The role of ACT score in mepolizumab discontinuation

Abstract Background Mepolizumab is a therapy for severe asthma. We have little knowledge of the characteristics of people in the US that discontinue mepolizumab in clinical care. Objective To investigate the real-world efficacy and time to clinical discontinuation of mepolizumab, we evaluated individuals with asthma started on mepolizumab at the Cleveland Clinic. We hypothesized that individuals that discontinue mepolizumab have more severe and uncontrolled asthma at baseline. Methods Between 2016 and 2022, patients who started on mepolizumab consented to be assessed over 18 months. At baseline, a questionnaire including demographic and medical history was collected. Laboratory findings such as ACT score, FENO (Fractional Excretion of Nitric Oxide), and spirometry were recorded. At the conclusion of the observation period, the participants were divided into two categories: Group A and Group B. Results Group B [N = 28] discontinued mepolizumab (p < 0.05) at an average of 5.8 months (SD 4.2 months). Group A [N = 129] stayed on the therapy for at least 1 year. A participant with an ACT score less than 13 has an odds ratio of 6.64 (95% CI, 2.1 − 26.0) of discontinuing mepolizumab therapy. For a male, the odds of discontinuing mepolizumab therapy is 3.39 (95% CI, 1.1–11.2). Conclusion In this real-world study, we find that high eosinophil count may not be adequate in screening which individuals will benefit from mepolizumab. Up to 17% of patients fail therapy within 6 months, with male sex and low ACT score increasing risk of mepolizumab discontinuation at Cleveland Clinic.


Introduction
The introduction of biologic therapy to the treatment of severe asthma represents an opportunity to carry out phenotype-specific intervention, reinforcing the need for precision in diagnosis (1).Though the Global Initiative for Asthma (GINA) does include a decision tree for how to choose a biologic agent for an individual with severe asthma, many individuals remain uncontrolled and can remain on a biologic agent for longer than necessary (2).People with severe and uncontrolled asthma experience the highest impact on quality of life and accompanying morbidity from asthma.Additionally, according to GINA, the cost of treating individuals with severe asthma is higher when compared to the cost of treating the general population for asthma (2,3), which underscores the benefits for precision medicine in asthma control and healthcare costs.
GlaxoSmithKline's (GSK) mepolizumab, a humanized, interleukin-5 antagonist monoclonal antibody, received approval from the US FDA (United States Federal Drug Administration) in 2015 for add-on treatment in adults with severe eosinophilic asthma (4).The clinical trials that evaluated mepolizumab in patients with asthma included three double-blind, randomized, placebo-controlled studies, including DREAM (Dose Ranging Efficacy And safety with Mepolizumab in severe asthma; NCT01000506), the Phase III MENSA (MEpolizumab as adjuNctive therapy in patients with Severe Asthma; NCT01691521) trial, and SIRIUS (SteroId ReductIon with mepoli-zUmab Study; NCT01691508) (4).The DREAM trial showed that a baseline blood eosinophil count and increased exacerbation frequency in the previous year were associated with good response to mepolizumab.The confirmatory trials required patients to have a blood eosinophil level greater than or equal to 150 cells/µL at screening or blood eosinophil levels greater than or equal to 300 cells/µL within one year of enrollment (5).SIRIUS, a 24-week clinical trial, demonstrated that mepolizumab has a significant glucocorticoid-sparing effect compared with placebo (5).In patients with baseline eosinophil counts greater than or equal to 150 cells/µL, the DREAM, MENSA, and MUSCA (Mepolizumab adjUnctive therapy in subjects with Severe eosinophiliC Asthma; NCT02281318) trials demonstrated that mepolizumab treatment reduces exacerbations, increases forced expiratory volume in one second (FEV 1 ) and improves asthma control and quality of life (4).MENSA, MUSCA, SIRIUS and DREAM are randomized controlled trials, and so do not have the same external validity that a real-world study has.While clinical trials are pivotal for FDA-approval of new therapies, real world applications of medicines can often reveal different outcomes than in prospective blinded trials (6).
To assess real-world effects of mepolizumab, GlaxoSmithKline (GSK) funded REALITI-A (REAL world effectiveness of mepolizumab in patient care-Asthma), a large prospective study with 368 participants.REALITI-A demonstrated a meaningful reduction in daily maintenance oral corticosteroid dose for people with asthma and a decrease in the rate of exacerbations.However, there was not a difference shown for FEV 1 (forced expiratory value in 1 second) % (4).Observational real-world cohorts from Europe also showed that mepolizumab reduces prednisone dependence and improves asthma control (7)(8)(9).Israeli and Australian studies demonstrate an additional FEV 1 % improvement (10,11).REALITI-A is the largest real-world trial which consists of individuals with asthma in the US.However, this industry-sponsored study, in which only 2% of participants discontinue mepolizumab, does not provide analysis on the time to discontinuation of mepolizumab.The results from these real-world studies suggest that more data are needed regarding the characteristics of individuals from the US who discontinue mepolizumab and the time to mepolizumab discontinuation for a person with very severe asthma.
To further investigate the real-world efficacy of mepolizumab, time to discontinuation of mepolizumab, and identification of clinical traits of those who discontinue mepolizumab, we evaluated people with asthma treated with mepolizumab as an add-on therapy at the Asthma Center of Cleveland Clinic over 18 months.The Asthma Control Test (ACT) is a validated questionnaire that we used to establish baseline symptomatic asthma control for each participant.We hypothesized that individuals who discontinue mepolizumab have less severe and uncontrolled asthma at baseline, and they will not demonstrate a change in their symptom control as defined by ACT score within a 6 month period.

