1 The American Journal of Medicine (2006) 119, 354.e e20 CLINICAL RESEARCH STUDY Is There Cross-Reactivity Between Penicillins and Cephalosporins? Andrea J. Apter, MD, MSc, a,b,c Judith L. Kinman, MA, b Warren B. Bilker, PhD, b,d Maximilian Herlim, b David J. Margolis, MD, PhD, b,d,e Ebbing Lautenbach, MD, MPH, MSCE, b,d,f Sean Hennessy, PharmD, PhD, b,c,d Brian L. Strom, MD, MPH b,c,d a Division of Pulmonary, Allergy Critical Care Medicine, Department of Medicine; b Center for Clinical Epidemiology and Biostatistics and Center for Education and Research on Therapeutics; c Leonard Davis Institute of Health Economics; d Department of Biostatistics and Epidemiology; e Department of Dermatology; f Division of Infectious Diseases, Department of Medicine, University of Pennsylvania, Philadelphia, Pa. ABSTRACT BACKGROUND: We sought to determine the risk of an allergic reaction to a cephalosporin exposure in those with prior penicillin reactions. METHODS: We conducted a retrospective cohort study using the United Kingdom General Practice Research Database. We selected all patients receiving a prescription for penicillin followed by a prescription for a cephalosporin and identified allergic-like events within 30 days after each prescription. Allergic events were defined by 2 sets of codes: 1 more restrictive, 1 more inclusive. Comparison was made with a population of patients receiving a prescription for a penicillin followed by a prescription for a sulfonamide antibiotic. RESULTS: A total of patients received a penicillin; (15%) received a subsequent cephalosporin. Among patients receiving a penicillin followed by a cephalosporin, the unadjusted risk ratio of an allergic-like event for those who had a prior event, compared with those who had no such prior event, narrowly defined, was 10.1 (confidence interval ). The absolute risk of anaphylaxis after a cephalosporin was less than 0.001%. The unadjusted risk ratio for sulfonamide antibiotic, rather than cephalosporin after penicillin allergic-like events was 7.2 (confidence interval ). CONCLUSION: Patients with allergic-like events after penicillin had a markedly increased risk of events after either subsequent cephalosporins or sulfonamide antibiotics. Cross-reactivity is not an adequate explanation for this increased risk, and the risk of anaphylaxis is very low. Thus, our data indicate that cephalosporins can be considered for patients with penicillin allergy Elsevier Inc. All rights reserved. KEYWORDS: Antibiotic allergy; Adverse drug reaction; Drug hypersensitivity; Penicillin; Cephalosporin Support was provided by the Agency for Healthcare Research and Quality Centers for Education and Research on Therapeutics Cooperative Agreement (grant HS10399) and the National Heart, Lung, and Blood Institute (HL K , Dr. Apter). Requests for reprints should be addressed to Andrea J. Apter, MD, MSc, 829 Gates Building, Hospital of the University of Pennsylvania, 3600 Spruce Street, Philadelphia, PA address: Whether patients allergic to penicillin are at increased risk for an allergic reaction to cephalosporins has remained unknown since the introduction of cephalosporins. 1-4 Spurring this question is the fact that antibiotics share a beta-lactam ring. Indeed, cross-reacting antibodies have been identified in the laboratory. 5 Hypersensitivity to cephalosporins also has been reported in patients with penicillin allergy with a prevalence ranging from 3% to 18%. 1,2,6-8 At least 1 reported fatality resulted from anaphylaxis attributed to a cephalosporin administered to a penicillin-allergic patient. 9 However, not all reports support the notion that patients allergic to penicillin are at increased risk for an allergic reaction to a cephalosporin. 10,11 No controlled epidemiologic studies address the /$ -see front matter 2006 Elsevier Inc. All rights reserved. doi: /j.amjmed
