British Journal of Anaesthesia 1997; 79: 59 63 Intradermal compared with prick testing in the diagnosis of anaesthetic allergy M. MCD FISHER AND C. J. BOWEY Summary We have tested the hypothesis that intradermal testing is a more effective method for determining the drug responsible for anaesthetic anaphylactic reactions than prick testing in 212 consecutive patients, aged more than 10 yr, referred to an anaesthetic allergy clinic over a 4-yr period. The study was a prospective, non-randomized design. Intradermal testing was conducted using a previously described method and diluted drugs, and prick testing using undiluted drugs (with the exception of opioid analgesics which were diluted 1:10). The tests were performed on individual patients forearms on the same occasion. Patients were followed-up to determine the results of subsequent anaesthesia and the difference between tests was analysed using kappa and tau statistics. There was 93% agreement overall between the paired tests. Which test detected the drug responsible was dependent on diagnostic criteria for positivity. The differences between the tests were not statistically significant. Using both tests improved predictability by 67% (tau 0.67, P 0.001). We conclude that in the absence of data to support one test being superior, other factors influence the choice of test. ing was cheaper, and the reduction in pain and trauma with prick testing makes it more suitable for children. However, there are no data available on the safety of subsequent anaesthesia based on the results of prick testing alone, and reliability with time has not been assessed. Intradermal testing may be easier for the infrequent user. Skin testing is valuable in the investigation of anaesthetic anaphylaxis whichever test is chosen. When there is doubt both tests should be performed. (Br. J. Anaesth. 1997; 79: 59 63). Key words Allergy. Skin, tests. Complications, anaphylaxis. The use of skin testing to verify the anaphylactic aetiology of an adverse event during anaesthesia, establish the drug responsible and predict the safety of alternative drugs has been validated in several studies. 1 7 The high incidence of radioimmunoassay (RIA) tests for specific IgE to induction agents and neuromuscular blocking drugs in patients with skin tests 8 9 suggests that these tests demonstrate IgE antibodies to drugs which give tests and that the reactions in which the tests are are IgE-mediated. Controversy persists as to the best method of skin testing. Intradermal testing has been used in the majority of studies, although three studies have demonstrated that prick testing has a similar efficacy. 10 12 In a prospective comparison of prick and intradermal testing in the same patients, Leynadier and colleagues 11 suggested that prick testing was the better test although seven of 50 reactors with intradermal tests had negative prick tests. Laxenaire and colleagues found no difference in the results in 34 patients using serial dilutions of drugs. 10 To evaluate these tests further, we have investigated prospectively 212 patients referred to an anaesthetic allergy clinic for investigation of possible anaesthetic allergy. Patients and methods Over a 4-yr period, we investigated 212 consecutive adult patients referred to the anaesthetic allergy clinic. Children less than 10 yr of age and patients who allegedly reacted to local anaesthesia were excluded. The hospital Ethics Committees did not require informed consent for testing as it was considered that both tests were valid diagnostic tests. For intradermal testing, drugs were freshly diluted in normal saline the day of the test and testing was performed using a design described previously. 2 Table 1 shows the dilutions used for the first 102 patients (group I) and the second 110 patients in whom dilutions of the neuromuscular blocking drugs were increased (group II) to determine if this affected the results of testing. All drugs used before the reactions were tested, and in the case of neuromuscular blocking agents the other available blockers were tested. This availability varied over the course of the study. Gallamine was used only if it had been administered before the reaction. MALCOLM MCD FISHER, MB, CHB, MD, FRCA, FANZCA, University of Sydney, Intensive Therapy Unit, Royal North Shore Hospital of Sydney, St Leonards, NSW, Australia 2065. C. JANE BOWEY, RN, Intensive Therapy Unit, Royal North Shore Hospital of Sydney, St Leonards, NSW, Australia 2065. Accepted for publication: February 26, 1997. Correspondence to M. F.
