PO.PS01.06 · 人群科学
亚洲人群中体质指数与腰围对头颈部癌症死亡风险的联合关联
Joint association of body mass index and waist circumference with head and neck cancer mortality risk in Asian population
作者与单位 Authors & Affiliations
摘要 Abstract
中文摘要
背景：低体质指数（BMI）与更高的头颈部癌症（HNC）死亡率相关，但全身性肥胖与中心性肥胖在亚洲人群中的作用尚不明确。大多数流行病学研究单独评估BMI或腰围（WC）。我们在亚洲前瞻性队列中探讨了BMI与WC对HNC死亡率的联合关联。
方法：我们对亚洲队列联盟（Asia Cohort Consortium）中随访了HNC病因特异性死亡的人群队列进行了汇总分析。BMI（kg/m²）采用亚太切点分类（<18.5、18.5—22.9、23.0—24.9、25.0—27.4、27.5—29.9、≥30），WC按性别特异性五分位数及WHO肥胖阈值（男性≥90 cm，女性≥80 cm）分类。我们还构建了BMI-WC联合分类及BMI-吸烟联合分类。多变量Cox模型估计风险比（HR）及95%置信区间（CI），校正年龄、性别、吸烟状态、饮酒、队列、糖尿病及高血压，在WC模型中额外校正身高；分析按吸烟状态分层。
结果：与正常BMI（18.5—22.9 kg/m²）相比，体重过轻的参与者（<18.5）HNC死亡风险高出两倍以上（HR 2.19，95% CI 1.75—2.74），而BMI 27.5—29.9者风险较低（HR 0.53，95% CI 0.37—0.77），中间分类显示相似的反向关联。WC呈现平行的反向模式；与最低五分位数相比，较高五分位数的HNC死亡率较低（例如Q4 HR 0.49，95% CI 0.33—0.72；p趋势<0.01）。在按吸烟分层的分析中，体重过轻在从不吸烟者（HR 2.56，95% CI 1.92—3.40）和吸烟者（HR 1.73，95% CI 1.19—2.52）中均与HNC死亡率强烈相关。以正常BMI的从不吸烟者为参照的BMI-吸烟联合模型显示，体重过轻的吸烟者风险显著升高（HR 3.76，95% CI 2.53—5.60），正常BMI的吸烟者风险也过高（HR 2.19，95% CI 1.70—2.83）。在BMI-WC联合分类中，体重过轻且WC正常者HNC死亡率较高（HR 2.29，95% CI 1.47—3.56），而BMI≥30.0且WC正常者风险较低（HR 0.67，95% CI 0.47—0.96）；加入WC并未实质性改变BMI-HNC的关联。
结论：在这些亚洲队列中，较小的体型——低BMI和小WC——始终与更高的HNC死亡率相关，而超重和中度肥胖则与较低的死亡率相关。BMI是比WC更强的预测因子，提示对于HNC死亡率而言，整体体质量可能比中心性脂肪堆积更为相关。这些发现凸显了将基线营养状态和吸烟史纳入风险分层及生存期护理的重要性。
查看英文原文 English abstract
Background: Low body mass index (BMI) has been linked to higher head and neck cancer (HNC) mortality, but the roles of general and central obesity in Asian populations are unclear. Most epidemiologic studies have evaluated BMI or waist circumference (WC) separately. We investigated the joint associations of BMI and WC with HNC mortality in Asian prospective cohorts.
Methods: We conducted a pooled analysis of population-based cohorts in the Asia Cohort Consortium with follow up for cause specific HNC mortality. BMI (kg/m 2 ) was categorized using Asia-Pacific cut offs (<18.5, 18.5-22.9, 23.0-24.9, 25.0-27.4, 27.5-29.9, ≥30), and WC into sex specific quintiles and WHO obesity thresholds (≥90 cm in men, ≥80 cm in women). We also constructed joint BMI-WC and BMI-smoking categories. Multivariable Cox models estimated hazard ratios (HRs) and 95% confidence intervals (CIs), adjusting for age, sex, smoking status, alcohol intake, cohort, diabetes, and hypertension, with additional adjustment for height in WC models; analyses were stratified by smoking.
Results: Compared with normal BMI (18.5-22.9 kg/m 2 ), underweight participants (<18.5) had more than a two fold higher risk of HNC mortality (HR 2.19, 95% CI 1.75-2.74), whereas those with BMI 27.5-29.9 had lower risk (HR 0.53, 95% CI 0.37-0.77), and intermediate categories showed similar inverse associations. WC showed a parallel inverse pattern; compared with the lowest quintile, higher quintiles had lower HNC mortality (e.g. Q4 HR 0.49, 95% CI 0.33-0.72; p trend<0.01). In smoking stratified analyses, underweight remained strongly associated with HNC mortality in both never smokers (HR 2.56, 95% CI 1.92-3.40) and smokers (HR 1.73, 95% CI 1.19-2.52). Joint BMI-smoking models using never smokers with normal BMI as the reference showed markedly elevated risk among underweight smokers (HR 3.76, 95% CI 2.53-5.60) and excess risk among normal BMI smokers (HR 2.19, 95% CI 1.70-2.83). In joint BMI-WC categories, underweight individuals with normal WC had higher HNC mortality (HR 2.29, 95% CI 1.47-3.56), whereas those with BMI ≥30.0 and normal WC had lower risk (HR 0.67, 95% CI 0.47-0.96); adding WC did not materially change BMI-HNC associations.
Conclusion: In these Asian cohorts, smaller body size-low BMI and small WC-was consistently associated with higher HNC mortality, whereas overweight and moderate obesity were associated with lower mortality. BMI was a stronger predictor than WC, suggesting that overall body mass may be more relevant than central adiposity for HNC mortality. These findings highlight the importance of incorporating baseline nutritional status and smoking history into risk stratification and survivorship care.
利益披露 Disclosure
S. Lee, None..
S. Park, None.