The interleukin 1 receptor antagonist anakinra to reduce disease severity of palmoplantar pustulosis in adults: APRICOT RCT and PLUM mechanistic study

Article history

The research reported in this issue of the journal was funded by the EME programme as project number 13/50/17. The contractual start date was in November 2015. The final report began editorial review in December 2020 and was accepted for publication in June 2021. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The EME editors and production house have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the final report document. However, they do not accept liability for damages or losses arising from material published in this report.

Copyright statement

Copyright © 2022 Cro et al. This work was produced by Cro et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaption in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – NIHR Journals Library, and the DOI of the publication must be cited.

2022 Cro et al.

Scientific background

Psoriasis is a common condition (estimated prevalence of 2% in the UK) that is known to have an impact on quality of life at a level comparable to other major diseases, including chronic heart disease and cancer. 3,4 Pustular forms of psoriasis are characterised by painful, intensely inflamed red skin studded with sheets of monomorphic, sterile, neutrophilic pustules. These pustules may be chronic. Pustular psoriasis is typically localised and involves the hands and feet [known as acral pustular psoriasis (APP)], although it can also occur more rarely as generalised, episodic and potentially life-threatening [generalised pustular psoriasis (GPP)]. 5,6 Some individuals may experience both forms throughout their life.

Although pustular psoriasis constitutes < 10% of all cases of psoriasis, it often ranks the highest of all psoriasis phenotypic variants in terms of symptoms (itch, pain and functional impairment, causing limited mobility and interference with daily living tasks and work). 7–9 Ultimately, the consequential impact is immense and equivalent to that of psychiatric illness and other major medical diseases. 10,11

Over the past decade, significant investment in novel therapies and the advent of biological therapies have revolutionised the treatment and management of plaque-type psoriasis. This has been primarily driven by scientific investigations of underlying genetic and immunological disease pathways. 12 By contrast, the treatment options for pustular psoriasis are currently profoundly limited. Super-potent (topical) corticosteroids, phototherapy, oral treatments [e.g. acitretin (Neotigason^®, Teva UK Ltd, Harlow, UK), methotrexate (Hospira UK Ltd, Maidenhead, UK) and ciclosporin (Capimune^®, Mylan, Canonsburg, PA, USA)] and targeted biologic therapies (notably tumour necrosis factor antagonists) are all used, although the evidence for their benefit is poor. 13 There is, therefore, a very significant unmet need in this patient group.

Rationale for the study

Recent evidence indicates that the molecular pathways underlying pustular psoriasis are distinct (from that observed with plaque-type disease) and involve the interleukin (IL) 36/IL-1 axis. Research has identified functionally relevant IL36RN mutations in both GPP and APP. 14–16 IL36RN encodes the IL-36 receptor antagonist IL-36Ra (this is an IL-1 family member that antagonises the pro-inflammatory activity of IL-36 cytokines). Disease mutations disrupt the inhibitory function of IL-36Ra, causing enhanced production of downstream inflammatory cytokines (including IL-1). 15,16 Indeed, in individuals with IL36RN mutations, IL-1 production has been shown to be significantly upregulated in response to IL-36 stimulation. 15 Furthermore, IL-1 is a cytokine that is known to sustain the inflammatory responses initiated by skin keratinocytes. 17

Interleukin 1 antagonists have previously shown therapeutic benefits in the treatment of IL-1-mediated diseases (many of which feature neutrophilic infiltration of the skin). 18 Furthermore, there has been research that suggests a key pathogenic role for IL-1 in pustular forms of psoriasis. 19,20

The model IL-1 antagonist proposed for the study was anakinra [Sobi (Swedish Orphan Biovitrum AB), Stockholm, Sweden]. Anakinra is an IL-1 receptor antagonist that is licensed for the treatment of rheumatoid arthritis and, during the timeline of this trial, periodic fever syndromes and Still’s disease. Anakinra was selected in preference of other licensed IL-1 antagonists for several reasons. It uniquely blocks the activity of both IL-1a and IL-1b. 18 Financially, it has the lowest drug acquisition cost (and this is of relevance to the NHS should anakinra show efficacy) and we had access to fully funded trial drugs through the manufacturer Sobi. Anakinra also possesses a rapid onset of action and an established safety profile (with > 70,000 patient-years’ exposure). Furthermore, there is early evidence of therapeutic benefit in participants with pustular psoriasis. 21

Hypothesis

We hypothesised that an IL-1 blockade would deliver therapeutic benefits in pustular forms of psoriasis. Therefore, this project aimed to investigate the clinical efficacy of an IL-1 blockade in palmoplantar pustulosis (PPP) (the most common form of pustular psoriasis) using the model IL-1 antagonist, anakinra, in a randomised, placebo-controlled trial with a two-staged adaptive design, followed by an open-label extension (OLE).

Study objectives

Design

This was a Phase IV, two-stage (stages 1 and 2), double-blind, randomised, placebo-controlled trial with an adaptive element followed by an OLE. 1 Participant data from both stages were included in the main stage 2 analysis.

Stage 1 compared treatment groups to ensure that there was sufficient efficacy and safety to progress to stage 2. The pre-planned interim analysis for stage 1 occurred after 24 participants had been randomised and followed up for 8 weeks. A decision to embark on stage 2 was made using stop/go efficacy criteria. Fresh pustule counts and PP-PASI scores at week 8 were compared between treatment groups to assess efficacy. If, at the end of stage 1, the placebo group did as well as, or better than, the anakinra group on both of the two outcomes, then the study would be stopped. However, because the anakinra group did better than the placebo group for at least one outcome, the study proceeded (to stage 2).

Furthermore, the primary outcome for stage 2 was chosen at the end of stage 1. The two candidate primary outcomes assessed were the fresh pustule count (across palms and soles) and the PP-PASI score. These were recorded at baseline and at weeks 1, 4, 8 and 12. To determine the efficacy of anakinra for PPP compared with placebo, the primary end point for stage 2 was prespecified to be the change in disease activity at 8 weeks (adjusted for baseline) measured using fresh pustule count (the default primary outcome) unless the PP-PASI score was judged to be more reliable and discriminating.

Stage 2 commenced with the PP-PASI score designated as the primary outcome (see Chapter 3, Stage 1). Stage 2 included the randomisation of a further 40 participants (64 in total). 1,22

Interventions

Participants were randomised (1 : 1) to receive (100 mg per day) either anakinra or placebo for 8 weeks, which was self-administered daily as a subcutaneous injection.

Participants who opted to take part in the OLE received a further 8 weeks of anakinra (100 mg per day) treatment, which was self-administered daily as a subcutaneous injection. The OLE was optional, and was offered to all participants who completed the 8-week treatment period and the 12-week follow-up visit.

Setting

The trial was set in 16 hospitals across England, Scotland and Wales.

Participants

Given that GPP is rare, episodic and potentially life-threatening, including GPP participants in a trial setting would have been difficult and potentially unethical (with the placebo group). Thus, the population designated for the study was participants with PPP. This chronic, localised form of pustular psoriasis involves the hands and/or feet, and is associated with significant disability. It is the most common form of pustular psoriasis, making recruitment feasible, and typically features chronic development of pustules so that we would expect to capture any treatment effect within the 8-week treatment period.

All participants were adults (aged ≥ 18 years) with a diagnosis of PPP that had been made by a trained dermatologist, with a disease duration of > 6 months and disease of sufficient impact and severity to require systemic therapy. To be randomised into the study, at the baseline visit participants had to exhibit at least moderate disease on the PPP-IGA, with evidence of active pustulation on palms and/or soles.

Women who were pregnant, breastfeeding or of child-bearing age and not on adequate contraception and men planning conception were excluded from taking part in the trial.

The specific inclusion criteria and exclusion criteria for the double-blind, placebo-controlled trial and the OLE are detailed below.

Identification and recruitment

Potentially eligible participants were identified by the following four methods:

1. In clinic at participating sites. Potentially eligible participants were identified in clinics and were approached directly by a member of the trial team and/or clinical care team, who explained the study to the patient and provided them with the patient information leaflet. Participants were then given as much time as they required (and at least 24 hours) to read the information leaflet and come to a decision about their participation.

2. Searching existing local health-care/medical databases at participating sites. Once sites were opened, local study teams identified potentially eligible participants through searching local clinic and pharmacy lists, electronic patient records, referral lists and letters, research databases and other lists (as appropriate). Potential participants were then contacted by their consultant and the research team (by letter, e-mail or telephone call), were invited to participate the study and were provided with the patient information leaflet.

3. Self-referral. Potential study participants identified themselves after becoming aware of the study. The study website (http://apricot-trial.com/; accessed 28 February 2020) included a specific page (which was taken down following the end of recruitment) on which participants could register through an interactive web-based patient recruitment questionnaire (see Appendix 1, Figure 19) for more information. These results were automatically sent to the trial manager and were used as the first line of eligibility screening. The trial manager/research nurse contacted potentially eligible patients by telephone and e-mail and invited them to participate (and provided them with the patient information leaflet if this had not already been downloaded by the patient from the trial website). If the patient remained interested in participating, then, with their consent, their contact details were provided to the trial team geographically closest to them to arrange a formal research consultation.

4. Participant identification centres (PICs). Potential study participants were identified at PICs following clinic visits or review of local clinic and pharmacy lists, electronic patient records, referral lists and letters, research databases and other lists. They were then contacted by their direct clinical care team (usually by letter, e-mail, telephone call or in person) and then invited to self-refer on the trial website (as detailed above) or (with their agreement) referred directly to the team at their chosen trial site for further information regarding participation.

Randomisation procedure

The randomisation service for the study was provided by the King’s Clinical Trials Unit (CTU). Following written consent at the screening visit, each participant was registered on the MACRO electronic case report form system version 4 (InferMed Macro) that generated a unique patient identification number (PIN). This unique PIN was then recorded on all source data worksheets and was used to identify the participants throughout the study.

At the baseline visit, randomisation occurred via a bespoke web-based randomisation system hosted by the King’s CTU (https://cturandomisation.iop.kcl.ac.uk/APRICOT/Login.aspx?ReturnUrl%20=%20%2fAPRICOT; accessed 1 September 2020). Authorised site staff were allocated a username and password for the randomisation system by the trial manager. An authorised staff member (typically the principal investigator or research nurse) logged into the randomisation system and entered in the patient’s details, including the unique study PIN.

Once a participant was randomised, the system automatically generated e-mails to key study staff. For example, an e-mail was sent to the local site pharmacy to alert the staff to a participant’s treatment group (treatment 1 or 2). Additional blinded and unblinded e-mails were generated from the randomisation system to notify key trial site staff (e.g. the chief investigator and trial manager) depending on their role in the study.

The randomisation sequence was generated using blocked randomisation, stratified by centre.

Blinding

Inclusion criteria for the double-blind treatment stage, placebo-controlled study

• Adults (aged ≥ 18 years) with a diagnosis of PPP made by a trained dermatologist, with disease of sufficient impact and severity to require systemic therapy.

• Disease of duration of > 6 months and not responding to an adequate trial of topical therapy, including very potent corticosteroids.

• Evidence of active pustulation on palms and/or soles to ensure sufficient baseline disease activity to detect efficacy.

• At least moderate disease on the PPP-IGA.

• In the case of women of child-bearing potential, being on adequate contraception (see Appendix 2, Contraception guidelines, for guidance) and not pregnant or breastfeeding.

• Written informed consent to participate.

