Abstract
Background:
The prevalence of drug allergies documented in electronic health records (EHRs) of large patient populations is understudied.
Objective:
We aimed to describe the prevalence of common drug allergies and patient characteristics documented in EHRs of a large healthcare network over the last two decades.
Methods:
Drug allergy data were obtained from EHRs of patients who visited two large tertiary care hospitals in Boston from 1990 to 2013. The prevalence of each drug and drug class was calculated and compared by sex and race/ethnicity. The number of allergies per patient was calculated and the frequency of patients having 1, 2, 3. . ., or 10+ drug allergies was reported. We also conducted a trend analysis by comparing the proportion of each allergy to the total number of drug allergies over time.
Results:
Among 1 766 328 patients, 35.5% of patients had at least one reported drug allergy with an average of 1.95 drug allergies per patient. The most commonly reported drug allergies in this population were to penicillins (12.8%), sulfonamide antibiotics (7.4%), opiates (6.8%), and nonsteroidal anti-inflammatory drugs (NSAIDs) (3.5%). The relative proportion of allergies to angiotensin-converting enzyme (ACE) inhibitors and HMG CoA reductase inhibitors (statins) have more than doubled since early 2000s. Drug allergies were most prevalent among females and white patients except for NSAIDs, ACE inhibitors, and thiazide diuretics, which were more prevalent in black patients.
Conclusion:
Females and white patients may be more likely to experience a reaction from common medications. An increase in reported allergies to ACE inhibitors and statins is noteworthy.
Keywords: adverse drug event, drug allergy, drug hypersensitivity, electronic health records, safety
Adverse drug events (ADEs) are common in clinical practice, affecting 15–25% of patients (1). Most allergy-related ADEs are unpredictable given what is known today and are immunologically mediated (2). Allergic drug reactions can range from mild (e.g. itching) to life-threatening (e.g. anaphylaxis). They account for about 5–10% of all ADEs (3).
Most electronic health record (EHR) systems have an ‘allergy’ module (or list), where clinicians can document a patient’s ‘allergies’ reported by the patient or observed by the care provider. Although the ‘allergy’ module often contains other types of ADEs, such as drug intolerance, drug toxicity, drug idiosyncrasy, and pseudoallergic reactions, it remains a rich and valuable data source for studying allergies. However, the use of large data sets from EHRs to describe trends across a broad spectrum of reported drug allergies is uncommon. Most previous epidemiological studies of drug allergies either relied on self-reported surveys or focused on a particular drug or drug class (4, 5). The most commonly studied drug allergens include antibiotics, such as penicillins, cephalosporins, and sulfonamides, and nonsteroidal anti-inflammatory drugs (NSAIDs) like aspirin (2, 6).
In addition, demographics and clinical characteristics of patients across all types of drug allergies have not been well documented in the USA. Few studies have explored the relationship between racial groups or sex on documented allergy rates. Also, only a few studies have described the characteristics of patients with drug allergies in a given large population (7, 8). For example, while Macy and Ho studied 478 283 patients with at least one allergy in EHRs, other studies included smaller populations (5, 7) or described a population with documented anaphylaxis only (9–12). Further, the proportional changes of different drug allergies over time have not been well studied.
In this study, we aimed to describe the prevalence of various drug allergies documented in EHRs in a large patient population across a large healthcare delivery network over the past two decades. In particular, we described and compared each drug allergy by sex and race/ethnicity. We also conducted a trend analysis by comparing the percentages of commonly reported drug allergies by year.
Methods
Settings and data collection
Drug allergy data were collected from the Partners Health-Care System, which is an integrated healthcare delivery network in the Greater Boston area funded by the Brigham and Women’s Hospital (BWH) and the Massachusetts General Hospital (MGH) and is composed by multiple community and specialty hospitals as well as community health centers. At Partners, patients’ ‘allergy’ information is documented by healthcare providers via the allergy module of EHR systems used at each affiliated institution and then integrated and stored in the Partners’ Enterprise-wide Allergy Repository (PEAR) (13). Each patient has a longitudinal, common allergy record in their EHRs shared across the entire healthcare network. Healthcare providers are also able to deactivate (remove) an allergy from a patient’s allergy list if the patient did not experience an adverse reaction during allergy work-up or accidental re-exposure. Our allergy data are similar to previous studies that also used allergy entries in EHRs, in that many ‘allergies’ documented in patient’s EHR allergy list are not immunologically mediated (7, 8) or confirmed by a diagnostic test; we therefore use the term ‘reported allergies’ in this study.
In this study, patients who visited BWH and/or MGH between 1990 and 2013 were included and their drug allergy information (also including unknown allergies and no-known allergies) was extracted from PEAR. A patient’s drug allergy information includes a list of ‘allergens’ (i.e. culprit drug), allergy status (active or deactivated), date/time reported or updated, and adverse reaction(s). Re-exposure to the same drug or drug class was not included in this analysis. We also collected all patients’ demographic information, including sex and race/ethnicity, when available in the EHR systems. Contrast media were not included in this analysis, as they represent a distinct group of drugs used for procedures or imaging studies.
While the majority (94%) of PEAR allergy records were entered and stored in a coded format, 6% were entered using free text (i.e. noncoded data). For free-text entries, we used natural language processing techniques in a semi-automated manner to normalize and ‘map’ these entries to the coded form (14). We then manually reviewed all of the mappings.
Data analysis
In EHRs, the documented drug allergen can be a specific drug or a drug class. We classified each allergen entry into a corresponding drug class using the American Hospital Formulary Service Pharmacologic-Therapeutic Classification for reporting prevalence (Tables 1 and 2), and then further manually classified some drug classes into broader ones (e.g. ‘cephalosporin antibiotics—1st generation’ and ‘cephalosporin antibiotics—2nd generation’ into cephalosporins) (15). Under each class, we analyzed and presented the most common drug allergies.
