Hormone Balance May Shape Blood Clot Strength in Men With Type 2 Diabetes

A single blood test that captures the balance between two powerful sex hormones may reveal something unexpected about how strongly a man’s blood clots. New research published in BMC Endocrine Disorders suggests that in men living with type 2 diabetes, the ratio of estradiol to testosterone is linked to the physical robustness of the clots their blood forms in the laboratory. The finding, drawn from a secondary cross-sectional analysis of 92 hospitalized men, adds a fresh hormonal dimension to the study of thrombosis risk in diabetes, a population already known to carry an elevated burden of cardiovascular disease. But the study also delivers a sobering caveat: the ratio, however intriguing, did not outperform testosterone alone at explaining clot strength, and its authors are careful to frame the result as hypothesis-generating rather than practice-changing.

The science at the heart of the study is thromboelastography, or TEG, a technique that goes far beyond the standard clotting time measured in routine coagulation panels. Instead of asking merely how long blood takes to clot, TEG measures the viscoelastic properties of a clot as it forms in a small cup of whole blood, tracking how quickly it develops, how strong it becomes, and how stable it remains. The headline parameter in this research is the maximum amplitude, abbreviated MA, which reflects the ultimate strength of the clot and depends heavily on platelet function and platelet-fibrin interactions. A higher MA signals a stiffer, more robust clot, a profile that clinicians describe as procoagulant or hypercoagulable. In patients with type 2 diabetes, previous work has associated testosterone deficiency with just such a procoagulant TEG signature, raising the question of whether the hormonal milieu more broadly, rather than testosterone in isolation, shapes clot behavior.

That question led the study’s author, Xiao-Jun Chen of the Department of Endocrinology at Shanghai Tenth People’s Hospital’s Chongming Branch and Anting Hospital in Shanghai, to examine the estradiol-to-testosterone ratio, calculated by dividing serum estradiol in picomoles per liter by total testosterone in nanomoles per liter. Estradiol, often thought of as a female hormone, is present in men as well, both produced directly and converted from testosterone through the enzyme aromatase. Because this conversion means estradiol and testosterone are biochemically intertwined, some investigators have proposed that the ratio between them may capture a more nuanced picture of relative androgen-estrogen balance than either hormone measured alone. Whether that added nuance translates into additional information about thrombosis risk has remained uncertain, and it is precisely the gap this analysis set out to probe.

The study population consisted of 92 hospitalized men with type 2 diabetes drawn from a previously published single-center cohort, with complete TEG data available for 84 of them. All participants had provided written informed consent under a protocol approved by the Ethics Committee of Chongming Branch, Shanghai Tenth People’s Hospital, and conducted in accordance with the Declaration of Helsinki. Clot testing was performed on a TCA-6000 analyzer using citrated whole blood with kaolin activation, a standard configuration in routine TEG work. The median estradiol-to-testosterone ratio in the cohort was 10.92, with an interquartile range spanning 5.97 to 17.23, indicating substantial variability in hormonal balance even within this relatively homogeneous group of diabetic men.

The first striking result emerged when participants were sorted into tertiles, or thirds, according to their hormone ratio. Across increasing tertiles of estradiol-to-testosterone, the median TEG maximum amplitude climbed steadily from 59.4 to 62.4 and then 65.9 millimeters, a graded relationship that reached statistical significance at P equal to 0.002. In other words, men whose blood showed a relatively higher estrogen signal compared with testosterone tended to form measurably stronger clots. Because MA is a direct readout of clot firmness, this pattern hints that a shift in hormonal balance toward relatively more estrogenic activity might accompany a more procoagulant state in diabetic men, though a cross-sectional snapshot like this cannot establish which way any causal arrow would point.

To test whether the association could survive statistical scrutiny, the analysis turned to multivariable linear regression, adjusting for age, body mass index, glycated hemoglobin, and fasting glucose, the usual suspects that might otherwise confound a hormone-clot relationship. After natural-log transformation of the ratio, the adjusted model estimated that each unit increase in log-transformed estradiol-to-testosterone was associated with a 2.85 millimeter increase in maximum amplitude, with a 95 percent confidence interval of 0.85 to 4.85. The conventional P value was 0.006, and the result held up under more demanding statistical treatment: a false-discovery-rate correction yielded an adjusted P of 0.024, and heteroscedasticity-consistent HC3 standard errors, which guard against uneven variability in the data, produced a P of 0.021. For an exploratory secondary analysis, that level of robustness is notable.

