Advertisement

A randomized multicenter study evaluating Xolair persistence of response after long-term therapy

Open AccessPublished:November 05, 2016DOI:https://doi.org/10.1016/j.jaci.2016.08.054

      Background

      Few data are available to assist clinicians with decisions regarding long-term use of asthma therapies, including omalizumab.

      Objective

      We sought to evaluate the benefit and persistence of response in subjects continuing or withdrawing from long-term omalizumab treatment.

      Methods

      Evaluating the Xolair Persistency Of Response After Long-Term Therapy (XPORT) was a randomized, double-blind, placebo-controlled withdrawal study that included subjects with moderate-to-severe persistent asthma receiving long-term omalizumab. Subjects were randomized by using a hierarchical dynamic randomization scheme to continue their same dose of omalizumab or withdraw to placebo and were then followed every 4 weeks for 1 year. The primary outcome was any protocol-defined severe asthma exacerbation. The secondary outcome was time to first protocol-defined severe asthma exacerbation. Exploratory outcomes included changes in Asthma Control Questionnaire and Asthma Control Test scores.

      Results

      Significantly more subjects in the omalizumab group (67%) had no protocol-defined exacerbation than in the placebo group (47.7%); an absolute difference of 19.3% (95% CI, 5.0%, 33.6%) represents a 40.1% relative difference. Time to first protocol-defined exacerbation analysis revealed a significantly different between-group exacerbation pattern that was consistent with the primary analysis. Subjects continuing omalizumab had significantly better asthma control (mean [SD] change from baseline to week 52: Asthma Control Test score, −1.16 [4.14] vs placebo, −2.88 [5.38], P = .0188; Asthma Control Questionnaire score, 0.22 [0.66] vs placebo, 0.63 [1.13], P = .0039). Discontinuation of omalizumab was associated with an increase in free IgE levels and an increase in basophil expression of the high-affinity IgE receptor. No safety concerns were noted.

      Conclusion

      Continuation of omalizumab after long-term treatment results in continued benefit, as evidenced by improved symptom control and reduced exacerbation risk.

      Key words

      Abbreviations used:

      ACQ (Asthma Control Questionnaire), ACT (Asthma Control Test), AE (Adverse event), EXCELS (Epidemiologic Study of Xolair (omalizumab): Evaluating Clinical Effectiveness and Long-term Safety in Patients with Moderate-to-Severe Asthma), Feno (Fractional exhaled nitric oxide), ICS (Inhaled corticosteroid), MESF (Molecules of equivalent, soluble fluorochrome units), OR (Odds ratio), SAE (Serious adverse event), XPORT (Evaluating the Xolair Persistency Of Response After Long-Term Therapy)
      Omalizumab (Xolair; Genentech, Inc, South San Francisco, Calif) is a recombinant, DNA-derived, humanized IgG mAb that selectively binds to human IgE and is approved for management of moderate-to-severe allergic asthma.
      Guidelines recommend consideration of the addition of omalizumab for patients with allergic asthma whose symptoms are not adequately controlled with high-dose inhaled corticosteroids (ICSs) plus a long-acting β-agonist.

      Global Initiative for Asthma. Global strategy for asthma management and prevention. Updated 2015. Available at: www.ginasthma.org. Accessed August 7, 2015.

      US Department of Health and Human Services, National Heart Lung and Blood Institute. Expert panel report 3: guidelines for the diagnosis and management of asthma. 2007. Available at: https://www.nhlbi.nih.gov/files/docs/guidelines/asthgdln.pdf. Accessed April 16, 2014.

      In vivo omalizumab selectively binds to unbound (free) human IgE at the same Fc site as the high-affinity IgE receptor (FcεRI), thereby blocking the interaction of IgE with FcεRI.
      • MacGlashan Jr., D.W.
      • Bochner B.S.
      • Adelman D.C.
      • Jardieu P.M.
      • Togias A.
      • McKenzie-White J.
      • et al.
      Down-regulation of Fc(epsilon)RI expression on human basophils during in vivo treatment of atopic patients with anti-IgE antibody.
      • Chanez P.
      • Contin-Bordes C.
      • Garcia G.
      • Verkindre C.
      • Didier A.
      • De Blay F.
      • et al.
      Omalizumab-induced decrease of FcεRI expression in patients with severe allergic asthma.
      The expression of FcεRI on mast cells and basophils is also reduced, resulting in a reduction of mast cell and basophil mediator release.
      • Pelaia G.
      • Gallelli L.
      • Renda T.
      • Romeo P.
      • Busceti M.T.
      • Grembiale R.D.
      • et al.
      Update on optimal use of omalizumab in management of asthma.
      Treatment with omalizumab reduces the risk of exacerbations in allergic asthma in clinical studies up to 1 year,
      • Bousquet J.
      • Cabrera P.
      • Berkman N.
      • Buhl R.
      • Holgate S.
      • Wenzel S.
      • et al.
      The effect of treatment with omalizumab, an anti-IgE antibody, on asthma exacerbations and emergency medical visits in patients with severe persistent asthma.
      • Busse W.
      • Corren J.
      • Lanier B.Q.
      • McAlary M.
      • Fowler-Taylor A.
      • Cioppa G.D.
      • et al.
      Omalizumab, anti-IgE recombinant humanized monoclonal antibody, for the treatment of severe allergic asthma.
      • Humbert M.
      • Beasley R.
      • Ayres J.
      • Slavin R.
      • Hébert J.
      • Bousquet J.
      • et al.
      Benefits of omalizumab as add-on therapy in patients with severe persistent asthma who are inadequately controlled despite best available therapy (GINA 2002 step 4 treatment): INNOVATE.
      • Solèr M.
      • Matz J.
      • Townley R.
      • Buhl R.
      • O'Brien J.
      • Fox H.
      • et al.
      The anti-IgE antibody omalizumab reduces exacerbations and steroid requirement in allergic asthmatics.
      but there are limited data regarding the long-term efficacy or persistence of a beneficial effect after its discontinuation.
      • MacGlashan Jr., D.W.
      • Bochner B.S.
      • Adelman D.C.
      • Jardieu P.M.
      • Togias A.
      • McKenzie-White J.
      • et al.
      Down-regulation of Fc(epsilon)RI expression on human basophils during in vivo treatment of atopic patients with anti-IgE antibody.
      • Chanez P.
      • Contin-Bordes C.
      • Garcia G.
      • Verkindre C.
      • Didier A.
      • De Blay F.
      • et al.
      Omalizumab-induced decrease of FcεRI expression in patients with severe allergic asthma.
      • Nopp A.
      • Johansson S.G.
      • Adédoyin J.
      • Ankerst J.
      • Palmqvist M.
      • Oman H.
      After 6 years with Xolair; a 3-year withdrawal follow-up.
      • Nopp A.
      • Johansson S.G.
      • Ankerst J.
      • Palmqvist M.
      • Oman H.
      CD-sens and clinical changes during withdrawal of Xolair after 6 years of treatment.
      • Saini S.S.
      • MacGlashan Jr., D.W.
      • Sterbinsky S.A.
      • Togias A.
      • Adelman D.C.
      • Lichtenstein L.M.
      • et al.
      Down-regulation of human basophil IgE and FC epsilon RI alpha surface densities and mediator release by anti-IgE-infusions is reversible in vitro and in vivo.
      The Evaluating the Xolair Persistency Of Response After Long-Term Therapy (XPORT; ClinicalTrials.gov identifier: NCT01125748) study was designed to determine whether subjects receiving long-term treatment with omalizumab demonstrate persistence of treatment effect after omalizumab discontinuation and to examine whether the benefits of omalizumab are sustained when therapy is continued.

