[Surgery for morbid obesity: 2. Complications. Results of a Technologic Evaluation by the ANAES]

The frequency of bariatric surgery has increased markedly in France in recent years, partly due to a better appreciation of the problem of morbid obesity but also due to the commercial introduction of adjustable gastric banding devices which can be placed by laparoscopic approach. Numerous complications of this surgery are known and require recognition to be appropriately treated. Studies of complications suffer from selection bias, methodologic flaws, and lack of follow-up. The incidence and type of complication are affected by the learning curve and surgical techniques. Postoperative mortality varies from 0.14% for laparoscopic gastric banding (LGB), to 0.31% for vertical banded gastroplasty (VBGP) and 0.35% for Roux-en-Y gastric bypass (GBP); pulmonary embolus accounts for 60-70% of deaths in all groups combined. Early post-operative complications vary with specific procedures. Abdominal wall complications, already frequent in an obese population, are decreased from 10% for open procedures to 6% for laparoscopic gastric banding. Both VBGP and GBP are now being done laparoscopically with increasing frequency. Complications specific to LGB include gastric perforation (0.3%), or port problems (5%). Complications with VBGP and GBP include fistula (1-3%), deep abscess, and pulmonary embolus (2%). Global early morbidity is 4.2% for LGB, and varies from 6.4%-22% for VBGP and 6.2%-11.3% for GBP depending on laparoscopic versus open approach. Late mechanical complications are also specific to type of surgery. Pouch dilatation is the most common late complication of LGB (6.3%) and seems related both to operative experience and to site of placement of the band; it has decreased with higher positioning of the band to leave a minimal gastric pouch and with dissection through the pars flaccida of the lesser omentum instead of directly along the muscular wall of the stomach. It usually requires reintervention. Erosion of the gastric band into the stomach (1.6%) is often asymptomatic and is suggested by late weight gain. With VBGP, disruption of a gastric staple line occurs in 12.1% and stenosis of the outlet with proximal dilatation in 6.5%; erosion of the calibrating band of Marlex or silastic occurs in 2.7%. With GBP, the disruption of a staple line across an intact stomach (23%) has become less of a problem with division of the gastric pouch from the distal stomach (2%). Stenosis of the gastrojejunostomy (3.7%) and marginal ulcer (3.5%) are not uncommon. The incidence of wound hernia, obstructive adhesions, and late cholecystectomy vary with the length and thoroughness of follow-up. Late functional complications such as vomiting, dysphagia, heartburn and esophagitis vary with the quality and length of follow-up study. GBP may cause diarrhea and dumping syndrome. Nutritional complications are more common with GPB than with purely restrictive procedures; iron, folate, and Vitamin B12 deficiency are the rule with GBP and require routine replacement therapy; iron deficiency has been noted even with LGB. ate death seems more related to co-morbidities than to the intervention itself. Thorough long-term follow-up study of complications is indispensable for assessment of outcomes and improvement of laparoscopic techniques. Even the less traumatic surgical approach of laparoscopic band placement should not be considered free of risk; strict adherence to pre-operative surgical indications should be maintained.