Activity of hepatitis B surface antigen and antibodies to hepatitis B e antigen

Sir We have read with great interest the paper by F ter Borg and coworkers (June 27, p 1914)1 on the relation between laboratory test results and histological inflammation activity in individuals positive for hepatitis B surface antigen (HBsAg) and antibody to hepatitis B e antigen (anti-HBe). Given the complexity of chronic hepatitis B, modelling techniques that identify multidimensional relations in clinical data are more suitable than logistic regression and other forms of analysis in which these relations are not apparent.2 A statistical data analysis network—the SMILES (similarity of least squares) model—has been developed for chronic hepatitis B patients to predict likelihood of response to interferon therapy.3 The models predict severity of inflammation and fibrosis, according to the criteria of V J Desmet et al.4 The models were developed from pretreatment data of 36 anti-HBe-positive5 and 82 HBe-positive3 patients with chronic hepatitis B, and included measurements of serum hepatitis B virus DNA, and aspartate and alanine aminotransferase, along with patients' age and sex. Most of our patients were selected for interferon treatment (persistent increased alanine amintransferase), whereas most of their patients had alanine aminotransferase values within normal limits.