Another important treatment is photochemotherapy with psoralen plus ultraviolet A (PUVA). Local PUVA treatment has been used in uncontrolled studies in palmoplantar conditions with varying success rates (1, 2, 4). Better results were observed in vesicular types of eczema compared with hyperkeratotic types (1, 2, 4).

MATERIALS AND METHODS

Trial design

Three treatments were investigated in parallel: (i) iontophoresis combined with local bath-PUVA therapy (“ionto+PUVA”); (ii) local bath-PUVA therapy only (“PUVA”); and (iii) topical corticosteroid therapy (“steroid”). The study was performed at 2 centres: University Medical Center Groningen and St Antonius Hospital, Nieuwegein. The study was observer-blinded.

Patients

Enrolment was open to patients with endogenous eczema and atopic dermatitis (AD). For AD the definition by the UK Working Party’s diagnostic criteria was used (10). All patients were patch-tested. Patients were excluded if they had contact allergy, which was considered relevant for the eczema (=patients indicated an exacerbation of dermatitis after exposure to the allergen). Psoriasis and fungal infections were ruled out. Patients with concomitant hand eczema were included, but dermatitis on other body sites led to exclusion. In addition, all patients were tested for type 1 allergy (by prick testing and/or specific IgE antibodies).

Further exclusion criteria were: systemic therapy (antihistamines excluded) for eczema within 3 months before study entry, (local) UV therapy within 3 months before study entry, history of abnormal UV sensitivity, phototoxic or cytotoxic drugs, history of (pre)malignant skin conditions, metal-containing device (cardiac pacemaker, orthopaedic implants, gynaecological devices), pregnancy, large erosions which could not be protected sufficiently by petrolatum, decreased skin sensitivity (polyneuropathy), or history of low compliance with therapy.

At inclusion the following characteristics were evaluated: duration of eczema, results of patch testing for contact allergy, Fitzpatrick skin type, presence of plantar hyperhidrosis (patient’s own view), smoking (if present, the number of pack years) and morphological type of eczema (vesicular or hyperkeratotic).

Patients were enrolled from July 2006 until January 2011. The study was performed in accordance with the Declaration of Helsinki and all applicable amendments. All patients gave written informed consent before enrolment and the study was approved by the ethics committee at each centre.

Randomization and treatment

Eligible patients were randomized to treatment via computer-generated random-number tables at each centre.

Tap-water iontophoresis. Tap-water iontophoresis was given 3 times weekly for 10 min using Hidrex® (GS Hidrex GmbH, Wuppertal, Germany). The direct current level was slowly increased, guided by the occurrence of tingling sensations. The maximum level was 30 mA. Iontophoresis was immediately followed by bath-PUVA.

Bath-PUVA. The psoralen bath was prepared by adding 8 ml 0.075 g/ml alcoholic stock solution of trimethylpsoralen to 3 l tap water to obtain a final trimethylpsoralen concentration of 0.2 mg/l. The soles were exposed to UVA radiation after soaking for 15 min. The initial UVA dose was 0.2 J/cm2 in patients with skin type I/II, and 0.3 J/cm2 in those with skin types III–VI. Dose increments were 0.2 J/cm2 in patients with skin type I/II, and 0.3 J/cm2 in those with skin types III–VI. The dose was not increased for one week in case of a slight erythema, followed by lower increments (0.1 J/cm2 for skin types I/II, and 0.2 J/cm2 for skin types III–VI). No treatment was given for one week in case of burning, followed by continuation with 70% of the last dose, and lower increments.

Irradiation was performed with Waldmann PUVA 180 L (Waldmann, Schwenningen, Germany) equipped with Sylvania F15W/T8. The irradiance at the surface of the patient’s skin was, on average, 16 mW/cm2 (Nieuwegein) and 16.4 mW/cm2 (Groningen).

Steroid treatment. This consisted of fluticasone propionate 0.05% cream or ointment (GlaxoSmithKline, London, UK) applied initially 2 times daily. This was tapered-off by the patients when the dermatitis showed a marked improvement on “GIP” (see below).

All patients used emollient ointments ad libitum, and antihistamines when necessary. All treatments were given for 8 weeks, followed by an 8-week follow-up period without topical steroid.

Efficacy assessments

Statistical analysis

Normally distributed data are listed as mean (standard deviation (SD)), non-normally distributed continuous data as median (q25, q75), nominal/ordinal data via frequencies and cross-tables.