2 354.e12 The American Journal of Medicine, Vol 119, No 4, April 2006 question of clinical cross-reactivity. 1,2 Nevertheless, physicians are cautioned about prescribing cephalosporins for penicillin-sensitive patients because cross-sensitivity among beta-lactam antibiotics has been clearly documented and may occur in up to 10% of patients with a history of penicillin allergy. 12 A national task force recently emphasized the importance of research to improve prevention and management of drug hypersensitivity reactions. 13 A major component of this research is understanding the risk of cephalosporin hypersensitivity in patients with penicillin allergy. With the increase in bacterial antimicrobial resistance, particularly multidrug resistance, therapeutic options are increasingly limited. Furthermore, avoidance of an antibiotic because of an allergy history may result in the selection of a less effective or more toxic antibiotic. To study the question of allergic cross-reactivity between penicillins and cephalosporins, we conducted a retrospective cohort study using the United Kingdom General Practice Research Database (GPRD), a database of electronic primary care medical records and prescriptions. To better understand etiology, we compared our results with analyses replacing cephalosporin with sulfonamide antibiotics because there is no known structural or immunologic cross-reactivity between penicillin and sulfonamide antibiotics. METHODS Data Source: General Practice Research Database GPRD contains the actual outpatient medical records from 687 general practitioner (GP) practices geographically representative of England and Wales, comprising 6% of this population. The database includes demographic information, prescription drug information, clinical events and diagnoses, and hospital admissions, as entered by the GP. Contributing GPs are required to meet specific recording standards to be considered Up to Standard (UTS) Prior studies suggest that the clinical information in GPRD is of high quality for epidemiologic studies The UTS data provide a window of observation for patients, that is, prescriptions written before or after the UTS dates were not considered in this study. CLINICAL SIGNIFICANCE Patients with allergic-like events following penicillin have a markedly increased risk of an allergic-like event following a subsequent cephalosporin. Such patients also have a similarly increased risk of an allergic-like event following a sulfonamide antibiotic. Crossreactivity is not an adequate explanation for this finding. Patients with allergic-like events following penicillin have a very low risk of anaphylaxis following a subsequent cephalosporin ( 0.001%). Our data indicate that cephalosporins can be considered for patients with penicillin allergy. Study Population We identified all patients in the GPRD UTS database from 1987 to September 2001 who received a prescription for penicillin (Figure 1). Within that population, we identified patients who also received a cephalosporin prescription at least 60 days after a penicillin prescription. The 60-day interval between the first penicillin and the subsequent cephalosporin prescription ensured that an event after the cephalosporin was not a late event associated with the earlier penicillin. This cohort was used for the primary analyses. Two subgroups were compared to calculate the risk of an allergiclike event after a cephalosporin. The study group consisted of patients who had an allergic-like event within 30 days after a penicillin prescription. The first penicillin prescription followed by an allergic-like event was called the index penicillin prescription. Comparison subjects never had an allergic-like event within 30 days after any penicillin prescription. For the comparison group, the index penicillin prescription was the first prescribed penicillin. Antibiotics Identification From the British National Formulary, we identified codes for cephalosporins (British National Formulary code 5.1.2) and classified them by generation (Appendix A). 28,29 We also identified codes for penicillin (British National Formulary code 5.1.1) 28 and sulfonamide antibiotics (Appendix A). Allergic-Like Events Identification There is no clinical gold standard for a drug hypersensitivity reaction. Thus, we constructed 2 sets of codes, 1 restrictive and 1 more inclusive, for an allergic-like event, recorded 1 to 30 days after an antibiotic prescription. Both sets included diagnoses like anaphylaxis, urticaria, angioedema, erythema multiforme, laryngeal spasm, drug-induced dermatitis, and toxic epidermal necrolysis (Appendix B). Because diagnostic codes like bronchospasm or asthma could represent an adverse drug reaction, but frequently refer to asthma independent of adverse drug reactions, we excluded such codes from the narrowly defined set. Bronchospasm, asthma, eczema, and adverse drug reactions were then included in the broadly defined set of codes. The results were comparable. Unless otherwise specified, the results presented are those using the narrow definition.
