J.C. Cappelleri

The clinical significance of quality of life assessments in oncology: a summary for clinicians

BackgroundA series of six manuscripts with an introduction appeared in the Mayo Clinic Proceedings, based upon the collective effort of 30 individuals with an interest and expertise in assessing the clinical significance of quality of life (QOL) assessments. The series of manuscripts described the state of the science of QOL assessments in oncology clinical research and practice and included extensive literature and theoretical justification for the continued inclusion of QOL in oncology clinical research and practice.ObjectivesThe purpose of this paper is to produce a summary of these articles and to supplement these works with additional information that was gleaned from subsequent meetings and discussions of this material. The primary aim of this paper is to present a cogent and concise description for clinicians to facilitate the incorporation of QOL assessments into oncology clinical research and practice. The theoretical discussion is supplemented with an example of how the various ideas can be operationalized in an oncology clinical trial.

Comparative effectiveness of antibiotics for the treatment of MRSA complicated skin and soft tissue infections

Abstract Objective: With a growing number of studies and comparators in MRSA skin infections, a unified framework for comparing treatments is needed for health technology assessment (HTA). The objective was to systematically assess the success rates of common antimicrobial agents for the treatment of complicated skin and soft tissue infections (cSSTIs) caused by MRSA. Methods: MEDLINE, EMBASE, and Cochrane databases were searched to identify published clinical trials in which dalbavancin, daptomycin, linezolid, telavancin, teicoplanin, tigecycline, and vancomycin were used to treat cSSTIs. Pooled efficacy estimates were generated from clinical and microbiological determinations of success for the MRSA-subgroups in cSSTI clinical trials using a Bayesian meta-analytic approach. Success rates for each antibiotic were reported with 95% Bayesian confidence intervals (called credible intervals [CrI]). In sensitivity analyses the impact of different model parameters and article quality were investigated. Results: Out of 36 identified studies, 14 studies on six antibiotics with 28 treatment arms (nu2009=u20091840) were included in the analysis. No MRSA data in cSSTI were found for teicoplanin. The pooled success rate and CrI95% for each agent was: vancomycin (74.7%; CrI95%: 64.1%–83.5%), dalbavancin (87.7%; CrI95%: 74.6%–95.4%), linezolid (84.4%; CrI95%: 76.6%–90.6%), telavancin (83.5%; CrI95%: 73.6%–90.8%), daptomycin (78.1%; CrI95%: 54.6%–93.2%) and tigecycline (70.4%; CrI95%: 48.0%–87.6%). Comparisons between antibiotics suggested differences versus vancomycin for linezolid (+9.7%; CrI95%: 4.4%–15.8%), dalbavancin (+13.1%; CrI95%: 1.0%–23.8%), and telavancin (+8.8%; CrI95%: 1.5–16.7%). The finding of lower vancomycin efficacy in MRSA cSSTI did not change in sensitivity analyses. Conclusion: The results of this meta-analysis suggest higher success rates for linezolid and the new glycopeptides (dalbavancin and telavancin) in MRSA-confirmed cSSTIs. The uncertainty margins reflect the study limitations including number of cases and indirect nature of the comparisons. This example of Bayesian meta-analysis for MRSA cSSTI provides a potential framework for comparisons that is useful for HTA and formulary decision-making.

THU0447 Treatment Adherence, Persistence, and Switch to Generic Nsaids in Patients with Osteoarthritis (OA) Prescribed Celecoxib: Us Retrospective Claims Database Analysis