      Methods

       Study design

      XPORT was a 52-week, randomized, double-blind, placebo-controlled multicenter study conducted in the United States from May 2010 to August 2013. The primary objective was to evaluate the persistence of response to omalizumab in patients with moderate-to-severe persistent allergic asthma who discontinued omalizumab therapy after long-term use. The secondary objective was to evaluate the safety of omalizumab discontinuation or continuation after long-term use. All subjects provided written informed consent for study participation. The study was conducted according to the International Conference on Harmonisation E6 Guidelines for Good Clinical Practice, US Food and Drug Administration regulations, and any applicable national laws. Visits conducted at the investigator's clinical site were scheduled at 4-week (±4 days) intervals over the 1-year study, beginning with the screening (baseline) visit at week 0. At screening, demographic information, concomitant medications, and medical/surgical/smoking history were documented. A physical examination, chest radiography, and spirometry were performed, and other clinical laboratory assessments were collected. Some centers performed fractional exhaled nitric oxide (Feno) measurements at baseline. In addition, each subject's personal best peak expiratory flow was documented. A checklist of prior asthma medications was completed, and the number of protocol-defined severe exacerbations per year in the previous 5 years was collected. Additional measures assessed at baseline included the Asthma Control Test (ACT) and the Asthma Control Questionnaire (ACQ).
      The majority of subjects in this study were recruited from a previously completed long-term, observational, epidemiologic study (Epidemiologic Study of Xolair [omalizumab]: Evaluating Clinical Effectiveness and Long-term Safety in Patients with Moderate-to-Severe Asthma [EXCELS]; ClinicalTrials.gov identifier: NCT00252135). EXCELS included subjects 12 years of age or older with moderate-to-severe asthma with a history of a positive skin test response or in vitro reactivity to a perennial aeroallergen.
      • Long A.A.
      • Fish J.E.
      • Rahmaoui A.
      • Miller M.K.
      • Bradley M.S.
      • Taki H.N.
      • et al.
      Baseline characteristics of patients enrolled in EXCELS: a cohort study.
      Subjects from EXCELS must have completed EXCELS and not discontinued omalizumab since completion of EXCELS. A protocol amendment based on investigator request added the following: subjects could be included in XPORT even if they had not participated in EXCELS but had received long-term (approximately 5 years) treatment with omalizumab before randomization, as substantiated by injection records. Subjects were randomized 1:1 to continue their pre-enrollment dose of omalizumab (omalizumab group) or were withdrawn from omalizumab and switched to placebo treatment (placebo group). Randomization was stratified for ICS use at baseline and level of asthma control at baseline (well controlled: ACT score ≥20 vs not well controlled: ACT score <20), for the omalizumab dosing interval that the subject received at baseline (2 vs 4 weeks), and for study site. A hierarchical dynamic randomization scheme was used to achieve approximate overall balance between treatment groups and within each stratum.
      At randomization, subjects entering the study from outside of EXCELS were stratified separately from the subjects who had participated in EXCELS to ensure that both subgroups were adequately balanced. An interactive voice response system was used for randomization.
      At each subsequent study visit, a physical examination and spirometry were performed, vital signs were measured, concomitant medications and unscheduled asthma-related medical visits were documented, and the ACT and ACQ were administered. Blood samples collected at baseline and weeks 12, 24, 36, and 52 were analyzed for total, free, and specific IgE levels; basophil FcεRI expression was measured by means of flow cytometry; and other clinical laboratory assessments were performed. Feno data were collected at weeks 12, 24, 36, and 52 in some centers.
      The primary efficacy end point was any protocol-defined severe asthma exacerbation, which was defined as a clinically significant worsening of asthma during the study period (52 weeks) that, in the clinical judgment of the investigator, required at least 1 of the following: (1) initiation of systemic corticosteroid treatment (tablets, suspension, and injection); (2) any clinically significant worsening of asthma requiring 3 or more days of treatment with a 20-mg or greater increase in the average daily dose of prednisone or a comparable dose of systemic corticosteroid in subjects already receiving chronic oral corticosteroids; or (3) a hospitalization or emergency department visit because of asthma requiring administration of systemic corticosteroids. The secondary efficacy outcome measure was the time to first protocol-defined severe asthma exacerbation. Protocol-defined exploratory outcome measures included the change from baseline in ACT scores, ACQ scores, spirometric results (FEV1), Feno values (measured in selected study sites with a hand-held portable device; NIOX MINO, Aerocrine [now Circassia], Solna, Sweden), free and total serum IgE concentrations, and basophil FcεRI expression. Safety outcome measures included the frequency and severity of adverse events (AEs) and serious adverse events (SAEs), vital signs, and results of clinical laboratory assessments, including hematology, serum chemistry, and urinalysis.
      After the primary end point was met at the first exacerbation, unblinding was permitted, and subjects could have their therapy altered at the discretion of the treating physician. Subjects receiving placebo at the time of exacerbation subsequently could be treated with omalizumab. Adjustments were made to include these subjects in the omalizumab exposure safety population.

