Interventions for chronic palmoplantar pustulosis

Abstract

Background

Chronic palmoplantar pustulosis (PPP) is a chronic inflammatory skin condition characterised by crops of sterile pustules (yellow pus spots) on the palms and soles which erupt repeatedly over months or years. The affected areas tend to become red and scaly; cracks may form and these are often painful. Many different treatments have been used for palmoplantar pustulosis but none is generally accepted as being reliably effective.

Objectives

To assess the effects of treatments for palmoplantar pustulosis, both in reducing disease severity and in maintaining remission once achieved.

Search methods

We searched the Cochrane Skin Group Specialised Register (January 2003), the Cochrane Central Register of Controlled Trials (The Cochrane Library Issue 1, 2003), MEDLINE (1966 to February 2003), EMBASE (1988 to February 2003). We also cross‐checked with the Salford Database of Psoriasis Trials and reference lists of articles. We also contacted authors included trials, members of the Cochrane Skin Group and dermatologists interested in psoriasis.

Selection criteria

Any randomised controlled trial in which patients with chronic palmoplantar pustulosis were randomised to receive one or more interventions.

Data collection and analysis

At least two reviewers independently assessed trial eligibility and quality. Study authors were contacted for additional information. Adverse effects information was collected from the trials.

Main results

Twenty‐three trials involving 724 people were included. There is evidence supporting the use of systemic retinoids (improvement rate difference 44%, 95 CI 28 to 59%), oral PUVA (improvement rate difference 44%, 95 CI 26 to 62%). However, a combination of PUVA and retinoids is better than the individual treatments. The use of topical steroid under hydrocolloid occlusion is beneficial. It would also appear that low dose ciclosporin, tetracycline antibiotics and Grenz Ray Therapy may be useful in treating PPP. Colchicine has a lot of side effects and it is unclear if it is effective and neither was topical PUVA (rate difference of 0.00, 95% CI ‐0.04 to +0.04). There is no evidence to suggest that short‐term treatment with hydroxycarbamide (hydroxyurea) is effective.

Authors' conclusions

Many different interventions were reported to produce "improvement" in PPP. There is, however, no standardised method for assessing response to treatment, and reductions in pustule counts or other empirical semi‐quantitative scoring systems may be of little relevance to the patient. This review has shown that the ideal treatment for PPP remains elusive and that the standards of study design and reporting need to be improved to inform patients and those treating them of the relative merits of the many treatments available to them.

Plain language summary

Treatments for chronic palmoplantar pustulosis (a skin disease where repeated crops of painful yellow pus spots form on the palms and soles)

Chronic palmoplantar pustulosis is a skin disease where repeated crops of painful yellow pus spots form on the palms and soles, often over many years. Many different treatments have been used including topical creams and ointments, drugs by mouth and ultraviolet radiation. The review of trials found that several treatments improve the symptoms of chronic palmoplantar pustulosis, although no treatment was shown to suppress the condition completely. Oral retinoid therapy (acitretin) appears to be helpful at relieving symptoms, particularly if combined with PUVA. Ciclosporin and tetracycline antibiotics can also provide some relief. Topical treatments were generally less helpful. As yet there is no ideal treatment for chronic palmoplantar pustulosis, though oral retinoids, particularly when combined with psoralens and ultraviolet radiation (PUVA), may help

Background

Description of the condition

Definition/biology

Chronic palmoplantar pustulosis (PPP) is a chronic inflammatory skin condition characterised by crops of sterile pustules (yellow pus spots) on the palms and soles which erupt repeatedly over months or years (Reitamo 1993). The affected areas tend to become red and scaly; cracks may form and these are often painful. As it has been thought to be a variant of psoriasis it has historically been known as chronic palmoplantar pustular psoriasis and is still widely known by this name.
 The relationship of PPP to psoriasis vulgaris is, however, not clear cut. Most people with the condition have no evidence of psoriasis elsewhere; the proportion that do varies in different series from 2% to 24% (Enfors 1971). Furthermore, recent genetic studies have failed to find any link between PPP and the major genetic susceptibility locus for psoriasis vulgaris, PSORS1 (Asumalahti 2002). Nevertheless, histologically it closely resembles psoriasis and it does occur in the presence of typical psoriasis elsewhere, or if this is a personal or family history of psoriasis. Many dermatologists are thus happy to regard it as a localised pustular variant of psoriasis. It must be differentiated from acute palmoplantar pustulosis (pustular bacterid) and from other pustular forms of psoriasis affecting the hands and feet (for example, acrodermatitis continua of Hallopeau or acropustulosis).

Causes

A strong association with smoking has been identified (Eriksson 1998; Kubeyinje 1997; Nymann 1996; O'Doherty 1985; Rosen 1988) but there is no evidence that, once the condition has developed, stopping smoking improves this form of psoriasis. PPP may also be associated with an arthritis of the anterior thorax (sternocostoclavicular, manubriosternal or sternocostal joints) (Edlund 1988; Jurik 1988; Le Loet 1991) or with abnormalities of thyroid function (Rosen 1988). A recent study from Sweden suggested that PPP is an autoimmune disease: circulating antibodies against nicotinic acetylcholine receptors of normal palmoplantar skin were detected in the sera of 42% of patients with PPP, whereas none were seen in patients with chronic hand eczema (Hagforsen 2002). There has been little investigation into the role of stress in PPP but in one small study (Saez‐Rodriguez 2002) anxiety scores were found to be higher in patients with PPP than in controls. The authors suggested that stress and exacerbations of PPP may be related.

Natural history

PPP tends to run a prolonged course with intermittent exacerbations followed by partial but rarely complete remissions. It may cause minimal symptoms but often represents a debilitating illness which interferes not only with work but also with everyday activities involving the hands and with mobility. It may also be socially stigmatising. A burning discomfort is common. Intolerable itch may occur. The affected areas have a tendency to crack and fissure and this may increase discomfort. Once PPP has developed it may persist for decades.

Description of the intervention

Therapy is unsatisfactory (Camp 1998). Topical treatments tend to be ineffective though some patients with mild disease may obtain relief from emollient creams or ointments alone. Many different systemic therapies (treatments administered remote from their site of action,usually by mouth or by injection) have been tried. None can reliably suppress the condition and none is curative. Interventions which have been advocated for palmoplantar pustulosis include coal tar, topical corticosteroids with or without polythene occlusion, photochemotherapy, oral tetracyclines, methotrexate, systemic retinoids and ciclosporin. Of these, dermatologists appear in general to have found topical psoralen photochemotherapy and systemic retinoids of most value (Camp 1998).

Why it is important to do this review

PPP is a disabling condition for sufferers and a frustrating condition for dermatologists to manage. Many treatments are used and some are potentially toxic. There are significant cost differences between different treatment methods but there is little information regarding their relative efficacy. The current review is intended to address these questions and, where appropriate, point to areas requiring further research.

Objectives

Primary

• To assess the effects of interventions for PPP in reducing disease severity and maintaining remissions

Secondary

• To examine the implications for dermatological practice of any recommendations made.

• To identify what further studies would be required to enable dermatologists to make rational choices between existing treatments.

The protocol also included an objective regarding healthcare costs but we were unable to find any data.

Methods

Criteria for considering studies for this review

Types of studies

Any randomised controlled trial in which patients with chronic PPP were randomised to receive one or more interventions.

Types of participants

Patients of any age who have been given a diagnosis of chronic PPP or chronic palmoplantar pustular psoriasis*. The clinical diagnosis was accepted if it was made by a dermatologist or a physician with an interest in skin disease.

* excludes acute PPP, acute pustular bacterid, acropustulosis, acrodermatitis continua of Hallopeau, pustular acrodermatitis, acrodermatitis perstans, all of which may be regarded as pustular forms of psoriasis affecting the hands and feet

Types of interventions

Reducing disease severity

1. Any topical therapy (including topical photochemotherapy) versus placebo or no treatment

2. Any systemic therapy (including systemic photochemotherapy) versus placebo or no treatment

3. Comparison of two or more topical therapies

4. Comparison of two or more systemic therapies

5. Comparison of systemic therapies with topical therapies

Maintaining reduction achieved

1. Any topical therapy (including topical photochemotherapy) versus placebo or no treatment

2. Any systemic therapy (including systemic photochemotherapy) versus placebo or no treatment

3. Comparison of two or more topical therapies

4. Comparison of two or more systemic therapies

5. Comparison of systemic therapies with topical therapies

Types of outcome measures

Primary outcomes

(a) Improvement in disease severity as assessed by objective severity assessment, preferably as measured by reduction in objective measure of disease severity (for example, predefined semi‐quantitative disease severity score).

(b) Clearance of disease as assessed by objective severity assessment, preferably as measured as an objective measure of disease severity (for example, predefined semi‐quantitative disease severity score).

(c) Patient satisfaction and quality of life scores: Improvement in patient satisfaction measures and quality of life assessment measures over the time course of the intervention.

Secondary outcomes

(a) Maintenance of reduction of disease severity from baseline as assessed by objective measure of disease severity (for example, predefined semi‐quantitative disease severity score).

(b) Maintenance of patients' satisfaction and quality of life scores.