3 Apter et al Penicillin and Cephalosporin Allergy 354.e13 Figure 1 Study protocol. The narrow definition of allergic-like event (ALE) is used. Primary Analyses The cohort was defined using Knowledge Manager (IMS Health, Frankfurt, Germany), a patient-level data warehouse and analysis tool developed for accessing GPRD. Analyses were performed using SAS Version 8.1 (SAS Institute Inc., Cary, NC) and STATA, Release 7.0 (StataCorp, College Station, Tex). Analyses were performed among all patients receiving 1 or more prescriptions for penicillin, reviewing the occurrence of allergic-like events 1 to 30 days after a prescription for penicillin and the probability of receiving a subsequent cephalosporin prescription. The primary analyses focused on all patients who received penicillin followed at least 60 days later by a cephalosporin prescription. We excluded events occurring on the prescription day because it was impossible to discern whether a drug given on the day of an event was responsible for an event or prescribed afterward. However, we included the antibiotic prescription day in a sensitivity analysis. The risk ratio (RR) was computed for the risk of an allergic-like event after the cephalosporin prescription in patients experiencing an allergic-like event after a prior penicillin exposure, compared with those with no events after the penicillin exposure. For perspective, we computed the RR for an allergic-like event after a sulfonamide antibiotic prescription (rather than a cephalosporin) for patients experiencing an allergic-like event within 30 days after a penicillin prescription. All ratios were computed with 95% confidence intervals (CIs). Logistic regression, which yields an odds ratio (OR) rather than an RR, was used to evaluate variables as potential confounders, examining each (eg, prior urticaria) individually. Confounders changing the point estimate of the OR of interest by 15% or more were included in the final adjusted model. 30 Age, sex, and cephalosporin generation were tested as effect modifiers of the relationship between exposure (ie, reaction to a penicillin) and outcome (ie, reaction to a cephalosporin) using the Breslow-Day Test for OR homogeneity. 31 Age was considered a continuous and categorical variable to allow for a nonlinear effect. Other potential confounders were the occurrence, before the index penicillin, of acute processes with manifestations similar to an acute hypersensitivity reaction, including diseases associated with a rash resembling an adverse drug reaction such as Epstein- Barr virus or vasculitis (Appendix C). We also adjusted for prior history of allergic-like events that together constitute the events under study such as urticaria or asthma (broad
4 354.e14 The American Journal of Medicine, Vol 119, No 4, April 2006 Table 1 Demographics of Subjects by Study Group or Comparison Group* Study Group* N 3920 Comparison Group N Total N Age (y) N % N % N % Sex Male Female *Proportions for the broad dataset were similar. Age was missing for 252 subjects. event codes only), both singly and as a group, and current use (within 30 days before the index cephalosporin prescription) of drugs known or suspected to be indicators of hypersensitivity/allergic reactions, such as corticosteroids or antihistamines. Finally, we accounted for the nonindependent nature of multiple patients from the same GP practice (ie, clustering) using the Huber-White sandwich estimator 32,33 for the variance of the model parameter estimates. 34 Sensitivity Analyses Several sensitivity analyses were performed to support the validity of our primary analysis. We included the antibiotic prescription day in the time interval for allergic-like events. We computed the RR limiting the interval to 10 (rather than 30) days after the penicillin prescription, the cephalosporin prescription, and both prescriptions. We excluded patients without 30 days of follow-up after the cephalosporin. Finally, we excluded those with penicillin prescriptions on the same day or within 30 days of the cephalosporin. We received approval from the Scientific and Ethical Advisory Group of the GPRD. The University of Pennsylvania Institutional Review Board determined that the protocol warranted exempt status. RESULTS Subjects Of patients receiving at least 1 prescription for penicillin (Figure 1), patients (using a narrow definition; patients using a broad definition of allergiclike event) received prescriptions for a penicillin followed by a cephalosporin. The demographics of these patients are presented in Table 1. Of the prescribed cephalosporins, 64% were first-generation antibiotics. Using either the broad or narrow definitions, 22% of the patients who had an allergic-like event after a penicillin were given a subsequent prescription for cephalosporin, compared with 16% (broad definition, 14%) who had never had an