JOURNAL OF OCULAR PHARMACOLOGY AND THERAPEUTICS Volume 29, Number 5, 2013 ª Mary Ann Liebert, Inc. DOI: 10.1089/jop.2012.0123

Evaluation of Adverse Events in Self-Reported Sulfa-Allergic Patients Using Topical Carbonic Anhydrase Inhibitors

Guilherme B. Guedes,1 Abraar Karan,1 Hylton R. Mayer,2 and M. Bruce Shields1

Abstract

Purpose: To investigate whether a self-reported history of allergy to sulfa-based drugs is a predictor for sub- sequent adverse reactions to topical carbonic anhydrase inhibitors (CAIs). Methods: A retrospective case-controlled cohort study via chart review was performed on 1,287 patients with a diagnosis of . The outcome measure was the development of an adverse reaction (either ocular, systemic, or both) within at least 30 days after receipt of 1 of 4 classes of topical glaucoma : CAIs ( and ), analogues, beta-adrenergic blockers, and alpha2-adrenergic agonists. Results: Patients with a self-reported history of sulfa allergy had significantly more ocular adverse reactions after the initiation of any of the topical antiglaucoma medications when compared to those patients with no reported allergies. Patients with a self-reported sulfa allergy and patients who self-reported other, nonsulfa-related al- lergies had similar rates of adverse reactions to most of the topical medications. The patients reporting a sulfa allergy who used topical CAIs did not have more adverse reactions compared with patients who reported having other, nonsulfa-related allergies who used topical CAIs. Self-reported sulfa-allergic patients had similar rates of adverse reactions to topical CAIs compared with topical prostaglandin analogues. Conclusion: It may be safe to use a topical CAI in patients who report a history of a sulfa allergy. Patients with allergies of any kind may be more likely to develop allergic reactions to other, unrelated drug classes.

Introduction quite different (Fig. 1). The nonsulfonamide medications, including dorzolamide, brinzolamide, and , llergic reactions to are most com- are useful antiglaucoma agents, but traditional practice has Amonly triggered by sulfa-based antibiotics, specifically been to avoid these medications in patients who report a those which contain the functional group— sulfa allergy. Avoiding these glaucoma agents may contrib- 1,2 SO2NH2. Accordingly, patients who report sulfa allergy ute to the progression of glaucomatous vision loss, or hasten likely experienced it in response to a previous sulfonamide the decision for more risky options to control glaucoma, such .1 The presence, or possibility, of a sulfa allergy can as incisional surgery. influence a patient’s health care in 2 significant ways: A At issue is whether a history of allergy to a sulfa-based sulfa-allergic patient may have an allergic reaction to an- antibiotic is a predictor for a subsequent allergic reaction to a other sulfonamide-containing medication, or a patient re- structurally different compound containing a sulfa group, porting a history of sulfa allergy may receive suboptimal care such as the nonsulfonylarylamines. To answer this question, of certain conditions in an effort to avoid the use of a sulfa- we conducted a retrospective case-controlled cohort study to containing medication.3 investigate adverse events in glaucoma patients with self- Certain nonantibiotic, sulfa-containing medications have reported sulfa allergy who were using sulfa-containing or traditionally been avoided in patients who report a history of nonsulfa-containing antiglaucoma medications. As with the sulfa-based antibiotic allergies, even though the chemical typical clinical experience, this retrospective study did not composition of the antibiotic sulfonylarylamine sulfonamide involve formal allergy testing to confirm self-reported med- and nonantibiotic, nonsulfonylarylamine compounds are ication allergies, nor were formal allergy tests used to

1Department of Ophthalmology, Yale University, New Haven, Connecticut. 2Eye Doctors of Washington, Chevy Chase, Maryland.

