August 2, 2023

Prediabetes as an Independent Risk Factor for Bone Fractures in Women

Investigating bone health and prediabetes during the menopausal transition
Having prediabetes before MT was associated with higher rates of fracture than those participants that did not have prediabetes.


Shieh A, Greendale G, Cauley J, Karvonen-Gutierrez C, Karlamangla A. Prediabetes and fracture risk among midlife women in the Study of Women’s Health Across the Nation,1980-2002. JAMA Network Open. 2023; 6(5): e2314835.

Study Objective

To examine whether prediabetes in premenopausal, midlife women is associated with postmenopausal bone fractures in those who do not develop overt diabetes. 

The second objective of this study was to assess if there was an association between prediabetes and bone fractures regardless of bone mineral density in midlife women undergoing menopausal transition (MT). 

Key Take-Away

Having prediabetes before MT was associated with higher rates of fracture than those participants that did not have prediabetes.


Longitudinal cohort observational study called the Study of Women’s Health Across the Nation (SWAN) 


Study of Women’s Health Across the Nation (SWAN) Cohort consisted of 3,302 community-dwelling women between the ages of 42 and 52 years of age. They were either in premenopause (exhibiting no change in menstrual bleeding) or early perimenopause (less predictable bleeding within a 3-month period).

The analysis group consisted of 1,690 women. The women’s ethnicity was as follows:

  • 437 Black women 25.9%
  • 197 Chinese women 11.7%
  • 215 Japanese women 12.7%
  • 841 White women 49.8%

The mean BMI at the start of the MT for the participants 27.6 kg/m2.

Mean bone mass density (BMD) for the lumbar spine (LS) was 1.059 g/cm2; and for the femoral neck (FN) it was 0.828 g/cm2.

SWAN Cohort participants were excluded if they no longer had an intact uterus and at least 1 ovary. They were also excluded if they were using hormone therapy or hormone-based contraception.

The SWAN Cohort came from 7 different clinical sites across the United States: Boston, Massachusetts; Chicago, Illinois; Detroit, Michigan; Pittsburgh, Pennsylvania; Los Angeles, California; Newark, New Jersey; and Oakland, California.

The SWAN Bone Cohort consisted of 2,365 women from 5 cities: Boston, Massachusetts; Detroit, Michigan; Pittsburgh, Pennsylvania; Los Angeles, California; and Oakland, California.

All SWAN Bone Cohort participants had 1 baseline visit at inception of the study and 16 consecutive follow-up visits at a median time of 1.1 years apart (interquartile range, QR, 1.0-1.4 years).

The authors defined the start of menopause transition (MT) as “the first visit in late perimenopause (less predictable menstrual bleeding at least once every 3-12 months).” For women who transitioned directly from premenopause or early perimenopause to postmenopause, the authors defined the start of MT as the first postmenopausal visit.

To be included in analysis, SWAN Bone Cohort participants needed to:

  • Complete at least 1 or more visits in the study 
  • Have at least 1 study visit after MT (to check for fractures)
  • Not be taking bone-beneficial medications including hormone therapy, calcitonin, calcitriol, bisphosphonates, denosumab, and parathyroid hormone (33 women were removed from the study)
  • Not be diagnosed with type 2 diabetes (94 women were removed from the study)

SWAN Bone Cohort participants were excluded if they:

  • Did not have any study visits before MT
  • Began taking bone-beneficial medications before MT
  • Had diabetes type 2 (a fasting blood glucose level of 126 mg/dl before or during MT)
  • Were taking any of the following metformin, sulfonylurea, meglitinide, thiazolidinedione, dipeptidyl peptidase 4 inhibitors, glucagonlike peptide-1 receptor agonists, or insulin
  • Had not had follow-up visits between when they started MT and when they got a fracture.

The mean of the follow up was 12 years.

Study Parameters Assessed

At the baseline visit, any pre-SWAN fractures were recorded. During the seventh visit, fractures were formally recorded again. Craniofacial and digital fractures were excluded. Traumatic fractures that occurred during a motor vehicle accident, rapid movement, playing sports, or from impact with heavy or fast-moving projectiles were also excluded.

All other traumatic and atraumatic fractures were included in the study.

Prediabetes (glucose levels between 100 and 125 mg/dl) at study visits was tracked:

  • Participants who never had a single measure of prediabetes had a score of 0 
  • Participants who had prediabetes values at every visit until MT had a score of 1 
  • Participants who had a prediabetes for at least 1 visit but not all the visit had a score between 0 and 1

Controls were set up for variables related to the potential of fractures at the age of MT:

  • Cigarette use at the age MT
  • BMI at age of MT
  • Ethnicity 
  • Study site
  • Bone-detrimental medications before the age of MT
  • Bone-detrimental medication during the study

There was an adjustment for BMD at MT in either the lumbar spine (LS) or femoral neck (FN). These were measured by dual x-ray absorptiometry.

Primary Outcome

Incident fractures during and after MT and prediabetic status before MT

Key Findings

The follow up from the start was 12 years.

