assess drug effects in clinical trials.

AIMS

Repeated adenosine monophosphate (AMP) challenges are used to assess drug efficacy in clinical trials of mild, steroid-naive asthmatics. Refractoriness has been reported after repeated challenges over short intervals. This study evaluated possible tachyphylaxis after repeated AMP challenges at 12 and 24 h in mild, steroid-naive asthmatics.

METHODS

This was an open, three-way crossover study. Twenty-six steroid-naive asthmatic subjects were randomized to the following AMP challenge regimens separated by 7–14 days: (A) challenge at 08.00 h, repeated 24 h later; (B) challenge at 08.00 h, repeated 12 and 24 h later; (C) challenge at 20.00 h, repeated 12 h later. Comparisons within day were assessed using 90% confidence intervals (CIs). Non-inferiority approach taken with 1 doubling concentration (DC) as a clinically relevant difference.

RESULTS

Regimen A: Significant increase in AMP reactivity at 24 h. Mean DC difference was 0.6 (90% CI 0.24, 0.96). Regimen B: No evidence of difference between AMP reactivity at 08.00 h and a repeated challenge 12 h later. Repeated challenge at 24 h caused a significant increase in provocation concentration (PC)₂₀ compared with 12 h (mean DC difference 0.48, 90% CI 0.02, 0.95) and 0 h (mean DC difference 0.82, 90% CI 0.49, 1.14 – the upper CI exceeds the criteria of 1 DC). Challenge regimen C: No difference between challenges; mean DC difference of 0.28 (90% CI −0.2, 0.76).

CONCLUSION

The small decline in AMP reactivity during repeated challenges was not consistently observed, and was small compared with the known effects of inhaled drugs.

Introduction

Bronchial hyperreactivity (BHR) in patients with asthma can be assessed by the inhalation of agents that directly or indirectly cause smooth muscle constriction [1]. Methacholine and histamine act directly on smooth muscle cells to cause bronchoconstriction. In contrast, indirect agents alter inflammatory cell or neuronal activity, in turn leading to modulation of smooth muscle tone. Adenosine monophosphate (AMP) is an indirect agent that interacts with purinoreceptors on the surface of mast cells and neurons [1, 2]. Mast cells release substances such as histamine that cause bronchoconstriction, whereas neuronal signalling regulates airway smooth muscle tone.

Inhaled corticosteroids (ICS) exert airway anti-inflammatory effects in asthma. AMP challenges are one established method for evaluating the effects of ICS on BHR in clinical trials [1, 3–6]. These studies often administer repeated challenges on different days, making between-day reproducibility an important issue, especially in trial design and power calculations. The magnitude and duration of ICS activity could be assessed using a serial AMP challenge regime, with repeated challenges within a 24-h period. Such a study design would be critically dependent on the within-day reproducibility of repeated challenges. There is evidence that repeated administration of AMP within 1 h causes tachyphylaxis [7–9]. This phenomenon may be due to the depletion of histamine stores or downregulation of purinoreceptor expression. However, the duration of tachyphylaxis appears to be short; first, tachyphylaxis induced by repetitive challenges resolved after 2–6 h [7, 8], and second, tachyphylaxis was not induced when challenges were repeated at 2-h intervals [9]. These studies had small sample sizes (≤10 subjects), included subjects who did not have asthma [8], and used different challenge methodologies to the European Respiratory Society (ERS)-recommended AMP challenge protocol [1]. In order to use serial AMP challenges with the ERS-recommended method within a 24-h period to assess anti-inflammatory agents in asthmatics, it is necessary to assess whether tachyphylaxis occurs. We have assessed repeated AMP challenges in steroid-naive patients with asthma, focusing on whether there is evidence of tachyphylaxis when challenges are repeated at 12- and 24-h intervals.

Methods

Subjects

Twenty-six patients (17 male, mean age 35 years) with stable steroid-naive asthma participated. Inclusion criteria were nonsmokers or ex/current smokers with a history of <10 pack-years and using short-acting β-agonists only for asthma treatment. Participants were required to have forced expiratory volume in 1 s (FEV₁) > 60% predicted and demonstrate a provocation concentration (PC)₂₀AMP of <80 mg ml⁻¹ at screening. Exclusion criteria included an exacerbation of asthma or upper respiratory tract infection within 4 weeks prior to the study. Written informed consent and local ethics committee approval were obtained.

Study design

This was an open, randomized, three-way crossover study. Subjects were randomized to the following AMP challenge regimens; (A) two challenges, first at 08.00 h and then repeated 24 h later at 08.00 h; (B) three challenges, first at 08.00 h and then repeated 12 and 24 h later at 20.00 and 08.00 h, respectively; and (C) two challenges, first at 20.00 h and then repeated 12 h later at 08.00 h. Study days were separated by 7–14 days.