Baseline characteristics were evaluated via analysis of variance, including treatment and centre as factors for normally distributed data or via Kruskal–Wallis test for non-normally distributed data.

This study describes time-dependent observations, which are correlated by definition. The appropriate analytical approach in such cases is the random intercept, random coefficient linear mixed model (12). Parameters, such as the eczema score and DLQI, are regressed on time (= weeks of treatment) and time is considered to be the random factor. Other relevant factors, such as “treatment group”, “centre” were introduced into the analysis as independent fixed factors, and their interaction with time. The latter informs on whether the differences between treatments and centres are constant over time.

Global impression by the patient and the patient’s view on plantar hidrosis are ordinal 4-point scale parameters. For these variables the generalized estimation equations analyse for time-dependent ordinal parameters were used.

SPSS 19 was used and p-values < 0.05 were considered significant. Evaluation is according to the intention-to-treat principle.

RESULTS

A total of 48 patients entered the study. Nineteen patients were allocated to iontophoresis with local PUVA therapy (“ionto+PUVA”), 14 to local PUVA therapy alone (“PUVA”), and 15 to topical fluticasone (“steroid”). Of these, 5 patients in the ionto+PUVA group, 6 patients in the PUVA group and 2 patients in the steroid group did not complete the trial (Fig. 1). Taken together, 27% of the patients discontinued after allocation.

Fig. 1. Participant flow through the trial. PUVA: psoralen plus ultraviolet A.

The 3 treatment groups did not differ in any of the baseline characteristics (Table I). The discontinued patients groups did not differ in any of the baseline characteristics (data not shown).

Two patients in the ionto+PUVA group and 3 patients in the PUVA group experienced burning during therapy. Burning in combination with the lack of efficacy was, for these patients, the main reason to withdraw from the study. One patient in each group had mild ery­thema. These side-effects disappeared after adjusting the UVA dose.

The mean cumulative values for UVA dose (± SD) in the ionto+PUVA group and PUVA group were 27.3 and 31.2 J/cm2, respectively. The mean cumulative value for iontophoresis dose in the into+PUVA group was 245 mA.

The time-dependent decrease in eczema score, broken down by treatment group, is shown in Fig. 2. The decrease in the eczema score over time (irrespective of the type of treatment) was highly significant (p < 0.001), from an initial 10.69 to 6.87 points at end-observation. The differences between the 3 treatments on the eczema score was not significant (p = 0.053), and the treatment*time interaction (p = 0.075) was also not significant. The latter indicates that the decrease in the eczema score over time was not different between the 3 treatment groups. The centre*time interaction was not significant (p = 0.778): the eczema score decreased by similar levels over time at both centres.

Fig. 2. Eczema score over time. PUVA: psoralen plus ultraviolet A; ionto: iontophoresis.

Neither the type of eczema (vesicular vs. hyperkeratotic) (p = 0.498) nor the presence of hidrosis (p = 0.127) predicted the decrease of the eczema score over time. The same was true for the variables age, sex, duration of eczema and smoking (data not given).

The time-dependent decrease in DLQI score broken down by the treatment group is shown in Fig. 3. The decrease in the DLQI score over time was highly significant (p < 0.001), from an initial 12.23 to 7.78 points at end-observation. However, neither the difference between the 3 treatments on the DLQI score was significant (p = 0.563), nor was the treatment*time interaction (p = 0.564). The type of eczema (vesicular vs. hyperkeratotic) (p = 0.042) was a weak significant predictor of the overall DLQI score. Hyperkeratotic eczema elicited a 3.15 higher DLQI score (95% CI 0.12–6.187). The eczema type*time and the eczema type*centre interactions were both non-significant (p = 0.554 and p = 0.937, respectively). Therefore, the eczema type had no influence on the effect of treatments. The presence of hyperhidrosis was a non-significant (p = 0.078) predictor of the DLQI score decrease over time. When DLQI over time is the dependent variable no influence was noted of treatment, time, centre, type of eczema and hyperhidrosis.

Fig. 3. Dermatology Life Quality Index (DLQI) over time. PUVA: psoralen plus ultraviolet A; ionto: iontophoresis.

The decrease in the GIP score over time was non-significant (p = 0.828), from an initial median 1 to 0 points at end-observation.

The degree of hidrosis as judged by the patients decreased over time only in the ionto+PUVA group (p = 0.143 for differences between groups).

