Volume 103, Issue 11 p. 2289-2295
ORIGINAL RESEARCH
Open Access

Antepartum vs postpartum amoxicillin oral challenge in pregnant patients with a reported penicillin allergy: A two-center prospective cohort study

Jeffrey Man Hay Wong

Corresponding Author

Jeffrey Man Hay Wong

Department of Obstetrics and Gynecology, University of British Columbia, Vancouver, British Columbia, Canada

Correspondence

Jeffrey Man Hay Wong, Oak Tree Clinic, 6th Floor, Room E600. 4500 Oak Street, Vancouver, BC V6H 3N1, Canada.

Email: [email protected]

Search for more papers by this author
Xiaoqing Liu

Xiaoqing Liu

Department of Obstetrics, Gynecology and Reproductive Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

Search for more papers by this author
Raymond Mak

Raymond Mak

Division of Allergy and Immunology, Department of Medicine, University of British Columbia, Vancouver, British Columbia, Canada

Division of Allergy and Immunology, Department of Pediatrics, University of British Columbia, Vancouver, British Columbia, Canada

Search for more papers by this author
Stephanie C. Erdle

Stephanie C. Erdle

Division of Allergy and Immunology, Department of Pediatrics, University of British Columbia, Vancouver, British Columbia, Canada

Search for more papers by this author
Colin Barber

Colin Barber

Section of Allergy and Clinical Immunology, Department of Internal Medicine, University of Manitoba, Winnipeg, Manitoba, Canada

Search for more papers by this author
Julianne van Schalkwyk

Julianne van Schalkwyk

Department of Obstetrics and Gynecology, University of British Columbia, Vancouver, British Columbia, Canada

Women's Health Research Institute, BC Women's Hospital and Health center, Vancouver, British Columbia, Canada

Search for more papers by this author
Melissa Watt

Melissa Watt

Section of Allergy and Clinical Immunology, Department of Internal Medicine, University of Manitoba, Winnipeg, Manitoba, Canada

Search for more papers by this author
Sudharsana Rao Ande

Sudharsana Rao Ande

Department of Obstetrics, Gynecology and Reproductive Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

Search for more papers by this author
Dozie Ochulor

Dozie Ochulor

Department of Obstetrics, Gynecology and Reproductive Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

Search for more papers by this author
Chelsea Elwood

Chelsea Elwood

Department of Obstetrics and Gynecology, University of British Columbia, Vancouver, British Columbia, Canada

Women's Health Research Institute, BC Women's Hospital and Health center, Vancouver, British Columbia, Canada

Search for more papers by this author
Vanessa Poliquin

Vanessa Poliquin

Department of Obstetrics, Gynecology and Reproductive Sciences, University of Manitoba, Winnipeg, Manitoba, Canada

Search for more papers by this author
First published: 10 September 2024

Abstract

Introduction

While 10% of pregnant individuals report a penicillin allergy, there is no established best practice for penicillin allergy delabeling in pregnancy. To better understand options for penicillin delabeling, we aimed to evaluate two penicillin allergy delabeling protocols in pregnancy regarding efficacy, adverse events, and patient satisfaction.

Material and Methods

From July 2019 to December 2022, we completed a two-center prospective cohort study, where each site recruited pregnant patients over 24 weeks gestational age with a reported penicillin allergy. One center offered antepartum amoxicillin oral challenges, either directly or after negative skin testing (i.e., antepartum oral challenge site). Our other centers completed a two-step approach with antepartum penicillin skin testing only and deferred oral challenges to the postpartum period (i.e., postpartum oral challenge site). Our primary outcome was the rate of penicillin allergy delabeling, defined as tolerating an antibiotic challenge with penicillin or amoxicillin. Univariate analyses were completed using chi-squared, Fisher's exact, and Wilcoxon rank tests.

Results

During the study period, 276 pregnant patients were assessed, with 207 in the antepartum oral challenge site and 69 in the postpartum oral challenge site. Among the 204 patients who completed antepartum oral challenges, 201 (98%) passed without reactions. Deferring oral challenges to the postpartum period led to a loss of follow-up for 37/53 (70%) of eligible individuals. Overall, 97% (201/207) of patients at the antepartum oral challenge site were delabeled from their penicillin allergy—compared to 38% (26/69) of patients referred to the postpartum oral challenge site (p < 0.0001). Three antepartum oral challenge reactions were noted, including two mild cutaneous reactions and a case of transient abdominal discomfort.

Conclusions

Antepartum amoxicillin oral challenge is a more effective method to delabel pregnant patients from their penicillin allergy. Deferral of oral challenges to the postpartum period introduces a significant barrier for penicillin allergy delabeling.

