Defining the roles of computed tomography and esophagogastroduodenoscopy in patients with caustic ingestions

For many patients with small-volume ingestions of common household products, the injury will be nonexistent or minimal, and they can either be managed at home or be discharged quickly following a brief hospital evaluation. Those patients ingesting larger volumes or higher-risk caustics may require hospitalization or even intensive care unit admission. For a much smaller proportion of patients, the ingestion will result in full-thickness injury with perforation requiring immediate life-saving surgery. The epidemiology of caustic ingestions was defined decades ago and likely remains true today. Most ingestions occur in children, are unsupervised with no intended self-harm, and are relatively benign when compared to intentional ingestions in adults, which tend to be more severe and even fatal [1–3].

Distinguishing among at-risk groups of patients presents a unique clinical challenge. What is the optimal modality to define the extent of their injuries? Unlike dermal burns, gastrointestinal injury is more occult. Surrogate markers of injury based on physical findings or reported signs and symptoms offer some insight [4–6], yet none have sufficient positive or negative predictive value to be applied to all patient groups and all potentially caustic xenobiotics.

Esophagogastroduodenoscopy (hereafter endoscopy) following caustic ingestion was first reported in the 1950s [7] and was already considered standard care in the 1960s [8]. By the time computed tomography (CT) scanners were used in patients who ingested caustics [9], Zargar and colleagues [10] had already published an endoscopic grading system that was linked to clinical outcomes. This grading system is still used today to standardize prospective trials [11–13] and compare patients in large retrospective evaluations [14–16].

In 2010, when Ryu and colleagues [17] proposed a CT grading system in Clinical Toxicology, a new era of assessment and research began. In their retrospective evaluation, CT outperformed endoscopy for predicting esophageal strictures in a small sample of patients. Unfortunately, although the grading system proposed by Ryu and colleagues [17] uses the same basic nomenclature (grades I-IV) as the Zargar criteria [10], the same numerical grade in the two systems represents different injuries, and therefore, these grading scales are not interchangeable. Subsequent investigations demonstrated the advantages and limitations of both CT and endoscopy [18–20]. In order to decide which modality is best utilized for a given patient, those benefits and limitations need to be clearly defined in the context of diagnosis, intervention, prognosis, and the healthcare model (Table 1).

Table 1. Comparisons of the benefits and limitations of endoscopy and computed tomography for evaluating patients with caustic ingestions.

Technique	Benefits	Limitations
Endoscopy	Directly visualizes the mucosal surface	Cannot visualize the serosal surface
	Assigns a Zargar grade to the injury	May be incomplete secondary to the inability to safely view past a severe injury
	Identifies candidates for corticosteroid therapy	May require transfer, especially for small children
	Allows for insertion of a nasogastric tube or stent under direct visualization	Requires specialized personnel to obtain the images
	May be more available than computed tomography in some practice locations	May require airway protection and anesthesia as a prerequisite in some practice locations
Computed tomography	Visualizes the entire gastrointestinal tract	Cannot directly visualize the mucosal surfaces
	Wide rapid availability in most high-income countries	Cannot be used for Zargar grading for corticosteroid therapy
	Easily identifies early perforation	Cannot facilitate insertion of a nasogastric tube or stent under direct visualization
	Identifies candidates for emergent surgery	Exposes the patient to radiation
	No specialized personnel required to obtain the images	May require transfer in some practice locations

In this issue of Clinical Toxicology, Kaewlai and colleagues [21] provide a unique window into another limitation of the use of computed tomography in patients with caustic ingestions. Computed tomographic images were obtained from 17 patients, 16 of whom ingested acid caustics. The gold standard was defined as the presence of an abnormality as agreed upon by four radiologists in at least six of eight interpretation sessions. The study then compared the diagnostic performance of eight clinicians (six radiologists with varied training and two acute care surgeons) who assessed the images in two separate sessions without knowledge of endoscopic or surgical findings. Kaewlai and colleagues [21] demonstrate significant intra- and interrater variability, which was improved somewhat if only the interpretations of the six radiologists were included in the analysis. The range in overall accuracy was between 52.9% and 80.4%, depending on the segment of the gastrointestinal tract being evaluated. Unfortunately, there is no similar study that evaluates the variability in interpretations of endoscopic images from patients who have ingested caustics.

How do we translate this information into clinical practice? Like any test, the results of either CT or endoscopy need to be applied in the clinical context of a patient and modified based on the quality of the images and the skill of the person who interprets those images. It is clear that eliminating endoscopy as a diagnostic consideration leads to some unnecessary surgical interventions that have serious consequences [18,20,22]. Similarly, the limitations of endoscopy as a sole modality are well described [20,23,24]. At the present time, the only reasonable conclusion is that both modalities have a complementary role, as was recently defined by Chen and colleagues [25].

In a healthcare model that has access to both endoscopy and CT, the immediate availability and ease of identifying early perforation make CT preferable for patients in extremis. If the interpretation is ambiguous, then endoscopy can help clarify the status of the patient. In contrast are patients with lower-risk ingestions, such as the kind that typically occur in young children. For these patients, the clinical considerations involve allowing feeding with a progression to a normal diet, a decision to initiate corticosteroid therapy, and/or bypassing a significant esophageal injury with a stent or feeding tube. As such, the use of endoscopy is preferred as it will usually provide the needed answer and spare the child radiation exposure. If the results of endoscopy are incomplete or inconclusive, then CT imaging will likely clarify the missing information. In clinical settings where only one diagnostic tool is rapidly available, it seems reasonable to use that modality and make decisions about patient transfer based on the data obtained, the clinical status of the patient, and the risks and expense of transport.

The present study by Kaewlai and colleagues [21] has clearly informed us about the real-world clinical limitations of CT imaging for patients with predominately acid caustic ingestions. It should be reproduced with data from patients who ingested alkaline caustics and with a different group of radiologists. Likewise, a comparison of the ability of gastroenterologists to interpret and grade endoscopic images would be highly valuable. For the present time, it appears that pending these and further investigations, both imaging modalities provide valuable perspectives.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

The author reported there is no funding associated with the work featured in this article.