HIV/AIDS News & Information

Epidemiologic Investigation
A total of 28 patients met the case definition; none had more than one abscess. The 28 case-patients received the injections associated with their abscesses during a 2-month period (April 20-June 22, 1995). Case-patients ranged in age from 13.6 to 79.8 years (mean = 43.2); 23 (82.1%) were female.

All case-patients had been given a corticosteroid injection to relieve allergy symptoms: Twenty-four (85.7%) received triamcinolone acetonide (ТА) (dose range: 30-80 mg), three (10.7%) received methylpred-nisolone acetate (dose: 80 mg), and for one case-patient, whether ТА or triamcinolone diacetate was given could not be determined. For two patients, we could not determine whether the injection of methylprednisolone or another injection given at the same time was the one associated with abscess formation. These case-patients were excluded from analyses involving medication. Abscesses began draining (spontaneously or surgically) a median of 75.5 (range: 32-211) days after the injection. Five (17.9%) case-patients were hospitalized for 1-13 days (median = 2 days).

We interviewed 126 (90%) of the 140 patients who had received IM injections on the same days that case-patients received the injection associated with their abscess. No significant differences with respect to sex, mean age, or proportion receiving a corticosteroid injection existed between the 14 control patients who could not be interviewed and the 126 patients who participated in the study.

When we compared case- and control-patients, abscess formation was not associated with immunosup-pressive conditions, age, or the person administering the injection. Being female, corticosteroid injection, injection from a multidose vial, and injection in the gluteal muscle were significantly associated with abscess formation (Table 1). When we adjusted the effects of the risk factors listed in Table 1 with a conditional logistic regression model, the effects of injection site and type of vial were inestimable in the model due to lack of variation between case- and control-patients within too many matched case-control groups. We removed these two risk factors and constructed a model with the remaining variables (Table 2). Only receipt of corticosteroid (OR = 17.8) and being female (OR = 7.4) were significantly associated with abscess formation. generic ampicillin

TABLE 1 —Comparison of Risk Factors for Abscess Formation in Case- and Control-Patients

ВС = benzalkonium chloride, Undef = undefined
Conditional maximum likelihood estimate of odds ratio; case- and control-patients matched on injection date.
! Exact 95% confidence interval.
!! Body mass index; only calculated for adults.

TABLE 2—Multivariable Analysis of Risk Factors for Abscess Formation

ВС = benzalkonium chloride
*Conditional maximum likelihood estimate of odds ratio; case- and control-patients matched on injection date.
! Exact 95% confidence interval.
!! Only calculated for adults.

Environmental and Laboratory Investigation
Cultures from cotton balls soaked in the demonstration benzalkonium chloride solution from the clinic grew P aeruginosa. The solution was prepared monthly by mixing approximately one part Zephiran® Chloride (17%) and two parts commercial normal saline in a glass canister and adding cotton balls. The canister was washed with soap between batches of benzalkonium chloride. None of the other environmental samples collected from the clinic were contaminated with P aeruginosa.

Unopened vials of ТА passed standard sterility tests conducted by the manufacturer. The FDA found no evidence of microbial contamination in 1 opened, multidose vial of ТА or in the 18 opened, multidose vials of other injectable medications. buy brand cialis

The abscesses of 23 case-patients were cultured; laboratory records showed that 20 (87.0%) grew P aeruginosa, and one grew Pseudomonas fluorescen s/putida. We obtained P aeruginosa isolates from three case-patients: All isolates were confirmed to be P aeruginosa by SLD. Isolates from two patients were identified as the same strain by MIS and pulse-field gel electrophoresis.