Certolizumab pegol and secukinumab for treating active psoriatic arthritis following inadequate response to disease-modifying antirheumatic drugs: a systematic review and economic evaluation

Article history

The research reported in this issue of the journal was commissioned and funded by the HTA programme on behalf of NICE as project number 15/06/04. The protocol was agreed in January 2016. The assessment report began editorial review in September 2016 and was accepted for publication in March 2017. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

none

Copyright statement

© Queen’s Printer and Controller of HMSO 2017. This work was produced by Corbett et al. under the terms of a commissioning contract issued by the Secretary of State for Health. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

Description of health problem

Psoriatic arthritis (PsA) is a chronic autoimmune disease closely associated with psoriasis of the skin and nails, but distinct from rheumatoid arthritis (RA). PsA is one of a family of inflammatory arthritis disorders called spondyloarthritis (or spondyloarthropathy), which also includes ankylosing spondylitis. 1 PsA is closely linked with inflammatory bowel disease, especially the form called Crohn’s disease. 2 Although any joint may be affected, PsA typically affects joints in the hands, feet and spine. Its course may be erratic, with flare-ups and remissions, but it can cause joint damage if it is not treated. Early diagnosis is important to avoid damage to joints. 3 Symptoms of arthritis include inflamed (swollen), stiff and painful joints; and psoriasis symptoms include patchy, raised red areas of inflamed skin with scaling. 4

The symptoms of psoriatic arthritis are similar to those of other forms of arthritis. The difference between PsA and RA is that the pattern of joint involvement is commonly asymmetrical, and involves the distal interphalangeal joints (in the hands and feet) and nail lesions. The following terms are used to present the patterns of PsA: oligoarthritis (four or fewer affected joints; 22–37% of patients); polyarthritis (five or more affected joints; 36–41% of patients); arthritis of distal interphalangeal joints (< 20% of patients); spondylitis (7–23% of patients); and arthritis mutilans (approximately 4% of patients). 5,6 Most patients with PsA will have developed psoriasis first (i.e. joint complications occur around 10 years after initial diagnosis of psoriasis), although joint involvement appears first in 19% of patients and concurrently with psoriasis in 16% of cases. 7

As PsA can affect both skin and joints, it can result in significant quality-of-life impairment, joint deformity and psychosocial disability. 7,8 A recent survey of patients with RA, PsA and axial spondyloarthritis found that disease burden in terms of patient-reported outcome measures was similar in PsA and axial spondyloarthritis patients, but significantly lower for the RA patients. 9 The physical and psychosocial problems experienced by patients affect their ability to perform paid work and everyday tasks; PsA has a substantial economic impact on the UK health-care system as a result of direct health-care costs as well as indirect costs, such as reduced work capacity. 10

Patients with PsA have a 60% higher risk of premature mortality than the general population, with cardiovascular disease being the leading cause of death. 11–13 The estimated reduction in life expectancy for patients with PsA is approximately 3 years,14 with a standardised mortality ratio of 1.62. A Canadian outpatient clinic study reported that mortality due to cardiovascular disease was 30% higher in patients with PsA than that in the general population. 12

Diagnosis

It is difficult to define PsA because there are no precise diagnostic criteria or diagnostic markers. 15 In general, diagnoses are primarily based on patient symptoms and physical examination. In most cases, Moll and Wright’s 1973 criteria16 have been used for diagnosis. Several classification criteria have have been introduced since Moll and Wright’s criteria, but none has been widely accepted or validated. In 2006, the multicentre Classification Criteria for Psoriatic Arthritis (CASPAR) study developed new classification criteria that are simple and have both a high sensitivity and a high specificity; they are currently a preferred method to define cases of PsA (Table 1). 17

Epidemiology

The exact prevalence of PsA is unknown, but estimates vary from 0.3% to 1% of the population. It has been estimated that in England, in 2013, there were around 53,900–161,600 people with PsA. PsA affects men and women equally, in contrast to RA, which is more common in women. 18

Psoriatic arthritis can develop at any time, including childhood,19 but normally it appears between the ages of 30 and 55 years. 18 Its development is a complex process involving both environmental and genetic factors. 20–22 Studies show a stronger genetic or family link to PsA than to other autoimmune rheumatic diseases. Around 40% of people who are diagnosed with PsA and psoriasis also have family members affected by the disease. 2

