Duodenal levodopa infusion represents an effective strategy to manage motor and non-motor complications in patients with advanced Parkinson's disease (PD). However, most published clinical series regard small numbers of patients and do not exceed 1 year follow-up. In this multi-national observational cohort study conducted in seven specialised PD clinics and university hospitals we assessed long-term safety and outcome of chronic treatment with intra-duodenal levodopa infusions in a large population of patients with advanced PD. The starting population consisted of 98 treated patients (safety population). We report clinical outcomes of 73 patients with subsequent efficacy assessment(s) (efficacy population) over a follow-up period up to 2 years. Follow-up periods and collection of clinical observations varied based on individual routine care program. At last follow-up there was a significant (p ≤ 0.05) reduction in duration of "Off" periods as well as dyskinesia duration and severity that was associated with an improvement of quality of life. Twenty three patients (25.3 % of the safety population) withdraw, due to adverse drug reaction (5), procedure and device related events (7), compliance (3) and lack of efficacy (8). The mean duration for last value reported after baseline (LV) was 608 ± 292 days (median: 697 days). Our results demonstrate significant and sustained benefit over a long observation period in motor complications and in quality of life following a change from oral pulsatile to continuous levodopa delivery. The relatively large number of withdrawals reflects the current use of duodenal levodopa infusion in very advanced PD patients.