Clinical history as a predictor of penicillin skin
Benjamin B. L. Wong, MD, FRCPC; Paul K. Keith, MD, FRCPC; and Susan Waserman, MD, FRCPC
Background: Up to 10% of the population reports an “allergy” to penicillin, whereas approximately 1.1% has positive
penicillin skin test results. Where penicillin skin tests are unavailable, some have advocated using history to decide whether to
use a penicillin-related antibiotic.
Objective: To determine if clinical history predicts penicillin skin test results.
Method: Retrospective medical record review of 94 consecutive patients who had previously taken penicillin referred for
penicillin allergy. Case histories were taken, penicillin skin tests performed, and an oral challenge recommended if skin test
results were negative.
Results: Of 91 cases studied, the average patient age was 27 years (range, 6 months to 82 years; 36% female). Fifty-two (57%)
experienced hives as their main adverse reaction. Sixteen (18%) had at least 1 positive test result. Of this group, 9 had hives as
their main symptom, whereas 1 had respiratory problems and 1 had angioedema. Most patients with positive skin test results had
experienced their reaction at least 3 years ago. Regression analysis showed that age, sex, and clinical history, including type of
reaction, time of reaction after penicillin ingestion, or time since the last reaction, were not associated with skin test positivity.
Seventy-two (96%) of the 75 patients who had negative skin test results underwent oral challenge. Seventy had negative
challenge results. The negative predictive value of a negative penicillin skin test result was 97%.
Conclusion: Clinical history was not predictive of subsequent penicillin skin test results.
Ann Allergy Asthma Immunol. 2006;97:169–174.
Penicillin allergy is a common problem in clinical practice.
Approximately 10% of the population reports an “allergy” to
penicillin,1,2yet there may be an up to 10-fold discrepancy
between a history of self-reported adverse reaction(s) and the
actual incidence of proven IgE-mediated allergy, which is
estimated to be 1.1%.1,3–5Penicillin skin testing, when it
includes a minor determinant mix, is useful because of its
high negative predictive value.6Whether clinical history can
be used to reliably predict penicillin skin test results is not
well studied and has been the subject of some controversy.7,8
Penicillin skin testing may not be readily available in many
practice settings. History may then become an important and
sometimes deciding factor in choosing a penicillin-based
antibiotic therapy.9Some have suggested that in low-risk
groups, a detailed history of the patient’s reaction to penicil-
lin may allow the clinician to exclude true penicillin allergy,
thus enabling this group to receive penicillin.10Others dis-
agree, stating that the low incidence of anaphylaxis to peni-
cillin, rather than the predictive value of clinical history, is
the reason patients tolerate penicillin safely.11Up to a third of
patients with a history suggestive of a non–IgE-mediated
reaction may have positive penicillin skin test results. This
implies that those with both convincing and less convincing
histories of penicillin allergy should undergo penicillin skin
testing before penicillin administration.12
It is our experience that most patients referred for evalua-
tion of penicillin allergy have negative skin test results and
tolerate penicillin on oral challenge. We have however also
noted that some individuals with positive skin test results
have a history that is not suggestive of immediate hypersen-
sitivity to penicillin. The objectives of this study are 3-fold:
(1) to survey the prevalence of penicillin skin test positivity
in patients referred to our Adverse Reactions Clinic for pen-
icillin allergy; (2) to characterize the demographics, preva-
lence, and distribution of symptoms in the cohort of patients
referred for penicillin allergy evaluation; and (3) to determine
if specific features of clinical history can be used to predict
penicillin skin test outcome. We also describe the results of
an oral challenge with penicillin in those patients who had
negative penicillin skin test results.
A retrospective medical record review of 94 consecutive
patients referred to the Adverse Reactions Clinic for penicil-
lin allergy was conducted during an 8-month period between
August 2000 and April 2001. The Adverse Reactions Clinic,
an outpatient regional tertiary referral center, is located at the
Firestone Institute of Respiratory Health, St. Joseph’s Hos-
pital, a McMaster University teaching hospital in Hamilton,
Ontario. Patients who did not have a previous adverse reac-
tion to penicillin (eg, those referred because of a positive
family history of penicillin allergy) and those with no known
Division of Allergy and Clinical Immunology, Department of Medicine,
McMaster University, Ontario, Canada.
Dr Wong is currently with Division of Allergy and Clinical Immunology,
Department of Medicine, North York General Hospital, University of
Received for publication January 29, 2006.
Accepted for publication in revised form February 8, 2006.
