Surgery for Obesity - Ernsberger

back to


by Paul Ernsberger, Ph.D.


"Well, the gold standard in medicine is the controlled clinical trial. We don't go subjecting 100,000 people to a surgical procedure without doing a controlled clinical trial or dozens of clinical trials, and then looking at the results. Do you know how many clinical trials have been published on weight-loss surgery or gastric bypass? Zero. None of them have compared it to clinical conservative treatment and found it to be superior for life expectancy or for anything else other than, you know, risk factors. A number of trials have been started, and the final results have never been reported. We have to ask, you know, why haven't we seen the final results? I think it's because it's bad news. "

(Paul Ernsberger on Donahue, 2002)

from Obesity & Health (renamed Healthy Weight Journal)
March-April 1991, pp. 24-25

As the eve of the 1991 NIH consensus conference on weight-loss surgery approaches, let us look back at the last NIH conference on this topic in December 1978. The panel gave its approval to intestinal bypass surgery, even though this operation was already coming under criticism for the long-term side-effects that it caused. As a result of recommendations by the consensus panel, intestinal bypass was accepted for health insurance coverage, which made possible tens of thousands of these operations. The legacy of the NIH panel's endorsement of intestinal bypass surgery is perhaps a hundred thousand patients worldwide, the majority of whom have suffered severe complications. Of the survivors, most now have had the operation undone. One can only hope that the legacy of this latest panel is more benign.

The panel members face a difficult task in evaluating the risks of the various surgical treatments for obesity. Nearly very surgical operation originates in a laboratory, where it is refined by extensive tests in animals. For example, coronary bypass surgery was the product of years of experimentation on dogs in which repeated measurements and detailed autopsies revealed potential complications and allowed surgeons to perfect the operation. In contrast, only three animal tests have ever been reported or gastric bypass or gastroplasty. In one test of gastric bypass in rats, many abnormalities were found at autopsy, including damage to the stomach, liver and pancreas from fibrosis (American Journal of Clinical Nutrition 40:293-302, 1984). We don't know whether these progressive abnormalities happen in human patients, because autopsy results have never been reported. Because of a lack of animal testing, the panelists will lack information on the biological effects of these operations.

Another difficulty the panelists will face is the bewildering variety of operations. As many as two dozen basic types of both gastroplasty and gastric bypass are in current use, along with several modifications of the original intestinal bypass. If we then consider variations on the basic methods, it can almost be said that no two surgeons do exactly the same procedure. The different types of operations differ in their safety and effectiveness. Operations such as gastric bypass that impair the absorption and processing of nutrients produce the greatest weight loss, and are most likely to produce lasting weight loss. With gastroplasty, weight regain is more common. After the most effective operation of this type, the vertical banded gastroplasty, 76% of patients failed to maintain weight loss after 30 months of follow-up (Surgery 98:700-707, 1985). In another study, 69% had fair-to-poor results in maintaining weight loss after 23 months and 22% suffered from obstruction of the narrow outlet from the stomach pouch (Mayo Clinic Proceedings 61:287-291, 1986). Weight regain after gastroplasty is usually the result of gradual stretching and enlargement of the stomach pouch or the narrow outlet from it. Because the stomach is almost infinitely expandable and adapts to increased pressure by growing larger, the operation is doomed to be eventually undone by natural adaptive processes of the patients' body. Although gastric bypass is more effective in maintaining loss of weight, there are more long-term complications, particularly nutritional deficiencies including anemia, pernicious anemia, osteoporosis, and neurological damage. These same complications were described 30 years ago as the long-term result of stomach surgery for ulcers. Consequently, these operations were long ago abandoned for treatment of ulcer. 

The stomach is not simply passive sac for storing ingested food, but plays a complex role in the processing of nutrients. Surgical procedures which interfere with the normal operation of the stomach inevitably cause multiple problems. Unfortunately, no controlled trials have ever been run which include physical examination of the patients for possible side effects by independent doctors not associated with the surgeon. In a rare instance of independent examination of bariatric surgery patient, a team of neurologists examined 500 patients who had received either gastric bypass or gastroplasty and found neurological complications (nerve or brain damage) in 5% of them (Neurology 37:196-200, 1987). The patients were usually examined within a year after surgery, so the incidence of long-term neurological deterioration could be much higher than 5%. Possible damage to organs other than brain and nervous system has not been put under rigorous independent evaluation.

