Contents lists available at ScienceDirect

Ann Allergy Asthma Immunol 114 (2015) 58e62

Omalizumab therapy for asthma patients with poor adherence toinhaled corticosteroid therapy

Leslie Hendeles, PharmD *,y; Yasmeen R. Khan, MD *; Jonathan J. Shuster, PhD z;Sarah E. Chesrown, MD y; and Mutasim Abu-Hasan, MD y

*Department of Pharmacotherapy and Translational Research, College of Pharmacy, University of Florida, Gainesville, Floriday Pulmonary Division, Department of Pediatrics, University of Florida, Gainesville, FloridazDivision of Biostatistics, Department of Epidemiology and Health Policy Research, College of Medicine, University of Florida, Gainesville, Florida

A R T I C L E I N F O

Article history:Received for publication July 31, 2014.Received in revised form September 23,2014.Accepted for publication October 14, 2014.

A

BmthO

Reprints: Leslie Hendeles, PharmD, PO Box 1004ville, FL 32610-0486; E-mail: hendeles@cop.ufl.edDr Khan is now an allergist in private practice inDisclosures: The study was funded by an investigand Dr Hendeles served as the Principal Invesnothing to disclose.Funding: This study was funded by an investigaPharmaceuticals Corporation and National InstituResearch Resources grant M01-RR00082 (Dr ShHealth Research Facilities Construction Program C

http://dx.doi.org/10.1016/j.anai.2014.10.0121081-1206/� 2015 American College of Allergy, A

B S T R A C T

ackground: Omalizumab, an anti-IgE monoclonal antibody, is administered by injection once or twiceonthly in offices and clinics. It offers a potential alternative intervention for patients with allergic asthmaat is not well controlled because of recalcitrant poor adherence to inhaled corticosteroid therapy.bjective: To assess the effect of omalizumab therapy by measuring airway responsiveness to adenosine, a

marker of allergic airway inflammation, and resource use.Methods: Patients (N¼ 17) aged 6 to 26 years (mean age, 16.4 years) with poorly controlled persistent allergicasthma, less than 50% adherence to inhaled corticosteroid therapy, a forced expiratory volume in 1 second(FEV1) of 60% predicted or higher, and adenosine provocation concentration that caused a decrease in FEV1 of20% (PC20) of 60 mg/mL or less were randomized to receive 4 months of omalizumab or placebo in a double-blind, crossover trial with a 3- to 4-month washout between treatments. Patients were instructed to continuetaking inhaled corticosteroids throughout the study. The PC20 was measured before and after each period.Results: Fifteen patients completed the study. The mean baseline PC20 was 14.1 mg/mL (95% CI, 10.8e18.4mg/mL). The fold change PC20 was 0.9 (95% CI, 0.5e1.7) during placebo and 3.1 (95% CI, 1.6e6.2) duringomalizumab treatment; the estimated ratio was 3.4 (95% CI, 1.2e9.3; P ¼ .02). Six patients required one ormore short courses of oral corticosteroids for asthma exacerbations during placebo, but none required thisintervention during omalizumab. During the study, the median prescription refills for inhaled corticosteroidswas 0.15 (95% CI, 0.00e0.33) canisters per month.Conclusion: Omalizumab therapy is an alternative for patients with more severe poorly controlled asthmain whom adherence does not improve with conventional interventions.Trial Registration: clinicaltrials.gov Identifier: NCT00133042.� 2015 American College of Allergy, Asthma & Immunology. Published by Elsevier Inc. All rights reserved.

Introduction increased risk of mortality from asthma.4 Conventional in-

Poor adherence to treatment with asthma controller medica-tions, such as inhaled corticosteroids (ICSs), is frequent amongasthma patients.1 This behavior may contribute to worsenedclinical outcomes, including increased need for short courses oforal corticosteroids,2 increased risk of hospitalization,3 and

86, University of Florida, Gaines-u.Ft Lauderdale, Florida.ator-initiated grant from Novartistigator. The other authors have

tor-initiated grant from Novartistes of Health National Center foruster) and National Institutes of06 grant RR17568.

sthma & Immunology. Published by E

terventions for poor adherence include removing barriers toadherence, home visits, patient education, and school-basedasthma care. However, in many patients these interventions maynot be successful.5

Previously, we tried to improve adherence in patients with highresource use by having a nurse make daily home visits to supervisecontroller medication administration.6 This reduced total hospi-talization days for asthma from 70 in the year before intervention to24 in the year after the intervention among 7 children with verypoor asthma control because of poor adherence to ICS therapy.However, Florida Medicaid and other third-party payers no longercover payments for this method of intervention. Accordingly, wehypothesized that once- or twice-monthly administration ofomalizumab (Xolair, Genentech Inc, South San Francisco, California,and Novartis Pharmaceuticals Corporation, East Hanover, NewJersey) would circumvent the challenge of daily adherence and,

lsevier Inc. All rights reserved.