What truly distinguishes the study, however, is the battery of sensitivity analyses deployed to make sure the finding was not an artifact of a single influential data point or a fragile modeling choice. The researchers excluded the participant with the maximum estradiol-to-testosterone value, and the estimate not only survived but grew slightly, to 3.40 with an HC3 P of 0.014. Winsorization, a technique that caps extreme values rather than deleting them, left the association intact at 3.00 with a P of 0.017. Screening out observations with excessive influence by Cook’s distance, a diagnostic that flags data points capable of distorting a regression, gave an estimate of 3.29 with a P of 0.002. Most convincingly, a leave-one-out procedure refit the model 84 separate times, each time removing a different participant, and the coefficient stayed positive in every refit, ranging from 2.36 to 3.44, with all HC3 P values below 0.05. The signal, in short, was not carried by any one man’s blood sample.

Yet the study’s most instructive result may be the one that did not favor the ratio. When the authors compared a model built on testosterone alone against the estradiol-to-testosterone model using the Akaike information criterion, a statistic that balances goodness of fit against model complexity, the testosterone-based model fit maximum amplitude slightly better, with an AIC of 525.2 versus 526.2. The difference is small, but its direction matters: the ratio, despite its theoretical appeal as an integrated index of hormonal balance, did not add explanatory power beyond what testosterone already provided. The author’s conclusion is accordingly measured. The estradiol-to-testosterone ratio should be viewed as a descriptive index of relative androgen-estrogen balance rather than an established independent thrombotic biomarker, and the association with clot strength should be treated as hypothesis-generating pending prospective confirmation.

That caution is well placed given the study’s design and scope. A cross-sectional analysis captures a single moment in time and cannot determine whether hormonal imbalance drives clot strength, whether an underlying metabolic or inflammatory state influences both, or whether reverse causation is at work. The cohort was modest in size, confined to hospitalized men with type 2 diabetes at a single center, and the analysis was secondary, meaning it repurposed data collected for a parent study rather than being designed from the ground up to test the hormone-clot question. Hormone levels also fluctuate with time of day, acute illness, and assay characteristics, all of which can introduce noise in a hospitalized population. Prospective studies that follow men over time, measure hormones repeatedly, and track actual thrombotic events would be needed to establish whether the ratio, or testosterone itself, has genuine predictive or clinical value.

Even so, the research opens a tantalizing line of inquiry at the intersection of endocrinology and hematology. Sex hormones are known to influence vascular biology in multiple ways, from effects on platelet reactivity to endothelial function and fibrinolysis, and diabetes amplifies cardiovascular risk through overlapping metabolic and inflammatory pathways. If a simple, widely available hormone panel could one day help stratify thrombotic risk in diabetic men, it would offer an accessible complement to existing risk scores. For now, the message is one of disciplined curiosity: a higher estradiol-to-testosterone ratio tracks with stronger clots in this population, the association is statistically durable, but the ratio does not beat testosterone alone, and no one should change clinical practice on the basis of a single exploratory analysis. The next chapter of this story will be written by prospective studies designed to answer the question this work could only raise.

Subject of Research: Association between the estradiol-to-testosterone ratio and thromboelastographic clot strength in men with type 2 diabetes

Article Title: Estradiol-to-testosterone ratio versus testosterone in relation to thromboelastographic clot strength in men with type 2 diabetes: a secondary cross-sectional analysis

Article References: Chen, X.-J. (2026). Estradiol-to-testosterone ratio versus testosterone in relation to thromboelastographic clot strength in men with type 2 diabetes: a secondary cross-sectional analysis. BMC Endocrine Disorders. https://doi.org/10.1186/s12902-026-02545-y

Image Credits: AI Generated

DOI: 10.1186/s12902-026-02545-y

Keywords: type 2 diabetes mellitus, estradiol-to-testosterone ratio, testosterone, thromboelastography, maximum amplitude, clot strength, hypercoagulability, cross-sectional analysis, hormonal balance, cardiovascular risk, statistical sensitivity analysis, Estradiol-to-testosterone