Article Text

Extended report
Application of composite disease activity scores in psoriatic arthritis to the PRESTA data set
  1. Oliver FitzGerald1,
  2. Philip Helliwell2,
  3. Philip Mease3,
  4. Aizad Mumtaz1,
  5. Laura Coates2,
  6. Ronald Pedersen4,
  7. Henk Nab5,
  8. Charles Molta4
  1. 1Department of Rheumatology, St Vincent's University Hospital, University College Dublin, Dublin, Ireland
  2. 2Leeds Institute of Molecular Medicine, Section of Musculoskeletal Disease, University of Leeds, Leeds, UK
  3. 3Swedish Medical Center and University of Washington School of Medicine, Seattle, Washington, USA
  4. 4Department of Statistics, Pfizer Inc., Collegeville, Pennsylvania, USA
  5. 5Department of Rheumatology, Pfizer Europe, Rome, Italy
  1. Correspondence to Professor Oliver FitzGerald, Bone and Joint Unit, St Vincent's University Hospital, Elm Park, Dublin, Ireland; oliver.fitzgerald{at}ucd.ie

Abstract

Objective This study aimed to compare the performances of the Modified Composite Psoriatic Disease Activity Index (mCPDAI) and the Disease Activity index for PSoriatic Arthritis (DAPSA) in an interventional study of etanercept in psoriatic arthritis.

Methods The components of the CPDAI and DAPSA were extracted using PRESTA (Psoriasis Randomized Etanercept STudy in subjects with psoriatic Arthritis) study data. Data for four of the five domains of the CPDAI—thus an mCPDAI—were available: joints, skin, dactylitis and enthesitis (spinal involvement was not assessed). Domains in the calculation of DAPSA were subjected to global assessment of pain, swollen and tender joint counts, and C reactive protein. Subjects were randomised to etanercept 50 mg weekly (n=373) or 50 mg twice weekly (n=379) for 12 weeks; all subjects then received etanercept 50 mg weekly for 12 weeks. The performance of the scores at baseline and on weeks 12 and 24 was compared between the two treatment regimens.

Results The mCPDAI and DAPSA could distinguish response to treatment comparing baseline and 12-week or 24-week values (p<0.0001). The mCPDAI, not DAPSA, could distinguish response between the two treatment groups at 12 weeks (p=0.0492), but not at 24 weeks. All domains evaluated contributed to the data variability of the mCPDAI; the most significant were dactylitis (r=0.64) and enthesitis (r=0.60).

Conclusion In psoriatic arthritis with severe skin involvement, the mCPDAI was able to distinguish treatment response between the two etanercept doses. DAPSA, while demonstrating improvement in both groups over time, was unable to distinguish response between the different doses of etanercept. Further studies are needed to confirm the sensitivity of both indexes.

View Full Text

Statistics from Altmetric.com

Introduction

Psoriatic arthritis (PsA) is a heterogeneous disorder that affects peripheral and axial joints, in addition to dactylitis and enthesitis; the majority of cases have skin involvement.1,,3 Although not all of these clinical features may occur simultaneously, it is important to be able to capture data on all in order to assess the complexity of the disease and its impact. A composite measure is one way of assessing all relevant clinical outcomes in one single instrument. By definition, it incorporates and, in some cases, combines several dimensions of disease status often by combining these different domains into a single score. Such indexes are more efficient than unidimensional instruments and generally give larger effect sizes in interventional trials.

The composite indexes used for evaluating rheumatoid arthritis—the American College of Rheumatology index4 and the 28-joint Disease Activity Score (DAS28)5—have historically been used for measuring PsA in clinical trials because they effectively assess peripheral articular involvement, pain and function. The PsA response criteria were specifically designed for use in randomised controlled trials in PsA and thus acquired the label of a PsA-specific response measure.6 However, they remain largely a peripheral articular responder index and, being a dichotomous response measure, may overestimate patient response. More recently, a novel approach to constructing a disease assessment measure in PsA has been proposed—the Composite Psoriatic Disease Activity Index (CPDAI).7 Based on work by the Group for Research and Assessment of Psoriasis and Psoriatic Arthritis (GRAPPA) to guide treatment decisions in PsA,8 involvement is assessed in five domains: peripheral joints, skin, entheses, dactyli and spine, with each domain scoring individually and contributing to the total score. Initial data suggest face and content validity and responsiveness to interventions.7

