We studied the performance of several meta-analysis methods in rare event settings, when the treatment effect is assumed to be homogeneous and baseline prevalences are either homogeneous or heterogeneous. We conducted extensive simulations that included the three most common effect sizes with count data: the odds ratio, the relative risk, and the risk difference. We investigated several important scenarios by varying the level of rareness, the value of the trials' arms unbalance, and the size of the treatment effect. We found that the Mantel-Haenszel method and the Binomial regression model provided the best results across all the scenarios investigated.The Peto method performed satisfactorily only when the true effect size was not too large and the degree of unbalance moderate. Inverse variance was the least reliable method. The use of a continuity correction factor slightly improved the performance of the inverse variance method but deteriorated that of the Peto and Mantel-Haenszel methods. A method based on median unbiased estimators of the probabilities provided similar results to those obtained when using the inverse variance method with a continuity correction. Therefore, when the treatment effect can be assumed to be homogeneous and for either homogeneous or heterogeneous baseline prevalences, we highly recommend using the Mantel-Haenszel method without continuity correction (for all the effect sizes) or the Binomial regression model (for the odds ratio only) to meta-analyze the data. K E Y W O R D Sfixed-effect methods, homogeneous treatment effect, meta-analysis, rare events, simulation study

Purpose Phyllodes tumors (PT) of the breast are rare fibroepithelial neoplasms. Information is controversial in the literature regarding to the optimal surgical management. Most studies suggested margins of at least 10 mm while some recent studies suggested narrower margins without an increased risk of local recurrences (LR) and distant metastases (DM). The objective of this systematic review was to identify and compare studies that assessed these different practices. Methods A systematic review was performed through five databases up to April 2019. Studies exploring the association between the width of margins, subtypes of PT, and the LR and DM rates were considered for inclusion. A statistical model for analyzing sparse data and rare events was used. Results Thirteen studies met eligibility criteria and were selected. Considering a threshold of 10 mm (margins < 10 vs margins ≥ 10 mm), the 5-year incidence rate of LR was estimated to be 5.22 vs. 3.63 (diff. −1.59) per 100 person-years for benign PT, 9.60 vs. 7.33 (diff. −2.27) for borderline PT, and 28.58 vs. 21.84 (diff. −6.74) for malignant PT. For DM, it was estimated to be 0.88 vs. 0.86 (diff. −0.02) for benign PT, 1.61 vs. 1.74 (diff. 0.13) for borderline PT, and 4.80 vs 5.18 (diff. 0.38) for malignant PT. The data for a threshold of 1 mm were not sufficient to draw any conclusions. Conclusion Irrespective of tumor grade, we found that DM was a rarer event than LR. Malignant PT had the highest incidence rate of LR and DM. This meta-analysis found a clear association between width of margins and LR rates. Whatever the tumor grade, surgical margins ≥ 10 mm guaranteed a lower risk of LR than margins < 10 mm. On the other hand, the width of margin did not influence the apparition of DM.

No abstract