The Endocrine Society has now published a timely evidence-based practice guideline on the evaluation and treatment of androgen deficiency in adult men. The guideline includes detailed information on evaluation of patients with suspected androgen deficiency, clinical pharmacology of currently available testosterone preparations, and monitoring of patients who are receiving testosterone supplements. Key recommendations are as follows:
The following is abstracted from an article by Guay in 2004 (Patient Care):

Signs of hypogonadism include anemia, diminished muscle mass, low bone density, thinning of body hair, testicular atrophy and increased central body fat. Symptoms include diminisehd libido, erectile dysfunction, reduced energy, diminished strength, cognitive impairment, and depressed mood. A total testosterone level < 200 ng/dl is considered low. A level < 300 with signs or symptoms of hypogonadism confers eligibility for a 3 month trial of testosterone replacement. Between 300-400 is a gray area; in patients with significant signs/symptoms of hypogonadism, a therapeutic trial may still be reasonable. Sex hormone binding globulin increases with age, and may elevate the total testosterone level even in the face of reduced bioavailable testosterone. Free testosterone levels vary greatly from lab to lab; total testosterone levels also exhibit significant lab variability. Preparations: Transdermal gels are convenient, yet costly ($200 per month). Transdermal patches are also available, but cause dematitis in about 40% of users. IM testosterone has been available for years, however serum levels vary greatly. The usual dose is 200 mg q 2 weeks or 100 mg im q week (more frequent dosing produces less of the roller coaster variability). Monitoring treatment: PSA, hematocrit, and digital rectal exam should be done at the outset. The goal of therapy is a testosterone level between 500 and 1000 ng/ml. PSA, hematocrit and DRE should be done at 6 months and a year, and then annually thereafter. PSA is likely to rise a little bit with testosterone therapy; however, a rise of 0.5 ng/ml after initiation of TT warrants a biopsy. The introduction of exogenous testosterone will suppress pituitary stimulation of the testes, and it may take 4-5 months after therapy is withdrawn before pituitary stimulation returns to normal - making testosterone supplementation an unnecessarily indefinite treatment in patients who may have had a transient fall in testosterone levels due to some physiologic stress. Risks of therapy: The biggest theoretical risk is that of prostate cancer. However, a review of 12 prospective studies examining the link between elevated levels of serum androgens and prostate cancer failed to show any clear correlation. Prostate volume increases about 15% in the first 6 months of treatment with testosterone; however, no significant change was found in PVR and flow rates. Polycythemia is a risk of TRT, especially in smokers. Infertility is a universal side effect of TRT. TRT in women: In light of possible risks for heart disease and stroke, a panel advised the FDA that more research was needed before approval of Intrinsa, a transdermal testosterone patch for women with low libido. Although not approved for use in women, testosterone has been used for decades. Studies have been few, and results have been uncompelling. Measuring testosterone levels in women is difficult, as the levels are so low to begin with (5-7 ng/dl). A free androgen index (ratio of total testosterone to SHBG) has been proposed, but this becomes unreliable if SHBG levels are low. Bottomline: need more research.