Effectiveness and predictors of response to somatostatin analogues in patients with gastrointestinal angiodysplasias: a systematic review and individual patient data meta-analysis

Background: Gastrointestinal angiodysplasias are vascular malformations that often cause red blood cell transfusion-dependent anaemia. Several studies suggest that somatostatin analogues might decrease rebleeding rates, but the true effect size is unknown. We therefore aimed to investigate the efficacy of somatostatin analogues on red blood cell transfusion requirements of patients with gastrointestinal angiodysplasias and to identify subgroups that might benefit the most from somatostatin analogue therapy. Methods: We did a systematic review and individual patient data meta-analysis. We searched MEDLINE, Embase, and Cochrane on Jan 15, 2016, with an updated search on April 25, 2021. All published randomised controlled trials and cohort studies that reported on somatostatin analogue therapy in patients with gastrointestinal angiodysplasias were eligible for screening. We excluded studies without original patient data, single case reports, small case series (ie, <10 participants), studies in which patients had a specific aetiology of gastrointestinal angiodysplasias, and studies in which somatostatin analogue therapy was initiated simultaneously with other treatment modalities. Authors of eligible studies were invited to share individual patient data. Aggregated data was used if individual patient data were not provided. The primary outcome was the mean reduction in the number of red blood cell transfusions during somatostatin analogue therapy, compared with baseline, expressed as the incidence rate ratio (IRR) and absolute mean decrease. We defined patients as either good responders (≥50% reduction in the number of red blood cell transfusions) or poor responders (<50% reduction). A mixed-effects negative binomial regression was used to account for clustering of patients and skewness in data. This study was registered in the International Prospective Register of Systematic Reviews (PROSPERO), number CRD42020213985. Findings: We identified 11 eligible studies (one randomised controlled trial and ten cohort studies) of moderate-to-high quality and obtained individual patient data from the authors of nine (82%) studies. The remaining two (18%) studies provided sufficient information in the published manuscript to extract individual patient data. In total, we analysed data from 212 patients. Somatostatin analogues reduced the number of red blood cell transfusions with an IRR of 0·18 (95% CI 0·14–0·24; p<0·0001) during a median treatment duration of 12 months (IQR 6·0–12·0) and follow-up period of 12 months (12·0–12·0), correlating with a mean absolute decrease in the number of red blood cell transfusions from 12·8 (95% CI 10·4–15·8) during baseline to 2·3 (1·9–2·9) during follow-up—ie, a reduction of 10·5 red blood cell transfusions (p<0·0001). 177 (83%) of 212 patients had a good response to somatostatin analogue therapy (defined as at least a 50% reduction in the number of red blood cell transfusions). Heterogeneity across studies was moderate (I2=53%; p=0·02). Location of gastrointestinal angiodysplasias in the stomach compared with angiodysplasias in the small bowel and colon (IRR interaction 1·92 [95% CI 1·13–3·26]; p=0·02) was associated with worse treatment response. Octreotide was associated with a better treatment response than lanreotide therapy (IRR interaction 2·13 [95% CI 1·12–4·04]; p=0·02). The certainty of evidence was high for the randomised controlled trial and low for the ten cohort studies. Adverse events occurred in 38 (18%) of 212 patients receiving somatostatin analogue therapy, with ten (5%) discontinuing this therapy because of adverse events. The most common adverse events were loose stools (seven [3%] of 212), cholelithiasis (five [2%]), flatulence (four [2%]), and administration site reactions (erythema, five [2%]). Interpretation: Somatostatin analogue therapy is safe and effective in most patients with red blood cell transfusion-dependent bleeding due to gastrointestinal angiodysplasias. Somatostatin analogue therapy is more effective in patients with angiodysplasias located in the small bowel and colon, and octreotide therapy seems to be more effective than lanreotide therapy. Funding: The Netherlands Organisation for Health Research and Development and the Radboud University Medical Center.