VOLUME 97, AUGUST, 2006169
exposure to penicillin were excluded. Patients who did not
successfully complete skin testing to the battery of reagents
listed herein were also excluded from the study. All patients
had a reaction to either oral penicillin or amoxicillin in the
past and were referred for penicillin allergy testing. All were
identified by a computer database, and information was ob-
tained and recorded on a standardized template. Statistical
analysis was performed using tools from Microsoft Excel
2000 (Microsoft Inc, Redmond, WA). Because of the retro-
spective and noninterventional nature of the study, ethics
board approval was not required.
All patients referred for penicillin allergy underwent
skin prick and intracutaneous tests to saline buffer (0.9%
sodium chloride diluent; Abbott Laboratories Inc, Missis-
sauga, Ontario) and to the following 4 reagents: (1) benzyl
penicillin (1,600 U/mL or 1.0 mg/mL at a concentration of
3.0 ? 10–3M Novopharm lot 03051352, Novopharm Lim-
ited, Stouffville, Ontario); (2) minor determinant mixture
(MDM) (1,600 U/mL, see below for the method of prep-
aration); (3) ampicillin (1.0 mg/mL or 2.0 ? 10–3M;
Novopharm lot 03051351); and (4) the major determinant
or penicilloyl polylysine (concentration 6.0 ? 10–3M;
Hollister-Stier Laboratories lot 31370, item 8365 ZA, Hol-
lister-Stier Laboratories LLC, Spokane, WA) (commer-
cially available previously as Prepen). Histamine (10 mg/
mL; Western Allergy Services/Quorum Pharmaceuticals,
Mississauga, Ontario) was used as a positive control.
The MDM stock solution was prepared by adding 3.1 mL
of sterile water without preservative to 1 million units of
penicillin G. One milliliter of this was transferred to a 5-mL
vial and 0.5 mL of sodium chloride was added. Then 0.1 N of
sodium hydroxide was added using a Millipore filter. The
mixture was incubated for 45 minutes at room temperature
followed by 48 hours at 4°C and then refrigerated. The
shelf-life for this stock solution was 28 days. The MDM was
prepared fresh before testing by the hospital pharmacy by
diluting 0.16 mL of stock MDM with 9.84 mL of 0.9%
sodium chloride. It was discarded after the clinic. This pro-
cess produces a solution that contains an adequate source of
Skin testing was performed by 2 experienced nurse clini-
cians at the Adverse Reactions Clinic. Histamine (10 mg/mL)
was used as a positive control. Saline was used as a negative
control. A positive skin prick and/or intracutaneous test result
was defined as a wheal 3 mm greater than the average
perpendicular diameter of the negative control. Patients who
had negative skin test results also underwent a 7-day graded
oral challenge at home with amoxicillin, taking a gradually
increasing daily dose until a 250-mg capsule was taken on the
seventh day, as confirmation that they are not allergic to
penicillin. Patients then contacted the clinic to report the
results of the oral challenge. For the purposes of this study, an
adverse reaction during the oral challenge was considered a
failed oral challenge.
A total of 94 patients were identified. Three patients were
excluded, because they were referred because of a family
history of penicillin allergy and did not have a personal
history of a penicillin reaction. All remaining 91 patients
completed penicillin skin tests. Of the 91 patients, a total of
16 patients (18%) were skin test positive, either on penicillin
skin prick test or intracutaneously, to 1 or more penicillin
skin test reagents. None of the 91 patients experienced any
adverse reactions to skin testing.
The average age of our cohort of patients was 27 years (SD,
22 years; range, 6 months to 82 years). Thirty-three (36%) of
the 91 patients were female. The breakdown of age and sex
among those who had positive and negative skin test results
is given in Table 1. There were no statistically significant
differences in age and sex between the groups.
Distribution of Positive Skin Test Results to Various
Sixteen patients tested positively to 1 or more skin test
reagents. Twelve patients tested positively to 1 reagent only.
(Seven patients tested positively to penicilloyl polylysine, 4
tested positively to penicillin G, and 1 tested positively to
MDM.) Three patients had 2 positive test results (to ampicil-
lin and penicilloyl polylysine), and 1 patient had 4 positive
test results (penicillin G, penicilloyl polylysine, MDM, and
ampicillin) (Fig 1).