Baseline characteristics
Cleveland Clinic's biologic registry includes individuals on omalizumab, mepolizumab, benralizumab, reslizumab, dupilumab, and tezepelumab.An asthma specialist assessed eligible participants in the registry and prescribed mepolizumab based on standard U.S guidelines at the time of study.These guidelines included an absolute blood count eosinophil ≥ 150 cells/µL or corticosteroid dependence and clinical evidence of uncontrolled asthma as demonstrated by two or more asthma exacerbations requiring prednisone, one or more hospitalizations or emergency medical visits or poor symptom control while being on an inhaled corticosteroid and a long-acting beta-agonist.Between 2016 and 2022, we consented 199 people who were prescribed mepolizumab by their provider.The Institutional Review Board (IRB # 8351) at Cleveland Clinic approved the study protocol and procedure.
To compare characteristics of individuals on mepolizumab compared to other biologics, numeric values were reported as mean with standard deviation if the data were symmetrical or median with interquartile range if the data were skewed.T-test was used for parametric data.Categorical variables were analyzed using chi-squared tests.We use Kruskal-Wallis test to compare participants on mepolizumab to those on other biologics.To compare mepolizumab responders and non-responders, we used student t-test.Comorbidities that that had 10%  or higher missing data were excluded.Several columns that had over 10% of "missing" values were excluded.

Longitudinal assessment
The study team assessed each participant at the baseline visit prior to first dose of mepolizumab, and then at 3, 6, 12, and 18 months (Figure 1).At the baseline visit, our team collected a questionnaire which included demographic information, asthma history, medical history, review of systems, asthma medications, and healthcare utilization from each participant (Figure 1).We also recorded peak flow, ACT score, F E NO (exhaled nitric oxide), spirometry and CBC, and we conducted similar assessments of the participants in subsequent visits (Supplementary Figure E1).These assessment tools allow us to establish baseline characteristics and subsequent changes in these characteristics over sequential visits.Peak flow and spirometry assess lung volume and lung function.F E NO is a measure of airway inflammation and indicates eosinophilic inflammation in the individual with asthma, and this is complemented by the eosinophil count from the CBC.These tools contribute to our understanding of each individual's asthma phenotype.
Student t-test was used to demonstrate the differences in their baseline characteristics between group A and group B at different time points.

Outcomes
The outcome of this retrospective study was discontinuation of mepolizumab therapy based on lack of clinical response.Clinical response was determined by the provider's medical reasoning which was informed by the patient's prednisone dependence and healthcare utilization while on mepolizumab.After 18 months of serial assessments, participants who started mepolizumab were divided into two categories:  Group A and Group B. We defined Group A participants as individuals who received mepolizumab for at least six months or received mepolizumab for 18 months without a change in therapy.Group B participants stopped the therapy within 6 months or had a clinical need to switch to a different biologic therapy within 18 months.Chart review excluded participants who stopped mepolizumab therapy due to other reasons.We used a Kaplan-Meier curve to compare time on the drug for Group A and Group B participants.Each responders' time was censored after 18 months.When data were censored, the total time on mepolizumab for a subject cannot be accurately determined past a certain date.Log rank was used to calculate P for the Kaplan-Meier curve.

