Case Presentation: A 75 year-old male with past medical history of aplastic anemia requiring immunosuppression and pRBC and platelet transfusions presented to the infusion center with hypotension. He was found unresponsive with BP of 59/36 prior to receiving blood products, and was transported to the ED after receiving 1L of normal saline.
On ED evaluation, the patient endorsed abdominal pain and bloody stools. History was notable for an allergy to penicillin described as hives. Vitals signs were 36.7C, BP 110/60, HR 58, RR 18, and physical exam showed slight abdominal distention without tenderness. Lab workup was notable for platelets of 8 x 109/L and white blood cell count of 2.3 x 109/L, with an absolute neutrophil count of 1.23 x 109/L.
The patient subsequently developed hypothermia, hypotension, and worsening mentation concerning for septic shock. He received 3L IV fluids and vancomycin. However, neither gram negative nor anaerobic bacterial coverage was ordered, due to indecision in the setting of charted penicillin allergy. Six hours after ED presentation, aztreonam was administered, and 21 hours after presentation anaerobic coverage with metronidazole was administered.
The patient developed worsening septic shock refractory to vasopressor support and was ultimately transitioned to comfort-focused care. Autopsy revealed findings consistent with neutropenic enterocolitis.
Discussion: Despite the patient’s immunocompromised state and sepsis from a likely intra-abdominal source, administration of antibiotics with appropriate coverage was delayed. This was multifactorial; however, his listed penicillin allergy likely contributed to his preventable death from neutropenic enterocolitis.
Penicillin allergy is common in the electronic medical record (~10-15%); yet, most patients (90-99%) with a chart history of penicillin allergy do not have true Type I hypersensitivity. This documentation inaccuracy is associated with longer hospital stays, higher rate of drug-resistant infections, and more surgical site infections.
Test-dose protocols offer an inpatient option for addressing charted penicillin allergy. In this protocol, patients with a reported history of itching, mild rash or unknown reaction to penicillin can be administered a small intravenous β-lactam test dose during a period of close bedside monitoring. If no reaction occurs, full dose is administered; the allergy can then be removed from the patient’s history should no reaction occur. This same test-dosing protocol can be utilized with a 3rd/4th generation cephalosporin in patients reporting a Type 1 hypersensitivity to penicillin.
Conclusions: Penicillin allergy is a common chart finding in hospital medicine, and inaccurate documentation can lead to poor outcomes and adverse events, as in this case. Type 1 hypersensitivity to β-lactams can be safely verified/refuted through inpatient test-dose protocols, and systems changes should be implemented to automate and support this diagnostic effort.