       Study subjects

      Subjects were eligible for study inclusion if they were aged 17 to 70 years inclusive, had continuous omalizumab exposure (having missed ≤25% of scheduled omalizumab doses) with omalizumab dose and dosing schedule consistent with US prescribing information,
      and received stable doses of other asthma therapies for 2 or more months before enrollment. Exclusion criteria included current participation in another clinical study; acute asthma exacerbation within 2 months of screening that required initiation of systemic corticosteroids, increase in systemic corticosteroid dose, doubling of ICS dose, or emergency department visit or hospitalization within 2 months of study entry; significant or unstable systemic disease; active lung disease other than asthma; 10 pack-year or longer smoking history; use of an experimental drug within 30 days before study screening; pregnancy, lactation, or any planned pregnancy during the study year; or increased serum IgE levels for reasons other than allergy.

       Treatments

      Study treatments (omalizumab or placebo) were administered subcutaneously after the subject's pre-enrollment dose and dosing interval unless there was a significant change in body weight since initial dosing and the omalizumab dose were adjusted, if necessary, based on US prescribing information. The minimum dose of omalizumab was 0.008 mg/kg/IgE (IU/mL) every 2 weeks or 0.016 mg/kg/IgE (IU/mL) every 4 weeks from weeks 0 through 48 (13 or 26 total doses). Doses of greater than 150 mg were divided among more than 1 injection site. Placebo treatment contained the same ingredients, excluding the omalizumab component.
      Use of specific concomitant medications was permitted: ICSs, long-acting β-agonists, leukotriene receptor antagonists, 5-lipoxygenase inhibitors, anticholinergics (oral, inhaled, and/or nasal), mast cell stabilizers, theophylline, and/or chronic oral corticosteroids (prednisone equivalent, 2-40 mg/d or 5-80 mg every other day for at least 4 weeks before screening). Allergen immunotherapy was permitted but not tracked because it was not anticipated to affect the study outcome.

       Analysis

      Sample size calculations were based on precision estimates. We estimated that a sample size of 90 subjects in each group (omalizumab and placebo) would have a precision of 0.11 (placebo) and 0.15 (treatment difference), assuming a persistence response of 0.40 for the placebo group and 0.70 for the omalizumab group and a 10% dropout rate before reaching the primary end point. The intent-to-treat population included all randomized subjects based on randomized treatment groups, regardless of actual treatment. The per-protocol population included all subjects who were not included in the major protocol violation listing based on blinded clinical review. The safety population was described in 2 ways: (1) at least 1 injection population (included all subjects who had received at least 1 injection [active drug or placebo]), and (2) omalizumab exposure population (included all subjects who received at least 1 injection of omalizumab while on study, including placebo-treated subjects who received omalizumab after a protocol-defined exacerbation and subsequent unblinding).
      The primary efficacy analysis included counts and proportions of subjects without any protocol-defined severe asthma exacerbations for each treatment group among all subjects, as well as in the subgroup of subjects receiving ICSs at baseline. Precision of the estimates was evaluated by including 95% CIs. The difference in the proportions of responders between the treatment groups was calculated in addition to a 95% CI by using normal approximation to the binomial distribution. Sensitivity analysis for the primary efficacy end point was conducted by repeating the analysis for the per-protocol population.
      The secondary efficacy analysis was performed for all subjects, as well as the subgroup of subjects receiving ICSs at baseline. For the time to first protocol-defined severe asthma exacerbation, the hazard ratio comparing the 2 treatment groups was calculated, and the 95% CI reported based on a Cox proportional hazards model with treatment was the only predictor in the model. For the exploratory outcome measures, analysis included descriptive summaries of continuous changes and the difference between study groups; P values were derived by using the Student t test. When an exploratory end point was categorical in nature, a χ2 test was used in conjunction with descriptive summaries of the counts and proportions within each study group to evaluate the statistical differences between the groups. AEs were summarized by the treatments actually administered for the controlled period of observation, as well as the uncontrolled period after a subject was unblinded.
      If a subject discontinued before the end of the 1-year treatment period and had not reached the primary end point, the subject was considered a nonresponder (imputed as “having a severe exacerbation”) for the primary end point. For the calculation of time to nonresponse, subjects were censored at the point of discontinuation, and the discontinuation date was used as the event time. For the exploratory outcome measures, if a subject discontinued before the end of the 1-year treatment period, the data after the point of discontinuation were imputed by using the last observation carried forward method. This method was chosen because the study had good subject completion rates.
      As a post hoc analysis, univariate logistic regression was used to determine whether the change in Feno value from baseline to week 12 (date of first measurement) was predictive of future exacerbations.

