Mathematical models for assessment of long‐term persistence of antibodies after vaccination with two inactivated hepatitis A vaccines

Very few studies with inactivated hepatitis A vaccines were designed for long‐term follow‐up of antibody persistence. Based on the serological data from these vaccine trials, mathematical models were developed to predict the decrease of anti‐hepatitis A virus (anti‐HAV) antibodies after vaccination. This study was designed to compare Avaxim (0–6 months) to Havrix 720 (0–1–6 months). In this paper, both groups of vaccinees are described considering the age, gender, and weight of the subjects at enrollment. For mathematical modelling, two different approaches were used: one starting the calculations from the geometric mean titres (GMTs) at each point in time, the other basing the calculations on individual anti‐HAV titres. Both vaccines are very immunogenic, although Avaxim shows a higher GMT at each point in time. When these data are used in mathematical models to predict the persistence of anti‐HAV antibodies, both vaccines (Avaxim and Havrix 720) show similar long‐term antibody kinetics. Antibody levels ≥ 20 mIU/ml are estimated to last on average for at least 10 years after completion of the full vaccination course. Ten years after the full course, approximately 53% of subjects are estimated to have antibody levels ≥ 20 mIU/ml. At 15 years, these levels will be maintained by about 34% of vaccinees. Avaxim and Havrix 720 show a similar long‐term profile of persistence of anti‐HAV. A mathematical model based on GMTs appeared to give equivalent results to a model based on individual serological data. The GMT method is easier to apply than the individual based method. However, the advantage of the latter method is the possibility of calculating confidence limits for the predicted values and making estimates of the percentage of subjects having a certain level of antibody titres at a certain time. J. Med. Virol. 60:1–7, 2000. © 2000 Wiley‐Liss, Inc.