Despite its efficacy, results of bariatric surgery are variable, with a substantial minority of patients

achieving significantly better than average weight loss and disease response rates, while a similar

minority experience suboptimal results. Anywhere from 10% to 30% of patients achieve <50% EWL,

which may be due to failure to lose weight initially, or weight regain after a period of adequate weight

loss. Mechanisms underlying such failures are poorly defined, and identification of accurate predictors

of response to surgery is an important area of research. While suboptimal results are often attributed to

lack of patient compliance with postoperative diet and exercise, these arguments ring false, echoing

similar arguments that attribute obesity itself to a lack of “willpower.” Instead, compelling evidence

implicates genetic and epigenetic mechanisms. Like obesity itself, responses to bariatric surgery are

highly hereditable. Ethnicity is a predictor of outcome, with Caucasians and Hispanics experiencing

greater weight loss than Blacks and Asians. Multiple single-nucleotide polymorphisms have been linked

to surgical outcome. Future research will identify clinical and genetic predictors of response and permit

optimization of allocation of surgical resources.

Perioperative Mortality and Morbidity

Morbidity and mortality associated with bariatric surgery have decreased over the past two decades as

surgeon experience increased and techniques evolved. Currently, gastric bypass and gastric band are

associated with mortality risks similar to hip replacement and laparoscopic cholecystectomy,

respectively. Perioperative mortality is highest for biliopancreatic diversion/duodenal switch, followed

by gastric bypass and sleeve gastrectomy, then gastric band (Table 46-3). Life-threatening perioperative

morbidity associated with gastric bypass and sleeve gastrectomy consists primarily of anastomotic

dehiscence, staple-line leaks, hemorrhage, thromboembolic events, and cardiac events. Life-threatening

perioperative morbidity associated with gastric band is rare and consists primarily of gastric

perforation, thromboembolic events, and cardiac events.

The most dire perioperative complication of gastric bypass is anastomotic dehiscence of the

gastrojejunostomy which usually occurs within 2 weeks of surgery. Incidences in most series range from

0.1% to 4%, although recent data suggest that anastomotic dehiscence rates are decreasing and highvolume centers report rates in the range of 0.1% to 0.2%. Signs and symptoms are similar to those of

anastomotic dehiscence after any intestinal operation, but may be less apparent in the obese. Persistent

tachycardia and abdominal pain are important but nonspecific sentinel signs. Upper GI radiographs may

be of diagnostic utility but lack sensitivity. Abdominal CT scan, even within a few days of surgery, may

be useful, as the lack of a fluid collection near the gastrojejunostomy has a high negative predictive

value for leak, while contrast extravasation has a high positive predictive value. Nonetheless, data are

conflicting regarding the sensitivity and specificity of radiologic studies for diagnosing anastomotic

dehiscence, and laparoscopic reexploration should be employed aggressively in suspected cases.

Operative management consists of washout, wide drainage, judicious attempt at repair, and distal

enteral access, usually with a gastrostomy tube in the remnant stomach, although a jejunostomy tube in

the distal Roux limb or biliary limb is an option if access to the remnant stomach is difficult.

Anastomotic dehiscence can be managed laparoscopically in some cases, reducing the high risk of

ventral hernia associated with laparotomy. Most will heal if sepsis is controlled.

Staple-line leaks are an equally dire event after sleeve gastrectomy, occurring anywhere from 1–2

weeks to many months after surgery. Leaks may occur at any point along the staple line, but when

involving the thicker tissue of the antrum may be less prone to heal, and interval conversion to gastric

bypass with resection of the involved portion of gastric sleeve may be necessary. Proper stapler

selection and use during primary operation with avoidance of overly narrow sleeves (current consensus

recommends 32- to 40-Fr diameter) reduce the risk of sleeve gastrectomy staple-line leaks, which

currently occur with an incidence of approximately 1% to 3%. Gastric band is only rarely associated

with perioperative septic complications, most often in the form of gastric perforation, which occurs in

0% to 0.5% of cases. Perforation, if not recognized, may be life-threatening and is a primary cause of

rare perioperative mortality and major morbidity after gastric band.

Table 46-2 Bariatric Surgery Long-Term Comorbidity Responses

1194

Table 46-3 Bariatric Surgery Perioperative Mortality and Morbidity

Perioperative hemorrhage after gastric bypass occurs with an incidence of 1% to 4%, often originates

from staple lines, and may be intraperitoneal or intraluminal. Intraluminal bleeding presents with

melena and is often self-limited, requiring only transfusion and observation. Infrequently, endoscopic or

operative intervention may be necessary. Hemorrhage after sleeve gastrectomy ranges from 1% to 4%,

is usually intraperitoneal from the gastric staple line, and may require operative intervention, although

observation in the hemodynamically stable patient is a reasonable initial course. Early small bowel

obstruction, and thromboembolic, pulmonary, and cardiac events comprise the majority of the

remainder of acute postoperative complications after bariatric surgery. Pathophysiology is not

substantially different than after nonbariatric GI operations, although the obese may be at higher risk,

reinforcing the importance of preoperative evaluation and preparation.