Abbreviation

  • PST
  • penicillin skin test.
  • Key message

    For pregnant patients with a reported low-risk penicillin allergy, antepartum oral challenge with amoxicillin is a safe and effective way to remove the allergy label. Postponing oral challenges to the postpartum period is unnecessary and can lead to a loss in follow-up.

    1 INTRODUCTION

    While up to 10% of pregnant patients report a penicillin allergy,1 there is no consensus on the best practices for allergy delabeling in pregnancy. This is especially problematic as up to 40% of pregnant individuals will be exposed to antibiotics during their pregnancy.2 Due to the use of second-line agents, pregnant patients with a penicillin allergy are at increased risk of postpartum endometritis,3 postoperative cesarean wound complications, and longer lengths of stay.4 Meanwhile, their newborns are more likely to require blood draws and extended hospitalizations.5

    Penicillin allergy delabeling is a safe and cost-effective practice in the pregnant population. Over 95% of pregnant individuals with a penicillin allergy can be delabeled during pregnancy, increasing their use of first-line antimicrobial therapy in pregnancy.6-8 All noted allergic reactions were mild, requiring only oral antihistamines and/or topical corticosteroids without adverse fetal outcomes. In patients who are successfully delabeled in pregnancy, there is an estimated lifetime cost savings of $58 780 per patient.9

    Historically, obstetrical centers limit antenatal allergy testing to penicillin skin testing (PST) only due to an aversion to challenging pregnant individuals with a potential allergen.10 This practice is especially problematic as a recent systematic review highlighted that the sensitivity of PSTs is under 50% in the non-pregnant population.11 However, as highlighted in the Dutch Working Party on Antibiotic Policy,12 PST is particularly useful in its high negative predictive value. In the existing literature, only one center implemented a direct oral challenge approach to pregnant individuals.6 Other penicillin allergy testing protocols in pregnancy universally completed a PST before an oral challenge, even in low-risk individuals.1, 7, 8 In contrast, in the general population, the American Academy of Allergy, Asthma, and Immunology and the Dutch Working Party on Antibiotic Policy concluded that direct oral challenges should be considered for patients with a history of distant (i.e., >5 years) and non-severe reactions to penicillins12, 13 The recent PALACE trial further affirmed the safety and efficacy of direct oral amoxicillin challenges in the low-risk, non-pregnant population.14

    Given the differences in penicillin delabeling practices, we developed a prospective cohort study to evaluate two allergy testing protocols involving (1) antepartum oral challenge in low-risk individuals and (2) antepartum penicillin skin testing with a deferred postpartum oral challenge. We hypothesize that both protocols have a low adverse event rate, with increased successful delabeling in the antepartum oral challenge protocol.

    2 MATERIAL AND METHODS

    Our prospective cohort study compares penicillin allergy testing between two Canadian centers in Vancouver, British Columbia, and Winnipeg, Manitoba, from July 2019 to December 2022. Both centers receive referrals from community obstetrical care providers (midwives, family physicians, and obstetricians) for pregnant individuals with a penicillin allergy.

    A standardized triage process was completed with the inclusion criteria for testing including:
    1. Pregnant individuals >24 weeks gestational age
    2. Reported history of allergy or reaction to penicillin

    Patients with a history of severe cutaneous adverse reactions (Steven–Johnson syndrome, drug reaction with eosinophilia and systemic symptoms, serum sickness) were excluded from penicillin allergy testing. Patients delabeled from their penicillin allergy by history alone, without needing allergy testing, were also excluded from the analysis.

    When patients present to the penicillin allergy clinics, clinicians and research personnel complete a standardized questionnaire via history-taking to capture baseline demographics, obstetrical data, and allergy history. The patients then underwent testing based on each site's standard practice—either in the antepartum oral challenge site (British Columbia) or the postpartum oral challenge site (Manitoba). Patients are deemed delabeled from their penicillin allergy when they complete a medication challenge with penicillin or amoxicillin without reaction.

    2.1 Antepartum oral challenge protocol in Vancouver, British Columbia (antepartum oral challenge site)

    Using the validated PENFAST tool,15 patients were deemed very low-risk (PENFAST score 0), low-moderate-risk (PENFAST 1–3) and high-risk (PENFAST 4–5) through history-taking by physicians. Very low-risk individuals would be offered a direct oral challenge with 500 mg amoxicillin in a monitored setting. Vital signs and baseline symptoms were noted before the challenge, and a repeat assessment was performed at 15- and 60-min post-challenge. Physicians and nurses were available for additional as-needed assessments during monitoring in case of a reaction. For patients with PENFAST 1–3, a PST was first completed with positive control, negative control, and penicillin testing using benzylpenicillin polylysine and penicillin G potassium. The PST would be evaluated in 15 min to determine whether a reaction occurred (defined as a wheal size >3 mm compared to negative control with erythema). If patients test negative for PST, an oral amoxicillin challenge will be completed in the same visit. Since the PENFAST tool is not well-validated in the pregnant population at study initiation, we elected for a conservative approach to complete PST in patients with a PENFAST score of 1 or 2. If a very low-risk patient preferred to receive a PST before an oral challenge would be completed, we would accommodate this request. High-risk patients were not tested in our clinic during pregnancy.

