Long-term trends in central obesity in England: an age-period-cohort approach

Central obesity measures, such as waist circumference (WC), waist-to-hip ratio (WHR), and waist-to-height ratio (WHtR) have previously outperformed body mass index (BMI) in predicting health risks. BMI has been shown to underdiagnose obesity in older adults.

We used data from the Health Survey for England (2005-2021) for 120,024 individuals aged 11-89 years, born in 1919-2008. High-risk classifications for WC, WHR, WHtR, and BMI were defined using established thresholds (World Health Organisation and the UK National Institute for Health and Care Excellence). Age, period (changes over time), and cohort effects were assessed using logistic regression with grouped variables to address the identification problem inherent in age-period-cohort (APC) models.

The prevalence of high-risk increased over time for all obesity measures. Central obesity measures showed a consistent linear increase with age until around 70 years of age. BMI exhibited an inverted U-shaped age trend. Obesity increased over time across all measures, while there was little evidence for a cohort effect. WHtR trends closely mirrored BMI at the population level but identified different high-risk individuals. The odds of high-risk WHtR increased with age, with odds ratios (OR) 4.91 (95% CI: 1.95-12.39) for females and 6.15 (95% CI: 2.24-16.89) for males by 85-89 years compared to 18-19 years. Period effects for WHtR showed ORs of 1.41 (95% CI: 1.16-1.72) for females and 1.25 (95% CI: 1.01-1.55) for males in 2019-2021 compared to 2005-2006.

Central obesity measures, particularly WHtR, could provide a more consistent reflection of age-related increases in obesity risk. The linear increase in high-risk with age for central obesity measures aligns better with known age-related increases in obesity-related comorbidities. Age plays a significant role in driving obesity trends meaning an aging population could leading to further increases in the prevalence of obesity.