Description:
<jats:title>Summary</jats:title><jats:p>Primary immune thrombocytopenia (<jats:styled-content style="fixed-case">ITP</jats:styled-content>) is an acquired immune‐mediated disorder with absence of any underlying cause. Corticosteroids are the standard initial treatment. Splenectomy is the main second‐line treatment. A trend to delay or avoid splenectomy has developed thanks to new agents like rituximab. Few studies have assessed the response rate to rituximab in chronic <jats:styled-content style="fixed-case">ITP</jats:styled-content>. We performed the first meta‐analysis of randomized clinical trials and observational studies on rituximab as an effective splenectomy‐avoiding option in adult chronic <jats:styled-content style="fixed-case">ITP</jats:styled-content>. Overall methods were adapted from published guidelines for meta‐analysis (meta‐analysis of observational studies in epidemiology and preferred reporting items for systematic reviews and meta‐analyses). Two haematologist investigators carried out study selection and data extraction independently, recording overall response rate (<jats:styled-content style="fixed-case">ORR</jats:styled-content>) and complete response (<jats:styled-content style="fixed-case">CR</jats:styled-content>) as primary assessment criteria. Of 364 records were identified through electronic databases. Of 19 retrospective or prospective observational studies were retained after removing duplicate studies and full‐text analyses. The <jats:styled-content style="fixed-case">ORR</jats:styled-content> was 57% (95% confidence interval [<jats:styled-content style="fixed-case">CI</jats:styled-content>]: 48–65), for 368 non‐splenectomized patients after rituximab; <jats:styled-content style="fixed-case">CR</jats:styled-content> was 41% (95% <jats:styled-content style="fixed-case">CI</jats:styled-content>: 0·33–0·51) for 346 patients. Results were stable for <jats:styled-content style="fixed-case">ORR</jats:styled-content> and <jats:styled-content style="fixed-case">CR</jats:styled-content> in all sub‐analyses. In univariate or multivariate mixed‐effect meta‐regression, age was the most relevant effect. According to our results, rituximab should be used in earlier in non‐splenectomized patients.</jats:p>