Measurement of disease

In 2016, the Group for Research and Assessment of Psoriasis and the Psoriatic Arthritis Outcome Measures in Rheumatology Organisation PsA working group updated the core set of domains to be assessed in clinical trials to reflect both patient and physician priorities. The domain set includes musculoskeletal disease activity (which now includes enthesitis, dactylitis and spine symptoms, in addition to peripheral arthritis), skin disease activity, patient global assessment, pain, physical function, health-related quality of life (HRQoL), fatigue and systemic inflammation. Four new items were added to the research agenda: stiffness, independence, treatment burden and sleep. 23

Many trials of PsA have used 20% improvement in the American College of Rheumatology criteria (ACR 20) as the primary outcome;24 the American College of Rheumatology (ACR) criteria were, however, developed to assess RA. The other outcome assessment tools that have commonly been used in clinical trials are:

• the Psoriatic Arthritis Response Criteria (PsARC), a multidomain measure which has similarities with ACR criteria but which was developed specifically for PsA

• the Psoriasis Area and Severity Index (PASI), to assess psoriasis

• the Health Assessment Questionnaire-Disability Index (HAQ-DI), to assess function (activities of daily living)

• various measures of enthesitis, dactylitis and radiographic progression of disease.

However, there are issues with some assessment tools:

• HAQ-DI concentrates on physical disability, which may not adequately capture disability in patients with predominantly skin disease. Consequently, there is less change in the context of treatment that has a predominant effect on the skin and not the joints. 25

• PASI has poor sensitivity to change and responsiveness when skin psoriasis is < 10% of body surface area (BSA) involvement. Furthermore, it has been stated that the correlation with quality-of-life measures is poor. 26 In addition, it is time-consuming and not practically very feasible in daily clinical practice.

• PsARC identifies only relative changes from baseline and overestimates the number of responders. 27 In general, PsARC placebo response rates are higher than other composite measures. 28

Current service provision

If PsA is not treated early, the inflammation can affect the whole body, which may lead to lasting joint and tissue damage. 2 The clinical management of PsA therefore aims to suppress joint, tendon and ligament inflammation, and to manage the skin symptoms of the disease. Current practice involves early diagnosis and early use of non-steroidal anti-inflammatory drugs (NSAIDs) and/or intra-articular corticosteroid injections. In patients who do not respond to these treatments, disease-modifying antirheumatic drugs (DMARDs) are then used [most commonly beginning with methotrexate (MTX)]. When conventional disease-modifying antirheumatic drugs (cDMARDs) are ineffective, or not tolerated, biologic therapies may be used; for example, anti-tumour necrosis factor (TNF) therapies, such as etanercept [(ETN) ENBREL^®; Pfizer Inc., New York City, NY, USA], infliximab [(INF) REMICADE^®; Merck Sharp & Dohme, Kenilworth, NJ, USA], adalimumab [(ADA) HUMIRA^®; AbbVie Inc., North Chicago, IL, USA] and golimumab [(GOL) SIMPONI^®; Merck Sharp & Dohme, Kenilworth, NJ, USA]. These anti-TNFs are approved by the National Institute for Health and Care Excellence (NICE). Anti-TNFs have been shown to slow the progression of joint damage when assessed radiographically. 29,30 Ustekinumab [(UST) STELARA^®; Janssen Pharmaceuticals, Inc., Horsham, PA, USA] – a different type of biologic therapy to anti-TNFs [being an interleukin (IL)-12/23 inhibitor] – is also recommended as a possible treatment, specifically when DMARDs have not worked well enough, provided that treatment with anti-TNFs is not suitable, or the patient has had an anti-TNF before. Apremilast [(APR) Otezla^®; Celgene Corporation, Summit, NJ, USA], a phosphodiesterase 4 inhibitor, is not currently approved by NICE.

Current NICE guidance relates to the treatment of patients who have had an inadequate response to two or more cDMARDs (administered either individually or in combination). The British Society for Rheumatology (BSR)’s guidelines make a provision for using a biologic after one DMARD in the presence of adverse prognostic factors; these are defined as five or more swollen joints in association with an elevated C-reactive protein (CRP) concentration for more than 3 months and structural joint damage due to disease. 31 Not all patients respond to an initial anti-TNF treatment, and in some patients the response diminishes over time. One observational study showed that one-third of PsA patients had switched to a second anti-TNF because of a lack of efficacy and side effects. 32 NICE does not specifically recommend switching anti-TNFs other than the guidance for UST, and switching decisions may depend on local Clinical Commissioning Group guidelines: in some parts of the country patients are allowed to switch from one anti-TNF to another.