allergic-like event after penicillin. Allergic-Like Events The absolute numbers of allergic-like events other than urticaria were very small (Table 2). Twenty-five patients experienced anaphylaxis after the penicillin, and of these, 1 subject experienced anaphylaxis after the cephalosporin. In all, anaphylaxis was reported for 13 subjects after the cephalosporin. In this analysis, 30 (70%) of 43 patients who Table 2 Description of Allergic-like Events after Penicillin and Cephalosporin Prescriptions* Reaction type Index penicillin N 3920 subjects N (% of reactions) Cephalosporin N 624 subjects N (% of reactions) Anaphylaxis 25 (0.64) 13 (2.08) 1 Angioedema 106 (2.70) 15 (2.40) 3 Laryngospasm 19 (0.48) 2 (0.32) 0 Urticaria 2904 (74.1) 461 (73.9) 30 Erythema multiforme 174 (4.44) 24 (3.85) 1 Toxic epidermal necrolysis 2 (0.05) 1 (0.16) 0 Dermatitis caused by an ingested drug 13 (0.33) 1 (0.16) 0 Adverse drug reaction attributed to an antibiotic 677 (17.3) 107 (17.2) 2 ALE allergic-like event. *Narrow definition: N Subjects with same ALE after both antibiotics
5 Apter et al Penicillin and Cephalosporin Allergy 354.e15 Table 3 Primary Results and Sensitivity Analyses Using Various Assumptions for the Risk of an Allergic-like Event after an Antibiotic* Primary analyses Penicillin followed by cephalosporin Penicillin followed by sulfonamide antibiotic Sensitivity analyses Including day of antibiotic prescription Limiting observation after penicillin to 1-10 d Limiting observation after cephalosporin to 1-10 d Limiting observation after both antibiotics to 1-10 d Removing patients who did not have 30 d of follow-up after cephalosporin prescription Removing patients with a penicillin prescription within 30 d of the cephalosporin prescription ALE after both penicillin and second antibiotic ALE allergic-like event; CI confidence interval. *Allergic-like events are narrowly defined. NO ALE after both penicillin and second antibiotic ALE after penicillin but not after second antibiotic ALE after second antibiotic but not after penicillin Unadjusted risk ratio (95% CI) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) experienced allergic-like events after both penicillin and cephalosporin had urticaria after both antibiotics. There was a substantial increase in risk of having an allergic-like event after cephalosporin in patients who had experienced an event after the earlier penicillin; 1.1% of the patients who had an event after penicillin also had an event after the cephalosporin, for an unadjusted RR of 10.0 (CI ) (Table 3). The unadjusted RR using the broad definition for events was 15.2 (CI ). Using the narrow definition, no potential confounders changed the OR of interest by 15% or more in the analysis of allergic-like events. For perspective, we also analyzed allergic-like events after prescriptions for a sulfonamide antibiotic rather than a cephalosporin (Table 3); 1.6% of the patients who had an event after penicillin also had an event after the sulfonamide. The unadjusted RR of an allergic-like event after a sulfonamide in patients who had an event after a preceding penicillin, compared with patients who had no such prior event, was 7.2 (CI ). The corresponding data using the broad codes were similar. Sensitivity Analyses The results of the sensitivity analysis are presented in Table 3. When the prescription day of the antibiotic is included or the time interval after the antibiotic prescription is varied, the RRs change numerically but are clinically similar to the primary unadjusted analyses. In summary, these analyses support our primary results. DISCUSSION We used a large electronic medical record database to measure the risk of an allergic-like event after a prescription for a cephalosporin in patients with a previous allergic-like event after penicillin. We found a substantially increased risk whether we used broad or narrow codes for allergic-like events. However, the risk was similar for those patients receiving a sulfonamide antibiotic after an earlier event following penicillin. These findings argue against specific immunologic cross-reactivity between penicillins and cephalosporins. Because there is no clinical gold standard for a drug hypersensitivity reaction, we conducted 2 analyses using both narrowly and broadly defined codes for allergic-like events. The narrowly defined set likely has fewer false-positive events but more false-negative events. The broader set has the opposite limitation: more false positives and fewer false negatives. Obtaining similar results from both analyses reduces the risk of an incorrect result because of misclassification of the allergic-like event. It is also noteworthy that both sets of codes are relevant to the clinician. That is, clinicians are often confronted with patients describing allergic events somewhat vaguely, reflecting the broader codes.