60 British Journal of Anaesthesia Table 1 Drug dilutions for skin testing and number of comparisons Drug Standard dilution Group 1 Intradermal test Group 2 Intradermal test Group 3 Times used Morphine 10 mg ml 1 1:100 000 1:100 000 1:10 9 Pethidine 50 mg ml 1 1:100 000 1:100 000 1:10 6 Papaveretum 20 mg ml 1 1:100 000 1:100 000 1:10 3 Fentanyl 100 µg/ml 1 1:100 1:100 Undilute 36 Thiopentone 25 mg ml 1 1:100 1:100 Undilute 175 Methohexitone 10 mg ml 1 1:100 1:100 Undilute 2 Propofol 10 mg ml 1 1:100 1:100 Undilute 20 Diazepam 5 mg ml 1 1:100 1:100 Undilute 3 Midazolam 5 mg ml 1 1:100 1:100 Undilute 1 Suxamethonium 50 mg ml 1 1:1000 1:500 Undilute 167 Alcuronium 5 mg ml 1 1:1000 1:500 Undilute 166 Atracurium 10 mg ml 1 1:10000 1:1000 Undilute 145 Tubocuraine 10 mg ml 1 1:10000 1:1000 Undilute 143 Gallamine 40 mg ml 1 1:1000 1:500 Undilute 4 Vecuronium 4 mg ml 1 1:1000 1:500 Undilute 146 Pancuronium 2 mg ml 1 1:1000 1:500 Undilute 144 Neostigmine 2.5 mg ml 1 1:1000 1:1000 Undilute 7 Atropine 0.6 mg ml 1 1:1000 1:1000 Undilute 10 Cephalothin 100 mg ml 1 1:100 1:100 Undilute 7 Cefotaxime 50 mg ml 1 1:100 1:100 Undilute 1 Haemaccel 3.5 mg ml 1 1:100 1:100 Undilute 5 Dextran 70 6 mg ml 1 1:100 1:100 Undilute 2 Histamine 1:50 1:50 1:50 Undilute 170 Droperidol 10 mg/2 ml 1:100 1:100 Undilute 3 Penicillin 1 000 000 u/5 ml 1:100 1:100 Undilute 3 Protamine 500 mg/10 ml 1:1000 1:1000 Undilute 2 Gentamicin 80 mg/2 ml 1:100 1:100 Undilute 3 For prick testing, the drug solution was placed on the ventral aspect of the forearm and the epidermis pricked through the solution using a 25-gauge needle, bevel uppermost at 45 C to the skin, to a depth at which the skin could be lifted. Morphine, papaveretum and pethidine were diluted 1:10 in saline as preliminary studies suggested that false s were caused by these drugs if undiluted. All other drugs were undiluted (table 1). For intradermal testing, a result was recorded for a persistent wheal of greater than 0.8 cm and for the prick testing a result was recorded for a wheal of greater that 0.4 cm. If the test was for a drug administered before the reaction being investigated the result was recorded as diagnostic, that is caused by that drug. The agreement between tests was assessed statistically using Kappa 13 14 and Tau 15 tests. All patients were given a letter describing the event and the results of testing, and details of subsequent anaesthesia were sought. The patients were divided in four groups: minor reactions 29 patients had reactions confined to the skin, or systemic reactions which resolved rapidly without treatment; severe reactions 135 patients had systemic reactions requiring treatment with vasoactive drugs; 129 of these were felt by the anaesthetist to have been life threatening; preoperative 13 patients were referred because of multiple allergies and were to undergo anaesthesia; not anaphylactic 35 patients had single system reactions in which another cause (assessed by an independent physician from the history) was likely. Results The results for all patients are shown in table 2. Overall there was agreement for 1294 (93%) comparisons and disagreement for 98 (7%); there was no significant difference between the agreement in the different groups. The highest diagnostic yield was in patients with severe reactions in which a diagnosis was made in 116 of 135 patients. The majority of these patients reacted to neuromuscular blocking drugs (93 patients), and the other drugs implicated by skin testing were induction agents (13 patients), induction agent with neuromuscular blocking agent (two patients), antibiotics (three patients) opioid analgesics (three patients) and Haemaccel (two patients). Intradermal testing in this group was for more drugs than prick testing overall, but less when Table 2 ing and intradermal testing in 212 patients: overall results Group No. of patients tested Drug responsible detected Agreement between tests No. of comparisons Prick Intradermal Both Either Neither Agree (%) Disagree (%) Minor reactions 29 117 8 5 4 9 20 113 (97%) 4 (3%) Severe reactions 135 965 108 112 104 116 19 890 (92%) 75 8(%) Preop. 13 82 1 1 0 2 11 75 (90%) 7 (9%) Not anaphylactic 35 228 0 0 0 0 35 216 (95%) 12 (5%) Total 212 1392 117 118 108 127 85 1294 (93%) 98 (7%)