Figure 1. Placebo or omalizumab (OMA) was administered every 2 or 4 weeks.Forced expiratory volume in 1 second (FEV1) was measured at every study visit; theprovocation concentration that caused a decrease in FEV1 of 20% (PC20) to adenosinemonophosphate challenge was measured before and after each treatment period.

Table 1Patient disposition

Variable No. ofpatients

Total screened 104Total screen failures 87Reasons for screen failureNo positive allergens 16FEV1 <60% predicted 12Combination of total body weight and IgE that would require dosage

of omalizumab >375 mg every 2 weeks12

Unable to perform ATS acceptable and reliable spirometry 9IgE <30 IU/mL 9Lack of evidence for poor asthma control 8PC20 >60 mg/mL 8Othera 13

Total randomized 17Total completed both treatment periods 15Discontinued because of relocation out of area 2

Abbreviations: ATS, American Thoracic Society; FEV1, forced expiratory volume in 1second; PC20, provocation concentration that caused a decrease in FEV1 of 20%.aIncludes failure to return, abnormal electrocardiography results, positive result forillicit drugs, abnormal laboratory values, or inability to withhold medications asrequired by the protocol.

L. Hendeles et al. / Ann Allergy Asthma Immunol 114 (2015) 58e62 59

thus, improve outcomes in patients whose asthma was not wellcontrolled because of poor adherence to ICS therapy.

Omalizumab is an anti-IgE monoclonal antibody that bindscirculating free IgE with a subsequent reduction in the number ofhigh-affinity receptors on mast cells and thereby decreases mastcell release of inflammatory mediators in asthma.7 Adenosinemonophosphate (AMP) enhances the release of inflammatory me-diators from activated mast cells, and airway responsiveness toAMP is amarker of allergic airway inflammation.8,9 By using AMP asa surrogate of clinical effectiveness, we were able to improve thepower of the study while minimizing the number of patientsneeded for a single-center study. As an indicator of asthma control,the number of prednisone bursts required during the study was asecondary outcome measure.

Methods

Patients

This study included patients (ages 6e26 years) with persistentasthma for whom ICSs were prescribed for at least 3 months, eitheralone or in combinationwith a long-acting b2-agonist or leukotrienemodifier. They had poor asthma control defined by any of thefollowing: forced expiratory volume in 1 second (FEV1) less than 80%predicted, short-acting b-agonist use more than 3 times per week,nocturnal symptomsmore than2 times permonth, exercise-inducedbronchospasm from activities of daily living, unscheduled physicianvisits or hospitalization for asthma, or more than 1 prednisone burstin previous 3 months. Other inclusion criteria were a pharmacyprescription refill history of less than 50% of prescribed doses of ICSfor 3months or longer; sensitization to one ormore indoor allergensoroutdoorAlternaria; total IgE level of 30 to 700 IU/mL for patients 12years or older or up to 1,300 IU/mL for those 6 to 12 years of age;baseline FEV1 of 60% predicted or more; and a 20% decrease in FEV1

after inhaling 60 mg/mL or more of AMP (ie, PC20 FEV1 �60 mg/mL).Patients were excluded if they had smoked in the past 12

months or had a smoking history of more than 10 pack-years, werepregnant or lactating, had a respiratory tract infection in the past 6weeks, or had an omalizumab dosage requirement of more than375 mg every 2 weeks.

This study was conducted under an investigator-sponsoredInvestigational New Drug Application approved by the US Foodand Drug Administration (No. 70,241) for use of AMP challenge andstudy of children younger than 12 years and was approved by theUniversity of Florida Institutional Review Board. All patients or par-ents gavewritten informed consent, and children gave verbal assent.

Study Design

This was a randomized, double-blind, 3-period, placebo-controlled, crossover study (Fig 1). Patients received omalizumabor placebo by subcutaneous injection every 2 or 4 weeks for 4months followed by a 3- to 4-month washout period and then 4months of the opposite treatment. FEV1 was measured at eachtreatment visit; AMP PC20 was measured before and after eachtreatment period (see eAppendix for details). Health care use wasrecorded at each visit, and prescription refill histories were ob-tained from their pharmacies after the screening visit and ondischarge from the study. Patients were not asked to measure peakflow or record symptoms in a diary because they were poorlyadherent to ICS therapy and it was assumed that they would notreliably record in a diary.

