Conflict of interest Dr Catherine Smith is invited speaker and receives grant/research support, not related to this study, from Abbott, Janssen-Cilag, Schering-Plough, Serono, Wyeth and Pfizer.
Funding source The authors acknowledge financial support from the Department of Health via the UK National Institute for Health Research (NIHR) Comprehensive Biomedical Research Centre award to Guy’s & St Thomas’ NHS Foundation Trust in partnership with King’s College London and King’s College Hospital NHS Foundation Trust.
Background Adult pityriasis rubra pilaris (PRP) type 1 is a rare chronic papulosquamous disorder with clinical and histological parallels with psoriasis. Treatment is challenging and recent case reports suggest a potential role for tumour necrosis factor (TNF) antagonists.
Objectives Our objective was to systematically review the literature for evidence of efficacy of TNF antagonists in the treatment of adult PRP.
Methods We performed a systematic search of the Cochrane library, EMBASE, Pubmed and MEDLINE databases. We defined diagnosis of PRP, classified clinical response and whether this was clearly attributed to TNF-antagonists. We also reviewed disease, treatment duration and follow up.
Results Sixteen articles were selected for detailed review. From these, 12 articles (13 cases) met our predefined criteria and were included in the systematic review. The authors identified two more cases from their personal archive. A total of 15 evaluable cases were included for analysis. Twelve showed complete response (CR) (80%) to TNF-antagonists with a mean time to maximal response of 5 months. In 10 of the CR cases (83%) this was clearly attributable to TNF antagonist therapy.
Conclusion These data indicate that TNF-antagonists may be of value in treating adult type 1 PRP refractory to other systemic agents but selective reporting bias, together with the lack of standard diagnostic criteria and established spontaneous resolution in PRP, prevent any firm recommendations on their place in management.
Classical adult pityriasis rubra pilaris (PRP) is a rare chronic papulosquamous disorder characterized by follicular hyperkeratotic papules and followed by widespread orange-red erythema with islands of sparing and palmoplantar keratoderma. Its prevalence is approximately 2.5 per million of population, or 1 case per 400 000 of the population1 occurring in all races and with an equal male to female ratio.2 The aetiology remains unclear.
PRP was classified by Griffiths in 1980 based on age and morphology into five clinical entities3 and later, a 6th HIV-related category was added to this classification (Table 1).4 The diagnosis remains primarily clinical as the histological features of PRP are largely non-specific. The eruption typically starts on the upper half of the body and spreads caudally usually progressing to erythroderma with characteristic small islands of sparing. Histopathological features helpful for diagnosis are alternating orthokeratosis and parakeratosis in both vertical and horizontal directions. Other features are hypergranulosis, thick suprapapillary plates, broad rete ridges, narrow dermal papillae and sparse superficial perivascular lymphocytic infiltrate.5 Treatment is challenging and options vary from topical steroids to a combination of systemic potent agents.
Table 1. The classification of pityriasis rubra pilaris
Commonest type (50% of cases). Spreads caudally. The patient is usually erythrodermic with diffuse thickening of the palms and soles. Ectropion often is present.
Long duration. Involves a more ichthyosiform pattern in association with hair loss and areas of eczematous changes.
Similar to type 1 but affects children in the first decade.
Affects children, with sharply demarcated areas of follicular hyperkeratosis and erythema on the knees and elbows. Usually does not progress.
Appears in the first few years of life and is chronic. Characterized by follicular hyperkeratosis, whereas erythema is not prominent feature. The skin on the hands and feet can appear scleroderma-like.
Symmetrical, pruritic eruption composed of erythematous and desquamating follicular papules. Is associated with late-onset acne conglobata.
In this report we systematically review the published literature on the treatment of PRP with TNF-antagonists to highlight the current practices and evidence for efficacy to enable clinicians to draw parallels for future use of TNF antagonists in the treatment of PRP.
Review of the literature
Materials and methods
Identification of articles
A systematic electronic literature search in The Cochrane Library, PUBMED, Medline and Embase was performed from January 1999 until October 2011. We used pityriasis rubra pilaris (MESH term) to identify all relevant articles. There was no limitation as to article language. The chronological limits were based on the fact that TNF- antagonists were introduced after 1999.
