Same age. Same waist. Same BMI. Different blood sugar, and the testosterone gap opened anyway.
Your drive went quiet somewhere in the last two years, and you noticed long before anyone else did. Strength is down. The weight sits around the middle and will not move. Someone looked at your waistline and told you that was the answer, and you believed it, because that is the answer everyone gives.
A 2024 study of 160 men in Alexandria, Egypt does not fit that story.
The lean men with type 2 diabetes came in lower than the obese men whose blood sugar was normal. The number you have been watching may not be the one doing the work.
How the comparison was built
The researchers recruited 160 men aged 27 to 45 at a diabetes clinic in Alexandria and split them into four groups of 40. Lean with type 2 diabetes, BMI under 25. Obese with type 2 diabetes, BMI 30 or above. Lean with normal blood sugar. Obese with normal blood sugar. Blood was drawn in the morning before 10 am, after 8 to 10 hours of fasting. Total testosterone, SHBG and HbA1c were measured. Free testosterone and HOMA-IR, a standard index of insulin resistance, were calculated.
The design choice that matters: the 80 men with diabetes and the 80 without were matched on the three variables that usually explain low testosterone. Age, median 43 against 42 years, p 0.21. BMI, median 27.45 in both groups, p 0.67. Waist circumference, median 94.5 against 95.5 cm, p 0.82. Same age, same shape. What separated them was glucose.
What the blood showed
Median total testosterone was 297.5 ng/dL in the men with diabetes and 510 ng/dL in the men without. Median calculated free testosterone was 6.15 ng/dL against 9.22 ng/dL. Both differences reached p below 0.001.
Then the four subgroups, where the study earns its title. Lean men with diabetes: median 302 ng/dL. Lean men without diabetes: 505 ng/dL. Obese men without diabetes: 510 ng/dL. Obese men with diabetes came in lowest of all, at 284.5 ng/dL.
Two numbers accounted for about half of it
In the regression model, HOMA-IR and HbA1c together accounted for approximately 51.3 percent of the variability in total testosterone, an adjusted R-squared of 0.513. Both coefficients were negative. Higher insulin resistance and higher HbA1c went with lower testosterone.
The gap between the groups on those two inputs was wide. Median HbA1c was 8 in the diabetes group and 5.2 in the non-diabetes group. Median HOMA-IR was 3.79 against 1.1.
Read that figure carefully. Accounting for 51.3 percent of variability is not the same as causing it. This is a statistical fit inside one sample of 160 men. It is not a mechanism, and it is not a prediction about you.
Free testosterone told a slightly different story
A second model looked at calculated free testosterone. Four variables accounted for roughly 45.1 percent of its variability: duration of diabetes, waist circumference, age and HOMA-IR. All four coefficients were negative.
Duration deserves a second look. Mean duration of diabetes here was 1.15 years, range 0 to 6. These were not men with decades of disease behind them. The signal showed up early.
The carrier protein fell too
SHBG is the protein that binds most of the testosterone circulating in your blood. Median SHBG was 21.7 nmol/L in the men with diabetes against 42.15 nmol/L in the men without, p below 0.001.
That closes an obvious loophole. When SHBG drops, total testosterone can drop with it while the free, active fraction holds steady, which would make the headline finding an artifact. The free fraction did not hold steady. It fell in the lean diabetic men against the lean controls, and in the obese diabetic men against the obese controls, both at p below 0.001.
Waist beat BMI, and neither was strong
Total testosterone correlated negatively with BMI at r -0.16, p 0.04, and with waist circumference at r -0.23, p 0.003. Free testosterone correlated with BMI at r -0.26 and with waist circumference at r -0.3, both at p below 0.001.
Waist circumference outperformed BMI in every pair. But be honest about the size of these correlations. An r of -0.16 is a faint signal. Body size was related to testosterone in this sample. It was not the main event.
What this study cannot tell you
- It is observational, case-control, measured at a single point in time. It shows association, not causation. Low testosterone has separately been reported as a risk factor for later type 2 diabetes, so the relationship may run either way. This design cannot settle which.
- It is 160 men at one hospital in one Egyptian city. The authors state plainly that their results were valid only among the population they studied. Post hoc power was estimated at 83 percent.
- Every participant was between 27 and 45 years old. It says nothing about men in their late fifties.
- Men taking insulin were excluded, because insulin interferes with the HOMA-IR calculation. The more advanced cases are absent from the data.
- Total testosterone was measured by immunoassay, not mass spectrometry. Free testosterone was calculated with the Vermeulen method, not measured directly.
That does not erase the finding. It bounds it.
The question worth asking
Every value in this study came from ordinary laboratory work. A fasting morning sample. Testosterone, SHBG, HbA1c, glucose, insulin. Nothing exotic. What the researchers did differently was read those results as one set instead of one line at a time.
If your testosterone has ever been checked in isolation, that result was blind to the two variables that accounted for about half the variation across these 160 men. What a specialist does with hormone and metabolic results read together is set out on the consultation page.
The men in this study were lean. It did not save them.