Hydrogen peroxide ingestions: The scope of the injury

To the Editor:

Ingestion of low concentrations of hydrogen peroxide (H₂O₂) is generally benign. However, serious toxicity such as venous/arterial gas embolism, hemorrhagic gastritis, and death, has been reported with higher concentrations.^1,2 A concentration of 35% or greater, sometimes used to treat various diseases such as viral infections, cancer, and Alzheimer's dementia, can easily be purchased at a health food store or online.³ There are limited data on the appropriate imaging studies and management strategies for suspected ingestions. We report H₂O₂ exposures over 10 years, managed by our regional poison center (RPC) in order to facilitate proper diagnostic workup.

After IRB approval, our RPC database was queried (Crystal Reports, Version 11) for all cases coded as H₂O₂ (January 1, 2001–December 31, 2010). A total of 760 cases of H₂O₂ exposure occurred during the study period. While most exposures were trivial gastrointestinal (GI) exposures (e.g. lick or taste), 23 were of interest. These cases were chosen to report because of suicide intent, exposure to 35% H₂O₂, or lower concentrations with high volumes (more than a mouthful) and/or symptomatology such as throat irritation, vomiting, or hematemesis. All patients of interest were evaluated in an emergency setting. The age range of these patients was 1–94 years (mean 38 years). The large majority were adults with only three 12-year-old patients. Volumes ingested ranged from mouthfuls⁷ to several gulps, swallows, or shots recorded. One patient reportedly ingested a “bottle” of 3% H₂O₂ with merely vomiting and abdominal pain, no further studies pursued (i.e. x-ray, CT scan, or EGD) and had a good outcome. Of our study population, 12 (52%) had oral exposure of < 35% H₂O₂ and 11 (48%) had ingested 35% H₂O₂. Ten (43%) patients had plain film x-rays (chest or abdominal x-ray) of which 9 (90%) were reported as normal with 1 (10%) reporting gastric distention (who ingested 35% H₂O₂). Eight of the 10 patients with x-rays, ingested 35% H₂O₂. One patient (who ingested 35% H₂O₂) had an abdominal computed tomography (CT) revealing portal venous air. This patient subsequently received hyperbaric oxygen therapy (HBOT) with complete resolution per repeat imaging. Finally, 9 (82%) patients who ingested 35% H₂O₂ received an esophagogastroduodenoscopy (EGD) of which 7 (78%) demonstrated merely mild inflammation in esophageal and/or gastric mucosa, and the other two were reported as normal. EGD was done in a subset of patients who ingested 35% H₂O₂ with symptoms such as throat irritation, vomiting, and/or hematemesis. EGD was recommended in three patients who did not have the procedure done either because of physician or patient refusal. No cerebrovascular accidents (CVAs) or deaths occurred in our study group.

Hydrogen peroxide can cause toxicity via three mechanisms: corrosive damage, lipid peroxidation, and oxygen formation.⁴ Once ingested, H₂O₂ readily reacts with tissue catalase found in mucous membranes, liver, kidneys, red blood cells, and bone marrow.⁴ This exothermic reaction rapidly decomposes H₂O₂ to water, heat, and oxygen. At standard temperature and pressure, 30 mL of 35% H₂O₂ will liberate 3.5 L of oxygen.⁵

Significant GI exposures to H₂O₂, like other caustic ingestions, have historically resulted in EGD being advocated to determine extent of mucosal injury. Likewise, plain films and CT scans have also been utilized to ascertain any abnormal accumulations of air. In our population, patients who consumed 35% H₂O₂ resulted in mild, if any, GI mucosal injury without significant morbidity. Likewise, another study reported a series of 11 patients, 10 of whom ingested 35% H₂O₂, with no significant GI morbidity (even in the context of hematemesis).² Additionally, all patients in this series had an abdominal CT and HBO therapy for treatment of hepatic gas emboli.² While HBOT is beneficial for CVA resulting from gas embolism, its role in gas embolism elsewhere remains questionable.² Our retrospective poison center study is limited in obvious ways. These data rely on voluntary reporting, which means cases could be missed. It is possible that patients with H₂O₂ ingestions having significant mucosal injury exists but, unfortunately, was not reported to our poison center. Additionally, cases are handled via telephone conferences, which at times could be second or third hand information and relies on correct and accurate coding from the poison center specialist. We solely relied on these patients’ records without actually visualizing the EGDs ourselves or speaking directly to the gastroenterologist who performed the study. Prospective data may help assign value to the various imaging studies regularly pursued in these patients. However, in light of the benign mucosal injury in our patient population as well as others, we have eliminated EGD as a standard diagnostic tool for future patients with GI exposure to H₂O₂.

Declaration of interest The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.