Adverse Reaction Associated With Previous ?-Lactam
Fifty-two patients (57%) experienced an urticarial-type erup-
tion, whereas 36 (40%) experienced a nonurticarial maculo-
papular eruption. Three (3%) complained of an exfoliating,
scaly rash. None of the patients had Stevens-Johnson syn-
drome or erythema multiforme. Eight patients (9%) had
shortness of breath and wheezing, 12 (13%) experienced
angioedema of the tongue and face, and 2 patients (2%)
experienced presyncope and had a decreased level of con-
sciousness. None had cardiac symptoms, whereas 8 (9%)
Table 1. Age Distribution Among Those Who Have Negative and
Positive Penicillin Skin Tests Results
No. (%) of negative
No. (%) of positive
170 ANNALS OF ALLERGY, ASTHMA & IMMUNOLOGY
complained of gastrointestinal symptoms such as nausea and
diarrhea. The distribution of symptoms between those who
had positive and negative skin test results is shown in Figure
2. Of the 16 patients who had positive skin test results, 9
patients (56%) had urticaria, 1 (6%) had angioedema, and 1
(6%) had respiratory symptoms. Four patients (25%) had a
maculopapular-type skin eruption. Only 4 patients (25%) had
their reactions within 24 hours of drug ingestion. None had an
immediate acute reaction (ie, in less than 1 hour) to penicillin.
Time of Onset of Reaction
The timing of the reaction was categorized as (1) acute,
defined as a reaction taking place within 6 hours of admin-
istration of the antibiotic; (2) accelerated, within 6 to 24
hours; (3) subacute, within 24 to 72 hours; and (4) delayed,
after 72 hours. Six patients (7%) had an acute reaction, 5
(5%) had an accelerated reaction, 3 (3%) had a subacute
reaction, and 10 (11%) had a delayed reaction. Most patients
(67 or 74%) could not recall exactly when their reaction
occurred after ingestion of penicillin. Some had been told
they had reactions as infants and could not recall the nature of
Time Since Last Reaction to Penicillin
Six patients (7%) recalled having a reaction less than 3
months before their clinic visit, whereas more than half of all
patients had a reaction to penicillin more than 3 years before
their clinic visit. Approximately 15% of patients could not
recollect when the reaction took place and could not be
categorized (Fig 3).
Reason for Penicillin Prescription
A large proportion of patients could not recall the nature or
site of the infection for which penicillin was prescribed (Fig
4). Approximately 20% of the patients recalled an otitis
media, and a remaining 25% had a respiratory infection.
Associated Features of Penicillin Allergy in Patients Who
Had Positive Skin Test Results
Of those 16 patients who had positive skin test results to
penicillin, Four patients (25%) had other non–penicillin-
based drug allergies, 5 patients (31%) had no other drug
allergies but evidence of atopy, and 2 patients (13%) had a
family history of penicillin allergy. The corresponding fig-
ures for those 75 patients who had negative penicillin skin
test results were 13 patients (17%), 24 patients (32%), and 13
patients (17%). There was no statistical difference between
the 2 groups.
Negative Predictive Value of a Penicillin Skin Test
Seventy-two (96%) of the 75 patients who had negative skin
test results proceeded to do an oral ingestion challenge with
amoxicillin. The other 3 patients were lost to follow-up. Of
the 72 patients who completed the oral challenge, 2 adult
patients had mild generalized urticaria on day 6 of the oral
challenge protocol (Table 2), after taking approximately one
Figure 2. Percentage of patients with their corresponding adverse symp-
toms, stratified according to skin test results. There were no significant
differences in the profile and types of adverse reactions between skin
test–positive and skin test–negative individuals. GI indicates gastrointestinal.
Figure 1. Distribution of positive skin test results. Sixteen patients had a
total of 22 positive skin test results, because some patients had more than 1
positive result. Prepen is the commercial name for penicilloyl polylysine;
Pen G indicates penicillin G.
Figure 3. Percentage of patients with their corresponding time elapsed
since last reaction, stratified according to skin test results. There were no
significant differences in the temporal profile of the most recent adverse
reaction between skin test–positive and skin test–negative individuals.
VOLUME 97, AUGUST, 2006 171
third of the 250-mg tablet (approximately 83 mg). Both of
these patients tested negatively to penicilloyl polylysine.
They were given the option of returning for additional skin
testing to rule out that the urticaria was due to an intercurrent
illness or some other cause; both deferred. It is also possible
that these reactions were lymphocyte dependent, which
would not be predicted by skin testing. They were told to
avoid penicillin-based antibiotics in the future. Based on this
cohort of 70 patients who had negative challenge results, the
negative predictive value of a negative penicillin skin test
result was 97%.
