Cataceous, the study is excellent in terms of terms of measurements.

However, the authors conclusion misrepresents the change in bone markers, which can be found in Table 4:

It is clear that bone formation markers carboxyterminal collagen and osteocalcin decreased in ALL enclomiphene groups and osteocalcin increased ONLY in the transdermal testosterone group.

The authors conclude that these changes are "not significant" because their treatment groups are rather small (~12 people each), and with such small samples it is not possible to discern such changes in the markers from random sampling fluctuations. However, if you take into account the fact that you see a consistent DECREASE of the two markers in all enclomiphene groups, the probability for random fluctuations to all align in the same direction 6 times is rather small (it's like flipping a coin 6 times and getting 6 heads).

The correct conclusion should have been "although the mean bone formation markers were reduced in all enclomiphene groups, these changes did not reach statistical significance due to the small sample sizes".

Medical people always make that error. "No (statistically) significant difference/change/effect" does not mean that there is no change. It means that the study could not prove a change, either because there really isn't one or because the sample size was too small and the study does not have the statistical power to claim the observed change is not random, which is what I think is happening in this study.