456 EVALUATION OF SULFA-ALLERGIC PATIENTS USING CAIS 457

FIG. 1. The sulfonamide antibiotics have an arylamine group (the R-NH2, which is circled, with the R group being an aromatic ring), while the carbonic anhydrase inhibitors (CAIs) (nonantibiotics) do not contain an aromatic amine.

confirm the allergic nature of patients’ adverse reactions.4 prostaglandin analogues, alpha2-adrenergic agonists, and For this reason, this study measured adverse reactions, beta-adrenergic blockers. not necessarily true allergic reactions, to topical glaucoma medications in patients who self-reported a sulfa allergy. We Methods also examined adverse events in glaucoma patients with unconfirmed, self-reported allergies to other nonsulfa- A retrospective case-controlled cohort study via chart re- containing medications who were using topical antiglaucoma view was performed on 1,287 patients with a diagnosis of medications, such as carbonic anhydrase inhibitors (CAIs), glaucoma who were seen by 2 of the authors (M.B.S. and 458 GUEDES ET AL.

H.R.M.) during the year of 2009. The study was approved by condition that there was a 1-month gap between start of the the Yale University School of Human Investigation drug usage. The 4 study groups were divided by which 1 of Committee. All charts were collected and reviewed from the the 4 medication classes a patient was using; the 4 control archives of the Yale Eye Center. Demographic information, groups were decided on this criterion as well. prescription-drug information, indications for all new pre- Because ascertaining true allergic events in the clinical scriptions, surgeries, clinical events, diagnoses, and physi- setting is often difficult, and no formal allergy testing was cians’ notes were documented. Inclusion criteria for the performed in this retrospective study, the primary outcome source population were (1) a clinical diagnosis of primary measure was the development of an adverse reaction (either open-angle, pigment dispersion, or exfoliation glaucoma, or ocular, systemic, or both) within at least 30 days of initiation ocular hypertension in at least 1 eye, (2) treatment with the of a topical . Adverse reactions were topical CAIs dorzolamide or brinzolamide, or with prosta- defined in this study as reactions significant enough to cause glandin analogues, beta-adrenergic blockers, or alpha2- discontinuation of treatment. Adverse ocular reactions in- adrenergic agonists, (3) at least 1 follow-up visit after 4 cluded itching, eyelid swelling, intolerable conjunctival hy- weeks of therapy. peremia, chemosis, mucoid discharge, significant papillary Exclusion criteria included (1) patients who started more or follicular reaction, or eyelid dermatitis. Adverse systemic than 1 topical medication at the same time, (2) patients who reactions included urticaria, anaphylactic shock, erythema underwent surgery less than 4 weeks after the initiation of multiform, asthma, and eczema. Statistical analysis was topical medical therapy, (3) patients using oral sulfa-based carried out using 2-proportion chi square tests with a sig- medications such as acetazolamide or . nificance level of 0.05. Cross testing was conducted across The typical practice to elicit an allergy history in the au- classes of drugs and across allergy histories. thors’ clinic was for a technician, resident, or attending physician to ask an open-ended question, ‘‘Do you have any Results allergies to medications?’’ The vast majority of patients simply reported having a sulfa allergy or allergy to some The study group consisted of 94 self-reported sulfa-allergic other medication. No efforts were made to interpret, char- patients from the study’s source population. The control acterize, or quantify the nature of a patient’s self-reported group consisted of 205 randomly selected patients from the allergies that were recorded in the chart. No confirmatory or study’s source population with glaucoma who self-reported formal allergy testing was performed to verify a patient’s no medication allergies or nonsulfa medication allergies. The self-reported allergy history. Only a few patients identified a study and control groups were evenly matched by race, named medication, such as Bactrim or . No efforts gender, and age (Table 1). were made to identify or classify details regarding the class Patients with previous self-reported sulfa allergy had of sulfa medication patients reported. The typical clinical significantly more ocular adverse reactions after the initia- practice of M.B.S. and H.R.M. was to discuss the higher tion of any of the topical antiglaucoma medications when possible risk of allergic reactions to topical CAIs in patients compared to those patients with no self-reported allergies who self-reported a sulfa allergy, when initiating therapy. (Table 2). Patients who reported a history of a sulfa allergy The physician’s decision to prescribe CAIs to these patients and patients who reported other, nonsulfa-related allergies was based on the determination that the benefits outweighed had statistically similar rates of reactions to most of the the potential risks of other intraocular pressure-lowering topical medications. alternatives. The self-reported sulfa-allergic patients using topical CAIs The study group comprised of patients from the cohort did not have more adverse events than patients who re- who self-declared a sulfa allergy. The control groups were ported other, nonsulfa-related allergies, and did not have a patients from the cohort who self-declared either no medi- higher rate of adverse reactions than did self-reported sulfa- cation allergies, or allergies to nonsulfa medications. To allergic patients using prostaglandin analogues (Table 3). randomize the controls, the control group patients were Patients who reported a history of a sulfa allergy developed chosen if they had Yale medical record numbers that were significantly more adverse reactions using topical alpha2- one before or one after each study group patient’s record adrenergic agonists than self-reported sulfa-allergic patients number. This randomization was formed on the assumption using a topical CAI (P < 0.05) (Table 4). Patients who reported that the order in which patients arrived at the clinic and were a sulfa allergy had a significantly higher rate of ocular ad- assigned medical record numbers followed a random pat- verse reactions after starting topical alpha2-adrenergic ago- tern. Each topical glaucoma medication used by the patient nists compared to both the group of patients who reported a was counted only once in the analysis. If a patient used history of nonsulfa-related medication allergies and the more than 1 topical glaucoma medication, it was under the group who reported no history of medication allergies.