6 354.e16 The American Journal of Medicine, Vol 119, No 4, April 2006 The absolute risk of all events, after cephalosporin intake in patients who had had an event after penicillin, whether using broad or narrow datasets, was vanishingly small. The rate of anaphylaxis was less than 0.001% using either definition of allergic event. One patient experienced anaphylaxis after both antibiotics (Table 2). Thus, although there is a substantial relative increase in risk, the absolute risk suggests these events are rare. Recently, we performed a related study examining the risk of an allergic event after a subsequent antibiotic prescription in patients with an allergic-like event after the first antibiotic prescription. 24,25 A retrospective cohort of 969 patients was used to examine whether cross-reactivity existed between sulfonamide antibiotics and sulfonamide non-antibiotics (eg, furosemide, hydrochlorothiazide, glyburide). 24 We found an association between an allergic-like event after a sulfonamide antibiotic and an allergic-like event after a subsequent sulfonamide non-antibiotic. This OR (6.6 [CI ]) was comparable to that of an allergic-like event after a penicillin for patients who had had an earlier event following a sulfonamide antibiotic (unadjusted OR 7.8 [CI ]). As in our current study, these findings did not support allergic cross-reactivity of 2 medication classes with structural similarity. Thus, examining different antibiotic combinations, we have shown that patients who have 1 allergic-like event are at risk for a subsequent event. Allergic cross-reactivity does not explain these phenomena. Instead, another mechanism may better explain these events, such as genetic predisposition. This is a study of clinical events documented in the outpatient medical record, not of patient report of allergy. It is in the outpatient setting that patients are most likely to be labeled as having an antibiotic allergy. 35 Many studies of drug allergy focus on hospitalized patients, who may have a different reaction rate than outpatients. Also, our study attempts to consider all events consistent with a hypersensitivity reaction and is not limited to dermatologic reactions, although these are among the most identifiable. Thus, we are attempting to capture the more serious clinical events related to drug hypersensitivity occurring in the outpatient population. It is important to recognize that although these events may be coded as allergic, we have no way of verifying an immunologic etiology. What we describe are signs and symptoms as the practicing clinician perceives them and makes medical decisions. Surprisingly, a higher proportion of patients who experienced an allergic-like event after penicillin subsequently received cephalosporin, compared with patients who experienced no allergic-like event after a prior penicillin, suggesting that practitioners substituted cephalosporin for penicillin in patients with prior events. Nevertheless, we cannot exclude the possibility that patients who did not receive cephalosporin after the index penicillin originally had events more consistent with drug allergy. Our study has limitations. Misclassification of an allergic-like event is a possibility, but the similarity of results from both the narrow and broad analyses supports our result. A limitation is the difficulty in studying time from prescription to the clinical event consistent with a hypersensitivity reaction. Medical records note patient contacts with their physician; it is possible that events were not reported immediately and/or not diagnosed. Also, prescriptions may not have been filled, or filled promptly, or once filled, not taken. To validate our methods, we repeated analyses with a second time interval of 10 days, examining the effect of the shorter surveillance period after either, then both, the index and subsequent antibiotic prescriptions. The RRs were not substantively changed. It is possible that patients had medical events before entering or after leaving a GPRD-participating practice, before or after their physician s participation in GPRD was terminated, or after September Such events were not captured in this study. Undercounting may have occurred if patients had a very severe reaction to penicillin, such as toxic epidermal necrolysis or anaphylaxis, before entering the GPRD database and subsequently avoided future prescriptions. However, these events are so rare they are unlikely to influence our results. Finally, the generalizability of the GPRD to other countries is unclear, given differences in population demographics. In conclusion, for patients who had an allergic-like event after penicillin, the risk of an allergic-like