Emerging biological therapies for treating chronic obstructive pulmonary disease: A pairwise and network meta-analysis

Emerging biological therapies for treating chronic obstructive pulmonary disease: A pairwise and network meta-analysis

Accepted Manuscript Emerging biological therapies for treating chronic obstructive pulmonary disease: A pairwise and network meta-analysis Paola Rogli...

1MB Sizes 0 Downloads 0 Views

Accepted Manuscript Emerging biological therapies for treating chronic obstructive pulmonary disease: A pairwise and network meta-analysis Paola Rogliani, Maria Gabriella Matera, Ermanno Puxeddu, Marco Mantero, Francesco Blasi, Mario Cazzola, Luigino Calzetta PII:

S1094-5539(18)30057-9

DOI:

10.1016/j.pupt.2018.03.004

Reference:

YPUPT 1716

To appear in:

Pulmonary Pharmacology & Therapeutics

Received Date: 28 February 2018 Revised Date:

28 March 2018

Accepted Date: 29 March 2018

Please cite this article as: Rogliani P, Matera MG, Puxeddu E, Mantero M, Blasi F, Cazzola M, Calzetta L, Emerging biological therapies for treating chronic obstructive pulmonary disease: A pairwise and network meta-analysis, Pulmonary Pharmacology & Therapeutics (2018), doi: 10.1016/ j.pupt.2018.03.004. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Emerging biological therapies for treating chronic obstructive pulmonary disease: a pairwise and network meta-analysis

Paola Rogliani1*, Maria Gabriella Matera2, Ermanno Puxeddu1, Marco

RI PT

Mantero3,4, Francesco Blasi3,4, Mario Cazzola1, Luigino Calzetta1

1. Department of Experimental Medicine and Surgery, University of Rome Tor Vergata, Rome, Italy.

SC

2. Department of Experimental Medicine, University of Campania Luigi Vanvitelli, Naples, Italy.

M AN U

3. Department of Pathophysiology and Transplantation, University of Milan.

4. Internal Medicine Department, Respiratory Unit and Regional Adult Cystic Fibrosis Center, IRCCS Fondazione Cà Granda Ospedale Maggiore Policlinico, Milan.

TE D

*Corresponding author: Paola Rogliani, Department of Systems Medicine, University of Rome Tor Vergata, Via Montpellier 1, 00133, Roma, Italy; telephone number: +39 06 2090 4656; e-mail: [email protected]

AC C

None

EP

Declaration of interest

Abbreviations

COPD: chronic obstructive pulmonary disease; CrI: Credible Interval; ERS: European Respiratory Society; FEV1: forced expiratory volume in 1 second; GRADE: Grading of Recommendations Assessment, Development, and Evaluation; IL: interleukin; mABs: monoclonal antibodies; MCID: minimal clinical important difference; MD: Mean Difference; OIS: optimal information size; PRISMA: Preferred Reporting Items for Systematic Reviews and MetaAnalyses; QoL: quality of life; RCTs; randomized clinical trials; RE: Relative Effect; RR: Risk Ratio; SE: standard error; SGRQ: St. George’s Respiratory

1

ACCEPTED MANUSCRIPT Questionnaire; SUCRA: surface under the cumulative ranking curve; Tc: Tcytotoxic; TGF-β: transforming growth factor; Th: T-helper; TNF: tumor necrosis factor; TSLP: thymic stromal lymphopoietin.

Acknowledgements

RI PT

None.

Conflict of interests

PR, MGM, EP, MC and LC have no relevant affiliations or financial

SC

involvement with any organization or entity with a financial interest in, or financial conflict with, the subject matter or materials discussed in the manuscript, including employment, consultancies, honoraria, stock ownership

royalties.

Funding sources

M AN U

or options, expert testimony, grants or patents received or pending, or

This study was supported by institutional funds (University of Rome “Tor

AC C

EP

TE D

Vergata”, Rome, Italy).

2

ACCEPTED MANUSCRIPT

Abstract Inflammation in chronic obstructive pulmonary disease (COPD) is often corticosteroid resistant and, thus, alternative anti-inflammatory approaches are needed. Since it is still not clear whether blocking specific proinflammatory factors may provide clinical benefit in COPD, we have

RI PT

performed a meta-analysis to quantify the impact of monoclonal antibodies (mABs) targeting the cytokine/chemokine-mediated inflammation in COPD.

A pairwise and network meta-analyses were performed by extracting data from randomized clicnial trials on COPD concerning the impact of mABs vs.

SC

placebo on the risk of exacerbation, forced expiratory volume in 1 second (FEV1), and St. George’s Respiratory Questionnaire (SGRQ).

Data on the interleukin (IL)-1β antagonist canakinumab, IL-1R1 antagonist

M AN U

MEDI8986, IL-5 antagonist mepolizumab, IL-5R antagonist benralizumab, IL-8 antagonist ABX-IL8, and TNF-α antagonist infliximab were found. Overall, mAB therapy had a moderate impact on the risk exacerbation, but not on FEV1 and SGRQ. The pairwise meta-analysis performed in eosinophilic patients, and the network approach, indicated that mepolizumab elicited a

TE D

beneficial effect against the risk of exacerbation, whereas benralizumab was more effective in improving both FEV1 and SGRQ. This study demonstrates that targeting the pathway activated by IL-5 may

Keywords

EP

have a beneficial impact in eosinophilic COPD patients.

AC C

COPD, monoclonal antibodies, benralizumab, mepolizumab, meta-analysis.

3

ACCEPTED MANUSCRIPT

1. Introduction Chronic obstructive pulmonary disease (COPD) is an inflammatory disorder in which the persistent respiratory symptoms and airflow limitation are due to a complex chronic inflammation that predominantly affects peripheral airways and lung parenchyma [1, 2].

RI PT

The inflammatory response to noxious particle and/or gases is prevalently orchestrated by the activation of inflammatory cells such as alveolar macrophages, neutrophils, T lymphocytes, and innate lymphoid cells recruited from the circulation. Furthermore, in some COPD patients there may also be

SC

increases in eosinophils. These inflammatory and structural cells, including epithelial and endothelial cells and fibroblasts, secrete several cytokines and chemokines [3]. These mediators have a relevant role in the pathogenesis

M AN U

and progression of COPD, since they recruit inflammatory cells from the circulation into the lungs, maintain inflammation, and lead to characteristic structural changes in the airways [2, 3].