      Results

       Subject population

      The intent-to-treat and safety populations included 176 subjects who were randomized to continuation of treatment with omalizumab (n = 88) or to withdrawal and placebo treatment (n = 88; Fig 1). Twenty subjects (9/88 [10.2%] and 11/88 [12.5%] in the omalizumab and placebo groups, respectively) discontinued the study without first experiencing a protocol-defined severe exacerbation and required an imputation of their primary end point. The omalizumab-exposed safety population (n = 121) included subjects continuing omalizumab (n = 88) and placebo-treated subjects (n = 33) who were given omalizumab after a protocol-defined exacerbation and subsequent unblinding.
      Figure thumbnail gr1
      Fig 1Subject disposition. Physician decision withdrawals took place because of protocol violations (eg, dosing not per US prescribing information and not meeting inclusion/exclusion criteria). Patient decision withdrawals were initiated by the subject for a variety of reasons that were not consistently captured.
      Most baseline demographics and disease characteristics were similar across treatment groups (Table I). The average age of subjects was 51.5 years, and most (69.9%) subjects were female. The majority (69.2%) of subjects had never smoked. Most of the subjects were from EXCELS (approximately 85% in each group). ICSs were used by 83.5% of subjects at baseline. Diabetes mellitus was noted in more subjects in the omalizumab group than in the placebo group (20.5% vs 13.6%), as was a higher exposure to biguanides (eg, metformin; 14.9% vs 6.8%). Regular use of nonsteroidal anti-inflammatory agents was also higher in the omalizumab group than in the placebo group (33.1% vs 19.3%).
      Table IBaseline demographics and disease characteristics (ITT population)
      CharacteristicPlacebo (omalizumab discontinuation [n = 88])Omalizumab continuation (n = 88)All subjects (n = 176)
      Age (y)
       Mean (SD)51.9 (13.3)51.1 (11.7)51.5 (12.5)
       Median (range)55.0 (19.0-70.0)53.0 (17.0-70.0)54.0 (17.0-70.0)
      Female sex, no. (%)60 (68.2)63 (71.6)123 (69.9)
      Race, no. (%)
       White73 (83.0)75 (85.2)148 (84.1)
       Black12 (13.6)9 (10.2)21 (11.9)
      Weight (kg), mean (SD)88.0 (20.3)85.6 (24.4)86.8 (22.4)
      BMI (kg/m2), mean (SD)32.4 (7.7)31.0 (8.3)31.7 (8.0)
      FEV1 (% predicted), mean (SD)77.7 (19.2)75.5 (18.8)76.6 (19.0)
      ACT score, mean (SD)21.3 (3.8)21.3 (3.7)21.3 (3.7)
      ACQ score, mean (SD)1.0 (0.8)1.0 (0.8)1.0 (0.8)
      Receiving ICSs at baseline, no. (%)74 (84.1)73 (83.0)147 (83.5)
      Smoking status,
      Placebo-treated (n = 87), omalizumab-treated (n = 85), and all (n = 172) subjects.
      no. (%)
       Current1 (1.1)4 (4.7)5 (2.9)
       Previous32 (36.8)16 (18.8)48 (27.9)
       Never54 (62.1)65 (76.5)119 (69.2)
      Dose interval
       Q2W29 (33.0)29 (33.0)58 (33.0)
       Q4W59 (67.0)59 (67.0)118 (67.0)
      Duration of asthma (y),
      Placebo-treated (n = 81), omalizumab-treated (n = 82), and all (n = 163) subjects.
      mean (SD)
      27.7 (16.1)27.3 (17.1)27.5 (16.6)
      Eosinophils (× 103/μL),
      Placebo-treated (n = 86) and omalizumab-treated (n = 86) subjects.
      mean (SD)
      0.25 (0.22)0.24 (0.17)NA
      Basophils (× 103/μL),
      Placebo-treated (n = 86) and omalizumab-treated (n = 86) subjects.
      mean (SD)
      0.06 (0.03)0.07 (0.04)NA
      Diabetes mellitus, no. (%)12 (13.6)18 (20.5)30 (17.0)
      BMI, Body mass index; ITT, intent to treat; NA, not available; Q2W, every 2 weeks; Q4W, every 4 weeks.
      Placebo-treated (n = 87), omalizumab-treated (n = 85), and all (n = 172) subjects.
      Placebo-treated (n = 81), omalizumab-treated (n = 82), and all (n = 163) subjects.
      Placebo-treated (n = 86) and omalizumab-treated (n = 86) subjects.
      Most (150/176 [85%]) subjects entered XPORT after participation in EXCELS. Baseline demographics and characteristics by EXCELS participation are provided in Table E1 in this article's Online Repository at www.jacionline.org.

       Primary efficacy end point

      A significantly greater number of subjects had no protocol-defined exacerbations in the omalizumab group (n = 59 [67.0%]) versus the placebo group (n = 42 [47.7%]). This absolute difference of 19.3% (95% CI, 5.0% to 33.6%) represents a 40.1% relative difference. The unadjusted odds ratio (OR) and 95% CI for omalizumab was 0.45 (0.24-0.83). The OR adjusted for age, sex, ICS use, EXCELS participation, number of severe exacerbations in year before study entry, and percent predicted FEV1 was similar (0.44; 95% CI, 0.23-0.82). Similar results were observed in the subgroup of subjects receiving ICSs at baseline: a greater number of subjects had no protocol-defined exacerbations in the omalizumab group (n = 51, 69.9% [95% CI, 59.3% to 80.4%]) vs the placebo group (n = 35, 47.3% [95% CI, 35.9% to 58.7%]). The overall and adjusted OR for subjects receiving ICSs at baseline was 0.39 (95% CI, 0.200-0.76). Per-protocol analysis supported these results (see Table E2 in this article's Online Repository at www.jacionline.org).

       Secondary and exploratory outcome measures

      Analyses with Cox proportional hazard modeling demonstrated that time to first protocol-defined exacerbation showed a significantly different between-group exacerbation pattern that was consistent with the primary analysis (Fig 2): it was longer in the omalizumab group compared with the placebo group (hazard ratio, 0.49; 95% CI, 0.28-0.86). These results were similar for the subgroup of subjects receiving ICSs at baseline. Symptom control was also significantly greater in the omalizumab group versus the placebo group, as evidenced by the mean (SD) change from baseline to week 52 in ACT and ACQ scores (Fig 3).
      Figure thumbnail gr2
      Fig 2Time to first protocol-defined asthma exacerbation.
      Figure thumbnail gr3
      Fig 3Mean (SD) change in ACT and ACQ scores from baseline to week 52.
      Total and free IgE levels remained stable from baseline to week 52 in the omalizumab group, whereas in the placebo group, total IgE levels decreased and free IgE levels increased (Table II). The mean change from baseline to week 52 in basophil expression of FcεRI was significantly greater in subjects in the placebo group compared with the omalizumab group (2136.36 [SD, 1087.65] molecules of equivalent, soluble fluorochrome units [MESF] vs 64.72 [SD, 348.65] MESF, P < .0001). There was no statistically significant difference between treatment groups in the change in FEV1 from baseline to week 52.
      Table IIChange in IgE concentrations from baseline to week 52
      Mean percentage (SD)PlaceboOmalizumabDifference (omalizumab – placebo)P value
      P values were derived from the Wilcoxon rank-sum test.
      Free IgE
      No direct comparison was possible (different units of measure).
      (ng/mL)
      n = 68

