Table 1.
Mendelian randomization estimates for the effect of testosterone on cardiovascular risk factors using variants in 10q21 (JMJD1C) gene region.
| Outcome | Estimate (SE) | p-Value |
|---|---|---|
| HDL-c | 0.085 (0.068) | p = 0.21 |
| LDL-c | 0.293 (0.073) | p < 0.0001 |
| Triglycerides | −0.229 (0.066) | p < 0.0001 |
| Adiponectin | 0.103 (0.070) | p = 0.14 |
| Systolic blood pressure | −2.236 (0.578) | p < 0.0001 |
| Diastolic blood pressure | −1.644 (0.331) | p < 0.0001 |
| BMI | −0.055 (0.056) | p = 0.33 |
| BMI (men) | 0.017 (0.074) | p = 0.82 |
| BMI (women) | −0.120 (0.070) | p = 0.084 |
| Height | 0.059 (0.046) | p = 0.20 |
| Height (men) | 0.135 (0.097) | p = 0.16 |
| Height (women) | −0.065 (0.089) | p = 0.47 |
| Male pattern balding | −0.051 (0.125) | p = 0.69 |
Mendelian randomization estimates for high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, triglycerides (all SD units), adiponectin (μg/ml, log-transformed), systolic and diastolic blood pressure (mmHg), body mass index (BMI, SD units) and height (SD units), and male pattern baldness (log odds ratio, positive values favour more balding). Estimates (standard errors) are changes in the outcome per unit increase in log-transformed testosterone (stepwise score). Genetic associations were measured in men and women unless otherwise indicated. Italics indicates associations at a nominal level of significance without correction for multiple testing (p < 0.05), bold indicates associations after correction for multiple testing (p < 0.05/7 = 0.007).