Statistical Analysis

A sample size of 16 patients was calculated (based on reportedreproducibility of AMPPC20 in 13 patientswith asthmawith a log SDof 0.4)10 to provide 95% power to detect a 2-fold difference in PC20

between treatments. The regression method of Shuster11 was usedto compare the treatments. This method takes the period 2 minusperiod 1 difference (irrespective of treatment assignment) andcompares the 2 treatment orderings. The effect size estimate issuperior to the 1-sample t test in that it is unbiased when the actualsample sizes assigned to the orderings differ, is more efficient, andadjusts for carryover effects. The dependent primary variable wasthe difference in the change in natural log final PC20 minus thebaseline (period 2 minus period 1). Note that 2 patients had post-dose values that could not be ascertained, except they were knownto be higher than 200 mg/mL, the highest AMP concentrationadministered. These values were assigned 200 mg/mL. FEV1 wascompared in the same way, except logs were not used. A fittedregressionmodel was used to determine the prognostic importanceof baseline PC20 and FEV1 on change in PC20 during active treatment.The number of corticosteroid bursts was compared by the Friedmantest. Median ICS use per month was assessed by the sign test forobtaining 95% CI. P < .05 was considered statistically significant.

Results

Patients

Of 104 patients screened, 17 were randomized and 15 completedboth treatment periods (Table 1). The most common reasons for

L. Hendeles et al. / Ann Allergy Asthma Immunol 114 (2015) 58e6260

screen failure included no positive allergens detected by blood testfor specific allergens (ImmunoCAP) (n ¼ 16), FEV1 less than 60%predicted (n ¼ 12), and a combination of body weight and total IgEthatwould require an omalizumab dosage higher than 375mg every2 weeks (n ¼ 12). The 2 randomized patients who did not completethe study discontinued because they moved out of the area.

Among the 17 patients randomized to treatment (10 femalepatients and 7 male patients), the mean (SD) age was 16.4 (5.5)years; 5 patients were aged 6 to 12 years. The mean baseline FEV1was 83.7% predicted, geometric mean PC20 was 14.1 mg/mL, meantotal IgE level was 427 IU/mL (range, 95e956 IU/mL), and doses ofomalizumab ranged from 300 to 375 mg (Table 2).

Primary and Secondary End Points

In the 15 patients who completed the study, the geometricmean PC20 increased from 10.8 to 33.9 mg/mL during omalizumabtreatment and decreased from 20.1 to 18.5 mg/ml during theplacebo period (Table 3). Thus, the primary end pointdgeometricfold change in PC20 from baseline to end of treatmentdwassignificantly improved with omalizumab vs placebo (3.1 vs 0.9,P ¼ .02; Fig 2). On the basis of the regression analysis, the pointinterval for the omalizumab-placebo ratio of fold changes(geometric means) was 3.4 (95% CI, 1.23e9.25). Per protocol, thewashout period was extended 1 month when the PC20 did notreturn to baseline after washout. This occurred in 5 of the 8patients who received omalizumab treatment first, comparedwith 2 of the 8 patients who received placebo first. Change in PC20during omalizumab treatment revealed no association withbaseline values of either FEV1 or PC20. In addition, there was nosignificant change in FEV1 during either treatment (the mean FEV1value increased by 6% during both placebo and omalizumabtreatment periods) (Table 3).

Six patients required at least one prednisone burst during pla-cebo administration (5 patients required 1 burst each and 1 patientrequired 2 bursts during the 4-month period); however, nonerequired prednisone during omalizumab treatment. One patientrequired an asthma-related emergency department visit whilereceiving placebo treatment (the same patient who required 2prednisone bursts during the placebo period and 1 burst duringwashout). There were no emergency department visits during theomalizumab treatment period and no asthma-related

Table 2Patient demographics and baseline characteristics

Characteristic Randomized patients (n ¼ 17)

Age, mean (SD), y 16.4 (5.5)Sex, No.Female 10Male 7

Race, No. (%)White 11 (64)African American 4 (24)Asian 1 (6)Hispanic 1 (6)

Weight, mean (SD), kg 63.0 (20.2)FEV1, mean (SD), % predicted 83.7 (11.8)PC20, geometric mean (95% CI), mg/mL 14.1 (10.8e18.4)Total IgE, mean (SD), IU/mL 427 (275)ICS refills per month,a median (95% CI) 0.17 (0.14e0.24)Calculated omalizumab dose for study, mean(SD), mg

313 (38)

Frequency, No.Every 2 weeks 10Every 4 weeks 7

Abbreviations: CI, confidence interval; ICS, inhaled corticosteroid; FEV1, forcedexpiratory volume in 1 second; PC20, provocation concentration that caused adecrease in FEV1 of 20%.aTwelve months before study entry.

hospitalizations throughout the study. We did not observe a sea-sonal pattern to these exacerbations.

