“What we see in our data is that indeed, early onset of menopause is associated with increased risk of type 2 diabetes, and this association is independent of potential intermediate risk factors: obesity, insulin, glucose, inflammation, but also of estradiol and other endogenous sex hormone levels,” said Taulant Muka, MD, PhD, in an interview.
Dr. Muka, a postdoctoral fellow at Erasmus Medical College, Rotterdam, the Netherlands, presented the analysis from a large, prospective cohort of menopausal women at the annual meeting of the North American Menopause Society.
Corne Bouman
Dr. Taulant Muka
In an interview, Dr. Muka noted that the study investigated whether the association between early age at natural menopause (ANM) and type 2 diabetes is independent of intermediate risk factors for type 2 diabetes, such as obesity. Finally, the study also assessed whether endogenous sex hormone levels play a role in the link between early ANM and type 2 diabetes.
Enrollment in the Rotterdam study has been continuing in waves since the 1990s, with enrollment for participants in this particular study cohort occurring in 1997 and with additional cohorts enrolled in 2000-2001 and 2006-2008.
“I think this is the first prospective study with such long follow-up data and with a broad adjustment for confounding factors. Previous studies have been mainly cross-sectional, providing conflicting results,” said Dr. Muka.
Using self-reported age at menopause, the investigators excluded from the study women whose menopausal status was not known, who were actually not menopausal, or whose menopause had not occurred naturally. The study population also excluded women with prevalent type 2 diabetes or for whom no information about diabetes follow-up could be found.
The remaining 3,210 women who were included in the study had a median age of 67 years and had reached menopause at a median of 49.9 years. Most women (82.6%) had an ANM of 45-55 years; ANM for 8.8% was 40-44 years, while just 2.3% had an ANM of under 40 years. Mean body mass index was 27.1 kg/m2.
Participants were considered to have incident diabetes based on several sources: a general practitioner’s records, hospital discharge paperwork, or glucose measurements from visits during the Rotterdam study. Participants also were classified as having diabetes if they used a hypoglycemic medication or had a fasting blood glucose level of at least 7 mmol/L (126 mg/dL) or, in the absence of a fasting blood glucose measurement, a nonfasting blood glucose of at least 11.1 mmol/L (200 mg/dL).
Dr. Muka said he and his coinvestigators identified and adjusted for a large number of variables, using a series of three Cox proportion hazard models.
The first model adjusted for age, which wave of enrollment (cohort) participants were in, hormone therapy status, age at menarche, and the number of pregnancies that reached at least 6 months’ gestation.
The second model used all of the factors in model 1, and added BMI and glucose and insulin levels. The third model used all of the factors in model 2, and also added total cholesterol level, systolic blood pressure, the use of lipid-lowering or antihypertensive medications, alcohol and tobacco use, educational level, prevalent cardiovascular disease, and C-reactive protein levels.
The association of early ANM with the risk of type 2 diabetes was statistically significant in all three models (P less than .001), with very similar hazard ratios (HRs) in all models. For the third, the most comprehensive model, the HR was 1.42 (95% confidence interval, 0.83-2.45).
Extensive sensitivity analyses were carried out, and the association held independent of physical activity level, smoking status, use of hormone therapy, and age. The investigators also used multivariable analyses to ensure that the effect was not mediated by serum levels of thyroid-stimulating hormone, dihydroepiandosterone (DHEA) and DHEA sulfate, estradiol, testosterone, sex hormone–binding globulin (SHBG), or androstenedione.
This study was the first in this population to obtain and adjust for serum sex hormone and SHBG levels, said Dr. Muka.
The prospective design of the study and the long follow-up period were study strengths, said Dr. Muka. Additionally, blood glucose readings taken at study visits, together with electronically linked pharmacy dispensing records, were used for incident diabetes diagnoses.
Study limitations, Dr. Muka said, included the possibility of survival bias, since enrollees “may represent survivors of early menopause who did not develop [type 2 diabetes] or die prior to enrollment.” However, he said, this would mean that “we have underestimated the association, so the risk would be even higher.” Also, all study participants were white, so the results cannot be extrapolated to nonwhite populations.
Dr. Muka said that the largely American audience for his presentation was interested in the fact that the association existed independent of BMI, an obesity marker, based on questions and comments following the talk. “Indeed, we stratified the analysis, and we didn’t find any difference between participants who were obese and nonobese.” Also, he said that there were queries about whether the analyses had corrected for DEXA measurements, in order to assess lean versus fat body composition more accurately. This analysis has not been done, but Dr. Muka plans to complete it on his return to Rotterdam, he said.
Up to 10% of women will reach menopause before the age of 45, said Dr. Muka, so this analysis has the potential to impact primary care for millions of women. “Women who undergo early menopause may be a target for type 2 diabetes prevention measures and might need to be screened for other cardiovascular risk factors like high blood pressure and dyslipidemia, since they are also at risk for cardiovascular disease.”
Dr. Muka reported no outside funding sources and had no relevant financial disclosures.
On Twitter @karioakes