(c) Relapse rates as measured by proportion of patients relapsing to baseline scores during continued treatment or following discontinuation of treatment.

Tertiary outcome measures

(a) Adverse events and side effects

Search methods for identification of studies

Electronic searches

We searched:

(a) The Cochrane Skin Group Specialised Register (March 2005) using the terms in Appendix 1.

(b)The Cochrane Central Register of Controlled Trials (CENTRAL) The Cochrane Library Issue 1,2005 using the terms in Appendix 2.

(c) MEDLINE (OVID) (1966 to March 2005) using the search strategy in Appendix 3.

(d) EMBASE (2003 to March 2005) using the search strategy in Appendix 4.

(e) Salford Database of Psoriasis Trials please see Appendix 5

Searching other resources

References from published studies

We scrutinised references in retrieved articles for evidence of additional studies.

Unpublished literature

We approached the authors of relevant trials, members of the Cochrane Skin Group and dermatologists interested in psoriasis and asked whether they were aware of any other relevant data.

Adverse events

We did not undertake a separate search for adverse events. We described adverse events reported in the included RCTs but did not analyse the data.

Language

We did not impose any language restrictions.

Data collection and analysis

Translation

We did not retrieve any non‐English language articles in this review.

Selection of studies

At least two authors (AMM, RJGC) examined all identified citations. If it was felt that the citation referred to a clinical trial of PPP, we obtained reprints and scrutinised these to determine whether the diagnostic criteria, randomisation methods and outcome measurements enabled a satisfactory comparison of interventions. If these details were unclear, we wrote to authors requesting further clarification. Non‐randomised clinical studies were excluded. All relevant studies were evaluated using Cochrane Collaboration guidelines as set out in the Cochrane Reviewers' Handbook and two members of the Review Group independently assessed these before being included in the review. Excluded studies and reasons for exclusion are listed in the 'Characteristics of excluded studies' table.

Data extraction and management

Two authors (AMM, RJGC) extracted data and this was then checked by a third (SH). If the details were unclear or were missing we attempted to contact the authors but did not receive any replies. Since a variety of scoring methods were used across the eligible trials, we elected to analyse dichotomous outcome data where possible. We examined each RCT to see whether this outcome data was reported or whether the outcomes could be transformed into dichotomous data (e.g. "improved/not improved" or "cleared/not cleared").

No attempt was made to extract or analyse continuous data, such as pustule counts, unless no other information was reported. If it was felt that the trials were sufficiently homogeneous in design, methodology and outcome, these data were pooled.

Assessment of risk of bias in included studies

We based a subjective assessment of methodological quality on the following four criteria, where each component was categorised as adequate, unclear, or inadequate:
 (a) concealment of allocation (we deemed adequate in any sequence where the treatment assignment cannot be foreseen);
 (b) method of generation of randomisation sequences;
 (c) blinding of outcome assessor, participant and care‐giver (we deemed adequate when outcome assessor, participant and/or care‐giver are unaware of the treatment assignment);
 (d) comparability of treatment groups at baseline.

We have also reported the handling of losses and attrition bias, and the period of the trials included in the review. A description of the methodological quality of each study is given based on these components

Measures of treatment effect

Rate differences with 95% confidence intervals (CI) were used as the main measure of effect to allow for comprehensibility of the results. We have also, however, reported rate ratios. We elected to use the term rate rather than risk, because we were interested in positive outcome measures. Newcombe's method was used to calculate confidence intervals for paired proportions from within‐patient designs (Newcombe 1998), since insufficient data could be extracted from the publications to allow for conditional effect measures to be calculated. However, the data were included in graphs when the confidence intervals obtained were very similar to those obtained with the correct paired analysis. We analysed separately and did not pool parallel group and within‐patient comparisons. We reported continuous data from within a single trial as mean differences with 95% CI.

Dealing with missing data

We examined the studies to ensure that intention‐to‐treat (ITT) analysis was used in all studies; the texts were scrutinised for drop‐outs and results were adjusted if necessary to count these as treatment failures. Whilst by convention it is assumed that participants who drop out of trials have not responded, it may be that participants drop out because their condition has cleared and they decide not to return to their physician. The numbers of drop‐outs involved were small and we did not undertake a sensitivity analysis. Within the treatment groups themselves, study designs were broadly similar (thus all phototherapy trials were within‐patient whereas retinoid comparisons were parallel group) and it was felt that it was reasonable to pool data and subject them to meta‐analyses. We used a random effects model for all comparisons of pooled data.

Subgroup analysis and investigation of heterogeneity

When there was statistical heterogeneity present, we did not pool data. Although different tetracycline antibiotics were used in the two studies included in the review, it was felt that the compounds were sufficiently similar to justify pooling the data. We did not perform any subgroup analyses as there was insufficient data.

Results

Description of studies

Results of the search

Twenty of the 23 reports included in the review were identified from the Cochrane Central Register of Controlled Trials. Our search of MEDLINE and EMBASE yielded 215 and 119 possible trials respectively and identified a further three reports. No further reports were identified by our validation checks.

Included studies

The 23 trials which were included in the review, involving 724 people, contained 27 randomised comparisons which fell into nine broad categories (seeResults section). Some reports contained more than one comparison. The majority of included RCTs were of retinoid therapy, photochemotherapy or combinations of both. We also found studies examining Grenz ray therapy, ciclosporin, tetracycline antibiotics, hydroxycarbamide (hydroxyurea), colchicine and topical corticosteroids. Within these therapeutic categories, it was possible to make comparisons and pool data as the populations, disease parameters and treatment methods appeared to be broadly similar. Outcome measures were very variable. Most studies used some semi‐quantitative method of scoring based on degrees of erythema, scaling, fissuring and pustulation as assessed objectively. Only one trial (Bhushan 2001) gave useful information from the patients' own assessments of treatment efficacy.

Excluded studies

Please see Characteristics of excluded studies.

Risk of bias in included studies

In general, we felt the methodological quality of trials was poor. Only two trials, both of non‐standard interventions, liarozole (Bhushan 2001) and Grenz ray therapy ( Lindeloef 1990) satisfied all four pre‐specified criteria and a further one trial (Ward 1976) satisfied three of the criteria. No trial examining standard interventions was considered by these criteria to be of high quality. Ten of the trials (Erkko 1998; Jansen 1978; Lassus 1983; Lassus 1988; Lawrence 1984; Layton 1991; Reitamo 1993; Rosen 1987; White 1985; White 1986) satisfied two of the four criteria and could be said to be of intermediate quality.

Allocation

In order to minimise bias, randomisation and concealment of allocation are essential components of a RCT. However, only five of the trials clearly demonstrated that concealment of allocation was adequate (Bhushan 2001; Erkko 1998; Jansen 1978; Lassus 1983; Ward 1976); and only four trials reported using a satisfactory randomisation method (Bhushan 2001; Lindeloef 1990; Ward 1976; White 1985). One study used a quasi‐randomised method according to date of birth (Rosen 1987) which was an unsatisfactory method. The other studies did not clearly state which randomisation method was used.

Blinding

Blinding of outcome assessment and detection bias

Double‐blinding was felt to be satisfactory in 17 of the 26 trials where either the outcome assessor, participant and/or care‐giver were unaware of the treatment assignment; although it should be noted that with some treatments (notably retinoids) the effects of treatment would be difficult to disguise.

Incomplete outcome data

Handling of losses and attrition bias

We decided to analyse all of the data based on the ITT principle. All studies presented data in such a way that an ITT analysis could be performed, but it was disappointing that many summarised the data on the basis of numbers finishing a trial.

Other potential sources of bias

Trial periods

Chronic palmoplantar pustulosis can persist for decades once established but during this time there may be periods of remission. It would appear sensible therefore to assess any intervention after a reasonable time period. Most of the trials used study periods of 8‐12 weeks (Bhushan 2001; Foged 1983; Lassus 1985; Lassus 1988; Layton 1991; Matsunami 1990; Rosen 1987; Ward 1976; White 1985; White 1986) which can be considered barely adequate. The five studies which lasted 4 months or more were Erkko 1998 (12 months), Jansen 1978 (4 months), Lassus 1983 (6 months), Lawrence 1984 (24 weeks) and Nielsen 1995 (12 months). Therefore we grouped data into initial response(<=16 weeks) and maintenance of remission. We analysed remission at three months and at six months.

Effects of interventions

(1) Retinoids

Of the eight trials examining the use of systemic retinoids, six compared etretinate with placebo, one (Lassus 1988) compared acitretin with etretinate and one (Bhushan 2001) compared liarozole with placebo (Analyses 1 to 3).

(a) Induction of remission

The pooled data from four trials compared the induction of remission by etretinate with placebo (Analysis 1.1 and Analysis 1.2). Forty‐eight of 67 (72%) patients who received retinoids as compared with 17 of the 60 (28%) patients who received placebo showed some improvement (rate difference 0.44, 95% CI 0.28 to 0.59; rate ratio 2.54, 95% CI 1.65 to 3.91). When using good or excellent response as an outcome measure, 26 of 67 (39%) patients who received retinoids as compared with 10 of 60 (17%) who received placebo achieved a good or excellent response (rate difference 0.22, 95% CI 0.07 to 0.36; rate ratio 2.31, 95% CI 1.20 to 4.46). The treatment period for these trials varied from 8 to 16 weeks and assessment of outcome was at the end of treatment period.