Exclusion criteria for the double-blind treatment stage, placebo-controlled study

• Previous treatment with anakinra or other IL-1 antagonists.

• A history of recurrent bacterial, fungal or viral infections that, in the opinion of the principal investigator, presented a risk to the patient.

• Evidence of active infection or latent tuberculosis (TB), human immunodeficiency virus (HIV) positivity or hepatitis B or C seropositivity.

• A history of malignancy of any organ system (other than treated, localised non-melanoma skin cancer), treated or untreated, within the past 5 years.

• Use of therapies with potential or known efficacy in psoriasis during or within the following specified time frame before treatment initiation (week 0, visit 1):

  • very potent topical corticosteroids within 2 weeks

  • topical treatment that is likely to impact signs and symptoms of psoriasis (e.g. corticosteroids, vitamin D analogues, calcineurin inhibitors, retinoids, keratolytics, coal tar, urea) within 2 weeks

  • methotrexate, ciclosporin, acitretin and alitretinoin (Toctino^®, Stiefel, Brentford, UK) within 4 weeks

  • phototherapy or psoralen and ultraviolet A radiation (PUVA) within 4 weeks

  • etanercept (Enbrel^®, Pfizer, New York, NY, USA) or adalimumab (Humira^®; AbbVie Inc., Chicago, IL, USA) within 4 weeks

  • infliximab (Remicade^®, Janssen Biotech Inc., Horsham, PA, USA) or ustekinumab (Stelara^®, Janssen Biotech Inc.) or secukinumab (Cosentyx^®, Novartis Pharmaceuticals UK Ltd, London, UK) within 3 months

  • other tumour necrosis factor (TNF) antagonists within 3 months

  • other immunosuppressive or immunomodulatory therapy within 30 days or five half-lives prior to treatment initiation, whichever was longer

  • any other investigational drugs within 30 days (or 3 months for investigational monoclonal antibodies) or five half-lives prior to treatment initiation, whichever was longer.

• Moderate renal impairment (defined as creatinine clearance of < 50 ml/minute).

• Neutropenia (defined as neutrophil count of < 1.5 × 10⁹/l).

• Thrombocytopenia (defined as platelet count of < 150 × 10⁹/l).

• Known moderate hepatic disease and/or raised hepatic transaminases [alanine aminotransferase (ALT) or aspartate aminotransferase (AST)], more than two times the upper limit of normal (ULN), at baseline. Participants who failed this screening criterion could still be considered following review by a hepatologist and confirmed expert opinion that study entry was clinically appropriate.

• Live vaccinations within 3 months prior to the start of study medication, during the trial and up to 3 months following the last dose.

• In the case of women, pregnancy, breastfeeding or being of child-bearing age and not on adequate contraception and, in the case of men, planning conception.

• Poorly controlled diabetes mellitus, cardiovascular disease, asthma and concomitant therapy that may interact with anakinra (e.g. phenytoin or warfarin) or any condition in which, in the opinion of the investigator, anakinra would a present risk to the patient.

• Unable to give written informed consent.

• Unable to comply with the study visit schedule.

• Diagnosis (or historic diagnosis) of either childhood- or adult-onset Still’s disease.

Inclusion criteria for the open-label extension

• Participation in the double-blind placebo-controlled study.

• Completion past visit 4 (week 8) of the double-blind placebo-controlled study.

• In the case of women of child-bearing potential, being on adequate contraception (see Appendix 2, Contraception guidelines for guidance) and not pregnant or breastfeeding.

• Written informed consent to participate.

Exclusion criteria for the open-label extension

• A history of recurrent bacterial, fungal or viral infections that, in the opinion of the principal investigator, presented a risk to the patient.

• Evidence of active infection or latent TB or of HIV positivity or hepatitis B or C seropositivity (required only for participants who were beyond visit 5, the double-blind treatment stage, of the placebo-controlled study).

• A history of malignancy of any organ system (other than treated, localised non-melanoma skin cancer), treated or untreated, within the past 5 years.

• Use of therapies with potential or known efficacy in psoriasis during or within the following specified time frame before treatment initiation (visit OLE 1):

  • methotrexate, ciclosporin, acitretin or alitretinoin within 4 weeks

  • phototherapy or PUVA within 4 weeks

  • etanercept or adalimumab within 4 weeks

  • infliximab, ustekinumab or secukinumab within 3 months

  • other TNF antagonists within 3 months

  • other immunosuppressive or immunomodulatory therapy within 30 days or 5 half-lives prior to treatment initiation, whichever was longer

  • any other investigational drugs within 30 days (or 3 months for investigational monoclonal antibodies) or 5 half-lives prior to treatment initiation, whichever was longer.

• Moderate renal impairment (defined as creatinine clearance of < 50 ml/minute).

• Neutropenia (defined as neutrophil count of < 1.5 × 10⁹/l).

• Thrombocytopenia (defined as platelet count of < 150 × 10⁹/l).

• Known moderate hepatic disease and/or raised hepatic transaminases (ALT/AST), more than two times the ULN, at baseline. Participants who failed this screening criterion could still be considered following review by a hepatologist and confirmed expert opinion that study entry was clinically appropriate.

• Live vaccinations within 3 months prior to the start of study medication, during the trial and up to 3 months following the last dose.

• In the case of women, pregnancy, breastfeeding or being of child-bearing age and not on adequate contraception or, in the case of men, planning conception.

• Poorly controlled diabetes mellitus, cardiovascular disease and asthma, and concomitant therapy that may interact with anakinra (e.g. phenytoin or warfarin) or any condition in which, in the opinion of the investigator, anakinra would present a risk to the patient.

• Inability to give written informed consent.

• Inability to comply with the study visit schedule.

• Having been previously invited to have the OLE therapy and declined.

• Diagnosis (or historic diagnosis) of either childhood-onset or adult-onset Still’s disease.

Concomitant medication, prohibited medication and rescue therapy information

The list of concomitant medication, prohibited medication and rescue therapy for the double-blind RCT is presented in Appendix 3, Table 36. The list of prohibited medication for the OLE is presented in Appendix 3, Table 37.

Participant pathway (trial procedures)

The participant pathway consisted of four periods: a screening period, a treatment period, a follow-up period and an optional OLE.

The overall study flow is detailed in Figure 1 and the detailed visit schedule is listed in Appendix 2, Tables 33–35.

FIGURE 1.

Trial flow chart. ITT, intention to treat.

The screening period, that is the period between the screening visit (visit 0) and baseline (visit 1), was a minimum of 5 days and a maximum of 3 months, and was used to assess eligibility and to taper off prohibited medicines (as part of the washout period for the study). Participants who failed the screening period (did not satisfy one or more eligibility criteria) had the option to be re-screened if clinically appropriate.

The treatment period (visits 1–4) was 8 weeks. At the start of the treatment period, eligible participants were randomised to receive the intervention (as described above).

The follow-ups (visits 5 and 6) at week 12 and 90 days post last treatment date were used to assess disease relapse off study treatment, to follow up any AEs that had been previously reported and to plan for post-treatment management of the participants’ condition.

If a participant decided to take part in the optional 8-week OLE, there were two possible pathways:

1. Participants who decided to take part in the OLE before or at the week 12 follow-up visit (visit 5) were to begin their 8-week OLE period directly after the week 12 follow-up (i.e. their OLE baseline visit could be on the same day as the week 12 follow-up visit). Their final follow-up visit would take place 90 days after their last dose of anakinra.

2. Participants who were beyond the week 20 follow-up visit (visit 6) may have been receiving another treatment for their PPP when they decided to take part in the OLE. These participants were required to have an OLE screening visit, a possible washout period (as per the study protocol) and an OLE baseline visit arranged once the required washout period was completed. A final follow-up visit was then conducted 90 days after the last dose of anakinra.

To achieve the exploratory objectives (mechanistic studies), all participants were invited to provide biological samples for use in exploratory laboratory tests. These were blood samples taken at visit 0, and longitudinal blood samples were taken at visits 1, 2, 4 and 5.

In addition, participants were invited to provide microbiopsy samples from the skin on the lateral edge of the base of their feet or palms prior to treatment initiation at baseline (visit 1) and then at visit 2 (approximately 1 week later). These samples were used to understand the underlying pathogenesis of pustular psoriasis and the mechanism by which anakinra may work, and to identify potential biomarkers of response.

Participant withdrawal

Participants had the right to withdraw from the study at any time and for any reason. The principal investigator also had the right to withdraw participants from the study drug in the event of intercurrent illness, AEs, serious adverse events (SAEs), suspected unexpected serious adverse reactions (SUSARs) or protocol violation, or for administrative reasons or other pertinent reasons.

Participants had to discontinue the investigational product (and non-investigational product at the discretion of the investigator) in the event of any of the following:

• withdrawal of informed consent (if the participant decided to withdraw for any reason)

• the occurrence of any clinical AE, laboratory abnormality or intercurrent illness that, in the opinion of the investigator, indicated that continued participation in the study was not in the best interest of the participant

• the need, in the principal investigator’s opinion, to administer concomitant medication not permitted by the trial protocol

• pregnancy (in which case the chief investigator was notified immediately).

If a participant decided to withdraw from the trial, all efforts were made to report the reason for withdrawal as thoroughly as possible and participants were encouraged to provide follow-up data for the remaining trial visits, but at a minimum were asked for outcome data and safety data (AE records) at week 8 and at 90 days post last dose follow-up. They were also asked if they were willing to provide trial-specific clinical data (i.e. outcome measures) and/or samples for mechanistic study, as per the remaining trial schedule. All data and samples collected up to the date of withdrawal were retained.

Safety bloods should have been taken as per the trial schedule for all participants and/or as considered appropriate by the principal investigator.

Significant amendments to the study protocol

For a summary of amendments, see Appendix 4, Table 38.

Statistics methodology

Statistical analysis

Trial oversight

Patient and public involvement

Patient and public involvement (PPI) has been central to the APRICOT study throughout its course. In this section, we summarise the different ways in which participants and the public have been involved and have had an impact on the study. A detailed description is provided in Appendix 7.

Reflections/critical perspective

The Psoriasis Association participated regularly in discussions about recruitment strategies and was extremely helpful in advertising the APRICOT study via social media, its magazine and its website. It helped to guide participants, directing any queries to the study website and e-mail address (where interested parties could self-refer). It also helped with the review and amendment of study materials. The PPI could have been enhanced by including more than one formal lay patient representative in the study infrastructure to more accurately reflect the PPP patient population and to enable more diverse conversations and guidance during the trial. A recent PPI event on psoriasis held during the COVID-19 pandemic over Zoom (Zoom Video Communications, San Jose, CA, USA) had more than 120 attendees; the question and answer format worked well, suggesting that virtual formats may be efficient and cost-effective ways of engaging a wider audience that would appeal to participants. We have opted to use this format to disseminate the findings once published and will also utilise the Psoriasis Association and its social medial channels to facilitate this.

Recruitment and participant flow

Between October 2016 and January 2020, a total of 64 eligible participants were enrolled: 33 were randomly allocated to the placebo group and 31 to the anakinra group (see Appendix 6, Tables 40 and 41, and Figure 20). An additional two consenting participants were randomised in error and never received any treatment and are excluded from all analysis. Figure 2 is the CONSORT flow chart for the trial, which summarises the participant flow through the trial. A total of 64 participants received treatment (placebo group, n = 33; anakinra group, n = 31), of whom 62 (placebo group, n = 31; anakinra group, n = 31) attended the week 8 visit and 61 attended the week 12 visit (placebo group, n = 31; anakinra group, n = 30). A total of 14 participants entered the OLE (placebo group, n = 9; anakinra group, n = 5).

