Greater long-term weight variability was associated with modestly higher subsequent triglyceride concentrations, though findings should not be interpreted as causal or independent of obesity.
Key Findings
Results
Each 1 SD increment in weight coefficient of variation (CV) was associated with 2.7% higher triglyceride concentrations in the core model.
β = 0.026, 95% CI: 0.007 to 0.045; p = 0.007
Core complete-case analysis included 5,191 participants
Model adjusted for baseline age, sex, weight slope, number of weight measurements, and working status in 2009
CV was standardized among 9,500 eligible participants
Marginal standardization yielded an adjusted difference in arithmetic mean triglyceride concentrations of 0.048 mmol/L per 1 SD increment in CV
Results
Each 5 kg increment in average successive variability (ASV) was associated with 10.7% higher triglyceride concentrations in the core model.
β = 0.102, 95% CI: 0.057 to 0.147; p < 0.001
Marginal standardization yielded an adjusted difference in arithmetic mean triglyceride concentrations of 0.200 mmol/L per 5 kg increment in ASV
ASV captures directional changes between successive weight measurements
This association was stronger in magnitude than that observed for CV
Results
After further adjustment for baseline BMI, the associations between weight variability and triglyceride concentrations were attenuated but remained statistically significant.
CV association after BMI adjustment: 3.6% higher triglycerides (95% CI: 1.7% to 5.6%)
ASV association after BMI adjustment: 5.6% higher triglycerides (95% CI: 1.0% to 10.3%)
BMI adjustment was conducted in the BMI-complete subsample
Both estimates remained positive and statistically significant after adjustment for adiposity
Results
Multiple imputation for missing data yielded estimates similar to the complete-case analysis.
Multiple imputation was used to assess robustness of findings to missing data
Results were consistent with the core complete-case analysis
This supports the stability of the association across analytic approaches
Results
A four-knot restricted cubic spline analysis supported an overall association between CV and triglyceride concentrations, with no strong evidence of nonlinearity.
Overall association p = 0.016
Test for nonlinearity p = 0.101, indicating the relationship was approximately linear
The spline analysis was conducted to evaluate the functional form of the CV–triglyceride relationship
Methods
The study population consisted of adults from the China Health and Nutrition Survey with at least three measured weights between 1989 and 2006 and a triglyceride measurement in 2009.
9,500 participants were eligible; 5,191 were included in the core complete-case analysis
Weight variability was assessed using two metrics: coefficient of variation (CV) and average successive variability (ASV)
Triglyceride concentrations were measured approximately 3 to 20 years after the weight variability exposure period
Covariates included baseline age, sex, weight slope, number of weight measurements, and working status in 2009
Conclusions
The authors caution that the findings should not be interpreted as causal or independent of obesity due to study design limitations.
The study used an observational design
Weight change has heterogeneous causes including illness, intentional weight loss, and changes in energy balance
Residual confounding remains a possibility
The authors explicitly state: 'the findings should not be interpreted as causal or independent of obesity'
What This Means
This research suggests that people whose body weight fluctuates more over many years tend to have higher triglyceride levels (a type of fat in the blood) compared to people whose weight stays more stable. The study followed thousands of Chinese adults, tracking their weight from 1989 to 2006 and then measuring their triglycerides in 2009. Two different ways of measuring weight instability — one looking at overall variability relative to average weight, and one tracking successive changes between measurements — both showed the same pattern: more weight swings were linked to higher triglyceride concentrations, even after accounting for which direction weight was trending (up or down) and other factors.
The size of the association was modest. For example, each standard deviation increase in weight variability was linked to roughly 2.7–3.6% higher triglyceride levels, and each 5 kg of average successive weight change was linked to about 5.6–10.7% higher triglycerides. These associations held up when the researchers accounted for body mass index (BMI), handled missing data using statistical techniques, and examined the shape of the relationship, which appeared roughly linear rather than showing a sharp threshold effect.
This research suggests that weight instability — not just weight itself — may be relevant to cardiovascular risk markers like triglycerides. However, because this was an observational study, it cannot prove that weight variability directly causes higher triglycerides. Weight can fluctuate for many different reasons, including illness, intentional dieting, or lifestyle changes, and these underlying causes could themselves affect triglyceride levels. The authors emphasize that residual confounding and the complexity of weight change make causal conclusions premature.
Check Your Own Numbers
Upload your bloodwork. We'll cross-reference your results against this study and 4,700 others.
Xu M, Wang H, Zhu L, Liu X. (2026). Long-Term Weight Variability Is Associated with Higher Triglyceride Concentrations After Accounting for Long-Term Weight Slope.. Nutrients. https://doi.org/10.3390/nu18162605