event after a subsequent cephalosporin is markedly increased, but the risk is similarly elevated for those patients after a sulfonamide antibiotic. Thus, cross-reactivity of penicillin and cephalosporin does not fully explain this increased risk, and the risk of anaphylaxis and other very serious adverse events is very low. Thus, our data indicate that cephalosporins can be considered for patients with penicillin allergy. References 1. Saxon A, Beall GN, Rohr AS, Adelman DC. Immediate hypersensitivity reactions to beta-lactam antibiotics. Ann Intern Med. 1987;107: Kelkar PS, Li JT. Cephalosporin allergy. N Engl J Med. 2001;345: Greenberger PA. Drug Allergy, Part B: Allergic Reactions to individual drugs: low molecular weight. In: Grammer LC, Greenberger PA, eds. Patterson s Allergic Diseases, 6th Edition. 1 vol. Philadelphia: Lippincott, Williams & Wilkins; 2002: Puchner TC Jr., Zacharisen MC. A survey of antibiotic prescribing and knowledge of penicillin allergy. Ann Allergy Asthma Immunol. 2002; 88: Zhao Z, Baldo BA, Rimmer J. Beta-lactam allergenic determinants: fine structural recognition of a cross-reacting determinant on benzylpenicillin and cephalothin. Clin Exp Allergy. 2002;32: Dash CH. Penicillin allergy and the cephalosporins. J Antimicrob Chemother. 1975;1: Girard JP. Common antigenic determinants of penicillin G, ampicillin and the cephalosporins demonstrated in men. Int Arch Allergy Appl Immunol. 1968;33: Thoburn R, Johnson JE 3rd, Cluff LE. Studies on the epidemiology of adverse drug reactions. IV. The relationship of cephalothin and penicillin allergy. JAMA. 1966;198: Pumphrey RS, Davis S. Under-reporting of antibiotic anaphylaxis may put patients at risk. Lancet. 1999;353:
7 Apter et al Penicillin and Cephalosporin Allergy 354.e Anne S, Reisman RE. Risk of administering cephalosporin antibiotics to patients with histories of penicillin allergy. Ann Allergy Asthma Immunol. 1995;74: Macy E, Burchette RJ. Oral antibiotic adverse reactions after penicillin skin testing: multi-year follow-up. Allergy. 2002;57: Anon. Physicians Desk Reference, 57th Edition. Montvale, NJ: Medical Economics Company; Adkinson NF Jr., Essayan D, Gruchalla R et al. Task force report: future research needs for the prevention and management of immunemediated drug hypersensitivity reactions. J Allergy Clin Immunol. 2002;109:S461-S Anon. Statistics of N. The General Practice Research Database: Information for Researchers. London: GPRD, Office for National Statistics, Department of Health; EPIC. The General Practice Research Database: A Guide for Researchers. London: EPIC; Lawson D, Sherman V, Hollowell J. The General Practice Research Database. Q J Med. 1998;91: Lewis JD, Brensinger C, Bilker WB, Strom BL. Validity and completeness of the General Practice Research Database for studies of inflammatory bowel disease. Pharmacoepidemiol Drug Saf. 2002;11: Jick H, Jick SS, Derby LE. Validation of information recorded on general practitioner based computerised data resource in the United Kingdom. BMJ. 1991;302: Jick H, Terris BZ, Derby L, Jick S. Further validation of information recorded on a general practitioner based computerized data resource in the United Kingdom. Pharmacoepidemiol Drug Saf. 1992;1: McKeever TM, Lewis SA, Smith C, et al. Early exposure to infections and antibiotics and the incidence of allergic disease: a birth cohort study with the West Midlands General Practice Research Database. J Allergy Clin Immunol. 2002;109: Margolis DJ, Bilker W, Knauss J, Baumgarten M, Strom BL. The incidence and prevalence of pressure ulcers among elderly patients in general medical practice. Ann Epidemiol. 2002;12: Margolis DJ, Bilker W, Santanna J, Baumgarten M. Venous leg ulcer: incidence and prevalence in the elderly. J Am Acad Dermatol. 2002; 46: Bashford JN, Norwood J, Chapman SR. Why are patients prescribed proton pump inhibitors? Retrospective analysis of link between morbidity and prescribing in the General Practice Research Database. BMJ. 1998;317: Strom BL, Schinnar R, Apter AJ, et al. Absence of cross-reactivity between sulfonamide antibiotics and sulfonamide nonantibiotics. N Engl J Med. 2003;349: Apter AJ, Kinman JL, Bilker W, et al. Re-prescription of penicillin following allergic-like events. J Allergy Clin Immunol. 