Share this article on:

Is Esophagogastroduodenoscopy Necessary in all Caustic Ingestions?

Gupta, Sandeep K.; Croffie, Joseph M.; Fitzgerald, Joseph F.

Journal of Pediatric Gastroenterology and Nutrition: January 2001 - Volume 32 - Issue 1 - p 50-53
Original Articles

Background It is unclear whether symptoms alone can identify patients with caustic ingestion who will benefit from esophagogastroduodenoscopy (EGD). The published data are contradictory. The purpose of the current study was to determine the relationship between initial symptoms and EGD findings in patients with caustic ingestion.

Methods Chart review of all caustic ingestions who underwent EGD during a 4-year period (December 1993 through November 1997).

Results Twenty-eight patients (15 girls; mean age, 2.7 years (range, 0.92–13.33) underwent EGD after caustic ingestion. Fourteen percent (4/28) of patients were asymptomatic, and findings on endoscopy were normal. Another 57% (16/28) had normal endoscopic findings, although all were symptomatic. Twenty-nine percent (8/28) of patients had esophageal injury on EGD, and all were symptomatic. Esophageal injury was graded as 1 (mucosal erythema), 2 (superficial burns; noncircumferential) or 3 (deep burns; circumferential). The injury was grade 1 in three of eight patients and grade 2 in two; all had one symptom each. Grade 3 injury was found in three of eight patients: two had two symptoms (drooling and vomiting, drooling and stridor), and one had one symptom (dysphagia). All patients with grade 3 injury subsequently underwent esophageal dilations. Follow-up information was secured for two of the three patients with grade 1 injury and both patients with grade 2 injury at 34.3 months (range, 24–50) after the ingestion, and all were asymptomatic. Of the 20 patients with absence of esophageal mucosal damage, follow-up data were available for l5 patients at 37.2 months (range, 7–63) after the event and all were well.

Conclusions All patients with clinically significant injury (grades 2 and 3) were symptomatic at initial assessment. No single symptom or combination of symptoms could identify all patients with esophageal injury. All asymptomatic patients had normal findings on endoscopic examinations. Esophagogastroduodenoscopy seems unnecessary in asymptomatic patients with alleged caustic ingestion. A larger, prospective study would be necessary to unequivocally answer this clinically important question.

Department of Pediatrics, Indiana University School of Medicine, James Whitcomb Riley Hospital for Children, Indianapolis, Indiana, U.S.A.

Received April 25, 1999;

revised February 3 and July 18, 2000; accepted July 21, 2000.

Presented in part at the Annual Meeting of the American Society for Gastrointestinal Endoscopy, New Orleans, Louisiana, May l998.

Address correspondence and reprint requests to Dr. Sandeep K. Gupta, Division of Pediatric Gastroenterology, Hepatology, and Nutrition, James Whitcomb Riley Hospital for Children, Indiana University School of Medicine, 702 Barnhill Drive, Room l740-G, Indianapolis, IN 46202-5225, U.S.A.

This article is accompanied by an editorial. Please see Kay and Wyllie, Caustic ingestions and the role of endoscopy. J Pediatr Gastroenterol Nutr 2001:32: 8–10.

Ingested caustic substances can cause esophageal injury. The resultant injury can be nonsignificant, life-threatening (perforation), or chronic (stricture-forming) (1). The most efficient technique for the evaluation of the integrity of the upper gastrointestinal tract mucosa after caustic ingestion is esophagogastroduodenoscopy (EGD) (2–4).

The indications for EGD in the evaluation of children with caustic ingestions are debated. It is generally agreed that the absence of oropharyngeal lesions does not completely exclude esophageal or gastric mucosal injury (5–7). Symptoms reported after caustic ingestion include nausea, vomiting, dysphagia, refusal to drink, drooling, abdominal pain, and stridor. The issue of whether symptoms can identify the patients who would benefit from EGD is unsettled. Published studies are contradictory. Gaudreault et al. (1) concluded that symptoms do not predict the presence or severity of esophageal damage, whereas Crain et al. (8) reported that the presence of two or more symptoms (vomiting, drooling, or stridor) is a better predictor of esophageal injury than only one symptom.