The median number of refills per month for a 1-month supply ofICSs was 0.15 (95% CI, 0.00e0.33) throughout the study, similar tothe 12 months before the study of 0.17 (95% CI, 0.12e0.33). Themedian paired difference (during minus before) was 0.04 (95%CI, �0.17 to þ0.17; P ¼ .99), reflecting continued poor adherence toICS therapy, despite instructions to continue ICS therapy.

Adverse Events

Three patients reported serious adverse events that resulted inemergency department visits: 2 during placebo administration(only 1 was asthma related) and 1 during the washout period. Noserious adverse events occurred during treatment with omalizu-mab. Nonserious adverse events were reported by 10 patientsduring both treatment periods: 2 patients only during placeboadministration and 4 patients only during omalizumab treatment.None of these adverse events were considered to be related tostudy treatment. None of the patients spontaneously offeredcomments about local injection site reactions. However, patientswere asked open-ended questions at each visit rather than specificquestions on whether they had experienced an injection site re-action from the previous visit.

Discussion

In this randomized, double-blind, crossover study in patientswith poor asthma control and prior evidence of poor ICS adherence,omalizumab significantly increased adenosine PC20, a marker ofairway inflammation, compared with placebo. In addition, none ofthe patients required prednisone for exacerbations of asthma duringomalizumab treatment, whereas 6 patients required this interven-tion while taking placebo. Interestingly, pharmacy refill rates forprescribed ICS therapy remained unchanged during the study,indicating persistently poor adherence. It is noteworthy that 12patients required 1 ormore emergency department visits for asthmain the year before this study, whereas only 1 patient required anasthma-related emergency department visit during the study(occurring during the placebo period). This reductionmay have beena result of providing a treatment plan that included supplying al-buterol metered-dose inhalers and prednisone to keep on hand,alongwith telephone access to a study coordinator duringweekdaysand to a study physician during nights and weekends who initiatedprednisone therapy for bronchodilator-unresponsive symptoms.

FEV1 did not significantly improve during the treatment periodwith omalizumab, which is consistent with other omalizumabclinical trials of similar duration.4,12,13 However, there was notmuch room for improvement because the mean baseline FEV1 was84% predicted. It is not known whether the reduction in asthmaexacerbations would decrease the rate of decline in lung functionthat may occur over time; however, long-term studies would helpanswer this question.

Prieto et al14 compared the effects of omalizumab on airwayresponsiveness to methacholine and AMP in patients with mild tomoderate allergic asthma in a randomized, placebo-controlled,parallel-group study. In that study, improvement in AMP PC20 wassignificantly greater in the omalizumab vs placebo group after 4weeks of treatment (PC20 increased by 1.92 doubling concentrationsin the omalizumab group vs 0.41 doubling concentrations in theplacebo group, P ¼ .02). However, after 12 weeks of treatment, theincreased PC20 in the omalizumab group was sustained, but im-provements in the placebo group were such that the differencebetween groups was no longer significant (increased PC20 frombaseline of 1.91 and 1.01 doubling concentrations in the omalizumaband placebo groups, respectively; P ¼ .24). This lack of significantdifference at 12 weeks may have been the result of too small of a

Table 3Individual and mean results for spirometry and adenosine challenge for the 15 patients who completed both the placebo and omalizumab treatment periods

Patient No. Placebo Omalizumab

FEV1, L PC20 FEV1, L PC20

Before End Change, % Before End Fold changea Before End Change, % Before End Fold changea

Order: omalizumabthen placebo1 1.85 1.96 6 146.4 19.93 0.1 1.90 1.72 �9 10.3 182.34 17.84 3.68 3.96 8 60.8 72.09 1.2 4.02 3.98 �1 39.8 30.46 0.89 2.08 2.65 27 17.1 20.36 1.2 2.47 2.43 �2 12.2 35.74 2.914b 2.21 2.42 10 16.6 15.89 1.0 2.01 1.99 �1 13.8 117.28 8.516 3.44 3.64 6 31.2 29.34 0.9 2.69 2.87 7 9.22 4.81 0.521 2.48 2.69 8 4.40 16.24 3.7 2.61 2.50 �4 9.67 4.93 0.536 4.42 4.31 �2 85.8 22.53 0.3 4.20 4.58 9 19.2 65.06 3.4

Order: placebo thenomalizumab3 2.63 2.89 10 19.7 54.23 2.8 2.70 2.73 1 65.2 >200 3.16b 4.13 4.26 3 12.3 7.80 0.6 3.59 4.82 34 3.10 15.51 5.010b 2.55 2.81 10 10.9 5.23 0.5 2.39 3.22 35 2.45 5.93 2.413b 3.14 3.25 4 9.30 52.78 5.7 2.88 3.14 9 5.17 141.30 27.327b 2.81 2.60 �7 5.50 2.64 0.5 2.27 2.40 6 1.05 4.45 4.230 3.23 3.19 �1 40.8 158.09 3.9 3.23 3.35 4 27.8 >200 7.234 2.44 2.26 �7 15.4 2.91 0.2 2.54 2.47 �3 20.2 90.51 4.538b 1.80 2.02 12 13.9 14.81 1.1 1.96 2.16 10 17.6 11.86 0.7

Mean (SD) 2.86 (0.80) 2.99 (0.76) 6 (9) 2.76 (0.71) 2.96 (0.91) 6 (13)Geometric mean(95% CI)

20.1 (11.7e34.6) 18.5 (9.8e35.1) 0.9 (0.5e1.7) 10.8 (5.9e19.7) 33.9 (15.0e76.8) 3.1 (1.6e6.2)c

Abbreviations: CI, confidence interval; FEV1, forced expiratory volume in 1 second; PC20, provocation concentration that caused a decrease in FEV1 of 20%.aFold change is the PC20 at the end divided by the PC20 at the beginning of the study period. The log of these values was used in the statistical analysis.bReceived prednisone during placebo. Patient 10 required 2 courses.cSignificantly greater than placebo (P ¼ .02).

L.Hendeles

etal./

Ann

Allergy

Asthm

aIm

munol

114(2015)

58e62

61

Figure 2. Geometric mean fold change (D) in adenosine provocation concentrationthat caused a decrease in FEV1 of 20% (PC20) from baseline to end of treatmentperiod for omalizumab (B) vs placebo (C). *P ¼ .02 for omalizumab vs placebo.

L. Hendeles et al. / Ann Allergy Asthma Immunol 114 (2015) 58e6262

sample size for the parallel design, in contrast to the crossoverdesign of our current study, which is statistically more powerful.

A potential limitation of this study is the observation that theAMP PC20 did not return to baseline after the washout period for5 of the 8 patients who received omalizumab first. Although a 3-month washout interval after omalizumab therapy was thoughtto be sufficient based on a previous report of airway respon-siveness to acetylcholine,15 it appears that a longer washoutperiod may have been needed. However, the use of AMP PC20 atthe start of each treatment period to calculate the change in PC20after 4 months and the regression analysis, which adjusts for acarryover effect, compensated for this. In addition, patients werenot asked to measure peak flow or report daily symptoms or useof albuterol, all important measures of asthma control.16 It wasour concern that data would be missing and make interpretationdifficult. Rather, the emphasis was on collection of objectivemeasures, such as airway responsiveness to adenosine and FEV1,along with intervention with prednisone and other health careuse as more reliable measures of impairment and risk. In retro-spect, it would have been important to measure exhaled nitricoxide during the study, a marker of eosinophilic airway inflam-mation.17 However, we did not have that capability until the endof this study.

There are few proven methods of improving adherence toasthmamedications. Some of them involve removing barriers, suchas cost of medication or remembering to take the medication, andothers focus on changing patient behavior through interviews.5

Because omalizumab is extremely expensive ($12,000 to $30,000per year), this intervention should only be considered when con-ventional methods of dealing with poor adherence fail and thepatient is at risk for severe outcomes. The goal is to circumvent thechallenge of requiring daily adherence to decrease resource use andpossibly even death in high-risk patients. In this circumstance, thepotential benefit far outweighs the risks. Local reactions at the in-jection site are uncommon; true anaphylaxis is rare,18 and the mostrecent observation study (Epidemiologic Study of Xolair [omalizu-mab]: Evaluating Clinical Effectiveness and Long-term Safety inPatients with Moderate-to-Severe Asthma) indicates that omali-zumab does not increase the risk of a malignant tumor.19 Never-theless, a cost-benefit analysis of this alternative is needed. Inconclusion, omalizumab is an alternative therapy for patients with

very poor asthma control who continue to have poor adherence toICSs after conventional interventions.

Acknowledgments

We thank Kathy Rice for word processing and editing the typescriptand remember the late Carmen Lowell, RT, who performed thechallenges.

Supplementary Data

Supplementary data related to this article can be found online athttp://dx.doi.org/10.1016/j.anai.2014.10.012.

References

[1] Sherman J, Patel P, Hutson A, Chesrown S, Hendeles L. Adherence to oralmontelukast and inhaled fluticasone in children with persistent asthma.Pharmacotherapy. 2001;21:1464e1467.

[2] Milgrom H, Bender B, Ackerson L, Bowry P, Smith B, Rand C. Noncomplianceand treatment failure in children with asthma. J Allergy Clin Immunol. 1996;98:1051e1057.

[3] Donahue JG, Weiss ST, Livingston JM, Goetsch MA, Greineder DK, Platt R.Inhaled steroids and the risk of hospitalization for asthma. JAMA. 1997;277:887e891.

