Stein et al's study (Reference Stein, Woolley and McPherson1999) provides essential reading for clinicians involved in treating anorexia nervosa or bulimia. However, in light of recent research concerning pregnancy and eating disorders (Reference Morgan, Lacey and SedgwickMorgan et al, 1999), three further areas of investigation seem to have been unexplored.

First, we have demonstrated that a third of women with bulimia developed postnatal depression, rising to two-thirds in women with previous anorexia nervosa. Second, Stein et al treat ‘eating disorders’ with unwarranted homogeneity; we found profound differences between pregnant women with bulimia and those with previous anorexia nervosa. The latter predicted postnatal depression, absence of breast-feeding and postpartum relapse into eating disorder. The most striking differences were apparent at a descriptive level, where women with previous anorexia nervosa appeared highly alienated from their infants. Third, the majority of pregnancies described in our study (Reference Morgan, Lacey and SedgwickMorgan et al, 1999) were unplanned due to mistaken beliefs regarding fertility.

It would be interesting if Stein et al were able to re-examine their data with reference to specific eating disorder diagnoses, presence of affective disorder and planning of pregnancy, all of which might generate maladaptive responses to the antecedents of conflict.