Frequently Asked Questions

By 22-year follow-up, 62.8% of participants with anorexia nervosa and 68.2% of participants with bulimia nervosa were recovered. Earlier in follow-up, at about 9 years, recovery was 31.4% for anorexia nervosa and 68.2% for bulimia nervosa, showing that recovery in bulimia nervosa occurred earlier, while recovery in anorexia nervosa continued to increase over the longer term.

Yes. In this cohort, 50.6% of participants with anorexia nervosa who had not recovered by the 9-year follow-up were recovered at 22 years. The authors concluded that recovery from anorexia nervosa continued over the long term rather than plateauing after the first decade.

Long-term recovery rates were similar by 22 years, but the timing differed. At 22 years, 62.8% of participants with anorexia nervosa and 68.2% of participants with bulimia nervosa were recovered, but by about 9 years recovery had reached 68.2% in bulimia nervosa versus 31.4% in anorexia nervosa. During the first wave, participants with anorexia nervosa recovered more slowly than those with bulimia nervosa (χ2₁ = 38.2; P < .001), and the median time to recovery for bulimia nervosa was 3.8 years.

Yes. For anorexia nervosa, recovery by the first follow-up period was strongly associated with recovery at 22 years, with an odds ratio of 10.5 (95% CI, 3.77-29.28; McNemar χ2₁ = 31.39; P < .01) when early recovery was defined as 52 consecutive weeks with a psychiatric status rating of 2 or less. The association remained present when early recovery was defined as 26 weeks (OR = 6.2; 95% CI, 2.60-14.61; P < .01) or 13 weeks (OR = 7.4; 95% CI, 2.90-18.83; P < .01).

Not when recovery was defined more stringently. In bulimia nervosa, early recovery was not associated with increased likelihood of recovery at 22 years when defined as 52 consecutive weeks (OR = 1.0; 95% CI, 0.49-2.05; P = 1.0) or 26 consecutive weeks (OR = 1.6; 95% CI, 0.77-3.46; P = .19). When early recovery was defined as 13 consecutive weeks, the association was modest (OR = 2.2; 95% CI, 1.01-4.88; P = .06).

Recovery was defined as an anorexia nervosa or bulimia nervosa psychiatric status rating of 2 or less for 52 consecutive weeks on the Longitudinal Interval Follow-Up Evaluation of Eating Disorders (LIFE-EAT-II). The study also examined shorter early recovery thresholds of 26 and 13 consecutive weeks when testing how early improvement related to long-term outcome.

Relapse remained clinically relevant over long-term follow-up. Among participants who had recovered by the first follow-up period, 10.5% of those with anorexia nervosa and 20.5% of those with bulimia nervosa were no longer recovered at 22 years.

Recovered participants had lower mean Eating Disorder Examination Questionnaire scores than nonrecovered participants in both anorexia nervosa and bulimia nervosa. In anorexia nervosa, recovery was associated with higher quality of life, while in bulimia nervosa only one quality-of-life domain was significantly higher in recovered than nonrecovered participants. The authors summarized this as normalized levels of eating disorder pathology in both groups, with broader quality-of-life improvement in anorexia nervosa.

Based on these data, the authors argued against palliative care for most patients. They found that recovery remained possible even after long illness duration, especially in anorexia nervosa, where recovery continued to accrue beyond the first decade; accordingly, they concluded that the findings support active treatment rather than palliative care for most patients with anorexia nervosa and bulimia nervosa.

This was a longitudinal follow-up study of women recruited from Boston-area outpatient eating disorder services between 1987 and 1991. The analytic sample included all surviving participants from the original cohort (N = 228), with 77% of survivors re-interviewed at 20 to 25 years and multiple imputation used to include missing Wave 2 data. Participants were interviewed every 6 to 12 months for a mean of 9.1 years during the first wave, then reassessed once at a mean follow-up of 22.1 years using the LIFE-EAT-II.

The main limitations are that the sample was predominantly white, Massachusetts-based, treatment-seeking, and first identified from 1987 through 1991, so the findings may not generalize to people who do not seek treatment or who received newer evidence-based interventions. The study also lacked specific treatment data, had potential bias because noncompleters may have had more severe illness despite multiple imputation, relied partly on retrospective recall in clinical interviews, and did not capture symptom course between the 9-year and 22-year follow-ups.