AMP challenge

Salbutamol was withheld 6 h prior to any visit. Smoking and caffeine were not permitted within 4 h before challenge or strenuous exercise within 12 h. FEV₁ was measured before and during the challenge using a dry wedge spirometer (Vitalograph®; Maids Moreton, UK). AMP challenges were performed only if the FEV₁ was ≥60% predicted. AMP was administered using a Mefar Dosimeter (Mefar-Bologna) calibrated to produce 10 µl output per inhalation. All inhalations were performed slowly from functional residual capacity to total lung capacity over 3 s followed by breath holding for 6 s. There were five inhalations at each dose level. FEV₁ was measured at 60 and 180 s and the highest recorded.

Initially, 0.9% saline control was used. Subjects with a post-saline decrease ≥10% on three consecutive occasions did not receive AMP. Subjects were then administered doubling concentrations (DC) of AMP from 0.32 to 320 mg ml⁻¹, until a ≥20% fall in FEV₁ was achieved or the highest concentration of AMP (320 mg ml⁻¹) was administered. The percentage fall in FEV₁ was plotted against the concentration of AMP on a logarithmic scale, and the provocative concentration of AMP required to reduce the FEV₁ by 20% of the post-saline baseline value (PC₂₀) was derived by linear interpolation.

Statistical methods

A power calculation was performed based on a within-subject standard deviation on the log₂ scale of 1.25 DC. A non-inferiority approach was used; a sample size of 20 subjects was estimated to provide 90% power to detect whether the upper confidence interval (CI) of the comparison 24 h (from challenge regimen A) compared with 0 h was <1 DC. Based on published data of the effects of drugs on AMP reactivity [3–6, 10], 1 DC was chosen as a clinically relevant difference between challenges.

If the highest concentration of AMP was administered before observing a 20% decrease in FEV₁, the PC₂₀ was recorded as 320 mg ml⁻¹. If the FEV₁ decreased by >20% after administration of the first concentration of AMP (0.32 mg ml⁻¹), a value of 0.32 mg ml⁻¹ was recorded as the PC₂₀. AMP PC₂₀ values were log₂ transformed before analysis and expressed as adjusted means after back transforming to the original scale. Between-day reproducibility was assessed by comparison of the first challenges at 08.00 h in challenge regimens A and B, using a 95% CI. If there was no significant difference observed between these challenges, it was decided a priori to pool these data at 08.00 h for analysis of within-day variability during challenge regimens A and B. Within-day variability was assessed in challenge regimens A, B and C using two-sided 90% CIs.

Results

At screening the mean FEV₁ was 3.2 l (86.6% predicted), and geometric mean PC₂₀ AMP was 10.9 mg ml⁻¹. Three of the 26 subjects were withdrawn during the study for the following reasons: exacerbation of asthma requiring a change in medication, an upper respiratory tract infection, and rhinitis. Data for these subjects were included in the analysis up to the time of withdrawal. For two other subjects, challenge regimen A data were incomplete due to technical problems experienced with the dosimeter.

One subject on four occasions had a 20% fall in FEV₁ after the first concentration of AMP was administered. Three subjects on a total of 14 occasions did not achieve a 20% fall in FEV₁ after the last AMP concentration was administered, although on each occasion the change in FEV₁ was close to a 20% fall, so it was decided, as per the a priori plan, that the highest concentration of AMP given was a reasonable estimate of the PC₂₀.

The adjusted geometric means for AMP challenges performed at 08.00 h on different days in challenge regimens A and B were very similar: 13.4 vs. 14.4 mg ml⁻¹, respectively. The mean DC difference between these challenges on different days was −0.1 (95% CI −0.57, 0.37), with the majority (14 of 21) of subjects with challenge results at 08.00 h within 1 DC (see Figure 1). The data at 08.00 h for regimens A and B was pooled for further analysis (see below) as per the predefined analysis plan.

Figure 1.

Doubling concentration difference (DCD) between challenges within regimens A, B and C. Also, differences between challenges at 08.00 h (0 h) during regimens A and B are presented in far right

Challenge regimen A

There was a numerical increase in the adjusted geometric mean PC₂₀ values for the repeated challenge at 24 h after an initial challenge at 08.00 h (see Figure 2). There was a small, statistically significant DC difference between the challenges (Table 1); however, the upper 90% CI did not reach the predefined criteria of a clinically important difference of 1 DC (Figure 3). Of 21 repeated challenges, 16 were within 1 DC (see Figure 1).

Figure 2.

Adenosine monophosphate (AMP) provocation concentration (PC)₂₀ (mg ml⁻¹) values during challenge regimens A and B. Error bars are 95% confidence limits

Table 1.

Within-day differences between challenges during regimens A, B and C

Challenge regimen	Time-point comparison	Adjusted mean DC difference	90% CI
A	24 vs. 0 h*	0.60	(0.24, 0.96)
B	12 vs. 0 h*	0.34	(−0.07, 0.74)
B	24 vs. 0 h*	0.82	(0.49, 1.14)
B	24 vs. 12 h	0.48	(0.02, 0.95)
C	24 vs. 12 h	0.28	(−0.20, 0.76)

^*0 h data pooled across regimens A and B.

