IMPORTANCE: Several trials demonstrated the impact of novel agent-based maintenance in newly diagnosed multiple myeloma (NDMM), but there is no current evidence demonstrating the superiority of one regimen over the other, owing to the lack of direct/indirect comparisons. OBJECTIVE: To analyze and compare the effectiveness of different maintenance regimens in NDMM via a network meta-analysis. DATA SOURCES: We performed 2 independent searches in PubMed and Cochrane databases, and then we identified all the records registered after 1999 and on or before November 20, 2017. STUDY SELECTION: By blinded review, we identified prospective phase 3 randomized trials evaluating novel agent-based maintenance in patients with NDMM; the included studies compared at least 2 maintenance approaches; comparators included placebo and no maintenance. From 364 screened records, 11 studies were included. DATA EXTRACTION AND SYNTHESIS: We followed (independent extraction) the guidelines provided by the PRISMA Report and the EQUATOR Network. The evidence was synthesized using a network meta-analysis (NMA). To allow comparison of all treatments, no maintenance was selected as common comparator and the effect of placebo was assumed to be the same as no treatment. The best option was identified by a Bayesian consistency model based on hazard ratio (HR), 95% credible interval (CrI), probability of being the best treatment (PbBT), and median ranking distribution (MedR). MAIN OUTCOMES AND MEASURES: Outcomes of interest were progression-free survival (PFS) and overall survival (OS). RESULTS: Eleven trials and 8 treatments including a total of 5073 participants were included. By PFS analysis, lenalidomide-based regimens (lenalidomide-prednisone, lenalidomide alone) were identified as the most effective options (HR, 0.39 [95% CrI, 0.28-0.53] and 0.47 [95% CrI, 0.39-0.55], respectively; MedR, 1 and 2; overall PbBT, 74%). Four treatments (thalidomide-interferon, thalidomide-bortezomib, bortezomib-prednisone, thalidomide alone) showed an HR in favor of maintenance. By OS analysis, lenalidomide alone was identified as the best option (HR, 0.76; 95% CrI, 0.51-1.16; MedR, 2; PbBT, 38%), followed by bortezomib-thalidomide and bortezomib-prednisone. Similar features were noticed in the restricted network including transplant trials, in the sensitivity analysis, and in most of the prognostic subgroups. CONCLUSIONS AND RELEVANCE: Based on PFS and OS results of this NMA, lenalidomide maintenance appears to be the best treatment option, by synthesizing the available evidence of novel agent-based maintenance in the past 20 years.