Relevant advances have been achieved for the identification of multiple cytokines and chemokines involved in chronic inflammatory disorders, but the

TE D

effort to characterize a specific inflammatory profile in patients with COPD produced conflicting results [3]. In fact, some studies suggested that COPD is mainly characterized by a T-helper (Th)-1 and T-cytotoxic (Tc)-1 subtype (Th1/Tc1) pattern, whereas further researches found a predominant Th2/Tc2

EP

phenotype [2]. Therefore, the cluster of cytokines and chemokines that orchestrate inflammation in COPD may involve tumor necrosis factor (TNF)-α,

AC C

interleukin (IL)-1β, IL-4, IL-5, IL-6, IL-8, IL-13 IL-17, IL-18, IL-23, IL-33, and thymic

stromal lymphopoietin (TSLP), and growth

factors such as

transforming growth factor (TGF)-β [2]. Little progress has been made in understanding the role of these factors in COPD, and how the multiple cytokine and chemokine network operate [3]. Nevertheless, the inhibition of these mediators, and the pathways that they activate, may represent a potential strategy to modulate the inflammatory processes that sustains the progression of COPD [4]. Moreover, since inflammation in COPD is largely corticosteroid resistant, alternative antiinflammatory approaches are strongly needed [5, 6]. Unfortunately, the

4

ACCEPTED MANUSCRIPT currently available alternative anti-inflammatory strategies approved for the treatment of COPD are often limited by several transient mild-to-moderate adverse events, that may reduce the compliance to treatment [1, 7]. As a consequence, to date there is a huge unmet medical need with regard to effective and well tolerated anti-inflammatory agents for treating patients with

RI PT

COPD [8]. Blocking cytokine and chemokines, their synthesis, antagonizing their receptors and targeting intracellular signalling pathways by using monoclonal antibodies (mABs) is a biological therapeutic approach that has been

SC

successful in the treatment of chronic inflammatory diseases, namely severe asthma, rheumatoid arthritis, and inflammatory bowel disease [6]. Indeed, mABs have advantages over small molecules, such as the high

M AN U

affinity and specificity for their targets, and the metabolic stability that allows them to be active for long time and have a long duration of action (from weeks to months). Furthermore, since their breakdown products are amino acids, they are not converted into toxic metabolites [9, 10]. Consequently, several biological inhibitors of inflammatory mediators have been developed for

TE D

targeting the cytokine/chemokine-mediated inflammation in COPD [2]. To date, it is still not clear whether blocking specific cytokines and chemokines may provide clinical benefit in COPD patients [3]. Therefore, we have performed a pairwise meta-analysis aimed to quantify the overall impact

EP

of mABs tested in randomized clinical trials (RCTs) with regard to functional and clinical outcomes of COPD. Furthermore, we have also carried out a

AC C

network approach to compare and rank the effectiveness of specific mABs on the risk of COPD exacerbation, lung function, and health-related quality of life in eosinophilic COPD pateints.

5

ACCEPTED MANUSCRIPT

2. Material and methods 2.1. Search strategy This pairwise and network meta-analysis has been registered in PROSPERO (registration

number:

CRD42017073945;

available

at

https://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD4201707

RI PT

3945), and performed in agreement with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement (Figure 1) [11]. Furthermore, this synthesis satisfied all the recommended items reported by

AC C

EP

TE D

M AN U

SC

the PRISMA-P 2015 checklist [12].

Figure 1. PRISMA flow diagram for the identification of studies included in the meta-analysis concerning the impact of mABs in COPD. mABs: monoclonal antibodies.

6

ACCEPTED MANUSCRIPT We undertook a comprehensive literature search for RCTs evaluating the impact of mABs in patients suffering from COPD, diagnosed by pulmonary function testing. The currently available mABs tested in COPD were searched. In particular,

RI PT

the IL-1β antagonist canakinumab, IL-1R1 antagonist MEDI8986, IL-5 antagonists mepolizumab and reslizumab, IL-5R antagonist benralizumab, the IL-8 antagonist ABX-IL8, TNF-α antagonists etanercept and infliximab, IL-13 antagonist lebrikizumab, IL-17 antagonist CNTO6785, and IL-33 AMG 282

SC

were searched for the mABs, and the terms “chronic obstructive pulmonary disease” and/or “COPD” were searched for the disease.

The search was performed in PubMed, Scopus, Embase, and Google Scholar

M AN U

[13] through October 2017, in order to provide for relevant published RCTs reported in English and published up to October 27, 2017. Further search was carried out on ClinicalTrials.gov, the EU Clinical Trials Register, and the European Respiratory Society (ERS) Congress Abstract Book in order to find not yet published RCTs. Citations of previous published reviews were

TE D

checked to select further pertinent RCTs, if any [2].

Two reviewers independently checked the relevant studies identified from literature searches and databases. The studies were selected in agreement with the previously mentioned criteria, and any difference in opinion about

EP

eligibility was resolved by consensus.

AC C

2.2. Study selection

RCTs reporting data concerning the impact of mABs vs. placebo on exacerbation, lung function and health-related quality of life (QoL) in COPD patients were selected, and included in the quantitative meta-analysis. No restriction on the duration of the treatment was applied. Two reviewers independently examined the clinical trials and any difference in opinion about eligibility was resolved by consensus.

2.3. Data extraction Due to the complexity of this meta-analysis, data have been extracted in agreement with the DECiMAL recommendations and the Cochrane Handbook 7

ACCEPTED MANUSCRIPT for Systematic Reviews of Interventions and [14, 15]. Data were extracted and checked for study characteristics and duration, doses of medications, patients characteristics and phenotypes, age, gender, smoking habits, COPD exacerbation, forced expiratory volume in 1 second (FEV1), St. George’s

RI PT

Respiratory Questionnaire (SGRQ), and Jadad score.