Features Predictive of Penicillin Allergy
Univariate and multivariate regression analyses were per-
formed to determine if features of clinical history could be
used to predict penicillin skin test results. None of the vari-
ables, age, sex, characteristics of history (hives, nature of
rash, respiratory symptoms, angioedema, syncope, gastroin-
testinal symptoms), time of onset of reaction, time elapsed
since the last reaction, adverse reactions associated with other
drug use, family history, or atopy, were found to be predictive
of penicillin skin test outcome.
Penicillin allergy is a common problem encountered in clin-
ical practice. Short of an oral challenge, which can incur
substantial risk, penicillin skin testing is the gold standard for
distinguishing between patients who have current IgE-medi-
ated allergy and those with non–IgE-mediated allergic reac-
tions to penicillin. In our clinical experience, some patients
have expressed concerns about the safety of penicillin skin
testing. Valyasevi and Van Dellen14reported that the sys-
temic reaction rate for all patients tested to penicillin reagents
was 0.12%; those who tested positive to penicillin had a 2.3%
reaction rate, with no fatalities. In this study, none of the 91
patients who underwent skin testing had an adverse reaction,
consistent with the notion that penicillin skin testing is gen-
erally a safe procedure.
Of the 16 patients who had a total of 22 positive penicillin
skin test results, 11 (69%) also tested positively to penicilloyl
polylysine. Our results are consistent with those of Gadde et
al,15who found that roughly 75% of patients who react on
penicillin skin tests react to penicilloyl polylysine. This arti-
cle thus provides corroborating evidence that of all the pen-
icillin reagents available, penicilloyl polylysine remains the
most sensitive reagent for penicillin skin testing. Only 1
patient, who had a history of urticaria, tested positive to
MDM alone. Skin testing without the use of a MDM would
have missed this patient.
The study demonstrated that clinical history, including
time elapsed since the last reaction, was not predictive of
subsequent skin test outcome. This finding is important in a
number of ways. First, it has been shown previously that
patients lose sensitivity to penicillin with time. Approxi-
mately 78% of penicillin allergic patients have negative skin
test results after 10 years of penicillin avoidance.16This
implies that more than 20% of patients still retain positive
skin test results after a decade. The point emphasized in this
study was that a remote but relatively recent history (defined
with a cutoff of roughly 3 years) of a penicillin reaction does
not significantly decrease the chance of a positive skin test
result, a surrogate marker for an IgE-mediated reaction to
penicillin. Positive penicillin skin test results decrease by
roughly 10% annually after a penicillin reaction,17and no
data are available to suggest that patients lose their sensitivity
more rapidly with time after 5 years. Because a number of
patients had atypical histories in this study, further break-
down of the group of patients who had their reaction more
than 3 years ago was not feasible. Based on this study, it
seems reasonable that a patient with a remote but relatively
recent history of penicillin allergy should not be casually
dismissed as someone who has “likely to have lost sensitivity
Second, some have suggested that in a low-risk group,
certain elements of the patients’ history, such as a maculo-
papular rash later in the course of taking penicillin, may allow
clinicians to exclude true penicillin allergy, making skin
testing unnecessary.10This study demonstrated that no spe-
cific features in the clinical history can be used to reliably
predict skin test outcome. Yet, in a recent survey of internists,
pediatricians, and allergists, the choice of whether to recom-
mend penicillin skin testing is in large part influenced by a
Table 2. McMaster Amoxicillin Oral Challenge Protocol
Day Ingestion challenge
1 Consume 1 tsp of water out of a 250-mL glass of water
containing roughly 1/100 (2.5 mg) of amoxicillin
Consume 6 tsp of water out of a 250-mL glass of water
containing roughly 1/100 (2.5 mg) of amoxicillin
Consume roughly 1/100 (2.5 mg) of an amoxicillin 250-mg
Consume roughly 1/30 of an amoxicillin 250-mg capsule
Consume roughly 1/10 of an amoxicillin 250-mg capsule
Consume roughly 1/3 of an amoxicillin 250-mg capsule
Consume one whole 250-mg amoxicillin capsule
Figure 4. Percentage of patients with their corresponding site of infection,
stratified according to skin test results. There were no significant differences
in the location between skin test–positive and skin test–negative individuals.
The category “Other” includes gastrointestinal, genitourinary, and central
nervous system infections.