Quite often PsA goes undetected and is sometimes not recognised and diagnosed by dermatologists or general practitioners (GPs). In the UK, rheumatologists manage the majority of patients with PsA, but patients with less severe joint disease may be managed by a dermatologist. However, patients with severe problems with joints and skin will tend to be managed by both rheumatologists and dermatologists.

Description of the technology under assessment

Certolizumab pegol (CZP; CIMZIA^®, UCB Pharma, Brussels, Belgium) is a biologic therapy (a monoclonal antibody that targets TNF) that is administered subcutaneously. Anti-TNFs target the activation of tumour necrosis factor alpha (TNF-α) and subsequently activation of downstream inflammatory processes, and as such have the potential to offer symptom control as well as altering disease progression. CZP in combination with MTX has a marketing authorisation in the UK for treating active PsA in adults when the response to previous DMARD therapy has been inadequate. CZP can be given as monotherapy if MTX cannot be tolerated or when continued treatment with MTX is inappropriate.

Secukinumab (SEC; COSENTYX^®, Novartis International AG, Basel, Switzerland), which is also administered subcutaneously, is a different type of biologic therapy to CZP, being a monoclonal antibody that targets the IL-17A cytokine molecule (rather than targeting TNF). SEC, alone or in combination with MTX, is indicated for the treatment of active PsA in adult patients when the response to previous DMARD therapy has been inadequate. SEC also has marketing authorisation from the European Medicines Agency for the treatment of ankylosing spondylitis and moderate–severe plaque psoriasis.

Previous National Institute for Health and Care Excellence appraisals

There have been no previous NICE technology appraisals (TAs) of CZP or SEC for PsA, although there have been several appraisals of other biologics for PsA: TA19933 (ETN, INF and ADA), TA22034 (GOL) and TA34035 (UST). APR, which is not a biologic, is not currently recommended by NICE.

A number of key areas of uncertainty and potential limitations of the evidence base were identified from the previous appraisals. These include:

• a lack of direct head-to-head trial evidence evaluating the relative efficacy and safety of the biologics

• some limitations in the external validity of the trial populations (i.e. the trial populations had some differences from populations seen in routine clinical practice)

• a lack of patient registry data for PsA

• the long-term effectiveness of biologics in controlling disease activity

• the prescription cost of biologics and the cost of treating psoriasis at different levels of severity

• the progression of (HAQ-DI) score (a measure of patient function) in patients on and off treatment, and the length of time biologics are assumed to be effective

• long-term progression of PsA with and without biologics

• a lack of an optimal outcome measure for PsA

• the rate of treatment withdrawal and the adverse effects associated with the long-term use of biologics

• a lack of evidence on the efficacy and safety of the sequential use of biologics.

This assessment has considered and attempted to address these limitations and areas of uncertainty using relevant evidence where available.

Overall objective of assessment

To determine the clinical effectiveness and cost-effectiveness within the NHS of CZP and SEC within their marketing authorisations for treating active PsA in adults for whom DMARDs have been inadequately effective.

Methods for reviewing clinical effectiveness

Quantity and quality of the identified evidence

A total of 1761 records were retrieved from the original December 2015 electronic database searches. The searches were updated on 28 April 2016, with a further 200 records available for screening. After screening titles and abstracts, full copies of 182 papers were assessed for inclusion in the review.

Two RCTs were excluded at the abstract stage for using unlicensed dosages (50 mg of ETN twice weekly,42 and 20 and 40 mg of APR43). Two RCTs were excluded at the full-paper stage: one did not report subgroup results for PsA44 and the other included only patients who were naive to MTX. 45 The FUTURE [Efficacy at 24 Weeks and Long Term Safety, Tolerability and Efficacy up to 2 Years of Secukinumab (AIN457) in Patients With Active PsA] 1 trial of SEC was excluded from the RCT short-term efficacy review as it used an unlicensed, very high, loading dose. It was, however, included as an open-label extension study as the impact of the initial high loading dose would probably be negligible at later time points. 46 Fifty open-label studies of comparator treatments were excluded as they did not relate to the latest (longest) duration of follow-up.