The Disease Activity index for REactive Arthritis (DAREA) is a composite score for disease activity that was originally developed for reactive arthritis.9 Following a principal component analysis, we extracted four components that are significant for PsA: swollen and tender joint counts, patient global score, pain score and C reactive protein (CRP) level. The application of DAREA to PsA disease has been validated, and it is now referred to as the Disease Activity index for PSoriatic Arthritis (DAPSA) when used in patients with PsA.10 11

The PRESTA (Psoriasis Randomized Etanercept STudy in subjects with psoriatic Arthritis) study was a large clinical trial (n=752) that examined the effects of two etanercept regimens on the skin and joint symptoms of subjects with psoriasis and PsA.12 This post hoc analysis was designed to compare the performance of a modified CPDAI (mCPDAI) with that of DAPSA in the PRESTA trial data set.

Methods

Study design

The design of the PRESTA trial has been previously reported in detail.12 Briefly, eligible subjects were 18 years of age or older, with moderate to severe clinically stable plaque psoriasis involving ≥10% body surface area and a dermatologist-assessed score of moderate or worse (≥3 on a scale of 0–5) on the Physician Global Assessment of psoriasis. In addition, subjects had active PsA (assessed by a rheumatologist), including ≥2 swollen joints and ≥2 tender joints at screening and baseline, subject-reported joint pain for ≥3 months prior to screening and a negative serum rheumatoid factor within 6 months of screening.

Subjects were randomly assigned to receive either etanercept 50 mg twice weekly or 50 mg weekly during the initial 12-week double-blind period, followed by a 12-week open-label period in which all subjects received etanercept 50 mg weekly. Subjects and investigators remained blinded to their treatment throughout the study.

Assessment instruments

Modified CPDAI

The CPDAI includes disease measured in five domains: joints, skin, dactylitis, enthesitis and spine; however, the PRESTA trial did not evaluate spinal involvement, so the spinal component was not included. Using the data obtained from the PRESTA study, we extracted four domains that were used to calculate the mCPDAI: joints (66 swollen joint count and 68 tender joint count; Health Assessment Questionnaire (HAQ)), skin (Psoriasis Area and Severity Index (PASI) and Dermatology Life Quality Index (DLQI)), dactylitis (a simple count of each digit involved) and enthesitis (number of tendons/fascia insertion sites showing enthesitis scored from 0 to 4, based on palpation of Achilles tendon and bilateral plantar fasciae insertion). The instruments used for dactylitis and enthesitis differ from those used in the original CPDAI description7 in order to reflect the data that were captured as part of the PRESTA trial. The mCPDAI domains were each scored using a 4-point scale from 0 (no disease activity) to 3 (most severe disease activity), giving an mCPDAI score range of 0–12.7

Disease Activity index for PSoriatic Arthritis

The DAPSA score is a simple arithmetic sum of the following: (1) patient-reported pain and global assessment on a 0–10 scale; (2) joint involvement, signified best by 66 swollen joint count and 68 tender joint counts as main variables; and (3) acute phase response, represented by CRP (mg/dl). The DAPSA value was calculated as the simple sum of five components extracted from the following PRESTA data set:

DAPSA=SJC66+TJC68+PtGA+pain+CRP.9

Statistical analysis

These analyses were performed on the baseline data of the PRESTA study, as well as on the 12-week and 24-week results of the modified intention-to-treat population (n=752) using the last-observation-carried-forward method for missing data imputation.

In view of the absence of spinal-related data in the PRESTA trial, an mCPDAI was calculated using the four remaining domains (each scored 0–3), giving a maximum score of 12. With the use of individual subject data, correlations between the DAPSA score and the mCPDAI score and various continuous disease measures employed Spearman's correlation coefficient (as various end points may not all have normal distributions). Within-group change from baseline in DAPSA and the mCPDAI was evaluated for statistical significance with paired t tests. Between-treatment comparisons of changes in DAPSA and the mCPDAI on weeks 12 and 24 were evaluated with analysis of covariance models, with baseline as covariate and with factors such as region and treatment group. Exploratory stepwise regression models for evaluating predictors of baseline DAREA and mCPDAI scores and week 24 change in DAPSA and mCPDAI scores used a p value of 0.15 to enter the model and a p value of 0.10 to stay in the final model.