      402.4 (464.1)
      n = 87

      10.9 (62.6)
      −391.4 (310.7)<.0001
       IQR of mean percentage0.0 to 689.9−22.9 to 25.7
       Median of mean percentage (range)298.9 (−79.8 to 2269.7)0.0 (−82.3 to 275.3)
      Total IgE
      No direct comparison was possible (different units of measure).
      (IU/mL)
      n = 88

      −42.8 (37.0)
      n = 87

      2.0 (42.9)
      44.9 (40.0)<.0001
       IQR of mean percentage−73.4 to 0.6−21.9 to 11.6
       Median (range) of mean percentage−57.9 (−89.6 to 103.0)−4.1 (−91.2 to 184.5)
      IQR, Interquartile range.
      P values were derived from the Wilcoxon rank-sum test.
      No direct comparison was possible (different units of measure).
      At baseline, there was no statistically significant difference in peripheral eosinophil counts between the placebo and omalizumab arms (P = .70), and the lack of a significant between-group difference was sustained at week 52 (P = .76). However, after omalizumab withdrawal, the difference in peripheral eosinophil count was statistically significantly higher in the placebo group with exacerbations compared with that in the group without exacerbations (P < .001). In contrast, pairwise comparisons in peripheral eosinophil counts between the omalizumab group without exacerbations versus the group with exacerbations was not statistically significant (Fig 4).
      Figure thumbnail gr4
      Fig 4Eosinophils over time by treatment cohort and exacerbation status. *At baseline, the difference in eosinophil counts was statistically significant for the placebo groups with and without exacerbations (P < .001).
      Seventy-eight subjects had Feno measurements (43 in the omalizumab cohort and 35 in the placebo cohort). There were no statistically significant differences between treatment groups at weeks 12, 24, 36, and 52, as measured by change from baseline in Feno values. However, a univariate logistic regression analysis revealed that an increase in Feno values from baseline to week 12 after omalizumab discontinuation was a positive predictor of exacerbations (P = .038). Similarly, when this regression analysis was repeated for the subset of subjects receiving an ICS at baseline, an increase in Feno values from baseline to week 12 after omalizumab discontinuation was again seen as a positive predictor of exacerbations (P = .047). For subjects in the placebo group who experienced an exacerbation, there was a sharp increase in Feno value from baseline to week 12 (Fig 5). Subjects in the omalizumab group who experienced an exacerbation also had an increase in Feno values from baseline to week 12, but it was less pronounced. Subjects in either treatment group who did not experience an exacerbation had no increase in Feno values from baseline to week 12. Rather, Feno values remained generally flat (placebo) or decreased (omalizumab).
      Figure thumbnail gr5
      Fig 5Feno (in parts per billion) values over time by treatment cohort and exacerbation status.

       Safety

      During the study, there were 33 subjects who transitioned from placebo to omalizumab; therefore AE rates were calculated per exposure time. Overall, the rate of AEs per 100 patient-years was similar in both treatment groups (placebo, 425.9; omalizumab, 413.2; Table III). The most common AEs reported (per 100 patient-years) were asthma (placebo, 71.8; omalizumab, 66.6), sinusitis (placebo, 30.8; omalizumab, 35.2), upper respiratory tract infection (placebo, 22.2; omalizumab, 27.4), and acute sinusitis (placebo, 25.7; omalizumab, 14.7). SAE rates per 100 patient-years were similar between the 2 treatment groups (placebo, 9.1; omalizumab, 8.3). The most common SAE reported (per 100 patient-years) was asthma (placebo, 6.8; omalizumab, 5.9). Two malignancy SAEs were reported in the study: adenocarcinoma of the colon in a subject in the omalizumab group and a mixed Müllerian tumor in a subject in the placebo group. One death was reported (placebo group) 7 months after the subject was withdrawn from the study for an SAE of a mixed Müllerian tumor that was diagnosed on day 129. No subjects in the omalizumab group experienced an AE leading to study withdrawal. Two subjects in the placebo group experienced an AE leading to study withdrawal (mixed Müllerian tumor and angioedema).
      Table IIITEAEs (event rate per 100 patient-years)
      TEAE (event rate per 100 patient-years)PlaceboAny omalizumab exposure
      Includes placebo-treated subjects who were treated with omalizumab after exacerbation.
      All subjects
      Total no. of TEAEs249422671
      Exposure time (patient-years)58.46102.14160.60
      Rate per 100 patient-years425.9413.2417.8
      Ear disorders
       Eustachian tube dysfunction0.02.91.9
      Eye disorders
       Allergic conjunctivitis0.02.91.9
      Gastrointestinal disorders
       Gastroesophageal reflux disease5.13.94.4
      General disorders
       Peripheral edema3.41.01.9
      Infections and infestations
       Oral candidiasis8.62.04.4
       Tooth infection1.72.92.5
       Otitis media3.42.93.1
       Bronchitis15.49.811.8
       Acute sinusitis25.714.718.7
       Nasopharyngitis10.33.96.2
       Pharyngitis3.42.02.5
       Sinusitis30.835.233.6
       Upper respiratory tract infection22.227.425.5
       Urinary tract infection12.012.712.5
       Viral gastroenteritis1.76.95.0
       Viral upper respiratory tract infection5.15.95.6
      Skin injuries
       Arthropod bite0.02.91.9
       Procedural pain3.42.93.1
       Contusion1.72.92.5
       Laceration0.02.91.9
      Metabolism and nutrition disorders
       Diabetes mellitus3.42.93.1
       Hypercholesterolemia0.02.91.9
       Vitamin D deficiency5.10.01.9
      Musculoskeletal and connective tissue disorders
       Arthralgia3.46.95.6
       Back pain5.10.01.9
       Pain in extremity3.41.01.9
       Osteoarthritis1.73.93.1
       Tendonitis3.40.01.2
      Nervous system disorders
       Headache5.16.96.2
       Anxiety1.72.92.5
       Depression5.11.02.5
       Insomnia6.81.03.1
      Reproductive disorders
       Postmenopausal hemorrhage3.40.01.2
      Respiratory, thoracic, and mediastinal disorders
       Asthma71.866.668.5
       Cough5.12.03.1
       Rhinitis allergic13.71.05.6
       Nasal polyps1.72.92.5
       Dysphonia1.72.92.5
       Oropharyngeal pain0.02.91.9
      Skin and subcutaneous tissue disorders
       Angioedema3.40.01.2
       Dermatitis contact1.75.94.4
       Eczema3.41.01.9
      Vascular disorders
       Hypertension5.12.93.7
      TEAE, Treatment-emergent adverse event.
      Includes placebo-treated subjects who were treated with omalizumab after exacerbation.

