Five-year outcome with gastric bypass: Roux limb length makes a difference.

BACKGROUND: Surgical reports have indicated that longer Roux limbs (150 cm) have greater or no effect on long-term weight loss in super-obese patients (body mass index [BMI]>or=50 kg/m2) and little effect in less obese patients.
METHODS: The weight loss outcomes through 5 years were compared in 3 sequential groups of patients, who underwent gastric bypass by 1 surgeon, and in whom the Roux limb lengths were different. Comparisons were made between 2 cohorts: those with a BMI of <50 (morbid obesity [MO]) and those with a BMI>or=50 kg/m2 (super obesity [SO]). Three groups of patients stratified by Roux limb lengths were compared: group 1, 41-61-cm Roux limb; group 2, 130-160-cm Roux limb; and group 3, 115-250-cm Roux limb (one third of small bowel). All comparisons were made using 2-way analysis of variance, and the interaction terms were not significant.
RESULTS: A comparable number of patients were in each group, and the average preoperative weights were similar; however, more than twice as many patients in groups 2 and 3 were SO than MO. The BMI loss and weight loss were similar in each group. The greater BMI cohort (SO) lost more weight than did the MO cohort (P<.001). The BMI change and weight change in the shorter Roux limb group were less than those in groups 2 or 3 (longer Roux limbs; P<.01-.05). This difference was established with the BMI by 18 months. The BMI change and weight loss were not different between groups 2 and 3, presumably because their mean Roux limb lengths were not different. A limited amount of weight gain or recidivism occurred in patients with 5 years of follow-up, and it was not different among the 3 groups.
CONCLUSION: The results of this study have shown that longer Roux limbs improve weight loss outcomes both early and late in SO patients but not in MO patients. Clinically used long lengths of Roux limbs are close enough to one third of the total small bowel length such that the weight outcomes were not different, and total length should not need to be measured operatively. The eventual changes attributed to recidivism were not affected by the Roux limb length.