    2.2 Two-stage protocol: Antepartum penicillin skin testing with postpartum oral challenge in Winnipeg, Manitoba (postpartum oral challenge site)

    Regardless of risk factors for continued penicillin allergy, patients first completed a PST with positive control, negative control, and penicillin testing using benzylpenicillin polylysine and penicillin G potassium. The PST would be evaluated in 15 minutes to determine whether a reaction occurred (defined as a wheal >3 mm compared to negative control with erythema). Patients who did not react to the PST would be permitted to receive antepartum/intrapartum penicillin and its derivatives in pregnancy if clinically required. The use of skin testing to allow for intrapartum penicillin use has been published in the past.10 Patients who did not receive antepartum/intrapartum penicillin were invited to return postpartum to complete an oral challenge with 250 mg amoxicillin by phone and email. Based on the American Academy of Allergy, Asthma, and Immunology guidelines, there is a lack of consensus on whether oral challenges should be completed with 250 or 500 mg amoxicillin.13 Since both are considered standard doses, the specific dose was left to the provincial allergists' discretion. Vital signs and baseline symptoms were noted before the challenge, with repeat assessments every 20 min until an hour had elapsed. Patients were deemed lost to follow-up if they had not returned for allergy testing after 6 months postpartum.

    Patients are deemed delabeled and non-allergic if no reactions occur with an antepartum, intrapartum, or postpartum drug challenge, and their electronic health records are updated to reflect such. Patients are advised to disclose any delayed hypersensitivity reactions to PST or oral challenges by contacting the clinics. After the initial visit, patients were invited to answer a 5-point Likert scale survey on their willingness and comfort to take penicillin antibiotics in the future. After the patient delivers, a chart review was also completed to capture pregnancy complications (eg antibiotics used in labor, infections).

    2.3 Statistical analyses

    Based on British Columbia's pilot study,6 2% of individuals will fail to be delabeled in the antepartum oral challenge protocol. If 8% of the postpartum group (fourfold) failed to be delabeled, this was determined to be a clinically significant difference. Assuming a confidence level of 95% with a desired power of 80%, a sample size of 408 (204 per group) was calculated. The antepartum oral challenge group tested more patients during the study period due to a larger referral population, and it was the first delabeling program in Canada. The analysis was completed with imbalanced groups, as the postpartum oral challenge site noted high rates of patients being lost to follow-up and terminated their study protocol.

    Statistical analyses were completed through SAS version 9.4 (Cary, North Carolina). The chi-squared tests (or the Fisher's exact tests if the expected number <5) were used to compare categorical variables from both sites. The Wilcoxon rank tests were conducted to compare continuous variables for both sites. A p-value of less than 0.05 was used as the cutoff for two-sided significance.

    3 RESULTS

    During the study period, 276 pregnant patients completed penicillin allergy testing, with 207 in the antepartum oral challenge site and 69 in the postpartum oral challenge site. Table 1 summarizes the two sites' demographics and their allergy history. While some symptoms can be consistent with anaphylaxis, no patients explicitly stated that they had a previous anaphylaxis reaction to penicillin.

