Centre for Exercise, Nutrition & Health Sciences, School for Policy Studies, University of Bristol, 8 Priory Road, Bristol, BS8 1TZ, UK.
Population Health Sciences, Bristol Medical School, University of Bristol, Canynge Hall, Bristol, BS8 2PN, UK.
BMC Public Health. 2024 May 27;24(1):1408. doi: 10.1186/s12889-024-18917-9.
To target public health obesity prevention, we need to predict who might become obese i.e. predictors of increasing Body Mass Index (BMI) or obesity incidence. Predictors of incidence may be distinct from more well-studied predictors of prevalence, therefore we explored parent, child and sociodemographic predictors of child/adolescent BMI z-score and obesity incidence over 5 years in Malaysia.
The South East Asia Community Observatory in Segamat, Malaysia, provided longitudinal data on children and their parents (n = 1767). Children were aged 6-14 years at baseline (2013-14) and followed up 5 years later. Linear multilevel models estimated associations with child BMI z-score at follow-up, adjusting for baseline BMI z-score and potential confounders. Predictors included parent cardiometabolic health (overweight/obesity, central obesity, hypertension, hyperglycaemia), and socio-demographics (ethnicity, employment, education). Logistic multilevel models explored predictors of obesity incidence.
Higher baseline BMI z-score predicted higher follow-up BMI z-score both in childhood to late adolescence (0.60; 95% CI: 0.55, 0.65) and early to late adolescence (0.76; 95% CI: 0.70, 0.82). There was inconsistent evidence of association between child BMI z-score at follow-up with parent cardiometabolic risk factors independent of baseline child BMI z-score. For example, maternal obesity, but not overweight, predicted a higher BMI z-score in childhood to early adolescence (overweight: 0.16; 95% CI: -0.03, 0.36, obesity: 0.41; 95% CI: 0.20, 0.61), and paternal overweight, but not obesity, predicted a higher BMI z-score in early to late adolescence (overweight: 0.22; 95% CI: 0.01, 0.43, obesity: 0.16; 95% CI: -0.10, 0.41). Parental obesity consistently predicted five-year obesity incidence in early to late adolescence, but not childhood to early adolescence. An adolescent without obesity at baseline with parents with obesity, had 3-4 times greater odds of developing obesity during follow-up (incidence OR = 3.38 (95% CI: 1.14-9.98, mother) and OR = 4.37 (95% CI 1.34-14.27, father) respectively).
Having a higher BMI z-score at baseline was a stronger predictor of a higher BMI z-score at follow-up than any parental or sociodemographic factor. Targeting prevention efforts based on parent or sociodemographic factors is unwarranted but early childhood remains a key period for universal obesity prevention.
为了针对公共卫生肥胖预防，我们需要预测哪些人可能会肥胖，即预测 BMI 增加或肥胖发病率的因素。发病率的预测因素可能与更广泛研究的患病率预测因素不同，因此我们在马来西亚探讨了父母、儿童和社会人口统计学因素与儿童/青少年 BMI z 评分和肥胖发病率的关系，随访时间为 5 年。
马来西亚塞格马特的东南亚社区观察站提供了儿童及其父母的纵向数据（n=1767）。儿童在基线时（2013-14 年）年龄为 6-14 岁，随访 5 年后。线性多层模型根据基线 BMI z 评分和潜在混杂因素调整了儿童随访时 BMI z 评分的关联。预测因素包括父母的心血管代谢健康状况（超重/肥胖、中心性肥胖、高血压、高血糖）和社会人口统计学因素（种族、就业、教育）。逻辑多层模型探讨了肥胖发病率的预测因素。
基线 BMI z 评分较高预测了儿童到青春期后期（0.60；95%置信区间：0.55，0.65）和青春期早期到后期（0.76；95%置信区间：0.70，0.82）的随访 BMI z 评分更高。儿童随访时 BMI z 评分与父母心血管代谢风险因素之间的关联证据不一致，独立于基线儿童 BMI z 评分。例如，母亲肥胖而不是超重，预测了儿童到青春期早期 BMI z 评分较高（超重：0.16；95%置信区间：-0.03，0.36，肥胖：0.41；95%置信区间：0.20，0.61），而父亲超重但不肥胖，预测了青春期早期到后期 BMI z 评分较高（超重：0.22；95%置信区间：0.01，0.43，肥胖：0.16；95%置信区间：-0.10，0.41）。父母肥胖一直预测青春期早期到后期的五年肥胖发病率，但不能预测儿童到青春期早期的肥胖发病率。基线时没有肥胖的青少年，如果父母肥胖，那么在随访期间肥胖的可能性增加 3-4 倍（发病率 OR=3.38（95%CI：1.14-9.98，母亲）和 OR=4.37（95%CI 1.34-14.27，父亲）。
与任何父母或社会人口统计学因素相比，基线时 BMI z 评分较高是预测随访时 BMI z 评分较高的更强预测因素。基于父母或社会人口统计学因素的预防措施是没有必要的，但儿童早期仍然是普遍肥胖预防的关键时期。
