Antenatal penicillin allergy evaluation is safe
Most pregnant patients who report penicillin allergy can be safely cleared to receive penicillin after evaluation and, when appropriate, supervised testing.
*Systematic review of cohort studies; Level 2a (OCEBM).

Citation

Reddeman L, Lim JWJ, Murphy KE, Fahmy D, Walsh C, Harris K. Safety, effectiveness, and acceptability of antenatal penicillin allergy evaluation: a systematic review. American Journal of Obstetrics & Gynecology. 2026; https://doi.org/10.1016/j.ajog.2025.08.033.

Background

Many pregnant patients report a penicillin allergy, but most are not truly allergic. Avoiding penicillin can lead to less effective antibiotics and worse outcomes; this review summarizes evidence on testing during pregnancy.

Patients

Pregnant people with an unverified self-reported penicillin allergy (19 studies; N=2085). Excluded: desensitization for confirmed allergy; case reports; studies with <5 participants; many studies excluded prior severe skin reactions and some excluded recent or high-risk reaction histories.

Intervention

Antenatal penicillin allergy evaluation, usually history review plus supervised testing (skin test and/or supervised penicillin dose by mouth).

Control

No evaluation or alternate evaluation approach (limited comparative data).

Outcome

Delabeling rate, patient acceptance, allergic reactions during testing, and pregnancy-related adverse events.

Follow-up Period

Immediate observation during testing (often ≥60 minutes); delayed reactions tracked variably by study.

Results

Outcome Overall finding
Penicillin allergy label removed (primary) 1956/2085 (93.81%)
Accepted evaluation when referred 1038/1705 (60.88%)
Mild immediate reaction during testing 17/1841 (0.92%)
Mild delayed reaction after testing 8/1841 (0.43%)
Severe allergic reaction during testing 3/1841 (0.16%): 2 severe allergic reactions (anaphylaxis) and 1 mild liver inflammation; no hospital transfers
Pregnancy-related harms linked to evaluation None reported

Limitations

All included studies were cohort studies (no randomized trials), with varied testing methods and inconsistent reporting of longer-term pregnancy and newborn outcomes. Referral and participation may favor lower-risk, better-resourced patients, limiting generalizability.

Funding

No funding reported; authors reported no conflicts of interest.

Clinical Application

Offer penicillin allergy evaluation during pregnancy (ideally one-visit models) to expand safe penicillin use; ensure testing occurs where severe reactions can be promptly treated.