Morphological characteristics and HPV genotype predict the treatment response in cutaneous warts Running head : Predicting treatment response in cutaneous warts

Background: Cutaneous warts have a cure rate after therapy of no more than approximately 50%. Recently, we developed and validated a standard assessment tool for warts (CWARTS) based on phenotypical characteristics. Objectives: To assess whether patient and morphological wart characteristics predict the HPV type in the specific wart and whether these characteristics as well as the HPV type predict a favourable treatment response. Methods: Photographs were used to score 9 morphological wart characteristics using the newly developed CWARTS tool. Genotyping of 23 wart-associated HPV types was performed using the HSL-PCR/MPG assay. The results were correlated with a favourable response to treatment with monochloroacetic acid, cryotherapy or a combination of cryotherapy and salicylic acid. Odds ratios were calculated using logistic regression in a Generalised Estimating Equations (GEE) model. Results: Black dots (capillary thrombosis) strongly predicted the presence of any HPV type in a wart. From all characteristics tested, the HPV type most strongly predicted the treatment response when the warts were treated with monochloroacetic acid or the combination of cryotherapy and salicylic acid with a significantly decreased treatment response if the warts contained HPVs of the alpha genus (HPV2, HPV27 or HPV 57). When cryotherapy alone was used for common warts, HPV type did not play a role, but cryotherapy was less effective when the wart showed callus and was located deeper in the skin. Conclusions: Morphological characteristics of the warts and the HPV genotype influence treatment outcome and thus potentially influence future treatment decisions for common and plantar warts. INTRODUCTION Warts are hyperkeratotic papules and plaques induced by human papillomaviruses (HPV). At present more than 350 HPV genotypes are recognized. Clinically, cutaneous warts are divided into two basic wart types: common warts (all locations except the soles of the feet) and plantar warts (on the soles of the feet). Cutaneous warts are a common ailment in children and adults with an average prevalence of 3-13%, ranging up to 33% in primary school population. Quality of life can be considerably affected due to physical and psychological distress. A wide variety of treatments is available of which none is very effective. Treatment results and HPV distribution differ between common and plantar warts. Monochloroacetic acid (MCA) and cryotherapy are most effective first line treatments in common warts, while for plantar warts MCA is most effective. Nevertheless, only approximately 50% of patients are cured when optimal treatment per wart type is chosen. A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. HPV2, 27 and 57 of the alpha genus and HPV1 of the mu genus are the most frequently found HPV types in cutaneous warts. In the past clinical morphological characteristics have been associated with some HPV genotypes. 13,14 Egawa found that HPV1 DNA sequences could be detected in warts with granular intracytoplasmic inclusion bodies, that punctate keratotic lesions with filamentous intracytoplasmic inclusion bodies were associated with HPV63 (mu genus) and that pigmented warts with homogenous intracytoplasmic inclusion bodies contained one of the gamma HPVs, i.e. HPV4, HPV60 or HPV65. More recently we have found that compared to HPV2, 27 and 57 positive warts, HPV1 containing warts usually occurred in children, preferentially on a plantar surface and with a short duration before presentation to the physician. Clinical morphological wart characteristics in relation to HPV type, however, thus far have not been systematically investigated. The HPV type in the wart may influence the treatment response. HPV1-positive warts had a much better response to treatment with salicylic acid, but also disappeared much faster with a wait and see policy than HPV2, 27 or 57 positive warts. In another study, the response to cryotherapy appeared to be unrelated to HPV genotype. Next to HPV genotype, location and duration of the warts and morphological wart characteristics could also influence treatment response. In a former study, morphological characteristics of warts were defined by a group of dermatologists and general practitioners and validated for interand intra-observer agreement. An assessment tool for diagnosing warts, i.e. CWARTS tool was developed which is implemented in our study (co-publication). The purposes of this study were 1) to assess whether patient and morphological wart characteristics can predict the HPV type in the specific wart and 2) whether the patient characteristics and morphological wart characteristics as well as the HPV type causing the wart can predict a favourable treatment response of common and plantar warts. For this purpose we performed a retrospective secondary analysis using photographs, HPV data and treatment