When Diets Fail...

Abstract & Commentary

By Barbara A. Phillips, MD, MSPH, Professor of Medicine, University of Kentucky; Director, Sleep Disorders Center, Samaritan Hospital, Lexington. Dr. Phillips reports no financial relationship to this field of study.

Synopsis: Compared with morbidly obese individuals who did not have surgery, those who underwent gastric bypass surgery had reduced total mortality after about 7 years, especially for deaths from diabetes, heart disease, and cancer. However, the rate of death from accidents and suicides was higher in the surgery group than in the control group.

Source: Adams TD, et al. NEJM. 2007;357:753-761.

This study is a report of a retrospective comparison of 7925 patients who underwent Roux-en-Y gastric bypass surgery between 1984 and 2002 with an equal number of markedly obese adults who were randomly selected from among applicants for a driver's license or identification card. The mean age of the subjects was 39.5 years, and 84% were women. The gastric bypass operations were all performed by a single Utah surgical practice of six experienced bariatric surgeons. The surgical treatment group and the control group were matched by sex, Body Mass Index (BMI), and age and year of surgery. The primary outcome was death from any cause. Secondary outcomes were death from various specific causes. The authors compared the risk of death in the two study groups with the use of Cox proportional-hazards regression analysis. Baseline was defined as the date of gastric bypass surgery or of license application. Survival time was computed as the difference between the baseline date and the date of death for decedents or January 1, 2003, for survivors. The groups were compared over about a 7 year period.

The rate of death from any cause was 40% lower in the surgery group than in the control group (P < 0.001), even after adjusting for covariates. In addition, death rates were lower for all diseases combined (52%, P < 0.001), diabetes (92%, P = 0.005), coronary artery disease (59%, P = 0.006), and cancer (60%, P = 0.001) in the surgical group. However, deaths not caused by disease (including suicide, accidents not related to drugs, poisonings of undetermined intent, and other deaths) were 1.58 times as great in the surgery group as in the control group (P = 0.04). After a mean follow-up of 7.1 years, 171 deaths from disease were prevented per 10,000 operations in the group that underwent surgery. After adjusting increased risk of non-disease-related deaths for those who underwent gastric bypass surgery, the authors calculated a net prevention of 136 deaths per 10, 000 surgeries. Those who had a BMI of at least 45 Kg/m2 had a statistically greater reduction in death risk after surgery than those whose BMI was less than 45 Kg/m2. Women had a slightly higher benefit than men in terms of reduced disease-related mortality following surgery.


Almost everyone who is currently practicing medicine is painfully aware of the global explosion of obesity and obesity-related illness.1,2 What can be done about it is not clear at all. Banning transfat, requiring physical education in schools, eliminating soft drink machines on campuses, requiring clearer nutritional information on food labels and other public health measures have had little impact. Those of us who address the obesity problem on a patient-by-patient basis are intensely aware of how few patients actually lose weight on their own. The issue is complicated by the plethora of dietary approaches our patients encounter as they desperately scan the internet and the lay press. So, when information comes along about an approach to weight loss that not only works but also saves lives, there is a great deal of interest and excitement. For example, the Wall Street Journal headlines declaim, "Lower weight from surgery shows benefit.3" This paper by Adams et al was prominently reviewed in the lay press, and our patients are asking about it.

This carefully done, long-term study demonstrates reduction in deaths after gastric bypass surgery in a large number of patients who were compared with slightly heavier controls. There is a companion paper in the same issue of the New England Journal which notes a similar reduction in mortality following bariatric surgery for obesity in the Swedish Obesity Study (SOS)4; however, only a minority of patients in the SOS study underwent gastric bypass, which was the surgical procedure used in the current study.

In the accompanying editorial5, Dr George Bray notes several caveats: "During the period in which these two studies were carried out, laparoscopic techniques largely replaced open operative techniques, allowing for a more rapid postoperative recovery, less surgical stress, and reduced mortality. Thus, future death rates associated with bariatric surgery should be lower than those reported by either Sjöström et al. or Adams et al. The lowest surgical mortality is seen among those patients whose surgeons have performed more than 50 operations and particularly more than 100 operations. This would suggest that the centers doing these procedures should provide optimal training and maintain health care facilities specifically for overweight patients."

There are other caveats to keep in mind. First, this wasn't a randomized, controlled trial, and, as the authors point out, "it is possible that severely obese persons who did not seek gastric bypass surgery were less likely to make healthy lifestyle choices, resulting in a rate of death higher than that for persons who underwent surgery."

There was also a troubling association with increased deaths due to accidents and suicides in the surgical group. The main categories of interest here are "accidents unrelated to drugs," "poisoning of undetermined intent," and "suicides." For the groups analyzed in this paper, the actual numbers of deaths in each of these 3 categories for the surgical vs nonsurgical groups respectively are 21 vs 17 for accidents, 9 vs 4 for poisoning, and 15 vs 5 for suicide. In their discussion, the authors note, "Reports reveal that a substantial number of severely obese persons have unrecognized presurgical mood disorders or post-traumatic stress disorder or have been victims of childhood sexual abuse. … Some centers for bariatric surgery recommend that all patients undergo psychological evaluation and, if necessary, treatment before surgery and psychologically related surveillance postoperatively." This is a useful observation, but doesn't really address why severely obese people who have surgery are more likely to have accidents or suicide.

In his editorial,5 Dr. Bray suggests it is time to re-evaluate the 16 year-old NIH guidelines about bariatric surgery for obesity.6 These guidelines emerged from a consensus conference recommending reserving bariatric surgery for people with BMI's over 40 kg/m2 or more than 35 kg/m2 with coexisting illnesses. The consensus guidelines also concluded that bariatric surgery is appropriate only if other forms of weight loss treatment fail. I have seen people die of obesity, and I agree that sometimes aggressive intervention is needed. The truth is, "diets" fail more often than not, and weight loss effected by medications is generally minimal and difficult to sustain. But my discussions with patients about surgical treatment of obesity will include the information that we still don't have the last word (only an RCT will establish that) and that there is an unexplained increase in accidents and suicides in those who choose to undergo weight loss surgery.


1. National Institutes of Health. Obes Res. 1998;6:Suppl. 2:51S-209S.

2. Obesity: preventing and managing the global epidemic: report of a WHO consultation. World Health Organ Tech Rep Ser. 2000;894:1-253.

3. Bulkeley WM. Wall Street Journal. Thursday, Aug 23, page D6.

4. Sjöström L, et al. N Engl J Med. 2007;357:741-752.

5. Bray GA. NEJM. 2007; 357:818-820.

6. Consensus Development Conference Panel. Ann Intern Med. 1991;115:956-961