BACKGROUND Studies showing high rates of alcohol and drug misuse and deliberate self-harm in bulimia nervosa have led some authors to call for a distinct diagnostic subgroup, sometimes termed "multi-impulsive bulimia". However, these studies have been uncontrolled and of clinic samples and may be subject to sampling bias.
METHOD One hundred and two women with DSM-III-R bulimia nervosa were compared with 204 normal controls and 102 controls with other psychiatric disorders, all recruited from the same community sample. Interview measures were used for diagnosis and for the assessment of alcohol and drug misuse and deliberate self-harm.
RESULTS Bulimia nervosa cases did not differ from either of the control groups in terms of current alcohol consumption. Bulimia nervosa cases used more illicit drugs than either control group, but loss of control over drug use was very uncommon. Bulimia nervosa cases had a higher rate of deliberate self-harm than the controls. Only six (6%) bulimia nervosa cases had two or more of these behaviours concurrently.
CONCLUSION Sampling bias is present in clinic-based studies of comorbidity in bulimia nervosa. Those with comorbid substance misuse and deliberate self-harm are probably heterogeneous in character, and their classification as a subgroup would therefore be premature.