Details of the numbers of other eligible full publications or conference abstracts that relate to open-label studies of the included RCTs and patient registry or safety studies are presented in Figure 1.

FIGURE 1.

Flow chart showing the number of studies identified and eligible for inclusion. OL, open label.

Characteristics of the randomised controlled trials included in the systematic review of short-term efficacy

Of the 19 included RCTs, 17 were placebo controlled: one of CZP,47 three of SEC (two of which were reported in one publication),48,49 one of GOL,50 two of INF,51,52 two of ETN,53,54 three of ADA,55–57 two of UST58,59 and three of APR. 60,61 The FUTURE 1 trial of SEC was excluded from the RCT short-term efficacy review as it used an unlicensed, and very high, loading dose. 46

Two trials compared active treatments: one compared SEC with UST62,63 and one compared INF, ETN and ADA. 64

Most studies were conducted mainly in Europe and North America. All but two53,64 were multicentre trials. Details of the trial durations, different phases and the dosing regimens of the main interventions studied are presented in Table 2. Details of all interventions studied are presented in Table 3. For some trials we excluded individual treatment arms from the systematic review (see Table 3). This was as a result of the doses not being licensed or recommended in the populations studied. Some included trials were excluded from the NMAs because of the populations being different from the other trial populations (see Table 3).

The design of many trials typically included a fully blinded, placebo-controlled phase followed by an ‘early escape’ crossover phase (from placebo to an active treatment) for non-responders, then finally crossover to active treatment for the remaining placebo participants. Non-response in this context related to failure to achieve prespecified minimum improvements (ranging between 5% and 20%) in tender joint count (TJC) and swollen joint count (SJC). All the trials using an early escape design ran for 16 weeks before patients were eligible for early escape. Trials then entered open-label extension phases (see Long-term effectiveness).

Table 4 describes the population characteristics of the included trials. Where available, this includes subgroup characteristics for patients who had never previously taken a biologic (i.e. biologic-naive populations) and patients who had previously taken a biologic (i.e. biologic-experienced populations). Biologic-experienced patients were available only for the more recent trials (those of SEC, CZP, UST and APR); in the earlier trials such patients were not eligible to participate. Trial sample sizes varied, with earlier trials tending to be smaller than more recent trials. Variation in sample size was also evident within treatments: the two trials of ETN had populations of 60 and 205,53,54 and the three trials of ADA had populations of 100, 207 and 315. 55–57,67 The duration of PsA ranged from 3 to 12 years across trials; the shortest durations (reported as medians) came from the UST PSUMMIT trials58,59,66 and the longest (reported as means) came from the Infliximab Multinational Psoriatic Arthritis Controlled Trial (IMPACT). 51,52 The duration of psoriasis ranged from 11 to 23 years, although this information was not available for the FUTURE 248 SEC and RAPID-PsA47 (Certolizumab Pegol in Subjects With Adult Onset Active and Progressive Psoriatic Arthritis) CZP trials. Although not reported in all trials, baseline CRP concentration levels were difficult to interpret as they appeared to have slightly skewed distributions, with means (range 10–31 mg/l) being generally higher than medians (range 7–15 mg/l).

Notwithstanding this limited heterogeneity, many key patient characteristics were broadly similar across trials, including mean ages (which ranged from 45 to 51 years), the proportion of male participants (around 50% for most trials), and TJCs and SJCs (TJC, range 18–29; SJC, range 9–18); an exception was the three-arm head-to-head trial, which had notably lower TJC and SJC. 64 The population in this trial, along with the PsA populations from the large SEC psoriasis trials,49 also had markedly higher baseline PASI scores than the other trials (typically around two to three times higher). The FUTURE 2 SEC trial had slightly higher baseline PASI scores than the other trials, most notably in the 150 mg treatment arm. The PsA populations from two of the SEC psoriasis trials49 also had lower baseline HAQ-DI scores (range 0.5–0.8 units) than the other trials (range 0.9–1.6 units). In light of these differences, the characteristics of the PsA patients in the SEC psoriasis trials were deemed to be too dissimilar to the other trials to be included in the NMAs. There were three of these psoriasis trials: Efficacy of Response and Safety of Two Fixed Secukinumab Regimens in Psoriasis (ERASURE), Full Year Investigative Examination of Secukinumab vs. Etanercept Using Two Dosing Regimens to Determine Efficacy in Psoriasis (FIXTURE) and Efficacy of Secukinumab Compared to Ustekinumab in Patients with Plaque-type Psoriasis (CLEAR; baseline data were not available for the PsA patients in CLEAR). To be eligible for the ERASURE, FIXTURE and CLEAR trials, patients had to have moderate–severe psoriasis based on a PASI score of > 12 units and BSA involvement of ≥ 10%. 49 In the trials only of patients with PsA, the proportion of patients with at least moderate psoriasis (i.e. PASI-evaluable patients, defined as a BSA involvement of ≥ 3%) ranged between 41% and 87%.

