Many Anorexia Deaths Linked To Alcoholism, Suicide

Anorexia nervosa should be on all psychiatrists’ radar screens, suggests a new study reported in the February Archives of General Psychiatry.

The study has produced several important findings, one of the investigators, David Herzog, M.D., told Psychiatric News. Herzog is a professor of psychiatry at Harvard Medical School and head of the Harvard Eating Disorders Center at Massachusetts General Hospital.

“The mortality rate from anorexia nervosa is extraordinarily high,” Herzog said, and the cause of death in anorexia nervosa is not always starvation.

Certainly, this is not the first study to document high death rates in subjects with anorexia. Some studies reported during the 1990s did so, but what makes the study by Herzog and his colleagues especially compelling is that it is based on more recent cases than those reported in earlier studies and on cases taken from the community, not just from inpatient populations.

Between 1987 and 1991, Herzog and his coworkers recruited 136 patients with anorexia nervosa and 110 patients with bulimia nervosa to participate in their investigation. Most of these patients were seeking outpatient treatment for their eating disorder at the Massachusetts General Hospital Eating Disorders Unit or at other Boston eating-disorder programs. Most of the patients received some form of treatment for their disorder.

Herzog and his colleagues then followed the fates of these 246 subjects through 1998 via follow-up interviews, phone calls, medical records, death certificates, and autopsy reports. “This was a carefully done study without subjects being lost to follow-up,” Herzog explained to Psychiatric News. “So we can trust our data set pretty well.”

By the end of 1998, 11 subjects had died—10 who had been diagnosed with anorexia and one with bulimia. Of the 10 anorexia subjects who died, the direct cause of death in four instances was suicide.

Also, the researchers scrutinized the data that they had collected about the 10 anorexia subjects at the time they were enrolled in the study and during the nine years or so that they were followed up to see whether any of the subjects’ characteristics at the start of or during the study might have predicted their deaths. They identified two characteristics: duration of illness and alcohol abuse. Indeed, of the 10 anorexia subjects who died, four had had a history of alcoholism at the start of the study, two more acquired alcohol problems during the study, and alcohol appeared to play a direct role in the deaths of two subjects.

These results have a number of crucial implications for clinical psychiatrists, Herzog and his team believe. First, anorexia nervosa (but not bulimia nervosa) appears to have a high risk of death associated with it—even when patients get treatment for it.

“The mortality rate in the anorexia subjects was close to 12 times that which you would see in a similar age and gender population,” Herzog told Psychiatric News. “So individuals continue to die from anorexia nervosa despite more information about it being available than in previous years. In other words, it continues to be a very serious illness.”

Second, suicide appears to be a frequent cause of death in anorexia patients. “The suicide rate in our subjects was some 60 times higher than you would see in a similar age and gender population,” Herzog observed.

Third, “the longer a patient has been ill when a clinician encounters him or her, the more concerned the clinician should be about the risk of death,” Pamela Keel, Ph.D., an associate professor of psychology at Harvard University and the lead investigator in the study, told Psychiatric News.

The anorexia subjects in the study had the disorder anywhere from three months to more than 25 years. While “it would be possible to calculate the mean duration of illness among subjects who died, this number would be specific to these women and would not necessarily reflect a number that clinicians should refer to,” said Keel.

Fourth, the combination of alcohol and anorexia is “a particularly lethal combination,” said Herzog. And as he and his team advised in their study, “Physicians treating patients with anorexia nervosa should carefully assess patterns of alcohol use during the course of care because one-third of [subjects] who had alcoholism and died had no history of alcohol use disorder at intake.”

Finally, “not all the anorexia subjects who died were severely underweight,” Herzog stressed. “In fact, a number were at about normal weight or even above normal weight. So just because one achieves normal weight in these subjects is not a reason to discontinue monitoring them.”

The study was funded by grants from the National Institute of Mental Health, the Harvard Eating Disorders Center, Eli Lilly and Company, and the Rubenstein Foundation.

An abstract of the study, “Predictors of Mortality in Eating Disorders,” is posted on the Web at http://archpsyc.ama-assn.org/issues/v60n2/abs/yoa10275.html. ▪

Arch Gen Psychiatry 2003 60 179