Table 1.
Drug allergy prevalence among the entire patient population by drug class and patient’s sex
| Drug class and allergen | Total (n = 1 766 328)*,† n (%) |
Male (n = 742 374) n (%) |
Female (n = 1 023 824) n (%)‡ |
|---|---|---|---|
|
| |||
| Penicillins | 225 957 (12.8) | 70 160 (9.5) | 155 797 (15.2) |
| Amoxicillin | 36 948 (2.1) | 10 998 (1.5) | 25 950 (2.5) |
| Sulfonamide Antibiotic | 130 029 (7.4) | 23 999 (3.2) | 106 030 (10.4) |
| Trimethoprim/Sulfamethoxazole | 21 682 (1.2) | 4224 (0.6) | 17 458 (1.7) |
| Opiates | 119 992 (6.8) | 25 595 (3.5) | 94 389 (9.2) |
| Codeine | 54 609 (3.1) | 9667 (1.3) | 44 942 (4.4) |
| Morphine | 22 908 (1.3) | 5397 (0.7) | 17 511 (1.7) |
| Oxycodone/Acetaminophen | 18 678 (1.1) | 3791 (0.5) | 14 887 (1.5) |
| Meperidine HCl | 12 938 (0.7) | 2478 (0.3) | 10 460 (1.0) |
| Oxycodone | 10 176 (0.6) | 2162 (0.3) | 8014 (0.8) |
| Hydrocodone/Acetaminophen | 7021 (0.4) | 1494 (0.2) | 5527 (0.5) |
| NSAIDs | 62 145 (3.5) | 17 317 (2.3) | 44 828 (4.4) |
| Aspirin | 28 738 (1.6) | 8290 (1.1) | 20 448 (2.0) |
| Ibuprofen | 11 927 (0.7) | 3374 (0.5) | 8553 (0.8) |
| Naproxen | 6447 (0.4) | 1526 (0.2) | 4921 (0.5) |
| Macrolides | 46 269 (2.6) | 8785 (1.2) | 37 484 (3.7) |
| Erythromycin | 25 905 (1.5) | 4547 (0.6) | 21 358 (2.1) |
| Azithromycin | 7072 (0.4) | 1672 (0.2) | 5400 (0.5) |
| ACE Inhibitors | 35 904 (2.0) | 14 290 (1.9) | 21 614 (2.1) |
| Lisinopril | 21 794 (1.2) | 8345 (1.1) | 13 449 (1.3) |
| Cephalosporins | 30 272 (1.7) | 7812 (1.1) | 22 460 (2.2) |
| Cephalexin | 10 232 (0.6) | 2524 (0.3) | 7708 (0.8) |
| HMG CoA Reductase Inhibitors (Statins) | 26 274 (1.5) | 10 962 (1.5) | 15 312 (1.5) |
| Atorvastatin | 11 400 (0.7) | 4878 (0.7) | 6522 (0.6) |
| Simvastatin | 8912 (0.5) | 3550 (0.48) | 5362 (0.52) |
| Fluoroquinolones | 22 417 (1.3) | 5156 (0.7) | 17 261 (1.7) |
| Ciprofloxacin | 12 080 (0.7) | 2709 (0.4) | 9371 (0.9) |
| Levofloxacin | 7942 (0.5) | 1927 (0.3) | 6015 (0.6) |
| Tetracyclines | 20 454 (1.2) | 4016 (0.5) | 16 438 (1.6) |
| Tetracycline | 7237 (0.4) | 1483 (0.2) | 5754 (0.6) |
| Selective Serotonin Reuptake Inhibitors (SSRIs) | 10 526 (0.6) | 2728 (0.4) | 7798 (0.8) |
| Phenothiazines | 9129 (0.5) | 1450 (0.2) | 7679 (0.8) |
| Prochlorperazine | 8974 (0.5) | 1438 (0.2) | 7536 (0.7) |
| Thiazide Diuretics | 7726 (0.4) | 2527 (0.3) | 5199 (0.5) |
| Hydrochlorothiazide | 7309 (0.4) | 2384 (0.3) | 4925 (0.5) |
| Lincosamide Antibiotics | 6298 (0.4) | 1276 (0.2) | 5022 (0.5) |
| Clindamycin | 6216 (0.4) | 1254 (0.2) | 4962 (0.5) |
| Other | 198 142 (11.2) | 59 158 (8.0) | 138 984 (13.6) |
| Acetaminophen | 7554 (0.4) | 1894 (0.3) | 5660 (0.6) |
Sex information was missing for 130 patients and not included in this table.
Number of allergies to a specific drug class may be less than the total number of individual drugs within the class due to patients having more than one documented drug allergy.
All drug class and individual drug allergies were significantly more prevalent among women, except for HMG CoA Reductase Inhibitors (Statins) (P = 0.30) and Atorvastatin (P = 0.10).
Percentages for Simvastatin were presented as hundredths to show significant difference between genders.
Table 2.