We conducted this study to determine the relationship between symptoms and EGD findings in children after caustic ingestion and to determine whether EGD is mandatory in all cases. We obtained follow-up data to examine long-term outcomes.

Back to Top | Article Outline


Hospital charts of all patients with alleged caustic ingestion who underwent EGD during a period of 4 years (December 1993 through November 1997) were reviewed. Data were collated on history, physical examination, and endoscopic findings. All study patients were evaluated by the Division of Gastroenterology, James Whitcomb Riley Hospital for Children, Indiana University Medical Center, Indianapolis, Indiana. The EGD was performed with a fiberendoscope within 24 hours of arrival at our institution. In patients with body weight less than 12 kg, a GIF-XP20 endoscope (Olympus America, Inc., Melville, NY, U.S.A.) was used, and in larger patients the GIF-100 Videoscope (Olympus) was used. Procedure reports and photographs from these EGDs were reviewed to document the appearance of the esophageal mucosa. Abnormal mucosal findings were classified as grade 1 (mucosal erythema), grade 2 (ulcerations with necrotic tissue and white plaques that were less than circumferential), or grade 3 (ulcerations, white plaques, and sloughing of the mucosa in a circumferential pattern) (9).

Patients were assessed a mean of 3 years after the ingestion to determine long-term outcome and possible occurrence of unexpected complications. They were initially contacted by telephone and mail. If these failed, the referring doctor was contacted for follow-up information.

Back to Top | Article Outline


Twenty-eight patients underwent EGD after caustic ingestion. Of these, 15 were girls (15/28; 54%). The mean age at ingestion was 2.7 years (range, 0.92–13.33). A variety of substances were ingested, with lye-containing cleaning agents being the most common. Other ingested items included bleach, hair relaxant, and hair dye.

All patients were brought to our institution within hours of ingestion, except one who was referred 48 hours after the incident. The majority of patients (24/28; 86%) were symptomatic at initial examination. The most common symptom was drooling (18/24 patients; 75%). Other symptoms were poor oral intake (n = 4), vomiting (n = 3), and stridor or wheezing (n = 3).

All patients underwent EGD within 12 to 24 hours of initial examination. All four asymptomatic patients (4/28; 14%) had normal findings on EGD. Another 16 symptomatic patients (16/28; 57%) had normal findings. Of those, l5/l6 had one symptom; and l/l6 had two symptoms (vomiting and stridor). Esophageal mucosal injury was documented in eight patients (8/28; 29%), and all were symptomatic at initial examination. Grade 1 injury was noted in three patients (3/8; 37.5%) and grade 2 in two (2/8; 25%). They exhibited a single predominant symptom (drooling in four and poor oral intake in one). Grade 3 injury was recorded in the remaining three patients (3/8; 37.5%). Two had two symptoms (drooling and vomiting; drooling and stridor), and one had a single symptom (poor oral intake). This last patient was transferred to our institution 48 hours after ingestion. The transfer was prompted by the symptom of inability to drink liquids.

One each of the three patients with grade 1 injury had ingested hair bleach (mild alkali [MA]), liquid ammonia (MA), and an alkali cleaner (strong alkali [SA]). Of the two patients with grade 2 injury, one had ingested a degreaser (SA) and the other a hair relaxant (SA). All three patients with grade 3 injury had ingested lye-containing compounds (e.g., drain opener; SA). Of the four asymptomatic patients with normal EGD, two had ingested bleach (MA), while one each had ingested a highly corrosive liquid hardener (SA) and a paint remover (MA). The l6 symptomatic patients with normal EGDs had ingested a variety of substances, including hair relaxant (SA; n = 4), household bleach (MA; n = 3), and lye-based cleaner (SA; n = 2). Of the remaining 7 of these l6 patients, one each ingested a dishwasher drying booster (MA), ammonia detergent (MA), hair bleach (MA), industrial strength detergent (SA), boric acid (mild acid), KOH-based detergent (SA), and oven cleaner (SA). The ingested items generally were in liquid or gel form.