172 ANNALS OF ALLERGY, ASTHMA & IMMUNOLOGY
patient’s clinical history. Puchner and Zacharisen18demon-
strated that only 2% to 8% of respondents recommended
penicillin skin testing when the primary adverse reaction was
a maculopapular rash, whereas 21% would recommend skin
testing if the same patient had a history of a systemic allergic
reaction. Solensky et al12demonstrated that up to a third of
patients who have positive penicillin skin test results have a
“vague” history that suggests a non–IgE-mediated reaction.
Therefore, Primeau and Adkinson19have recommended that
even patients with a low-risk history, such as maculopapular
rash, should undergo penicillin skin testing before penicillin
administration, because these patients may still have positive
penicillin skin test results. This study supports the recom-
mendation of Primeau and Adkinson, because 25% of the
patients with positive skin test results had a vague history of
a nonurticarial maculopapular “rash” (Fig 2).
Recently, there has been increased concern about inappro-
priate antibiotic use and the rising incidence of multiantibi-
otic-resistant organisms. This concern was highlighted in 2
recent studies. In the first study, Puchner and Zacharisen18
asked what physicians would do when faced with an other-
wise healthy ambulatory patient with sinusitis or bronchitis
and with a history of maculopapular rash from penicillin
within the last 10 years. Sixty-one percent of respondents
chose an alternative (nonpenicillin, noncephalosporin) anti-
biotic instead of referring the patient to an allergist for pen-
icillin skin testing. Of these respondents, 53% chose a mac-
rolide antibiotic. This is concerning, because macrolides may
not provide effective first-line coverage for the common
pathogens that cause sinusitis.20Since up to 90% of patients
who claim to be penicillin allergic may have negative skin
test results, false labeling of patients occurs, compounding
this problem. In the second study, Arroliga et al3demon-
strated that in the intensive care unit (ICU), 81.5% of patients
who had negative skin test results to penicillin and who were
receiving antibiotics for therapeutic reasons were subse-
quently switched to a ?-lactam–based antimicrobial. The
study concluded that “penicillin skin testing is a safe, reliable,
and effective strategy to reduce the use of non–?-lactam
antimicrobials in patients who are labeled penicillin allergic
and admitted to the ICU.” Therefore, the introduction of a
practice guideline for penicillin skin testing may increase the
appropriate use of penicillin skin testing and reduce antibiotic
costs for patients with a history of penicillin allergy and
infections with penicillin- or penicillin-related susceptible
pathogens.21This may also reduce the incidence of prophy-
lactic vancomycin in patients suspected of having an allergy
to penicillin or cephalosporins who are undergoing elective
orthopedic surgery22and may help prevent the emergence of
This study has a number of limitations. One limitation is
the relatively small sample size. Another limitation is the
retrospective nature of the study. There may be recall bias
among patients, because many had experienced their reac-
tions more than a few years earlier. There was also a rela-
tively high prevalence of “vague” symptoms. The somewhat
limited and unsystematic documentation of history and symp-
toms on patients’ records may potentially limit the power of
statistical analysis. Finally, some have suggested that the use
of a positive skin test result as a surrogate marker for true
penicillin allergy may overestimate the true incidence of
penicillin allergy due to false-positive results. However, an
oral challenge of all patients with positive skin test results is
not ethically feasible. The use of a penicillin skin test remains
the “closest test to a criterion standard for allergy.”7
In conclusion, 18% of patients who presented to the Ad-
verse Reactions Clinic had positive penicillin skin test results.
The most common adverse reaction to penicillin was gener-
Table 3. Characteristics of Patients With Positive Penicillin Skin Test Results
Time elapsed since
Positive skin test results
Within 24 hours
Within 24 hours
After 7 days
After 7 days
Vague, maybe within 24 hours
Within 24 hours
After 7 days
Within 24 hours
Abbreviations: MDM, minor determinant; MP, nonurticarial maculopapular rash; Prepen, penicilloyl polylysine or the major determinant; Resp,
respiratory symptoms; Vague, vague history.
VOLUME 97, AUGUST, 2006173
alized urticaria. Of the reagents used in skin testing, the major
determinant (penicilloyl polylysine or Prepen) was most com-
monly positive. The negative predictive value of a negative
penicillin skin test result was 97%. Demographics and fea-
tures of history were similar between patient groups with
positive and negative skin test results; regression analysis did
not demonstrate a relationship between history and skin test
results. In our study, clinical history was not a predictor of
penicillin skin test outcome.
We would like to thank the Adverse Reactions Clinic at the
Firestone Institute of Respiratory Health, Ontario, Canada,
and our nurses Ann Bartlett, Marilyn Dawson, Mary Pierce,
and Lynn Darlington for their participation in this study.
