Background: Bariatric results expressed in the relative measure excess weight loss (%EWL) vary significantly by initial body mass index (BMI): the heavier the patient, the lower the %EWL. We examine if this variation is caused by using a wrong outcome measure and argue that no relative weight loss measure can express bariatric or metabolic goals unequivocally.
Methods: Nadir weight loss results after laparoscopic gastric bypass in 168 women with initial BMI ≥35 to <60 kg/m2 are calculated for %EWL and 61 different relative measures using the formula 100% × (initial BMI - nadir BMI) / (initial BMI-a), with a ranging from -30 to +30. Standard deviations are compared mutually and with those reported in the literature. For each relative measure, the significance of any variation by initial BMI is determined with the Mann-Whitney U test.
Results: Mean initial BMI was 44.9 ± 6.7 (35.0-59.7) kg/m2. Mean nadir BMI was 28.8 ± 5.8 (18.5-44.4) kg/m2. Mean nadir excess BMI loss (%EBL; a = 25) was 87.0 ± 28.0 (19.4-155.1)%. Mean nadir (total) weight loss (%TWL; a = 0) was 35.9 ± 8.5 (9.5-57.1)%. Mean nadir %EWL was 77.3 ± 22.8 (17.7-135.2)%. The smallest variation coefficient was 23.7% at a ranging from -1 to +3, including %TWL (a = 0). This is lower than variation coefficients of %EWL results in our series and in the literature. Variation by initial BMI is significant using relative measures with a ≥3, including %EBL and %EWL (both p < 0.0001) and not significant with a <3, including %TWL (p = 0.13).
Conclusions: In contrast to their widespread use, %EBL and %EWL are not suited for comparing different patients or nonrandomized groups. They cause variation by initial BMI, which disappears using %TWL. In general, absolute terms should be preferred for bariatric outcome and goals. The power of bariatric procedures is best represented by their mean %TWL value.