1.1. Analysis.

Comparison 1 retinoid versus placebo, Outcome 1 improvement.

1.2. Analysis.

Comparison 1 retinoid versus placebo, Outcome 2 good or excellent response.

(b) Maintenance of remission

Two trials comparing etretinate with placebo to maintain remission were included (Analysis 1.3 and Analysis 1.4). Thirteen of 21 (62%) who received retinoids maintained clinical remission at three months as compared with 5 of 24 (21%) who received placebo (rate difference 0.42, 95% CI 0.16 to 0.68; rate ratio 3.06, 95% CI 1.29 to 7.27). Lassus 1983 noted that side effects were seen frequently in patients receiving etretinate and that this led to discontinuation of treatment in 20% of them. Treatment periods were 6 months (Lassus 1983) and 12 weeks (White 1986) following a 4 to 16 week period of high dose retinoids during which remission was induced.

1.3. Analysis.

Comparison 1 retinoid versus placebo, Outcome 3 in remission at three months.

1.4. Analysis.

Comparison 1 retinoid versus placebo, Outcome 4 in remission at six months.

(c) Retinoids compared

Lassus 1988 found no difference in efficacy between etretinate and acitretin, as measured by reduction in the number of pustules (mean difference 3.24 95% CI ‐29.13 to 35.61; Analysis 2.1).

2.1. Analysis.

Comparison 2 retinoids compared, Outcome 1 reduction in pustules.

Note: Retinoids are teratogenic (may induce birth defects) and must therefore be strictly avoided in early pregnancy. As there is concern that etretinate, the first systemic retinoid to be released, is retained in fat stores for prolonged periods after discontinuation of therapy, a safer alternative was sought. The active moiety of etretinate was thought to be its hydrolysis product, acitretin, which is much more rapidly eliminated from the body than the parent compound. As a result of comparative clinical studies both in psoriasis vulgaris and in PPP showing little difference in efficacy and because of the presumed greater safety of acitretin, etretinate was withdrawn from the market. It was subsequently shown, however, that esterification of acitretin to etretinate may occur in vivo, particularly in the presence of alcohol. The disadvantages of systemic retinoid therapy for women of child‐bearing potential therefore remain. It seems reasonable to conclude that the effects of the two drugs in PPP are broadly similar and that the conclusions from trials of etretinate may be extrapolated to acitretin.

(d) Induction of remission using liarozole

Bhushan 2001 conducted a small study showing marked improvement in four out of seven patients receiving liarozole compared with one out of eight patients receiving placebo (rate difference 0.45, 95% CI 0.01 to 0.88; rate ratio 4.57, 95% CI 0.66 to 3.89; Analysis 3.1). The study incorporated analysable data for individual patients' own assessments of efficacy: five out of seven patients receiving liarozole and three out of eight receiving placebo indicated some possible improvement of symptoms at the end of the intervention period (change in visual analogue score, rate difference 0.34, 95% CI ‐0.13 to 0.81; rate ratio 1.90, 95% CI 0.69 to 5.23), although this did not reach statistical significance.

3.1. Analysis.

Comparison 3 liarozole versus placebo, Outcome 1 induction of remission.

Note: Liarozole is an imidazole‐like compound which inhibits the cytochrome P450‐mediated breakdown of all‐trans retinoic acid (RA). This inhibition causes elevation of plasma and cutaneous levels of RA and thus liarozole acts in a similar way to retinoid drugs. It is sometimes termed a "retinoid mimetic" for this reason. A commercial decision was taken not to market liarozole.

Conclusion: Established but modest efficacy of retinoids when compared to placebo.

(2) PUVA (psoralen plus ultraviolet A radiation)

Four studies compared PUVA with placebo, one compared topical PUVA with systemic PUVA, two compared PUVA with etretinate and one compared short term PUVA with no intervention after induction of remission with steroid under occlusion (Analyses 4 to 8).

(a) Oral PUVA

Two within‐patient studies of oral PUVA compared treated versus untreated hands and/or feet. Murray 1980 showed that 22 of 22 (100%) sides improved with oral PUVA compared to 13 of 22 (59%) sides receiving placebo. Rosen 1987 showed improvement in 9 of 14 (64%) and 2 of 14 (14%) sides with oral PUVA and placebo respectively (Analysis 4.1). Pooling these data gives an overall estimate of rate difference for improvement of 0.44 (95% CI 0.26 to 0.62; Analysis 4.1) and a rate ratio for improvement of 2.07 (95% CI 1.41 to 3.02).

4.1. Analysis.

Comparison 4 PUVA versus placebo, Outcome 1 improvement.

When using clearance as the outcome measure, Murray 1980 found 12 of 22 (55%) sides cleared with oral PUVA compared with 0 of 22 (0%) sides receiving placebo; Rosen 1987 found that 3 of 14 (21%) sides responded to PUVA compared with 0 of 14 (0%) sides receiving placebo (pooled rate difference for the two studies 0.42, 95% CI 0.25 to 0.58, Analysis 4.2).

4.2. Analysis.

Comparison 4 PUVA versus placebo, Outcome 2 clearance.

(b) Topical PUVA

One within‐patient study (Layton 1991) of topical PUVA compared treated versus untreated areas of the hands and/or feet in 27 patients with 44 matched areas. Layton 1991 showed 30 of the 44 areas (68%) receiving topical PUVA improved as compared with 29 of the 44 areas (66%) receiving placebo (rate difference 0.02, 95% CI ‐0.17 to 0.22; rate ratio 1.03, 95% CI 0.77 to 1.39, Analysis 4.1).

When using clearance as the outcome measure, Layton 1991 found none of the 44 areas receiving topical PUVA and none of the 44 areas receiving placebo cleared (rate difference 0.00, 95% CI ‐0.04 to +0.04). In the within‐patient study carried out by Matsunami 1990, 1 of 10 (10%) sides receiving topical PUVA cleared compared to 0 of 10 sides receiving placebo (rate difference 0.1, 95% CI ‐0.19 to +0.40). Pooling these gives an overall estimate of rate difference of 0.00 (95% CI ‐0.04 to +0.04, Analysis 4.2).

Note: These confidence intervals have been correctly calculated using the paired data but are very similar to those calculated by RevMan and shown in the graphs, which are not based on paired analysis.

Topical and oral PUVA data from the four trials were not pooled in the final analysis as the results are very different.

(c) Topical PUVA versus systemic PUVA

Lassus 1985 compared topical PUVA with systemic PUVA . The study in general showed lower response rates for all PUVA modalities than other studies. Four of 33 cleared completely with topical 8‐MOP cream and UV‐A, 0 of 18 cleared completely with trioxalen soak PUVA (Analysis 6.1), and 0 of 13 cleared completely with systemic PUVA (Analysis 5.1). It was not possible to calculate the proportions of patients in each group showing improvement. Overall, 4 of 51 (8%) cleared with topical PUVA compared with 0 of 13 (0%) who received systemic PUVA (rate difference 0.08, 95% CI ‐0.05 to 0.21; rate ratio 2.42, 95% CI 0.14 to 42.38, Analysis 5.1).

6.1. Analysis.

Comparison 6 topical 8‐MOP cream PUVA versus trioxsalen soak PUVA, Outcome 1 clearance.

5.1. Analysis.

Comparison 5 topical PUVA versus systemic PUVA, Outcome 1 clearance.

(d) PUVA for maintenance following induction of remission by topical corticosteroids

Nielsen 1995 found that, following induction of remission with topical corticosteroid under occlusion, three out of nine patients receiving short term PUVA did not relapse within one year compared with 6 out of 13 patients who did not receive further intervention (rate difference 0.13, 95% CI ‐0.54 to 0.28; rate ratio 0.72, 95% CI 0.24 to 2.16,Analysis 7.1).

7.1. Analysis.

Comparison 7 PUVA versus no intervention after induction of remission with steroid plus occlusion, Outcome 1 no relapse within one year.

(e) PUVA versus retinoids

Two studies compared PUVA with etretinate (Lassus 1985; Rosen 1987). Lassus used topical and systemic PUVA and Rosen used oral PUVA. The study by Rosen showed no benefit of retinoids over PUVA but the data from Lassus, who had generally disappointing results with PUVA, appears to demonstrate an advantage with retinoids (Analysis 8.1 and Analysis 8.2). As the results are so different, the data were not pooled.

8.1. Analysis.

Comparison 8 retinoid versus PUVA, Outcome 1 good or excellent response.

8.2. Analysis.

Comparison 8 retinoid versus PUVA, Outcome 2 clearance.

Discrepancies in the conclusions reached by these studies do not obviously appear to be due to differences in the populations studied. The failure of the studies by Layton 1991 and Matsunami 1990 to show any benefit of topical PUVA over placebo is consistent with the findings in the study of Lassus 1985 where the apparently best treatment, topical 8‐MOP phototherapy, produced only a 12% clearance rate with confidence intervals crossing the line of no difference and where topical PUVA overall (8‐MOP or trioxsalen bath PUVA) cleared only 4 of 51 people treated. What did appear to be different in Lassus' study, however, was the failure of any patient to improve with systemic PUVA, in contrast to the clearance rates achieved in the studies of Murray 1980 and Rosen 1987 where 39% patients cleared with systemic PUVA.

