TNF therapy is effective for all aspects of psoriatic disease, but these drugs are costly and the long-term effects are unknown. Further, methotrexate causes concern with long-term adverse events. The purpose of this pilot study was to test the feasibility of drug withdrawal from patients with psoriatic arthritis, in stable low disease state. We examined the availability of patients, their willingness to participate, study procedures, and the proportion of patients in the withdrawal arm who relapsed during the study. Low disease state was defined by minimal disease activity criteria (MDA), and relapse by failure to achieve these criteria. Patients in the withdrawal group underwent a phased withdrawal of medication where the last treatment added was the first withdrawn. Assessments were monthly for 3 months before study exit. Seventy-two patients were invited to participate, of which 57 were found to be eligible. Twenty-six (36.1 %) subsequently attended the screening visit but 9 failed eligibility criteria so that 17 patients (29.8 % of the 57 eligible patients, 95 % confidence interval (CI) 18.4, 43.4 %) were randomised at a ratio of 2:1 in favour of the withdrawal arm (11 withdrawals, 6 standard care). Six patients experienced a flare, all of whom were in the withdrawal arm (relapse rate 54.6 %, 95 % CI 23.4, 83.3 %). Four of the flares were apparent from visit 3 (8 weeks after starting withdrawal). Given the high relapse rate, an alternative trial design of partial treatment withdrawal, possibly including a patient preference arm, is recommended.