Table 1. Demographic Data by Self-Declared Allergy History

Sex Race

Allergy history Total number Male Female Age (years) Caucasian African American Asian Hispanic

S (only or + other 94 29 (30.8%) 65 (69.2%) 66.75 ( – 16.12) 74 (78.7%) 14 (14.9%) 1 (1.1%) 5 (5.3%) allergies) Other allergies 47 16 (34.1%) 31 (65.9%) 68.25 ( – 15.58) 39 (82.9%) 5 (10.7%) 0 (0%) 3 (6.4%) No allergy 158 61 (38.6%) 97 (61.4%) 63.75 ( – 18.54) 118 (74.8%) 25 (15.9%) 1 (0.6%) 14 (8.7%) EVALUATION OF SULFA-ALLERGIC PATIENTS USING CAIS 459

Table 2. Summary of Adverse Reactions to Topical Antiglaucoma Medications

Self-reported Total Patients with Patients with Rate of history patients no reaction adverse reactiona adverse reactions

CAI Sulfa allergies 28 19 9 0.32 Other allergies 10 5 5 0.50 No allergies 40 36 4 0.10 Sulfa allergies 29 23 6 0.21 Other allergies 17 13 4 0.24 No allergies 47 44 3 0.06 Alpha2-adrenergic agonist Sulfa allergies 15 7 8 0.53 Other allergies 9 8 1 0.11 No allergies 34 27 7 0.21 Beta-adrenergic blocker Sulfa Allergies 22 19 3 0.14 Other allergies 11 10 1 0.09 No allergies 37 37 0 0.0

aThis includes ocular and systemic adverse reactions. CAI, carbonic anhydrase inhibitor.

The self-reported sulfa-allergic study group had a significantly one or more nitrogen groups at the sulfonamido-N1 position, lower number of adverse reactions to topical beta-adrenergic chemical structures that are not present in CAIs.7,8 blockers when compared with self-reported sulfa-allergic A previous study by Lee et al. that evaluated self-reported patients using any other medication class (P < 0.05). sulfa-allergic patients who were using oral nonbacteriostatic The rates of systemic (nonocular) reactions to topical an- acetazolamide and/or found that tiglaucoma medications were not significantly different be- 7% of patients developed urticaria and no patients devel- tween the medication classes. The only systemic reactions oped a severe allergic reaction.8 Similarly, a study by Strom that patients experienced were rash, skin swelling, erythema et al. found that patients taking a nonantibiotic sulfonamide multiform, and severe vomiting. No patients experienced who had a self-reported sulfa allergy had equal rates of al- . lergic-type reactions when compared to those patients who self-reported allergies to any type of medication.3 Discussion Our retrospective study measured adverse events that were significant enough to necessitate discontinuation of There is little published biochemical or clinical evidence topical medical therapy. No efforts were made to differen- to support the assumption that prior allergic reactions to tiate the features of patients’ self-reported allergies, and no sulfonamide-based drugs increases susceptibility to allergic patients in our study had formal testing to determine if they reactions from all sulfa-based drugs.5 There are significant had true allergies to sulfa or other medication, or true allergic structural differences between sulfonamide antibiotics and reactions. In agreement with previous reports that evaluated other sulfonamide medications such as CAIs. While sulfon- oral medications (that also did not confirm true allergic re- amide antibiotics contain an arylamine (R-NH2 where R is actions), our study, evaluating patients using topical glau- an aromatic ring), the chemical structure that has been pro- coma medications, found that patients who self-reported a posed to be responsible for the sulfonamide hypersensitivity sulfa allergy had more adverse events than patients with no reactions, other sulfonamide-containing medications do not self-reported allergies. In addition, the self-reported sulfa- have this aromatic amine.6 In addition, sulfonamide antibi- allergic patients had equal amounts of adverse reactions otics contain a 5- or 6-member aromatic heterocyclic ring and to sulfa-based medications compared with patients who