Conclusions: Established but modest efficacy of oral PUVA when compared to placebo. No proven benefit of topical PUVA. No definite benefit of retinoids over PUVA or vice versa. No proven advantage of using short‐term PUVA after induction of remission with topical steroid under occlusion.

(3) Retinoids and PUVA combined

Three studies examined the use of retinoids combined with PUVA (Re‐PUVA). The design of two of these was complex in that patients were randomised into two parallel groups (oral retinoid or placebo) and then every patient received PUVA to one hand, foot or side (hand and foot) only (Matsunami 1990; Rosen 1987 ‐ Analysis 9.1 and Analysis 10.2). In the third study (Lawrence 1984) all patients received PUVA but were similarly randomised into retinoid and placebo groups (Analysis 9.1).

9.1. Analysis.

Comparison 9 retinoid +PUVA versus PUVA alone, Outcome 1 clearance.

10.2. Analysis.

Comparison 10 retinoid + PUVA versus retinoid alone, Outcome 2 clearance.

(a) Re‐PUVA versus PUVA

Matsunami 1990 reported that six of 10 (60%) sides cleared with topical PUVA‐etretinate compared with 1 of 10 (10%) sides with topical PUVA alone (rate difference 0.50, 95% CI 0.14 to 0.86; rate ratio 6.00, 95% CI 0.87 to 41.22, Analysis 9.1). Lawrence 1984 found that all eight patients treated with PUVA‐etretinate cleared compared with five out of nine patients treated with PUVA alone and Rosen 1987 found that the PUVA‐treated sites cleared in 14 of 23 patients randomised to etretinate compared with 3 of 14 patients randomised to placebo. If these data are combined then 22 of 31(71%) cleared with oral PUVA‐etretinate compared with 8 of 23 (35%) with oral PUVA alone (rate difference 0.41, 95% CI 0.18 to 0.64 (Analysis 9.1); rate ratio 2.26, 95% CI 1.24 to 4.11, Analysis 9.1). If the topical and oral PUVA data is added then it can be shown that 28 of 41 (68%) cleared with PUVA‐etretinate compared with 9 of 33 (27%) receiving PUVA alone (rate difference 0.44,95% CI 0.24 to 0.63; rate ratio 2.66, 95% CI 1.49 to 4.75; Analysis 9.1).

(b) Re‐PUVA versus retinoid

Rosen 1987 showed that 17 of 23 (74%) sites were much improved or cleared with PUVA‐etretinate compared with 10 of 23 (43%) sites with etretinate alone (rate difference 0.30, insufficient detail available to calculate 95% CI; worst case p=0.17, Analysis 10.1). It was not possible to analyse lesser responses than clearance in the small study by Matsunami 1990, where 6 of 10 (60%) sites cleared with Re‐PUVA as against 2 of 10 (20%) sites with retinoid alone (rate difference 0.40, rate ratio 3.00). Clearance was achieved in 14 of Rosen's 23 Re‐PUVA‐treated sites (61%) compared with 3 of 23 (13%) sites which received no PUVA (rate difference 0.48, rate ratio 4.67).

10.1. Analysis.

Comparison 10 retinoid + PUVA versus retinoid alone, Outcome 1 good or excellent response.

Note: The within patient data from Rosen 1987 and Matsunami 1990 could not be appropriately analysed in RevMan but more suitable methods gave very similar results in this case. Insufficient details were available to calculate correct confidence intervals; the worst case p values were 0.01 and 0.29 respectively. Since the estimates from the two studies were similar, this gives reasonable evidence of benefit of re‐PUVA over retinoid from the two studies combined (rate difference 0.45, rate ratio 4.00,Analysis 10.2).

Conclusions: Established increased efficacy of combination of retinoids and PUVA over either treatment modality when used alone.

(4) Ciclosporin

Two studies (Erkko 1998 and Reitamo 1993) examined the use of ciclosporin in low doses (Analysis 11.1). Due to heterogeneity of the results (p=0.03), data were not pooled. Erkko 1998 found that 13 of 27 receiving 1mg/kg/day ciclosporin improved compared with 6 of 31 receiving placebo (rate difference 0.29, 95% CI 0.05 to 0.52; rate ratio 2.49, 95% CI 1.10 to 5.64). Reitamo 1993 used a higher dose (2.5mg/kg/day) and found that 17 of 20 receiving ciclosporin improved compared with 4 of 20 receiving placebo (rate difference 0.65, 95% CI 0.42 to 0.88; rate ratio 4.25, 95% CI 1.74 to 10.41). Erkko found that 7 of 27 receiving ciclosporin still had improvement 12 months into treatment compared with 2 of 31 receiving placebo (rate difference 0.19, 95% CI 0.01 to 0.38; rate ratio 4.02, 95% CI 0.91 to 17.73, Analysis 11.2).

11.1. Analysis.

Comparison 11 low dose ciclosporin versus placebo, Outcome 1 improvement.

11.2. Analysis.

Comparison 11 low dose ciclosporin versus placebo, Outcome 2 improvement lasting 12 months.

Conclusions: Good evidence of improvement but not of clearance with low dose ciclosporin. Little evidence to support the use of long‐term low dose ciclosporin.

(5) Tetracycline antibiotics

Two studies (Thomsen 1973 and Ward 1976) examined the use of a tetracycline (tetracycline for one month and clomocycline for three months respectively). Pooled data showed improvement in 30 of 80 (38%) tetracycline treatment courses compared with 10 of 80 (13%) placebo courses (rate difference 0.25, 95% CI 0.13 to 0.37; Analysis 12.1).

12.1. Analysis.

Comparison 12 tetracycline antibiotic versus placebo, Outcome 1 improvement.

Note: These were both crossover studies and so could not be analysed appropriately within RevMan. For Thomsen 1973 the first period data were extracted whereas both periods are included for Ward 1976 and so this is not a paired analysis. However, the confidence interval for a correct paired analysis is very similar to that shown in RevMan. The difference between placebo and tetracycline antibiotic is less obvious in the three month study with clomocycline than the one lasting one month with tetracycline, but an advantage is still apparent. Thomsen 1973 stated that in some cases of improvement the lesions disappeared completely, but in most cases no more than an 'improvement' was recorded. Unfortunately the figures for this were not disclosed.

Conclusion: Some evidence of modest improvement with tetracycline antibiotics.

(6) Hydroxycarbamide

One study (Hattel 1974) compared hydroxycarbamide with placebo in 13 patients in a short cross‐over trial. Unfortunately numbers of responders in each arm of the study were not disclosed. The investigators found no significant difference in scores for disease severity between placebo and intervention periods.

Conclusion: There is a lack of evidence to suggest that short‐term treatment with hydroxycarbamide is effective.

(7) Colchicine

Two cross‐over studies (Thestrup‐Pedersen 84 and Mann 1982) examined the use of colchicine. Both studies reported high rates of side effects related to the use of colchicine, including diarrhoea, headaches and nausea (Mann 1982). None of the patients 'responded' in the trial reported by Mann 1982. There were 13 documented 'improvements' in that reported by Thestrup‐Pedersen 84: 10 of 27 patients improved when receiving colchicine compared with 3 of 27 when receiving placebo (rate difference 0.26, insufficient details to calculate 95% CI, worst case 0.004 to 0.44). There were no clearances in this trial (rate difference 0, 95% CI ‐0.12 to +0.12). Confidence intervals reported here are calculated by the method described by Newcombe 1998. The data for the two studies were not pooled due to clinical heterogeneity.

Conclusion: Some evidence to suggest a possible modest benefit from colchicine at the expense of a high rate of side effects.

(8) Grentz ray therapy

Lindeloef 1990 found that 13 of 17 sides receiving Grenz ray (low voltage X‐ray therapy) showed greater improvement than the contralateral side; by comparison only 1 of 17 sides receiving placebo showed greater improvement than the Grenz ray‐treated side (p=0.002, Analysis 13.1). The authors reported that none of the patients achieved clearance (Analysis 13.2).

13.1. Analysis.

Comparison 13 Grenz ray therapy versus "sham" radiotherapy, Outcome 1 superior response.

13.2. Analysis.

Comparison 13 Grenz ray therapy versus "sham" radiotherapy, Outcome 2 clearance.

Conclusion: Some evidence of improvement, but not of clearance, from Grenz Ray therapy.

(9) Topical corticosteroids

Kragballe 1991 found that 12 of 19 (63%) sides receiving medium strength corticosteroid under hydrocolloid occlusion cleared completely compared with 3 of 19 (16%) sides receiving highly potent corticosteroid twice daily (p=0.001, Analysis 14.1, rate difference 0.47, 95% CI 0.20 to 0.75). These are within patient data but are not fully reported to allow correct analysis: the lower limit of this confidence interval could potentially be 0.09.

14.1. Analysis.

Comparison 14 steroid + occlusion versus steroid alone, Outcome 1 clearance.

