National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, Maryland, USA.
Obesity (Silver Spring). 2010 Sep;18(9):1821-6. doi: 10.1038/oby.2009.472. Epub 2010 Jan 14.
When examining health risks associated with the BMI, investigators often rely on the customary BMI thresholds of the 1995 World Health Organization report. However, within-interval variations in morbidity and mortality can be substantial, and the thresholds do not necessarily correspond to identifiable risk increases. Comparing the prevalence of hypertension, diabetes, coronary heart disease (CHD), asthma, and arthritis among non-Hispanic whites, blacks, East Asians and Hispanics, we examine differences in the BMI-health-risk relationships for small BMI increments. The analysis is based on 11 years of data of the National Health Interview Survey (NHIS), with a sample size of 337,375 for the combined 1997-2007 Sample Adult. The analysis uses multivariate logistic regression models, employing a nonparametric approach to modeling the BMI-health-risk relationship, while relying on narrowly defined BMI categories. Rising BMI levels are associated with higher levels of chronic disease burdens in four major racial and ethnic groups, even after adjusting for many socio-demographic characteristics and three important health-related behaviors (smoking, physical activity, alcohol consumption). For all population groups, except East Asians, a modestly higher disease risk was noted for persons with a BMI <20 compared with persons with BMI in the range of 20-21. Using five chronic conditions as risk criteria, a categorization of the BMI into normal weight, overweight, or obesity appears arbitrary. Although the prevalence of disease risks differs among racial and ethnic groups regardless of BMI levels, the evidence presented here does not support the notion that the BMI-health-risk profile of East Asians and others warrants race-specific BMI cutoff points.
当研究与 BMI 相关的健康风险时，研究人员通常依赖于 1995 年世界卫生组织报告的常规 BMI 阈值。然而，发病率和死亡率在区间内的变化可能很大，而且这些阈值不一定与可识别的风险增加相对应。通过比较非西班牙裔白人、黑人和东亚裔及西班牙裔人群中高血压、糖尿病、冠心病（CHD）、哮喘和关节炎的患病率，我们研究了 BMI 与健康风险之间的关系在小的 BMI 增量上的差异。该分析基于国家健康访谈调查（NHIS）的 11 年数据，样本量为 337375 人，包括 1997-2007 年合并的成人样本。该分析采用多元逻辑回归模型，采用非参数方法对 BMI 与健康风险的关系进行建模，同时依赖于狭义的 BMI 类别。在四个主要种族和族裔群体中，BMI 水平的升高与慢性疾病负担的增加有关，即使在调整了许多社会人口特征和三个重要的健康相关行为（吸烟、身体活动和饮酒）后也是如此。对于除东亚裔以外的所有人群，与 BMI 在 20-21 范围内的人相比，BMI<20 的人患某些疾病的风险略高。使用五种慢性疾病作为风险标准，将 BMI 分为正常体重、超重或肥胖的分类似乎是任意的。尽管无论 BMI 水平如何，不同种族和族裔群体的疾病风险患病率都有所不同，但这里提供的证据并不支持东亚裔和其他人的 BMI-健康风险特征需要特定种族的 BMI 截止值的观点。