response of a large randomized controlled trial (WARTS2-trial). First, characteristics as obtained with the CWARTS tool as well as patient characteristics were analysed for association with HPV genotype. In the second part of this study we investigated phenotypical characteristics and HPV genotype and related them to treatment response. MATERIALS AND METHODS This is a retrospective secondary analysis of a selected subpopulation of the WARTS-2 trial. In brief, the WARTS-2 trial was a multicentre randomized parallel-group superiority trial to compare the effectiveness of MCA to most effective usual treatments, i.e. in common warts MCA was compared with cryotherapy and in plantar warts MCA was compared with a combination of cryotherapy and salicylic acid. For the treatment with MCA every 2 weeks a saturated concentration of 76% was used which was applied and subsequently covered with tape by the general practitioner or practice assistant until all warts were cured; for cryotherapy every two weeks three subsequent freeze-thaw cycles were applied to common warts and to plantar warts cryotherapy was combined with daily selfA cc ep te d A rt ic le This article is protected by copyright. All rights reserved. administration of petroleum jelly containing 40% salicylic acid. A favourable treatment response was reached when all warts were cured at 13 weeks. From 41 general practices immunocompetent patients with new cutaneous warts (common or plantar) aged 4 years and older were recruited. Common warts were defined as warts on the hands and all other locations except the soles of the feet which were defined as plantar warts. More details of the treatment protocol of the WARTS-2 trial and the outcome assessment have been published before. The study protocol was approved by the Medical Ethical committee of the Leiden University Medical Centre and conducted according to the principles of the Declaration of Helsinki (version 2008) and the Medical Research Involving Human Subjects Act (WMO). Photographs were taken by a trained research nurse using the Dino-Lite digital microscope (ANMO Electronics Corporation, Taiwan), present to a fixed scale of ‘20’ resulting in a scale of 5.7:1 and using a standard object which assured a standard camera-wart distance from a selection of the patients who were included in the WARTS-2 trial. At an earlier stage, a clinical assessment tool, further referred to as CWARTS tool had been developed and validated for interand intra-observer agreement by respectively 18 and 6 physicians with a different set of pictures based on the score of nine dichotomized morphological characteristics including: 1. Arrangement (confluent/multiple or solitary); 2. Level (elevated or skin level); 3. Aspect (rough/lobed or smooth/not lobed); 4. Border (sharply or not sharply circumscribed); 5. Colour (lighter/skin colour/yellow or red); presence of 6. White skin flakes; 7. Black dots (capillary thrombosis); 8. Border erythema and 9; Callus (copublication). In short, interand intra – observer agreement for most characteristics was moderate to strong with average intraclass correlation coefficients of respectively 0.51 and 0.42. For the current study, the CWARTS tool was used by two researchers (SCB and KEH) to score the photographs of the warts for these nine morphological characteristics and in case of a different score consensus was reached. Wart size and wart clearance were assessed by trained research nurses during home visits. In order to assess HPV genotype, swabs were taken and analysed using the HSL-PCR/MPG assay (Labo Biomedical Products BV, Rijswijk, the Netherlands). All known cutaneous wartassociated virus were assessed, i.e.: wart-associated HPV types from the alpha(HPV2, 3, 7, 10, 27, 28, 29, 40, 43, 57, 77, 91 and 94), gamma(HPV4, 48, 50, 60, 65, 88 and 95), mu(HPV1 and 63) and nu-genus (HPV41). This sensitive and specific assay has been previously described and evaluated. Morphological differences between wart types, i.e. plantar or common warts were calculated using two sided Chi square test for dichotomous variables and student t-test for continuous variables. The correlations of morphology and patient characteristics with HPV genotype and treatment response were calculated using a Generalised Estimating Equation (GEE) model with robust standard errors, thereby correcting the standard error for possibly incorrect assumptions regarding the covariance structure in the data, induced by the multiple measurements in patients, to estimate univariable and multivariable odds ratios with all nine morphological characteristics, age, sex, and A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. duration and location of the warts in the model. Subsequently, stepwise logistic regression with backward elimination was manually performed to identify the factors that significantly contributed to the model. The linearity assumption for age and duration of warts was tested by first centring the variables age and duration of warts (centring in general decreases the