Conclusion: Evidence for successful induction of remission with topical corticosteroid under hydrocolloid occlusion.

Discussion

Summary of main results

Few high quality studies were identified in the review. The majority were short‐term studies in a chronic disease subject to relapses and remissions. Many studies used non‐standardised semi‐quantitative scores as outcome measures and few reported patient satisfaction measures. Documented superiority over placebo does not necessarily equate to worthwhile benefit for the patient. It is also of course possible that failure to report negative trials has resulted in an overestimate of benefit for some of the interventions examined in this review. The heterogeneity of outcome measurement and methodology makes pooling of statistical data difficult. Within these trials other distinguishing variables were identifiable:

Patient factors

Most trials demonstrated that groups were matched for age, sex, length and history of previous treatment and site of lesions. The proportion of patients with evidence of concomitant psoriasis elsewhere varied considerably and it is not known whether this may influence responses to therapy. As PPP is a chronic relapsing‐remitting condition, many patients had received previous treatments for their disease. It is not known whether the response of patients to treatment for an initial episode differs from the response seen in those who have received multiple previous treatments.

Treatment factors

• Oral Treatments: different drugs, doses and formulations of tetracyclines and retinoids

• Topical PUVA: different concentrations, formulations and methods of application of psoralen; different irradiation doses and regimens.

• Systemic PUVA: different regimens of irradiation

• Trial period: different periods of treatment and different 'washout' period prior to trial

Site of involvement

Whether or not there is a difference between hands and feet in the probability of response to treatment is a question which was not discussed by most researchers. In an uncontrolled trial which could not be included in this review (Agren‐Jonsson 1985), palmar disease appeared to respond much more readily than plantar disease to treatment with systemic PUVA.

The Review Manager software is not designed to handle both phases from cross‐over trials. The rate differences and 95% confidence intervals for such trials (Hattel 1974; Mann 1982; Thestrup‐Pedersen 84 and Ward 1976) which appear in the text were calculated according to the method of Newcombe (Newcombe 1998); however, those which appear in the figures in this review differ only slightly from these.

Potential biases in the review process

Trial size

Most studies involved few participants. In a disease with poorly defined treatment outcome measures, many of the studies were underpowered.

Authors' conclusions

Implications for practice.

With the exception of corticosteroids under occlusion, there is little evidence for efficacy of topical therapy in PPP. The efficacy of topical corticosteroids is increased by occlusion with hydrocolloid gel dressings: such therapy may be expected to affect clearance within 12 days in about two thirds of people with PPP. This remission may, however, be short‐lived. There is some evidence to suggest that additional short‐term PUVA does not prevent relapse. We found no eligible studies examining other topical therapies such as tar, anthralin, calcipotriol or tazarotene. Grenz Ray therapy has been shown to produce moderate improvement in about two‐thirds of patients with PPP. There are no studies examining the use of this as an adjunct to other interventions. There is little evidence from RCTs to support the use of topical PUVA.

Systemic photochemotherapy and systemic retinoids are both of value for PPP. Systemic PUVA can induce clearance in up to two fifths of people with PPP though larger studies are needed to confirm this figure. Systemic retinoids (etretinate and acitretin) at a dose of 0.5 mg/kg bw/day produce objective improvement in about two thirds of people with PPP. A good or excellent response is seen in about two fifths. Many patients who do improve will not, however, achieve satisfactory control of disease either because of an insufficient response or because of intolerance of retinoid side effects. Continuation of retinoid therapy after induction of remission reduces the incidence of relapse. Limited data suggest that liarozole might be similarly efficacious in producing objective improvement in PPP. Comparison between retinoids and PUVA suggests that retinoids are more likely to induce clearance but there is little difference in the proportions of people who will show improvement. A combination of systemic retinoids and PUVA produces clearance in about two‐thirds of people with PPP, and compares favourably with PUVA or retinoid therapy alone.

There is evidence that ciclosporin and tetracycline antibiotics may be of some benefit for PPP. Ciclosporin at low dose (1 to 2.5 mg/kg bw/day) produces moderate objective improvement in about two‐thirds of patients over one month. The evidence which suggests that about one‐fourth of patients may be expected to maintain the improvement gained for one year needs confirmation by further studies. The response rates reported in uncontrolled studies and case series using higher doses of ciclosporin suggest that the doses used in published studies may be suboptimal; further studies using higher doses are required to establish whether this is so. Tetracycline antibiotics produce objective improvement in about half of patients with PPP. Clearance is rarely seen.

There is a paucity of evidence supporting the use of colchicine in PPP. There is no evidence to suggest that short term treatment with hydroxycarbamide is effective in PPP. There are no randomised controlled trials to date that support the use of methotrexate, although one uncontrolled prospective study claimed benefit in only a third of treated patients, mainly those people with evidence of psoriasis at other sites (Thomsen 1971).

Implications for research.

The clinical outcomes of trials for PPP are not standardised and this makes it difficult to compare studies. Although many investigators use changes in pustule counts or clinical scores of erythema and scaling, these on their own are insufficient. It should always be possible to dichotomise the results into categories such as improved/not improved or clear/not clear. Of equal importance are the assessments of patients receiving the treatment both in terms of efficacy and acceptability.

We suggest that future trials should report specifically:

• numbers of patients randomised into each treatment arm as well as the numbers completing each arm according to protocol

• numbers of randomised patients who achieve clearance

• numbers of randomised patients who show objective improvement

• numbers of randomised patients who report significant benefit

• numbers of randomised patients who regard the intervention as acceptable or "worth‐while"

We also suggest that:

• the presence or absence of evidence of psoriasis at other body sites should be recorded

• trial design should incorporate separate stratification for this

• data should be presented in such a way that differential response rates of hand and feet can be identified

Further studies of systemic PUVA against placebo may help to determine more precisely the proportion of people who may be expected to clear with PUVA and to clarify the differences in response rates to systemic PUVA reported by Murray 1980 and Rosen 1987 on the one hand and Lassus 1988 on the other. There is much uncertainty as to the efficacy of topical PUVA in PPP and, ideally, a treatment arm using topical PUVA should be included in order to determine whether this modality has any place in its management. A study of ciclosporin at a dose of 3.0 to 5.0 mg/kg bw/day would help to establish the value of this drug for PPP.

A comparison of liarozole with acitretin would help to establish whether the former can achieve response rates to compare with acitretin and, if so, whether it offers benefits in terms of reduced side effects.

A comparison of methotrexate with acitretin would establish whether there is a place for the former in the management of PPP.

In view of the encouraging reports regarding the use of oral pimecrolimus in the treatment of psoriasis, and of its apparent improved side‐effect profile over ciclosporin, a study of this drug as a treatment for PPP would seem appropriate. Although there have been anecdotal reports that the new biological agents which have been introduced for psoriasis may be beneficial in PPP, RCTs should be performed to evaluate their efficacy.

What's new

Date	Event	Description
20 January 2016	Amended	Contact author information (email) updated.

History

Protocol first published: Issue 1, 1999
 Review first published: Issue 1, 2006

Date	Event	Description
16 April 2015	Amended	Edited the published note about the updating of the review.
16 October 2013	Amended	This review is going to be updated. We have written a published note to say that because the updating team is completely new and the original review is very old, a new protocol and then a new review will be written.
17 October 2008	Amended	Converted to new review format.
6 October 2005	New citation required and conclusions have changed	Substantive amendment.
11 March 2005	New search has been performed	Minor update.

Notes

This review is being updated by way of a new protocol and then a review, as the scope of the review has substantially expanded. The citation for the new protocol is as follows: Obeid G, Do G, Katsahian S, Kirby L, Hughes C, Le Cleach L. Interventions for chronic palmoplantar pustulosis (Protocol). Cochrane Database of Systematic Reviews 2015, Issue 4. Art. No.: CD011628. DOI: 10.1002/14651858.CD011628.

The protocol for this review was originally published under the title: Interventions for chronic palmoplantar pustular psoriasis. Since the evidence that this disease is a variant of psoriasis has become increasingly weak, the authors and editors have elected to use the alternative and increasingly widely used name, chronic palmoplantar pustulosis.