2.4. Outcomes

The primary endpoints of this quantitative synthesis were the impact of mABs on the risk of exacerbations and lung function in both eosinophilic and non-

SC

eosinophilic COPD patients, compared to placebo and across specific mABs. The secondary endpoint was the impact of mABs on the health-related QoL measured via SGRQ in both eosinophilic and non-eosinophilic COPD

M AN U

patients, compared to placebo and across specific mABs.

Exploratory endpoint was the safety profile in eosinophilic COPD patients, compared to placebo and across specific mABs.

2.5. Quality score, risk of bias and evidence profile

TE D

The Jadad score, with a scale of 1 to 5 (score of 5 being the best quality), was used to assess the quality of the RCTs concerning the likelihood of biases related to randomization, double blinding, withdrawals and dropouts [16]. Two reviewers independently assessed the quality of individual studies, and any

EP

difference in opinion about the quality score was resolved by consensus. The risk of publication bias was assessed fro primary endpoints by applying

AC C

the funnel plot and Egger’s test through the following regression equation: SND = a + b × precision, where SND represents the standard normal deviation (treatment effect divided by its standard error [SE]), and precision represents the reciprocal of the standard error. Evidence of asymmetry from Egger’s test was considered to be significant at P<0.1, and the graphical representation of 90% confidence bands are presented [16]. The optimal information size (OIS) was calculated as previously described [7], and the quality of the evidence has been assessed in agreement with the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system [17].

8

ACCEPTED MANUSCRIPT 2.6. Data analysis The data extracted by the overall COPD population were analyzed via pairwise approach, and a subset analysis was performed on COPD patients characterized by an eosinophilic phenotype (blood eosinophil count ≥150 per mm3 at screening or ≥300 mm3 during the previous year). After that, a network

RI PT

meta-analysis was performed by using exclusively the data extracted from RCTs in which COPD patients who had an eosinophilic phenotype were enrolled.

Since the follow-up duration was not consistent among the RCTs included in

SC

this meta-analysis, the data have been normalized as a function of personseason, where one season lasts three months [18]. This method involved the conversion of the measures into a common metric (events per person-time)

M AN U

prior to meta-analyze the data, leading to increased estimates of effect, precision, and clinical interpretability of results [15, 19].

Results of the pairwise meta-analysis are expressed as Risk Ratio (RR) and Mean Difference (MD), and 95% confidence interval (95%CI). Since data were selected from a series of studies performed by researchers operating

TE D

independently, and a common effect size cannot be assumed, we used the random-effects model to perform the pairwise meta-analysis in order to balance the study weights and adequately estimate the 95%CI of the mean distribution of drugs effect on the investigated variables [16]. The mathematics

EP

behind the fixed-effects model are much simpler that those of the randomeffects model, and the results of this quantitative synthesis cannot be

AC C

generalized via fixed-effects model since the included studies were quite dissimilar [20]. Therefore, the greater the degree of difference among the studies incorporated in the analysis, the more important it becomes to employ the random-effects model [21]. The network meta-analysis was performed to indirectly compare the effect of specific mABs in COPD patients characterized by an eosinophilic phenotype. A full Bayesian evidence network was used (chains: 4; initial values scaling: 2.5; tuning iterations: 20.000; simulation iterations: 50.000; tuning interval: 10), and the convergence diagnostics for consistency and inconsistency was assessed via the Brooks-Gelman-Rubin method, as previously reported [22]. Due to the characteristics of parameters besides the available data, the just 9

ACCEPTED MANUSCRIPT proper non-informative distributions specified the prior densities, in agreement with the Bayesian Approaches to Clinical Trials and Health-Care Evaluation [23, 24]. Since the distributions were sufficiently vague, the reference treatment, study baseline effects, and heterogeneity variance were unlikely to have a noticeable impact on model results. In this condition, GeMTC software

RI PT

automatically generates and runs the required Bayesian hierarchical model and selects the prior distributions and starting values as well, via heuristically determining a value for the outcome scale parameter (i.e. outcome scale S) [25, 26]. The posterior mean deviance of data points in the unrelated mean

SC

effects model were plotted against their posterior mean deviance in the consistency model in order to provide information for identifying the loops in the treatment network where evidence was inconsistent [27]. Results of the

M AN U

network meta-analysis are expressed as Relative Effect (RE) and 95% Credible Interval (95%CrI). The probability that each intervention arm was the most effective was calculated by counting the proportion of iterations of the chain in which each intervention arm had the highest mean difference, and the surface under the cumulative ranking curve (SUCRA), representing the

TE D

summary of these probabilities, was also calculated. The SUCRA is 100% when a treatment is certain to be the best, and 0% when a treatment is certain to be the worst [28].

OpenMetaAnalyst [29] and GeMTC [25] software were used for performing

EP

the meta-analysis, GraphPad Prism (CA, US) software to graph the data, and GRADEpro to assess the quality of evidence [17]. The statistical significance was assessed for P<0.05, and moderate to high levels of heterogeneity were

AC C

considered for I2>50%.

10

ACCEPTED MANUSCRIPT

3. Results 3.1. Studies characteristics Results obtained from 2,453 COPD patients (1,375 were treated with a mAB and 1,078 with placebo) were selected from 7 studies carried out on canakinumab [30], MEDI8986 [31], mepolizumab [32, 33], benralizumab [34],

RI PT

ABX-IL8 [35], and infliximab [36]. Only the studies that investigated mepolizumab and benralizumab reported also information on sputum and/or blood eosinophil count (1,434 patients had an eosinophilic phenotype) [3234].

SC

Three studies on TNF-α antagonists were excluded from this quantitative synthesis because were observational studies (non-RCTs) [37], did not evaluate the outcomes considered by the endpoints of this analysis [38], or

M AN U

results were not suitable for meta-analysis since data were presented as medians with ranges [39]. Data from RCTs carreid out in COPD patients treated with IL-13 (NCT02546700), IL-17 (NCT01966549), and IL-33 mABs [40] were not available. The mABs against IL-4, IL-6, IL-23, and TGF-β were not tested in COPD patients, as well as the IL-5 antagonist reslizumab.