Drug allergy prevalence among the entire patient population by drug class and patient’s race/ethnicity
| Drug class and allergen | White (n = 1 241 012) n (%)* |
Black (n = 125 798) n (%) |
Hispanic (n = 143 335) n (%) |
Asian (n = 71 860) n (%) |
Other (n = 32 218) n (%) |
|---|---|---|---|---|---|
|
| |||||
| Penicillins | 177 216 (14.3) | 10 516 (8.4) | 10 735 (7.5) | 5468 (7.6) | 2864 (8.9) |
| Amoxicillin | 30 156 (2.4) | 1132 (0.9) | 1433 (1.0) | 905 (1.3) | 389 (1.2) |
| Sulfonamide Antibiotic | 108 464 (8.7) | 4163 (3.3) | 2895 (2.0) | 2486 (3.5) | 1185 (3.7) |
| Trimethoprim/Sulfamethoxazole | 17 599 (1.4) | 1050 (0.8) | 874 (0.6) | 373 (0.5) | 203 (0.6) |
| Opiates | 99 901 (8.1) | 5082 (4.0) | 3540 (2.5) | 1243 (1.7) | 1047 (3.3) |
| Codeine | 45 980 (3.7) | 2024 (1.6) | 1171 (0.8) | 473 (0.7) | 459 (1.4) |
| Morphine | 19 261 (1.6) | 906 (0.7) | 675 (0.5) | 183 (0.3) | 208 (0.7) |
| Oxycodone/Acetaminophen | 15 239 (1.2) | 1033 (0.8) | 669 (0.5) | 215 (0.3) | 180 (0.6) |
| Meperidine HCl | 11 233 (0.9) | 293 (0.2) | 186 (0.1) | 95 (0.1) | 79 (0.3) |
| Oxycodone | 8171 (0.7) | 682 (0.5) | 456 (0.3) | 156 (0.2) | 81 (0.3) |
| Hydrocodone/Acetaminophen | 5911 (0.5) | 265 (0.2) | 268 (0.2) | 79 (0.1) | 63 (0.2) |
| NSAIDs | 44 552 (3.6) | 4969 (4.0) | 4859 (3.4) | 1954 (2.7) | 1005 (3.1) |
| Aspirin | 19 069 (1.5) | 2852 (2.3) | 2833 (2.0) | 1014 (1.4) | 596 (1.9) |
| Ibuprofen | 8255 (0.7) | 1066 (0.9) | 1038 (0.7) | 441 (0.6) | 205 (0.6) |
| Naproxen | 5036 (0.4) | 362 (0.3) | 354 (0.3) | 163 (0.2) | 54 (0.2) |
| Macrolides | 39 340 (3.2) | 1434 (1.1) | 845 (0.6) | 668 (0.9) | 373 (1.2) |
| Erythromycin | 21 981 (1.8) | 852 (0.7) | 394 (0.3) | 349 (0.5) | 216 (0.7) |
| Azithromycin | 5925 (0.5) | 224 (0.2) | 222 (0.2) | 140 (0.2) | 55 (0.2) |
| ACE Inhibitors | 27 302 (2.2) | 3207 (2.6) | 1935 (1.4) | 1128 (1.6) | 373 (1.2) |
| Lisinopril | 16 650 (1.3) | 1919 (1.5) | 1166 (0.8) | 662 (0.9) | 214 (0.7) |
| Cephalosporins | 25 564 (2.1) | 805 (0.6) | 745 (0.5) | 546 (0.8) | 277 (0.9) |
| Cephalexin | 8710 (0.7) | 264 (0.2) | 222 (0.2) | 152 (0.2) | 81 (0.3) |
| HMG CoA Reductase Inhibitors (Statins) | 22 090 (1.8) | 1157 (0.9) | 817 (0.6) | 395 (0.6) | 209 (0.7) |
| Atorvastatin | 9645 (0.8) | 507 (0.4) | 306 (0.2) | 175 (0.2) | 91 (0.3) |
| Simvastatin | 7432 (0.6) | 445 (0.4) | 348 (0.2) | 156 (0.2) | 75 (0.2) |
| Fluoroquinolones | 19 111 (1.5) | 566 (0.5) | 602 (0.4) | 359 (0.5) | 170 (0.5) |
| Ciprofloxacin | 10 236 (0.8) | 304 (0.2) | 342 (0.2) | 201 (0.3) | 95 (0.3) |
| Levofloxacin | 6798 (0.6) | 216 (0.2) | 206 (0.1) | 113 (0.2) | 63 (0.2) |
| Tetracyclines | 16 877 (1.4) | 868 (0.7) | 501 (0.4) | 438 (0.6) | 173 (0.5) |
| Tetracycline | 6031 (0.5) | 276 (0.2) | 133 (0.1) | 118 (0.2) | 56 (0.2) |
| Selective Serotonin Reuptake Inhibitors (SSRIs) | 8687 (0.7) | 377 (0.3) | 731 (0.5) | 158 (0.2) | 86 (0.3) |
| Phenothiazines | 7321 (0.6) | 528 (0.4) | 430 (0.3) | 136 (0.2) | 93 (0.3) |
| Prochlorperazine | 7226 (0.6) | 524 (0.4) | 421 (0.3) | 132 (0.2) | 86 (0.3) |
| Thiazide Diuretics | 5956 (0.5) | 679 (0.5) | 372 (0.3) | 229 (0.3) | 74 (0.2) |
| Hydrochlorothiazide | 5625 (0.5) | 641 (0.5) | 350 (0.2) | 226 (0.3) | 71 (0.2) |
| Lincosamide Antibiotics | 5336 (0.4) | 239 (0.2) | 172 (0.1) | 114 (0.2) | 51 (0.2) |
| Clindamycin | 5247 (0.4) | 231 (0.2) | 163 (0.1) | 117 (0.2) | 50 (0.2) |
| Other | 159 720 (12.9) | 9964 (7.9) | 9047 (6.3) | 4126 (5.7) | 2030 (6.3) |
| Acetaminophen | 5635 (0.5) | 561 (0.5) | 459 (0.3) | 209 (0.3) | 104 (0.3) |
All drug class and individual drug allergies were significantly more prevalent among white patients expect NSAIDs, ACE inhibitors, and thiazide diuretics.
Reported drug allergy prevalence rates were calculated as the number of patients with an active allergy to a drug or drug class divided by the total number of the patient population. We then calculated and compared the prevalence rates by sex (i.e. male and female) and racial/ethnic group (i.e. white, black, Asian, Hispanic, and other). We calculated the number of drug allergies documented per patient for the total population and reported the frequency of patients having 1, 2, 3. . ., or 10+ drug allergies. The average number of reported drug allergies per patient was compared by sex and race/ethnicity.