Figure 3.

Mean doubling concentration difference (DCD) for within-day differences during challenge regimens A, B and C. Error bars are 90% confidence limits

Challenge regimen B

There were numerical increases in AMP PC₂₀ when challenges were repeated at 12 and 24 h after an initial challenge at 08.00 h (see Figure 2). The PC₂₀ at 12 h was not statistically significantly different compared with 0 h (Table 1, Figure 3). In contrast, the PC₂₀ at 24 h was significantly higher compared with 12 h and 0 h (Table 1). The upper 90% CI of the change at 24 h compared with 0 h (1.14 DC) also exceeded the predefined criteria of 1 DC for a clinically relevant difference (Figure 3). Of 25 repeated challenges, the majority were within 1 DC (Figure 1): 15 challenges at 12 h compared with 0 h, 17 challenges at 24 h compared with 0 h, 18 challenges at 24 h compared with 12 h.

Challenge regimen C

The geometric PC₂₀ means for challenges at 20.00 h followed by 08.00 h on day C were 14.3 vs. 17.4 mg ml⁻¹, respectively. This increase was not statistically significant (Table 1, Figure 3). Twelve of the 25 repeated challenges were within 1 DC (Figure 1).

Discussion

We observed no evidence of significant tachyphylaxis, defined a priori as the upper CI of the difference between challenges being >1 DC, for the majority of AMP challenges repeated at 12- or 24-h intervals in this study. The results during regimens A and B at the 24-h time point showed decreases in AMP reactivity that were statistically significant, as the lower 90% CI was greater than zero. However, only for the comparison of 24 h with 0 h during regimen B did the upper 90% CI exceed the predefined criteria of 1 DC for a clinically relevant difference. Overall, it can be concluded that the magnitude of any tachyphylaxis was small, and most likely to be observed after three consecutive challenges at 12-h intervals.

ICS are known to decrease AMP reactivity [3–6]. For example, after a single dose of fluticasone propionate 1000 µg, the DC difference at 2 h has been reported as 2.7 [5] and 2.9 [3], whereas at 14 h this has declined to 1.5 [3]. Similarly, a single dose of fluticasone 250 µg caused a 2.2-DC increase in reactivity at 2 h [5], whereas multiple doses of ICS can cause a >3-DC increase [4, 6]. The leukotriene antagonist montelukast causes a decrease in BHR of 1 DC [10]. For the current study, we therefore chose a non-inferiority criterion of 1 DC as a difference that was clinically relevant. Previous studies that have observed tachyphylaxis have used shorter intervals between challenges compared with the current study, usually ≤1 h, and have demonstrated that tachyphylaxis lasts for only 2–6 h [7–9]. The current study is consistent with these previous findings, as challenges repeated at 12- or 24-h intervals did not show significant tachyphylaxis in the majority of challenge regimens.

Diurnal variation is well recognized in asthma, and may cause changes in AMP reactivity [11]. In the current study, the adjusted PC₂₀ geometric mean for the first challenge in the morning (08.00 h) during regimens A and B was very similar to the first challenge in the evening (20.00 h) during regimen C [mean (95% CI) 13.9 (7.4, 26.3) vs. 14.3 (7.4, 27.6) mg ml⁻¹, respectively]. Overall, this suggests that diurnal variation in AMP reactivity in the current study did not occur.

Our a priori primary analysis plan was to use 90% CI to decide whether the upper threshold crossed a predefined limit of 1 DC, and the power calculation for this study was based on this concept. The 90% CI showed that there was no consistent pattern of significant tachyphylaxis after repeated challenges. The individual plots showed that the majority of repeated challenges were within 1 DC, further supporting our conclusion that repeated AMP challenges do not suffer from significant tachyphylaxis.

We used the ERS-recommended AMP challenge method [1]. This is a five-breath dosimeter method with the results presented as PC₂₀. However, clinical trials of anti-inflammatory drugs commonly express the ‘doubling dose’ effect of the drug on AMP challenge, even though the data are PC₂₀. So the terminology ‘doubling dose’ is not strictly correct; it should, in our opinion, be ‘doubling concentration’, which we have shortened to ‘DC’.

We studied mild, steroid-naive asthmatics, as these are the subjects often enrolled into clinical trials using AMP challenges to evaluate drug effects. Our results are relevant to this population group, but may differ in subjects with more severe disease or those using ICS.

The results of this study support the use of repeated AMP challenges at 12- or 24-h intervals in asthma clinical trials of anti-inflammatory drugs, e.g. to evaluate the duration of action. There was some evidence of a decline in AMP reactivity after three repeated challenges at 12-h intervals, but the magnitude of this effect was small compared with the known effects of ICS. This methodology study should encourage multiple AMP challenges to be performed at 12- or 24-h intervals without fear of clinically significant tachyphylaxis occurring.