We retrieved all cases with definite or likely diagnosis of PRP type 1. A definite diagnosis of adult type PRP meant all three of the following criteria were met: (i) a compatible clinical history; (ii) typical clinical features and/or diagnostic histopathology findings; and (iii) clinical photos consistent with PRP. A likely diagnosis met two out of three of these criteria. If less than two criteria were available in the paper the case was regarded as not assessable because of insufficient evidence.
In the absence of defined treatment response criteria or objective disease severity tools for PRP, we defined and classified clinical response as follows (i) Complete response (CR) was defined by the use of any of the following terms : ‘dramatic or spectacular improvement’, ‘near clearance’, ‘almost clear’ and ‘complete regression’ supported by photographic evidence showing adequate body surface area; (ii) Partial response (PR) definition was based on the physician’s global evaluation of ‘partial clearing or ‘partial response’ and/or photographic evidence showing 50% of skin disease improvement; and (iii) Poor or no response was defined by the use of these terms in the article. Articles which had no sufficient information to classify clinical response were deemed unassessable.
The next aspect we reviewed was whether the clinical response could be clearly assigned to the TNF-antagonists or the presence of other confounders played a significant role. If a person has been on immunomodulatory drugs for over 3 months, at a stable dose and disease severity, when a TNF antagonist is introduced, we assign any subsequent response to the TNF antagonist. On the other hand, if systemic therapy (e.g. acitretin, ciclosporin) has been introduced just prior TNF therapy, any subsequent clinical response could not be clearly assigned to the TNF antagonist.
We reviewed the abstracts of all 172 citations yielded by the electronic search, excluded 156 articles after title and abstract review. Only cases treated with TNF-antagonists and type 1 PRP were included. From the 16 articles retrieved for full text review (total of 25 cases), 12 papers (13 cases) met the predefined criteria and were included in the systematic review. We also added the authors’ two cases (total of 15 cases). Four papers, which included 12 cases, were defined as not assessable and excluded (Fig. 1). Disagreements were resolved by discussion between authors.
Data was abstracted and tabulated in a two tables created for this systematic review. The first table (Table 2) summarizes the demographics including age, gender, diagnosis (definite or likely), disease duration prior treatment with TNF-antagonists, duration of treatment and follow up. The second table (Table 3) summarizes anti-TNF agent used, concomitant oral immunosuppressant, compartment of improvement (erythroderma, scaling and/or palmoplantar keratoderma) clinical response and if this can be clearly assigned to the TNF-antagonist used. Clinical response was subclassified to primary and maximal. By primary we mean the point of initial skin improvement and by maximal the point of best skin response.
Table 2. Table summarising the demographic characteristics including age, gender, diagnosis (definite or likely), disease duration prior treatment with TNF-antagonists, duration of treatment and follow up of the patients included in the review
Mean age in years (range, median)
Mean disease duration prior TNF antagonists in months (range, median)
Mean treatment duration with TNF antagonists in months (range, median)
Mean follow up after initiation of treatment with TNF antagonists in months (range, median)
TNF, tumour necrosis factor.
54.6 (30–65, 59)
Definite (3) Likely (2)
13.8 (1–54, 8)
7.7 (1.5–9, 8.5)
8.9 (97.5–9, 8.5)
55.8 (24–79, 56)
Definite (5) Likely (5)
7.2 (0.5–24, 7)
2.8 (0.5–7.5, 2)
8.2 (2–24, 6)
All cases (15)
55.4 (24–79, 56)
Definite (8) Likely (7)
9.4 (0.5–54, 7)
4.4 (0.5–8.5, 3.5)
8.5 (2–24, 7.5)
Table 3. Table summarizing the treatment, response and compartment of improvement
Concomitant oral immunosuppressant and when introduced
Clinical response clearly assigned to TNF blockade?