2004;113; van Staa TP, Wegman S, de Vries F, Leufkens B, Cooper C. Use of statins and risk of fractures. JAMA. 2001;285: Hennessy S, Strom BL. Statins and fracture risk. JAMA. 2001;285: British National Formulary 43. March London: British Medical Association, Royal Pharmaceutical Society of Great Britain; Gilbert D, Moellering RJ, Sande M. The Sanford Guide to Antimicrobial Therapy, 30th Edition. Hyde Park, VT: Antimicrobial Therapy Inc.; Rothman K, Greenland S. Modern Epidemiology. Philadelphia: Lippincott Williams and Wilkins; Breslow NE, Day NE. Statistical Methods in Cancer Research. Volume II-The Design and Analysis of Cohort Studies. Lyon: International Agency for Research on Cancer; Huber PJ. The Behavior of Maximum Likelihood Estimates Under Non-standard Conditions. Berkeley, CA: University of California; White H. A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrika. 1980;48: StataCorp. STATA Statistical Software: Release 7.0. College Station, TX: Stata Press; Huovinen P, Cars O. Control of antimicrobial resistance: time for action. The essentials of control are already well known. BMJ. 1998; 317: Lis Y, Mann R. The VAMP Research multi-purpose database in the U.K. J Clin Epidemiol. 1995;48: Gelfand JM, Margolis DJ, Dattani H. The UK General Practice Research Database. In: Strom BL, ed. Pharmacoepidemiology, 4th Edition. Chichester: John Wiley & Sons, Ltd; 2005:
8 354.e18 The American Journal of Medicine, Vol 119, No 4, April 2006 Appendix A Penicillins, Cephalosporins, and Sulfonamide Antibiotics Selected from British National Formulary Groups 28 BNF Penicillins BNF Sulphonamides and trimethoprim BNF Benzyl and phenoxymethyl penicillins Benethamine penicillin BNF Antimalarials Benzathine penicillin Dapsone Benzylpenicillin Sulphadoxine Penicillin Sulphonamide antibiotic Penicillin G Sulfphaguanidine Penicillin V BNF Sulphonamide and trimethoprim BNF Penicillinase-resistant penicillin Sulphaloxate Cloxacillin Sulphamoxole Flucloxacillin Sulphadiazine Methicillin sodium Cotrimoxazole Teocillin Phthalylsulphathiazole Succinylsulphathiazole BNF Broad-spectrum penicillins Sulfametopyrazine Amoxicillin Sulfasuxidine Amoxicillin clavulanic acid Sulfphacetamide Amoxicillin Sulfphadiazine Sulphadimethoxine BNF Anti-pseudomonal penicillins Sulphadimidine Ticarcillin Sulphafurazole Piperacillin Sulphaguandidine Sulphamerazine BNF Mecillinam Sulphamethazine Pivmecillinam Sulphamethizole Sulphamethoxypyridazine BNF Cephalosporins and Cephalomycins 29 Sulphaphenazole 1st generation Sulphapyridine Cefadroxil Sulphathiazole Cephalexin Sulphaurea Cefazolin Cephradine BNF Antileprotics Cephalexin Dapsone 2nd generation Cefaclor BNF Cefamandole Sulphanilamide nasal preparation Cefoxitin Cefprozil Cefuroxime 3rd generation Cefixime Cefodizime Cefotaxime Cefpodoxime Cefsulodin Ceftazidime Ceftibuten Ceftizoxime Ceftriaxone BNF British National Formulary. *BNF5.1.2 includes monobactams and carbapenems, but the numbers are small and their exclusion does not change the result.
9 Apter et al Penicillin and Cephalosporin Allergy 354.e19 Appendix B Diagnoses Included in the Definition of Hypersensitivity/Allergic Reaction* Narrow definition Allergic urticaria Anaphylactic shock Angioedema Dermatitis caused by drugs Erythema multiforme Laryngeal spasm Shock unspecified Shock without mention of trauma Stevens-Johnson Syndrome Toxic epidermal necrolysis Upper respiratory tract hypersensitivity reactions Urticaria Urticaria unspecified Selected codes for adverse drug reactions: Drug allergy Allergy drug by mouth Allergic drug reaction not specified (NOS) Hypersensitivity NOS Drug hypersensitivity NOS Adverse reaction to eye anti-infectives and other eye drugs Adverse reaction to anti-infectives and other enteric drugs Additional diagnoses in the broad definition Asthma Eczema Unspecified adverse effect of a drug, medicinal, and biologic substance NOS not otherwise specified. *Diagnoses are recorded using OXMIS (Oxford Medical Indexing System) codes that are cross-referenced to Read diagnostic codes. 36,37 Prescriptions are recorded using 2 drug code schemes (Multilex and Prescription Prescribing Authority of the National Health Service). 36,37 OXMIS and READ diagnostic codes are available on request.
10 354.e20 The American Journal of Medicine, Vol 119, No 4, April 2006 Appendix C Potential Confounding Variables Demographic variables: Age Sex Preexisting hypersensitivity/adverse reactions: See Appendix B Potential confounding drugs: Antihistamines Systemic and topical corticosteroids Other drugs used for asthma HIV human immunodeficiency virus. Potential confounding diagnoses: Allergic rhinitis Amyloid disease Arteritis Behçet disease Churg-Strauss disease Collagen vascular disease Erythromelalgia Giant cell arteritis Hay fever Mixed connective tissue disease Polyarteritis nodosa Polychondritis Polymyalgia rheumatica Raynaud s syndrome Rheumatoid arthritis Scleroderma Spondyloarthritis Systemic lupus erythematosus Takayasu s arteritis Vasculitis Wegener s granulomatosis Cytomegalovirus HIV infection Epstein-Barr virus infection Cellulitis Asthma Eczema