The percentage of each suspected drug alleren class was calculated for each year by dividing the number of allergies documented for that drug class by the total number of documented drug allergies in the same year. We then conducted a trend analysis by comparing these percentages over the past two decades. Drug classes in each of four broader groups (antibiotics, analgesics, antihypertensives, and others) were presented together on separate time line graphs for comparison.
Statistical analyses were conducted using SAS statistical software (version 9.3; SAS Institute, Inc Cary, NC, USA). Chi-square tests were used to compare prevalence rates across patient groups, t-tests were used to compare means (e.g. number of reported drug allergies across patient groups), and P < 0.05 was considered statistically significant. This study was approved by the Partners Institutional Review Board.
Results
General description of the patient population
Among the total of 1 766 328 patients included in this study, 1 023 824 (58.0%) were female, 742 374 (42.0%) were male, and 130 patients’ sex information was missing. The majority of the patients were white (70.3%, n = 1 241 012), followed by Hispanic (8.1%, n = 143 335), black (7.1%, n = 125 798), Asian (4.1%, n = 71 860), and other (1.8%, n = 32 218). There were 152 105 (8.6%) patients whose race/ethnicity information was unknown. In total, there were 1 222 164 active drug allergy records. More than one-third of the population (35.5%, n = 626 871) had at least one drug allergy documented.
Prevalence rates of drug allergies
Table 1 shows the prevalence of the top 14 reported drug allergen classes. Individual drugs that made up 0.5% of all reported drug allergies were included in the table. Among the entire patient population (n = 1 766 328), the most commonly reported allergens were penicillins (12.8%), sulfonamide antibiotics (7.4%), and opiates (6.8%). Other frequently reported drug allergies included NSAIDs (3.5%), macrolides (2.6%), angiotensin-converting enzyme (ACE) inhibitors (2.0%), and cephalosporins (1.7%). The most commonly reported individual drugs were codeine (3.1%), amoxicillin (2.1%), and aspirin (1.6%). Although selective serotonin re-uptake inhibitor (SSRI) antidepressants were a commonly reported drug class, no individual SSRI was documented in more than 0.5% of all drug allergy records.
Drug allergy prevalence rates by sex
Almost all of the reported drug allergy classes and individual drug allergies were significantly more prevalent among women (Table 1). Antibiotic drug classes were among the most prevalent reported drug allergies, including penicillins (15.2% in women vs 9.5% in men, P < 0.01), sulfonamide antibiotics (10.4% vs 3.2%, P < 0.01), macrolides (3.7% vs 1.2%, P < 0.01), cephalosporins (2.2% vs 1.0%, P < 0.01), fluoroquinolones (1.7% vs 0.7%, P < 0.01), and tetracyclines (1.6% vs 0.5%, P < 0.01).
Reported NSAID allergies were more prevalent in women (4.4% vs 2.3%, P < 0.01). Aspirin, the most commonly reported NSAID allergy in the population was found in 2.0% of women and 1.1% of men (P < 0.01). Ibuprofen was considerably more prevalent in women (0.8% vs 0.5%, P < 0.01).
The overall prevalence rates of reported statin allergies were comparable between women and men (1.5% vs 1.5%, P = 0.30), as were the rates for reported atorvastatin allergy (0.6% vs 0.7%, P = 0.10). However, reported simvastatin allergy was significantly more prevalent among women (0.52% vs 0.48%, P < 0.01). No reported allergies to any drug or drug class were significantly more prevalent in men compared to women.
Prevalence rates by race/ethnicity
White patients had significantly more reported drug allergies than any other racial group (Table 2). Penicillins (14.3% in whites vs 7.9% in nonwhites, P < 0.01), sulfonamide antibiotics (8.7% vs 2.9%, P < 0.01), and opiates (8.1% vs 2.9%, P < 0.01) were among the most prevalent reported drug classes in whites. Codeine (3.7% vs 1.1%, P < 0.01) was the most prevalent reported individual drug in whites, followed by amoxicillin (2.4% vs 1.0%, P < 0.01) and erythromycin (1.8% vs 0.5%, P P < 0.01).
Some reported allergens were significantly more prevalent in black patients. These included NSAIDs (4.0% in blacks vs 3.5% in non-blacks, P < 0.01), ACE inhibitors (2.6% vs 2.1%, P < 0.01), and thiazide diuretics (0.54% vs 0.45%, P < 0.01). Commonly reported individual drugs within these classes were also more prevalent in black patients, including aspirin (2.3% vs 1.6%, P < 0.01), ibuprofen (0.9% vs 0.7%, P < 0.01), lisinopril (1.5% vs 1.23%, P < 0.01), and hydrochlorothiazide (0.5% vs 0.4%, P < 0.01). Reported naproxen allergy, however, was significantly more prevalent in white patients (0.4% in whites vs 0.3% in non-whites, P < 0.01). No reported drug class allergies or individual drug allergies were significantly more prevalent in any other race/ethnicity (i.e. Hispanic, Asian, or other).
Number of drug allergies per patient
Among the 626 871 patients who had at least one drug allergy documented in their EHRs, 57.4% had only one drug allergy, 21.8% had two, and 9.5% had three (Table 3). On average, patients with at least one allergy had 1.95 drug allergies documented. Females and white patients had significantly higher numbers of allergies compared to other groups. On average, female patients had 2.09 drug allergies documented compared to 1.63 drug allergies documented in male patients (P < 0.01). White patients had 2.00 drug allergies documented on average (vs 1.71 in non-whites, P < 0.01), 1.83 in blacks (vs 1.97 in non-blacks, P < 0.01), 1.64 in Hispanics (vs 1.98 in non-Hispanics, P < 0.01), 1.59 in Asians (vs 1.97 in non-Asians, P < 0.01), and 1.67 in others (vs 1.96 in nonothers, P < 0.01).
