Sterile thigh abscesses at vaccination site

Within nine months of adding HBV vaccine to the EPI, the Ministry of Health received a report of 17 cases of unusual sterile thigh abscesses at the site of HBV vaccination. No organism had been isolated. Officials were concerned about the safety of the recombinant HBV vaccine.

All the hospitals in the health region provided information about the number of births and HE V vaccination of newborns. A review of admission and operation logbooks in all hospitals revealed 27 sterile thigh abscesses with no growth on culture.

The abscesses were situated on the anterior or anterolateral side of the thigh. They were not tender, not red and intramuscular, containing yellow liquefied pus. The mean age of the infants at the time of incision and drainage (I&D) was 4.5 months. Median healing time after I&D was 4.5 weeks. No organism was grown using the four standard growth media. However, acid-fast bacilli were seen in one of two examined under Zeil Nelson staining.

Of these 27 cases, 24 were born in one hospital (3 per 1000 live births), compared with none (0/1000 live births) from the seven other hospitals (P<0.01). The birthplaces of two were unknown and one was born outside the region.

Within the involved hospital, all 24 abscesses were in newborns vaccinated in the delivery room (3 per 1000), compared with none vaccinated in the nursery. In the delivery room, BCG syringes were usually loaded prior to delivery and kept in the refrigerator. HBV vaccine and Vitamin K were loaded into syringes as needed. Syringes for all three newborn injections were the same size, but usually with different needle lengths. The syringes were not labeled.

A case-control study showed that none of the four lots of HBV used was associated with the thigh abscesses (odds ratios [OR] = 1.0, 2.1, 1.0, 0.8; P = NS for each). These lots had also been used on hospital staff and on infants vaccinated in the nursery.

Newborns who had been vaccinated by one nurse had a very high risk of developing an abscess (OR = 124; 95% confidence interval = 24, 751). In one month, 33% of the infants she vaccinated later developed abscesses. Because of language differences, communication between this nurse and other nurses was difficult.

As a control measure, the hospital transferred all newborn vaccinations to the nursery and enforced labeling of all syringes used for routine newborn vaccination. Continuing surveillance at the hospital has revealed five new cases, all vaccinated before beginning the control measures, and one case vaccinated elsewhere.

BCG vaccine is given intradermally in the upper left arm at the insertion of the deltoid muscle and HBV is given intramuscularly at the anterolateral side of the thigh. Vitamin K is given in the opposite thigh. Several lines of evidence suggest that these abscesses were caused by accidental intramuscular injection of BCG into the thigh.

The long incubation period, prolonged healing period and clinical nature of the abscess suggest a slow-growing organism like BCG. Although acid-fast bacilli were not routinely sought in microbiological testing, they were seen in stained preparations from one case. An obvious source of the organism was present in the delivery room. Poor identification of the syringe contents and poor communication between nurses made a mix-up more possible. The problem was specific to the delivery room of one hospital and to one nurse, which implies an error in vaccine administration rather than a defect in the vaccine.

The investigation also identified three suspect cases from other hospitals. Regional health authorities should detect and report new cases of abscesses following routine vaccination. This can identify other sites where vaccination practices need improvement.