[4] Suissa S, Ernst P, Benayoun S, Baltzan M, Cai B. Low-dose inhaled cortico-steroids and the prevention of death from asthma. N Engl J Med. 2000;343:332e336.

[5] Apter AJ, Bender BG, Rand CS. Adherence. In: Adkinson NF, Bochner BS,Busse WW, Holgate ST, Lemanske RF, Simons FE, eds. Middleton’s Allergy. 7thed. Philadelphia, PA: Elsevier Inc; 2009:1473e1483.

[6] Sherman JM, Baumstein S, Hendeles L. Intervention strategies for childrenpoorly adherent with asthma medications; one center’s experience. ClinPediatr. 2001;40:253e258.

[7] Hendeles L, Sorkness CA. Anti-immunoglobulin E therapy with omalizumabfor asthma. Ann Pharmacother. 2007;41:1397e1410.

[8] Van Den Berge M, Meijer RJ, Kerstjens HA, et al. PC20 adenosine 5’-mono-phosphate is more closely associated with airway inflammation in asthmathan PC20 methacholine. Am J Respir Crit Care Med. 2001;163:1546e1550.

[9] De Meer G, Heederik D, Postma DS. Bronchial responsiveness to adenosine 5’-monophosphate (AMP) and methacholine differ in their relationship withairway allergy and baseline FEV1. Am J Respir Crit Care Med. 2002;165:327e331.

[10] Crimi N, Palermo F, Polosa R, et al. Effect of indomethacin on adenosine-induced bronchoconstriction. J Allergy Clin Immunol. 1989;83:921e925.

[11] Shuster JJ. Design and analysis of experiments. In: Ambrosius WT, ed. Topics inBiostatistics. Totowa, NJ: Humana Press; 2007:235e259.

[12] Solèr M, Matz J, Townley R, et al. The anti-IgE antibody omalizumab reducesexacerbations and steroid requirement in allergic asthmatics. Eur Respir J.2001;18:254e261.

[13] Busse W, Corren J, Lanier BQ, et al. Omalizumab, anti-IgE recombinant hu-manized monoclonal antibody, for the treatment of severe allergic asthma.J Allergy Clin Immunol. 2001;108:184e190.

[14] Prieto L, Gutiérrez V, Colás C, et al. Effect of omalizumab on adenosine 5’-monophosphate responsiveness in subjects with allergic asthma. Int ArchAllergy Immunol. 2006;139:122e131.

[15] Noga O, Hanf G, Kunkel G. Immunological and clinical changes in allergicasthmatics following treatment with omalizumab. Int Arch Allergy Immunol.2003;131:46e52.

[16] National Asthma Education and Prevention Program. Expert Panel Report 3:Guidelines for the Diagnosis and Management of Asthma. Full Report 2007.www.nhlbi.nih.gov/guidelines/asthma/asthgdln.htm. Accessed April 14, 2014.

[17] Dweik RA, Boggs PB, Erzurum SC, et al. American Thoracic Society Committeeon Interpretation of Exhaled Nitric Oxide Levels (FENO ) for Clinical Appli-cations. An official ATS clinical practice guideline: interpretation of exhalednitric oxide levels (FENO) for clinical applications. Am J Respir Crit Care Med.2011;184:602e615.

[18] Cox L, Platts-Mills TA, Finegold I, Schwartz LB, Simons FE, Wallace DV.American Academy of Allergy, Asthma & Immunology; American College ofAllergy, Asthma and Immunology. American Academy of Allergy, Asthma &Immunology/American College of Allergy, Asthma and Immunology JointTask Force Report on omalizumab-associated anaphylaxis. J Allergy ClinImmunol. 2007;120:1373e1377.

[19] Long A, Rahmaoui A, Rothman KJ, et al. Incidence of malignancy in patientswith moderate-to-severe asthma treated with or without omalizumab.J Allergy Clin Immunol. 2014;134:560e567.

L. Hendeles et al. / Ann Allergy Asthma Immunol 114 (2015) 58e62 62.e1

eAppendix. Adenosine Challenge Procedure

University of Florida Asthma Research Laboratory

A. Equipment1. SpirometryandchallengeareperformedwithaKoKoDigidoser

(Ferraris Respiratory, Louisville, CO). The adenosine challengeis a modification of the American Thoracic Society (ATS) 5-breath dosimeter method1 and the procedure of Taylor et al.2

2. Aerosol is generated with a “characterized” DeVilbiss 646Nebulizer (No. 492249, Ferraris Respiratory). Nebulizers arenumbered, and the same nebulizer is used for all challengesfor a given patient. The nebulizer is designed to deliver par-ticleswith an aerodynamicmassmedian diameter of 1.8 mmata constant air pressure of 30 psi and a constant inspiratoryflow (approximately 0.5 L/s). The Digidoser is connected to acompressed air cylinder through an adjustable pressure valve.The output (in milliliters per minute), inscribed on thenebulizer, is entered into the challenge software program. Thesoftware is set to deliver the dose for 0.6 second. With thissystem, the output of the nebulizer cannot be determinedaccurately or calibrated according to the ATS guidelines.3

