Development and validation of the K-VSCOR for scoring Koebner's phenomenon in vitiligo/non-segmental vitiligo

Abstract

The relation of vitiligo/non-segmental vitiligo (NSV) to Koebner's phenomenon is variably appreciated. Our objective was to develop and validate a simple clinical score for Koebner's phenomenon (KP) in patients with vitiligo/NSV. The study population was composed of 351 individuals in the development sample and 285 patients in the validation sample. Seven variables were independently associated with the presence of KP: disease duration of more than 3 yr, forehead + scalp areas, eyelids, wrists, genital + belt areas, knees and tibial crests. The score computed by the weighted sum of the rounded coefficients of these seven variables ranged from 0 to 56 (mean 38.39 ± 22.93). The probability of having KP was computed as follows: exp (-2.37 + 0.1*score)/exp [1 + (-2.37 + 0.1*score)]. When applying the score to each patient in the validation and the development sample, the score maintained adequate discrimination and calibration (AUC-ROC = 0.78), arguing that KP can be adequately predicted using our score. Further studies should evaluate KP assessed by the K-VSCOR in clinical practice with the aim to determine its association with clinical profile, course and treatment response of vitiligo.