Statistical analysis
We performed all analyses with R, version 4.1.2(R Project for Statistical Computing, Vienna, Austria),

Inferential analysis
A receiver operative curve (ROC) with C-index was used to determine optimal cutoff for discrete variables with p < 0.05 between responders and non-responders.Binomial logistic regression was utilized to obtain odds ratio with confidence intervals.In this analysis, less than 4% of data were missing and were deemed missing completely at random.

Results
Baseline Characteristics Of the participants who started on mepolizumab (n = 199), the top reason for exclusion is lack of recorded follow-up visits (n = 27).Participants (n = 15) are excluded for medication side effects, loss to follow up, or change in medication indication (Figure 2a).157 (70.4%) of the participants who started on mepolizumab at their baseline visits are included in this analysis (Figure 2b).The group that started mepolizumab and the group that started other biologic agents have similar baseline demographic characteristics and medical comorbidities (Table 1) except for a higher prevalence nasal polyps (Mepolizumab: n = 69 versus [vs] Other Biologics: n = 16, p < 0.05) and sinusitis (Mepolizumab: n = 112 vs Other Biologics:n = 30, p < 0.05) in participants placed on mepolizumab.Participants starting mepolizumab are also more prednisone dependent (Mepolizumab: n = 98 vs Other Biologics: n = 24, p < 0.001).There are no differences in asthma control score, eosinophil counts, and daily oral prednisone use between the two groups (Table 1).

Longitudinal assessments
For the participants placed on mepolizumab (n = 157), there is gradual attrition over visits 3, 6, 12, and 18 months (Figure 3).Despite attrition, median ACT scores improve over sequential visits on mepolizumab (p < 0.05) (Figure 4A).FEV 1 improves rapidly in most individuals between the baseline visit and 3 months, however it is not sustained (p < 0.05) (Figure 4B).There is a striking decline in prednisone use between visit 1 and subsequent visits (p < 0.05) (Figure 4C).

Outcomes
Of the 157 individuals on mepolizumab, 129 are classified as Group A and 28 as Group B (Table 2).Group B participants are on mepolizumab therapy for an average of 5.8 months (standard deviation [SD] 4.2 months, p < 0.05) (Figure 5).Participants in Group A demonstrated an improvement in their ACT scores at each subsequent visit (Figure 6A); whereas participants in Group B did not demonstrate a change in ACT score through subsequent visits (Figure 6B).Group B participants have lower median baseline ACT of 9 (interquartile range [IQR], 4) compared to a median baseline ACT of 13 (IQR, 9) for responders (p < 0.05) (Figure 7).There are no other differences between the baseline characteristics of the two groups (Table 2).Sub-analysis of people with nasal polyps demonstrates that there is no difference in the people who discontinue mepolizumab and those who stay on the medication (Table 3).

Inferential statistics
Based on the observed difference in ACT of the groups, the area under the curve (AUC) of the receiver  operative curve analysis was 0.68 for an ACT cutoff score of 13 (Figure 8).Binomial logistic regression controlled for age, BMI (Body Mass Index), F E NO, and eosinophils.This model demonstrated if an individual has an ACT score less than 13, the odds ratio of being in Group B is 6.64 (95% CI, 2.1-26.0)(Figure 9).Men have an odds ratio of 3.39 (95% CI, 1.1-11.2) for being in Group B (Figure 9)