Table 3.
Number of drug allergies documented per patient by race and sex among patients with at least one drug allergy
| Overall n (%) |
Sex* n (%) |
Race† n (%) |
||||||
|---|---|---|---|---|---|---|---|---|
| Number of drug allergies | Overall (n = 626 871) |
Female (n = 433 164) |
Male (n = 193 659) |
White (n = 494 893) |
Black (n = 31 280) |
Hispanic (n = 28 522) |
Asian (n = 14 900) |
Other (n = 7322) |
|
| ||||||||
| Mean (95% CI) | 1.95 (1.95–1.95) | 2.09 (2.09–2.10) | 1.63 (1.62–1.63) | 2.00 (2.00–2.01) | 1.83 (1.81–1.84) | 1.64 (1.62–1.65) | 1.59 (1.57–1.61) | 1.70 (1.67–1.73) |
| 1 | 359 500 (57.4) | 230 978 (53.3) | 128 493 (66.3) | 275 425 (55.7) | 19 396 (62.0) | 19 172 (67.2) | 10 212 (68.5) | 4757 (65.0) |
| 2 | 136 340 (21.8) | 97 893 (22.6) | 38 438 (19.9) | 109 907 (22.2) | 6248 (20.0) | 5384 (18.9) | 2845 (19.1) | 1461 (20.0) |
| 3 | 59 352 (9.5) | 45 333 (10.5) | 14 013 (7.2) | 49 043 (9.9) | 2562 (8.2) | 2027 (7.1) | 919 (6.2) | 557 (7.6) |
| 4 | 29 458 (4.7) | 23 360 (5.4) | 6097 (3.2) | 24 733 (5.0) | 1296 (4.1) | 838 (2.9) | 409 (2.7) | 233 (3.2) |
| 5 | 15 929 (2.5) | 13 030 (3.0) | 2898 (1.5) | 13 430 (2.7) | 685 (2.2) | 464 (1.6) | 214 (1.4) | 127 (1.7) |
| 6 | 9246 (1.5) | 7745 (1.8) | 1501 (0.8) | 7807 (1.6) | 396 (1.3) | 240 (0.8) | 122 (0.8) | 75 (1.0) |
| 7 | 5555 (0.9) | 4741 (1.1) | 812 (0.4) | 4694 (1.0) | 232 (0.7) | 149 (0.5) | 64 (0.4) | 33 (0.5) |
| 8 | 3597 (0.6) | 3100 (0.7) | 497 (0.3) | 3077 (0.6) | 138 (0.4) | 81 (0.3) | 37 (0.3) | 24 (0.3) |
| 9 | 2328 (0.4) | 1999 (0.5) | 329 (0.2) | 2009 (0.4) | 99 (0.3) | 51 (0.2) | 23 (0.2) | 11 (0.2) |
| 10+ | 5566 (0.9) | 4985 (1.2) | 581 (0.3) | 4768 (1.0) | 228 (0.7) | 116 (0.4) | 55 (0.4) | 44 (0.6) |
Sex information was missing for 48 patients.
Race information was missing for 49 954 patients.
Percentages of drug allergen classes reported each year
In recent years, reported penicillin allergies have declined but still remained the most commonly reported antibiotic allergy (Fig. 1A). Penicillins still made up 16% of reported drug allergies in 2013. The prevalence of sulfonamide antibiotic allergies also declined but this remained the second most commonly reported antibiotic allergy (8% in 2013). After an initial increase (3.2% in 1992 and 3.8% in 1993 for macrolides and 2.2% in 1992 and 2.7% in 1993 for cephalosporins), the prevalence rates for other antibiotics remained fairly constant. The prevalence of reported tetracyclines, fluoroquinolones, and clindamycin allergies remained constant over the study period and are of the least reported antibiotic allergies.
Figure 1.
(A–D) Display percentages of common drug allergies over all reported drug allergies annually between 1990 and 2013. (A) Antibiotic drug allergies. (B) Analgesic drug allergies. (C) Antihypertensive drug allergies. (D) Other drug allergies.
Until 1993, NSAIDs were the more commonly reported analgesic allergy (Fig. 1B). Recently, however, opiates have become the more frequently reported allergy. Opiates have made up 10–15% of reported drug allergies in the last 13 years, and NSAIDs have only made up 5–8%.
Thiazides remained fairly constant throughout the years and made up one of the lowest reported drug classes out of the most frequent (<1% every year) (Fig. 1C). In contrast, ACE inhibitors have been steadily increasing starting from 2000 (1.4% in 2000 and 3.6% in 2013).
Phenothiazines and SSRI prevalence rates have been relatively constant (phenothiazines with a slight decrease and SSRIs with a slight increase) (Fig. 1D). Statins, however, have been reported more frequently over the years and have doubled since 2003 (1.5% in 2003 and 3.4% in 2013).
Discussion
This study represents one of the largest EHR-based reports of drug allergy prevalence and associated patient characteristics in a large health system (7). We found that antibiotics continue to represent a large proportion of the medications involved. Also, females and white patients had more drug allergies documented than other patients.
Our findings with respect to the medications involved are consistent with prior studies. A study of self-reported drug allergy surveys in an adult Portuguese population found that the most frequently reported drugs were penicillins or other beta-lactams, followed by aspirin or other NSAIDs (5). Penicillins were also the most commonly reported drug class allergy in our results. Cephalosporins and NSAIDs were among our top drug classes. Amoxicillin was the most common penicillin drug allergy, which is consistent with other reports (16).