      Discussion

      This is the first randomized controlled trial to rigorously compare the risks and benefits of continuing versus discontinuing omalizumab after long-term use in patients with allergic asthma. The results demonstrate that compared with subjects withdrawn to placebo, continuation of omalizumab resulted in sustained benefit in both domains of asthma control: risk (prevention of exacerbations) and impairment (symptom control). Less clear, however, is the extent to which a persistence of these benefits exists after omalizumab is discontinued, which was the primary rationale for conducting the study. Had the study shown equal and extremely high or extremely low percentages of subjects in both arms without exacerbations, it would be easier to infer that omalizumab had a persistent effect or had lost its efficacy, respectively. Neither of these extremes was seen in this study. Instead, the results are different for the 2 groups, with 47.7% of the placebo group experiencing no exacerbations and 67.0% of the continuation group experiencing no exacerbations. The relatively large percentage of subjects in the placebo arm with no exacerbations (47.7%) suggests a persistence of benefit in some subjects because these subjects presumably, but not necessarily, experienced frequent exacerbations before starting omalizumab.
      To definitively answer the question regarding persistence, one would need to know the exacerbation rates and ACT/ACQ scores for the population before omalizumab initiation; then one could compare the exacerbation rates and ACT/ACQ scores after omalizumab withdrawal with the rates and scores before omalizumab initiation to see whether there was a difference. Unfortunately, this information is not available for the current study, and this lack of pretreatment data limits any conclusions. The study is further limited by the fact that even if exacerbation rates before omalizumab initiation were known, exacerbation rates could not be calculated for the placebo group because subjects who experienced at least 1 exacerbation were allowed to be unblinded and switched to known study drug.
      The findings regarding symptom control also suggest the persistence of benefit from omalizumab in some subjects. ACT/ACQ scores at baseline indicate that subjects' symptoms were reasonably well controlled on long-term omalizumab therapy, but presumably, this was not the case before omalizumab initiation. The fact that ACT/ACQ scores had a statistically significant but relatively small degree of deterioration in the placebo group might also be an indication of persistence of benefit. Furthermore, because the data reflect mean group changes, it is likely that subjects with a persistence benefit experienced minimal or no deterioration, whereas others had a more significant worsening.
      It is possible that subjects experienced an improvement in their asthma independent of omalizumab treatment. However, a persistence effect from omalizumab would be consistent with the study from Nopp et al
      • Nopp A.
      • Johansson S.G.
      • Ankerst J.
      • Palmqvist M.
      • Oman H.
      CD-sens and clinical changes during withdrawal of Xolair after 6 years of treatment.
      in which most of 18 adult atopic subjects treated with omalizumab for 6 years remained clinically stable up to 14 months after omalizumab discontinuation.
      Several immunologic changes were observed in the 52 weeks after omalizumab discontinuation, including an increase in free IgE levels and an increase in basophil expression of FcεRI. These immunologic marker changes correlated with the observed worsening of symptom control and increased risk for exacerbation after discontinuation of omalizumab. These findings are consistent with previous studies and reflective of the current understanding of omalizumab's mechanism of action.
      • MacGlashan Jr., D.W.
      • Bochner B.S.
      • Adelman D.C.
      • Jardieu P.M.
      • Togias A.
      • McKenzie-White J.
      • et al.
      Down-regulation of Fc(epsilon)RI expression on human basophils during in vivo treatment of atopic patients with anti-IgE antibody.
      • Chanez P.
      • Contin-Bordes C.
      • Garcia G.
      • Verkindre C.
      • Didier A.
      • De Blay F.
      • et al.
      Omalizumab-induced decrease of FcεRI expression in patients with severe allergic asthma.
      • Pelaia G.
      • Gallelli L.
      • Renda T.
      • Romeo P.
      • Busceti M.T.
      • Grembiale R.D.
      • et al.
      Update on optimal use of omalizumab in management of asthma.
      • Saini S.S.
      • MacGlashan Jr., D.W.
      • Sterbinsky S.A.
      • Togias A.
      • Adelman D.C.
      • Lichtenstein L.M.
      • et al.
      Down-regulation of human basophil IgE and FC epsilon RI alpha surface densities and mediator release by anti-IgE-infusions is reversible in vitro and in vivo.
      Although data from the predefined outcome measures do not clearly identify subpopulations of subjects receiving omalizumab who can be withdrawn from treatment without loss of asthma control, the data regarding eosinophil counts and Feno measurements do provide some insights for clinicians. At baseline, there was no difference in peripheral blood eosinophil counts between the omalizumab and placebo groups. However, when the 2 treatment arms were divided into subgroups based on whether they experienced an exacerbation, a higher baseline peripheral blood eosinophil count was observed in subjects in the placebo group who experienced an exacerbation. This suggests that subjects with persistently higher blood eosinophil counts during omalizumab treatment are more likely to have an exacerbation when omalizumab is discontinued. In contrast, this difference was not seen in subjects who had exacerbations while maintaining omalizumab, suggesting a different mechanism for these exacerbations.
      The relationship between Feno values and exacerbation risk identified in this study is more complex. In the group of subjects in which omalizumab was withdrawn, an increase in Feno values at week 12 is predictive of increased risk of exacerbation. However, in the group that continued omalizumab, Feno values remained relatively stable, irrespective of asthma exacerbation. This suggests that asthma exacerbations in subjects receiving long-term omalizumab might have triggers not reflected in Feno changes. This is consistent with the peripheral blood eosinophil observation previously discussed. Furthermore, for subjects receiving omalizumab who persistently have a relatively high Feno value, it might be even more important to maintain omalizumab therapy because continuation of therapy sustains benefit in terms of exacerbation risk reduction. Although a relationship between changes in Feno values and exacerbations can be evident earlier than 12 weeks, this analysis was limited to looking at the first Feno measurement 12 weeks after discontinuation of long-term omalizumab therapy.
      Evidence from other studies suggests that peripheral blood eosinophil counts
      • Malinovschi A.
      • Fonseca J.A.
      • Jacinto T.
      • Alving K.
      • Janson C.
      Exhaled nitric oxide levels and blood eosinophil counts independently associate with wheeze and asthma events in National Health and Nutrition Examination Survey subjects.
      • Schleich F.N.
      • Louis R.
      Importance of concomitant local and systemic eosinophilia in uncontrolled asthma.
      • Zeiger R.S.
      • Schatz M.
      • Li Q.
      • Chen W.
      • Khatry D.B.
      • Gossage D.
      • et al.
      High blood eosinophil count is a risk factor for future asthma exacerbations in adult persistent asthma.
      • Zeiger R.S.
      • Schatz M.
      • Li Q.
      • Chen W.
      • Khatry D.B.
      • Gossage D.
      • et al.
      The association of blood eosinophil counts to future asthma exacerbations in children with persistent asthma.
      and Feno values
      • Malinovschi A.
      • Fonseca J.A.
      • Jacinto T.
      • Alving K.
      • Janson C.
      Exhaled nitric oxide levels and blood eosinophil counts independently associate with wheeze and asthma events in National Health and Nutrition Examination Survey subjects.
      are predictive of more severe asthma. In this study the findings regarding Feno values and eosinophil counts are consistent with the post hoc analysis of the Study of Omalizumab (Xolair) in Subjects with Moderate to Severe Persistent Asthma (EXTRA),
      • Hanania N.A.
      • Wenzel S.
      • Rosén K.
      • Hsieh H.J.
      • Mosesova S.
      • Choy D.F.
      • et al.
      Exploring the effects of omalizumab in allergic asthma: an analysis of biomarkers in the EXTRA study.
      which showed that omalizumab treatment was associated with greater effects in 3 high-biomarker subgroups (exhaled Feno values, blood eosinophil counts, and blood periostin levels).
      Thus, a consistent picture is emerging. Withdrawal from omalizumab after long-term therapy results in reversal of the effects on IgE and basophils, which correlate with deterioration in asthma control. Additionally, although it is difficult to predict which patients can be safely withdrawn from omalizumab therapy, Feno values and peripheral blood eosinophil counts might be useful for guidance in making treatment decisions regarding omalizumab.
      Clinical implications
      Patients continue to benefit from ongoing use of omalizumab after long-term treatment. Feno values and blood eosinophil counts might be useful biomarkers to guide treatment decisions.
      We would like to acknowledge the contributions of Jan Canvin, MD, formerly of Novartis Pharmaceuticals Corporation, to the design of the study and Ben Ortiz, MD, formerly of Novartis Pharmaceuticals Corporation, for assistance with data review and interpretation. We thank the clinical trial investigators, staff, and subjects who participated in the study.