In the FUTURE 2 (SEC)48 and RAPID-PsA (CZP)47 trials, the biologic-experienced and biologic-naive subgroups were broadly similar except that the biologic-experienced subgroups tended to have slightly higher TJCs and SJCs, and slightly longer durations of PsA.

All the trials of ETN, INF, ADA and GOL and one UST trial58 (nine in total) excluded patients who had previously received an anti-TNF, so their populations comprised entirely biologic-naive patients (Table 5). In the remaining trials, where reported, the proportion of biologic-experienced patients ranged from 15% to 58%. Of the trials that allowed recruitment of biologic-experienced patients, the RAPID-PsA trial47 was more selective than the FUTURE 2,48 PSUMMIT 259,66 and PALACE trials. 60,61,65 RAPID-PsA47 was the only trial that excluded patients with primary failure of a previous anti-TNF (primary failure was defined as no response within the first 12 weeks of treatment with the anti-TNF). (See Appendix 2, which details the eligibility criteria for all trials.) The results for the RAPID-PsA biologic-experienced subgroup may therefore be somewhat inflated when compared with the other trials reporting results for this subgroup.

ERASURE and FIXTURE trials

As the focus of the ERASURE and FIXTURE trials49 was on patient populations with psoriasis (subgroups of which also had PsA), fewer outcomes that were relevant to this assessment were evaluated. Patients recruited into in the ERASURE and FIXTURE trials had more severe psoriasis but lower baseline HAQ-DI scores than the patients recruited into the FUTURE 2 trial and into the other trials included in the systematic review (see Table 4). The FIXTURE trial was one of the very few identified in the systematic review that compared different biologics (SEC with ETN).

Table 16 and Figure 2 (in which data from the ERASURE and FIXTURE trials have been pooled) illustrate SEC’s superiority over placebo for the PASI outcomes. In the FIXTURE trial at 12 weeks, 300 mg of SEC was statistically significantly more effective than 50 mg of ETN twice weekly in terms of patients achieving a PASI 75 response (RR 1.86, 95% CI 1.24 to 2.81) and a PASI 90 response (RR 2.42, 95% CI 1.20 to 4.88). Changes from baseline in the HAQ-DI scores were greater in SEC- and ETN-treated patients in ERASURE and FIXTURE trials than with placebo.

TABLE 16 - Efficacy outcomes in the ERASURE and FIXTURE trials at 12 weeks49

Standard errors not reported.

Trial	Treatment	Number of PsA patients	PASI 50	PASI 75	PASI 90	HAQ-DI change from baselinea
ERASURE49	300 mg of SEC	57	–	38 (67%)	30 (53%)	–0.35
	150 mg of SEC	46	–	32 (70%)	20 (43%)	–0.18
	Placebo	68	–	3 (4%)	0 (0%)	–0.08
FIXTURE49	300 mg of SEC	50	–	36 (72%)	22 (44%)	–0.41
	150 mg of SEC	49	–	29 (59%)	19 (39%)	–0.19
	50 mg of ETN	44	–	17 (39%)	8 (18%)	–0.29
	Placebo	49	–	1 (2%)	1 (2%)	0.02

FIGURE 2.

Forest plot of the efficacy of 300 mg of SEC vs. placebo for PASI 75 at 12 weeks in PsA patients with moderate–severe psoriasis. 49 df, degrees of freedom; M–H, Mantel–Haenszel.

Evaluating the secukinumab and certolizumab pegol trial results in comparison with other treatments

In order to more fully evaluate the clinical efficacy of SEC and CZP, the trial results of these two newer biologics need to be compared with each other and with the results of the older biologics (and APR). However, this is not straightforward for two reasons. First, there is variation across trials with respect to previous biologic use.

• The populations recruited to clinical trials have changed over time, with earlier trials excluding biologic-experienced patients and later trials including such patients.