Background OA is the most common and costly chronic musculoskeletal conditions. Celebrex® (celecoxib), which will become generic in the US in 2015, is a COX-2 selective inhibitor for pain relief for OA. Literature to date does not address rates of treatment switch from celecoxib to generic NSAIDs. Objectives The study aimed to measure celecoxib adherence, persistence, and rates of switch to generic NSAIDs and understand comparative rates of treatment-emergent events for patients who switched from celecoxib to generic NSAIDs vs persistent patients; also compared were before- versus after-switch among switched patients. Methods The retrospective claims database analysis used MarketScan® (2009–2013) to extract a cohort of adult OA patients (ICD-9-CM: 715.xx) who were new users of celecoxib prescription. Included patients had 12 months continuous enrollment before (pre-index period, without celecoxib) and ≥6 months after the index date with ≥2 celecoxib claims post-index. Adherence was measured by medication possession ratio (MPR). Persistence was defined as time to the first prescription gap of ≥30 days. Treatment switch required a fill for generic NSAIDs within 30 days of the end of celecoxib day supply. The likelihood of developing gastrointestinal (GI), cardiovascular (CV), renal and neuro-muscular (rheumatoid arthritis, ankylosing spondylitis, low back and other back/neck pain, fibromyalgia, arthritis/other arthropathies, diabetic/peripheral neuropathy) events while taking celecoxib was compared between switched and persistent celecoxib users in propensity score matched cohorts. Similar methods were used to compare outcome events before versus after switch among the switched patients. Cox proportional hazard models were employed to identify characteristics associated with time-to-switch. Results Of 65,530 included patients (mean ± SD: age 61±11.9; 62.5% female), 6,783 (10.4%) were persistent on celecoxib after 13.6±7.7 months of follow-up (maximum 44.4 months). The majority (54,554, 83.3%) discontinued celecoxib without switching and 3,475 (5.3%) switched to generic NSAIDs (median time to switch or discontinuation 2.96 months). Persistent users were more adherent (MPR≥0.8) than switched (92.9% vs 78.5%, p<.001). Treatment-emergent events were higher for switched than persistent patients while taking celecoxib (Figure 1a). Among switched patients, post-switch incidence of treatment-emergent events were higher than pre-switch (Figure 1b). More likely to switch sooner were younger patients from Northeast region with more pre-index inpatient days, pre-index pain medications use, less adherence to celecoxib (MPR<0.8), no pre-switch bleeding, myocardial infarction or renal disorders, and no treatment-emergent GI, CV or neuro-muscular events. Conclusions Within the cohort, relatively few patients switched from celecoxib to generic NSAIDs, although half discontinued in 3 months of celecoxib initiation. There was a higher incidence of GI, CV, renal and neuro-muscular events for switched than persistent patients. Among switched patients, higher incidence of post-switch vs pre-switch events suggested that these events were more likely to occur post-celecoxib when taking generic NSAIDs. Disclosure of Interest X. Ji Grant/research support from: Pfizer Inc., Employee of: Pharmerit International, J. Cappelleri Shareholder of: Pfizer Inc., Employee of: Pfizer Inc., C. Solem Grant/research support from: Pfizer Inc., Employee of: Pharmerit International, S. Liu Grant/research support from: Pfizer Inc., Employee of: Pharmerit International, A. Shelbaya Shareholder of: Pfizer Inc., Employee of: Pfizer Inc., C. Walker Shareholder of: Pfizer Inc., Employee of: Pfizer Inc.

Assessment of pruritus in atopic dermatitis: validation of the Severity of Pruritus Scale (SPS)

Introduction: Pruritus, or itch, is a central feature of atopic dermatitis (AD) and is often cited as the most bothersome symptom. The Severity of Pruritus Scale (SPS) is a 4-point rating scale adapted from the Atopic Dermatitis Severity Index to provide a measure of pruritus severity within a 24-hour recall period. The objective of this analysis was to assess the qualitative and quantitative validity of the SPS in AD. Methods: Content validity of the SPS was evaluated qualitatively in US patients with AD. Psychometric properties of the SPS were evaluated using data from 2 phase 3 trials conducted to investigate crisaborole in mild to moderate AD (study AD-301: NCT02118766; study AD-302: NCT02118792). Results: Fourteen patients were included in the qualitative analysis, considered adequate because of the single-item, single-concept nature of the SPS. Itch was the most prevalent symptom, and the SPS was easily understood and completed in US English (n=9) and US Spanish (n=5). The psychometric analysis used data from 1344 patients. Test-retest reliability analysis identified an intraclass correlation coefficient of 0.54 for a single SPS observation, and at least 0.70 when at least 2 SPS observations were averaged. SPS scores correlated with Investigator’s Static Global Assessment (ISGA) scores, measures of quality of life, and 4 of 5 signs of AD (Pearson correlations, ≥0.40 at day 29). The difference in score between the no disease group/clear (ISGA 0) and the severe disease group (ISGA 4) was 0.80. The clinically important difference was estimated to be 0.20 and the clinically important response was estimated to be at least a 0.19-point reduction from baseline. Discussion: The SPS is an appropriate tool to assess itch in patients with AD and is a valid and reliable measure of pruritus severity.