    TABLE 1. Demographics and allergy information from each study site.
    Antepartum oral challenge site (n = 213) Postpartum oral challenge site (n = 69)
    Gestational age at testing (weeks) 33.7 ± 2.6 35.8 ± 4.3
    Body mass index (kg/m2) 26.6 ± 5.1 30.1 ± 4.9
    Nulliparous patients 107/208 (51.4%) 29/63 (46%)
    Cigarette use 1/203 (0.5%) 4/58 (6.9%)
    Alcohol use 1/203 (0.5%) 6/58 (10.3%)
    Street drugs 3/184 (1.6%) 4/56 (7.1%)
    Gestational diabetes 30/190 (15.8%) 5/69 (7.3%)
    GBS at delivery 38/171 (22.2%) 16/56 (30.4%)
    Preterm delivery 14/183 (7.7%) 6/65 (9.2%)
    Fever in labor 16/189 (8.5%) 4/46 (8.7%)
    Delivery mode
    Cesarean section (elective) 31/189 (16%) 9/64 (14%)
    Cesarean section (emergency) 56/189 (30%) 8/64 (13%)
    Vaginal (assisted) 22/189 (12%) 8/64 (13%)
    Spontaneous vaginal delivery 80/189 (42%) 39/64 (61%)
    Years since penicillin allergy
    <5 years 12/199 (6%) 7/68 (10%)
    5–10 years 19/199 (10%) 8/68 (12%)
    >10 years 168/199 (84%) 53/68 (78%)
    Allergy reaction history
    Rash 102/213 (48%) 34/69 (49%)
    Urticaria 57/213 (31%) 18/69 (26%)
    Angioedema 12/213 (5.6%) 4/69 (5.8%)
    Breathing difficulties 8/213 (3.8%) 1/69 (1.5%)
    Vomiting 9/213 (4.2%) 3/69 (4.3%)
    Cannot recall 38/213 (18%) 6/69 (8.7%)
    PENFAST score
    0 169/213 (79%) 52/69 (75%)
    1 25/213 (12%) 12/69 (17%)
    2 14/213 (6.6%) 3/69 (4.3%)
    3 5/213 (2.3%) 2/69 (2.9%)

    3.1 Effectiveness in delabeling patients from penicillin allergy label

    Overall, 97% (201/207) of patients referred to the antepartum oral challenge site were delabeled from their penicillin allergy—compared to only 38% (26/69) of patients referred to the postpartum oral challenge site (RR: 2.58; 95% CI 1.90–3.49; p < 0.0001). At the postpartum oral challenge site, 63/69 (91%) of pregnant patients had a negative skin test and were cleared to obtain penicillin in labor if needed. Only 10/63 (16%) required penicillin antibiotics in labor—and the remaining 53 patients were invited to complete postpartum testing. Unfortunately, only 16/53 (30%) of individuals returned postpartum to complete their oral challenge, which were all negative (see Figure 1).

    Details are in the caption following the image
    Flowchart of penicillin allergy evaluation of pregnant women at our two sites.

    In the low-risk population, adverse events during oral challenges were rare in both antepartum and postpartum patients (2% vs. 0%). The reactions that occurred in the antepartum oral challenge site included two mild immediate cutaneous reactions and one episode of abdominal discomfort, which settled with antihistamines. No epinephrine administration was required. All three reactions during the oral challenge had a preceding negative PST due to pre-existing risk factors and did not receive a direct oral challenge.

    3.2 Patient's willingness and comfort in taking penicillin post-allergy visit

    There is a universally high degree of willingness to take penicillin for an infection after completing either a penicillin skin test (96%) or an oral challenge (99%). However, pregnant individuals completing antepartum oral challenges were more likely to express comfort in taking penicillin than individuals tested with an antepartum PST alone (99% vs. 91%; p = 0.0039).

    4 DISCUSSION

    Our study highlights that antepartum oral challenges are effective in delabeling penicillin allergies in pregnancy for low-risk individuals with low rates of adverse events. At a site where antepartum oral challenges are offered, the delabeling rate is 97%, whereas if oral challenges are only offered postpartum, the rate falls to 38%. This highlights that deferring oral challenges to the postpartum period leads to a significant loss of follow-up and failure to complete the allergy delabeling process. However, even in the postpartum oral challenge group, the antepartum skin tests allowed 14% of patients to receive appropriate antibiotics in labor.

    In pregnancy, cohort trials also demonstrated that oral challenges have low rates of adverse events, with successful delabeling reaching over 90% in all studies.6-8, 16 However, no previous trials had a comparator group between oral challenges and penicillin skin tests in pregnancy. Our study continues to highlight that antepartum oral challenges can successfully delabel most low-risk individuals. Reactions to oral challenges include mild cutaneous reactions and transient abdominal pain. Moreover, all patients with oral challenge reactions completed a PST prior, further reinforcing the limited sensitivity of a PST. The gold standard for penicillin allergy testing in low-risk individuals is shifting toward oral challenges rather than PST. From a safety perspective, for individuals with a PENFAST score lower than 3, a recent randomized controlled trial demonstrated that direct oral challenges are non-inferior to sequential PST followed by oral challenges.14 Direct oral challenges are less costly and more time-efficient in adult populations.17