FIGURE 2.

The CONSORT flow chart. a, An additional two participants were randomised, making a total of 66 randomised; however, these two participants were randomised in error, as they were ineligible. They were not offered treatment and were immediately withdrawn and excluded from all analysis. b, One participant withdrew in week 1. One participant was lost to follow-up and withdrawn post week 4. c, One participant withdrew at week 8. Note: numbers withdrawn from the trial are cumulative.

Stage 1

Recruitment began in October 2016 and the pre-planned interim stage 1 analysis was performed in January 2018 after the first 24 participants had been followed up for 8 weeks (placebo group, n = 13; anakinra group, n = 11). The unadjusted PP-PASI score averaged over weeks 1–8 was 16.2 points (SD 11.1 points) in the placebo group and 12.9 points (SD 7.9 points) in the anakinra group. The baseline-adjusted treatment group difference in PP-PASI scores averaged over weeks 1–8 was –1.2 points (95% CI –5.5 to 3.1 points), and the point estimate was in favour of anakinra. The unadjusted fresh pustule count averaged over weeks 1–8 was 47.2 points (SD 59.4 points) in the placebo group and 61.6 points (SD 76.9 points) in the anakinra group. The baseline-adjusted treatment group difference in the fresh pustule count averaged over weeks 1–8 was 16.5 points (95% CI –51.0 to 49.6 points), and the point estimate was in favour of placebo. Given that one outcome was in favour of anakinra (PP-PASI score), the trial met the criteria to progress to stage 2.

The mean difference in agreement between the fresh pustule count assessed at sites and the fresh pustule count assessed centrally using photography was 27 pustules (95% CI –131 to 185 pustules). The ICC was used as a measure of agreement for fresh pustule count between the site assessor and the photographic central assessor, and was found to be 0.13 pustules. The mean difference in agreement in PP-PASI scores between the first and the second site assessors was –0.56 points (95% CI –5.9 to 4.8 points). The ICC for fresh pustule count between two independent site assessors was 0.73 pustules. Overall, the DMC decided unanimously to recommend PP-PASI score as the primary outcome variable for stage 2 because this outcome was judged to be more reliable than fresh pustule count; the agreement (ICC) between the two PP-PASI independent assessors was considerably higher (0.73 points) than the agreement between the site and the central photo fresh pustule count assessments (0.13 points). Additional results from stage 1, including the unadjusted standardised mean differences between the treatment groups for the stage 1 outcomes by time point, can be found in Appendix 9, Table 42, and Figures 23 and 24.

Baseline characteristics

Table 1 summarises baseline demographics by randomised group. The baseline characteristics of the placebo and anakinra treatment groups were generally well matched, including demographics and the severity and impact of participants’ disease.

The mean age of the participants was 50.8 years; 84% were female and 92% were of white ethnicity. Current smokers made up 55% of participants and ex-smokers made up 22%. The mean disease severity at baseline, as measured by the PP-PASI, was 17.8 points (SD 10.5 points). The median fresh pustule count including both the palms and soles was 27.0 pustules (IQR 15.0–49.0 pustules) and total pustule count across the palms and soles was 9.0 pustules (IQR 45.0–169.0 pustules).

Withdrawals from treatment and from the study

During the trial, a total of 11 participants (17%) withdrew from study treatment: six from the placebo group and five from the anakinra group (Table 2).

Retention in the study was high (see Figure 2). Only three participants (5%) who withdrew from treatment also withdrew entirely from the study. One participant who withdrew from treatment in the placebo group prior to the end of week 1 did not attend any further follow-up appointments and was withdrawn from the study because of non-compliance with the visit schedule. Two further participants who withdrew from treatment early continued in the trial immediately following treatment cessation, but were later withdrawn: one placebo participant was withdrawn post week 4 prior to week 8 due to loss to follow-up and one anakinra participant was withdrawn at the week 8 visit prior to week 12 due to a wish to start other therapies.

Temporary treatment discontinuations, after which trial treatment was recommenced, occurred more frequently in the anakinra group than in the placebo group. There was a total of nine participants recorded to have temporarily discontinued treatment, with 3 out of 33 (9%) in the placebo group and 6 out of 31 (19%) in the anakinra group (Table 3). Temporary discontinuations were mainly as a result of AEs, except for one discontinuation in the anakinra group from an individual who wanted to start other treatments because their condition was worsening. The larger number of discontinuations in the anakinra group was driven by injection site reactions. The temporary treatment numbers include one participant in the placebo group and one in the anakinra group who later permanently withdrew from treatment and who are also included above.

Adherence to treatment

Adherence to trial treatment was recorded by using the responses to daily text messages [from a short message service (SMS)], self-reporting from participants using a paper trial diary (issued at the baseline visit and checked at each study visit) and verbal self-recall at study visits. For those who used the SMS service (see Table 4), a daily SMS was sent out to enquire about whether or not the participant had administered their dose that day and required a response of ‘yes’. Unfortunately, an operational incident (which was discovered in January 2019 and was rectified during mid-March 2019) meant that the SMS service was not utilised by all participants (six participants were affected during this period). After excluding the known treatment withdrawals using SMS (Table 4), out of a maximum of seven injections per week, the placebo group reported, on average, 5.7 injections per week at both week 1 and week 8, and the anakinra group reported, on average, 5.8 injections per week at week 1 and 5.4 injections per week at week 8.

Table 4 also summarises the average number of injections received weekly, as self-reported at each clinical visit; these adherence data were self-reported using either a paper trial diary or verbal self-recall: it was not possible to separate the two methods of measurement. For those with self-reported adherence data (see Table 4), after excluding withdrawals, an average of 6.5 injections was reported per week in week 1 and 6.4 in week 8 in the placebo group, compared with 6.9 at week 1 and 6.7 at week 8 in the anakinra group. When combining the SMS and self-recalled adherence data (taking the mean weekly adherence where both measurements were reported per participant), reported adherence was similar in both groups, with the average number of injections received per week being 6.3 at week 1 and 6.2 at week 8 in the placebo group and 6.7 at week 1 and 6.4 at week 8 (after excluding withdrawals) in the anakinra group. However, when including the withdrawals data from the participants who were known to receive no injections following treatment withdrawal or during temporary withdrawal periods to get an overall picture of adherence per week (Table 5), the adherence was a little lower in the placebo group than in the anakinra group: an average of 6.1 injections per week at week 1 and 4.8 injections per week at week 8 in the placebo group, compared with 6.7 at week 1 and 5.3 at week 8 in the anakinra group.

Given that the number of participants with data on adherence available varied by week, we calculated an overall level of adherence per participant to further explore the overall levels of adherence across the 8-week treatment period. For each participant, using the obtained weekly data on adherence, an overall measure of total adherence was calculated as the proportion of injections received relative to the total number of injections planned over the 8-week treatment period (total planned = 8 × 7 = 56). See Table 14 for a summary of overall compliance levels based on receiving 50–90% of the planned injections. At least 50% of the total planned injections were received by 79% of the placebo group and 81% of the anakinra group. At least 90% of the total planned injections were received by 61% of the placebo group and 48% of the anakinra group.

Non-randomised treatment use

Rescue therapies are listed in Appendix 3, Rescue therapy. Over the 8-week treatment period, eight participants in the placebo group and 11 in the anakinra group were initiated on rescue therapy. After the treatment period, an additional three participants in the placebo group and one in the anakinra group were initiated on rescue therapy prior to week 12 (Table 6). The proportion starting rescue medication was similar in both groups, but more participants in the anakinra group than in the placebo group received this earlier.

Over the 8-week treatment period, three (9%) participants in the placebo group and three (10%) in the anakinra group were initiated on prohibited therapy. After the treatment period, an additional two participants in the placebo group and two in the anakinra group were initiated on prohibited therapy prior to week 12 (Table 7).

A larger number of ‘other topical treatments’ (excluding topical rescue therapy and prohibited treatments) were used by participants in the anakinra group (n = 12) than by participants in the placebo group (n = 7) over the 8-week treatment period (Table 8). Following the treatment period, an additional one (3%) participant in the placebo group and one (3%) in the anakinra group were recorded as having initiated other topical therapy prior to week 12. A summary of the other topical treatments used is given in Table 8. The use of emollients may not have been accurately reported because it was captured by self-recall on a concomitant medication form. It is likely that the use of emollients reported is an underestimate. Furthermore, the site of use for all other topical treatments, as site of PPP or otherwise (e.g. for injection site reactions), cannot be distinguished for these reported uses. For simplicity, we have included all of the topical treatment uses reported on the concomitant medication data form. Therefore, these data on other topical treatments should be interpreted cautiously.

Loss to follow-up and missing data

Loss to follow-up was low (see Figure 2). Appendix 10, Table 43, summarises the completeness of the primary PP-PASI scores outcome. Only four participants had missing primary end-point data at week 8 (6%), which is lower than the 15% factored into the sample size calculation. This meant that a total of 60 (94%) participants provided week 8 PP-PASI (primary outcome) data.

Primary outcome: Palmoplantar Pustulosis Psoriasis Area Severity Index

Figures 3 and 4 display the individual participant PP-PASI profiles over time by treatment group. Figure 5 and Table 9 summarise the mean PP-PASI score outcome by time point and treatment group, with the unadjusted mean treatment group differences. In both treatment groups, the mean PP-PASI score was lower at week 8 than at baseline. The unadjusted mean difference in PP-PASI scores at week 8 for those in the anakinra group compared with those in the placebo group was –1.4 points (95% CI –6.0 to 3.2 points), for which the point estimate was in favour of anakinra.

FIGURE 3.

Placebo participant PP-PASI profiles over time. The raw PP-PASI values for each participant are plotted on the y-axis against the time point on the x-axis.

FIGURE 4.

Anakinra participant PP-PASI profiles over time. The raw PP-PASI values for each participant are plotted on the y-axis against the time point on the x-axis.

FIGURE 5.

The PP-PASI over the 12-week follow-up period. The unadjusted mean PP-PASI is plotted on the y-axis, against the time point on the x-axis for each treatment group. The error bars represent 95% CIs for the unadjusted treatment group means.

Primary analysis

The primary analysis included a total of 62 participants who had at least one post-baseline follow-up (placebo group, n = 32; anakinra group, n = 30). A mixed linear regression model, including a random intercept for participant and centre, was used to adjust the week 8 treatment group difference for baseline PP-PASI scores. The adjusted mean treatment group difference in the week 8 PP-PASI scores for those in the anakinra group compared with those in the placebo group was –1.65 points (95% CI –4.77 to 1.47 points; p = 0.300). The adjusted mean treatment group difference in the week 4 PP-PASI scores for those in the anakinra group compared with those in the placebo group was –0.72 points (95% CI –3.86 to 2.43 points) and the adjusted mean treatment group difference in the week 1 PP-PASI scores for those in the anakinra group compared with those in the placebo group was 0.22 points (95% CI –2.88 to 3.31 points) (see Appendix 10, Figures 25 and 26).

