randomization was described and appropriate. Another point was
given for the description of dropouts and withdrawals, for a total
of 5 points. The inclusion and exclusion of studies, data extraction,
and validation, and the assessment of methodologic quality were
performed independently by 2 reviewers (GH, EE). Disagreements
in the evaluation of individual trials were largely attributable to
reading errors, and were resolved through discussion.