TE D

All the included studies were randomized and double-blinded, were published between 2004 and 2017, the period of treatment ranged from 24 to 56 weeks, and were characterized by a Jadad score ≥3. More details on studies

AC C

EP

characteristics are reported in Table 1.

11

ACCEPTED MANUSCRIPT

Table 1. Patient demographics, baseline and study characteristics.

Study and year

Stu dy dur ati on (w ee ks)

An aly sed pati ent s

Study characteristics

Dasgupt a et al., 2017

NCT01463644

Proof of principle, single centre, randomised, double-blind, placebocontrolled, parallel group.

24

18

Mepolizumab (750 mg, monthly, IV)

Pavord et al., 2017

NCT02105948 (METREX), NCT02105961 (METREO)

Multicenter, randomized, placebocontrolled, double-blind, parallel group.

52

1,5 10

Mepolizumab (100 mg and 300 mg, monthly, SC)

Calverle y et al., 2015

NCT01448850

Multicenter, randomized, double-blind, placebo-controlled, parallel group.

52

324

MEDI8968 (600 mg IV on day 1, then 300 mg SC every 4 weeks)

Disease characteristics

Moderate to severe COPD (postbronchodilator FEV1/FVC <0.7; FEV1 <60% predicted)

Moderate to severe COPD (postbronchodilator FEV1/FVC <0.7; FEV1 > 20% and ≤80% predicted)

EP

TE D

M AN U

SC

Drugs and doses

RI PT

Clinical trial identification

Phe noty pe a

Eosi noph ilic phen otyp e Both eosi noph ilic and noneosi noph ilic phen otyp es (ME TRE X); eosi noph ilic phen otyp e (ME TRE O)

66 .0

7 2. 0

NA

39. 0

46.8

3

65 .4

6 4. 6

27. 6

>10

45.3

4

Moderate to very severe COPD

NA

62 .9

6 7. 9

NA

NA

39.1

4

63 .8

8 1. 5

19. 0

49. 4

47.1

4

NA

>20

NA

3

46. 6

>10

43.4

4

Moderate to severe COPD (postbronchodilator FEV1/FVC <0.7; FEV1 <80% predicted)

45

147

Canakinumab (initial dose of 1 mg/kg IV, four weeks later a dose of 3 mg/kg, and another dose of 3 mg/kg two weeks later; thereafter, doses of 6 mg/kg every four weeks until completion of the treatment period)

COPD (post-bronchodilator FEV1/FVC <0.7; FEV1 ≤80% predicted)

NA

63 .7

24

234

Infliximab (3 mg/kg or 5 mg/kg at weeks 0, 2, 6, 12, 18, and 24)

Moderate to severe COPD

NA

65 .1

56

Novartis, 2011

NCT00581945

Exploratory, randomized, double-blind, placebo controlled, parallel group.

Rennard et al., 2007

NA

Multicenter, randomized, double-blind, placebo-controlled, parallel group.

AC C

Benralizumab (100 mg SC; the first three doses were given at weeks 1, 4, and 8, followed by doses at weeks 16, 24, 32, 40, and 48)

Randomised, double-blind, placebocontrolled, parallel group.

Jada d scor e

M al e ( % )

101

NCT01227278

Postbronc hodil ator FEV1 (% predi cted)

Ag e (y ea rs)

Both eosi noph ilic and noneosi noph ilic phen otyp es

Brightlin g et al., 2014

Sm oki ng his tor y (pa ckyea rs)

Cu rre nt sm ok ers (%)

5 9. 8 5 9. 0

12

ACCEPTED MANUSCRIPT

Mahler et al., 2004

NA

Pilot, randomized, double-blind, placebo controlled, parallel group.

12

119

ABX-IL8 (loading dose of 800 mg IV administered at the baseline visit, two doses of 400 mg administered monthly thereafter) 3

COPD and chronic bronchitis (postbronchodilator FEV1/FVC <0.7; FEV1 ≥30% and ≤70% predicted)

NA

64 .0

5 4. 5

NA

68. 5

42.0

5

3

AC C

EP

TE D

M AN U

SC

RI PT

a. the eosinophilic phenotype was identified on the basis of blood eosinophil counts (≥150 eosinophils per mm at screening or ≥300 eosinophils mm during the previous year). COPD: chronic obstructive pulmonary disease FEV1: forced expiratory volume in 1 second FVC: forced vital capacity IV: intravenous NA: not available SC: subcutaneous

13

ACCEPTED MANUSCRIPT 3.2. Meta-analysis 3.2.1.

Primary endpoints

The overall pairwise meta-analysis indicated that mABs induced a weak although significant (P<0.05) reduction of the risk of exacerbation in the overall COPD population, compared with placebo (RR 0.90, 95%CI 0.83 –

RI PT

0.98; I2 21%) (Figure 2A). The subset pairwise meta-analysis performed by including exclusively eosinophilic patients inducated that mepolizumab, but not benralizumab, was effective in reducing the risk of COPD exacerbation (Figure 2B). The network meta-analysis did not evidence any significant

SC

difference between mepolizumab and benralizumab with regard to the protection against COPD exacerbation in eosinophilic patients (mepolizumab vs. benralizumab: RE 0.77, 95%CrI 0.37 – 1.68). However, a further analysis

M AN U

performed by plotting the data of SUCRA analysis and the rank of being the best treatment showed that mepolizumab was more effective than benralizumab, and that benralizumab was weakly more effective than placebo (Figure 2C).

