BACKGROUND: Natalizumab (NTZ) is sometimes discontinued in patients with multiple sclerosis, mainly due to concerns about the risk of progressive multifocal leukoencephalopathy. However, NTZ interruption may result in recrudescence of disease activity. OBJECTIVE: The objective of this study was to summarize the available evidence about NTZ discontinuation and to identify which patients will experience post-NTZ disease reactivation through meta-analysis of existing literature data. METHODS: PubMed was searched for articles reporting the effects of NTZ withdrawal in adult patients (⩾18 years) with relapsing-remitting multiple sclerosis (RRMS). Definition of disease activity following NTZ discontinuation, proportion of patients who experienced post-NTZ disease reactivation, and timing to NTZ discontinuation to disease reactivation were systematically reviewed. A generic inverse variance with random effect was used to calculate the weighted effect of patients' clinical characteristics on the risk of post-NTZ disease reactivation, defined as the occurrence of at least one relapse. RESULTS: The original search identified 205 publications. Thirty-five articles were included in the systematic review. We found a high level of heterogeneity across studies in terms of sample size (10 to 1866 patients), baseline patient characteristics, follow up (1-24 months), outcome measures (clinical and/or radiological), and definition of post-NTZ disease reactivation or rebound. Clinical relapses were observed in 9-80% of patients and peaked at 4-7 months, whereas radiological disease activity was observed in 7-87% of patients starting at 6 weeks following NTZ discontinuation. The meta-analysis of six articles, yielding a total of 1183 patients, revealed that younger age, higher number of relapses and gadolinium-enhanced lesions before treatment start, and fewer NTZ infusions were associated with increased risk for post-NTZ disease reactivation (p ⩽ 0.05). CONCLUSIONS: Results from the present review and meta-analysis can help to profile patients who are at greater risk of post-NTZ disease reactivation. However, potential reporting bias and variability in selected studies should be taken into account when interpreting our data.