BACKGROUND No study has looked at differences of pooled estimates-such as meta-analyses-of corresponding study documents of the same intervention. In this study, we compared meta-analyses of human papillomavirus (HPV) vaccine trial data from clinical study reports with trial data from corresponding trial register entries and journal publications. METHODS We obtained clinical study reports from the European Medicines Agency and GlaxoSmithKline, corresponding trial register entries from ClinicalTrials.gov and corresponding journal publications via the Cochrane Collaboration's Central Register of Controlled Trials, Google Scholar and PubMed. Two researchers extracted data. We compared reporting of trial design aspects and 20 prespecified benefit and harm outcomes extracted from each study document type. Risk ratios were calculated with the random effects inverse variance method. RESULTS We included study documents from 22 randomized clinical trials and 2 follow-up studies with 95,670 healthy participants and non-HPV vaccine comparators (placebo, HPV vaccine adjuvants and hepatitis vaccines). We obtained 24 clinical study reports, 24 corresponding trial register entries and 23 corresponding journal publications; the median number of pages was 1351 (range 357 to 11,456), 32 (range 11 to 167) and 11 (range 7 to 83), respectively. All 24 (100%) clinical study reports, no (0%) trial register entries and 9 (39%) journal publications reported on all six major design-related biases defined by the Cochrane Handbook version 2011. The clinical study reports reported more inclusion criteria (mean 7.0 vs. 5.8 [trial register entries] and 4.0 [journal publications]) and exclusion criteria (mean 17.8 vs. 11.7 and 5.0) but fewer primary outcomes (mean 1.6 vs. 3.5 and 1.2) and secondary outcomes (mean 8.8 vs. 13.0 and 3.2) than the trial register entries. Results were posted for 19 trial register entries (79%). Compared to the clinical study reports, the trial register entries and journal publications contained 3% and 44% of the seven assessed benefit data points (6879 vs. 230 and 3015) and 38% and 31% of the 13 assessed harm data points (167,550 vs. 64,143 and 51,899). No meta-analysis estimate differed significantly when we compared pooled risk ratio estimates of corresponding study document data as ratios of relative risk. CONCLUSION There were no significant differences in the meta-analysis estimates of the assessed outcomes from corresponding study documents. The clinical study reports were the superior study documents in terms of the quantity and the quality of the data they contained and should be used as primary data sources in systematic reviews. SYSTEMATIC REVIEW REGISTRATION The protocol for our comparison is registered on PROSPERO as an addendum to our systematic review of the benefits and harms of the HPV vaccines: https://www.crd.york.ac.uk/PROSPEROFILES/56093_PROTOCOL_20180320.pdf: CRD42017056093. Our systematic review protocol was registered on PROSPERO on January 2017: https://www.crd.york.ac.uk/PROSPEROFILES/56093_PROTOCOL_20170030.pdf. Two protocol amendments were registered on PROSPERO on November 2017: https://www.crd.york.ac.uk/PROSPEROFILES/56093_PROTOCOL_20171116.pdf. Our index of the HPV vaccine studies was published in Systematic Reviews on January 2018: https://doi.org/10.1186/s13643-018-0675-z. A description of the challenges obtaining the data was published on September 2018: https://doi.org/10.1136/bmj.k3694.

中文翻译：

---

人乳头瘤病毒 (HPV) 疫苗的好处和危害：临床研究报告的试验数据与相应的试验登记条目和期刊出版物的比较。


背景 没有研究关注同一干预措施的相应研究文件的汇总估计值的差异，例如荟萃分析。在这项研究中，我们将临床研究报告中的人乳头瘤病毒 (HPV) 疫苗试验数据与相应试验登记条目和期刊出版物中的试验数据进行了荟萃分析。方法 我们从欧洲药品管理局和葛兰素史克获得了临床研究报告，从 ClinicalTrials.gov 获得了相应的试验登记条目，并通过 Cochrane 合作组织的对照试验中央登记册、Google Scholar 和 PubMed 获得了相应的期刊出版物。两名研究人员提取了数据。我们比较了试验设计方面的报告和从每种研究文件类型中提取的 20 个预先指定的益处和危害结果。采用随机效应逆方差法计算风险比。结果 我们纳入了 22 项随机临床试验和 2 项随访研究的研究文件，涉及 95,670 名健康参与者和非 HPV 疫苗比较者（安慰剂、HPV 疫苗佐剂和肝炎疫苗）。我们获得了24份临床研究报告、24份相应的试验注册条目和23份相应的期刊出版物；页数中位数分别为 1351 页（范围 357 至 11,456）、32 页（范围 11 至 167）和 11 页（范围 7 至 83）。所有 24 份 (100%) 临床研究报告、无 (0%) 试验注册条目和 9 份 (39%) 期刊出版物均报告了 2011 版 Cochrane 手册定义的所有六种主要设计相关偏差。临床研究报告报告了更多包容性标准（平均 7.0 vs. 5.8 [试验登记条目] 和 4.0 [期刊出版物]）和排除标准（平均 17.8 vs. 11.7 和 5.0），但主要结局较少（平均 1.6 vs. 3.5 和 1.0）。2) 和次要结果（平均 8.8 对比 13.0 和 3.2）比试验登记条目高。已发布 19 项试验注册条目（79%）的结果。与临床研究报告相比，试验登记条目和期刊出版物包含 7 个评估益处数据点（6879 对 230 和 3015）的 3% 和 44%，以及 13 个评估危害数据点（167,550）的 38% 和 31%。对比 64,143 和 51,899）。当我们将相应研究文件数据的汇总风险比估计值与相对风险比进行比较时，荟萃分析估计值没有显着差异。结论 相应研究文件的评估结果的荟萃分析估计没有显着差异。临床研究报告在其所含数据的数量和质量方面是优质的研究文件，应作为系统评价的主要数据来源。系统审查注册 我们的比较方案已在 PROSPERO 上注册，作为我们对 HPV 疫苗的益处和危害进行系统审查的附录：https://www.crd.york.ac.uk/PROSPEROFILES/56093_PROTOCOL_20180320.pdf：CRD42017056093 。我们的系统评价方案于 2017 年 1 月在 PROSPERO 上注册：https://www.crd.york.ac.uk/PROSPEROFILES/56093_PROTOCOL_20170030.pdf。 2017 年 11 月，两项协议修正案在 PROSPERO 上注册：https://www.crd.york.ac.uk/PROSPEROFILES/56093_PROTOCOL_20171116.pdf。我们的 HPV 疫苗研究索引于 2018 年 1 月发表在《系统评论》上：https://doi.org/10.1186/s13643-018-0675-z。 2018 年 9 月发布了获取数据挑战的描述：https://doi.org/10.1136/bmj.k3694。