MARKERS OF ANDROGEN DEFICIENCY IN PATIENTS WITH CHRONIC SPINAL CORD INJURY: DEVELOPMENT OF A PREDICTIVE NOMOGRAM

Spinal cord injury is a condition with a high prevalence of comorbidities, including androgen deficiency which is accompanied by metabolic alterations and reduced sexual desire. Low testosterone (T) may be the biochemical expression of hypogonadism; however, such a diagnosis is challenging in SCI as the clinical features of androgen deficiency may represent direct or indirect consequences of the disability. We aimed to identify clinical/biochemicals markers to predict androgen deficiency in men with chronic SCI, in order to conceive a nomogram supporting clinical decision for T levels assessment.

124 patients with chronic clinically stable SCI underwent neurological and clinical evaluations.

Leisure Time Physical Activity was registered by a validated questionnaire. After 12 h of fasting, serum T, SHBG, insulin and other biochemical parameters were measured.

Androgenic deficiency was found in 46 patients, showing more comorbidities and higher BMI, glycaemia, insulin, HOMAi and triglycerides compared to subjects with normal T levels. The group with androgen deficiency exhibited lower total and free calculated T as well as a lower LTPA. Univariate logistic regression detected several predictors; however, an adjusted multiple logistic regression analysis confirmed an independent association with low T only for the HOMAi, BMI and LTPA. According to the ROC curve, our nomogram discriminated subjects with androgen deficiency with a sensitivity of 89.5% and a specificity of 69%. The internal validation test proved a non-significant discrepancy between the probability predicted and that observed.

Low LTPA, high BMI, and high HOMA-index are independent predictors of low T levels in men with chronic SCI; these routine assessments could therefore represent, in combination, a useful criterion to rule out whether to test T level in a given patient. Until the actual significance of androgen deficiency in such population will be clarified, an intervention on these modifiable risk factors would represent a safer approach than T replacement.

Every author has any conflict of interest to declare.