All 20 patients with a normal endoscopic appearance of the esophageal mucosa were discharged home after the endoscopy. Follow-up information could be obtained in 75% (l5/20), and all were well at 37.2 months (range, 7–63) after the event. Two of the three patients with grade 1 injury and both patients with grade 2 injury were reached 34.3 months (range, 24–50) after the ingestion and were asymptomatic. All three patients with grade 3 injury needed multiple esophageal dilations after the initial insult. Overall, follow-up information was obtained for 79% (22/28) of patients.

Back to Top | Article Outline


Earlier literature promotes prompt endoscopic evaluation of all patients with alleged caustic ingestion (1–4,10). Recently, various investigators have questioned this blanket recommendation (11–14). In 1992, Gorman et al. (11) published results of a 2-year prospective study of caustic ingestions conducted in seven poison centers around the United States. The study included 336 patients (age range, 3 months to 80 years) of which 88 were referred for esophagoscopy. They concluded that endoscopy was not indicated in asymptomatic patients after unintentional ingestion. In 1994, Nuutinen et al. (12) retrospectively reviewed their experience in 98 patients with caustic ingestion, during a 14-year period. They too concluded that EGD was not routinely necessary after caustic ingestions. Rather, the decision on EGD should be based upon presence of drooling and dysphagia during follow-up of 1 to 2 days. It should be noted that this study was conducted in Finland where “lye has not been sold . . . without special authorization since 1969.” In 1995, Christesen (13) published a retrospective study of 115 children with caustic ingestion during a period of 18.5 years and concluded that endoscopy is not needed in children who are asymptomatic after caustic ingestion but should be performed in symptomatic children, especially after ingestions of lye or ammonia water. In 1997, Cox and Eisenbeis (14) reported a study in which they examined the utility of endoscopy in 26 children after hair relaxant ingestion and opined that endoscopic examination is not indicated in asymptomatic patients.

Contrary to these reports are the results of a multicenter retrospective study by Gaudreault et al. (1), published in 1983. They evaluated the findings in 378 patients with caustic ingestion evaluated over a period of 11 years. Of the 313 patients with grade 0 to 1 esophageal injury, 70 (22%) were asymptomatic. Significant esophageal damage was noted in 65 patients, of whom 10 (15%) were asymptomatic. In fact, an esophageal stricture developed in one asymptomatic patient after negative esophagoscopy, which is difficult to reconcile in a cause-and-effect manner. They opined that absence of symptoms does not exclude significant esophageal damage.

We were prompted to conduct this study because we felt that a number of our patients with caustic ingestion demonstrated normal findings on EGD, even though the rate of abnormal EGD findings in our study (28.5%) is similar to that in other emergency room–based studies (1,8,9,11,15). The EGDs were performed on the recommendation of our local poison center, which takes into consideration the ingestant and patient's symptoms. All our asymptomatic patients had normal endoscopic findings, supporting the findings of others (8,11,13, 14). In concert with the findings of Crain et al. (8) and Gorman et al. (11), all our patients with clinically significant esophageal injury (i.e., grade 2 or 3) were symptomatic. There was, however, no single symptom or group of symptoms that identified all patients with potentially serious esophageal injury, unlike the report by Crain et al. (8). One patient with grade 3 injury to the esophageal mucosa was referred 48 hours after ingestion because of refusal to drink. This symptom strongly suggests significant esophageal mucosal injury (1).

It has been questioned whether EGD should be performed in the immediate postingestion period or when there is persistence of symptoms over the next 1 to 2 days (12). There is no definitive therapy available for patients with significant esophageal injury, and the available data on the use of steroids in this setting is contradictory (9,16). On the other hand, 15% of patients with significant esophageal damage reported by Gaudreault et al. (1) were asymptomatic early. Endoscopy within 12 hours is currently recommended (17).

We have not reported in detail the relationships between the findings on physical examination and EGD. Our data are in keeping with the impression that absence of oropharyngeal lesions does not exclude esophageal or gastric mucosal disease (5–7). We had patients with normal findings on oropharyngeal examinations and grade 2 and 3 mucosal injury.

A shortcoming of our study is the relatively small number of patients, especially in the asymptomatic group. By itself, it would seem hasty to conclude that children who are asymptomatic after caustic ingestion should not undergo EGD. That our findings are in concert with those of other recent reports (8,11–14) may dilute this criticism. The crux is how to separate the symptomatic patient who will have normal EGD findings from one who will have abnormal findings. Our data are not able to differentiate these groups. Lastly, although the entire esophagus was endoscopically examined in all our patients, the depth of esophageal injury is difficult to assess precisely, and this could, theoretically, result in misgrading. This would be limited to grades 2 and 3, and both these grades were taken to represent significant injury.

We conclude that asymptomatic patients with alleged unintentional caustic ingestion do not necessarily have to undergo EGD. Endoscopic examination should be performed when symptoms are present. A prospective, controlled study, possibly multicenter to capture enough patients, is needed to further examine the issue of evaluation of these children. Further, the endoscopic findings with different ingestants should be documented.

Back to Top | Article Outline


The authors thank Vicki Haviland for her expert attention to this work; Donna Morley-Pitts, Stacie Spears, and Nicole Spears for their assistance in obtaining follow-up information; and James Mowry, Ph.D., Indiana Poison Center, for his help and personal communication.

Back to Top | Article Outline


1. Gaudreault P, Parent M, McGuigan MA, et al. Predictability of esophageal injury from signs and symptoms: a study of caustic ingestions in 378 children. Pediatrics. 1983; 71:767–70.
2. Ferguson MK, Migliore M, Staszak VM, et al. Early evaluation and therapy for caustic esophageal injury. Am J Surg. l989;157:116–20.
3. Moore WR. Caustic ingestions: pathophysiology, diagnosis, and treatment. Clin Pediatr. 1986; 25:192–6.
4. Friedman EM. Caustic ingestions and foreign bodies in the aerodigestive tract of children. Pediatr Clin North Am. 1989; 36:1403–10.
5. Yarington Jr, CT Heatly CA. Steroids, antibiotics, and early esophagoscopy in caustic esophageal trauma. NY State J Med. 1963; 63:2960–3.
6. Viscomi GJ, Beekhuis GJ, Whitten CF. An evaluation of early esophagoscopy and corticosteroid therapy in the management of corrosive injury of the esophagus. J Pediatr. 1961; 59:356–60.
7. Borja AR, Ransdell Jr, HT Thomas TV, et al. Lye injuries of the esophagus: analysis of ninety cases of lye ingestion. J Thorac Cardiovasc Surg. 1969; 57:533–8.
8. Crain EF, Gershel JC, Mezey AP. Caustic ingestions: symptoms as predictors of esophageal injury. Am J Dis Child. 1984; 138:863–5.
9. Anderson KD, Rouse TM, Randolph JG. A controlled trial of corticosteroids in children with corrosive injury of the esophagus. N Engl J Med 1990; 323:637–40.
10. Ashcraft KW, Simon JL. Accidental caustic ingestion in childhood—a review: pathogenesis and current concepts of treatment. Texas Med. 1972; 68:86–8.
11. Gorman RL, Khin-Maung-Gyi MT, Klein-Schwartz W, et al. Initial symptoms as predictors of esophageal injury in alkaline corrosive ingestions. Am J Emerg Med. 1992; 10:189–94.
12. Nuutinen M, Uhari M, Karvali T, et al. Consequences of caustic ingestions in children. Acta Paediatr. 1994; 83:1200–5.
13. Christesen HBT. Prediction of complications following unintentional caustic ingestion in children: is endoscopy always necessary? Acta Paediatr. l995;84:1177–82.
14. Cox III, AJ Eisenbeis JF. Ingestion of caustic hair relaxer: is endoscopy necessary? Laryngoscope. 1997; 107:897–902.
15. Adam JS, Birck HG. Pediatric caustic ingestion. Ann Otol Rhinol Laryngol. 1982; 91:656–8.
16. Howell JM, Dalsey WC, Hartsell FW, et al. Steroids for the treatment of corrosive esophageal injury: a statistical analysis of past studies. Am J Emerg Med. 1992; 10:421–5.
17. Gryboski JD. Traumatic injury of the esophagus. In: Walker et al, eds. Pediatric Gastrointestinal Disease. St. Louis: Mosby; 1996:430–53.

Caustic ingestion; Endoscopy; Pediatrics

© 2001 Lippincott Williams & Wilkins, Inc.