The study was performed in accordance with the ethical principles of the Declaration of Helsinki. The protocol and its amendments received independent ethics committee or institutional review board approval and regulatory review and approval before site initiation and recruitment of subjects. All elements of informed consent were explained to eligible subjects, and adequate time was allowed for questions and for subjects to make voluntary decisions. No subject underwent procedures specific for the protocol until he or she had signed and dated an approved informed consent form.

Role of the funding source

Wyeth Research, which was acquired by Pfizer Inc. in October 2009, sponsored the PRESTA trial and was responsible for the collection and analysis of the data for that study. The authors were involved in the design of this validation study, interpretation of data, manuscript preparation and decision to publish. All statistical analyses were performed by the Global Biostatistics and Programming Department of Pfizer Inc. The corresponding author had full access to all the data in the study and had the final responsibility for the decision to submit the report for publication.

Results

Subjects

The demographic and baseline clinical characteristics of the two treatment groups from PRESTA are shown in table S1. At baseline, there were no significant differences in demographics between the groups. Furthermore, there were no differences in the duration or severity of psoriatic disease. Subjects in the PRESTA trial achieved significant improvement from baseline in skin, joint and entheseal disease components on week 24, regardless of treatment regimen.12

The mCPDAI and DAPSA could distinguish response to treatment comparing baseline and 12-week or 24-week values (p<0.0001 for all). The mCPDAI, not DAPSA, could distinguish response between the two treatment groups at 12 weeks (p=0.049 and p=0.738 for the mCPDAI and DAPSA, respectively). Neither of the indexes could distinguish response in the two treatment groups at 24 weeks (p=0.65 and p=0.34 for the mCPDAI and DAPSA, respectively) (table 1 and 2).

Table 1

Descriptive statistics for the mCPDAI on weeks 12 and 24

Table 2

Descriptive statistics for DAPSA on weeks 12 and 24

Content validity

All domains contributed to the data variability of the mCPDAI at baseline, with dactylitis (r=0.636) and enthesitis (r=0.600) being the most significant (table 3). Joint scores contributed most to the variability in the DAPSA score (swollen joint count r=0.8, total joint count r=0.91) (data not shown).

Table 3

Spearman correlations with the mCPDAI sum at baseline, on week 12 and on week 24 (raw value correlations)

Predictors of response

Exploratory stepwise regression models showed that enthesitis, HAQ, dactylitis and DLQI were the most significant predictors of baseline mCPDAI scores (table 4). For change in the mCPDAI from baseline, stepwise regression revealed that a change in enthesitis, DAS28, HAQ, dactylitis and DLQI was the most significantly associated (table 5). For change in DAPSA (data not shown), a change in all of the five components was significantly associated.

Table 4

Final stepwise regression results at baseline: end points in the final model (the change in the mCPDAI sum from baseline is predicted by the change in predictors from baseline)

Table 5

Final stepwise regression results on week 24: end points in the final model (the change in the mCPDAI sum from baseline is predicted by the change in predictors from baseline)

Discussion

In this study, the performance of two composite disease activity scores, mCPDAI and DAPSA, is tested in the large PRESTA data set. While both the mCPDAI and DAPSA could distinguish response to treatment at 12 weeks, only the mCPDAI could distinguish between the two doses of etanercept used. The data showed that neither score could distinguish between the treatment groups on week 24. However, the design of the PRESTA study dictated that between weeks 12 and 24, all patients received identical open-label doses of etanercept. Meaningful differences between treatment groups in the first 12 weeks would likely have dissipated by the end of the second 12 weeks.

Notably, recruited subjects had substantial levels of disease at baseline with mean PASI of 20 and 19 in the etanercept twice weekly and etanercept once weekly groups, respectively, and mean swollen joint counts of 12 and 13. Providing the mCPDAI with good content validity, all domains assessed contributed to data variability. The results described support the conclusion that the mCPDAI is currently the more appropriate composite disease activity measure, even in the absence of the spinal component, to include in the setting of randomised clinical trials.

It is likely that these results reflect the inclusion in the mCPDAI of all the domains involved in PsA, including psoriasis, arthritis, enthesitis and dactylitis—thus not focusing solely on joint-related features. In the original PRESTA publication,12 the higher dose of etanercept contributed to improved skin responses, but no added benefit was observed in joint-related outcome measures. In particular, by including a skin domain, the mCPDAI is well placed to reflect improvement in all of the ways in which patients with PsA may be affected by their disease. When the spinal component is added in, even greater sensitivity may be expected. It appears that the CPDAI, either modified without a spinal component or as a full set of five domains,7 is a valuable tool for the clinic. In contrast, DAPSA largely reflects the joint domain although the patient outcome measures (visual analogue scale for pain and arthritis) may also reflect activities in the entheseal, dactylic and spinal domains to a certain extent. The lack of specific questions beyond joint involvement appears to impair the ability of DAPSA to discriminate the differences in the treatment arms.