Table 3. Comparing Rates of Adverse Reaction After Topical Antiglaucoma Medication Use by Self-Reported Allergy History

Drug Self-reported allergy history CAI Prostaglandin analogue Alpha2-adrenergic agonist Beta-adrenergic blocker

Prior sulfa allergya No significant No significant Significantly higher in No significant vs. other allergy difference (P > 0.05) difference (P > 0.05) prior sulfa group (P < 0.05) difference (P > 0.05) Prior sulfa allergy Significantly higher Significantly higher Significantly higher Significantly higher vs. no allergy in prior sulfa in prior sulfa in prior sulfa in prior sulfa group (P < 0.05) group (P < 0.05) group (P < 0.05) group (P < 0.05) Other allergy Significantly higher Significantly higher No significant No significant vs. no allergy in other allergy in other allergy difference (P > 0.05) difference (P > 0.05) group (P > 0.05) group (P = 0.05)b

aSome patients have sulfa allergy as well as other allergies. bP-value = 0.052. 460 GUEDES ET AL.

Table 4. Comparing Rate of Adverse Reactions Between Drug Classes in Self-Reported Sulfa Allergic Patients Using Topical Antiglaucoma Medications

Drug CAI Prostaglandin analogue Alpha2-adrenergic agonist Beta-adrenergic blocker