Late Morbidity

Late morbidity after bariatric surgery is similar to that associated with complex nonbariatric GI

operations, may occur months or years later, and often presents with abdominal pain. Gallbladder

disease and small bowel obstruction from adhesions and ventral/trocar hernias may occur after all

bariatric operations. Other complications are operation specific. While definitions vary, overall late

morbidity rates for gastric bypass range from 5% to 10%. Many late complications of gastric bypass

result from Roux-en-Y anatomy. Internal hernias occur in 0.5% to 9% of patients, and are categorized as

mesenteric (common to all Roux-en-Y reconstructions, in which bowel passes through the mesenteric

split in the Roux limb), mesocolic (specific to retrocolic reconstructions, in which bowel passes through

the mesocolic defect created for the Roux limb), and Petersen’s (more often associated with antecolic

than retrocolic reconstructions, in which bowel passes under the Roux limb and its mesentery) (Fig. 46-

7).26 Although closure of internal hernia defects at primary operation is straightforward and

recommended, long-term clinical follow-up data do not demonstrate that this practice reduces hernia

rates compared with series in which routine closure was not performed, possibly because of lack of scar

tissue formation at the sites of closure, and/or loss of mesenteric adipose tissue with weight loss leading

to late patency of previously closed defects. Patients often present years after surgery with a history of

severe episodic abdominal pain. Diagnosis may be challenging, as obstipation may be absent and

imaging normal even in the face of incarceration, especially if the biliopancreatic limb, which is out of

continuity with the alimentary stream, is involved. A low threshold for operative exploration must be

maintained in any patient with unrelenting abdominal pain after gastric bypass. Operative reduction and

closure of internal hernia defects can usually be performed via a laparoscopic approach.

1195

Figure 46-7. Internal hernias. Internal hernias may occur after any Roux-en-Y reconstruction, and include mesenteric, Petersen’s,

and mesocolic hernias. Weight loss after gastric bypass and loss of adipose tissue in the intestinal mesentery may predispose to

internal herniation.

Gastrojejunal anastomotic ulcer occurs in 0.5% to 25% of patients after gastric bypass, with the

majority of series reporting incidences between 2% and 8%. Tobacco or NSAID use increases risk, but

ulcers may occur in the absence of these risk factors. Diagnosis by upper endoscopy is usually

straightforward. Upper GI radiography is advisable to rule out an associated gastrogastric fistula,

especially if an ulcer is recalcitrant to medical treatment or is associated with weight regain. Most

ulcers heal with high-dose antacid therapy, and once healed, most surgeons recommend maintenance

antacid therapy for life, although data supporting this practice are sparse. Rarely, ulcers recalcitrant to

medical therapy require resection with anastomotic revision. Large pouches are a risk factor for ulcers

due to retention of parietal cells in the distal gastric pouch that do not participate in negative inhibition

of gastrin secretion, thus rationalizing pouch resection and anastomosis revision as a treatment strategy.

Gastrogastric fistula occurs with an incidence of 0% to 2% and may be associated with ulcers.

Gastrogastric fistula incidence has decreased substantially since the advent of divided gastric bypass.

Gastrogastric fistula often results from incomplete division of the gastric pouch from the gastric

remnant at primary operation. Stenosis of the gastrojejunostomy typically presents with intolerance of

solid foods and vomiting within 1 to 3 months of surgery with a wide range of reported incidences from

<1% to >20%, but most commonly 3% to 7%. Stenosis is usually easily treated with endoscopic

dilation, although rare cases may require anastomotic revision. Retrograde intussusception at the

jejunojejunostomy is a rare cause of abdominal pain after gastric bypass with a reported incidence of

0.1%. Pathogenesis is likely related to motility disorders resulting from disruption of enteroenteric

nerves and divorce of the small intestine from duodenal pacemaker plexi with subsequent development

of aberrant ectopic pacemakers in the involved intestinal limbs. Diagnosis may be challenging and

imaging unreliable. Treatment consists of either plication/pexy of involved limbs or resection and

revision of the jejunojejunostomy, with the latter approach associated with lower rates of recurrence.