• The RAPID-PsA trial was more selective than the FUTURE 2,48 PSUMMIT 259,66 and PALACE trials60,61,65 in recruiting its biologic-experienced patients: only in RAPID-PsA were patients with primary failure of a previous biologic excluded (see Characteristics of the randomised controlled trials included in the systematic review of short-term efficacy).

Second, placebo response rates have increased markedly over time across the trials included in this review. This issue is key when interpreting RRs because, although RRs are easy to interpret clinically, their ceilings (maximum values) are limited by baseline response rates. For example, in the FUTURE 2 trial48 the placebo response rate for PsARC was (confidential information has been removed) in the biologic-naive subgroup. This high rate meant that the maximum possible RR would be (confidential information has been removed); this maximum result is lower than some of the actual RRs for other biologics presented in Table 29, which compares unadjusted RRs across the trials in the NMAs. Comparisons between treatments using odds ratios (ORs) and that adjust for the varying placebo rates were therefore necessary (see Chapter 4).

Examination of the trial baseline characteristics across trials offers no clear reason as to why placebo response rates in biologic trials have increased over time. The PsARC placebo response rates increased most markedly from 2013 onwards, starting with the PSUMMIT trials. 60,61,65 One theory is that patient and clinician expectations have increased over time (i.e. more caution and lower expectations when the first biologics were trialled, and more confidence about the likely benefits in more recent trials). Subjective patient- and clinician-reported outcomes such as PsARC and ACR may be prone to such expectation effects. This theory might also explain why, within trials, higher placebo response rates are observed in the biologic-naive subgroups than in biologic-experienced subgroups, where treatment expectations might be lower. Coupled with this is the trend – beginning with the PSUMMIT trials – for increases in the number of active treatment arms offered in trials: typically there was one active arm in the early trials and two or more active arms in more recent trials (e.g. the FUTURE 2 SEC trial had three active treatment arms: 75, 150 and 300 mg). Patients in the more recent trials might therefore also be more confident and optimistic about the likelihood that they are receiving an active treatment.

Ideally the different treatments would be compared in head-to-head trials. However, only one trial identified in the systematic review compared two or more biologics directly in a PsA population. The Atteno et al. trial64 compared INF, ETN and ADA. It reported that patients on INF and ADA showed the greatest improvement in terms of PASI (statistically significantly better than ETN), whereas patients on ETN showed the greatest improvement in TJC (statistically significantly better than INF and ADA) and HAQ-DI (statistically significantly better than ADA). However, the reliability of this study’s results are limited somewhat by its small size (100 patients were randomised in total). This trial also did not report its methods clearly (see Table 6), and was rated as being at high risk of bias (although blinding would be difficult to achieve in such a trial). Finally, by reporting results only at the 52-week time point, the results of this trial could not be included in our NMAs.

Long-term effectiveness

Review of anti-tumour necrosis factor patient registry studies

Review of the natural history of psoriatic arthritis: registry and cohort study data

A total of four publications33,98–100 analysing patterns of natural disease progression in registries or long-term cohort data were found and are shown in Table 36. These were reviewed in order to gain an understanding of the manner in which disease progresses in patients who do not receive anti-TNF therapy, despite being eligible for treatment. Owing to the now ubiquitous nature of anti-TNFs and only recent recognition of PsA as a separate and distinct form of arthritis, information on the long-term uncontrolled progression of the disease is scarce. Two of the studies33,100 found in the literature search were different analyses of the same data set derived from the Norfolk Arthritis Register (NOAR): one was a 2-year prospective cohort study99 and the other a retrospective analysis of a Canadian single-site patient registry. 98

The studies explore changes in functional disability in terms of HAQ-DI score and bone erosion as measures of disease activity and progression over time. There is a great deal of variability between the three cohorts under observation in terms of both baseline characteristics and patterns of disease. It should be noted that disease classification of the NOAR cohort99,100 was performed retrospectively and both studies analysing the 79 patients emphasise that they are unlikely to be representative of PsA patients, preferring instead to refer to them as having polyarthritis plus psoriasis. The Morgan et al. 100 study analysed the change in median cohort HAQ-DI score over 5 years in 79 patients, finding an increase of 0.125 units over the observation period, indicating a small increase of 0.025 units in HAQ-DI score every year. The patients in this analysis may or may not have been treated with DMARDs over this period. The analysis in Rodgers et al. 33 includes only those patients who had previously received two or more DMARDs at each time point, finding an annual HAQ-DI score change of –0.060 units per year over the first 2 years (n = 24), and an annual increase of 0.077 units over years 3 to 5 (n = 52). This represents a faster progression of disease than that found in the Morgan et al. 100 study, but is inconsistent and derived from a small cohort of varying size.