Diagnosis of anaesthetic allergy 61 Table 3 Prick vs intradermal testing for severe reactors. Changing diagnostic criteria to any response to prick test No. Intradermal Both Either Neither Agree Disagree Criterion I Wheal>4mm 135 108 112 104 116 19 890 (92%) 75 (8%) Criterion II Any wheal 135 115 112 106 112 14 855 (89%) 110 (11%) Table 4 Effects of altering intradermal test concentration on reliability of tests in patients with severe reactions No. Intradermal Both Either Neither Agree Disagree Test I Normal concn 74 58 63 57 64 10 518 (93.50) 36 (6.50) Test II Increased concn 61 50 49 47 52 9 372 (90.50) 39 (9.50) Table 5 Results of subsequent anaesthesia in patients tested by prick and intradermal testing. Group n Subsequent GA uneventful Minor reactions 29 10 0 Severe reactions Drug determined 116 65 1 Skin tests negative 19 17 0 Preoperative 13 12 0 Not anaphylaxis 35 20 0 Subsequent prick test results were considered if wheal size was less than 4 mm (table 3). This increase with prick testing was not statistically significant. The agreement between intradermal and prick teasing was substantial in the severe group (kappa 0.7) and excellent overall (kappa 0.8). In spite of this degree of agreement, the tau analysis suggested a 50% reduction in predictive error if both tests were used in the severe group (tau 0.5, P 0.001) and a reduction in predictive error of 67% if both tests were used in all patients (tau 0.67, P 0.001). The majority of disagreement in reactions to neuromuscular blocking drugs was not in the drug responsible but in the detection of cross sensitivity to other neuromuscular blocking agents. In the severe group of patients, increasing the concentration of neuromuscular blocking agent did not alter significantly the diagnostic yield of the test (kappa 0.73 vs kappa 0.68) (table 4). Details of subsequent anaesthesia were obtained in 124 patients (table 5). One patient had a second reaction. This patient, who reacted to pancuronium, had prick and intradermal tests to pancuronium and a subsequent reaction to which we attributed to alcuronium on the basis of history. Both tests were negative for alcuronium after both reactions. One patient had a minor systemic reaction after testing with thiopentone and histamine only, and this responded to adrenaline 0.2 mg s.c. Discussion The major problem with studies of this nature is that there is no absolute measure with which to compare the results. 16 Thus when there is disagreement it is not possible to tell which test is correct, without subjecting the patient to the hazards of challenge. As RIA tests are less sensitive and specific than skin tests, 16 they are not suitable to discriminate between tests which disagree, although when for an individual drug they add support to the skin test diagnosis. We have shown previously that basing subsequent anaesthesia on diagnosis and detection of cross s ens itivity from intradermal tes ting is us ually reliable, with only three anaphylactic reactions in subsequent anaesthesia in 234 patients who had lifethreatening reactions before testing. 16 The data from this study confirm that in patients with anaphylaxis during anaesthesia, subsequent anaesthesia based on the results of prick and intradermal testing is usually, but not always, safe. There have been no published studies showing the safety or otherwise of subsequent anaesthesia based on the results of prick tests alone. Galletly 17 suggested that the dilutions chosen in our studies for intradermal testing (which were chosen by determining dilutions that did not cause wheals as a result of direct histamine release in volunteers) were greater than necessary. This could be a source of false negative results. Increasing the concentrations of neuromuscular blocking agents for intradermal testing did not significantly alter the diagnostic value of the test but increased the discrepancy between prick and intradermal tests for other neuromuscular blocking agents. The majority of this increase was related to reactions to the potent cutaneous histamine releasers tubocurarine and atracurium. The diagnostic criteria for positivity were fulfilled more often by intradermal testing than by prick testing when a result for prick testing was a wheal greater than 4 mm. Moneret Vautrin and Laxenaire 18 have suggested that any wheal should be considered for prick testing and, using this criterion, prick testing produced a diagnostic result more commonly than intradermal testing. It cannot be determined however whether this was an increase in false s or true allergy detected. Both tests
62 British Journal of Anaesthesia produced more diagnoses than each test alone, and the statistical analysis suggested the safety of subsequent anaesthesia would be improved by performing both tests. There is little to choose between each test on the basis of the above data in which a single intradermal test was used for each drug, rather than the serial dilutions used in the French study which showed 100% agreement in 38 patients. 10 It is possible that using serial dilutions rather than a single injection would increase the diagnostic value of intradermal 10 18 testing even further. ing has advantages over intradermal testing: it is cheaper than intradermal testing (by approximately 25% in our unit); the reduced pain and trauma of prick testing makes it more suitable for use in children; and it is said to be safer. 18 There are no data to support this view for anaesthetic drugs. Indeed the low incidence of systemic reactions to skin testing after anaesthetic reactions (three in more than 1200 patients in our series) would imply that a very large study is necessary to show a significant difference. It has been suggested that intradermal tests are more likely to produce false s than prick tests. 