Appendices

Appendix 1. The Cochrane Skin Group Specialised Register Search Strategy

((PALMOPLANTAR or PUSTULAR)and PSORIASIS)OR PALMOPLANTAR AND (PUSTULOSIS or (PUSTULAR and BACTERID))

Appendix 2. The Cochrane Central Register of Controlled Trials (CENTRAL) search strategy

PSORIASIS
 PSORIASIS*:ME
 (#1 or #2)
 (((PALMOPLANTAR or PUSTULAR) OR (PALMOPLANTAR and PUSTULOSIS)) OR (PUSTULAR AND BACTERID))
 (#3 and #4)

Appendix 3. MEDLINE (OVID) search strategy

1. randomized controlled trial.pt.
 2. controlled clinical trial.pt.
 3. randomized controlled trials.sh.
 4. random allocation.sh.
 5. double blind method.sh.
 6. single‐blind method.sh.
 7. or/1‐6
 8. (animal not human).sh.
 9. 7 not 8
 10. Clinical trial.pt.
 11. exp clinical trials/
 12. (clin$ adj25 trial$).ti,ab.
 13. ((singl$ or doubl$ or trebl$ or tripl$) adj25 (blind$or mask$)).ti,ab.
 14. Placebos.sh.
 15. Placebo$.ti,ab.
 16. Random$.ti,ab.
 17. Research design.sh.
 18. or/10‐17
 19. 18 not 8
 20. 19 not 9
 21. Comparative study.sh.
 22. exp evaluation studies/
 23. follow up studies.sh.
 24. Prospective studies.sh.
 25. (control$ or prospectiv$ or volunteer$).ti,ab.
 26. or/21‐25
 27. 26 not 8
 28. 27 not (9 or 20)
 29. 9 or 20 or 28
 30. (palm$ or plant$ or sole$ or bacterid).mp. and (pustul$ or psoria$).ti,ab.
 31. 29 and 30

Appendix 4. EMBASE search strategy

1. random$.mp.
 2. factorial$.mp.
 3. crossover$.mp.
 4. placebo$.mp. or PLACEBO/
 5. (doubl$ adj blind$).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
 6. (singl$ adj blind$).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
 7. assign$.mp.
 8. volunteer$.mp. or VOLUNTEER/
 9. Crossover Procedure/
 10. Double Blind Procedure/
 11. Randomized Controlled Trial/
 12. Single Blind Procedure/
 13. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12
 14. psoriasis.mp. or exp PSORIASIS/
 15. exp PUSTULAR PSORIASIS/ or pustular.mp.
 16. exp Pustulosis Palmoplantaris/ or palmoplantar.mp.
 17. (pustular and bacterid).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
 18. (palmoplantar and pustul$).mp. [mp=title, abstract, subject headings, drug trade name, original title, device manufacturer, drug manufacturer name]
 19. 16 or 17 or 18
 20. 14 and 19
 21. 15 or 20
 22. 21 and 13
 23. limit 22 to yr=2003 ‐ 2005
 24. from 23 keep 1‐12

Appendix 5. Salford Database of Psoriasis Trials

We undertook a validation check by cross‐checking with the Salford Database of Psoriasis Trials for trials containing the keyword term PALMOPLANTAR PUSTULAR PSORIASIS. The Salford Database is derived from searches of MEDLINE (from 1966 to 1999) and EMBASE (from 1988 to 1999) using the terms STUDY or TRIAL* or RANDOM* in the text, COMPAR* in the title or CLINICAL‐TRIAL in the subject heading; the Cochrane Central Register of Controlled Trials (CENTRAL); and the European Dermato‐Epidemiology Network (EDEN) Psoriasis Trials Database. We screened and retrieved all retrieved studies by disease type and by intervention including PALMOPLANTAR PUSTULAR PSORIASIS.

Data and analyses

Comparison 1. retinoid versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 improvement	4	127	Risk Difference (M‐H, Fixed, 95% CI)	0.44 [0.28, 0.59]
2 good or excellent response	4	127	Risk Difference (M‐H, Fixed, 95% CI)	0.22 [0.07, 0.36]
3 in remission at three months	2		Risk Difference (M‐H, Fixed, 95% CI)	Subtotals only
3.1 in partial remission at three months	2	45	Risk Difference (M‐H, Fixed, 95% CI)	0.42 [0.16, 0.68]
3.2 in complete remission at three months	2	45	Risk Difference (M‐H, Fixed, 95% CI)	0.19 [‐0.03, 0.42]
4 in remission at six months	1	25	Risk Difference (M‐H, Fixed, 95% CI)	0.35 [‐0.02, 0.72]

Comparison 2. retinoids compared.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 reduction in pustules	1	60	Mean Difference (IV, Fixed, 95% CI)	3.24 [‐29.13, 35.61]

Comparison 3. liarozole versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 induction of remission	1		Risk Difference (M‐H, Fixed, 95% CI)	Subtotals only
1.1 good or excellent response (objective)	1	15	Risk Difference (M‐H, Fixed, 95% CI)	0.45 [0.01, 0.88]
1.2 improvement in visual‐analogue scale (subjective)	1	15	Risk Difference (M‐H, Fixed, 95% CI)	0.34 [‐0.13, 0.81]

Comparison 4. PUVA versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 improvement	3		Risk Difference (M‐H, Fixed, 95% CI)	Subtotals only
1.1 improvement with oral psoralen plus UVA	2	72	Risk Difference (M‐H, Fixed, 95% CI)	0.44 [0.26, 0.62]
1.2 improvement with topical psoralen plus UVA	1	88	Risk Difference (M‐H, Fixed, 95% CI)	0.02 [‐0.17, 0.22]
2 clearance	4		Risk Difference (M‐H, Fixed, 95% CI)	Subtotals only
2.1 clearance with oral psoralen and UVA	2	72	Risk Difference (M‐H, Fixed, 95% CI)	0.42 [0.25, 0.58]
2.2 clearance with topical psoralen and UVA	2	108	Risk Difference (M‐H, Fixed, 95% CI)	0.02 [‐0.04, 0.08]

Comparison 5. topical PUVA versus systemic PUVA.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 clearance	1		Risk Difference (M‐H, Fixed, 95% CI)	Subtotals only
1.1 bath Trioxsalen versus Oral 8‐MOP	1	31	Risk Difference (M‐H, Fixed, 95% CI)	0.0 [‐0.12, 0.12]
1.2 topical 8‐MOP versus oral 8‐MOP	1	46	Risk Difference (M‐H, Fixed, 95% CI)	0.12 [‐0.03, 0.27]
1.3 Topical PUVA overall versus Oral 8‐MOP	1	64	Risk Difference (M‐H, Fixed, 95% CI)	0.08 [‐0.05, 0.20]

Comparison 6. topical 8‐MOP cream PUVA versus trioxsalen soak PUVA.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 clearance	1	51	Risk Difference (M‐H, Fixed, 95% CI)	0.12 [‐0.01, 0.26]

Comparison 7. PUVA versus no intervention after induction of remission with steroid plus occlusion.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 no relapse within one year	1	22	Risk Difference (M‐H, Fixed, 95% CI)	‐0.13 [‐0.54, 0.28]

Comparison 8. retinoid versus PUVA.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 good or excellent response	1	37	Risk Difference (M‐H, Fixed, 95% CI)	‐0.21 [‐0.53, 0.11]
2 clearance	2		Risk Difference (M‐H, Fixed, 95% CI)	Totals not selected

Comparison 9. retinoid +PUVA versus PUVA alone.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 clearance	3	74	Risk Difference (M‐H, Fixed, 95% CI)	0.44 [0.24, 0.63]
1.1 retinoid + topical PUVA versus topical PUVA alone	1	20	Risk Difference (M‐H, Fixed, 95% CI)	0.5 [0.14, 0.86]
1.2 retinoid + systemic PUVA versus systemic PUVA alone	2	54	Risk Difference (M‐H, Fixed, 95% CI)	0.41 [0.18, 0.64]

Comparison 10. retinoid + PUVA versus retinoid alone.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 good or excellent response	1	46	Risk Difference (M‐H, Fixed, 95% CI)	0.30 [0.03, 0.57]
2 clearance	2	66	Risk Difference (M‐H, Fixed, 95% CI)	0.45 [0.25, 0.66]

Comparison 11. low dose ciclosporin versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 improvement	2		Risk Difference (M‐H, Fixed, 95% CI)	Totals not selected
2 improvement lasting 12 months	1	58	Risk Difference (M‐H, Fixed, 95% CI)	0.19 [0.01, 0.38]

Comparison 12. tetracycline antibiotic versus placebo.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 improvement	2	160	Risk Difference (M‐H, Fixed, 95% CI)	0.25 [0.13, 0.37]

Comparison 13. Grenz ray therapy versus "sham" radiotherapy.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 superior response	1	34	Risk Difference (M‐H, Fixed, 95% CI)	0.71 [0.48, 0.94]
2 clearance	1	34	Risk Difference (M‐H, Fixed, 95% CI)	0.0 [‐0.11, 0.11]

Comparison 14. steroid + occlusion versus steroid alone.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 clearance	1	38	Risk Difference (M‐H, Fixed, 95% CI)	0.47 [0.20, 0.75]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Bhushan 2001.

Methods	Randomised controlled trial; double‐blind, parallel group. Allocation code not broken until study completely finished.
Participants	15 patients with palmoplantar pustulosis. age: 42‐74 yrs, median 60.4 yrs. sex: 12 female, 3 male
Interventions	Group 1. liarozole Group 2. placebo study period: 12 weeks
Outcomes	(1) Modified PASI (PPPASI) score based on erythema, scaling and pustulation (2) Patient assessment of treatment efficacy with Visual Analogue Score (3) Responders defined as patients showing overall clearance or marked improvement as judged by the physician or a decrease in PPPASI of more than 70% compared to baseline
Notes	No drop‐outs.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Erkko 1998.

Methods	Randomised controlled trial; double‐blind, parallel group. Allocation code not broken until month 12.
Participants	58 patients with palmoplantar pustulosis with at least 20 fresh pustules at entry. age: 21‐70 yrs sex: 42 female, 16 male
Interventions	Group 1. ciclosporin 1mg/kg bw/day Group 2. placebo study period: one month
Outcomes	(1) Reduction in fresh pustule count by 50% or more at one month. (2) Duration of remission (up to 12 months) as defined by above criterion.
Notes	In patients who achieved specified outcome by one month, treatment with the initial treatment and dose was continued under blinded conditions for up to 12 months or until relapse or withdrawal. No drop‐outs.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Foged 1983.

