Reported β-Lactam and Other Antibiotic Allergies in Solid Organ and Hematopoietic Cell Transplant Recipients

Abstract

Background: Patients with reported β-lactam antibiotic allergies (BLAs) are more likely to receive broad-spectrum antibiotics and experience adverse outcomes. Data describing antibiotic allergies among solid organ transplant (SOT) and hematopoietic cell transplant (HCT) recipients are limited.

Methods: We reviewed records of adult SOT or allogeneic HCT recipients from 1 January 2013 to 31 December 2017 to characterize reported antibiotic allergies at time of transplantation. Inpatient antibiotic use was examined for 100 days posttransplant. Incidence rate ratios (IRRs) comparing antibiotic use in BLA and non-BLA groups were calculated using multivariable negative binomial models for 2 metrics: days of therapy (DOT) per 1000 inpatient days and percentage of antibiotic exposure-days.

Results: Among 2153 SOT (65%) and HCT (35%) recipients, 634 (29%) reported any antibiotic allergy and 347 (16%) reported BLAs. Inpatient antibiotics were administered to 2020 (94%) patients during the first 100 days posttransplantation; average antibiotic exposure was 41% of inpatient-days (interquartile range, 16.7%-62.5%). BLA patients had significantly higher DOT for vancomycin (IRR, 1.4 [95% confidence interval {CI}, 1.2-1.7]; P < .001), clindamycin (IRR, 7.6 [95% CI, 2.2-32.4]; P = .001), and aztreonam in HCT (IRR, 9.7 [95% CI, 3.3-35.0]; P < .001), and fluoroquinolones in SOT (IRR, 2.9 [95% CI, 2.1-4.0]; P < .001); these findings were consistent when using percentage of antibiotic exposure-days.

Conclusions: Transplant recipients are frequently exposed to antibiotics and have a high prevalence of reported antibiotic allergies. Reported BLA was associated with greater use of β-lactam antibiotic alternatives. Pretransplant antibiotic allergy evaluation may optimize antibiotic use in this population.

Keywords: allergy; hematopoietic cell transplantation; solid organ transplantation; β-lactams.

Figure 1.

A–F, Prevalence of reported antibiotic allergies among solid organ transplant/hematopoietic cell transplant (HCT) recipients by transplant type. Patients with reported antibiotic allergies to >1 class are counted in both classes. Of 2153 patients, 634 (29%) reported at least 1 allergy. Of those 634 patients, 129 (20%) reported antibiotic allergies to 2 different classes, 48 (8%) reported allergies to 3 classes, 14 (2%) patients reported allergies to 4 classes, 2 (0.3%) patients reported allergies to 5 classes, and 1 (0.1%) patient reported allergies to 7 classes of antibiotics. In (A), patients who reported multiple β-lactam drug allergies were counted only once. Twenty-six of 347 (7%) reported both a penicillin and cephalosporin allergy, 1 (0.2%) reported both a cephalosporin and carbapenem allergy, 2 (0.6%) reported penicillin and carbapenem allergies, and 1 (0.3%) reported allergies to all 3 classes. Other allergy classes were not presented due to low numbers (see Supplementary Table 2 for full results): carbapenems (n = 16), tetracyclines (n = 37), aminoglycosides (n = 10), macrolides (n = 56), clindamycin (n = 20), linezolid (n = 2), metronidazole (n = 8), nitrofurantoin (n = 16), rifamycins (n = 3), daptomycin (n = 2), trimethoprim (n = 1), aztreonam (n = 5), polymyxins (n = 5), telavancin (n = 1). *Computed using Fisher exact test, remainder tested by χ ² test.

Figure 2.

Mean percentage of total antibiotic exposure-days that each antibiotic or class was given. *Significant differences in antibiotic use per antibiotic exposure-days by β-lactam allergy group in the multivariable adjusted model. Abbreviations: Amox-clav, amoxicillin-clavulanic acid; Amp-sul, ampicillin-sulbactam; BLA, β-lactam allergy; Pip-tazo, piperacillin-tazobactam.

Figure 3.

Results from univariate and multivariable negative binomial models examining days of therapy/1000 inpatient-days (A) and percentage of antibiotic exposure-days (B) for each antibiotic. Estimates shown in black are adjusted for age, transplant type, and diagnosis of cystic fibrosis, as appropriate; estimates shown in gray are unadjusted. Points represent incidence rate ratio estimates and horizontal bars represent 95% confidence intervals (CIs). Clindamycin and aztreonam are shown separately to allow viewing of high incidence rate ratio (IRR) estimates; upper limit of the CI may extend beyond the plot (represented by arrow). Aztreonam use was only measured in hematopoietic cell transplant (HCT) recipients, and fluoroquinolone use only measured in solid organ transplant (SOT) recipients. The adjusted IRR for ceftazidime differed by SOT vs HCT group; model estimates are presented separately. Abbreviations: Amox-clav, amoxicillin-clavulanic acid; Amp-sul, ampicillin-sulbactam; IRR, incidence rate ratio; Pip-tazo, piperacillin-tazobactam.