      Appendix

      Table E1Baseline demographics and disease characteristics by participation in EXCELS
      EXCELS subjectsNon-EXCELS subjects
      Placebo (omalizumab discontinuation [n = 75])Omalizumab continuation (n = 75)Placebo (omalizumab discontinuation [n = 13])Omalizumab continuation (n = 13)
      Age (y)
       Mean (SD)53.0 (12.0)50.8 (12.3)45.3 (18.1)53.4 (8.0)
       Median (range)56.0 (21.0-70.0)53.0 (17.0-70.0)49.0 (19.0-70.0)54.0 (35.0-66.0)
      Female sex, no. (%)55 (73.3)55 (73.3)5 (38.5)8 (61.5)
      Race, no. (%)
       White61 (81.3)62 (82.7)12 (92.3)13 (100.0)
       Black11 (14.7)9 (12.0)1 (7.7)0 (0.0)
       Other3 (4.0)4 (5.3)0 (0.0)0 (0.0)
      Weight (kg), mean (SD)88.7 (21.1)85.0 (23.2)84.4 (14.2)89.0 (31.2)
      BMI (kg/m2), mean (SD)33.0 (7.9)31.0 (8.5)29.0 (5.0)30.9 (7.2)
      Receiving ICSs at baseline, no. (%)63 (84.0)64 (85.3)11 (84.6)9 (69.2)
      Smoking status,
      EXCELS placebo-treated (n = 74) and omalizumab-treated (n = 72) subjects.
      no. (%)
       Current1 (1.4)3 (4.2)0 (0.0)1 (7.7)
       Previous28 (37.8)15 (20.8)4 (30.8)1 (7.7)
       Never45 (60.8)54 (75.0)9 (69.2)11 (84.6)
      Dose interval
       Q2W24 (32.0)24 (32.0)5 (38.5)5 (38.5)
       Q4W51 (68.0)51 (68.0)8 (61.5)8 (61.5)
      Q2W, Every 2 weeks; Q4W, every 4 weeks.
      EXCELS placebo-treated (n = 74) and omalizumab-treated (n = 72) subjects.
      Table E2Subjects without a severe exacerbation (per-protocol population)
      PlaceboOmalizumabDifference (omalizumab-placebo)
      All subjectsn = 71n = 70
       No. (%)38 (53.5)51 (72.9)19.4%
       95% CI41.9% to 65.1%62.4% to 83.3%3.7% to 34.9%
      Subjects receiving an ICS at baselinen = 58n = 59
       No. (%)31 (53.4)51 (74.6)21.2%
       95% CI40.6% to 66.3%63.5% to 85.7%4.2% to 38.1%

      References

      1. Xolair (omalizumab) prescribing information. Genentech, Inc: Novartis Pharmaceuticals Corporation, South San Francisco (CA)2014
      2. Global Initiative for Asthma. Global strategy for asthma management and prevention. Updated 2015. Available at: www.ginasthma.org. Accessed August 7, 2015.

