Objective  To investigate the relationship between body mass index (BMI) and its changes and the incidence of metabolic syndrome (MS) for providing a reference to MS prevention and control.

  Methods  Totally 9 280 adult residents aged ≥ 18 years were recruited from 12 counties/prefectures/districts of Guizhou province with multistage cluster random sampling for a baseline survey including face-to-face interview, physical examination and laboratory detections from November 2010 through December 2012; subsequent follow-up surveys were carried out in all the residents from December 2016 to June 2020. The association of BMI and its change with MS among the residents were analyzed by the end of follow-up.

  Results  The follow-up time was 5.05 – 9.53 years (averagely 7.01 ± 1.03 years) for the 4 292 adult residents finally included in the analysis. During the follow-up period, totally 860 MS cases were identified; the MS incidence rate (1/1 000 person-years) was 28.60 for all the adult residents and the MS incidence rate was 19.13 for the residents with low weight, 24.32 for the residents with normal BMI, 42.10 for those being overweight, and 53.46 for those being obese, respectively. After adjusting for sex, age, place of residence, smoking, alcohol drinking, intakes of fruits and vegetables (< 400 g/d), oil consumption (> 25 g/d), intake of salt (> 6 g/d), sedentary time, and physical activity, multivariate Cox proportional hazard regression analysis showed that one unit BMI increment at the baseline survey was correlated with a 58% increased risk of MS incidence (hazard risk HR = 1.58, 95% confidence interval 95%CI: 1.43 – 1.75); the residents being overweight and obese were at increased risk of MS compared to the residents with normal BMI, with the HR (95% CI) of 1.70 (1.47 – 1.98) and 2.33 (1.73 – 3.14). Restricted cubic spline analysis indicated that MS risk increased gradually with the increment of BMI from an initial value of 21.94 kg/m². After adjusting for the confounding factors mentioned above and the baseline BMI, multivariate Cox proportional hazard regression demonstrated that compared to those with stable BMI, the MS risk was significantly lower in the residents with decreased BMI (HR = 0.61, 95%CI: 0.48 – 0.78) and higher in the residents with increased BMI (HR = 2.07, 95%CI: 1.74 – 2.46). For the residents with normal BMI, being overweight, and being obese at the baseline survey, a BMI decrement of more than one unit (kg/m²) was associated with reduced MS risk, with the HR (95%CI) of 0.65 (95%CI: 0.45 – 0.93), 0.47 (95%CI: 0.33 – 0.69), and 0.36 (95%CI: 0.17 – 0.77); while for the residents with normal BMI and being overweight at the baseline survey, a BMI dynamic increment of more than one unit (kg/m²) was correlated with increased MS risk, with the HR (95%CI) of 2.57 (95%CI: 2.05 – 3.22) and 1.62 (95%CI: 1.18 – 2.21), respectively. Restricted cubic spline analysis revealed that MS risk increased gradually with the increment of BMI from an initial value of 0.845.

  Conclusion  Both high baseline BMI and dynamic increment in BMI were significantly associated with increased risk of MS incidence and maintaining a normal BMI is an effective measure to prevent MS among adult residents in Guizhou province.