The overall pairwise meta-analysis indicated that the treatment with mABs did not significantly (P>0.05) increase FEV1 in COPD patients (+15 ml, 95%CI -26

TE D

– 55, vs. placebo; I2 30%) (Figure 2D). Although this result was confirmed by the subset pairwise meta-analysis performed in eosinophilic patients, in this population a signal of effectiveness was detected for benralizumab with

EP

regard to the increase in FEV1 (+277 ml, 95%CI -10 – 563, vs. placebo) (Figure 2E). The network meta-analysis carried out in eosinophilic patients

AC C

showed that benralizumab had a numerical, albeit not significant, greater beneficial impact on FEV1 than mepolizumab (benralizumab vs. mepolizumab: RE 268, 95%CrI -48 – 617). The combined SUCRA and ranking plot confirmed that benralizumab was more effective than mepolizumab, and that mepolizumab was almost as effective as placebo concerning the increase in FEV1 (Figure 2F). 3.2.2. Data

Secondary endpoint on

SGRQ

were

available for only benralizumab, MEDI8986,

mepolizumab and ABX-IL8. The overall pairwise meta-analysis showed that mAB therapy did not significantly (P>0.05) improve the change from baseline 14

ACCEPTED MANUSCRIPT in SGRQ (-0.20 units, 95%CI -1.47 – 1.07, vs. placebo; I2 0%) (Figure 2G). Although also the subset pairwise meta-analysis showed that neither benralizumab nor mepolizumab induced a significant beneficial impact on SGRQ in eosinophylic patients (Figure 2H), the network approach indicated that benralizumb was more effective than mepolizumabm and that

RI PT

mepolizumab was more effective than placebo in reducing the SGRQ (Figure 2I).

3.2.3.

Exploratory endpoint

SC

No significant difference (P>0.05) was detected across benralizumab, mepolizumab and placebo with regard to the overall risk of adverse event in eosinophilic COPD patients (benralizumab vs. mepolizumab: RE 1.04,

M AN U

95%CrI 0.32 – 3.57; benralizumab vs. placebo: RE 1.39, 95%CrI 0.47 – 4.50; mepolizumab vs. placebo: RE 1.34, 95%CrI 0.93 – 1.97). The subset analysis of the risk of serious adverse events confirmed that no significant (P>0.05) difference exists across benralizumab, mepolizumab and placebo concerning the safety profile (benralizumab vs. mepolizumab: RE 2.34, 95%CrI 0.76 –

TE D

7.12; benralizumab vs. placebo: RE 1.85, 95%CrI 0.65 – 5.20; mepolizumab

AC C

EP

vs. placebo: RE 0.79, 95%CrI 0.56 – 1.13).

15

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 2. Impact of mAB therapy on COPD exacerbation (A, B, and C), change from baseline in trough FEV1 (D, E, and F), and SGRQ (G, H, and I). Data have been analyzed via pairwise meta-analysis (effectiveness of mABs vs. placebo) in the overall population (A, D and G) and in COPD patients characterized by eosinophilic phenotype (B, E and H). Results of the network meta-analysis are reported for patients characterized by eosinophilic phenotype (C, F and I), via the ranking plot of the network across the effectiveness of mABs and placebo: treatments have been plotted on X-axis according to SUCRA (score of 1 being the most effective) and on Y-axis according to the rank of being the best treatment (score of 1 being the most effective). CI: Confidence Interval; COPD: chronic obstructive pulmonary disease; FEV1: forced expiratory volume in 1 second; HD: high dose; LD: low dose; mABs: monoclonal antibodies; NA: not available SGRQ: St. George’s Respiratory Questionnaire; SUCRA: surface under the cumulative ranking curve.

16

ACCEPTED MANUSCRIPT 3.3. Bias and quality of evidence No heterogeneity was detected in the pairwise meta-analysis for the impact of mABs on both COPD exacerbation (I2 21%, P=0.26) and

FEV1 (I2 30%,

P=0.19), as well as in the subset analysis carried out by considering exclusively eosinophilic patients (exacerbations: I2 0%, P=0.91; FEV1: I2 9%,

RI PT

P=0.35). The consistency/inconsistency analysis of the network meta-analysis indicated that all the points fit adequately with the line of equality for both the risk of COPD exacerbation (Pearson r 0.99, slope 1.05, P<0.001) and change

SC

from baseline in FEV1 (Pearson r 0.97, slope 0.99, P<0.001) (Figure 3A and B).

Overall, this meta-analysis met a reasonable OIS to ensure a very good low

M AN U

risk of observing an overestimated intervention effect due to random errors (power 90%, relative risk reduction 20%; control group risk 15% - 40%). The visual inspection of funnel plot showed no asymmetry, and Egger’s tests did not identify any significant bias with regard to the impact of mABs on both COPD exacerbations and change in FEV1 (Figure 3C - F).

TE D

The GRADE approach indicated high quality of evidence (++++) for both the

AC C

EP

risk of COPD exacerbation and the change from baseline in FEV1 (Table 2).

17

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 3. Publication bias assessment via consistency/inconsistency plot (A and B), funnel plots (C and D) and Egger’s test (E and F) for the impact of mABs on the primary endpoints (COPD exacerbation, left panels; change from baseline in trough FEV1, right panels). COPD: chronic obstructive pulmonary disease; mABs: monoclonal antibodies; SND: standard normal deviate.

18

ACCEPTED MANUSCRIPT Table 2. GRADE evidence profile: impact of mAB therapy vs. placebo on the risk of COPD exacerbation and change from baseline in FEV1. Quality assessment

Quality of the

studies

Study design

Risk of bias

Inconsistency Indirectness Imprecision Other considerations

evidence (GRADE)

Outcome: risk of COPD exacerbation

RI PT

№ of

All plausible residual 6

Randomised

Not

trials

serious

Not serious

Not serious

Not serious

confounding would

⨁⨁⨁⨁

reduce the

HIGH

Outcome: change from baseline in FEV1

SC

demonstrated effect

All plausible residual

Randomised

Not

trials

serious

Not serious

Not serious

Serious a

confounding would

M AN U

6

⨁⨁⨁⨁

reduce the

HIGH

demonstrated effect

GRADE Working Group grades of evidence

High quality: we are very confident that the true effect lies close to that of the estimate of the effect

Moderate quality: we are moderately confident in the effect estimate (the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different)

Low quality: our confidence in the effect estimate is limited (the true effect may be substantially different from the estimate of the effect) Very low quality: we have very little confidence in the effect estimate (the true effect is likely to be substantially different from the estimate of effect)

AC C

EP

TE D

a. The lower 95%CI for the effect estimate overlaps the Mean Difference value of 0 (no effect) COPD: chronic obstructive pulmonary disease FEV1: forced expiratory volume in 1 second GRADE: Grading of Recommendations Assessment, Development, and Evaluation mABs: monoclonal antibodies

19

ACCEPTED MANUSCRIPT

4. Discussion The overall results of this quantitative synthesis indicate that mABs provide a weak, although significant, beneficial effect when administered to moderateto-severe

COPD

patients.