Sensitivity analysis

A sensitivity analysis was conducted to explore the impact of the missing data on the primary PP-PASI scores outcome. In each treatment group, one participant was missing PP-PASI follow-up data over weeks 1–8; in addition, in each treatment group, one participant was missing week 8 data only and one participant in the placebo group and two in the anakinra group were missing interim week 4 follow-up data (see Appendix 10, Table 43). The primary analysis assumes that the missing responses are MAR conditional on the variables included in the model: treatment group, observed PP-PASI scores at baseline, weeks 1, 4 and 8, and centre. The sensitivity analysis explored the robustness of the primary analysis results to various MNAR assumptions.

The estimated intervention effect (the baseline adjusted mean treatment group difference in PP-PASI scores at week 8) did not change when it was assumed that the rate of change of the PP-PASI score between the observed and the unobserved cases for each week unobserved was 0.039 to 0.39 points worse (a higher PP-PASI score outcome) than that predicted under MAR (Table 10), corresponding to outcomes that were worse by 0–100% of the unadjusted mean rate of change in the PP-PASI scores observed over 8 weeks.

Supplementary analysis

Exploratory analysis for the Palmoplantar Pustulosis Psoriasis Area Severity Index

A total of 63 participants, (placebo group, n = 32; anakinra group, n = 31) were included in the analysis of the primary outcome up to 12 weeks. The adjusted mean treatment group difference in the week 12 PP-PASI scores for those in the anakinra group compared with those in the placebo group was –2.42 points (95% CI –5.97 to 1.13 points; p = 0.182). The results are presented in Figure 6.

FIGURE 6.

The PP-PASI scores over the 12-week follow-up period by treatment group: mixed-model estimates. The baseline adjusted mean PP-PASI score is plotted on the y-axis against the time point on the x-axis for each treatment group. The error bars represent 95% CIs for the adjusted treatment group means, estimated from the primary linear mixed model with week 12 data added in.

Secondary outcomes

Secondary investigator-assessed outcomes

Fresh pustule count (palms and soles)

The mean fresh pustule counts by visit are shown in Figure 7 and are summarised in Table 16. A total of 61 participants, (placebo group, n = 32; anakinra group, n = 29) were included in the analysis of the fresh pustule count. One individual in the anakinra group [identification (ID) = 100029] had a baseline fresh pustule count of 799 and it was known that their pustule counts were measured incorrectly; therefore, this participant was excluded from the analysis. Sensitivity analysis is performed below including this participant. The adjusted mean difference in the week 8 fresh pustule count for those in the placebo group compared with those in the anakinra group was 2.94 pustules (95% CI –26.44 to 32.33 pustules; p = 0.844), where the point estimate is in favour of the placebo group. The adjusted mean difference in the week 4 fresh pustule count for those in the anakinra group compared with those in the placebo group was 9.98 pustules (95% CI –19.40 to 39.36 pustules), and the adjusted mean difference in the week 1 fresh pustule count for those in the anakinra group compared with those in the placebo group was –2.41 pustules (95% CI –31.76 to 26.94 pustules).

FIGURE 7.

Fresh pustule count over the 12-week follow-up period. The unadjusted mean fresh pustule count is plotted on the y-axis against the time point on the x-axis for each treatment group. The error bars represent 95% CIs for the unadjusted treatment group means.

TABLE 16 - Fresh pustule count over time by treatment group

One individual (ID = 100029) in the anakinra group had a baseline fresh pustule count of 799 and it was known that their pustule counts were measured incorrectly; therefore, this participant was excluded from the analysis. Sensitivity analysis is performed including this participant.

Time point	Placebo group (N = 33)		Anakinra group (N = 31)		Total (N)	Unadjusted mean difference, anakinra – placebo (95% CI)
	n	Mean (SD)	n	Mean (SD)
Baseline	33	36.1 (33.1)	30	39.8a (46.3)	63	–
Week 1	30	38.0 (45.8)	29	39.6 (58.2)	59	1.6 (–25.6 to 28.8)
Week 4	31	35.7 (45.5)	28	48.0 (64.3)	59	12.2 (–16.6 to 41)
Week 8	31	36.9 (79.5)	28	42.4 (65.1)	59	5.5 (–32.6 to 43.6)
Week 12	29	36.8 (62.7)	27	30.6 (43.0)	56	–6.2 (–35.2 to 22.8)

In the sensitivity analysis, which included the individual from the anakinra group who had an outlying pustule count at baseline (n = 799), the adjusted mean treatment group difference in the week 8 fresh pustule count for those in the anakinra group compared with those in the placebo group (baseline) was 3.08 pustules (95% CI –27.8 to 34.0 pustules; p = 0.845; n = 62). The adjusted mean difference in the week 4 fresh pustule count for those in the anakinra group compared with those in the placebo group was 2.99 pustules (95% CI –27.9 to 33.9 pustules) and the adjusted mean difference in the week 1 fresh pustule count for those in the anakinra group compared with those in the placebo group was 4.59 pustules (95% CI –26.29 to 35.5 pustules; p = 0.771).

Fresh pustule count: exploratory analysis by palms and soles separately

Fresh pustule counts on the palms are summarised in Table 17. A total of 62 participants, (placebo group, n = 32; anakinra group, n = 30) were included in the analysis of the fresh pustule count on the palms. The adjusted mean treatment group difference in the week 8 fresh pustule count of the palms for those in the anakinra group compared with those in the placebo group was 4.07 pustules (95% CI –5.78 to 13.92 pustules; p = 0.418), where the point estimate was in favour of the placebo group.

TABLE 17 - Fresh pustule count on palms over time by treatment group

One individual in the anakinra group (ID = 100029) had a baseline palm fresh pustule count of 209 and it was known that their pustule counts were measured incorrectly; therefore, this participant was excluded from the analysis. Sensitivity analysis is performed including this participant.

Time point	Placebo group (N = 33)		Anakinra group (N = 31)		Total (N)	Unadjusted mean difference, anakinra – placebo (95% CI)
	n	Mean (SD)	n	Mean (SD)
Baseline	33	10.2 (19.2)	30	10.2a (16.5)	63	–
Week 1	31	11.1 (29.5)	29	10.7 (16.3)	60	–0.4 (–12.8 to 12.0)
Week 4	31	8.9 (20.9)	28	11.7 (25.7)	59	2.8 (–9.4 to 15.0)
Week 8	31	7.0 (14.7)	29	10.8 (19.2)	60	3.9 (–4.9 to 12.7)
Week 12	29	8.5 (21.8)	27	9.6 (22.4)	56	1.1 (–10.7 to 13.0)

In the post hoc sensitivity analysis, which included the one individual from the anakinra group who had an outlying pustule count on the palms (ID 100029, n = 209), the adjusted mean treatment group difference in the week 8 fresh pustule count of the palms for those in the anakinra group compared with those in the placebo group was 1.83 pustules (95% CI –8.88 to 12.54 pustules; p = 0.738).

Fresh pustule counts on the soles are summarised in Table 18. A total of 62 participants (placebo group, n = 32; anakinra group, n = 30) were included in the analysis of the fresh pustule count on the soles.

TABLE 18 - Fresh pustule count on soles over time by treatment group

One individual in the anakinra group (ID = 100029) had a baseline soles fresh pustule count of 590 and it was known that their pustule counts were measured incorrectly; therefore, this participant was excluded from the analysis. Sensitivity analysis is performed including this participant.

Time point	Placebo group (N = 33)		Anakinra group (N = 31)		Total (N)	Unadjusted mean difference, anakinra – placebo (95% CI)
	n	Mean (SD)	n	Mean (SD)
Baseline	33	25.9 (23.4)	30	29.6a (43.2)	63	–
Week 1	31	26.2 (34.4)	29	28.9 (50.2)	60	2.7 (–19.4 to 24.9)
Week 4	31	26.8 (38.1)	28	36.3 (61.1)	59	9.4 (–16.9 to 35.7)
Week 8	31	29.9 (69.1)	28	31.4 (61.2)	59	1.5 (–32.7 to 35.7)
Week 12	29	28.3 (46.0)	28	20.3 (36.0)	57	–8.1 (–30.1 to 13.9)

The adjusted mean treatment group difference in the week 8 fresh pustule count of the soles for those in the anakinra group compared with those in the placebo group was –1.42 pustules (95% CI –27.33 to 24.48 pustules; p = 0.914).

In the post hoc sensitivity analysis, which included the one individual from the anakinra group who had an outlying pustule count on the soles (ID = 100029), the adjusted mean treatment group difference in the week 8 fresh pustule count of the soles for those in the anakinra group compared with those in the placebo group was –0.02 pustules (95% CI –26.7 to 26.69 pustules; p = 0.999).

Total pustule count

The mean total pustule count scores by visit are shown in Figure 8 and are summarised in Table 19. A total of 61 participants (placebo group, n = 32; anakinra group, n = 29) were included in the analysis of the total pustule count. The adjusted mean difference in the week 8 total pustule count for those in the anakinra group compared with those in the placebo group was –30.08 pustules (95% CI –83.20 to 23.05 pustules; p = 0.267), where the point estimate was in favour of the anakinra group.

FIGURE 8.

Total pustule count over the 12-week follow-up period. The unadjusted mean total pustule count is plotted on the y-axis against the time point on the x-axis for each treatment group. The error bars represent 95% CIs for the unadjusted treatment group means.

TABLE 19 - Total pustule count over time by treatment group

One individual in the anakinra group (ID = 100029) had a baseline total pustule count of 1186 and it was known that their pustule counts were measured incorrectly; therefore, this participant was excluded from the analysis. Sensitivity analysis is performed including this participant.

Time point	Placebo group (N = 33)		Anakinra group (N = 31)		Total (N)	Unadjusted mean difference, anakinra – placebo (95% CI)
	n	Mean (SD)	n	Mean (SD)
Baseline	33	116.9 (96.4)	30	154.3a (198.7)	63	–
Week 1	30	111.8 (112.1)	29	149.7 (191.5)	59	37.9 (–43.5 to 119.4)
Week 4	31	104.6 (107.3)	28	144.5 (163.3)	59	39.9 (–31.5 to 111.2)
Week 8	31	114.2 (171.8)	28	111.4 (129.3)	59	–2.8 (–82.7 to 77.2)
Week 12	29	112.4 (155.0)	27	107.2 (107.3)	56	–5.2 (–77.2 to 66.7)

The adjusted mean difference in the week 4 total pustule count for those in the anakinra group compared with those in the placebo group was 15.82 pustules (95% CI –37.26 to 68.89 pustules). The adjusted mean difference in the week 1 total pustule count for those in the anakinra group compared with those in the placebo group was 11.06 pustules (95% CI –41.98 to 64.10 pustules).

In post hoc sensitivity analysis, which included the one individual from the anakinra group who had an outlying total pustule count of 1186 at baseline (ID = 100029), the adjusted mean treatment group difference in the week 8 total pustule count for those in the anakinra group compared with those in the placebo group was –27.98 pustules (95% CI –84.51 to 28.54 pustules; p = 0.332). The adjusted mean difference in the week 4 total pustule count for those in the anakinra group compared with those in the placebo group was 1.73 pustules (95% CI –54.73 to 58.19 pustules). The adjusted mean difference in the week 1 total pustule count for those in the anakinra group compared with those in the placebo group was 23.59 pustules (95% CI –32.85 to 80.04 pustules).