CAI No significant Significantly higher in No significant difference (P > 0.05) alpha agonist difference (P > 0.05) group (P < 0.05) Prostaglandin analogue Significantly higher in No significant alpha agonist difference (P > 0.05) group (P < 0.05) Alpha2-adrenergic agonist Significantly higher in alpha agonist group (P < 0.05) Beta-adrenergic blocker self-reported allergies to nonsulfa-based medications. Our sensitivity, and individual medication intolerance, which is study found that self-reported sulfa-allergic patients had the why we elected to record adverse events that were signifi- highest rate of local adverse reactions to alpha2-adrenergic cant enough to warrant cessation of therapy. Utilizing ad- agonists, and the lowest rate of local adverse reactions to verse events may result in an over-estimation of the actual topical beta-adrenergic blockers, similar to previous reports incidence of true allergic events. of adverse reactions to topical glaucoma medications in pa- In conclusion, we found no significant difference in the tients who report no medication allergies. rate of adverse events in self-reported sulfa-allergic patients There is uncertainty regarding the underlying cause for who were using topical sulfa-containing CAIs compared the results seen in this study, primarily the finding that pa- with self-reported sulfa-allergic patients who were using tients who reported a previous allergy to any medication topical prostaglandin analogues or beta-adrenergic blockers. were more likely to have an adverse event to topical medi- Patients who reported allergies to any kind of medication, cations compared with patients who reported no medication sulfa and nonsulfa containing, were more likely to develop allergies. It is possible that some patients are less tolerant of an adverse reaction to topical antiglaucoma medications the medication side effects, and more likely to voice dis- than patients who reported no allergies to medications. Fi- comfort or report adverse events. It is also possible that some nally, self-reported sulfa-allergic patients were most likely to patients may have an immunologic basis for hypersensitivity have an adverse reaction to topical alpha2-adrenergic ago- to various chemical compounds, including medication ex- nists, and were most tolerant of topical beta-adrenergic cipients, such as preservatives. It would be valuable to fur- blockers, compared with the other classes of medications. ther investigate the nature and significance of cross-reactivity The results of this study suggest that it may be reasonable events for true and self-reported medication allergies. to initiate therapy with a sulfa-containing topical anti- This study has a number of limitations. The retrospective glaucoma CAI in a patient who reports a sulfa allergy. It nature of the study makes it difficult to control for many would be prudent, however, for clinicians to investigate the variables; most notably, clinicians may have avoided pre- nature and severity of prior adverse reactions to sulfa med- scribing topical CAIs in patients who reported a history of ications. Clinicians should exercise caution before re- severe reactions to sulfa medications (e.g., Stevens Johnson’s commending the use of sulfa-containing CAIs in patients or anaphylaxis), only prescribing topical CAIs for sulfa- with sulfa allergies or any allergy history, inform patients of allergic patients who reported a remote or mild allergic reac- the low, but potentially significant possibility of adverse re- tion. There was a greater number of women than men in our actions, and follow patients carefully for the development of study, and a preponderance of Caucasians, but these differ- adverse reactions. ences were consistent between the various groups. We at- tempted to identify a large case-controlled cohort in an effort Disclosure of Funding to reduce random sampling errors and bias. It is possible that an allergic or adverse reaction could develop later than a Unrestricted departmental grant from Research to Prevent month after the initiation of therapy. The 30-day window Blindness. was selected because most hypersensitivity reactions to drugs occur within several weeks of administration and a Author Disclosure Statement shorter study period may better control confounders.5 It is possible that patients developed a reaction to the cumulative No competing financial interests exist. exposure to multiple or chronic medications, but this cu- mulative exposure experience should have been similar References across all the study groups. The self-reported nature of pa- 1. Macy, E., and Poon, K.-Y. Self-reported antibiotic allergy tients’ allergies is an inherent limitation, but it represents the incidence and prevalence: age and sex effects. Am. J. Med. limits of a typical personal medical history taken in a daily 122:778.e1–778.e7, 2009. clinical practice. It is likely that this cohort includes many 2. Johnson, K.K., Green, D.L., Rife, J.P., and Limon, L. Sulfona- patients who do not have a true sulfa allergy because con- mide cross-reactivity: fact or fiction? Ann. Pharmacother. firmatory or formal allergy testing was not done to ascertain 39:290–301, 2005. the verity of the self-declared sulfa allergy history. It is dif- 3. Strom, B.L., Schinnar, R., Apter, A.J., Margolis, D.J., Lau- ficult to distinguish between a true allergic reaction, chemical tenbach, E., Hennessy, S., Bilker, W.B., and Pettitt, D. Absence EVALUATION OF SULFA-ALLERGIC PATIENTS USING CAIS 461

of cross-reactivity between sulfonamide antibiotics and sul- 8. Knowles, S., Shapiro, L., and Shear, N.H. Should be fonamide nonantibiotics. N. Engl. J. Med. 349:1628–1635, 2003. contraindicated in patients who are allergic to sulfonamides? 4. Prescott, W.A., and Kusmierski, K.A. Clinical importance of Revisiting the meaning of ‘‘sulfa’’ allergy. Drug Safety 24:239– carbapenem hypersensitivity in patients with self-reported 247, 2001. and documented allergy. Pharmacotherapy 27:137– Received: June 13, 2012 142, 2007. 5. Stock, J.G. Sulfonamide hypersensitivity and acetazolamide. Accepted: September 25, 2012 Arch. Ophthalmol. 108:634–635, 1990. 6. Shepherd, G.M. Hypersensitivity reactions to drugs: evalua- Address correspondence to: tion and management. Mt. Sinai J. Med. 70:113–125, 2003. Dr. Hylton R. Mayer 7. Lee, A.G., Anderson, R., Kardon, R.H., and Wall, M. Eye Doctors of Washington Presumed ‘‘Sulfa Allergy’’ in patients with intracranial 2 Wisconsin Circle, Suite 200 hypertension treated with acetazolamide or furosemide: Chevy Chase, MD 20815 cross-reactivity, myth, or reality? Am. J. Ophthalmol. 138:114– 118, 2004. E-mail: [email protected] Copyright of Journal of Ocular Pharmacology & Therapeutics is the property of Mary Ann Liebert, Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.