Once a surgical technique has been developed in the animal laboratory, normally the next step is to run controlled clinical trials comparing long-term outcomes for patients and an untreated control group. Coronary bypass was tested in this way and it was proven that heart patients undergoing surgery lived longer than comparable patients getting only non-surgical treatment. No controlled clinical trials have ever been run for weight-loss surgery, except for one Danish trial of gastric bypass that showed that patients undergoing surgery experienced more health problems than comparable patients who were put on very-low-calorie diets (New England Journal of Medicine 310:352-356, 1984; Danish Medical Bulletin 37:359-370, 1990). Gastric bypass or gastroplasty does result in improved levels of blood pressure. cholesterol, and blood sugar. These reduced risk factors might translate into long-term disease prevention over the patients' lifespan if (and only if) their reduced weight can be maintained for life.

Do the benefits of weight loss exceed the risks of major surgery and the side-effects of tampering with the digestive system? In considering this question, the consensus panel must first ask whether weight loss is permanent. If a patient loses 100 pounds only to gain it back five years later, then her risk factors will return to dangerous levels. Because of the harmful effects of losing and regaining weight (the "yo-yo syndrome"), some risk factors, especially blood pressure, may be worse after regain of weight than they were at the start. When the risks of surgery and long-term complications are taken into account, then it become apparent that the net outcome for the patient who regains weight is highly negative. The typical patient receiving these operations is a woman in her 30's. To be certain of reducing her risk for heart disease later in life, weight loss must be assured for at least 10 years and preferably 20.

The consensus panel must also consider the benefit side of the risk-benefit equation. In order to evaluate the benefits of weight loss in extremely obese persons, let us start with the assumption that the poor health suffered by fat people can be completely reversed by weight loss. This may not be a valid assumption, especially because it is now dear that many fat people are burdened by a genetic defect that allows runaway weight gain. This same genetic defect may also lead to diabetes, high blood pressure and heart disease. Currently we do not know whether the health problems of the obese are directly caused by the fatty deposits themselves or whether they result from defective genetic machinery. If the latter is true, then weight loss will not completely erase the excess risk associated with obesity, because weight loss will not change a person's genetic makeup. Setting this argument aside, let us consider the actual risk faced by extremely obese persons. The median life expectancy from age 25 as a function of body mass index is shown in the illustration above (see original article). These data were taken from the world's largest epidemiological study, which tracked 1.8 million Norwegians for 10 years (Acta Medica Scandinavica Supplementum 679:1-56, 1984). Let us consider the lifespan of women, since 90-95% of all patients undergoing weight-loss surgery are female. Women who are neither underweight nor overweight have a life expectancy of about 79 years. Morbid obesity begins at a body mass index of 35. Women with a body mass index of 40 and above are shown at the far right. Their life expectancy is reduced by 5 years, which is equivalent in risk to light cigarette smoking. However, even these extremely obese women still have a longer life expectancy that normal-weight men. Several conclusions can be made from this graph: first, the typical "morbidly obese" woman in her 30's considering weight-loss surgery faces another four decades of life, which means that weight loss maintenance and surgical complications must be evaluated over the very long term. Second, given that the maximum benefit from weight-loss is a 5-year prolongation of life, the risks from surgery must be kept very low. Third, surgery should clearly be reserved for the most obese patients (body mass index over 40) and the ongoing trend for surgeons to make exceptions to the "100-pound rule" and operate on thinner patients must be deplored.

The challenge awaiting the NIH consensus panel is formidable. They will lack a base of knowledge of the biological effects and the medical consequences of these operations, especially over the long term. The vast array of variations of these operations complicates any evaluation. Operations that are more effective are also less safe; none of the surgical procedures seem to be both safe and effective. Risks and benefits must be compared over the remaining 40 years of life expectancy of the patient, but for many procedures patients are tracked for two years or less, and many times only weight loss is recorded without independent evaluations of the patients' overall health. Hopefully the 1991 panel will not repeat the mistakes of the 1978 panel, which failed to confront the epidemic of complications from intestinal bypass.

Paul Ernsberger, Ph.D.
Associate Professor of Medicine, Pharmacology and Neuroscience
Case Western Reserve School of Medicine
10900 Euclid Avenue
Cleveland, OH 44106-4982