Membranous nephropathy (MN) is an autoimmune glomerular disease and a leading cause of primary nephrotic syndrome in adults. In primary MN, autoantibodies—most notably against the M-type phospholipase A2 receptor (PLA2R)—drive immune complex deposition along the glomerular capillary wall. Standard immunosuppressive therapies and rituximab (RTX), a type I anti-CD20 monoclonal antibody, have improved outcomes but 20%–40% of patients may not respond to RTX, creating a need for alternative approaches.
Obinutuzumab is a humanized type II anti-CD20 monoclonal antibody with augmented antibody-dependent cellular cytotoxicity and enhanced direct B-cell killing compared with type I agents. Multiple small studies and case series have reported efficacy of obinutuzumab in refractory MN, but results have varied between investigations. The present single-arm meta-analysis synthesized available clinical data to quantify efficacy (overall, complete, partial, and immunologic remission rates) and safety outcomes in patients with refractory MN treated with obinutuzumab.
The review followed PRISMA guidelines and the protocol was registered (PROSPERO CRD420251231344). Four databases (Cochrane Library, Web of Science, Embase, PubMed) were searched through November 12, 2025, using MeSH terms and free-text terms including obinutuzumab and membranous nephropathy. Eligible studies enrolled patients with refractory MN—defined in the review as persistent high or unchanged PLA2R antibody levels after adequate first-line immunosuppression, or ongoing nephrotic syndrome with low serum albumin for more than 6 months. Interventions included obinutuzumab; outcomes required reporting at least one of OR, CR, PR, IR, or adverse events.
Study designs accepted were single-arm clinical trials, cohort studies, and case series. Two reviewers independently screened records and extracted data including study characteristics, remission rates, and continuous laboratory measures (24-hour urine protein or UPCR, serum albumin, serum creatinine, eGFR). When needed, mean changes and standard deviations were calculated or estimated following Cochrane Handbook methods. Methodological quality was appraised with the MINORS instrument. Statistical pooling used R (v4.5.2); pooled proportions with 95% CIs were calculated for remission and AE rates. For continuous measures, MD was used when units were uniform (serum albumin, eGFR, serum creatinine) and SMD when proteinuria units varied across studies. Heterogeneity was evaluated with I² and subgroup analyses were planned by geographic region and follow-up duration.
The initial search returned 232 records; after de-duplication and screening, 12 studies met inclusion criteria, comprising 222 patients. Included reports were published between 2020 and 2025. Nine studies originated from China; three were from the United States, Australia, and India. Two investigations used a cohort design and the remainder were single-arm studies. Mean patient age ranged from 38.5 to 58.0 years; the proportion of males varied between 60% and 94.9%. Mean disease duration ranged from 9 to 85 months. Prior treatments reported across studies included glucocorticoids, RTX, cyclophosphamide, calcineurin inhibitors, and mycophenolate mofetil. Obinutuzumab dosing regimens varied (most commonly two 1 g infusions two weeks apart) with some studies administering supplemental doses during follow-up based on clinical parameters.
Pooled remission estimates across the 12 included studies were:
Reported individual-study OR rates ranged from 76% to 100% in highlighted examples; however, inter-study variability was present. Subgroup analyses by region and follow-up duration were performed in the meta-analysis methodology but specific subgroup numerical results beyond those pooled rates were not detailed in the abstract.
Pooling continuous endpoints showed a large reduction in proteinuria after obinutuzumab treatment, with SMD = −1.2 (95% CI −1.37 to −1.03), indicating a clinically meaningful decline in protein excretion despite heterogeneous measurement methods (g/24h vs UPCR). Serum albumin increased with a pooled mean difference of 12.5 (95% CI 9.91–15.08), reflecting improved oncotic protein levels post-treatment. No significant change in eGFR was observed (MD = 3.23, 95% CI −3.76 to 10.21). Change in serum creatinine results were collected but pooled MD for SCr was not highlighted in the abstract-level summary.
The pooled adverse event rate across included studies was 29% (95% CI 15%–48%). Reported serious infectious events included two fatal pneumonia-related outcomes: one case of severe COVID-19 pneumonia and one case of pneumonia-induced respiratory failure. Incidences of infusion-related reactions and other adverse events were collected across studies, with heterogeneity in reporting. The authors note a low incidence of severe adverse events overall but emphasize the limited evidence base and variability in AE reporting.
Methodological quality was assessed using the MINORS instrument. The body of evidence consisted predominantly of uncontrolled retrospective single-arm trials and case series; two cohort studies were included. Given the non-randomized and frequently retrospective designs, risks of selection bias and confounding are present. The authors state that remission rates derived from these uncontrolled data may be overestimated and caution interpretation accordingly.
Key limitations identified by the authors include the uncontrolled single-arm nature of most included studies, retrospective designs, small sample sizes, geographic concentration of studies (nine of 12 from China), heterogeneous dosing regimens and follow-up durations, and variability in outcome measurement methods (proteinuria reported as g/24h or UPCR). These factors limit causal inference and generalizability. The authors call for large, multicenter randomized controlled trials to confirm efficacy and further define safety.
In pooled uncontrolled data from 12 studies (222 patients) with refractory MN, obinutuzumab was associated with high pooled rates of overall and immunologic remission, substantial reductions in proteinuria, and increased serum albumin, with no significant change in eGFR and a pooled AE rate of 29%. The evidence derives from non-randomized single-arm studies and case series; therefore, findings should be considered preliminary. The authors recommend randomized controlled trials to validate these results and to better characterize the risk–benefit profile of obinutuzumab in refractory MN.