The clinical presentation and differential diagnosis of abdominal pain after gastric bypass is diverse

and diagnostic imaging may be unreliable. The clinician is not infrequently faced with a gastric bypass

patient who presents with persistent pain in the absence of a clear cause after thorough diagnostic

1196

evaluation. Exploratory laparoscopy in such patients provides a diagnosis in up to 90% of cases. Even in

the absence of intraoperative findings other than patent internal hernia defects without incarceration,

closure of defects provides relief in the majority of patients. Less commonly, in the absence of a clear

cause, pain is presumed to result from atypical motility disorders associated with Roux-en-Y anatomy

and may be difficult to treat (Algorithm 46-1).

Late complications after gastric band are common, afflicting 10% to 50% of patients, may occur years

after primary operation, and include band slippage, band erosion, and development of GERD and

esophageal motility disorders. Treatment often requires band explant, with late explant rates of 30% to

50% in some series. While long-term data are evolving, late complications after sleeve gastrectomy

appear to be uncommon, with infrequent reports of sleeve stenosis (0.5% to 3%) and development of

GERD.

Bariatric operations alter absorption of macro- and micronutrients. Protein and lipid malabsorption,

steatorrhea, and protein wasting are rare after modern gastric bypass and virtually absent with gastric

band and sleeve gastrectomy, but afflict anywhere from 2% to 20% of patients after biliopancreatic

diversion/duodenal switch. The elimination of the pyloric emptying mechanism in gastric bypass is

associated with dumping and hypoglycemic syndromes, the underlying mechanisms of which are not

well understood. Early dumping occurs within 10 to 30 minutes of a meal, and is thought to result from

intraluminal fluid shifts and acute intravascular volume depletion secondary to rapid emptying of

hyperosmolar gastric contents into the small intestine. Early dumping is associated with tachycardia and

diaphoresis. Aberrations in gut hormone secretion, including elevated secretion of vasoactive intestinal

polypeptide, enteroglucagon, pancreatic polypeptide, and GLP-1, have been associated with a

predisposition to early dumping. Late dumping occurs hours after a meal and is thought to be due to

hypersecretion of insulin with subsequent reactive hypoglycemia; excess secretion of GLP-1 has also

been implicated. Dumping and hypoglycemia typically respond to a low-carbohydrate diet and

preprandial acarbose therapy. In cases recalcitrant to these measures, pharmacotherapy with agents that

reduce insulin secretion, including octreotide, diazoxide, and calcium channel blockers, may be of

benefit. Surgical banding of the gastric pouch to delay emptying has been employed with variable

efficacy.

Micronutrient deficiencies are underreported after bariatric surgery. The reported prevalences of

deficiencies of virtually all micronutrients after surgery span a broad range from 10% to >80% for

biliopancreatic diversion/duodenal switch and gastric bypass. Deficiencies are less common after sleeve

gastrectomy and gastric band. An understanding of the anatomic sites and mechanisms of micronutrient

absorption affected by gastric bypass foregut diversion guides diagnosis and management. Anemia,

neurologic, dermatologic, and GI symptoms are common clinical manifestations of many micronutrient

deficiencies (Table 46-4). Acute deficiencies may involve dramatic presentations (e.g., thiamine

deficiency and Wernicke encephalopathy). Chronic occult micronutrient deficiencies are associated with

increased risk of chronic disease. Chronic magnesium deficiency, for example, is associated with an

increased risk of type II diabetes. Chronic selenium deficiency has been associated with an increased

risk of cancer and cardiovascular disease. The most common deficiencies associated with gastric bypass

involve calcium, vitamin D, folate, iron, and B12. Vomiting due to stenosis or gastroenteritis increases

the risks of severe micronutrient deficiencies, and nutritional status should be monitored carefully in

any bariatric surgery patient with significant vomiting. It is important to note that obesity is associated

with multiple micronutrient deficiencies independent of bariatric surgery, which should be identified

and corrected in the preoperative period, and which may resurface in the postoperative period.

Patient Selection, Choice of Operation, Preoperative Preparation

Current criteria for candidacy for bariatric surgery are based on a 1991 NIH consensus conference, and

include BMI ≥40, or BMI ≥35 with at least one serious comorbidity of obesity. While widely used,

these criteria are based on expert opinion rather than definitive epidemiologic data, and were

established over two decades ago. Recent data suggest that bariatric surgery provides benefits to

diabetic patients with BMI 30 to 34.9.27 While consensus is evolving, the American Society for

Metabolic and Bariatric Surgery (ASMBS) had issued a statement in support of bariatric surgery in this

low BMI patient population, and the FDA, as the regulatory body that authorizes device use, approved

gastric band for this subgroup in 2011. As data accrue, candidacy criteria will expand. Good outcomes

have been reported in carefully selected patients with organ transplants, end-stage renal disease,

compensated cirrhosis, HIV, and adolescents. Active uncontrolled cardiac, pulmonary, or psychiatric

disease and uncompensated liver disease are generally considered absolute contraindications.

1197