Palmoplantar Pustulosis – Investigator’s Global Assessment

The PPP-IGA ratings over the 12-week follow-up period are summarised in Table 20. A total of 63 participants (placebo group, n = 32; anakinra group, n = 31) were included in the analysis of the PPP-IGA. There was no evidence for statistical superiority in the odds of a higher PPP-IGA rating between the treatment groups at week 8 (OR 0.54, 95% CI 0.13 to 2.19; p = 0.384), for which the point estimate favoured the anakinra group. There was also no evidence for statistical superiority in the odds of a higher PPP-IGA rating between the treatment groups at week 4 (OR 0.64, 95% CI 0.17 to 2.44) and week 1 (OR 1.26, 95% CI 0.33 to 4.78). No participants achieved a clear rating on the PPP-IGA at week 8: 0 out of 28 (0%) in the placebo group compared with 0 out of 30 (0%) in the anakinra group.

TABLE 20 - The PPP-IGA ratings over time by treatment group

PPP-IGA rating	Placebo group	Anakinra group
Baseline
Mild, n (%)	4 (12)	4 (13)
Moderate, n (%)	16 (48)	17 (55)
Severe, n (%)	13 (39)	10 (32)
Median (IQR)	3.0 (3.0–4.0)	3.0 (3.0–4.0)
Week 1
Mild, n (%)	4 (13)	6 (20)
Moderate, n (%)	20 (65)	14 (47)
Severe, n (%)	7 (23)	10 (33)
Median (IQR)	3.0 (3.0–3.0)	3.0 (3.0–4.0)
Week 4
Almost clear, n (%)	0 (0)	1 (4)
Mild, n (%)	7 (23)	6 (21)
Moderate, n (%)	18 (58)	17 (61)
Severe, n (%)	6 (19)	4 (14)
Median (IQR)	3.0 (3.0–3.0)	3.0 (2.5–3.0)
Week 8
Almost clear, n (%)	2 (7)	1 (3)
Mild, n (%)	4 (14)	6 (20)
Moderate, n (%)	12 (43)	17 (57)
Severe, n (%)	10 (36)	6 (20)
Median (IQR)	3.0 (3.0–4.0)	3.0 (3.0–3.0)
Week 12
Almost clear, n (%)	3 (10)	0 (0)
Mild, n (%)	7 (24)	9 (33)
Moderate, n (%)	12 (41)	13 (48)
Severe, n (%)	7 (24)	5 (19)
Median (IQR)	3.0 (2.0–3.0)	3.0 (2.0–3.0)

Time to response (75% reduction in fresh pustule count)

A total of 28 participants achieved a 75% reduction in fresh pustule count compared with their baseline count during the 12-week follow-up period: 15 out of 31 (48%) in the placebo group and 13 out of 30 (43%) in the anakinra group (Figure 9). A total of 61 participants (placebo group, n = 31; anakinra group, n = 30) were included in the analysis of the time to response. There was no evidence of statistical superiority of anakinra in the time to response between the treatments (HR 0.58, 95% CI 0.22 to 1.50; p = 0.263), although the point estimate favoured the anakinra group.

FIGURE 9.

Time to response by treatment group.

Time to relapse (return to baseline fresh pustule count)

A total of 39 participants returned to their baseline fresh pustule count during the 12-week follow-up period: 19 out of 31 (61%) in the placebo group and 20 out of 30 (67%) in the anakinra group (Figure 10). A total of 61 participants (placebo group, n = 31; anakinra group, n = 30) were included in the analysis of the time to relapse. The time to relapse between the treatment was found to be similar (HR 0.94, 95% CI 0.50 to 1.78; p = 0.853).

FIGURE 10.

Time to relapse by treatment group.

Development of disease flare

A disease flare is defined as a > 50% deterioration in PP-PASI score compared with baseline. The denominator for the analysis of development of disease flare includes the number of participants followed up over the full treatment period, that is up to week 8. Four participants in the placebo group experienced disease flare compared with two in the anakinra group [12.9% vs. 6.9%, unadjusted difference in proportions –6.0% (95% CI –20.98% to 8.97%)]. A mixed-logistic regression model was used to adjust the treatment group difference for baseline PP-PASI scores and centre. There was no evidence of statistical superiority in the odds of disease flare for anakinra relative to placebo (OR 0.55, 95% CI 0.08 to 3.71; p = 0.542), with the point estimate in favour of anakinra (i.e. those receiving the anakinra treatment are less likely to develop disease flare).

Plaque type psoriasis

The mean PASI scores by visit are shown in Figure 11 and summarised in Table 21. A total of 38 participants (placebo group, n = 20; anakinra group, n = 18) had plaque psoriasis evident at one or more follow-up visits and were included in the analysis of the PASI. The adjusted mean difference in the week 8 PASI scores for those in the anakinra group compared with those in the placebo group was –0.41 (95% CI –0.96 to 0.15; p = 0.151), for which the point estimate was in favour of the anakinra group. The adjusted mean difference in the week 4 PASI scores for those in the anakinra group compared with those in the placebo group was –0.08 (95% CI –0.61 to 0.45).

FIGURE 11.

The PASI scores over the 12-week follow-up period. The unadjusted mean PASI scores are plotted on the y-axis against the time point on the x-axis for each treatment group. The error bars represent 95% CIs for the unadjusted treatment group means.

TABLE 21 - The PASI scores over time by treatment group

Time point	Placebo group (N = 33)		Anakinra group (N = 31)		Total (N)	Unadjusted mean difference, anakinra – placebo (95% CI)
	n	Mean (SD)	n	Mean (SD)
Baseline	19	2.1 (5.4)	16	1.1 (1.6)	35	–
Week 4	18	0.9 (1.5)	18	0.9 (1.4)	36	–0.1 (–1.1 to 0.9)
Week 8	16	0.8 (1.7)	15	0.9 (1.1)	31	0.0 (–1.0 to 1.1)
Week 12	17	0.9 (1.5)	18	0.4 (0.8)	35	–0.4 (–1.2 to 0.4)

Palmoplantar Pustulosis Psoriasis Area Severity Index-50 and Palmoplantar Pustulosis Psoriasis Area Severity Index-75

In the post hoc analysis, the treatment group difference in PP-PASI-50 and PP-PASI-75 scores at week 8 was explored because these two outcomes were reported in two recently published randomised controlled trials investigating interventions in PPP. 37,38

Five participants in the placebo group experienced PP-PASI-50, compared with six in the anakinra group (16.1% vs. 20.7%; unadjusted difference in proportions –4.6%, 95% CI –15.1% to 24.2%). There was no evidence for statistical superiority in the odds of PP-PASI-50 for those in the anakinra group compared with those in the placebo group (OR 1.68, 95% CI 0.35 to 8.19; p = 0.520), for which the point estimate was in favour of the anakinra group. One placebo group participant experienced PP-PASI-75, compared with zero participants in the anakinra group (3.2% vs. 0.0%; unadjusted difference in proportions –3.2%, 95% CI –9.4% to 3.0%). Owing to the small numbers of events, it was not possible to adjust the treatment effect on PP-PASI-75 by baseline PP-PASI score and centre, as a mixed-logistic regression model did not converge.

Palmoplantar Pustulosis Area Severity Index pustule subscales

In the post hoc analysis, we explored the treatment group difference in the PP-PASI pustule subscale, separately for palms and soles, at week 8 (Table 22). There was no evidence for statistical superiority in the odds of a higher PP-PASI pustule rating on the palms across treatment groups for those in the anakinra group compared with those in the placebo group (OR 2.51, 95% CI 0.56 to 11.28; p = 0.231), for which the point estimate was in favour of placebo. There was no difference in the odds of a higher PP-PASI pustule rating on the soles across treatment groups for those in the anakinra group than for those in the placebo group (OR 1.63, 95% CI 0.49 to 5.46; p = 0.426), for which the point estimate was similarly in favour of placebo (a more severe pustule subscale rating for those in the anakinra group relative to those in the placebo group).

TABLE 22 - The PP-PASI pustule subscale

PP-PASI pustule subscale rating	Number of participants (%)
	Placebo group	Anakinra group
Palm (worst pustule subscale across left and right palm)
Baseline
None	12 (36)	10 (32)
Slight	9 (27)	9 (29)
Moderate	7 (21)	10 (32)
Severe	1 (3)	2 (6)
Very severe	4 (12)	0 (0)
Week 8
None	14 (45)	11 (37)
Slight	10 (32)	9 (30)
Moderate	5 (16)	8 (27)
Severe	2 (6)	2 (7)
Very severe	0 (0)	0 (0)
Sole (worst pustule subscale across left and right sole)
Baseline
None	3 (9)	4 (13)
Slight	5 (15)	2 (6)
Moderate	5 (15)	9 (29)
Severe	12 (36)	13 (42)
Very severe	8 (24)	3 (10)
Week 8
None	3 (10)	2 (7)
Slight	6 (195)	8 (28)
Moderate	11 (35)	8 (28)
Severe	9 (29)	9 (31)
Very severe	2 (6)	2 (7)

Secondary participant-assessed outcomes

Palmoplantar Quality of Life Instrument

The mean Palmoplantar Quality of Life (PP-QoL) scores by visit are shown in Figure 12 and are summarised in Table 24. A total of 62 participants (placebo group, n = 31; anakinra group, n = 31) were included in the analysis of the PP-QoL. The adjusted mean difference in the week 8 PP-QoL scores for those in the anakinra group compared with those in the placebo group was 1.27 (95% CI –3.04 to 5.57; p = 0.564), for which the point estimate was in favour of the placebo group.

FIGURE 12.

The PP-QoL scores over the 12-week follow-up period. The unadjusted mean PP-QoL scores are plotted on the y-axis against the time point on the x-axis for each treatment group. The error bars represent 95% CIs for the unadjusted treatment group means.

TABLE 24 - The PP-QoL scores over time by treatment group

Time point	Placebo group (N = 33)		Anakinra group (N = 31)		Total (N)	Unadjusted mean difference, anakinra – placebo (95% CI)
	n	Mean (SD)	n	Mean (SD)
Baseline	33	46.4 (13.8)	31	45.5 (14.8)	64	–
Week 8	31	40.2 (16.0)	31	41.4 (13.9)	62	1.2 (–6.4 to 8.8)
Week 12	29	40.5 (15.5)	28	40.8 (14.8)	57	0.3 (–7.7 to 8.3)

Dermatology Life Quality Index

The mean DLQI scores by visit are shown in Figure 13 and are summarised in Table 25. A total of 62 participants (placebo group, n = 31; anakinra group, n = 31) were included in the analysis of the DLQI scores. The adjusted mean difference in the week 8 DLQI scores for those in the anakinra group compared with those in the placebo group was 0.52 (95% CI –2.04 to 3.07; p  = 0.692), for which the point estimate was in favour of the placebo group.

FIGURE 13.

Dermatology Life Quality Index scores over the 12-week follow-up period. The unadjusted mean DLQI scores are plotted on the y-axis against the time point on the x-axis for each treatment group. The error bars represent 95% CIs for the unadjusted treatment group means.