    The postpartum oral challenge protocol was established as hospital administration was concerned about exposing pregnant individuals to a potential allergen during pregnancy—therefore, antepartum testing was limited to PSTs. However, our study highlights the impracticality of requesting postpartum patients to return to the clinic for further allergy testing. When oral challenges were postponed to postpartum, 70% of eligible individuals were lost to follow-up. This finding is particularly unfortunate, as all postpartum patients who returned to the clinic passed their oral challenges without any reactions. From the general obstetrical literature, up to 40% of postpartum patients miss their follow-up at the 6-week mark with their obstetrical care provider,18 so our loss to follow-up is not unexpected. Additional visits for penicillin delabeling will increase the healthcare system's costs and patients' out-of-pocket expenditures. The two-step allergy testing approach further disadvantages individuals with limited access to health care, including those with lower socioeconomic status and those in remote communities. Our postpartum oral challenge site is the only center in the province to offer antenatal allergy testing, serving a population of over 1.4 million, including rural and remote communities. Providing a single-visit option for delabeling penicillin allergies will facilitate more equitable access to penicillin allergy testing.

    Aside from the inability to fully delabel patients from their penicillin allergy, there are harms associated with this failure, specifically cost and patient satisfaction. In a publicly funded healthcare system, there is a substantially higher cost to complete a PST for every patient regardless of their history, as skin testing reagents and materials are costly. Successfully delabeling someone from their penicillin allergy during pregnancy is also linked with lifetime cost savings of CAD $58 780.9 Our post-visit survey also demonstrated that patients are more comfortable taking penicillin after an oral challenge than a PST. A major determining factor for the sustainability of allergy delabeling is the patient's comfort in re-exposure to penicillin in the future.19

    While including both sites allows us to compare two distinct allergy testing protocols, it also led to some limitations. Given that this is not a controlled trial, there are differences between the populations in different provinces. However, the baseline allergy risk between both groups is comparable—and the differences in demographics are not expected to impact our outcome. The number of participants at each site is also not balanced. One of the reasons is that the antepartum oral challenge site started its allergy program first. Additionally, the postpartum oral challenge site noted the high number of patients lost to follow-up, which prompted a change in their allergy testing strategy to include antepartum oral challenges. Further research in penicillin allergy testing for pregnant patients should focus on scaling up penicillin allergy testing in pregnancy. Specifically, given the limitations of allergists in Canada, other care delivery models should be examined. For example, examining care models where alternative care providers such as obstetricians, pharmacists, and/or family physicians perform oral challenges in pre-screened low-risk pregnant individuals should be considered. Lastly, we had low rates of penicillin allergy delabeling by history alone compared to non-obstetrical literature.20 Since delabeling by history is understudied in obstetrical literature, our providers and patients were likely more conservative in our approach. We advocate that further research on the safety of delabeling penicillin allergies by history alone in the pregnant population should be completed.

    5 CONCLUSION

    Antepartum oral challenges are more effective in delabeling a penicillin allergy in pregnant individuals than a two-stage protocol involving antepartum skin testing followed by a postpartum oral challenge. The postpartum oral challenge relies on individuals to return, which leads to a loss of follow-up. As allergy testing is moving toward oral challenges, pregnant patients should not be left behind in this advancement.

    AUTHOR CONTRIBUTIONS

    All authors meet the International Committee of Medical Journal Editors (ICMJE) authorship criteria. Jeffrey Man Hay Wong is the article's primary author, who contributed to research design, statistical analysis, manuscript drafting, and editing. Xiaoqing Liu was involved in completing the statistical analysis and editing the manuscript. Raymond Mak, Stephanie Erdle, and Julianne van Schalkwyk helped collect patient data, reviewed the data analysis, and edited the manuscript. Melissa Watt, Sudharsana Rao Ande, and Dozie Ocholor were the study administrators involved in curating the data, collecting data, and editing the manuscript. Chelsea Elwood and Vanessa Poliquin are both the senior authors of the article and involved with the conception of the research topic, research design, assembling the research team, funding acquisition, and reviewing the analysis and manuscript. Chelsea Elwood and Vanessa Poliquin both oversaw Jeffrey Man Hay Wong's work as the primary author.

    FUNDING INFORMATION

    Health Sciences Center Foundation and the B.C. Women's Hospital and Health Center. Jeffrey Man Hay Wong receives salary support from the Canadian HIV Trials Network, as part of his Postdoctoral Research Fellowship Award.

    CONFLICT OF INTEREST STATEMENT

    None.

    ETHICS STATEMENT

    The Manitoba study site received the University of Manitoba Research Ethics Board approval to complete this study (HS23603) on May 26, 2020, and the British Columbia site received approval from the University of British Columbia Children's & Women's Research Ethics Board (H21-00508) on August 5, 2022. The Manitoba approval required informed consent, which was obtained from each patient. The British Columbia approval deemed the project a quality improvement study, where informed consent was determined to be unnecessary.