Antibiotics have been reported as the majority of drug allergies documented, and we found that antibiotics accounted for seven of the top 14 drug classes (8). In a study of reported antibiotic allergy incidence and prevalence using outpatient medical records, sulfonamide antibiotics were found to be the most prevalent antibiotic allergy among both men and women, followed by penicillins (8). Our results, however, showed the reverse; penicillins were the most commonly reported antibiotic allergy followed by sulfonamide antibiotics. Our prevalence rates for both penicillins and sulfonamide antibiotics were substantially higher than reported in the prior antibiotic allergy study (8). Past reports have found that ciprofloxacin was the most prevalent fluoroquinolone antibiotic allergy, which was consistent with our findings (4). Our higher observed reported antibiotic allergy rates may be due to higher overall antibiotic exposure in tertiary care hospitals (8, 17).
NSAIDs were the 4th most common drug allergy reported in our data, with aspirin being the most common individual drug reported in the class. Some studies have also found that aspirin was the most common NSAID allergy, and others have found it to be naproxen (5, 18). In a 10-month prospective cohort study on adverse cutaneous drug reactions (ACDRs) in hospitalized patients, naproxen was found to be the 4th drug most frequently associated with ACDRs (18). In our data, naproxen (0.4%) was much less commonly reported compared to aspirin (1.6%). Ibuprofen was found to be the most common NSAID involved with drug hypersensitivity reactions in a 6-year study of a large population, but our data showed reported ibuprofen allergies (0.7%) to be less common than reported aspirin allergies (19).
Two classes of antihypertensives were frequently documented as drug allergies, ACE inhibitors and thiazide diuretics. Antihypertensives have been reported to be frequent causes of hospitalizations due to adverse reactions (20). Common adverse reactions include bradykinin-induced cough for ACE inhibitors and electrolyte imbalances (e.g. hyponatremia and hypokalemia) for thiazide diuretics (21, 22). Although beta-blockers and calcium channel blockers can also cause allergies and ADRs, we did not find these drug classes to be as common as ACE inhibitors and thiazide diuretics in our study (20).
Like many other studies, we found that reported drug allergies were significantly more prevalent among females (5, 7, 8, 23). Females were also found to have more drug allergies documented per patient than men on average. All drug classes and individual drugs were more commonly reported in women with the exception of statins, which were comparable in both genders. This exception may be because males have been more likely to be prescribed a statin compared to females, shown by studies of statin prescribing patterns in recent years (24, 25). An increased usage of statins could have reflected a higher prevalence of reported statin allergies. Also, as many patients may be intolerant to statins due to their common nonimmune-mediated adverse events (e.g. myalgias and myositis), many of these reported allergies are not ‘true’ immune-mediated reactions (26).
Our data showed that most of the top drugs documented were more prevalent in white patients compared to other races/ethnicities. White patients also had more drug allergies documented per patient than any other race/ethnicity. NSAIDs, ACE inhibitors, and thiazide diuretics, however, were significantly more prevalent in black patients. The high ACE inhibitor prevalence may be due to the higher risk and prevalence of ACE inhibitor-induced angioedema in black patients (27). The racial difference in the reported drug allergies may be caused by provider documentation bias as reported in other studies (28–30).
Over one-third of the population had at least one reported drug allergy. This is a higher prevalence rate compared to other studies that reported 20–25% (7, 31). More than half of the patients with reported drug allergies had only one drug allergy. However, we found a much higher number of patients with multiple drug allergies compared to literature (7). Macy et al. reported that multiple drug intolerance syndrome (MDIS; commonly defined as reactions to three or more chemically unrelated medications) was present in 2.1% of their patient population at Kaiser Permanente (7, 32, 33). We found that 7.4% of our total population had MDIS. This may be attributed to our population being an urban tertiary care referral population that may be sicker (and therefore have more medication exposure) than the Kaiser Permanente population, as the largest risk factor for reported drug allergy is drug exposure.
Time prevalence trends showed that reported penicillin and sulfonamide antibiotic allergies declined over time, but still remained the most commonly reported allergies of all antibiotic classes. Macy et al.(8) reported temporal trends of antibiotic usage in the Kaiser Permanente San Diego area, which showed a decline in penicillin usage from 1995 to 2007. The prevalence of reported opiate allergies has been consistently high compared to reported NSAID allergies since about 1995. This is consistent with a reported increase in opioid prescribing and use in past years (34, 35). While thiazide diuretics remained relatively stable throughout the years, reported ACE inhibitor allergies have more than doubled since 2000. This may be because of an increase in usage due to expansion of the drug group’s indications (36). Reported statin allergies have also doubled since 2003. Studies on UK primary care databases showed a steady increase in statin prescriptions due to statins being proven to be the most effective treatment for hyperlipidemia (26, 37). These consistencies between reported drug usage and reported drug allergies show that increased usage may be a risk factor for drug allergies.
It is worth mentioning that systems such as VigiBase (the World Health Organization Global Individual Case Safety Reports database) collect spontaneous ADE reports (38). Compared to the allergy entries in EHRs, the type and severity of reported adverse drug reactions and reported pharmacological groups may be different. For example, the majority of ADEs in VigiBase were reported for nervous system medications, followed by cardiovascular medicines (38).
