Objective Percentiles for the body mass index (BMI) offer a possibility to epidemiologically assess the linear weight criterion of 85% average body weight commonly used for the diagnosis of anorexia nervosa. Method: BMI values corresponding to 85% average body weight were calculated and assessed with percentiles derived from epidemiological studies in both the United States and Germany. The underweight range was characterized epidemiologically. Results: The weight criterion used for the diagnosis of anorexia nervosa corresponds to BMI values between the 5th and 10th centiles in both populations. In epidemiological terms the lowest BMI values in individuals aged 10 years and older occur during adolescence. In the general population BMI values <16 kg/m2 are rarely observed. Upon the use of higher BMI cutoffs in the underweight range females clearly predominate. The BMI increase associated with the 5th or 10th centile in the age range between 18 and 30 years is quite low suggesting that many underweight females in the general population gain only minimal weight during this age span. Discussion: The diagnostic, epidemiological, and therapeutic implications for anorexia nervosa are discussed. © 1996 by John Wiley & Sons, Inc.

OBJECTIVE: Within the past decades prevalence rates for obesity among children and adolescents have increased in different populations. The hypothesis of this study is that the degree of adiposity in clinical study cohorts of extremely obese children and adolescents increased within the past decade. DESIGN: In six different study cohorts of the time period from 1985±1995 body mass indices (BMIs) of obese children and adolescents who were treated as inpatients at a specialized children's hospital were evaluated. For this purpose body heights, body weights, ages and sex of all inpatients of three referring agencies were retrospectively assessed biannually. RESULTS: In these six cohorts a signi®cant BMI-increase from 1985±1995 of 1.9 kg/m 2 (P`0.0001) for constant sex, age and referring agencies was found. Comparisons of the quartiles and the ninth decile in both sexes did not show any systematic increase at the ®rst quartile. In contrast, BMI-increases at the ninth decile were approximately 5 kg/m 2 for males and 2.5 kg/m 2 for females. CONCLUSION: Within the decade studied a signi®cant BMI-increase was detectable in this clinical population. This effect is especially discernible in the most extreme weight groups and in males.

No abstract

This study explored the hypothesis that body weight is reduced in male children and adolescents with schizoid personality disorder or Asperger's disorder. The body weights of 33 consecutively admitted male subjects with one of these disorders were retrospectively assessed with percentiles for the body mass index (BMI). The mean percentile (+/- SD) for the BMI was 31.6 +/- 27.6 and differed significantly from the expected value of 50 (P<0.001). Ten subjects had a BMI of < or = 10th age percentile. Post hoc comparisons revealed that BMI percentiles were (a) reduced to a similar extent in patients with schizoid personality disorder and Asperger's disorder and (b) reduced to a greater extent in patients with abnormal eating behaviour. During childhood and adolescence both diagnoses are associated with an increased risk of being underweight. Population-based BMI percentiles are useful for detecting associations between specific psychopathological syndromes and body weight.

No abstract