In 2015, Dan et al. published the systematic review and meta-analysis “Effect of Prednisolone Administration on Patients with Unexplained Recurrent Miscarriage and in Routine Intracytoplasmic Sperm Injection: A Meta-Analysis”.¹ This meta-analysis showed that prednisolone therapy resulted in higher live birth rate (RR 1.58, 95% CI 1.23−2.02) and lower miscarriage rate (RR 0.42, 95% CI 0.28−0.61) in patients with unexplained recurrent pregnancy loss (RPL). Consequently, the paper was cited over 20 times providing possible evidence for prednisolone administration in daily practice.

For the meta-analysis, Dan et al. included two studies from Tang et al.^{2, 3} and a study from Gomaa et al.⁴ to provide information on the Live Birth Rate (LBR) and successful pregnancy (defined as an ongoing pregnancy beyond 20 weeks gestation). Remarkably, the publication by Tang et al from 2009³ concerns a protocol publication and the presented numbers in the publication are a calculation of the needed sample size for adequate power, not the actual results of the trial.

By interpreting these numbers as actual events, the data was included in the meta-analysis, influencing the final outcome and conclusion of the meta-analysis significantly. If only the authentic data from the feasibility trial by Tang et al.² were used, the accurate conclusion would have reflected a Risk Ratio of 1.5 with a 95% confidence interval of 0.79−2.86, indicating an absolute difference of 20% (95% CI: −10% to +50%); statistically insignificant. To make things even more confusing, the alignment of Tang's et al. references is presented inaccurately; in their presented forest plot (Figure 1). Tang AW 2013 is referred to as the study protocol, though published in 2009 (and vice versa). Finally, the plot itself contradicts their claim prednisolone is effective, due to an exchange between the control- and prednisolone group.

In conclusion, we must assert that the execution of the meta-analysis of Dan et al. and subsequent conclusions drawn in the review are flawed. Recognizing fallibility as a human trait, we attempted to communicate these irregularities to Dan et al. on several occasions, yet regrettably, we have not received any response to date. Given the critical importance of fostering evidence-based medicine and evidence-based interventions—in this case, the prescription of prednisolone to women experiencing unexplained recurrent pregnancy loss in an attempt to improve Live Birth Rate or prevent miscarriage—we must keep highlighting these discrepancies.

With this letter-to-the-editor we aim to elicit a response from the authors. In the event of no response, we urge the editorial board of the American Journal of Reproductive Immunology to take necessary measures to prevent the dissemination of incorrect information to patients and doctors, potentially averting detrimental effects.

2015 年，丹等人。发表了系统评价和荟萃分析“泼尼松龙给药对不明原因复发性流产和常规胞浆内单精子注射患者的影响：荟萃分析”。¹这项荟萃分析显示，对于不明原因的复发性流产 (RPL) 患者，泼尼松龙治疗可提高活产率（RR 1.58，95% CI 1.23−2.02）并降低流产率（RR 0.42，95% CI 0.28−0.61）。 ）。因此，该论文被引用了 20 多次，为日常实践中使用泼尼松龙提供了可能的证据。

对于荟萃分析，Dan 等人。包括 Tang 等人的两项研究。^2、3以及 Gomaa 等人的一项研究。⁴提供有关活产率 (LBR) 和成功妊娠（定义为妊娠 20 周以上的持续妊娠）的信息。值得注意的是，Tang 等人 2009 年^{3 的}出版物涉及方案出版物，出版物中提供的数字是对足够功效所需样本量的计算，而不是试验的实际结果。

通过将这些数字解释为实际事件，将数据纳入荟萃分析，从而显着影响荟萃分析的最终结果和结论。如果只有唐等人的可行性试验的真实数据就好了。^{如果使用2}，准确的结论将反映风险比为 1.5，95% 置信区间为 0.79−2.86，表明绝对差异为 20%（95% CI：−10% 至 +50%）；统计上微不足道。让事情变得更加混乱的是，唐等人的结盟。参考文献提供不准确；在他们提出的森林图中（图 1）。Tang AW 2013 被称为研究方案，尽管其发布于 2009 年（反之亦然）。最后，由于对照组和泼尼松龙组之间的交流，情节本身与他们声称泼尼松龙有效的说法相矛盾。

总之，我们必须断言 Dan 等人的荟萃分析的执行。审查中随后得出的结论是有缺陷的。认识到易犯错误是人类的一种特征，我们试图将这些违规行为传达给 Dan 等人。曾多次提出，但遗憾的是，我们迄今尚未收到任何答复。鉴于促进循证医学和循证干预措施的至关重要性（在本例中，为经历不明原因反复流产的女性开出泼尼松龙处方，以试图提高活产率或预防流产），我们必须不断强调这些差异。

通过这封致编辑的信，我们的目的是引起作者的回应。如果没有回应，我们敦促《美国生殖免疫学杂志》编辑委员会采取必要措施，防止向患者和医生传播不正确的信息，从而可能避免有害影响。