A limitation to this study is that patients tend to report and clinicians tend to document other types of adverse drug reactions in the EHR allergy module including common side-effects (e.g. diarrhea), intolerances (e.g. gastro-intestinal upset), and immune-mediated hypersensitivities (e.g. anaphylaxis). Whether the reaction was immune-mediated or whether a diagnostic test was performed was often not documented. It is likely that patients may report drug allergies that have never been confirmed and may not be ‘true’ allergies. Further, the quality of the allergy entries in EHRs often depends on clinicians’ expertise and interpretation, which may result in incongruent or unspecific data. Additionally, in this study, prevalence rates were calculated as the number of patients with a reported allergy divided by the total number of the patient population; however, it would also be important to consider patients’ drug exposure by using prescription data. This would allow us to compare results against prescriptions as correlations between gender, race/ethnicity, time trends, and drug allergy reports can probably be explained by different patterns in prescription and drug intake in different subpopulations. This represents an important area of future research. Another limitation is that data were only taken from two institutions in the Boston area, with urban populations that may have more comorbid patients. These patients may not be representative of a broader patient population.
In summary, we found that females and white patients were at a higher risk of experiencing an adverse drug event for most of the commonly reported drugs found in our patient population. The increasing number of reported drug allergies to some of these commonly reported drugs may be due to an increased exposure and usage over time. Clinicians should be wary of this, especially in females and white patients, as many of these commonly reported drugs are the foundation of modern-day medical therapies. In addition, geographical, environmental, and patient population characteristics may contribute to these differences. Future studies should consider these factors in their analysis.
Funding
This study was funded by the Agency for Healthcare Research and Quality (AHRQ) grant R01HS022728.
Abbreviations
- ACE
angiotensin-converting enzyme
- ADE
adverse drug event
- AHFS
American Hospital Formulary Service
- BWH
Brigham and Women’s Hospital
- EHR
electronic health record
- MDIS
multiple drug intolerance syndrome
- MGH
Massachusetts General Hospital
- NSAID
nonsteroidal anti-inflammatory drug
- PEAR
Partners’ Enterprise-wide Allergy Repository
- SSRI
selective serotonin reuptake inhibitor
Footnotes
Conflict of interest
The authors declare that they have no conflicts of interest.
References
- 1.Gandhi TK, Weingart SN, Borus J, Seger AC, Peterson J, Burdick E et al. Adverse drug events in ambulatory care. N Engl J Med 2003;348:1556–1564. [DOI] [PubMed] [Google Scholar]
- 2.Warrington R, Silviu-Dan F. Drug allergy. Allergy Asthma Clin Immunol 2011;7(Suppl 1):S10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Riedl MA, Casillas AM. Adverse drug reactions: types and treatment options. Am Fam Physician 2003;68:1781–1790. [PubMed] [Google Scholar]
- 4.Kulthanan K, Chularojanamontri L, Manapajon A, Dhana N, Jongjarearnprasert K. Cutaneous adverse reactions to fluoroquinolones. Dermatitis 2011;22:155–160. [PubMed] [Google Scholar]
- 5.Gomes E, Cardoso MF, Praca F, Gomes L, Marino E, Demoly P. Self-reported drug allergy in a general adult Portuguese population. Clin Exp Allergy 2004;34:1597–1601. [DOI] [PubMed] [Google Scholar]
- 6.Pawankar R, Canonica GW, Holgate ST, Lockey RF. The Burden of Allergic Diseases. In. World Allergy Organization editor. WAO White Book on Allergy: 2011–2012: Executive Summary. Milwuakee, Wisconsin: World Allergy Organization, 2011: 5. [Google Scholar]
- 7.Macy E, Ho NJ. Multiple drug intolerance syndrome: prevalence, clinical characteristics, and management. Ann Allergy Asthma Immunol 2012;108:88–93. [DOI] [PubMed] [Google Scholar]
- 8.Macy E, Poon KYT. Self-reported antibiotic allergy incidence and prevalence: age and sex effects. Am J Med 2009;122:778. [DOI] [PubMed] [Google Scholar]
- 9.Wood RA, Camargo CA Jr, Lieberman P, Sampson HA, Schwartz LB, Zitt M et al. Anaphylaxis in America: the prevalence and characteristics of anaphylaxis in the United States. J Allergy Clin Immunol 2014;133:461–467. [DOI] [PubMed] [Google Scholar]
- 10.Thong BY, Tan TC. Epidemiology and risk factors for drug allergy. Br J Clin Pharmacol 2011;71:684–700. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jerschow E, Lin RY, Scaperotti MM, McGinn AP. Fatal anaphylaxis in the United States, 1999–2010: temporal patterns and demographic associations. J Allergy Clin Immunol 2014;134:1318–1328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Harduar-Morano L, Simon MR, Watkins S, Blackmore C. A population-based epidemiologic study of emergency department visits for anaphylaxis in Florida. J Allergy Clin Immunol 2011;128:594–600. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kuperman GJ, Marston E, Paterno M, Rogala J, Plaks N, Hanson C et al. Creating an enterprise-wide allergy repository at Partners HealthCare System. AMIA Annu Symp Proc 2003;2003:376–380. [PMC free article] [PubMed] [Google Scholar]
- 14.Goss FR, Plasek JM, Lau JJ, Seger DL, Chang FY, Zhou L. An evaluation of a natural language processing tool for identifying and encoding allergy information in emergency department clinical notes. AMIA Annu Symp Proc 2014;2014:580–588. [PMC free article] [PubMed] [Google Scholar]
