Methodologic pitfalls in assessing the efficacy of recombinant follicle-stimulating hormone versus human menopausal gonadotropin in assisted reproduction.