In

particular,

the

pairwise

meta-analysis

demonstrates that mAB therapy induced a moderate reduction of the risk of

RI PT

COPD exacerbation, but did not significantly modulate neither the change in FEV1 nor SGRQ, compared to placebo.

However, the analysis of the effect estimates resulting from the analysis of eosinophilic COPD population evidences that, although in a non-significant

SC

manner, mepolizumab induced a numerical and appreciable benefit with regard to the reduction of the risk of COPD exacerbation, by reaching a RR of almost 0.75 [41], and in agreement with the minimal clinical important

M AN U

difference (MCID) suggested for this outcome [18]. Furthermore, when administered in eosinophilic COPD patients, benralizumab elicited a strong signal increase in trough FEV1 of 277 ml at week 56, and extensively fulfilled the suggested MCID for lung function [18]. In eosinophilic COPD patients both mepolizumab and benralizumab had a similar safety profile, that was not

TE D

significantly different compared to placebo.

In order to better characterize the real impact of mAB therapy in the eosinophilic COPD population, we have carried out a network meta-analysis by using the placebo arm as the common intervention among the RCTs

EP

included in this study. Such an analysis permitted to perform an indirect comparison of the investigated treatments that were not studied in a head-to-

AC C

head fashion [42]. To date, only data on benralizumab and mepolizumab are available in eosinophilic COPD patients. Intriguingly, the findings obtained by the network synthesis confirm the beneficial effect of mepolizumab on the risk of exacerbation and of benralizumab on both lung function and health-related QoL. Furthermore, the network approach permits to identify a rank of effectiveness with regard to the impact on the risk of COPD exacerbation (mepolizumab > benralizumab > placebo), change in FEV1 (benralizumab >> mepolizumab ≃ placebo), and SGRQ (benralizumab > mepolizumab > placebo).

20

ACCEPTED MANUSCRIPT Indeed, the treatment with mABs targeting the IL-5 pathway seems to have a positive impact in eosinophilic COPD patients. Surprisingly, we have found that binding the circulating IL-5 with mepolizumab had a protective effect against the risk of COPD exacerbation, whereas blocking the α subunit of the IL-5R with benralizumab had a prevalent effect on lung function.

RI PT

By a pharmacological point of view, these functional and clinical findings are consistent with the strong expression of IL-5R demonstrated in human airway smooth muscle [43]. Such an airway tissue–selective expression may lead to IL-5-induced bronchial hyperresponsiveness independent of eosinophil influx,

SC

and explain the beneficial impact of reslizumab on trough FEV1.

Persistent eosinophilic inflammation appears to be a common feature of some patients with COPD, and may be associated with an increased risk of

M AN U

exacerbation [44]. Therefore, since IL-5 stimulates prevalently the maturation, growth, activation, and survival of human eosinophils and basophils, the inactivation of circulating IL-5 via a selective mAB may explain the protective effect of mepolizumab against COPD exacerbations [43, 45]. As expected, this quantitative synthesis confirms that the effectiveness of

TE D

benralizumab and mepolizumab was greater in eosinophilic patients compared with non-eosinophilic patients [33, 34]. In any case, although targeting IL-5 and IL-5R depletes sputum and blood eosinophil count, the few currently published data provide conflicting results with regard to the direct

EP

association between the reduction of eosinophil count and the real improvements in lung function, risk of exacerbation and health-related QoL in

AC C

COPD patients. This suggests that eosinophils may not have a pivotal pathobiological role in patients with smoking-related COPD, and that eosinophilia may have a less important role in exacerbations of COPD than in asthma [32, 34].

Although the currently available data are scarce, this meta-analysis indicates that ABX-IL8, canakinumab, infliximab, and MEDI8968 have no clinical impact on COPD exacerbations, lung function and health-related QoL in COPD. Among the studies on TNF-α antagonists excluded by the quantitative metaanalysis, but suitable for a qualitative synthesis, one confirmed the lack of effectiveness of infliximab in improving the lung function in patients with mild-

21

ACCEPTED MANUSCRIPT to-moderate COPD, compared with placebo [39]. Another RCT on patients with acute COPD exacerbation failed to demonstrate the superiority of etanarcept with regard to lung function and clinical endpoints, compared with standard treatment with prednisone [38]. Furthermore, an observational study reported that etenarcept, but not infliximab, reduced the rate of COPD

RI PT

hospitalization among patients with COPD receiving TNF-α antagonists to treat their rheumatoid arthritis [37]. The results of these two RCTs [38, 39] confirmed those obtained by our meta-analysis, whereas the conflicting findings resulting from the study of Suissa and colleagues [37] may be

SC

explained by considering the potential biases of observational studies on drug effectiveness. The confounding by indication could have occurred if patients with more severe COPD, and thus at greater risk of being hospitalised for

M AN U

COPD, preferentially received a TNF-α antagonist. Moreover, the confounding by contraindication could also have biased the results of the observational study, since COPD patients with more severe rheumatoid arthritis, who were more likely to receive an anti-TNF-α agent, were less likely to be labelled as COPD when they were hospitalised [37].

