BackgroundCompared with levodopa, dopamine agonists (DAs) as initial treatment are associated with lower incidences of motor complications in early Parkinson's disease (PD). There is no strong evidence that a given DA is more potent in lower incidences of motor complications than another. ObjectiveWe performed a network meta-analysis of levodopa versus DAs as monotherapy in early PD to access the risk of motor complications. MethodsDatabases were searched up to June 2022 for eligible RCTs. Levodopa and four DAs (pramipexole, ropinirole, bromocriptine and pergolide) were investigated. The incidences of motor complications and efficacy, tolerability and safety outcomes were analyzed. ResultsNine RCTs (2112 patients) were included in the current study. The surface under the cumulative ranking curve (SUCRA) indicated that levodopa ranked first in the incidence of dyskinesia (0.988), followed by pergolide, pramipexole, ropinirole, and bromocriptine (0.704, 0.408, 0.240, 0.160). Pramipexole was least prone to wearing-off (0.109) and on-off fluctuation (0.041). Levodopa performed best in improvements of UPDRS-II, UPDRS-III, and UPDRS-II + III (0.925, 0.952, 0.934). Bromocriptine ranked first in total withdrawals and withdrawals due to adverse events (0.736, 0.751). Four DAs showed different adverse events profiles. ConclusionIn the two non-ergot DAs, ropinirole is associated with a lower risk of dyskinesia while pramipexole is associated with lower risks of wearing-off and on-off fluctuations. Our research may facilitate head-to-head research, larger sample sizes, long following-up time RCTs to confirm the findings of this network meta-analysis.