      3. US Department of Health and Human Services, National Heart Lung and Blood Institute. Expert panel report 3: guidelines for the diagnosis and management of asthma. 2007. Available at: https://www.nhlbi.nih.gov/files/docs/guidelines/asthgdln.pdf. Accessed April 16, 2014.

        • MacGlashan Jr., D.W.
        • Bochner B.S.
        • Adelman D.C.
        • Jardieu P.M.
        • Togias A.
        • McKenzie-White J.
        • et al.
        Down-regulation of Fc(epsilon)RI expression on human basophils during in vivo treatment of atopic patients with anti-IgE antibody.
        J Immunol. 1997; 158: 1438-1445
        • Chanez P.
        • Contin-Bordes C.
        • Garcia G.
        • Verkindre C.
        • Didier A.
        • De Blay F.
        • et al.
        Omalizumab-induced decrease of FcεRI expression in patients with severe allergic asthma.
        Respir Med. 2010; 104: 1608-1617
        • Pelaia G.
        • Gallelli L.
        • Renda T.
        • Romeo P.
        • Busceti M.T.
        • Grembiale R.D.
        • et al.
        Update on optimal use of omalizumab in management of asthma.
        J Asthma Allergy. 2011; 4: 49-59
        • Bousquet J.
        • Cabrera P.
        • Berkman N.
        • Buhl R.
        • Holgate S.
        • Wenzel S.
        • et al.
        The effect of treatment with omalizumab, an anti-IgE antibody, on asthma exacerbations and emergency medical visits in patients with severe persistent asthma.
        Allergy. 2005; 60: 302-308
        • Busse W.
        • Corren J.
        • Lanier B.Q.
        • McAlary M.
        • Fowler-Taylor A.
        • Cioppa G.D.
        • et al.
        Omalizumab, anti-IgE recombinant humanized monoclonal antibody, for the treatment of severe allergic asthma.
        J Allergy Clin Immunol. 2001; 108: 184-190
        • Humbert M.
        • Beasley R.
        • Ayres J.
        • Slavin R.
        • Hébert J.
        • Bousquet J.
        • et al.
        Benefits of omalizumab as add-on therapy in patients with severe persistent asthma who are inadequately controlled despite best available therapy (GINA 2002 step 4 treatment): INNOVATE.
        Allergy. 2005; 60: 309-316
        • Solèr M.
        • Matz J.
        • Townley R.
        • Buhl R.
        • O'Brien J.
        • Fox H.
        • et al.
        The anti-IgE antibody omalizumab reduces exacerbations and steroid requirement in allergic asthmatics.
        Eur Respir J. 2001; 18: 254-261
        • Nopp A.
        • Johansson S.G.
        • Adédoyin J.
        • Ankerst J.
        • Palmqvist M.
        • Oman H.
        After 6 years with Xolair; a 3-year withdrawal follow-up.
        Allergy. 2010; 65: 56-60
        • Nopp A.
        • Johansson S.G.
        • Ankerst J.
        • Palmqvist M.
        • Oman H.
        CD-sens and clinical changes during withdrawal of Xolair after 6 years of treatment.
        Allergy. 2007; 62: 1175-1181
        • Saini S.S.
        • MacGlashan Jr., D.W.
        • Sterbinsky S.A.
        • Togias A.
        • Adelman D.C.
        • Lichtenstein L.M.
        • et al.
        Down-regulation of human basophil IgE and FC epsilon RI alpha surface densities and mediator release by anti-IgE-infusions is reversible in vitro and in vivo.
        J Immunol. 1999; 162: 5624-5630
        • Long A.A.
        • Fish J.E.
        • Rahmaoui A.
        • Miller M.K.
        • Bradley M.S.
        • Taki H.N.
        • et al.
        Baseline characteristics of patients enrolled in EXCELS: a cohort study.
        Ann Allergy Asthma Immunol. 2009; 103: 212-219
        • Malinovschi A.
        • Fonseca J.A.
        • Jacinto T.
        • Alving K.
        • Janson C.
        Exhaled nitric oxide levels and blood eosinophil counts independently associate with wheeze and asthma events in National Health and Nutrition Examination Survey subjects.
        J Allergy Clin Immunol. 2013; 132 (e1-5): 821-827
        • Schleich F.N.
        • Louis R.
        Importance of concomitant local and systemic eosinophilia in uncontrolled asthma.
        Eur Respir J. 2014; 44: 1098-1099
        • Zeiger R.S.
        • Schatz M.
        • Li Q.
        • Chen W.
        • Khatry D.B.
        • Gossage D.
        • et al.
        High blood eosinophil count is a risk factor for future asthma exacerbations in adult persistent asthma.
        J Allergy Clin Immunol Pract. 2014; 2: 741-750
        • Zeiger R.S.
        • Schatz M.
        • Li Q.
        • Chen W.
        • Khatry D.B.
        • Gossage D.
        • et al.
        The association of blood eosinophil counts to future asthma exacerbations in children with persistent asthma.
        J Allergy Clin Immunol Pract. 2015; 3: 283-287.e4
        • Hanania N.A.
        • Wenzel S.
        • Rosén K.
        • Hsieh H.J.
        • Mosesova S.
        • Choy D.F.
        • et al.
        Exploring the effects of omalizumab in allergic asthma: an analysis of biomarkers in the EXTRA study.
        Am J Respir Crit Care Med. 2013; 187: 804-811