TABLE 25 - Dermatology Life Quality Index Scores over time by treatment group

Time point	Placebo group (N = 33)		Anakinra group (N = 31)		Total (N)	Unadjusted mean difference, anakinra – placebo (95% CI)
	n	Mean (SD)	n	Mean (SD)
Baseline	33	13.9 (7.2)	31	15.1 (7.0)	64	–
Week 8	31	10.5 (6.9)	31	12.5 (8.3)	62	2.0 (–1.9 to 5.9)
Week 12	29	11.4 (7.7)	27	12.2 (8.8)	56	0.7 (–3.7 to 5.2)

EuroQol-5 Dimensions, three-level version, utility index

The mean EuroQol-5 Dimensions, three-level version (EQ-5D-3L), scores by visit are shown in Figure 14 and are summarised in Table 26. A total of 62 participants (placebo group, n = 31; anakinra group, n = 31) were included in the analysis of the EQ-5D-3L utility index scores. A mixed-linear regression model was used to adjust the treatment group difference for baseline EQ-5D-3L score and centre. The adjusted mean difference in the week 8 EQ-5D-3L utility index score for those in the anakinra group compared with those in the placebo group was –0.09 (95%CI –0.23 to 0.06; p = 0.227), for which the point estimate was slightly in favour of the placebo group.

FIGURE 14.

The EQ-5D-3L utility index scores over the 12-week follow-up period. The unadjusted mean EQ-5D-3L scores are plotted on the y-axis against the time point on the x-axis for each treatment group. The error bars represent 95% CIs for the unadjusted treatment group means.

TABLE 26 - The EQ-5D-3L utility index scores over time by treatment group

Time point	Placebo group (N = 33)		Anakinra group (N = 31)		Total (N)	Unadjusted mean difference, anakinra – placebo (95% CI)
	n	Mean (SD)	n	Mean (SD)
Baseline	33	0.4 (0.4)	31	0.5 (0.4)	64	–
Week 8	31	0.6 (0.4)	31	0.5 (0.4)	62	0.0 (–0.2 to 0.2)
Week 12	30	0.5 (0.4)	29	0.5 (0.3)	59	0.0 (–0.2 to 0.2)

Treatment acceptability

The number of participants who strongly agreed that the treatment was worthwhile was larger in the anakinra group (n = 12/29, 41%) than in the placebo group (n = 4/14, 14%) (Table 27). The mean difference in the proportion strongly agreeing that the treatment was worthwhile for those in the anakinra group compared with those in the placebo group was 27.1% (95% CI 5.0% to 49.2%).

TABLE 27 - Treatment acceptability

Level of agreement with the statement ‘the treatment was worthwhile’	Number of participants (%)
	Placebo group	Anakinra group
Strongly agree	4 (14)	12 (41)
Agree	8 (29)	7 (24)
Neither agree nor disagree	7 (25)	6 (21)
Disagree	5 (18)	3 (10)
Strongly disagree	4 (14)	1 (3)
Total	28	29

Figures 15 and 16 display the PP-PASI profiles over time for the participants who strongly agreed that the treatment was worthwhile by treatment group, which can be compared with the PP-PASI profiles across all trial participants in Figure 3. One participant who received anakinra, who strongly agreed that the treatment was worthwhile, had the highest baseline PP-PASI score across the trial of 58 points and saw the largest decrease in PP-PASI score to 28.8 points at week 8.

FIGURE 15.

The PP-PASI profiles for the placebo group participants who strongly agreed that the treatment was worthwhile.

FIGURE 16.

The PP-PASI profiles for the anakinra group participants who strongly agreed that the treatment was worthwhile.

Table 28 summarises the overall levels of compliance with the 8-week treatment (proportion of total planned injections received) by treatment group, and strong agreement that the treatment was worthwhile. Among those in the anakinra group receiving at least 90% of the planned total injections over the 8-week treatment period, there was a higher proportion of participants who strongly agreed that the treatment was worthwhile (67%) than those who did not (37%), indicating a potential relationship between treatment adherence and acceptability.

TABLE 28 - Proportions of compliers for ≥ 50% to ≥ 90% planned injections received by treatment acceptability

All individuals in whom compliance data were missing were assumed to be non-compliant in accordance with the APRICOT statistical analysis plan.

Compliancea	Number of participants (%)
	Placebo group		Anakinra group
	Strongly agreed	Did not strongly agree	Strongly agreed	Did not strongly agree
≥ 50% of injections	3 (75)	23 (82)	9 (75)	16 (84)
≥ 60% of injections	3 (75)	21 (75)	9 (75)	15 (79)
≥ 70% of injections	2 (50)	20 (71)	9 (75)	15 (79)
≥ 80% of injections	2 (50)	20 (71)	9 (75)	14 (74)
≥ 90% of injections	2 (50)	18 (64)	8 (67)	7 (37)

Safety outcomes

Safety monitoring

Table 29 summarises the types of AEs by treatment group. A full listing of the non-SAEs and reactions by MedDRA-preferred term and treatment group is given in Appendix 5, Table 39. Figures 17 and 18 summarise the non-SAEs by MedDRA system organ class. There was a notably larger number of participants experiencing general disorders and administration site conditions in the anakinra group than in the placebo group (anakinra group, n = 23, 74%; placebo group, n = 4, 12%).

TABLE 29 - Summary of safety events by type and treatment group

Relatedness to IMP not available.

Cellulitis, c-reactive protein increased and nausea.

Injection site reaction and nasopharyngitis.

Event	Placebo group		Anakinra group		Total
	Participants (n)	Events (n)	Participants (n)	Events (n)	Participants (n)	Events (n)
Total non-serious AEs	26	84	29	114	55	198
AE	24	52	24	66	48	118
AR	10	30	26	48	36	78
Unexpected adverse reaction (subset of adverse reaction)	2	3b	2	2c	4	5
Unclassifiablea	1	2	0	0	1	2
Total SAEs	0	0	0	0	0	0
SAE	0	0	0	0	0	0
SAR	0	0	0	0	0	0
SUSAR (subset of SAR)	0	0	0	0	0	0
Total	26	84	29	114	55	198

FIGURE 17.

Adverse events and reactions by MedDRA System Organ Class. This figure displays the proportions of individuals experiencing each type of event by treatment group on the left-hand side, the relative treatment group difference expressed as relative risk with 95% CI in the middle and numbers of participants experiencing each event and event totals on the right-hand side.

FIGURE 18.

Volcano plot of AEs and reactions by MedDRA System Order Class. In the volcano plot, the x-axis represents the difference in proportions of participants experiencing each category of AE between the treatment groups (placebo and anakinra). Risk difference of < 0 favours placebo. The y-axis represents the p-value from a Fisher’s exact test on a negative log-scale; smaller p-values are situated higher up the y-axis. The centre of each circle indicates the co-ordinates for a particular category of AE and the size of the circle is proportional to the total number of events for both treatment arms combined. AE categories have been labelled where p < 0.2.

Examining the events at the preferred term level reveals that the difference is the result of a larger number of injection site reactions in the anakinra group (see Appendix 5, Table 39); there were 20 injection site reactions among 19 participants (61%) in the anakinra group compared with one injection site reaction (3%) in the placebo group. Injection site reactions led to temporary treatment interruption in 5 out of 31 (16%) of the anakinra group participants (see Table 3); 3 out of 31 (10%) of the anakinra group participants stopped treatment permanently because of injection site reactions (including one participant who had previously temporarily stopped because of these reactions; see Table 2).

Table 30 summarises the number of AEs that related to an infection by treatment group. The total number of events relating to an infection was larger in the anakinra group (37 events) than in the placebo group (23 events). The number of participants who were prescribed medication for an AE relating to an infection was similar across the groups (anakinra group, n = 11; placebo group, n = 9). Table 31 summarises the blood assessments. At week 8, as expected, there was a fall in the total white cell count, driven by a fall in the neutrophil count (which in two participants was > 50% from baseline) and platelet counts in the anakinra group, but no participants experienced a clinically significant change. This is consistent with the known effects on full blood count. Collectively, this AE profile in people with PPP is comparable to that reported for licensed indications, such as rheumatoid arthritis.

TABLE 30 - Numbers prescribed medication for harm events related to an infection by event type and treatment group

Event type	Total number of events related to infection		Prescribed medication
			Events (n)		Participants (n)
	Placebo group	Anakinra group	Placebo group	Anakinra group	Placebo group	Anakinra group
AE	14	24	6	9	6	8
AR	9	13	4	5	3	5
Unclassified (non-serious)	0	0	0	0	0	0
SAE	0	0	0	0	0	0
SAR	0	0	0	0	0	0
Total	23	37	10	14	9	11

TABLE 31 - Blood values

Adjusted for baseline blood value and centre using linear mixed model.

Time	Placebo group		Anakinra group		Total (n)	Unadjusted mean difference, anakinra – placebo (95% CI)	Adjusteda mean difference, anakinra – placebo (95% CI)	p-value
	n	Mean (SD)	n	Mean (SD)
Neutrophil count (× 109/l)
Baseline	32	5.0 (1.7)	29	4.9 (1.5)	61
Week 8	30	5.1 (1.7)	31	4.3 (2.3)	61	–0.8 (–1.8 to 0.3)
Week 8 change	30	0.2 (1.3)	29	–0.7 (2.1)	59	–0.9 (–1.8 to 0.0)	–0.9 (–1.7 to 0.01)	0.053
Total white cell count (× 109/l)
Baseline	32	8.3 (2.0)	29	8.0 (2.1)	61
Week 8	30	8.4 (2.6)	31	7.4 (2.6)	61	–1.0 (–2.3 to 0.3)
Week 8 change	30	0.3 (1.7)	29	–0.7 (2.2)	59	–1.0 (–2.0 to 0.1)	–1.0 (–2.01 to 0.00)	0.051
Haemoglobin (g/l)
Baseline	32	139.7 (8.9)	29	137.7 (10.3)	61
Week 8	30	138.4 (7.1)	31	140.5 (9.2)	61	2.1 (–2.2 to 6.3)
Week 8 change	30	–0.1 (5.1)	29	2.2 (6.4)	59	2.3 (–0.7 to 5.3)	2.0 (–0.6 to 4.7)	0.129
Platelets (× 109/l)
Baseline	32	283.8 (74.3)	29	272.3 (65.6)	61
Week 8	30	279.9 (68.5)	31	254.1 (57.8)	61	–25.8 (–58.2 to 6.7)
Week 8 change	30	2.0 (27.8)	29	–22.2 (33.8)	59	–24.2 (–40.3 to –8.1)	–25.3 (–39.6 to –11.1)	< 0.001
CRP (mg/l)
Baseline	26	5.0 (5.7)	27	6.2 (7.6)	53
Week 8	9	3.2 (2.4)	8	3.1 (2.2)	17	–0.1 (–2.5 to 2.3)
Week 8 change	9	–1.4 (2.0)	7	0.0 (0.6)	16	1.4 (–0.2 to 3.1)	1.05 (–0.5 to 2.6)	0.174

Open-label extension

A total of 14 out of 64 (22%) participants entered the optional OLE: nine placebo and five anakinra participants. Appendix 11, Table 46, summarises the baseline characteristics of all participants in the original double-blind period against those of participants entering the OLE period. Table 32 summarises the outcomes from the 8-week OLE, including by first-time exposure period. Participants entering the OLE had improved disease severity relative to baseline and required no washout period. Only eight participants provided 8-week follow-up data. On average, investigator-assessed outcome improved over the 8-week open-label period; however, these results should be interpreted with caution because there was no control in this phase and the results are based on a small self-selecting group of participants. Furthermore, because the OLE started part-way through the overall trial (in July 2019, when 40 participants had already completed the RCT component), many had already been moved onto other treatments and did not want to participate.

A total of 26 non-serious AEs were recorded over the OLE (see Appendix 11, Table 47). A total of five injection site reactions occurred among five participants (5/14, 36%) in the OLE.

