Allergy to penicillin

SIR,-Dr S J Surtees and colleagues' report on allergy to penicillin is misleading.' Though their conclusion that most patients with minor or vague reactions to penicillin with a negative result of a radioallergosorbent test to penicillin can be given oral penicillin safely is probably true, this is likely to apply equally to patients who have not had such tests performed. It is important to emphasise that radioallergosorbent tests for penicillin allergy are of little value. There is extensive evidence from studies of large numbers ofpatients (table) that skin tests give both false positive and false negative reactions.2 These studies show that although higher proportions of patients with a positive skin test result (in the examples shown in the table 0%, 7%, and 50%) than with a negative skin test result (0%, 0 15%, 0 5%, and 4%) react to subsequent challenge with penicillin, most patients in both groups do not. Clearly, in individual patients the tests do not predict outcome. Only studies such as these in which challenge tests were performed can be used to validate the results of skin tests or radioallergosorbent tests. Studies such as those quoted by Dr Surtees and colleagues' in which no challenge tests were performed and which simply relate the results of skin tests or radioallergosorbent tests to the history cannot be used to validate these tests or to exclude or confirm allergy. Unfortunately, Dr Surtees and colleagues did not challenge their subjects who had positive results on radioallergosorbent testing, and there were only four of these. Comparison of methods of testing for penicillin IgE antibodies-skin testing and radioallergosorbent testing-is not straightforward as various allergens (major and minor penicillin determinants) can be used for skin tests whereas the radioallergosorbent test can be done only for the major penicilloyl determinant. A proportion of reactions are due to IgE antibodies to minor determinants. In general the more skin test reagents that are used the more positive reactions are obtained. Full skin testing is therefore more sensitive than radioallergosorbent testing. If only the major determinant allergens (and these are the only ones routinely available) are used, however, the results of skin testing and radioallergosorbent testing correlate fairly well.' The benzylpenicilloyl radioallergosorbent test gives a positive result in 60-95% of patients with a positive skin test result to penicilloyl-polylysine. The data shown in the table, with the exception of those from the study in which nine skin test extracts were used testing both major and minor determinants, are therefore likely to apply to results of radioallergosorbent tests. Few studies ofchallenge in patients with positive results on radioallergosorbent testing have been reported, and fewer patients with positive than negative results have been challenged. Of critical importance, when considering an individual patient, is the fact that a positive or negative result on radioallergosorbent testing or skin testing cannot be relied on to predict outcome. The only absolutely certain test is to challenge the patient with the drug, and this must be done using the appropriate preparation and route of administration starting with small doses (extremely small in patients with a history of a serious reaction to parenteral penicillin) and working up to a full dose (for a parenteral drug) or therapeutic course (for example five days' treatment) for an oral preparation. Patients developing minor reactions, such as rashes, may do so after several days' treatment. Our own practice is to see patients and obtain a detailed history. A good history is probably of more value in managing these patients than results ofskin or radioallergosorbent tests. Only in patients who have a history of serious (anaphylactic) reactions or who need treatment with penicillin (for a disease for which no other antibiotic is suitable) is a challenge performed. The number falling into this category is extremely small compared with the many referred with putative penicillin allergy. As a challenge test would still be necessary in this group of patients, irrespective of the result of the skin test or radioallergosorbent test to the major determinant, there is no indication to do these tests routinely. I recognise that the vast majority of patients with suspected penicillin allergy are indeed not sensitive and that patients with a negative result to penicilloyl radioallergosorbent testing are at low risk of reactions to penicillin; this test alone, however, does not exclude the possibility of anaphylactic reactions. At best the radioallergosorbent test can be said to have weak inferential value only when it gives a positive result. The