TE D

The previously reported lack of effectiveness of the IL-1β antagonist canakinumab, IL-1R1 antagonist MEDI8986, and IL-8 antagonist ABX-IL8 [2] has been confirmed by this meta-analysis, at least with regard to the impact on the risk of exacerbation, lung function and health-related QoL in COPD

EP

patients. The high quality of evidence of this meta-analysis indicates that the overall results on the impact of mABs with regard to the primary endpoints are

AC C

reliable and free from detectable publication bias. Nevertheless, the currently available data on the specific mABs are still limited and further research is needed to support the real benefit induced by the investigated emerging biological therapies for treating COPD. In this regard, the network approach would provide more refined estimates if results on direct comparison between different mABs were available but, unfortunately, no head-to-head RCTs have been carried out. A limitation of this study is related with the fact that the overall meta-analysis included mABs targeting different pathways tested in the overall COPD population. Furthermore, although a subset

analysis was performed in

eosinophilic COPD patients, the data from only one RCT (NCT01227278) 22

ACCEPTED MANUSCRIPT were available for benralizumab, whereas those for mepolizumab were extracted from three RCTs (NCT01463644, NCT02105961, NCT02105948). We cannot exclude that these limitations may have influenced the results of this meta-analysis. Finally, this meta-analysis highlights the strong medical need of data from

RI PT

further large RCTs, especially on IL-5 and IL-5R mABs, that would enrol specific COPD patients, such as those with eosinophilia and COPD. In particular, when results of ongoing phase 3 studies on benralizumab (NCT02138916 and NCT02155660) will be available, this quantitative

SC

synthesis would be updated to further improve the strength of evidence for the

AC C

EP

TE D

M AN U

investigated outcomes.

23

ACCEPTED MANUSCRIPT

5. Conclusions Cytokines and chemokines play a pivotal role in the pathogenesis and development of COPD. Nevertheless, nowadays the mABs directed against these mediators have provided only little or no evidence of therapeutic effectiveness with regard to functional and clinical outcomes of COPD, such

RI PT

as exacerbation rate, lung function and health-related quality of life. Although this scenario confirms the complexity and heterogeneity of COPD in which no dominant cytokine and/or chemokines seems to be involved in the disease, this meta-analysis offers for the first time the quantitative synthesis that

SC

targeting the pathway activated by IL-5 may have a beneficial impact in eosinophilic COPD patients. In fact, mABs acting on IL-5 and IL-5R reduce the differentiation and maturation of eosinophils, as well as their survival in the

M AN U

tissues. This leads to a fast decrease in blood eosinophilia, and reduction of eosinophils in the lung and bone marrow.

Since the umbrella term COPD includes many endotypes characterized by several underlying mechanisms and phenotypes, we cannot rule out that a broad-spectrum anti-inflammatory approach is more likely to be clinically

TE D

effective compared with blocking specific cytokines and chemokines [46]. Nevertheless, our findings provide the evidence that the emerging biological therapies represented by mABs targeting cytokines, chemokines, and their receptors should be investigated in restricted and well defined populations, in

EP

order to provide therapeutic strategies precisely tailored to the specific

AC C

patient's requirements [47].

24

ACCEPTED MANUSCRIPT

6. References

AC C

EP

TE D

M AN U

SC

RI PT

[1] GOLD. Global Initiative for Chronic Obstructive Lung Disease. Global strategy for diagnosis, management, and prevention of COPD – 2018 Report (accessed March 26, 2018). 2018:Available at http://goldcopd.org/wpcontent/uploads/2017/11/GOLD-8-v6.0-FINAL-revised-20-Nov_WMS.pdf. [2] Matera MG, Page C, Rogliani P, Calzetta L, Cazzola M. Therapeutic Monoclonal Antibodies for the Treatment of Chronic Obstructive Pulmonary Disease. Drugs. 2016;76:1257-70. [3] Barnes PJ. Inflammatory mechanisms in patients with chronic obstructive pulmonary disease. J Allergy Clin Immunol. 2016;138:16-27. [4] Caramori G, Adcock IM, Di Stefano A, Chung KF. Cytokine inhibition in the treatment of COPD. Int J Chron Obstruct Pulmon Dis. 2014;9:397-412. [5] Barnes PJ. Development of new drugs for COPD. Curr Med Chem. 2013;20:1531-40. [6] Barnes PJ. Corticosteroid resistance in patients with asthma and chronic obstructive pulmonary disease. J Allergy Clin Immunol. 2013;131:636-45. [7] Rogliani P, Calzetta L, Cazzola M, Matera MG. Drug safety evaluation of roflumilast for the treatment of COPD: a meta-analysis. Expert Opin Drug Saf. 2016:1-14. [8] Cazzola M, Page CP, Calzetta L, Matera MG. Emerging anti-inflammatory strategies for COPD. Eur Respir J. 2012;40:724-41. [9] Fellner RC, Terryah ST, Tarran R. Inhaled protein/peptide-based therapies for respiratory disease. Molecular and cellular pediatrics. 2016;3:16. [10] Chames P, Van Regenmortel M, Weiss E, Baty D. Therapeutic antibodies: successes, limitations and hopes for the future. Br J Pharmacol. 2009;157:22033. [11] Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA Statement. Open Med. 2009;3:e123-30. [12] Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Systematic reviews. 2015;4:1. [13] Rogliani P, Calzetta L, Cavalli F, Matera MG, Cazzola M. Pirfenidone, nintedanib and N-acetylcysteine for the treatment of idiopathic pulmonary fibrosis: A systematic review and meta-analysis. Pulm Pharmacol Ther. 2016;40:95-103. [14] Pedder H, Sarri G, Keeney E, Nunes V, Dias S. Data extraction for complex meta-analysis (DECiMAL) guide. Systematic reviews. 2016;5:212. [15] Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. Available from www.cochrane-handbook.org. The Cochrane Collaboration. 2011. [16] Calzetta L, Rogliani P, Matera MG, Cazzola M. A Systematic Review With Meta-Analysis of Dual Bronchodilation With LAMA/LABA for the Treatment of Stable COPD. Chest. 2016;149:1181-96. [17] Guyatt G, Oxman AD, Akl EA, Kunz R, Vist G, Brozek J, et al. GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin Epidemiol. 2011;64:383-94.

