In October 1990, a hospital (DMCH) in eastern Saudi Arabia reported 17 infants who had developed sterile thigh abscesses where they had received their first dose of recombinant hepatitis B vaccine (HBV). The hospital raised concerns about HBV safety. From operating room log books we identified all pediatric incision and drainage procedures in seven local hospitals including DCMH for the preceding year. We defined a case as a thigh abscess which yielded no organisms in routine culture from a child below 14 months old. We interviewed parents and reviewed medical and vaccination records. For each case we chose five control infants at random from the same month in DMCH delivery room log book. We compared cases with controls for HBV lot, attending staff, and work load. We interviewed and tested nurses on vaccination methods. We identified 27 cases of thigh abscess from October 1989 to October 1990. Birthplace was available for 24 affected infants, all of whom were born in DMCH and given HBV, BCG and vitamin K in the delivery room (3 cases/1000 HBV). All abscesses were intramuscular, with a median 7 weeks incubation and a median of 3 weeks for healing. All affected infants had BCG scars. The case-control comparison gave odds ratios (OR) of 1.04, 2.09, 0.97, and 0.75 for different lots of HBV and 95% confidence interval for all OR included 1.0. Abscesses were associated with vaccination by one delivery room nurse (OR=122, 95% CI 24-751). However, this nurse's knowledge and vaccination techniques were above the median of her co-workers. HBV and BCG were given in the delivery room with unlabelled 1 ml syringes. BCG was prepared 10 syringes at a time. The nurse associated with the abscesses spoke a different language than the nurses who often filled the syringes. After we transferred newborn vaccination responsibilities to specifically trained nurses in the nursery, no new cases were seen to occur at DMCH. Rather than a defect in HBV, we believe that the absence of association with any HBV lot, the clinical presentation, the localization to one nurse in one hospital, the possibility of confusing syringes, and the disappearance of the problem after correction of vaccination practices all point to inadvertent administration of BCG into the thigh as the cause of these abscesses.
Conclusion: