In people with chronic spontaneous urticaria, a complete response to antihistamine therapy is associated with low disease activity.

The goal of this study was to find out the determinants of a full response to standard-dosed and higher-dosed antihistamine therapy in chronic spontaneous urticaria (CSU)-affected people.

Medical records of 475 CSU subjects (120 of whom were under the age of 18) from 3 separate facilities were examined. Based on 17 clinical factors, 15 machine learning models and conventional statistical techniques were employed to predict full response to standard-dose and higher than standard-dose antihistamine therapy.

The urticaria activity score (UAS), which was used to measure CSU disease activity, was the only clinical indicator that could predict for all age groups a full response to standard-dose and higher-than-standard-dose antihistamine therapy using conventional statistics and machine learning models. According to receiver operating characteristic (ROC) studies, for standard-dosed (area under the ROC curve [AUC] = 0.69) and higher than standard-dosed (AUC = 0.79) antihistamine treatments, the ideal cut-off values of disease activity to predict full response was UAS <3 and UAS <4, respectively.

Machine learning models also found lower total immunoglobulin E (IgE < 150 IU/mL) as a predictor of complete response to standard-dose antihistamine therapy and lower C-reactive protein (CRP < 3.4 mg/mL) as a predictor of complete response to higher-dose antihistamine treatment.

Those with a UAS of less than 3 are highly likely to have a complete response to standard-dosed antihistamine therapy, while patients with a UAS of less than 4 are highly likely to have a complete response to higher than standard-dosed antihistamine therapy. With both machine learning models and conventional statistics, low CSU disease activity is the sole constant predictor of a complete response to antihistamine therapy across all age groups.