Discussion
Utilizing this real-world cohort of people with severe asthma, we find that most people on mepolizumab demonstrate clinical improvement shown by an increase in ACT score by 18 months, a decrease in prednisone use as soon as 3 months, and an initial increase in FEV 1 % by 3 months.Our findings also corroborate the REALITI-A trial and other real-world  studies but additionally provide a unique perspective on mepolizumab discontinuation in the United States.This large US real-world mepolizumab study shows that despite having eosinophils ≥ 150 cells/µL and history of recurrent asthma exacerbations, two variable have an effect on early discontinuation of mepolizumab: male sex or a baseline ACT score < 13.Our findings also show that if mepolizumab is discontinued, the discontinuation will occur within 6 months of starting the drug; those that stay on mepolizumab and demonstrate an improvement in ACT score are on mepolizumab for at least an average of 1 year.The results suggest that poorly controlled asthma at baseline and men with very symptomatic asthma may have a different biologic endotype less responsive to anti-IL5 therapy, which supports our hypothesis.A sub-analysis of Group B, which includes those individuals initially excluded for side effects, demonstrates that sex remains a critical variable for discontinuing mepolizumab at 6 months.Additionally in the sub-analysis, those with a lower baseline ACT score are also more likely to discontinue mepolizumab at 6 months.These findings in the intention-to-treat sub-analysis bolster the initial findings (Table 4).Group B participants do not do as well on mepolizumab compared to Group A participants.Predictive biomarkers for response to mepolizumab treatments include blood eosinophil counts as the strongest predictor for the reduction in exacerbation rates with mepolizumab compared to placebo (12).However, in our study, eosinophil counts are similar among responders and non-responders.Therefore, participants of Group B may have an additional complex neutrophilic or pauci-granulocytic mechanism for their asthma because pure Type 2 (T2) inflammation-mediated asthma should respond well to mepolizumab (13).
Our findings support sex as a risk factor for mepolizumab discontinuation.Though 31% of the participants in our study on mepolizumab are male, they are found to have an odds ratio of 3.39 of failing mepolizumab therapy.The Australian Mepolizumab Registry also found that males are more likely to fail mepolizumab therapy (10).Sex hormones play a role in asthma severity, and androgens such as testosterone may reduce asthma incidence and symptoms (14).Therefore, the finding that males have higher odds of failing mepolizumab therapy is unexpected and may suggest underlying sex-based mechanism for this subset.
The most interesting difference between our study and the industry-sponsored studies is the BMI of our patient population.Our study has a higher median BMI when compared to the BMI of REALIT-A study participants (median, IQR: BMI = 27.5 [8.7] kg/m 2 ), which is the pharmacy-sponsored real-world study for mepolizumab.BMI values of Group B trend higher than the BMI values of Group A (BMI, Group A: 32.5 [7.8] kg/m 2 vs Group B: 31 [12], p = 0.513 kg/ m 2 ).Obesity, which causes changes to normal lung physiology, innate immune function, and changes in the gut microbiome, is a major risk factor and disease in asthma for adults (15).In this study, Obesity causes mechanical changes such as lung compression and reduction in lung volume (15).As BMI increases, there is a higher risk for airway hyperreactivity (16).Additionally, the Western dietary pattern, which promotes obesity, might also affect the development of asthma via changes in the gut microbiome (15).We believe that the median BMI in our study is representative of individuals who are placed on mepolizumab in the real-world.
While this study has many strengths such as strong external validity because it is a real-world study and has a diverse demographic population, it has three limitations.In a real-world study, there are no controls and no randomization.The first limitation in our study is that participants who started on mepolizumab compared to those that started on other biologic agents have higher a prevalence of chronic rhinosinusitis with nasal polyposis (CRSwNP).After the randomized controlled trial, SYNAPSE, mepolizumab was approved for nasal polyps in the setting of severe chronic rhinosinusitis (17).However, those with greater type 2 inflammation often have concomitant CRSwNP, so it is not surprising that clinicians chose mepolizumab for those with this comorbidity (18).Since the people started on mepolizumab have a higher incidence of CRSwNP in this study, these people likely have greater Th2 inflammation.Therefore, the population on mepolizumab in our cohort has a more severe eosinophilic asthma at baseline than the people placed on other biologics.Those on mepolizumab do have higher prednisone utilization, which supports greater Th2 inflammation (Prednisone, Mepolizumab: 62% vs Other Biologics: 36%, p < 0.001).Our sub-analysis of the cohort with CRSwNP demonstrated that there was no difference in the people that stayed on mepolizumab and those that discontinued it early (Table 3).
A second limitation is that the individuals in Group A have either milder asthma or better controlled at baseline as they have a higher baseline ACT score and a better baseline FEV 1 (72% vs 62% predicted, p = 0.07).A possibility for the underperformance of Group B with mepolizumab is that Group B participants may not have well controlled asthma at baseline due to lack of appropriate inhaler therapy.However, we know that all patients in this study are on high-dose inhaled corticosteroid in addition to a second long-acting controller medication, therefore, their medications are similar to the Group A participants' inhaler regimen.This possibility does not explain the low performance on mepolizumab for Group B participants.
Finally, since this was a real-world study, there is also attrition between the baseline visit and visit 18, which reduces the power of our study.If an individual left the study to receive their asthma management at another facility, their information was not recorded.This further reduces the power of our study.It is remarkable that given the limitations, we still were able to corroborate the findings of DREAM, MENSA, MUSCA, SIRIUS, and REALITI-A, and we were also able to find a strong relationship between ACT score and mepolizumab response.
In conclusion, upto 17% of patients discontinue therapy within 6 months, with male sex and low ACT score increasing risk of early discontinuation.One mechanism for low ACT scores in the people that discontinue mepolizumab early could be related to an altered immune response associated with higher BMI favoring Th2 low pathways.This phenotype could be less responsive to biologics targeting Th2 high pathways, despite having a high baseline eosinophil count.Further studies are needed to determine which, if any, biologics might perform better in those people who suffer with severe asthma and poor control.