Effect of crisaborole topical ointment, 2%, on atopic dermatitis–associated pruritus: an extended analysis of 2 phase 3 clinical trials

Introduction: Pruritus is an essential feature of atopic dermatitis (AD) and is widely considered the most distressing symptom. Crisaborole ointment is a nonsteroidal phosphodiesterase 4 inhibitor for the treatment of mild to moderate AD. The efficacy of crisaborole for AD-associated pruritus was assessed in 2 phase 3 trials using the Severity of Pruritus Scale (SPS). Post hoc validation of the SPS identified that 1 SPS observation provided inadequate test-retest reliability. Therefore, extended analyses were conducted using at least 2 SPS observations for robust assessment of pruritus in the phase 3 crisaborole trials. Methods: Data were analyzed from 2 identically designed, vehicle-controlled, double-blind, phase 3 trials designed to investigate the efficacy and safety of crisaborole in AD patients aged 2 years and above (AD-301: NCT02118766; AD-302: NCT02118792). At least 2 SPS observations were averaged for acceptable test-retest reliability. Results: At least 2 baseline observations were available for 569 patients in AD-301 and 561 patients in AD-302. Median time to pruritus improvement (SPS score ⩽1 with at least 1-point improvement from baseline) was shorter with crisaborole than with vehicle (AD-301: 5.0 vs. 10.0u2009d, P=0.0003; AD-302: 6.0 vs. 9.0u2009d, P=0.0087). At week 4, more crisaborole-treated patients than vehicle-treated patients experienced pruritus improvement (AD-301: 37% vs. 21%, P<0.0001; AD-302: 34% vs. 21%, P=0.0006), mean pruritus scores were lower with crisaborole than with vehicle (AD-301: 0.97 vs. 1.28, P<0.0001; AD-302: 1.08 vs. 1.35, P<0.0001), and more crisaborole-treated patients than vehicle-treated patients experienced a clinically important pruritus response (AD-301: 75% vs. 57%, P<0.0001; AD-302: 72% vs. 64%, P=0.0828). Conclusions: These extended analyses show that patients treated with crisaborole experienced rapid and clinically relevant improvement in AD-associated pruritus.

THU0300 Network meta-analysis of tofacitinib vs bdmards or apremilast for the treatment of tnf inhibitor-naÏve patients with psoriatic arthritis

