ARTICLE IN PRESS

assignment was blinded to patients and all medical person-nel (except the operating team) during follow-up, there-fore reducing confounding factors in health care consump-tion. Second, we did not use decision-analytic method-ologies which require simplified assumptions and may not reflect clinical practice. Third, the duration of follow-up was one-year, which is also a limitation. Although most studies on AS have a six or 12 months follow-up period, it seems reasonable to suppose that longer follow-up would more accurately define differences in health care consump-tion. However, HrQoL outcomes may become con-founded by additional co-morbidity as follow-up extends and therefore affect the QALY calculation and coinciding ICER. During the trial, there was no statistical difference in co-morbidity between both treatment groups w2x. Fur