Distinct weight gain patterns in midlife women tracked with future diabetes risk
Midlife women in the United States did not follow a single predictable path as body size changed across the menopause transition. In the Study of Women’s Health Across the Nation, investigators identified separate trajectories for BMI and waist circumference, and those patterns carried very different diabetes risks over time.
Among 1,406 women followed across 10 annual visits from 1996 to 2008, 90 participants, or 6.4%, developed diabetes during the later follow-up period. The study used the first five visits to classify body mass index and waist circumference trajectories, then linked those patterns to diabetes occurring during visits 6 through 10. The design matters because it moves beyond a single baseline weight measurement and instead captures how adiposity evolves during the years when menopause-related metabolic shifts often accelerate.
A single number on a clinic scale can miss the biology unfolding underneath.
BMI trajectories carried a graded rise in diabetes hazard
Four BMI trajectory groups emerged from the longitudinal modeling: low stable, medium increasing, high increasing, and very high increasing. The low stable group made up 37.4% of participants, while the medium increasing group accounted for 34.8%, the high increasing group for 19.5%, and the very high increasing group for 8.3%.
Diabetes risk rose steeply across those categories. Compared with the reference pattern, the hazard ratio for diabetes was 3.83 in the medium increasing group, 6.97 in the high increasing group, and 14.22 in the very high increasing group. The pattern is hard to ignore: as BMI moved upward more sharply over time, diabetes incidence climbed in parallel. In practical terms, gradual and moderate weight gain during midlife was not benign, and the highest BMI trajectory identified a group at especially elevated risk.
These findings fit the metabolic changes that accompany menopause. Declining endogenous oestradiol is linked with central fat redistribution, reduced energy expenditure, and worsening insulin secretion and insulin sensitivity. In that setting, upward drift in BMI may reflect more than excess weight alone; it may signal a shift toward a more diabetogenic hormonal and metabolic state.
Waist circumference separated central adiposity from overall body size
Waist circumference provided a second lens on risk, and the signal was at least as strong. Four waist circumference trajectory groups were identified: increasing, normal increasing, high increasing, and very high stable. The two largest groups were normal increasing at 38.4% and increasing at 35.9%, followed by high increasing at 18.7% and very high stable at 6.9%.
Here too, diabetes hazards rose with adverse patterns. The hazard ratio was 2.26 in the increasing group, 8.12 in the high increasing group, and 11.30 in the very high stable group. Central adiposity appears to have captured a clinically important dimension of risk that BMI alone can miss. Visceral fat is metabolically active tissue, not passive storage; it is tied to insulin resistance, inflammation, and a more atherometabolic profile, which makes waist size a biologically plausible predictor of future diabetes.
Combined BMI and waist trajectories identified the highest-risk profiles
The analysis also combined BMI and waist circumference into joint trajectory groups, and those profiles sharpened the risk stratification further. Four multi-trajectory groups emerged. Group 1, representing 34.5% of women, showed high increasing BMI with increasing waist circumference. Group 2, 33.1%, had low stable BMI with normal increasing waist circumference. Group 3, 22.1%, had medium increasing BMI with high increasing waist circumference. Group 4, 10.3%, had very high increasing BMI with very high increasing waist circumference.
Higher diabetes rates appeared in groups 1, 3, and 4. Even without the paper providing a single combined hazard ratio in the abstract, the direction is clinically coherent: women whose overall body mass and central adiposity both worsened were the ones most likely to develop diabetes later. This combined approach reflects real-world physiology better than either measure alone, because two women with the same BMI may carry very different amounts of abdominal fat and therefore very different metabolic risk.
Menopause-related adiposity changes have direct bedside implications
The study’s core message is straightforward: midlife weight trajectories are not interchangeable. A woman whose BMI remains stable and whose waist circumference changes little may face a very different future than a peer whose weight and abdominal girth both rise over successive annual visits. For clinicians, the work supports serial measurement rather than isolated screening. Repeated BMI and waist circumference assessment can identify women whose risk is shifting before hyperglycemia becomes clinically apparent.
The biological rationale is strong. Menopause is associated with central fat redistribution and deteriorating insulin sensitivity, and these changes are precisely the ones that waist circumference is designed to capture. BMI remains useful, but it can blur the distinction between fat mass and lean mass and cannot localize adiposity. Waist circumference adds the missing anatomical signal, especially when paired with longitudinal tracking.
This is a population that may benefit from earlier, more individualized prevention. Weight management interventions around midlife could have meaningful impact if deployed before the highest-risk trajectory groups emerge. The data do not suggest that every woman gains weight in the same way, or that every gain carries the same metabolic consequence. They show a more complicated and more actionable reality: distinct body composition trajectories mark distinct diabetes futures.
What this cohort can and cannot establish
This was a secondary analysis of observational data, so the results identify associations rather than proving causation. The study also focused on trajectories during specific visit windows, which means the findings describe patterns within this cohort and time frame rather than a universal rule for all women. Still, the signal is consistent across both BMI and waist circumference: more adverse upward trajectories aligned with more diabetes.
The sample included 1,406 women, and 90 developed diabetes during follow-up. Those numbers are sufficient to reveal broad trajectory patterns, but they do not substitute for an intervention trial that tests whether targeting women in the high-risk groups can actually prevent diabetes. Even so, the current analysis gives clinicians something concrete to work with: repeated anthropometric measurements in midlife are not routine paperwork. They are a window into future metabolic risk.
Future prevention strategies will need to determine whether serial waist measurement, in particular, should be incorporated more systematically into menopausal care, especially for women whose BMI is rising and whose abdominal adiposity is expanding in tandem.