18 This would favour the use of intradermal tests, as a false test inconveniences the anaesthetist and does not carry the risk of anaphylaxis of a false negative test. 8 Intradermal testing is more reliable in propofol reactions. 19 Intradermal testing has been validated in repeat studies over time 20 21 but no data are available for prick testing. The study of Leynadier, Sansarricq and Dry 21 suggested that the persistence of tests was insufficiently reliable on which to base subsequent drug selection and further preoperative skin tests should be performed, but our own study 20 and the anaesthetic follow-up of allergic patients, 16 suggest that the tests persist for up to 30 yr and preoperative testing is only mandatory if untested drugs to which the patient may show cross sensitivity are to be used. One other practical factor worthy of consideration is ease of performance. We believe that the anaesthetist who is not regularly testing patients would find intradermal testing easier to perform and interpret. It is important to emphasize that history is also very important in the follow-up of patients with possible anaesthetic allergy. In 19 patients with a severe reaction in whom a diagnosis was not made with skin testing, a diagnosis could be made from history or RIA testing in 17 of 19 cases. In six patients with a negative skin test, the diagnosis could be made from the history as only one drug was administered within 10 min of the reaction. Two patients had negative cutaneous responses to histamine controls and had a diagnosis made by RIA testing. One patient who had previous skin and RIA testing was re-tested 7 yr later and all tests were negative; we concluded that sensitivity had been lost. Two patients had strongly RIA and skin tests for neuromuscular blocking agents other than the one administered and one patient who had reacted to decamethonium and suxamethonium on separate occasions had strongly RIA to suxamethonium and skin tests to decamethonium only. She was not re-tested with decamethonium as it is no longer available. (She has subsequently had a intradermal test to suxamethonium 3 yr later.) In these three patients we suggested that all drugs and the one used should be avoided. Three other patients had intradermal tests implicating a drug used, but not reaching the criterion for positivity in this study, and were warned against the drugs suggested. One patient had a negative mast cell tryptase test and negative skin tests, making an anaphylactic aetiology unlikely. One patient, who was tested only for thiopentone and histamine as thiopentone was the sole agent used, had negative skin tests but a mild anaphylactic reaction. In only two patients who had severe reactions could no aetiological agent be found. Both of these patients and 15 of the others who did not the meet criteria for skin tests have undergone subsequent uneventful anaesthesia. In patients referred before operation, one patient had a prick test but not intradermal test to three drugs and one patient had a intradermal test but not a prick test to three neuromuscular blocking agents. These drugs were avoided for anaesthesia which was uneventful in all patients. We conclude that patients who have s evere reactions during anaesthesia which might be allergic should undergo skin testing. There was no significant difference in the reliability of diagnostic yield of intradermal or prick testing, and the choice of test could be based on other factors such as age of the patient, cost, ease of performance, unreliability of prick testing with propofol, lack of data on reliability of prick testing with time and outcome of subsequent anaesthesia in allergic patients when drug choice is based on prick testing. If the results of either test is negative or equivocal, intradermal and RIA testing should be performed. Where there are conflicting results all drugs which give tests by either method should be avoided. The results of skin testing should be communicated in a letter stating drugs given, what happened, results and conclusions. The results of subsequent anaesthesia should be documented in the patient s letter and the file letter. In reactions to neuromuscular blocking agents, the results agreed better for detecting the drug responsible for a reaction than determining cross sensitivity to other neuromuscular blocking drugs. It is sensible to suggest avoiding all neuromuscular blocking agents that may be as cross 16 22 sensitivity occurs in 60% of patients. Safe subsequent anaesthesia is usual but not guaranteed when patients, who have had adverse reactions which may be allergic, receive drugs suggested to be safe on the basis of skin tests and history. Acknowledgement We acknowledge the help and expertise of Caro Badcock, medical statistician, Royal North Shore Hospital.