Exploratory objectives: mechanistic studies

Summary findings

This trial assessed the use of an IL-1 receptor antagonist, anakinra, in adult participants with PPP using a novel two-stage adaptive trial design that enabled confirmation of primary outcome within the trial and opportunity for early stopping. In the randomised double-blind phase, there was no evidence of treatment benefit after 8 weeks of anakinra compared with placebo on the primary PP-PASI outcome. Similarly, there was no evidence of statistical superiority of anakinra on the secondary objective of investigator-assessed outcomes or participant-assessed outcomes. In keeping with the known safety profile of anakinra, neutrophil and platelet counts were smaller following treatment, but the difference did not reach clinical significance; there were no SAEs, but there were a larger number of injection site reactions with anakinra relative to placebo.

Interpretation and clinical relevance

Failing to demonstrate efficacy in a trial may be because of a number of factors. First, the study was underpowered. This study size was small (n = 64) and established to detect a large effect size of 0.9 SDs. At baseline, the observed SD for the PP-PASI (n = 64) was 10.5 points; therefore, according to post hoc calculations, 0.9 SDs is approximately equivalent to a change of 9.5 points in the PP-PASI. A standardised effect size was chosen because this was calculated prior to the conformation of the primary outcome for stage 2 and made with input from clinicians and participants. Thus, estimates for some of the secondary outcomes lacked precision. However, we had high follow-up rates and a robust primary outcome demonstrating no benefit, which was supported by secondary outcomes.

Second, the lack of treatment effect may be because of inadequate drug exposure: dose and/or duration. With respect to dose, the (100-mg daily subcutaneous injection) dose of anakinra used in this trial was the same as that used for other licensed indications in adults. Anakinra is licensed for use to treat rheumatoid arthritis and Still's disease (specialist use only) in the UK. 40 Nevertheless, a daily self-administered subcutaneous injection that is often associated with injection site reactions places a significant burden on participants and adherence was variable across participants. Approximately 80% of the anakinra group received ≥ 50% of the total planned daily doses over the 8-week treatment period, but just under half of the participants in the anakinra group received ≥ 90% of the total planned. Notably, the CACE, which estimates the causal effect of treatment for the population of eligible participants who would be able to comply with the treatment schedule, indicated that, compared with the primary ITT effect of –1.65, the treatment effect for an individual who could adhere to at least 90% of total planned injections was just over double, at –3.80, although this finding did not reach statistical significance. The 8-week duration of treatment was selected as being long enough to see an effect on pustules, the expected main target for anakinra, and, based on PPI feedback, the maximum reasonable duration for a placebo-controlled trial requiring daily injections. Nevertheless, the anakinra observed treatment effect (as measured by the PP-PASI) was maintained and slightly increased at 12 weeks (albeit remaining insignificant). Individual participant PP-PASI profiles in the anakinra group over 12 weeks show improvements beyond week 8.

In addition, two randomised trials investigating the effectiveness of guselkumab (a mAb directed against the IL-23 subunit p19) and secukinumab (mAb directed against IL-17) in PPP published during our trial support the notion that treatment duration may be relevant because the treatment benefits for both therapeutic agents, although modest, improved consistently through to 52 weeks (primary outcomes at 16 weeks). 37,38 Thus, taken together (along with the notable finding that a greater proportion of participants in the anakinra group than in the placebo group strongly agreed that the treatment was worthwhile), it seems possible that for people to be able to adhere to treatment, and, if treatment had been given for longer, there may be some treatment benefit with anakinra. If this is the case, have we missed a clinically relevant treatment benefit?

Overall, across primary and secondary outcomes there was no evidence for statistical superiority of anakinra at 8 weeks in the ITT analyses. No clinical differences in usage of rescue or prohibited treatments were observed between the treatment groups, and the analysis to estimate the treatment effect if rescue and prohibited treatment had not been used did not suggest that this was likely to have influenced the observed results. Although the CACE was suggestive that poor adherence may have contributed to the observed lack of efficacy, adherence rates are likely to be even lower in routine clinical practice. Thus, we can confidently conclude from this trial that there is no evidence for clinically relevant benefit with 8 weeks of anakinra for those with PPP. Based on the results of mechanistic studies, it is also reasonable to conclude that genetically determined IL-1 upregulation is not a major disease driver in this condition.

Strengths and limitations

Methodologically this trial has made a number of novel contributions that are potentially generalisable beyond this specific indication. First, as proof-of-concept data and safety information were limited in this rare disease because of the small PPP population, a novel two-stage adaptive trial design with prespecified progression criteria was adopted. 1 This allowed the trial to stop after 24 participants if the results of the interim analysis flagged a concern for safety or if there was no signal for efficacy. A conventional approach to the design of the interim stage would have resulted in a prohibitively large sample size. Second, as there were no validated outcomes to measure disease change for PPP, this design enabled the most reliable outcome with the best distribution properties out of two prespecified candidate outcomes to be used as the primary outcome for the main trial analysis. The PP-PASI was unanimously selected by the independent DMC to be the primary trial outcome following an assessment of reliability and its distributional properties in comparison with the fresh pustule count. There was a large degree of variability in fresh pustule count values between independent site and central assessors from photographs and much less in PP-PASI outcomes between two independent site assessors. Selecting PP-PASI as the primary outcome was supported by the main analysis findings in which the uncertainty around the fresh pustule count treatment effect remained (with wide 95% CIs).

In addition to providing evidence on the efficacy and safety of anakinra, this trial also provides important data on the natural history of PPP and change in disease severity over time. Notably, improvements in outcomes over time were observed in both treatment groups during the trial, a trend observed in other recent placebo-controlled trials of targeted interventions in PPP. 37,38

This might be in part because of a selection bias towards less severe or unstable participants entering the trial: the study was placebo controlled with no guaranteed access to anakinra for the majority of the recruitment period, as well as stipulated washout periods for all interventions with potential or known effectiveness in PPP. The trial cohort was predominantly female, white and current/ex-smokers. However, this is consistent with the observed clinical bias in those with PPP. PPP shows a marked sex bias, with women accounting for 60–90% of affected individuals and is associated with smoking, with up to 90% of participants self-identifying as smokers at the time of diagnosis. 41 Higher prevalence rates have also been reported in white people than in other ethnic groups. 42

Currently validated outcome measures in PPP are lacking, and validation of end points is required. The obtained trial data include measurements of pustule counts, PP-PASI, Investigator Global Assessment and various patient-reported outcome measures. As a result, APRICOT has also provided a valuable source data that will be used in future research to validate outcome measures in PPP. On a related note, it was intriguing that a greater proportion of participants in the anakinra group than in the placebo group strongly agreed that the treatment was worthwhile. We did identify a higher rate of compliance among those who strongly agreed in the anakinra group relative to those who did not strongly agree. But this difference raises the question as to whether or not there are any outcomes that were not measured, such as pain or fatigue, that would be important to explore further in future research.

The results of the OLE should be interpreted cautiously. The OLE was added to the trial primarily to aid recruitment into the trial. Only around 20% of total participants optionally entered this phase of the trial. A greater proportion of participants entering this phase were from the placebo group (27%) relative to the anakinra group (16%). However, overall there was a highly self-selective group entering this phase of the trial, without washout. Moreover, 8-week follow-up data were obtained for only just over half of those entering this phase.

Mechanistic findings

Mechanistic studies provided evidence for potential involvement of IL-1 pathways in a subset of affected individuals. For instance, whole-exome sequencing identified a small number of PPP cases with mutations in IL-1-related genes.

However, because of the small size of the data set and the small number of responders in the trial, it was not possible to draw any definitive conclusions about the genotypes of responders and non-responders.

Bulk RNA sequencing highlighted important differences between the activity of IL-1 in perilesional and lesional skin. Although the immune landscape of the former was dominated by Th1 and Th2 cell activation, the inflammatory milieu of the lesional skin was mostly characterised by neutrophil infiltration. In this context, there was strong evidence for IL-1β activity in lesional skin, whereas other cytokines were dominant in perilesional samples. This suggests that IL-1β is a likely driver of skin pustulation but may be less important for the maintenance of chronic inflammation.

Given that some participants agreed to provide skin samples after treatment initiation, further experiments could be carried out to monitor the expression of IL-1 signature genes at week 1. Differences between responders and non-responders could then be investigated in the light of compliance data.

Finally, the mechanistic studies were accompanied by the recruitment of the PLUM cohort. To our knowledge, this is one of the largest and best-characterised data sets to be held in a single research centre. It has enabled and established a productive partnership with the European Rare And Severe Psoriasis Network (ERASPEN), leading to the publication of two research papers. 41,43 It has also enabled the identification of a novel disease gene that has been described in a recent article. 44

Since the inception of this trial, IL-36 per se has been further validated as a potential therapeutic target in pustular psoriasis, with the first proof-of-concept phase 1 study of the IL-36 receptor antagonist monoclonal antibody BI655130 showing efficacy in GPP in a series of seven participants. 45 As there is only limited in vitro evidence to suggest that an IL-1 blockade may abrogate IL-36 signalling, the fact that we have not demonstrated therapeutic efficacy with anakinra does not preclude IL-36 being of pathogenic importance in localised forms of pustular psoriasis such as PPP. 46

However, although clinical trials have shown that an IL-36 blockade ameliorates the symptoms of GPP, limited efficacy in PPP has recently been shown. 45,47,48 Equally, the biological therapies, particularly those targeting the canonical IL-23/IL-17 pathway, which deliver such impressive clearance rates in plaque psoriasis show only modest benefit, with two recent randomised controlled trials37,38 reporting data for secukinumab and guselkumab, respectively. Thus, our findings alongside the recent published findings in IL-36 suggest a need to determine other drug targets. There are two explanations: (1) both are targeting the wrong pathway or (2) poor drug exposure may be contributing to poor outcomes.

Palmoplantar pustulosis remains an area of high unmet need. We recommend that further research is conducted to (1) identify new drug targets, (2) determine the contributory role of drug exposure (including pharmacokinetics and adherence) and (3) validate outcome measures in PPP.

Conclusion

There was no evidence for clinically relevant benefit with 8 weeks of anakinra in treating PPP and by inference that IL-1 blockade over 8 weeks is not a useful intervention.

General acknowledgements

First and foremost, we would like to thank the participants who participated in APRICOT and made this study possible. They all enthusiastically engaged in the research process, were incredibly charitable with their time and in many cases generously donated mechanistic samples.

A particular thanks to Rosemary Wilson and Angela Pushpa-Rajah, who previously managed the study and were central figures in the trial’s development and operation, respectively.

Our sincere gratitude to all the independent members of the TSC (Professor Edel O’Toole, Professor Hervé Bachelez and Dr Stephen Kelly) and the DMC (Professor Deborah Symmons, Dr Mike Ardern-Jones and Professor Simon Skene), who all provided their time and expertise in helping to construct the study and then made sure that it was run with integrity by helping to interpret all the findings throughout the whole duration of the trial.

A special thanks to Mr David Britten, who was our patient representative on the TSC and who throughout the whole study offered invaluable insight in participant outreach and retention, contributed to vital discussions on strategy and helped to form various trial-related materials. Thanks also to Giselle Folloni, who was a passionate advocate for the study on social media.

An additional thanks to Emma Gray and Aysar Al-Rawi, who were the study clinical research associates and provided a comprehensive data monitoring service and general study support.

Thank you to Sobi: its generous support enabled the study to go ahead and for the OLE to be successfully implemented. Sobi kindly provided anakinra and matched placebo for the study.

