Bariatric Outcomes Longitudinal Database (BOLD) suggests excess weight loss and excess BMI loss to be inappropriate outcome measures, demonstrating better alternatives.

Excess weight loss (%EWL) results show significant variation by initial body mass index (BMI): the heavier the patient, the lower the result. It is unclear whether this effect originates from the arithmetic construction of this outcome measure or from any true difference in bariatric impact on heavier and lighter patients. Outcome measures generating inappropriate variation would actually be unsuited for bariatric reports with possible implications on existing bariatric evidence. Nadir weight loss results after laparoscopic gastric bypass in 7,212 women from the Bariatric Outcomes Longitudinal Database (BOLD) are calculated for %EWL and 26 different relative measures using the formula 100% × (initial BMI - nadir BMI)/(initial BMI - a), with reference point 0 ≤ a ≤ 25. Variations by initial BMI and deviations in results using each relative measure are compared. Mean initial BMI, 47.5 kg/m(2). Mean nadir BMI, 28.9 kg/m(2). Mean nadir results of the lighter (A) and heavier (B) halves (2× n=3, 606) by reference point, a=25 (excess BMI loss (%EBMIL)), 95.3% (A)-77.2% (B); a=ideal weight (%EWL), 81.7% (A)-70.4% (B); a=10, 49.6% (A)-49.7% (B); a=0 (total weight loss (%TWL)), 37.7% (A)-40.2% (B). Variation by initial BMI is significant using all relative measures, except those with a = 9 (p = 0.396) and a = 10 (p = 0.504). The smallest variation coefficient is 21.5% at 8 ≤ a ≤ 14. Gastric bypass works equally effective for all female patients, reducing the part of body mass above 10 kg/m(2) by an average of 49.6%, regardless of their initial BMI. In contrast, %EWL and %EBMIL generate variation by initial BMI, providing lighter patients inappropriately with better results, potentially interfering with the significance of bariatric evidence. These two most widespread used outcome measures in bariatric surgery are therefore actually unsuited for comparing patients or groups. This BOLD data makes a strong argument for abandoning %EWL and %EBMIL altogether and reporting bariatric relative outcome as %TWL only.