  • 56 women had fractures before MT.
  • 136 sustained fractures during the MT or after MT.
  • 225 women had prediabetes and sustained fractures ( 11.1%).
  • 111 of the women without diabetes also sustained fracture (7.6%).
  • 33 started bone-benefiting medications and were suppressed from the study.

Participants who had prediabetes at 50 % visits before MT were 49% more likely to have a fracture after MT or in postmenopause than those with no prediabetes at any MT.

Participants who had prediabetes at every visit before MT were 120% more likely to have a fracture after MT or in postmenopause than those with no prediabetes at any MT.

There was an observed fracture hazard of 6.3 per 1,000 person-years for women with no prediabetes in any visits before MT. 

There was an increase in absolute fracture hazard of 3 per 1,000 person-years for women who had prediabetes in half of the pre-MT visits.

For those women who had prediabetes during every pre-MT visit, the increase in fracture hazard was 7 per 1,000 person-years.

Interestingly, this fracture increase was independent of BMD and overt diabetes type 2 . These finding suggest that the increase in fracture risk with prediabetes may not be associated with BMD, but may represent a separate mechanism altogether.

Of the women in the study who were noted as having prediabetes at a least 1 visit before MT, a larger percentage were Black, Chinese, or Japanese participants.


Greendale reported receiving grants from the National Institutes of Health (NIH) during the conduct of the study. Karvonen-Gutierrez reported receiving grants from the NIH during the conduct of the study and grants from the NIH outside the submitted work. Karlamangla reported receiving grants from NIH during the conduct of the study and personal fees from OptumRx outside the submitted work. No other disclosures were reported.

Practice Implications & Limitations

The most practical implication from this study revolves around simple education and testing for our female patients.

It is imperative that women in the pre-midlife years be informed of their pre-diabetes or diabetes status when they come into our offices. According to the CDC, a crude estimate between 2017-2020 showed that approximately 38% of Americans aged 18 years or older were in the prediabetic range based on fasting glucose or HbA1c levels tested. Of that 38%, only 19% reported being told by their medical professional about the diagnosis.1

Neglecting to share patient diagnosis with patients is not an acceptable standard of care, especially when prediabetes correction through lifestyle changes has been well documented.

Running fasting glucose and HbA1c, along with other standard preventative medicine panels, on each patient and communicating the results to them can be one of the most eye-opening aspects of our treatment plans with our patients.

Communicating the diagnosis of prediabetes and the associated health complications that arise from untreated prediabetes, such as the initiation and progression macrovascular disorders including CVD, stroke, peripheral vascular disease and, now, increased risk of bone fracture in midlife, can improve patients’ compliance to prediabetes treatment plans. Helping patients undertake lifestyle changes, including diet, exercise, stress management, and weight loss when indicated, can reverse prediabetes and prevent future disease processes.2

The second clinical implication from this study is to start conversations about bone health with our female patients when they are in their 30-40s.  This is where true prevention can begin and be reversed especially if they have been diagnosed with prediabetes.

According to a paper published the Journal of Women’s Health, conversations about declining bone health usually occur with medical professionals in post-menopausal women approximately 10 years after MT, once a bone fracture has already occurred.3

Having conversations about bone health only after a fracture has occurred is a missed opportunity to discuss prevention practices. Creating medical conversations in young adult women before menopause can create healthier bone health after MT. Sharing that prediabetes is part of the conversation about bone health will be a surprise twist for many patients who likely have no idea about the association of prediabetes with fracture after menopause.

Finally, it was noted in the study that women of color, specifically Black, Chinese, and Japanese women, had higher incidence of prediabetes in the study. There needs to be particular attention paid in educating these women about current and future outcomes related to prediabetes and fractures.

A new body of research shows a dangerous racial bias toward Black women in the healthcare system, causing Black women’s health concerns to not be taken seriously. In some instances, the neglect has been life-threatening.4

Making sure to do simple testing for prediabetes and sharing the results and treatment options with our Black women patients is essential to change that narrative and demonstrate to Black women patients that their health is being taken seriously by healthcare providers.

Research shows that racial discrimination, which has a long history in the United States, has negative implications for Black women’s health today. Understanding that Black women have higher incidence of prediabetes and sharing their results and treatment options during visits begins to correct the racial bias that has been noted in this research.5

A limitation noted by the study’s authors is that while there is recognition that rates of diabetes and fractures differed by race and ethnicity, there was not testing to show associations between prediabetes and fractures differed by race and ethnicity.

Categorized Under


  1. Prevalence of prediabetes among adults | Diabetes | CDC. Accessed August 1, 2023.
  2. Prediabetes and cardiovascular disease: pathophysiology and interventions for prevention and risk reduction. Endocrinol Metab Clin North Am. 2018; 47(1): 33-50.
  3. McPhee C, Aninye IO, Horan L. Recommendations for improving women's bone health throughout the lifespan. J Womens Health (Larchmt). 2022;31(12):1671-1676.
  4. Sacks TK. Performing Black womanhood: a qualitative study of stereotypes and the healthcare encounter. Critical Public Health. 2018; 28:1:59-69.
  5. How discrimination can harm black women’s health. News. Published October 31, 2018. Accessed August 1, 2023.