DISCUSSION

This study investigated the efficacy of iontophoresis combined with local bath-PUVA, and the efficacy was compared with that of topical steroids and local bath-PUVA only. There were no significant differences between these 3 treatments for the decrease in the eczema score and DLQI values. DLQI values in these chronic FE patients were high, particularly for hyperkeratotic eczema. Eczema type or hyperhidrosis had no influence on the effects of treatments with respect to the eczema score and DLQI.

An eczema severity score was chosen as the primary efficacy parameter for dermatitis severity (9). This composite instrument measures several objective and 2 subjective characteristics of hand eczema. It is not validated for FE, but no validated instruments to assess FE severity are available. DLQI is a well-known measure for estimating quality of life, validated for various dermatological conditions (11) but not specifically for FE.

Local bath-PUVA is a widely accepted and effective second-line modality in the treatment of hand and foot eczema (13). Therefore, we decided not to include a placebo arm in this study on ethical grounds. The high disease burden and the well-known effects of the reference therapies eliminate the need for a placebo-controlled study.

The high drop-out rate of 27% may be considered a potential source of bias, as most of the drop-outs were due to non-efficacy or side-effects not outweighing efficacy. Such subjects may negatively influence the overall efficacy estimates of the 3 therapies. Excluding these subjects from the analysis would falsely aug­ment therapeutic efficacy estimates. Fortunately, the statistical method used to evaluate the data is not sensitive to premature drop-out. Per subject the decline over time of the parameters is estimated and a minimum of 2 data-points is needed for that. So a late drop-out (after 2 data-points), due to, for example, non-efficacy, provides an estimate of decline and, when this is low or negative, the overall group efficacy is influenced by this. Once again, the discontinued patients did not differ in any of the baseline characteristics.

In previous studies better results were observed for vesicular, as opposed to hyperkeratotic, eczema types (1, 2, 4). In hyperkeratotic eczema bath-PUVA resulted in a moderate improvement, whereas oral PUVA had a significantly better response, presumably because of impaired penetration of 8-methoxypsoralen through the thickened epidermis (4).

In a small open observational study on the combination of iontophoresis with local PUVA (8) clearance or marked improvement was observed in 4 out of 5 treatments, but the current better designed investigation could not confirm the alleged superiority of the combination therapy. In another (right–left) study on vesicular hand eczema, iontophoresis had a favourable effect on vesicles and pruritus scores, but not on erythema and scaling scores (6). In that study the same iontophoresis device was used as in our study, and was given daily for 3 weeks. No mention was made of the cumulative mA values or of the presence of hyperhidrosis (6). It is difficult to compare the outcome of that study with that of our study because of lack of details.

Comparison of the results of the above-mentioned studies is hampered by the fact that different scoring systems were used in small unbalanced groups, evaluated by statistical methods that did not meet the appropriate standards. In contrast, our results are based on the best available scoring systems and are analysed by appropriate statistics, which were interpreted strictly. These powerful analyses point towards the notion that there is no additional advantage of local PUVA with or without iontophoresis over local steroid therapy.

A significant benefit of combined iontophoresis–bath-PUVA in comparison with bath-PUVA only and topical steroid therapy could not be demonstrated in this study. All treatments induced improvement. These results may be because of the study design using a comparison between 3 active treatments. Introducing a placebo group is unethical given the high disease burden in our patients. Moreover, it makes more sense to compare the efficacy of the novel therapy with a first-line treatment (in this case topical steroids) (17). To the best of our knowledge, there are only 2 comparative studies on palmoplantar eczema without a placebo group (4, 14). In those studies too, no differences were observed between the treatment groups, which consisted of UVA therapy only. Our study has demonstrated that in each treatment group there were responders and non-responders. This is probably caused by the fact that patients behave differently by genetic mechanisms and also by the fact that they may become resistent to previously given therapies.

In dermatological practice topical steroid treatment is regarded as a first choice. Therefore, local bath-PUVA may be offered when patients are resistent to local steroids.

ACKNOWLEDGEMENTS

We thank J. L. van Velde, Dermatologist, St Antonius Hospital, Nieuwegein and N. Drukker, Nurse Practitioner, University Medical Center Groningen, for their help in assessing the efficacy parameters. We also thank M. J. van Es, MD, Groningen, for assistance in recruiting the patients and data entry.

The authors declare no conflicts of interest

REFERENCES