Adverse reactions to antibiotic drugs are common, and up to 25% of patients who are treated with antimicrobial agents report allergic reactions to antimicrobials [1]. Penicillin is the most frequently implicated agent; approximately 10% to 20% of hospitalized patients have a history of penicillin allergy [1,2]. As the use of semisynthetic penicillins (amoxicillin and ampicillin) and cephalosporins has increased, these agents are becoming a major cause of allergic responses [3]. Generally, non-beta-lactam antibiotics such as vancomycin and fluoroquinolones are commonly used in patients with a history of beta-lactam allergy to avoid allergic reactions. The use of alternative antibiotics is associated with an increased number of infections caused by multidrug-resistant bacterial strains, which is associated with higher costs for care, morbidity, and mortality [4-7]. To solve this problem, identifying patients in the clinical setting who are at high risk for immediate IgE-mediated reactions is essential. Antibiotic skin testing (AST) is well-known to be a safe and reliable method for detecting immediate allergic reactions to beta-lactams. Although skin testing is a routine antibiotic test in Korean hospital wards, methods vary from hospital to hospital. The purpose of this study was to examine the current practice of AST in Korea.

We selected 12 hospitals that employed allergy specialists. To represent general practices in Korea, we tried to avoid deviations in grade and location of the hospitals, so three secondary hospitals (J to L, Table 1) and nine tertiary hospitals (A to I, Table 1) were enrolled. Among them, one secondary hospital (J, Table 1) and five tertiary hospitals (C to G, Table 1) were located in Seoul, and two secondary hospitals (K and L, Table 1) and four tertiary hospitals (A, B, H, and I, Table 1) were located in the province. Questionnaires were sent to the allergists, and the responses were collected either by e-mail or fax. The questionnaire included items such as the type and concentration of the antibiotics tested, the testing methods, and the interpretation of the test results. The questions were as follows:

1) Do you routinely conduct AST before using antibiotics in your wards? If so, how many and which antibiotics do you test? 2) Clarify the details of the methods for AST, e.g., testing method (skin prick testing vs. intradermal testing), concentrations and doses of antibiotics tested, test site, and whether a control substance is tested. 3) Clarify the method for AST interpretation, e.g., time lag between testing and interpretation and criteria for positivity.

The response rate was 100%. In all hospitals, the AST was a routine procedure before starting certain antibiotics in general wards and was usually conducted by a nurse without regard to allergic reaction history. However, the number of antibiotics tested varied from three to five. Among 12 hospitals, the ASTs for penicillins and cephalosporins were performed in 10 and 11 hospitals, respectively. In one hospital (E, Table 1), ASTs for cephalosporins were not routinely performed based on the idea that no study has demonstrated the clinical utility of cephalosporin skin testing for screening in subjects without a drug allergy history. Moreover, in two hospitals (H and I, Table 1), ASTs for penicillins were not performed because penicillins were never prescribed in these hospitals for fear of an adverse reaction. For other beta-lactam antibiotics, semisynthetic penicillins such as ampicillin and amoxicillin were tested in eight hospitals, carbapenems such as imipenem and meropenem in nine hospitals, and monobactams such as aztreonam in one hospital. Additionally, ASTs for vancomycin were conducted in four hospitals, and ASTs for aminoglycosides such as streptomycin were carried out in three hospitals. The details are shown in Table 1. Antibiotics other than those listed above were used without prior skin tests in all hospitals.

The ASTs were performed on the volar forearm using the intradermal method. However, skin tests using a control solution (e.g., saline) were not routinely conducted with the AST. Skin tests with saline were performed as a negative control in only two hospitals (F and G, Table 2), following the AST and for suspicious, false-positive results cases. The injection amounts varied among the hospitals, ranging from 0.02 to 0.3 mL (Table 2). The concentrations tested also varied among hospitals (Table 1). For example, testing concentrations for penicillin G showed a 10,000 fold difference, ranging from 5 IU/mL to 50,000 IU/mL. However, the time lag (15 to 20 minutes) between testing and interpretation was similar in all hospitals. The criteria for positive results on the AST varied and were inconsistent among the hospitals: either a wheal or erythema larger than standard in six hospitals, only a wheal regardless of erythema in four hospitals, and only erythema in one hospital. Only one hospital considered both a wheal and erythema for interpreting the results. Additionally, the wheal and erythema standards for a positive response varied among the hospitals (Table 2).

We examined the current practice of AST in 12 Korean general hospitals with different grades and locations. We emphasize that what we examined in this study was the routine testing performed for screening in all patients in general wards and not what allergy specialists do. Unexpectedly, the AST protocols varied from hospital to hospital with regard to the types of antibiotics tested, testing concentrations and amounts, and the interpretation method. Therefore, discussing the similarities and differences between the current practices and the commonly recommended procedures in the medical literature is warranted.