25

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

[18] Calzetta L, Matera MG, Braido F, Contoli M, Corsico A, Di Marco F, et al. Withdrawal of inhaled corticosteroids in COPD: A meta-analysis. Pulm Pharmacol Ther. 2017;45:148-58. [19] Guevara JP, Berlin JA, Wolf FM. Meta-analytic methods for pooling rates when follow-up duration varies: a case study. BMC Med Res Methodol. 2004;4:17. [20] DeCoster J. Meta-analysis notes. http://www.stat-help.com/meta.pdf. Accessed June 14, 2017. 2004. [21] Turner JR, Durham TA. Meta-methodology: conducting and reporting metaanalyses. J Clin Hypertens (Greenwich). 2014;16:91-3. [22] Calzetta L, Rogliani P, Ora J, Puxeddu E, Cazzola M, Gabriella Matera M. LABA/LAMA combination in COPD: a meta-analysis on the duration of treatment. DOI: 10.1183/16000617.0043-2016. Eur Respir Rev. 2016;0:1-11. [23] Lu G, Ades AE. Assessing evidence inconsistency in mixed treatment comparisons. Journal of the American Statistical Association. 2006;101:447-59. [24] Spiegelhalter DJ, Abrams KR, Myles JP. Bayesian approaches to clinical trials and health-care evaluation: John Wiley & Sons; 2004. [25] van Valkenhoef G, Lu G, de Brock B, Hillege H, Ades AE, Welton NJ. Automating network meta-analysis. Research synthesis methods. 2012;3:285-99. [26] Valkenhoef G, Dias S, Ades AE, Welton NJ. Automated generation of node-splitting models for assessment of inconsistency in network meta-analysis. Research synthesis methods. 2016;7:80-93. [27] Dias S, Welton NJ, Sutton AJ, Caldwell DM, Lu G, Ades AE. Evidence synthesis for decision making 4: inconsistency in networks of evidence based on randomized controlled trials. Med Decis Making. 2013;33:641-56. [28] Cazzola M, Calzetta L, Rogliani P, Matera MG. Tiotropium formulations and safety: a network meta-analysis. Doi: 10.1177/2042098616667304. Therapeutic Advances in Drug Safety. 2016:2042098616667304. [29] Wallace BC, Dahabreh IJ, Trikalinos TA, Lau J, Trow P, Schmid CH. Closing the Gap between Methodologists and End-Users: R as a Computational Back-End. Journal of Statistical Software. 2012;49:1-15. [30] Novartis. Safety and Efficacy of Multiple Doses of Canakinumab (ACZ885) in Chronic Obstructive Pulmonary Disease (COPD) Patients. ClinicalTrials.gov Identifier: NCT00581945. Available at https://clinicaltrials.gov/ct2/show/NCT00581945. Accessed August 4, 2017. 2011. [31] Calverley PM, Sethi S, Dawson M, Ward C, Newbold P, Van Der Merwe R. A Phase 2 Study Of Medi8968, An Anti-Interleukin-1 Receptor I (il-1ri) Monoclonal Antibody, In Adults With Moderate-To-Very Severe Chronic Obstructive Pulmonary Disease (COPD). Am J Respir Crit Care Med. 2015;191:A3964. [32] Dasgupta A, Kjarsgaard M, Capaldi D, Radford K, Aleman F, Boylan C, et al. A pilot randomised clinical trial of mepolizumab in COPD with eosinophilic bronchitis. Eur Respir J. 2017;49. [33] Pavord ID, Chanez P, Criner GJ, Kerstjens HAM, Korn S, Lugogo N, et al. Mepolizumab for Eosinophilic Chronic Obstructive Pulmonary Disease. N Engl J Med. 2017. [34] Brightling CE, Bleecker ER, Panettieri RA, Jr., Bafadhel M, She D, Ward CK, et al. Benralizumab for chronic obstructive pulmonary disease and sputum

26

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

eosinophilia: a randomised, double-blind, placebo-controlled, phase 2a study. The lancet Respiratory medicine. 2014;2:891-901. [35] Mahler DA, Huang S, Tabrizi M, Bell GM. Efficacy and safety of a monoclonal antibody recognizing interleukin-8 in COPD: a pilot study. Chest. 2004;126:92634. [36] Rennard SI, Fogarty C, Kelsen S, Long W, Ramsdell J, Allison J, et al. The safety and efficacy of infliximab in moderate to severe chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2007;175:926-34. [37] Suissa S, Ernst P, Hudson M. TNF-alpha antagonists and the prevention of hospitalisation for chronic obstructive pulmonary disease. Pulm Pharmacol Ther. 2008;21:234-8. [38] Aaron SD, Vandemheen KL, Maltais F, Field SK, Sin DD, Bourbeau J, et al. TNFalpha antagonists for acute exacerbations of COPD: a randomised doubleblind controlled trial. Thorax. 2013;68:142-8. [39] van der Vaart H, Koeter GH, Postma DS, Kauffman HF, ten Hacken NH. First study of infliximab treatment in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2005;172:465-9. [40] Carroll J. Genentech snags a PhII-ready IL-33 asthma/COPD drug from Amgen. Available at: http://www.fiercebiotech.com/biotech/genentech-snags-aphii-ready-il-33-asthma-copd-drug-from-amgen. Acceded August 4, 2017. 2016. [41] Guyatt GH, Oxman AD, Kunz R, Brozek J, Alonso-Coello P, Rind D, et al. GRADE guidelines 6. Rating the quality of evidence--imprecision. J Clin Epidemiol. 2011;64:1283-93. [42] Jansen JP, Naci H. Is network meta-analysis as valid as standard pairwise meta-analysis? It all depends on the distribution of effect modifiers. BMC Med. 2013;11:159. [43] Rizzo CA, Yang R, Greenfeder S, Egan RW, Pauwels RA, Hey JA. The IL-5 receptor on human bronchus selectively primes for hyperresponsiveness. J Allergy Clin Immunol. 2002;109:404-9. [44] George L, Brightling CE. Eosinophilic airway inflammation: role in asthma and chronic obstructive pulmonary disease. Ther Adv Chronic Dis. 2016;7:34-51. [45] Flood-Page PT, Menzies-Gow AN, Kay AB, Robinson DS. Eosinophil's role remains uncertain as anti-interleukin-5 only partially depletes numbers in asthmatic airway. Am J Respir Crit Care Med. 2003;167:199-204. [46] Barnes PJ. New anti-inflammatory targets for chronic obstructive pulmonary disease. Nat Rev Drug Discov. 2013;12:543-59. [47] Cazzola M, Calzetta L, Rogliani P, Matera MG. The Challenges of Precision Medicine in COPD. Mol Diagn Ther. 2017;21:345-55.

27