Background Tofacitinib is an oral JAK inhibitor for the treatment of psoriatic arthritis (PsA). Objectives To perform a systematic literature review (SLR) and network meta-analysis (NMA) to evaluate the efficacy of tofacitinib 5 and 10u2009mg BID relative to biologic disease-modifying antirheumatic drugs (bDMARDs) or a targeted synthetic DMARD (apremilast [APR]) in tumour necrosis factor inhibitor-naïve (TNFi-N) patients with active PsA. Methods The SLR identified randomised controlled clinical trials (RCTs) evaluating tofacitinib, bDMARDs or APR to treat patients with active PsA who were TNFi-N. Outcomes included American College of Rheumatology (ACR)20 response and change from baseline in Health Assessment Questionnaire-Disability Index (ΔHAQ-DI), Dactylitis Severity Score (ΔDSS) and Leeds Enthesitis Index (ΔLEI). Treatment effects were only evaluated during placebo (PBO)-controlled trial phases. The Bayesian NMA (with non-informative priors) was conducted using WinBUGs. The binomial logit model was used for ACR20. ΔHAQ-DI, ΔDSS and ΔLEI were analysed using the normal identify link model. A fixed-effect model was fitted to the data. Median treatment rankings represent data from each iteration of the model from which inferences are based, following model convergence. Results The SLR identified 25 RCTs and 21 were included in the NMA (see treatments in table 1). All trials allowed methotrexate use. PBO-controlled treatment durations ranged from 12–24 weeks. In general, patient characteristics were similar across trials. All treatments were associated with improvements in ACR20 and ΔHAQ-DI vs PBO. Tofacitinib 5u2009mg BID was associated with substantially decreased odds ratios (ORs) for ACR20 vs golimumab 50 and 100u2009mg Q4W, etanercept 25u2009mg BW, infliximab 5u2009mg/kg and secukinumab 150u2009mg QW-Q4W (table 1); ORs for all remaining comparators were not substantially different. Tofacitinib 10u2009mg BID was associated with a substantially increased OR for ACR20 vs APR 20u2009mg BID. Etanercept was associated with an improvement in ΔHAQ-DI vs tofacitinib 5 and 10u2009mg BID. There was no difference in ΔHAQ-DI for tofacitinib vs other bDMARDs. For ACR20, tofacitinib 5 and 10u2009mg BID were median ranked 14 (95% credible interval: 8, 17) and 9,5, 14 respectively, among 18 comparators. For ΔHAQ-DI, tofacitinib 5 and 10u2009mg BID were median ranked 114, 13 and 8,2, 13 respectively, among 14 comparators. Two studies evaluated ΔDSS and ΔLEI; there were no substantial differences in ΔDSS and ΔLEI for tofacitinib 5 and 10u2009mg BID vs adalimumab 40u2009mg Q2W and ixekizumab 80u2009mg Q2W and Q4W.Abstract THU0300 – Table 1 Fixed-effect NMA data for ACR20 and ΔHAQ-DI in TNFi-N patients with PsA – comparison of bDMARDs or apremilast vs tofacitinib 5 and 10 mg BIDa,b Conclusions Based on the NMA of published RCTs in TNFi-N patients with PsA, tofacitinib 5 and 10u2009mg BID had similar efficacy vs many, but not all, bDMARDs and APR in improving ACR20 and ΔHAQ-DI. Acknowledgements Study sponsored by Pfizer Inc. Medical writing support was provided by P Scutt of CMC and funded by Pfizer Inc. Disclosure of Interest D. Gladman Grant/research support from: AbbVie, Amgen, Bristol-Myers Squibb, Celgene, Eli Lilly, Janssen, Novartis, Pfizer Inc, UCB, Consultant for: AbbVie, Amgen, Bristol-Myers Squibb, Celgene, Eli Lilly, Janssen, Novartis, Pfizer Inc, UCB, A.-M. Orbai Grant/research support from: Celgene, Eli Lilly, Horizon, Janssen, Novartis, Pfizer Inc, UCB, Consultant for: Eli Lilly, Janssen, Novartis, Pfizer Inc, UCB, J. Gomez-Reino Grant/research support from: AbbVie, MSD, Pfizer Inc, Roche, Consultant for: Pfizer Inc, Speakers bureau: AbbVie, Biogen, Bristol-Myers Squibb, Janssen, MSD, Pfizer Inc and Roche, S. Chang-Douglass Employee of: Decision Resources Group, E. Leoncini Employee of: Decision Resources Group, H. Burton Employee of: Decision Resources Group, K. Kanik Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, T. Hendrikx Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, J. Cappelleri Shareholder of: Pfizer Inc, Employee of: Pfizer Inc, M.-A. Hsu Shareholder of: Pfizer Inc, Employee of: Pfizer Inc