1. Kerr JR. Penicillin allergy: a study of incidence as reported by
patients. Br J Clin Pract. 1994;48:5–7.
2. Solensky, R; Earl HS, Gruchalla RS. Lack of penicillin resen-
sitization in patients with a history of penicillin allergy after
receiving repeated penicillin courses. Arch Intern Med. 2002;
3. Arroliga ME, Radojicic C, Gordon SM, et al. A prospective
observational study of the effect of penicillin skin testing on
antibiotic use in the intensive care unit. Infect Control Hosp
4. Arroliga ME, Wagner W, Bobek MB, et al. A pilot study of
penicillin skin testing in patients with a history of penicillin
allergy admitted to a medical ICU. Chest. 2000;118:
5. Langley JM, Halperin SA, Bortolussi R. History of penicillin
allergy and referral for skin testing: evaluation of a pediatric
penicillin allergy testing program. Clin Invest Med. 2002;25:
6. Macy E, Burchette RJ. Oral antibiotic adverse reactions after
penicillin skin testing: multi-year follow-up. Allergy. 2002;57:
7. Macy E. How predictive is a history of penicillin allergy?
8. Whitmore SE. How predictive is a history of penicillin allergy?
9. Solensky R, Earl HS, Gruchalla RS. Clinical approach to pen-
icillin-allergic patients: a survey. Ann Allergy Asthma Immunol.
10. Salkind AR, Cuddy PG, Foxworth JW. Is this patient allergic to
penicillin? an evidence-based analysis of the likelihood of pen-
icillin allergy. JAMA. 2001;285:2498–2505.
11. Macy E, Richter PK, Falkoff R, et al. Skin testing with Penicilloate
and Penilloate prepared by an improved method: amoxicillin oral
challenge in patients with negative skin test responses to penicillin
reagents. J Allergy Clin Immunol. 1997;100:586–591.
12. Solensky R, Earl HS, Gruchalla RS. Penicillin allergy: preva-
lence of vague history in skin test-positive patients. Ann Allergy
Asthma Immunol. 2000;85:195–199.
13. Ressler C, Neag PM, Mendelson LM. A liquid chromatographic
study of stability of the minor determinants of penicillin allergy:
a stable minor determinant mixture skin test preparation.
J Pharm Sci. 1985;74:448–454.
14. Valyasevi MA, Van Dellen RG. Frequency of systematic reac-
tions to penicillin skin tests. Ann Allergy Asthma Immunol.
15. Gadde J, Spence M, Wheeler B, et al. Clinical experience with
penicillin skin testing in a large inner-city STD clinic. JAMA.
16. Sullivan TJ, Wedner MJ, Schatz GS, et al. Skin testing to detect
penicillin allergy. J Allergy Clin Immunol. 1981;68:171–180.
17. Blanca M, Torres MJ, Garcia JJ, et al. Natural evolution of skin
test sensitivity in patients allergic to beta-lactam antibiotics. J
Allergy Clin Immunol. 1999;103:918–924.
18. Puchner TC, Zacharisen MC. A survey of antibiotic prescribing
and knowledge of penicillin allergy. Ann Allergy Asthma Im-
19. Primeau MN, Adkinson NF. Penicillin allergy: clinical quanda-
ries. Ann Allergy Asthma Immunol. 2002;88:1–3.
20. Slavin RG. Nasal polyps and sinusitis. In: Middleton E Jr, Reed
CE, Ellis EF, et al, eds. Allergy: Principles and Practice. 5th ed.
St Louis, MO: CV Mosby; 1998:1032.
21. Forrest DM, Schellenberg RR, Thien VV, et al. Introduction of
a practice guideline for Penicillin skin testing improves the
appropriateness of antibiotic therapy. Clin Infect Dis. 2001;32:
22. Li JT, Markus PJ, Osmon DR, et al. Reduction of vancomycin
use in orthopedic patients with a history of antibiotic allergy.
Mayo Clin Proc. 2000;75:902–906.
23. Lee CE, Zembower TR, Fotis MA, et al. The incidence of
antimicrobial allergies in hospitalized patients: implications re-
garding prescribing patterns and emerging bacterial resistance.
Arch Intern Med. 2000;160:2819–2822.
Requests for reprints should be addressed to:
Benjamin Wong, MD, FRCPC
Department of Medicine
North York General Hospital
4001 Leslie St
Toronto, Ontario, Canada M2K 1E1
174ANNALS OF ALLERGY, ASTHMA & IMMUNOLOGY