It might be argued that it is not possible to have a meaningful composite score that encompasses dermatological and articular domains. Detractors might argue that there is generally a poor relationship between disease severity and activity in the two domains. If this is indeed the case, there may frequently be patients in whom the dermatological disease is extensive and severe alongside minimal joint disease. The opposite scenario may also be common. In these cases, a composite score in which one of the domains dominates might under-represent the severity of other domains and—if the score were used to judge treatment interventions—might deny the introduction of effective systemic treatments. To counter this possibility, the mCPDAI applies a rule stating that if the patient scores a ‘severe’ grading in any one of the five domains, then he/she is eligible for systemic interventions (such as tumour necrosis factor inhibitors), irrespective of his/her total mCPDAI score. An alternative scoring system for the mCPDAI might be to use a score in which inactive domains are omitted and the total score is divided by the number of active domains. In this way, if a person had no tender or swollen joints, no dactylic digits and no spinal involvement, his/her composite score would reflect only skin and entheseal diseases.

The PRESTA data set, while extensive, included neither information on spinal involvement nor data on the Bath Ankylosing Spondylitis Disease Activity Index or the Ankylosing Spondylitis Quality of Life. For this reason, it was not possible to calculate the full score range of 0–15. If these data had been available, the mCPDAI would have possibly performed even better as a discriminator between the two treatment regimens, although the use of higher doses of etanercept in axial spondyloarthropathy has not been fully investigated. Furthermore, as the PRESTA data did not include a non-active treatment comparator, it was not possible to investigate the performance of the mCPDAI or DAPSA as outcome measures in comparison with placebo. In addition, the subjects entered into the PRESTA study were rather unique, as most patients with PsA seen by rheumatologists have more modest PASI scores. For this reason, comparing DAPSA and the mCPDAI in a typical rheumatology office might have given different results. The ongoing study to develop a new composite measure and to compare existing instruments (GRAPPA Composite Exercise (GRACE)) will enable such a comparison.13

The instruments used to measure dactylitis and enthesitis in this study were different from those used in the original CPDAI description.7 This reflected the data actually collected as part of PRESTA. While we acknowledge this difference, we doubt that it contributes in any significant way to the conclusions of this study. For further validation of these composite measures, it will be important to ensure that all of the appropriate measures are included in future randomised controlled trials of patients with PsA.13

Both composite instruments studied in this paper will need further development. For example, correlation of individual domain scores with baseline status or other analyses could be performed. Such analyses are beyond the scope of the current study. As already mentioned, the GRACE study will permit a comparison of these measures in a different population in which the articular disease is more prominent. Furthermore, the GRACE project will generate data on responsiveness and external validity, and will enable the calculation of cut-offs for high and low disease activities and the magnitude of response. Future work might also include a consensus approach to the cut-offs for severity used in each of the domains of the CPDAI—a project already initiated by GRAPPA.

Conclusions

The mCPDAI, which reflects domains such as skin, enthesitis and dactylitis, as well as joint disease, was able to distinguish global treatment response between the two different etanercept regimens at 12 weeks in a group of patients with moderate to severe skin and joint diseases. Further studies are needed to determine if the mCPDAI is an improvement over other measures in patients with PsA who were recruited from rheumatological or other settings.

Acknowledgments

The authors thank Robyn Boyle (Precise Publications LLC) and Patricia McChesney (UBC Scientific Solutions) for assistance in drafting and revising the manuscript with the direction of the authors (funded by Pfizer Inc), in accordance with the standards of the International Committee of Medical Journal Editors.

References

View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

    • Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Funding Wyeth Research, acquired by Pfizer Inc.

  • Competing interests OFG, PH, PM and LC received money for consultancy and/or money for travel/accommodations/meeting expenses from Abbott, Centocor, Janssen Cileg and Pfizer. AM received money for Abbott Newman clinical research fellowship. RP, HN and CM are employees or former employees of Wyeth Research and Pfizer Inc.

  • Ethics approval Ethics approval was obtained from an independent ethics committee or from the institutional review board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.