Methods	Randomised controlled trial; double‐blind, parallel group.
Participants	50 patients with palmoplantar pustular psoriasis. age: 26‐78, median 55 yrs sex: 8 male, 42 female
Interventions	Group 1. etretinate 1mg/kg bw/day Group 2. placebo study period: 8 weeks
Outcomes	(1) Semiquantitative severity score based on :erythema, desquamation, fissures, pustules, pain and difficulty in walking (2) Patient assessment of treatment efficacy (3) Clinician‐assessed efficacy and tolerance
Notes	9 patients withdrew (4 etretinate, 5 placebo)
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Hattel 1974.

Methods	Randomised controlled trial; double‐blind, cross‐over
Participants	13 patients with palmoplantar pustulosis. age: 17‐72 yrs sex: gender not stated
Interventions	Group 1. hydroxycarbamide 0.5mg three times daily Group 2. placebo study period: 3 weeks
Outcomes	Semiquantitative severity score based on pustule count, erythema, thickening and scaling
Notes	No drop‐outs
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Jansen 1978.

Methods	Randomised controlled trial; double‐blind, parallel group. Concealed allocation code used.
Participants	19 patients with palmoplantar pustular psoriasis. age: 19‐72, mean 44 yrs sex: 13 female, 7 male
Interventions	Group 1. etretinate 25mg ‐ 100mg daily (individually tailored according to tolerance) Group 2. placebo study period: 4 months
Outcomes	Semi‐quantitative scoring for pruritus, desquamation, erythema, pustules, vesicles, burning and fissuring (score 0‐4)
Notes	One drop‐out (72 year old woman taking retinoid developed a transient ischaemic attack)
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Kragballe 1991.

Methods	Randomised controlled trial; open left‐right comparison.
Participants	19 patients with palmoplantar pustular psoriasis. age: 18‐71, mean 44 yrs sex:16 male, 3 female
Interventions	Group 1: hydrocolloid gel dressing (Actiderm) over triamcinolone acetonide 0.1% cream applied every third day for four weeks Group 2. clobetasol 0.1% cream applied twice daily for four weeks study period: 4 weeks treatment, followed by 4 weeks treatment with emollient alone
Outcomes	Semi‐quantitative scoring for erythema, thickness, scaling pustules and pruritus
Notes	Treatments not blinded. No drop‐outs.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Lassus 1983.

Methods	Randomised controlled trial; double‐blind, parallel group. Code concealed until treatment completed.
Participants	25 (of 40) patients with palmoplantar pustular psoriasis who achieved complete remission following therapy with etretinate 0.9‐1.0 mg/kg bw/day for up to 16 weeks age: mean 33 yrs sex: 26 female, 14 male in initial (40 patient) group.
Interventions	Group 1. low‐dose etretinate (20‐30mg/day) Group 2. placebo study period: 6 months.
Outcomes	Maintenance of complete remission.
Notes	4 drop‐outs (all in placebo arm of trial)
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

Lassus 1985.

Methods	Randomised controlled trial; open, parallel group
Participants	84 patients with palmoplantar pustular psoriasis. age:median age 51 yrs sex: 63 female, 21 male
Interventions	Group 1. systemic PUVA Group 2. topical 8‐MOP + UVA Group 3. topical trioxsalen + UVA Group 4. etretinate 0.9‐1.0 mg/kg bw/day for 2 weeks, then 0.6‐0.7 mg/kg bw/day study period: 12 weeks
Outcomes	Reductions in fresh pustule counts, erythema, infiltration and scaling.
Notes	Treatments not blinded. Semi‐quantitative scoring system for erythema, infiltration, and scaling. 3 drop‐outs from etretinate group due to side effects. 4 on systemic PUVA, 5 on local methoxsalen and 9 on local trioxsalen treatment discontinued after 8 weeks due to lack of effect.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Lassus 1988.

Methods	Randomised controlled trial; double‐blind, parallel group.
Participants	60 patients with palmoplantar pustular psoriasis. age: 22‐76 yrs sex: 42 female, 18 male
Interventions	Group 1. acitretin 30 mg daily for 4 weeks then variable dose Group 2. etretinate 30 mg daily for 4 weeks then variable dose study period: 8 weeks
Outcomes	Reductions in fresh pustule counts, erythema, infiltration and scaling.
Notes	Semi‐quantitative scoring system for erythema, infiltration, and scaling. 2 drop‐outs: one from etretinate group (did not return after 2 weeks) and one from acitretin group (cleared after 4 weeks).
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Lawrence 1984.

Methods	Randomised controlled trial; double‐blind, parallel group.
Participants	17 patients with palmoplantar pustular psoriasis. age: mean 56yrs sex: 16 female, 4 male
Interventions	Group 1. PUVA and etretinate Group 2. PUVA and placebo study period: up to 20 weeks
Outcomes	(1) fresh pustule count (2) degree of erythema/scaling (0) none, (1) mild, (2) moderate, (3) severe. Assessment at weeks 10,20 or at clearance of the rash. Length of clinical remission (assessed by patients 24 weeks after finishing treatment) .
Notes	Side effects of etretinate were also carefully documented. One patient (on placebo) did not complete the study due to (unrelated) diverticulitis.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Layton 1991.

Methods	Randomised controlled trial; single‐blind (? double‐blind) left‐right comparison.
Participants	27 patients with 44 matched areas on hands and/or feet of palmoplantar pustular psoriasis for at least one year. age: not stated sex: not stated
Interventions	Group 1. topical photochemotherapy using 0.75% 8‐methoxypsoralen emulsion applied 10 minutes prior to UVA exposure starting at 1 J/cm 2 and increasing gradually to 15 J/cm 2. Group 2. base emulsion with UVA as above study period: 8 weeks
Outcomes	(1) pustule count (2) degree of erythema, scaling and fissuring. (3) subjective symptoms as assessed by the patients.
Notes	Randomisation method not clearly stated; method of blinding of observers unclear. No drop‐outs.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Lindeloef 1990.

Methods	Randomised controlled trial; double‐blind (patient and evaluating doctor) left‐right comparison
Participants	17 patients with palmoplantar pustulosis; age: 26‐84 yrs (median 54 yrs) sex: not stated
Interventions	Side 1. Grenz ray (10kV) 4Gy weekly to one side Side 2. "sham" treatment administered to contralateral side study period: 6 weeks
Outcomes	Semiquantitative scores for erythema, scaling, pustulation, disease extent as assessed by dermatologist and for itching and overall symptoms as assessed by patient
Notes	2 drop‐outs (one elderly patient withdrew due to illness and one patient developed generalised psoriasis)
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Mann 1982.

Methods	Randomised controlled trial; double‐blind, cross‐over
Participants	12 patients with palmoplantar pustulosis age: 50‐72 yrs sex: not stated
Interventions	Group 1. colchicine 0.5mg twice daily then matching placebo Group 2. placebo then colchicine 0.5mg twice daily study period: 2 x 6 weeks
Outcomes	(1) Number of yellow pustules (2) Number of brown lesions (3) Degree of scaling (4) Degree of erythema
Notes	Throughout the trial participants applied fluocinolone acetonide 0.01% cream twice daily. 3 drop‐outs: two on colchicine and one on placebo.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Matsunami 1990.

Methods	Randomised between patient comparison and left/right within patient comparison. controlled trial. See Notes.
Participants	20 patients with palmoplantar pusular psoriasis. mean ages 52 and 58 yrs for groups 1 and 2 respectively
Interventions	Group 1. Re (+) received etretinate orally and PUVA topically to right hand & foot. Group 2. Re (‐) received just PUVA topically to right hand & foot. All patients in each group received topical PUVA on the right hand and right foot. The left hand and left foot served as PUVA (‐) controls. study period: 20 weeks
Outcomes	(1) Severity of lesions was evaluated using a semiquantitative score for the following seven factors: (a) erythema (b) infiltration (c) pustule count (d) vesicle count (e) desquamation (f) local itching (g) pain (2) Overall effect of treatment.
Notes	Hand/foot not randomised No placebo/ blinding. No drop‐outs.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Murray 1980.

Methods	Randomised controlled trial; ?blinded, left‐right comparison.
Participants	22 patients with palmoplantar pustular psoriasis. mean age 47.8 yrs (male) and 52.9 yrs (female) sex: 16 female, 6 male
Interventions	Group 1. sides treated with oral 8‐mop PUVA Group 2. sides untreated with PUVA study period: 7.5 weeks
Outcomes	(1) worse (2) no change (3) improved (4) much improved (5) clear improved = less pustulation much improved = no pustulation with residual erythema and scaling
Notes	Second part of study cannot be included as it is uncontrolled. No dropouts.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Nielsen 1995.

