Analysis of Outcome Measures in Hidradenitis Suppurativa 

Lead InvestigatorJohn Frew, Rockefeller University
Title of Proposal Research:Analysis of Outcome Measures in Hidradenitis Suppurativa
Vivli Data Request: 3320
Funding Source: None
Potential Conflicts of Interest: None
Summary of the Proposed Research: 

There are a plethora of outcome measures for Hidradenitis Suppurativa (HS). The HiSCR (Hidradenitis Suppurativa Clinical Response) is the most well-known and widely used in clinical trials, however Physician Global Assessment (PGA), International Hidradenitis Suppurativa Severity Score (IHS4:developed by the European Hidradenitis Suppurativa Foundation Investigator Group) and the Acne Inversa Severity Index (AISI) have been recently developed since HiSCR was established. Given the multiple Phase 2 clinical trials currently active (and planned) in multiple monoclonal antibodies and antibody targets,  (IL-17, IL-23, IL-1 family) and small molecule therapies (JAK Inhibitors), comparing results across studies will be vital for patients and physicians to choose the most appropriate drug for their patients. 

One issue vexing the interpretation of efficacy in Hidradenitis Suppurativa clinical trials is the heterogeneity of disease. It is clear from the collated PIONEER trial data (PIONEER 1, PIONEER 2 and PIONEER OLE), as well as from analyses of cytokine and immunohistochemistry data (work from our research group- currently unpublished but in press), that different states of disease in HS respond to different therapies. 

This is likely due to the fact that in Hurley Stage 1 and 2 disease- TNF-alpha is the most prominent cytokine amenable to intervention, however feed-forward mechanisms invoking Th17 pathways emerge as the disease worsens to demonstrate psoriasiform hyperplasia and early upregulation of TGF-Beta – leading to early sinus tract formation. 

As sinus tract formation and scarring are well established, a continual inflammatory loop mediated by fibroblasts, keratinocytes (via the role of ICAM-1 and TGF-Beta) perpetuates a keratinocyte-mediated inflammatory cascade which is less responsive to TNF-alpha inhibition. It is hypothesized by our group that this disease state would be more responsive to IL-1 or small molecular inhibition. Early scarring may be more susceptible to intervention with IL-17 inhibition, but once scarring and fistulisation is well established, IL-17 inhibition may only work to reduce discharge. 

The relevance of this inflammatory dichotomy to clinical trials and outcome measures is that, in contrast to HiSCR (which does not take into account draining fistulae), other outcome measures weight draining fistulae very highly. For example, the IHS4 weights draining fistulae 4 times higher than number of nodules.  

Whilst all registered studies on use HiSCR as the primary outcome measure- other outcome measures are listed as secondary outcomes. Given the high weighting of draining fistulae in these other measures, one can hypothesize that relatively modest reductions in draining fistulae counts will lead to dramatic decreases in scoring for the IHS4 and AISI. This would result in an overestimation of effect, to the detriment of drugs such as Adalimumab, which can be effective in reducing draining fistulae, but which is not reflected by the HiSCR Data. 

Statistical Analysis Plan:

The statistical analysis will proceed to duplicate the methods used by Kimball et al in the study from which the data is being requested.

  • Effect measure of interest: Change in outcome measure score (IHS4/ AISI/ from Week 0 to Week 12)
  • Description of the population to be analyzed: All Patients in PIONEER 1 and PIONEER 2
  • Endpoints (variables) to be analyzed: Nodule Count, Abscess Count, Draining Fistulae Count, Smoking Status, BMI, Hurley Stage, Family History, Modified Sartorius Score, DLQI score, Patient Global Assessment Score

In addition to following the methods of analysis in Kimball et al, we will proceed by examining the distributional qualities of all measures, identifying outliers, attending to transformations needed to satisfy analytic assumptions, and examine and resolve any missing data issues through the implementation of various imputations methods as needed.

Depending upon distributional properties of the outcome measure of interest, we will initially compute means, standard deviations, minimum and maximums (for continuous measures), or sample sizes and proportion of responses (for categorical and count measures), both overall and by group (i.e. Adalimumab vs control), for baseline and follow-up time points. Group differences and baseline, and at follow-up, will be tested using t-tests (or non-parametric equivalent) and chi-squared tests.

We will then proceed to compute the observed treatment effects by regressing the Week 12 outcome variable on the primary independent dummy variable for treatment (i.e. 1= Adalimumab, 0=Control) along with the baseline value of the outcome serving as the sole covariate, using either OLS, Logistic, or Poisson regression (for Continuous, binary, or count data, respectively). We will then adjust these observed measures (i.e. mean differences, Odds Ratio’s or differences in counts) by identified additional covariates (e.g. Gender, Race, Age), and then explore the moderation effects by examining the difference in effects under stratified analyses, and interaction models in regression to test for significance of any observed moderation effects.

Requested Studies:
A Phase 3 Multicenter Study of the Safety and Efficacy of Adalimumab in Subjects With Moderate to Severe Hidradenitis Suppurativa – PIONEER II 
Sponsor: AbbVie
Study ID: NCT01468233 
Sponsor ID: M11-810 

A Phase 3, Open-Label Study of the Safety and Efficacy of Adalimumab in Subjects With Moderate to Severe Hidradenitis Suppurativa – PIONEER (Open-Label Extension) 
Sponsor: AbbVie
Study ID: NCT01635764
Sponsor ID: M12-555 

A Phase 3 Multicenter Study of the Safety and Efficacy of Adalimumab in Subjects With Moderate to Severe Hidradenitis Suppurativa – PIONEER I 
Sponsor: AbbVie
Study ID:  NCT01468207
Sponsor ID: M11-313 

 Public Disclosures:

  1. Clinical Response Rates, Placebo Response Rates and Significantly Associated Covariates Are Dependent Upon Choice of Outcome Measure in Hidradenitis Suppurativa: A Post-Hoc Analysis of PIONEER 1 and 2 Individual Patient Data
    Frew, John W. et al.
    Journal of the American Academy of Dermatology, Volume 0, Issue 0
    doi: 10.1016/j.jaad.2019.12.044
  2. Frew, J., Jiang, C., Singh, N., Grand, D., Navrazhina, K., Vaughan, R. and Krueger, J. (2020), Dermal Tunnels Influence Time to Clinical Response and Family History Influences Time to Loss of Clinical Response in Hidradenitis Suppurativa Patients Treated with Adalimumab.
    Clinical and Experimental Dermatology. Accepted Author Manuscript.
    doi: 10.1111/ced.14448
  3. Frew J, Jiang C, Singh N, Navrazhina K, Vaughan R, Krueger J, Quantifying the Natural Variation in Lesion Counts Over Time in Untreated Hidradenitis Suppurativa: Implications for Outcome Measures and Trial Design, JAAD International (2020)
    doi: 10.1016/j.jdin.2020.09.005
  4. Frew, J. W., Singh, N., Jiang, C. S., Vaughan, R., & Krueger, J. G. (2021). The Impact of Body Mass Index Upon the Efficacy of Adalimumab in Hidradenitis Suppurativa. Frontiers in Medicine.
    doi: 10.3389/fmed.2021.603281