Our gratitude also extends to Ange Cape and the Guy’s Hospital Pharmacy Manufacturing Unit for processing, labelling and co-ordinating the distribution of the study IMP.

Thank you to Caroline Murphy and the King’s Clinical Trials Unit staff for their assistance with developing the electronic case report form and for their support in delivering the trial.

Finally, a general thanks goes out to everyone whose efforts and support in the conduct and implementation of APRICOT allowed us to successfully complete the study and, consequently, this final report.

The APRICOT Team acknowledges the support of the National Institute for Health Research Clinical Research Network (NIHR CRN).

The study was developed with support from the UK Dermatology Clinical Trials Network (UK DCTN). The UK DCTN is grateful to the British Association of Dermatologists and the University of Nottingham for financial support of the network.

This study was supported by the UK Clinical Research Collaboration-registered King’s Clinical Trials Unit at King’s Health Partners, which is part-funded by the NIHR Biomedical Research Centre for Mental Health at South London and Maudsley NHS Foundation Trust and King’s College London and the NIHR Evaluation, Trials and Studies Co-ordinating Centre.

Thank you to the BADBIR pharmacovigilance team for providing MedDRA coding for AEs. MedDRA^® terminology is the international medical terminology developed under the auspices of the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH). MedDRA^® trademark is registered by IFPMA on behalf of ICH.

Collaborators

Contributions of authors

Suzie Cro (https://orcid.org/0000-0002-6113-1173) (Research Fellow and Clinical Trials Statistician, Study Statistician) designed the statistical analysis plan; carried out the statistical analyses for the DMC and the final stage 2 analysis; contributed to interpretation of the data; was a member of the TMG and TSC; drafted the report; and reviewed the full report.

Victoria Cornelius (https://orcid.org/0000-0002-0080-1065) (Reader in Medical Statistics and Head of Statistics and Trial Methodology, Study Statistician) designed the novel adaptive study and statistical analysis plan; contributed to the grant application, the interpretation of data and the report; obtained grant funding; carried out the statistical analyses for stage 1; was a member of the TMG and TSC; and reviewed the full report.

Francesca Capon (https://orcid.org/0000-0003-2432-5793) (Reader in Inflammation Genetics, Head of Laboratory) contributed to the grant application, was the head of the central laboratory facility; was the scientific lead; was responsible for design, delivery and interpretation with mechanistic studies; edited reports for publication; and drafted relevant parts of this report.

Jonathan Barker (https://orcid.org/0000-0002-9030-183X) (Professor of Medical Dermatology and Honorary Consultant Dermatologist, Senior Co-investigator) contributed to the grant application and recruitment and provided clinical guidance and expertise with critical input into the mechanistic and trial findings.

David Burden (https://orcid.org/0000-0001-7395-9931) (Consultant Dermatologist) contributed to the grant application; provided clinical guidance and expertise, operational leadership and important contributions to the conception and design of the trial and subsequent acquisition of data; and assisted with the interpretation of findings and critical appraisal of all related publications (including this report).

Christopher Griffiths (https://orcid.org/0000-0001-5371-4427) (Foundation Professor of Dermatology) contributed to the grant application; provided clinical guidance and expertise, operational leadership and important contributions to the conception and design of the trial and subsequent acquisition of data; and assisted with the interpretation of findings and critical appraisal of all related publications (including this report).

Helen Jane Lachmann (https://orcid.org/0000-0001-8378-2498) (Professor in Medicine and Honorary Consultant Nephrologist) contributed to the grant application and to the design and conduct of the study throughout the duration of the trial, and provided clinical guidance and expertise, particularly with the use of anakinra.

Helen McAteer (https://orcid.org/0000-0002-2887-5533) (Chief Executive at Psoriasis Association, Patient Representative) contributed to the grant application and to the design and conduct of the study, including the development of patient-facing literature and material, participation in TMG meetings and communication and dissemination of study progress and results; was the lead PPI representative for the trial and provided guidance.

Prakash Patel (https://orcid.org/0000-0001-9971-0130) (Clinical Trials Manager) was the trial manager for the project; was involved with the implementation, oversight and management of the study; was a member of the DMC and TMG; edited reports for publication; and was responsible for co-authoring and reviewing this report.

Andrew Pink (https://orcid.org/0000-0001-5151-5539) (Consultant Dermatologist, Senior Coinvestigator) was a principal investigator for the study, provided clinical guidance and expertise, and contributed to recruitment.

Nick Reynolds (https://orcid.org/0000-0002-6484-825X) (Professor Dermatology and Consultant Dermatologist) contributed to the grant application and to recruitment; was a principal investigator for the study; provided clinical guidance, expertise, operational leadership, including with mechanistic studies and important contributions to the conception and design of the trial and subsequent acquisition of data; and assisted with the interpretation of findings and critical appraisal of all related publications (including this report).

Richard Warren (https://orcid.org/0000-0002-2918-6481) (Professor of Dermatology and Therapeutics and Honorary Consultant Dermatologist) contributed to the grant application and to recruitment; was a principal investigator for the study; provided clinical guidance, expertise, operational leadership, and important contributions to the conception and design of the trial and subsequent acquisition of data; was a member of the TMG; and assisted with the interpretation of findings and critical appraisal of all related publications (including this report).

Catherine Smith (https://orcid.org/0000-0001-9918-1144) (Professor of Dermatology and Therapeutics and Consultant Dermatologist, Chief Investigator) co-conceived the project and led with the grant application, contributed widely to the design and conduct of the trial including recruitment and ongoing development, was responsible for the overall management and oversight of the study throughout its duration, reviewed data analyses, provided clinical guidance and expertise and a critical review of all reports for publication, and co-authored this report.

Publications

Cornelius V, Wilson R, Cro S, Barker J, Burden D, Griffiths C, et al. A small population, randomised, placebo-controlled trial to determine the efficacy of anakinra in the treatment of pustular psoriasis: study protocol for the APRICOT trial. Trials 2018;19:465.

Cro S, Smith C, Wilson R, Cornelius V. Treatment of pustular psoriasis with anakinra: a statistical analysis plan for stage 1 of an adaptive two-staged randomised placebo-controlled trial. Trials 2018;19:534.

Cro S, Patel P, Barker J, Burden D, Griffiths C, Lachmann H, et al. A randomised placebo controlled trial of anakinra for treating pustular psoriasis: statistical analysis plan for stage two of the APRICOT trial. Trials 2020;21:158.

Twelves S, Mostafa A, Dand N, Burri E, Farkas K, Wilson R, et al. Clinical and genetic differences between pustular psoriasis subtypes. J Allergy Clin Immunol 2019;143:1021–6.

Cro S, Cornelius VR, Pink AE, Wilson R, Pushpa-Rajah A, Patel P, et al. Anakinra for palmoplantar pustulosis: results from a randomized, double-blind, multicentre, two-staged, adaptive placebo-controlled trial (APRICOT). Br J Dermatol 2022;186:245–56.

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration.

Patient data

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it’s important that there are safeguards to make sure that it is stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

Disclaimers

This report presents independent research. The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, the MRC, NETSCC, the EME programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the EME programme or the Department of Health and Social Care.

Contraception guidelines

Women of child-bearing potential were eligible to participate in the study following confirmation of agreement to remain abstinent during the period of IMP dosing and for at least 4 weeks after the last dose. Abstinence was acceptable if it was in line with the preferred and usual lifestyle of the patient. Alternatively, confirmation of the use of single or combined contraceptive methods that resulted in a failure rate of < 1% per year during the IMP dosing and for at least 4 weeks after the last dose of study treatment was also acceptable.

Examples of contraceptive methods with a failure rate of < 1% per year include tubal ligation, male sterilisation, hormonal implants, combined contraceptives (oral/injection) and certain intrauterine devices. Alternatively, two methods (e.g. two barrier methods, such as a condom and a cervical cap) may be used to achieve a failure rate of < 1%. Barrier methods must always be supplemented with use of spermicide.

Topical therapy

Emollient therapy was permitted throughout the trial.

For injection sites

To treat the common side effect of an injection site reaction, the use of topical mild corticosteroid (e.g. hydrocortisone up to 2.5%) or antihistamine cream/ointment was permitted.

For plaque psoriasis

Use of emollients was recommended as the first-line intervention, but mild–moderate topical corticosteroids were permitted at the discretion of the investigator as a second-line intervention for plaques at sites other than the hands and feet. Gloves should have been worn for application.

For palmoplantar pustulosis

Prohibited medication for the initial double-blind treatment stage

Any therapy likely to have efficacy in PPP or psoriasis or to compound the potential immunosuppressive effects of anakinra was prohibited and stipulated washout periods should have been adhered to. If treatment with any of the prohibited treatments was essential, the patient should have notified the study team and they should have been withdrawn from the trial.

Prohibited medication for the open-label extension

Stipulated washout periods should have been adhered to. Concomitant topical treatment (only) was allowed only during the OLE stage. If treatment with any of the prohibited systemic treatments (as indicated in Table 37) was essential, the patient should have notified the study team and they should have been withdrawn from the trial and anakinra should have been discontinued.

Introduction

Pustular psoriasis had been identified as an area of unmet need during the development of the NICE guidelines49 on the management of psoriasis, which had substantial input from participants and the public, with the lack of effective/safe interventions in pustular psoriasis being highlighted as a research gap. In addition, the Psoriasis Association, the largest and most important UK patient organisation for people with psoriasis, had also made a major commitment to this area by funding two PhD (Doctor of Philosophy) studentships at King’s College London investigating disease pathogenesis. The study group, therefore, invited Helen McAteer, Chief Executive of the Psoriasis Association, to partner with the study group to ensure that we effectively engaged with participants and the public in the design, implementation, evaluation and communication of programme of research.

Aim

To develop a trial PPI infrastructure that would:

• advise on study design and ethics issues

• develop patient support materials and questionnaires

• facilitate shared learning and reflection from the study

• advise on the best methods to disseminate research outputs and review articles for publication in the lay press.

Methods

Impact of patient and public involvement on the study results

Discussion and conclusions

The PPI experience in APRICOT was astoundingly excellent and decisive, and enabled the study to recruit to target.

The input gained from the PPI was important in designing the study and also in shaping the trial once running. Fundamental changes to the trial design (e.g. the removal of some visits in substantial amendment 4 and the addition of the OLE as part of substantial amendment 12) were heavily informed by PPI. The same can be said for development of the study-related literature and patient-facing documentation and the amendments made to them throughout the study.

Patient and public involvement was also crucial in the promotion of the trial, which ultimately generated study awareness and helped to enhance recruitment.

Beyond the study, Helen McAteer and the Psoriasis Association have provided useful guidance on how to disseminate the study results and logistical support in doing so.

Reflections/critical perspective

The PPI network had a completely positive effect on the study and there were no negative factors of the involvement of the individuals and their conduct.

One possible suggestion that could have enhanced the PPI would be to seek to include more than one formal lay patient representative to the study infrastructure to reflect the PPP patient population more accurately and to enable more diverse conversations and guidance. A recent PPI event held during the pandemic over Zoom (Zoom Video Communications, San Jose, CA, USA) on psoriasis had more than 120 attendees; the question and answer format worked well, suggesting that virtual formats may be an efficient, and cost-effective, way of engaging a wider audience that would appeal to participants. We are exploring this format for our PPI results dissemination strategy.