AB0742 Evolution of Biologic Treatment, Satisfaction with Control of Disease, and Patient-Reported Outcomes in European Patients with Psoriatic Arthritis: A Real-World Study

Background Patients with psoriatic arthritis (PsA) commonly receive therapy with conventional synthetic disease-modifying antirheumatic drugs (csDMARDs), either alone or in combination with biologics such as tumour necrosis factor inhibitors.1 Objectives To assess the evolution of real-world biologic treatment patterns in patients with PsA by comparing patient data from two distinct patient surveys from 2014 and 2011. Methods This was a retrospective, cross-sectional analysis of data from two independent surveys (Adelphi Real World PsA Disease Specific Programmes) conducted in Q1 2011 and Q3 2014. Rheumatologists provided patient demographics, clinical details, treatment history and levels of satisfaction with disease control. Patient-reported outcomes included satisfaction with disease control, EuroQol-5 Dimension, Work Productivity and Activity Impairment (WPAI) and Alternative Health Assessment Questionnaire Disability Index (without aids/devices, HAQ-DI). Statistical differences were assessed using Mann-Whitney-U tests and Fishers exact tests for quantitative and categorical outcomes, respectively. Missing data were not imputed. Results Patients with PsA receiving biologic treatment were similar in age, sex and body mass index in 2014 (n=325) and 2011 (n=326). The two independent cohorts showed no difference in time since diagnosis, severity of PsA prior to initiation of current treatment, disease progression prior to initiation of current treatment and time on current biologic treatment. The 2014 cohort had milder PsA vs the 2011 cohort (mild/moderate/severe: 72.6%/24.9%/2.5% vs 59.8%/35.0%/5.2%, respectively; P=0.0004), fewer patients who had ever experienced an acute episode (36.4% vs 56.6%, respectively; P<0.0001) or experienced an acute episode in the last 12 months (21.9% vs 31.4%, respectively; P=0.0074) and more patients with better current disease status (improving/stable/deteriorating slowly/deteriorating rapidly/unstable: 31.4%/56.7%/1.6%/7.2%/3.1% vs 27.3%/52.2%/14.4%/3.1%/3.1%, respectively; P=0.045). Patients did not receive biologic treatment any earlier in 2014 (median [interquartile range]; 36 [12, 72] months in 2014 vs 33 [15, 63] in 2011, P=0.2261). The number of csDMARDs received prior to initiation of biologic treatment did not differ significantly in 2014 vs 2011 (P=0.1551). Both physicians and patients were more satisfied with control of PsA in 2014 vs 2011 (81.5% vs 71.8%, P=0.0016 and 90.9% vs 67.3%, P=0.0002, respectively). The HAQ-DI score was significantly lower, indicating better health, in 2014 vs 2011 (0.38 [0.00, 0.88] vs 0.63 [0.27, 1.00], respectively; P=0.0065). WPAI scores showed nonsignificant improvements in outcomes in 2014 compared with 2011. Conclusions There were no significant differences in severity of PsA at initiation of treatment or treatment patterns in the patient samples analysed in 2014 vs 2011; however, patients disease status was improved and both patients and physicians were more satisfied with the control of PsA in 2014 than in 2011. Further investigation is required to elucidate the reasons behind the improved outcomes observed over time. References Soriano ER, et al. Best Pract Res Clin Rheumatol. 2014;28:793–806. Disclosure of Interest J. Lucas Employee of: Adelphi Real World, R. Wood Employee of: Adelphi Real World, J. Piercy Employee of: Adelphi Real World, J. C. Cappelleri Shareholder of: Pfizer, Employee of: Pfizer, B. Kola Shareholder of: Pfizer, Employee of: Pfizer, A. Chhabra Shareholder of: Pfizer, Employee of: Pfizer, R. Vasilescu Shareholder of: Pfizer, Employee of: Pfizer, M.-A. Hsu Shareholder of: Pfizer, Employee of: Pfizer