3. Spirometer calibration is performed with a standard 3-L sy-ringe. The KoKo filter inlet is attached to the outlet on thesyringe, and the syringe handle is pulled all the way out. Theprocedure begins by allowing the Pneumotach to registerbaseline airflow when lying on the table without movement(nulling the device). During the calibration process, at least 3syringe push/pull efforts are performed. When prompted bythe KoKo software, the syringe handle is pushed all the wayin at a low air flow (approximately 2 L), and the syringehandle is pulled out. The above step is repeated for medium(4 L/s) and high flow (8 L/s). The software notifies the oper-ator if the calibration is acceptable or asks for the procedureto be repeated. The calibration report is printed and saved.

B. Solutions1. Lyophilized adenosine monophosphate (AMP) disodium salt

is acquired from Merck Biosciences AG (Läufelfingen,Switzerland). It is provided in sterile vials that contain 3.8 g ofactive adenosine-50-monophosphate disodiumsalt. The aden-osine product is obtainedunder FoodandDrugAdministrationinvestigator-sponsored Investigational New Drug No. 70,241.

2. AMP solutions for bronchoprovocation are prepared by theInvestigational Drug Unit of Shands Hospital PharmacyDepartment under sterile conditions by injecting 8 mL ofsterile saline into a vial containing 3.8 g of AMP, producing astock solution of 380mg/mL. Further dilutions are thenmadeto produce the following concentrations: 200, 100, 50, 25,12.8, 6.4, 3.2, 1.6, 0.8, and 0.4 mg/mL.

3. Each concentration is contained in a final volume of 2 mL,dispensed in unit dose syringes, and stored at 4�C in theAsthma Research Laboratory until use or 5 months, which-ever occurs first.4

C. Baseline Spirometry1. Before the AMP challenge, baseline spirometry is performed

with the patient wearing a nose clip. The patient may beseated or standing as long as position is kept constant foreach challenge in the same patient. Spirometric efforts will berepeated (up to 5 times) until at least 3 ATS acceptable ef-forts5 have been obtained, with the goal of the 2 highestvalues not differing by more than 0.1 L.

A ATS criteria judged by the KoKo software:

� Back extrapolation less than 5% of forced vital capacityor <0.15 L (whichever is greater)

� Noearly termination of exhalation (shouldhave at least 1secondofplateauorat least 6 secondsof exhalation time)

A ATS criteria that need to be judged by the technician:� No cough during the first second of exhalation� No hesitation during the effort that results in a cessa-tion of airflow

� No glottic closure or variable effort� No leak� No mouthpiece obstruction (mechanical, tongue, orteeth)

2. If there is a progressive decrease from the first maneuver butthe last 2 (of the 5) stabilize, the higher of the last 2 is selectedas the stabilized baseline FEV1. If the last 2 differ bymore than0.1 L, additional efforts, up to a total of 8, are performed until2 consecutive FEV1 values arewithin 0.1 L. The higher of the 2is then selected as the stabilized baseline. If the goal of 2values within 0.1 L is not achieved, the higher of the last 2efforts is accepted as the stabilized baseline.

3. If the stabilized baseline FEV1 is 60% predicted or greater, thechallenge is commenced. If not, the patient is rescheduledonce. If this criterion is not met during the second visit, thepatient is discontinued from the study.

D. Challenge1. A total of 2 mL of diluent is added to the nebulizer bowl

through a 0.22-mm Millipore filter, and the top is screwed ontightly with the ball facing the mouthpiece.

2. The patient holds the nebulizer by the Digidoser handle butnot by the bowl and wears a nose clip. The mouthpiece isplaced in the patient’s mouth, and the patient performs anynumber of normal tidal breaths. For the dosing breaths, thetechnician triggers the dosimeter during exhalation into themouthpiece by pressing the spacebar. The Digidoser auto-matically discharges during the next inhalation. The patientis coached to inhale slowly and deeply to total lung capacity(TLC). At TLC, the patient holds his/her breath for 5 secondsand then exhales slowly into the mouthpiece for 5 seconds.This will be repeated 4 more times for a total of 5 breaths.These breaths do not need to be consecutive (ie, 1 or 2 non-nebulized breaths may be taken in between).

3. The FEV1 is measured at 90 and 150 seconds after the fifthexhalation. At least 1 acceptable FEV1 must be obtained. If anacceptable FEV1 is not obtained,1 or 2 additional attempts aremade (for a maximum of 3-4 attempts). The highest accept-able FEV1 is accepted as the postdiluent FEV1. It should nottake more than 2 minutes to perform these maneuvers. Ifafter a total of 4 attempts an acceptable value is not obtained,the challenge is discontinued.