Acitretin (50 mg/day, increased to 75 mg/day) week 24 to today. Cyclosporine (3 mg/kg/day) week 8 to 0
Acitretin (50 mg/day) week 8 to today
Acitretin (50 mg/day) week 0–4
No response (14)
MTX (10 mg/week) week 6–22
Erythroderma resolved PPK reduced
Complete resolution of erythema and follicular hyperkeratosis
O’ Kane D.
MTX (15 mg/week reduced to 7.5 mg at week 0) week -8 to week 4
Complete resolution of erythroderma
Acitretin (50 mg/day) week 20 to 2
Erythema and scaling subsided
Resolution of erythema and scale
Etanercept (dose reduced to 25 mg/twice weekly after week 12)
Regression of erythroderma and PPK
Resolution of erythema
Erythema and hyperkeratosis faded
A total of 15 cases were included in the systematic review. These included five women and 10 men, with a median age of 59 years (30–65) and 56 years (24–79) respectively. Eight of the 15 cases had a definite diagnosis of PRP and seven a likely diagnosis. The mean disease duration for all the cases included in this systematic review was 9 months. Out of the 15 cases, 12 showed complete response (80%), two partial and one no response. From the 12 cases that showed complete response, in 10 of them (83%), this was clearly and solely attributable to TNF antagonist therapy. In the rest of the patients they have received methotrexate or acitretin for up to 7 months in one case, which may have contributed to improvement. Half of the cases that showed complete response had a definite diagnosis of PRP. Out of all the cases that showed clinical response (PR or CR) the mean time point for primary response was 5 weeks (1–14 weeks). Out of the cases that showed a complete response the mean time point for maximal response was 5 months (4–52 weeks).
Collectively Infliximab,6–13 Etanercept14,15 and Adalimumab16–18 were used individually or in combination with oral immunosuppressants such as methotrexate and acitretin. A combination of one anti-TNF agent with an oral immunosuppressant (acitretin and methotrexate being the commonest) appears to be the treatment of choice in most cases. However, in 40% of cases (six patients) a TNF-antagonist was used as monotherapy with good results. Infliximab is the most commonly used agent, possibly as the experience with its use for other condition such as inflammatory bowel disease and psoriasis has been established over a longer period of time. Infliximab was administered following the psoriasis protocol at week 0, 2, 6 and every 8 weeks thereafter. The standard dose of 5 mg/kg was used. No severe immediate or long term side effects were reported and treatment was usually well tolerated.
This review attempts to critically assess the reported cases of PRP type 1 treated with TNF-antagonists. It provides some evidence that TNF-antagonists may be of value in treating adult type 1 PRP refractory to immunomodulatory agents. The most evidence exists for infliximab in combination with acitretin, with half of the infliximab receiving cases showing a complete response. However, this may substantially overestimate the benefit of TNF antagonist therapy given the high risk of selective reporting and publication of positive outcomes. It is important to note that all cases were severe/refractory to other therapies. Out of the evaluable cases only one that showed no response was reported.9
There are a number of additional factors that also preclude making definite conclusions about the efficacy or otherwise of TNF antagonists in PRP. Firstly, the diagnosis itself can be difficult. In the cases we reviewed only eight had a definite diagnosis, as defined earlier in the Methods section. Specifically only in seven cases was a skin biopsy taken, with most of the earlier reported cases having no histological evidence of PRP. Secondly, we know that 80% of PRP individuals clear within 3 years2 although a single case of spontaneous resolution after 20 years has been reported.19 There is therefore the possibility that in some of the cases reported, the resolution of disease may not necessarily be attributable to treatment, but rather represent spontaneous resolution. The cases reported were followed up for short period of time, with only two reporting follow-up over 1 year.
Thirdly, the effect of confounders is important. Most cases were treated with oral immunomodulatory agents prior initiation of TNF antagonists and in nine cases (60%) the patients received concomitant oral immunosuppressants.
In conclusion, there is some, but potentially highly biased evidence, that TNF antagonist therapy may be of value in type 1 PRP. Prospective clinical data from well-designed clinical studies of adult type PRP are therefore needed to establish accurately the efficacy of TNF antagonists.