A prospective cohort study of progression in early arthritis carried out by Kane et al. 99 found that HAQ-DI score changed from 0.71 units at baseline to 0.4 units at 1 year and remained as such until the end of the 2-year observation period, representing a decrease of 0.31 units. This decrease is likely to be explained by the increase in uptake of DMARDs, as 12% of patients were receiving DMARD treatment at baseline, compared with 59% at 1 year and 56% at 2 years. This was the only study that recorded radiographic progression, finding consistent increases across all measures between baseline and 2 years, despite the simultaneous drop in HAQ-DI score. The Sharp erosion score increased from 1.2 units at baseline to 3 units at 2 years, demonstrating how HAQ-DI score change may not reflect progressive radiographic damage, particularly during early disease.

The study by Husted et al. 98 was the longest and largest study of natural history of PsA, with 341 patients included and observed for up to 10 years. This study found that the patient population exhibited several patterns of disease progression, rather than just universal consistent deterioration over time. Patients were assigned to one of three disability states based on their HAQ-DI score. These were as follows: ‘no disability’ (a HAQ-DI score of < 0.5 units), ‘moderate disability’ (a HAQ-DI score of 0.5–1.5 units) and ‘severe disability’ (a HAQ-DI score of 1.51–3.0 units). The transition of patients between groups was recorded over the course of the observation period to ascertain the direction of change in their symptoms. Forty-six per cent remained in the same disability group over the course of the study, with 28% of these in the no disability state, 12% in the moderate state and 6% in the severely disabled state. A total of 26.7% of patients made a single transition between disability groups, reflecting steady improvement or deterioration, and 27.3% experienced two or more transitions between disability states. Although this methodology may reveal broad patterns of disease progression, it appears to be insensitive to change within groups and weights HAQ-DI score change near thresholds more highly (e.g. a patient with a baseline HAQ-DI score at the lower end of a Markov group can experience a significant worsening of their disability without progressing into the next group). Mean HAQ-DI score change between consecutive assessments was 0.55 units (± 0.32 units) for those moving from a lower to a higher state and –0.57 units (± 0.36 units) for those moving to a lower state, with assessments being on average 1.29 years apart. In those patients who did not move between groups, the mean HAQ-DI score change was –0.002 units (± 0.215 units). A more complete picture of patterns of disease progression would have been possible had there been more Markov states. The mean HAQ-DI change for the majority of patients at any one time was effectively zero, but this may conceal significant within-group changes in either direction. Greater age was associated with a slower improvement in HAQ-DI score in those in the moderate and severe disability groups, and disability worsened more slowly in males than in females. Time since PsA diagnosis was related to more frequent transition between disability states, and there was no association between PASI score and transition between disability states. In summary, this study indicates that functional disability (as measured via the HAQ-DI) in PsA is generally stable over time in the majority of patients, but there are groups who exhibit patterns of more rapidly worsening or improving symptoms at certain periods, with some experiencing fluctuating deterioration and improvement over time.

Owing to the paucity of observational data on natural history of PsA, it is difficult to produce accurate estimates of yearly disease progression rates without anti-TNF therapy. None of the included studies can claim to provide accurate long-term estimates on uncontrolled disease progression. It is clear from the largest cohort that functional disability deteriorates over time, but the course of HAQ-DI progression is not constant or predictable. Therefore, it is unclear if an average rate of HAQ-DI change is a useful statistic, as this change is neither constant nor generalisable to the patient population. The Kane et al. study99 does show that, despite reductions in functional disability in early-stage disease under DMARD therapy, radiographic progression continues to occur, which theoretically will ultimately result in worsening disability in the long term; however, because of the lack of large and long-term observational studies, HAQ-DI change over time in uncontrolled PsA is yet to be properly measured.