- 15.American Society of Health-System Pharmacists. American Hospital Formulary Service – Pharmacologic Therapeutic Classification. In. 2015. ed. AHFS Drug Information: American Society of Health-System Pharmacists. [Google Scholar]
- 16.Gamboa PM. The epidemiology of drug allergy-related consultations in Spanish Allergology services: Alergologica-2005. J Investig Allergol Clin Immunol 2009;19 (Suppl 2):45–50. [PubMed] [Google Scholar]
- 17.Ingram PR, Seet JM, Budgeon CA, Murray R. Point-prevalence study of inappropriate antibiotic use at a tertiary Australian hospital. Intern Med J 2012;42:719–721. [DOI] [PubMed] [Google Scholar]
- 18.Hernandez-Salazar A, Rosales SP, Rangel-Frausto S, Criollo E, Archer-Dubon C, Orozco-Topete R. Epidemiology of adverse cutaneous drug reactions. A prospective study in hospitalized patients. Arch Med Res 2006;37:899–902. [DOI] [PubMed] [Google Scholar]
- 19.Dona I, Blanca-Lopez N, Torres MJ, Garcia-Campos J, Garcia-Nunez I, Gomez F et al. Drug hypersensitivity reactions: response patterns, drug involved, and temporal variations in a large series of patients. J Investig Allergol Clin Immunol 2012;22:363–371. [PubMed] [Google Scholar]
- 20.Lee S, Hess EP, Nestler DM, Bellamkonda Athmaram VR, Bellolio MF, Decker WW et al. Antihypertensive medication use is associated with increased organ system involvement and hospitalization in emergency department patients with anaphylaxis. J Allergy Clin Immunol 2013;131:1103–1108. [DOI] [PubMed] [Google Scholar]
- 21.Israili ZH, Hall WD. Cough and angioneurotic edema associated with angiotensin-converting enzyme inhibitor therapy. A review of the literature and pathophysiology. Ann Intern Med 1992;117:234–242. [DOI] [PubMed] [Google Scholar]
- 22.Clayton JA, Rodgers S, Blakey J, Avery A, Hall IP. Thiazide diuretic prescription and electrolyte abnormalities in primary care. Br J Clin Pharmacol 2006;61:87–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Park MA, Matesic D, Markus PJ, Li JT. Female sex as a risk factor for penicillin allergy. Ann Allergy Asthma Immunol 2007;99:54–58. [DOI] [PubMed] [Google Scholar]
- 24.Johansen ME, Green LA, Sen A, Kircher S, Richardson CR. Cardiovascular risk and statin use in the United States. Ann Fam Med 2014;12:215–223. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Pauff BR, Jiroutek MR, Holland MA, Sutton BS. Statin prescribing patterns: an analysis of data from patients with diabetes in the national hospital ambulatory medical care survey outpatient department and national ambulatory medical care survey databases, 2005–2010. Clin Ther 2015;37:1329–1339. [DOI] [PubMed] [Google Scholar]
- 26.Molokhia M, McKeigue P, Curcin V, Majeed A. Statin induced myopathy and myalgia: time trend analysis and comparison of risk associated with statin class from 1991–2006. PLoS ONE 2008;3:e2522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Brown NJ, Ray WA, Snowden M, Griffin MR. Black Americans have an increased rate of angiotensin converting enzyme inhibitor-associated angioedema. Clin Pharmacol Ther 1996;60:8–13. [DOI] [PubMed] [Google Scholar]
- 28.Burgess DJ, Gravely AA, Nelson DB, van Ryn M, Bair MJ, Kerns RD et al. A national study of racial differences in pain screening rates in the VA health care system. Clin J Pain 2013;29:118–123. [DOI] [PubMed] [Google Scholar]
- 29.Saha S, Freeman M, Toure J, Tippens KM, Weeks C, Ibrahim S. Racial and ethnic disparities in the VA health care system: a systematic review. J Gen Intern Med 2008;23:654–671. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Williams EC, Lapham GT, Hawkins EJ, Rubinsky AD, Morales LS, Young BA et al. Variation in documented care for unhealthy alcohol consumption across race/ethnicity in the Department of Veterans Affairs Healthcare System. Alcohol Clin Exp Res 2012;36:1614–1622. [DOI] [PubMed] [Google Scholar]
- 31.Lee CE, Zembower TR, Fotis MA, Postel-nick MJ, Greenberger PA, Peterson LR et al. The incidence of antimicrobial allergies in hospitalized patients: implications regarding prescribing patterns and emerging bacterial resistance. Arch Intern Med 2000;160:2819–2822. [DOI] [PubMed] [Google Scholar]
- 32.Schiavino D, Nucera E, Roncallo C, Pollastrini E, De Pasquale T, Lombardo C et al. Multiple-drug intolerance syndrome: clinical findings and usefulness of challenge tests. Ann Allergy Asthma Immunol 2007;99:136–142. [DOI] [PubMed] [Google Scholar]
- 33.Blumenthal KG, Saff RR, Banerji A. Evaluation and management of a patient with multiple drug allergies. Allergy Asthma Proc 2014;35:197–203. [DOI] [PubMed] [Google Scholar]
- 34.Manchikanti L, Helm S 2nd, Fellows B, Janata JW, Pampati V, Grider JS et al. Opioid epidemic in the United States. Pain Physician 2012;3(Suppl):ES9–ES38. [PubMed] [Google Scholar]
- 35.Mazer-Amirshahi M, Mullins PM, Rasooly I, van den Anker J, Pines JM. Rising opioid prescribing in adult U.S. emergency department visits: 2001–2010. Acad Emerg Med 2014;21:236–243. [DOI] [PubMed] [Google Scholar]
- 36.Psaty BM, Manolio TA, Smith NL, Heckbert SR, Gottdiener JS, Burke GL et al. Time trends in high blood pressure control and the use of antihypertensive medications in older adults: the Cardiovascular Health Study. Arch Intern Med 2002;162:2325–2332. [DOI] [PubMed] [Google Scholar]
- 37.Stone NJ, Robinson JG, Lichtenstein AH, Bairey Merz CN, Blum CB, Eckel RH et al. 2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation 2014;129(25 Suppl 2):S1–S45. [DOI] [PubMed] [Google Scholar]
- 38.Aagaard L, Strandell J, Melskens L, Petersen PS, Holme Hansen E. Global patterns of adverse drug reactions over a decade: analyses of spontaneous reports to VigiBase. Drug Saf 2012;35:1171–1182. [DOI] [PubMed] [Google Scholar]