An increase in life expectancy, contracted HIV infections in the elderly, comorbidity, and a discrepancy between real and biological age make BMD loss an unavoidable phenomenon in each patient. Both HIV and antiretrovirals amplify the age effect on BMD. Among N(n)RTIs, the effect oftenofovir on BMD seems to exceed the effect of other antiretrovirals, but we can find studies which confirm this observation, and other studies which doubt the conclusions and/or minimize the implications. This disagreement exists because there are a number of biases in all these studies that, unfortunately, compromise the conclusions. In the same analyses are men and women, different BMIs, different ages, different times of HIV infection before entering the trial, resulting in different biological ages, and especially different baseline BMD. All these biases together create a pool that is not homogeneous, having many variables, which complicates the correct interpretation of the results, especially in comparative studies between two pools. BMD in the studies is expressed as percentage loss during the follow-up, and not as the variation of absolute values of BMD: this feature may be deceptive regarding the real loss of BMD. Another bias is the approximate use, when defining BMD loss, of T-score even in non-menopausal women and in men younger than 50. Finally, another bias is the use of a different DXA machine to determine BMD. Prospective studies with a longer follow-up and without all these biases are urgently needed.