Figure 1 .
Figure 1.flowchart for mepolizumab.this flowchart delineates information collected at visits at baseline, 3, 6, 12, and 18 months.Questionnaires were collected at every visit.aCt, medication use, and spirometry results were also recorded at every visit.Definitions: f E no = fractional exhaled nitric oxide; CBC = Complete Blood Count; aCt = asthma Control test

Figure 2 .
Figure 2. a. mepolizumab registry.the registry has 199 people initiated on mepolizumab between the years of 2016 and 2022.a total of 27 people were removed for having only one visit recorded within the database.15 people were removed for stopping mepolizumab for reasons other than clinical failure. of the remaining 157 participants, 129 people are clinically considered mepolizumab responders as defined by an asthma specialist.figure 2(B) is a pie chart demonstrating the reasons people were removed from this study.most people that were removed had only a single visit recorded or were lost to follow up.

Figure 3 .
Figure 3. attrition in study.the number of participants per visit is shown in this bar graph which shows attrition over subsequent visits.

Figure 4 .
Figure 4. mepolizumab outcomes.Panel a demonstrates median aCt scores.Each visit has an increase in aCt scores.Panel B demonstrates the greatest increase in median fEV 1 % occurs between visit 1 and visit 3. Panel C shows the decrease in prednisone use for participants while on mepolizumab.Panel D demonstrates the number of individuals on prednisone during spirometry.the asterisk (*) is for p < 0.05.

Figure 5 .
Figure 5. time to Discontinuation.time to discontinuation of mepolizumab or censorship between the two cohorts is different with p < 0.05. the Kaplan-meier curve demonstrates the average time someone from Group B is on mepolizumab and the average time to censorship for Group a. Group B stay on mepolizumab for a mean of 5.8 months (standard deviation [sD] 4.2 months).Group a are on mepolizumab for a mean of 12.3 months (sD 4.6 months) to the time of censorship.the tick marks for group a represent when our database stops recording the participant's information though they may still be on mepolizumab.

Figure 6 .
Figure 6.a. mean aCt score in Group a and B over 12 months.this graph depicts the mean aCt score with standard deviation of Group a, which is the group of people that remained on mepolizumab.through subsequent visits for Group a, the mean aCt score of the group significantly improves from the baseline aCt score.Graph 6B demonstrates the mean aCt score for Group B does not significantly change between the baseline visit and 12 months.

Figure 7 .
Figure 7. Baseline median aCt scores in Group a and Group B. Group B present with a remarkably lower baseline aCt score prior to starting mepolizumab therapy compared to those who stay on mepolizumab for longer than 6 months (aCt: median [IQr], Group B, 9 [ 4]; Group a, 13 [IQr9]).Definitions: aCt = asthma Control test

Figure 8 .Figure 9 .
Figure 8. receiver operative Curve in asthma.receiver operating curve (roC) for aCt score in all participants started on mepolizumab showed an aCt cutoff score of 13.Definitions: roC = receiver operating Curve; auC = area under the Curve

Table 1 .
Baseline characteristics of all participants.

Table 2 .
Baseline Characteristics of Individuals on mepolizumab.
Characteristics of Group a and Group B are delineated in this table.those in Group B have remarkably lower baseline aCt score (aCt, median [IQr], aCt: 13

Table 3 .
sub-analysis of nasal polyps and mepolizumab.