Diagnosis of anaesthetic allergy 63 References 1. Sage D. Intradermal drug testing following anaphylactoid reactions during anaesthesia. Anaesthesia and Intensive Care 1981; 9: 381 386. 2. Fisher M. Intradermal testing after anaphylactoid reaction to anaesthetic drugs: practical aspects of performance and interpretation. Anaesthesia and Intensive Care 1984; 12: 115 120. 3. Thacker MA, Gibbs JM. A hypersensitivity screening clinic following untoward reactions to anaesthesia. New Zealand Medical Journal 1984; 97: 232 234. 4. Galletly DC, Treuren BC. Anaphylactoid reactions during anaesthesia. Seven years experience of intradermal testing. Anaesthesia 1985; 40: 329 333. 5. Moscicki RA, Sockin SM, Corsello BF, Ostro MG, Bloch KJ. Anaphylaxis during induction of general anesthesia: Subsequent evaluation and management. Journal of Allergy and Clinical Immunology 1990; 86: 325 332. 6. Facon A, Grzybowski M, Divry M, Tsicopoulos A, Scherpereel P. Allergy to anesthetic agents: analysis of 96 recent cases. Cahiers d Anesthesiologie 1988; 36: 97 100. 7. Vervloet D, Arnaud A, Vellieux P, Kaplanski S, Charpin J. Anaphylactic reactions to muscle relaxants under general anesthesia. Journal of Allergy and Clinical Immunology 1979; 63: 348 353. 8. Fisher MM, Baldo BA. Role of IgE in anaphylactoid reactions during anaesthesia. Annales Francaises d Anesthesie et de Reanimation 1985; 4: 133 136. 9. Harle DG, Baldo BA, Fisher MM. Detection of IgE antibodies to suxamethonium after anaphylactoid reactions during anaesthesia. Lancet 1984; 1: 930 932. 10. Moneret-Vautrin DA, Laxenaire MC, Widmer S, Hummer M. The value of prick tests in the detection of anaphylaxis caused by muscle relaxants. Annales Francaises d Anesthesie et de Reanimation 1987; 6: 352 355. 11. Leynadier F, Sansarricq M, Didier JM, Dry J. s in the diagnosis of anaphylaxis to general anaesthetics. British Journal of Anaesthesia 1987; 59: 683 689. 12. Pepys J, Pepys EO, Baldo BA, Whitwam JG. Anaphylactic/anaphylactoid reactions to anaesthetic and associated agents. Skin prick tests in aetiological diagnosis. Anaesthesia 1994; 49: 470 475. 13. Landis JR, Koch GG. An application of hierarchical kappatype statistics in the assessment of majority agreement among multiple observers. Biometrics 1977; 33: 363 374. 14. Thompson WD, Walter SD. Kappa and the concept of independent errors. Journal of Clinical Epidemiology 1988; 41: 969 970. 15. Goodman LA, Kruskal WH. Measures of Association for Crossclassification. New York: Springer-Verlag, 1996. 16. Fisher MM, Baldo BA. Anaphylaxis during anaesthesia: current aspects of diagnosis and prevention. European Journal of Anaesthesiolgy 1994; 11: 263 284. 17. Galletly DC. Comparative cutaneous histamine release by neuromuscular blocking agents. Anaesthesia and Intensive Care 1986; 14: 365 369. 18. Moneret Vautrin DA, Laxenaire MC. Skin tests in diagnosis of allergy to muscle relaxants and other anaesthetic drugs. Monographs in Allergy 1992; 30: 145 155. 19. Laxenaire MC, Mata Bermejo E, Moneret Vautrin DA, Gueant JL. Life-threatening anaphylactoid reactions to propofol (Diprivan). Anesthesiology 1992; 77: 275 280. 20. Fisher MM, Baldo BA. Persistence of allergy to anaesthetic drugs. Anaesthesia and Intensive Care 1992; 20: 143 146. 21. Leynadier F, Sansarricq M, Dry J. Reproducibility of intradermal tests after anaphylaxis caused by muscle relaxants. Presse Medicale 1987; 16: 523 525. 22. Fisher MM, Munro I. Life-threatening anaphylactoid reactions to muscle relaxants. Anesthesia and Analgesia 1983; 62: 559 564.