We thank Dr Flegal for her interest in our recent study. We understand her concerns about self-reported anthropometric data and acknowledge the limitation that weight was not measured in all 10 cohorts, as we noted in the Discussion section of the article. Reassuringly, validation studies have shown that the absolute difference between reported and measured weight is relatively small. In the anthropometric validation study from the Million Women Study cohort, Wright et al. (1) documented a mean (SD) reported weight of 68.4 (11.9) kg and measured weight of 69.8 (12.6) kg among 40 000 women with both measures at the same time point. The calculated regression dilution ratio for weight in that study was 1.02, and the authors concluded that “reporting errors are likely to generate very little bias in estimates of associations with disease outcomes” (1). In her letter, Dr Flegal points out a small but statistically significant difference in the hazard ratios for self-reported and measured data from one subset of studies in the Global BMI Mortality Collaboration Cohort (2). However, the difference in the results obtained using self-reported and measured body mass index when all 239 studies were included was not statistically significant. Furthermore, the same conclusions were drawn regardless of whether the weight was self-reported or measured, a modest increased risk of death for overweight and a greater risk for obesity.

中文翻译：

---

回复Flegal。

我们感谢Flegal博士对我们最近的研究的兴趣。我们了解她对自我报告的人体测量数据的担忧，并承认存在局限性，因为我们在本文的“讨论”部分中指出，在所有10个队列中均未测量体重。令人放心的是，验证研究表明，报告的重量和测量的重量之间的绝对差相对较小。在“百万妇女研究”队列的人体测量学验证研究中，Wright等人。（1）记录了4万名妇女的平均（SD）报告体重为68.4（11.9）千克，实测体重为69.8（12.6）千克，两种措施均在同一时间点进行。在该研究中，计算得出的体重的回归稀释比为1.02，作者得出结论：“报告错误可能与疾病结局的相关性估计产生很小的偏差”（1）。Flegal博士在信中指出，来自全球BMI死亡率协作研究组（2）的一项研究的自我报告和测得数据的危险比存在细微但统计学上显着的差异。但是，当包括全部239个研究时，使用自我报告和测得的体重指数获得的结果差异无统计学意义。此外，无论体重是自行报告还是测量，得出的结论都是相同的，超重导致的死亡风险适度增加，肥胖症的风险较高。当包括全部239个研究时，使用自我报告和测量的体重指数获得的结果差异无统计学意义。此外，无论体重是自行报告还是测量，得出的结论都是相同的，超重导致的死亡风险适度增加，肥胖症的风险较高。当包括全部239个研究时，使用自我报告和测量的体重指数获得的结果差异无统计学意义。此外，无论体重是自行报告还是测量，得出的结论都是相同的，超重导致的死亡风险适度增加，肥胖症的风险较高。