Methods	Randomised controlled trial; single‐blind, crossover. Patients could not be blinded.
Participants	22 patients with palmoplantar pustular psoriasis. age: 32‐72, mean 52 yrs sex: 20 female, 2 male
Interventions	Group 1. clobetasol propionate ointment under hydrocolloid gel dressing changed every 4 days for 12 days . When PPP lesions recurred the same treatment was repeated followed by PUVA administered to the soles twice weekly for three weeks. Group 2. treatments in opposite sequence study period: 1 year
Outcomes	(1) Severity of lesions was assessed at the end of the glucocorticoid treatment and at the end of the PUVA treatment according to the following scores; 0‐ smooth and uniform skin on soles 1‐ scaling and erythema 2‐scaling, erythema and 1‐10 pustules on each sole 3‐ scaling, erythema, fissuring and > 10 pustules on each sole.
Notes	One patient did not tolerate oral psoralen and was withdrawn.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Reitamo 1993.

Methods	Randomised controlled trial; double‐blind, parallel group.
Participants	40 patients with palmoplantar pustular psoriasis. age: 18‐70, mean 41 yrs sex: 28 females, 12 male
Interventions	Group 1. oral ciclosporin 2.5 mg /kg bw/day Group 2. placebo study period: 4 weeks
Outcomes	Reduction in fresh pustule count by 50% or more.
Notes	Erythema, infiltration, and scaling were also assessed but were not published. 1 drop‐out in ciclosporin group and one in placebo group.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Rosen 1987.

Methods	Quasi‐randomised controlled trial; Placebo‐controlled 'randomised' paired comparison and between patient comparison
Participants	37 patients with PPPP previously treated (including with potent topical steroids) without satisfactory response
Interventions	Parallel group quasi‐randomisation by year of birth into etretinate and placebo groups Left‐right quasi‐randomisation by date of birth into PUVA or no PUVA (left‐right comparison of either feet or hands depending on which set of extremities was most severely affected) Group 1. oral etretinate 0.6mg/kg + PUVA Group 2. oral etretinate 0.6mg/kg Group 3. placebo + PUVA Group 4. placebo
Outcomes	Assessment by patient and investigator on a four point scale: 'cleared' meant an excellent result with no desquamation or pustulation, erythema and slight residual infiltration were allowed; 'much improved' meant a very good result, but some residual desquamation pustulation and infiltration; 'somewhat improved' meant a substantial, easily recognised improvement.
Notes	Randomisation by year of birth (even or odd). 5 drop‐outs in etretinate group, 2 drop‐outs in placebo group.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	High risk	C ‐ Inadequate

Thestrup‐Pedersen 84.

Methods	Randomised controlled trial; double‐blind, crossover
Participants	27 patients with palmoplantar pustulosis age: 19‐80, median 58 yrs sex: 5 male, 22 female
Interventions	Group 1. colchicine then placebo Group 2. placebo then colchicine study period: 2 x 6 weeks Patients weighing less than 75kg received 1.5g of colchicine daily. Those weighing over 75kg received 2g of colchicine daily.
Outcomes	(1) pustule formation (2) redness (3) scaling
Notes	If the patients condition actually deteriorated during the trial the were treated with potassium permanganate bath weekly and the daily application of 0.1% bethametasone valerate with 1% chinoform. One patient stopped treatment due to side effects and did not receive placebo. 3 had reduction in their dosage.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Thomsen 1973.

Methods	Randomised controlled trial; double‐blind, crossover
Participants	40 patients with palmoplantar pustulosis. age: 16 ‐ 78 yrs sex: not stated
Interventions	Group 1. tetracycline 250mg twice daily then placebo then tetracycline 250mg twice daily Group 2. placebo then tetracycline 250mg twice daily then placebo study period: 3 x 4 weeks
Outcomes	Positive response: indicating that the lesions had disappeared completely or had considerably diminished with pustules no longer present. Negative response: indicating no response.
Notes	two patients stopped taking the drug after two courses because of side effects.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Ward 1976.

Methods	Randomised controlled trial; double‐blind, crossover. Manufacturer concealed allocation of supplies.
Participants	60 patients with palmoplantar pustular psoriasis. age: 48 female median age 52 yrs 12 male median age 46 yrs sex: 48 female, 12 male
Interventions	Group 1. clomocycline 170mg three times daily for 2 weeks then twice daily for 10 weeks followed by placebo Group 2. placebo then clomocycline 170mg three times daily for 2 weeks then twice daily for 10 weeks study period: 2 x 3 months
Outcomes	(1) pustulation ‐ more, unchanged, less, much less, none. (2) erythema, scaliness and the area involved were considered together ‐ worse, unchanged, less or clear. At 12 weeks those who were improved were classified as responders. If no change or worse were classified as non‐responders.
Notes	20 patients failed to complete: 6 did not attend, 14 had side effects ( 13 of which were on clomocycline)
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Low risk	A ‐ Adequate

White 1985.

Methods	Randomised controlled trial; double‐blind (doctor and patient), parallel group.
Participants	20 patients with palmoplantar pustular psoriasis. age: 16 female average age 57 yrs 4 male average age 60 yrs sex: 16 female, 4 male
Interventions	Group 1. etretinate 1mg/kg/day Group 2. placebo study period: 10 weeks
Outcomes	(1) semi quantitative score based on overall objective assessment by the doctor (2) pustule count
Notes	Groups matched for age but not for disease duration. 1 drop‐out (etretinate group)
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

White 1986.

Methods	Randomised controlled trial; double‐blind, parallel group
Participants	20 patients with palmoplantar pustular psoriasis who had received etretinate 70 mg daily for 4 weeks age: median age group 1 55 yrs median age group 2 50 yrs sex: 18 female, 2 male
Interventions	Group 1. 30mg etretinate daily Group 2. placebo study period: 12 weeks
Outcomes	(1) Number of fresh pustules counted (2) degree of scaling/ erythema‐ (0)‐none, (1) mild, (2) moderate, (3) severe
Notes	Initially all patients received the same treatment for a 4 week period. They were then randomised to receive either etretinate or placebo. 2 drop‐outs in etretinate group and 2 drop‐outs in placebo group.
Risk of bias
Bias	Authors' judgement	Support for judgement
Allocation concealment (selection bias)	Unclear risk	B ‐ Unclear

Bw: Bi‐weekly
 MOP: Methoxypsoralen
 PASI: Psoriasis Area and Severity Index
 PPPASI: Palmoplantar Pustular Psoriasis Area and Severity Index
 PUVA: psoralen plus ultraviolet A radiation
 UVA: ultraviolet A radiation

Methods: Unless otherwise stated, no details were provided of method of allocation concealment and blinding.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Abel 1980	Uncontrolled study of topical PUVA.
Agren‐Jonsson 1985	Uncontrolled study of oral PUVA.
Allenby 1966	Uncontrolled study of topical methotrexate applied under polythene occlusion at night.
Coleman 1989	Retrospective uncontrolled study of topical PUVA.
Davis 1998	Uncontrolled study of topical PUVA in a variety of palmoplantar dermatoses.
De Rie 1995	Uncontrolled study of topical PUVA in patients with non‐pustular palmoplantar psoriasis.
Duweb 2001	Study of the effect of occlusion on calcipotriol efficacy in palmoplantar non‐pustular psoriasis.
Fredriksson 1979	Randomised unblinded comparison of two different etretinate dose regimens.
Goette 1984	Uncontrolled case series of intralesional triamcinolone injection.
Jansen 1981	Uncontrolled case series of topical PUVA .
Kumar 1997	Non‐randomised parallel group comparison of crude coal tar and salicylic acid ointments in non‐pustular palmoplantar psoriasis.
Meinardi 1990	Uncontrolled open study of ciclosporin.
Morison 1978	Non‐randomised and unblinded study of oral PUVA.
Schröder 1989	30 patients were randomised to acitretin or placebo for four weeks only, after which all patients received acitretin for a further 20 weeks. Data for the four week initial comparative phase of the study were not analysable.
Shephard 1998	Non‐randomised left‐right comparison of 8‐MOP bath PUVA versus 8‐MOP lotion PUVA in patients with a variety of palmoplantar dermatoses, including some with PPP.
Thomsen 1971	Not placebo‐controlled. 25 patients with pustulosis palmaris et plantaris were treated with methotrexate 25mg weekly for 2 months.
Thune 1982	Partly randomised study of etretinate.
Vahlquist 1985	Comparison of acitretin with etretinate. The clinical response to treatment was not presented in such a way that patients with palmoplantar pustulosis could be assessed separately.
Volden 1992	Uncontrolled study of clobetasol propionate lotion under hydrocolloid occlusive dressing.
Wilkinson 1979	Uncontrolled observational study of topical PUVA.

Contributions of authors

Sacha Marsland and Robert Chalmers produced the review with helpful comments and criticisms from Chris Griffiths. Sally Hollis gave advice throughout the drafting process and provided help with statistical analysis. Jo Leonardi‐Bee gave statistical advice and made revisions on the final version of the review.

Sources of support

Internal sources

• Salford Royal Hospitals NHS Trust, UK.

External sources

• No sources of support supplied

Declarations of interest

Christopher Griffiths has been a paid Consultant and has received speaker's fees and research grants from Novartis, the manufacturer of ciclosporin.

Edited (no change to conclusions)