Review of adverse effects of certolizumab pegol, secukinumab and comparators

Framework of analyses

The evidence synthesis was undertaken using WinBUGS (version 1.4.3; MRC Biostatistics Unit, Cambridge, UK). WinBUGS is a Bayesian analysis software tool that, through the use of Markov chain Monte Carlo methods, evaluates posterior distributions for the parameters of interest given likelihood functions derived from data and prior probabilities (uninformative priors were used throughout). There were few individual studies on each treatment; therefore, fixed-effect models were used across studies in all analyses. Parameter estimates for all functional parameters were reported from the models. These differ by outcome, and further details are presented in Methods. Treatment effects were expressed in relation to placebo. Owing to the sparse evidence imposing a high level of uncertainty over estimates of functional parameters, point estimates are medians throughout. Some models assumed exchangeability across treatments within a class, that is, different treatments of the same class were assumed to be similar, rather than equal. Within such models we reported the relative effectiveness estimates for each treatment (called shrunken estimates), rather than the class means, allowing us to represent any residual differences across treatments.

The validity of a NMA depends on an assumption of homogeneity/exchangeability between all the trials included in the network [i.e. that there are no essential differences between the methods, populations and interventions being studied, and that any differences are a result of chance (as in a standard meta-analysis)]. The lack of homogeneity/exchangeability between studies involving one of the treatments of interest and studies involving the other treatments of interest may generate inconsistency. Checking for consistency in the current network was not possible because of the lack of trials that directly compare active agents. Our examination of the study details and patient characteristics (see Chapter 3, Characteristics of the randomised controlled trials included in the systematic review of short-term efficacy) identified that the trials of the newer agents (SEC, CZP, UST and APR) included biologic-experienced patients as well as biologic-naive patients. Given that it is evident from large observational data sets (see Chapter 3, Review of anti-tumour necrosis factor patient registry studies) that efficacy response rates in biologic-experienced patients are lower than in biologic-naive patients, it was considered inappropriate to conduct an ‘all-patients’ NMA for any outcome, and that, instead, biologic-naive and biologic-experienced patients should be analysed separately. Therefore, separate analyses (separate networks) for treatment-naive and treatment-experienced patients were constructed for each of the four outcomes: one each for PsARC, HAQ-DI conditional on PsARC, PASI 50, 75 and 90, and ACR 20, 50 and 70 responses. A summary of the trials reporting data on each of these outcomes is presented in Table 39. It should be noted that the NICE scope112 for the present appraisal subdivides biologic-naive patients into those who have not responded to one cDMARD and those who have not responded to two cDMARDs. However, sufficient data were not available for these further levels of subgroup analysis.

As discussed in Chapter 3, Evaluating the secukinumab and certolizumab pegol trial results in comparison with other treatments, another important difference between the included trials is the observed results in the placebo arms, particularly for PsARC (see Table 40), PASI outcomes (see Table 50) and ACR (see Table 56). Our investigations on trial designs and patient characteristics did not identify any clear reasons for such differences, other than that placebo response rates appear to have increased over time. This observation (termed ‘placebo creep’) has been made in several other areas of clinical research and its impact on indirect treatment comparisons has been discussed. 113 In the current review, across all trials, the PsARC placebo response rates are high, but are much higher in more recently conducted trials, and this has implications when interpreting unadjusted effect estimates. This is because the ceilings (maximum values) of RRs are limited by baseline response rates. For example, in the FUTURE 2 trial,48 the placebo response rate for PsARC in the biologic-naive subgroup was (confidential information has been removed), which meant that the maximum possible RR would be (confidential information has been removed); this maximum result is lower than some of the actual RRs for other biologics (see Table 40). Higher placebo rates therefore appear to dilute effect estimates somewhat. This is also demonstrated by the examining the RRs moving up the ACR outcome thresholds from ACR 20 to ACR 70, which generally increase (see Table 29). However, it is not clear exactly how these varying placebo rates will affect treatment effects when calculated using ORs. The evidence synthesis – which was based on ORs – therefore explored a potential relationship between baseline risk and relative effectiveness. The NMA explored scenarios where a metaregression on baseline risk (i.e. placebo response) was implemented for PsARC, PASI and ACR outcomes, which imposes an interaction effect between baseline risk and relative effectiveness. 114 Further details of these analyses are presented below. Given that HAQ-DI scores are modelled conditional on PsARC response, such an interaction effect was deemed to be less relevant, and a metaregression model was not implemented on HAQ-DI.

Psoriatic Arthritis Response Criteria response

Health Assessment Questionnaire-Disability Index changes conditional on Psoriatic Arthritis Response Criteria response/non-response