4. AMP challenges are performed only if the postdiluent FEV1decreases by less than10%. If it decreases by 10%ormore belowthe stabilized baseline value, the test is rescheduled. If thishappens again during the rescheduled visit, the patient is notchallenged with adenosine and discontinued from the study.

5. Five minutes after the beginning of the first inhalation, thefirst AMP concentration (0.4 mg/mL) is administered asdescribed above.

6. Both 90 and 150 seconds after the end of the fifth exhalation,spirometry is performed. One acceptable FEV1 measurementmust be obtained. If an acceptable FEV1 measurement is notobtained, 1 or 2 additional attempts are made (for amaximum of 3-4 attempts).

7. If the FEV1 has not decreased 20% compared with the diluentFEV1 and the technician feels thinks it is safe to proceed to thenextdose level, thenextconcentrationofAMP isadministered5minutes after the beginning of the first inhalation of the pre-viousdose.ThehigherFEV1of the2willbe recorded foranalysis.

8. Doubling concentrations of AMP are administered at 5-minuteintervalsuntil a20%decrease inFEV1occurscomparedwith thediluent FEV1 or the maximum concentration is administered.

eReferences

[1] Crapo RO, Casaburi R, Coates AL, et al. Guidelines for methacholine and exercisechallenge testinge1999. Am J Respir Crit Care Med. 2000;161:309e329.

[2] Taylor DA, Jensen MW, Kanabar V, et al. A dose-dependent effect of the novelinhaled corticosteroid ciclesonide on airway responsiveness to adenosine-5’-monophosphate in asthmatic patients. Am J Respir Crit Care Med. 1999;160:237e243.

[3] Kahn YR, McDonough P, Cockcroft DW, David BE, Hendeles L. Nebulizer outputfor methacholine challenges with the KoKo Digidoser [letter]. J Allergy Clin

L. Hendeles et al. / Ann Allergy Asthma Immunol 114 (2015) 58e6262.e2

During screening themaximumisusually100mg/mL,whereasduring studies with drugs expected to increase PC20 themaximum is 200 mg/mL.

E. Safety Precautions1. The challenge is not performed if a screening 12-lead elec-

trocardiogram demonstrates a PR interval greater than 200milliseconds or a QRS interval greater than 120 milliseconds.

2. Once the AMP challenge is completed, 4 puffs of albuterolfrom a metered-dose inhaler attached to a valved-holdingchamber are administered. This dose is repeated at 15-min-ute intervals, if necessary. The patient is dischargedwhen his/her FEV1 has returned to 90% or greater of that day’s initialbaseline value and the patient is judged to be clinically stable.

3. During the adenosine challenge, all standard precautions forpatient safety are followed. A physician is in the building andimmediately available by beeper. Medication to treat severebronchospasm and oxygen is immediately available. Patientsare not left unattended during the procedure once theadministration of adenosine begins.

F. Data Analysis1. The following formula is used to calculate the percent

decrease:

Immunol. 2005;116:924e926.

[4] Martinez-Garcia MA, Perpiñá-Tordera M, Vila V, Compte-Torrero L, De Diego-Damiá A, Macián-Gisbert V. Analysis of the stability of stored adenosine 5’-monophosphate used for bronchoprovocation. Pulm Pharmacol Ther. 2002;15:157e160.

[5] American Thoracic Society. Standardization of spirometry:1994 update. Am JRespir Crit Care Med. 1995;152:1107e1136.

% decrease in FEV1

¼ ðbest diluent FEV1Þ � ðbest FEV1 after adenosineÞðbest diluent FEV1Þ

� 100%

2. The KoKo Software uses the following equation to calculatethe PC20:

PC20 ðmg=mLÞ ¼ antilog

"logC1þðlogC2� logC1Þ ð20�R1Þ

ðR2�R1Þ

#

C1 ¼ next to the last concentration (<20% FEV1 decrease)C2 ¼ last concentration (>20% FEV1 decrease)R1 ¼ % fall FEV1 after C1, and R2 ¼ % decrease FEV1 after C2

G. Sanitizing Procedure1. After each challenge the nebulizer is disassembled, washed in

dish soap water, and rinsed.2. The nebulizer is then placed in isopropyl alcohol for 10 mi-

nutes, rinsed, and allowed to air dry. Chlorox and Cidex mustnot be used because these solutions can affect the weldedstraw and baffle of the nebulizer and thus alter performance.

3. The nebulizer bowl and body are inscribed with an identifi-cation number. Care is taken to make sure that the matchingparts are assembled with the ball facing the mouthpiece.