Ten-year outcome of anorexia nervosa with teenage onset

Abstract

Objectives: To study prospectively the relatively long-term outcome in a representative sample of teenage onset anorexia nervosa (AN) in respect of overall outcome, physical health, psychiatric disorders, and personality disorders and to investigate familial factors among first-degree relatives. Method: A group of 51 cases (48 female, 3 male) with AN, recruited by community screening (including the total AN population (minus one individual) from one birth cohort), with a mean age of onset of 14 years, was contrasted with an age-, sex-, and school-matched group of 51 comparison cases on various measures of outcome at a mean age of 24 years (10 years after reported AN onset). All except 3 telephone interviewees in the AN group were seen personally. All 102 cases had originally been examined at age 16 years and followed up at 21 years. At 24 years all probands were blindly assessed by the author using the Structured Clinical Interview for DSM-III-R for personality disorders (SCID-II), the Yale-Brown Obsessive Compulsive Scale (Y-BOCS), and the Asperger Syndrome Diagnostic Interview (ASDI), and were examined neurologically/ neurodevelopmentally. Another psychiatrist performed interviews in respect of psychiatric disorders (SCID-I) and overall outcome (Morgan-Russell assessment schedule, Global Assessment of Functioning scale, GAF). Concerning familial factors, the author, who was blind to diagnostic group, prepared case notes detailing those factors. Another psychiatrist, also blind to diagnostic group status, assigned psychiatric diagnoses on the basis of information obtained from the case notes. Results: The mortality rate was zero in both groups at 10-year follow-up. One in four in the AN group had a clinically significant eating disorder (ED). Three females still had AN. Three quarters of the AN group admitted to past or present binge-eating and/or purging. Lifetime diagnoses of affective disorders and obsessive-compulsive disorder were overrepresented in the AN group. Cluster C personality disorders (PDs), obsessive-compulsive personality disorder (OCPD) in particular, were the most common PDs at 10-year follow-up. Autism spectrum disorders (ASD) were significantly more common in the AN group, affecting 18%. Outcome according to the Morgan-Russell outcome categories was considered poor in 27%, intermediate in 29% and good in 43%. According to the GAF, half the AN group had severe psychiatric symptoms and/or a poor psychosocial functioning. Those were cases with either current ED or childhood onset ASD/OCPD. Physical disorders/complaints were significantly overrepresented among the AN cases. Dysdiadochokinesis occurred almost exclusively among AN subjects. Social impairment in first degree relatives and deaths in the close family were overrepresented in the AN group. Conclusions: The ten-year outcome of adolescent onset AN was favourable in about half of the AN cases. Thirty percent of the AN group were diagnosed with a PD, most frequently cluster C, and, as in most clinical outcome studies half the AN group reported poor psychosocial functioning. The consistency over time of OCPD and/or ASD in a subgroup of AN cases was in accordance with our hypothesis of persisting personality characteristics of restraint and rigidity associated with a poor overall outcome.