Management of caustic injury: better to evaluate patient as a “whole” rather than only through a “hole”!

The prediction of outcome of caustic ingestion is traditionally based on endoscopic grading [1]. Endoscopic grading predicts short-term outcomes such as systemic complications, including respiratory failure and immediate mortality, and long-term outcomes such as nutritional autonomy and future stricture formation [2]. However, endoscopic grading is inaccurate in predicting depth of necrosis, which leads to both overestimation and underestimation of severity resulting in either inappropriate emergency surgery or nonoperative management, respectively, jeopardizing optimal outcomes. Other drawbacks of endoscopic severity assessment are interobserver variation, inability to perform endoscopy in all cases, and risk of overestimation beyond 48 hours of corrosive ingestion owing to the presence of edema and submucosal hemorrhage [2].

To overcome the limitations of endoscopy, an endoscopic ultrasound (EUS)-based scoring system was proposed, the rationale being that assessment of transmural involvement and destruction of the esophageal muscle layer would predict stricture formation [3]. Although EUS could be performed safely within 24 hours of caustic ingestion, there is no added advantage over conventional endoscopy in predicting early or late complications. Rather, bronchoscopy could detect tracheobronchial involvement and hence is essential for surgical planning in cases of emergency surgery [4].

Therefore, composite scores incorporating clinical and endoscopic parameters were proposed to improve prognostic evaluation and triage of patients who have ingested a corrosive substance [5]. In this issue of Endoscopy, Tosca et al. propose a prognostic score based on analysis of nearly 450 patients prospectively included in derivation and validation cohorts over 22 years [6]. The score incorporates clinical parameters such as chemical nature of corrosive (acid), inflammatory markers (white blood cell count), and systemic involvement (metabolic acidosis indicating transmural involvement), as well as standard endoscopic parameters. This led to a higher accuracy than endoscopy alone (P = 0.04), with high sensitivity, specificity, positive and negative predictive values (93.3 %, 92.7 %, 72.7 %, 98.5 %, respectively), and area under the curve (AUC 0.976, 95 % confidence interval [CI] 0.973–0.979; P < 0.001) in predicting poor evolution, defined as intensive care unit admission, urgent surgery, or death, in the acute stage. This score is novel as it incorporates clinical, biochemical, and endoscopic indices for prediction of poor evolution in the acute stage.

“This study is the first step toward an algorithm for the prediction of outcomes of corrosive injury and it reiterates the importance of taking into account multiple factors rather than endoscopy alone.”

However, there are some caveats regarding use of this new prognostic model. First, the main predictive model has been derived based on both morphological (endoscopic changes) and functional (neutrophil percentage, which indicates inflammation, and bicarbonate, which indicates metabolic status). The functional parameters here, which are continuous variables, have been dichotomized into categorical variables based on receive operating characteristics (ROC). This is a standard method for deriving predictive scores. However, in the context of this study, the dichotomization could introduce variability in the performance characteristics (sensitivity, specificity, positive and negative predictive values) when applied to a larger population in real-world practice. This is because a bicarbonate level of say 15 mmol/L will indicate a higher degree of acidosis than a level of 20 mmol/L, even though both values are below the cutoff value. The same holds true for neutrophil percentage. Nevertheless, the fairly narrow 95 %CI in accuracy and ROC implies that the variability is unlikely to be affected substantially. Having said that, in a realistic sense, instead of a dichotomous grouping, a graded grouping of the continuous variables with smaller class intervals might still have been a better option. Second, although validation has been done, further validation in an Eastern population is warranted prior to use of this model, given that acidic corrosive injury is more common in Asia [1] [5]. Another drawback of the study is omission of computed tomography (CT) grading in this scenario, which is valuable in the acute setting. In fact, World Society of Emergency Surgery consensus conference (2015) recommended emergency CT scan along with endoscopy as important investigations in the management algorithm for caustic ingestion [7]. In patients with endoscopic grade IIIb caustic injury, the use of CT is helpful in selecting patients for emergency esophagectomy. This approach leads to improved survival and functional outcomes with a reduction in cost of the treatment. In this study, as the authors mention, the exclusion of CT grading could be due to the fact that many patients in real-life settings do not undergo CT because of the mild nature of the injury as ascertained by other variables. This scoring system could be helpful in triaging patients with predicted worse outcomes and selecting patients who would benefit from a CT scan. The drawbacks of the CT scan are difficulty in interpretation, risk of contrast-associated renal injury, scarcity of normative data in children with corrosive injury, along with radiation risk.

The majority ( > 80 %) of patients in the derivation and validation cohorts in the Tosca et al. study had endoscopic severity of grade II or less. Although missing value imputation was done in cases of variables with more than 5 % missing values, this is one of the drawbacks of such analysis, as the amount of missing data is variable for different predictors. This was obvious given the fact that not all investigations were conducted in mild cases.

Despite these limitations, the current study has several strengths. It reminds us of the importance of taking account of multiple factors ([Fig. 1]) when predicting poor outcomes in corrosive ingestion, rather than relying on endoscopy alone. Endoscopy, CT, and EUS findings, and clinical and biochemical parameters all have their inherent limitations when used alone; hence, a composite score incorporating all these factors could be additive in prediction of both short-term and long-term sequelae of corrosive injury that has substantial morbidity and mortality. This study by Tosca et al. is the first step toward such algorithms, which need to be incorporated into the management of corrosive injuries. The authors have done excellent work by simultaneously deriving predictors and validating them using a large number of patients who were evaluated prospectively over decades.

Fig. 1 Overall assessment of patients with caustic ingestion.

In summary, composite prognostic scores are better than endoscopy alone in predicting short-term adverse outcomes. However, the findings need to be validated in other populations. Such composite scores are complementary to cross-sectional and endoscopic imaging but could not replace them. Moreover, such scores could be useful in deciding which patients require urgent cross-sectional imaging (CT). It remains to be seen whether such scores could predict long-term outcomes such as stricture formation better than endoscopy or CT. More data and validation in other populations are required prior to incorporation of such composite scores into the evidence-based management algorithm of caustic injury.

Publication History

Article published online:
27 July 2021

• References

• 1 Zargar SA, Kochhar R, Mehta S. et al. The role of fiberoptic endoscopy in the management of corrosive ingestion and modified endoscopic classification of burns. Gastrointest Endosc 1991; 37: 165-169
• 2 Chirica M, Bonavina L, Kelly MD. et al. Caustic ingestion. Lancet 2017; 389: 2041-2052
• 3 Chiu HM, Lin JT, Huang SP. et al. Prediction of bleeding and stricture formation after corrosive ingestion by EUS concurrent with upper endoscopy. Gastrointest Endosc 2004; 60: 827-833
• 4 Benjamin B, Agueb R, Vuarnesson H. et al. Tracheobronchial necrosis after caustic ingestion. Ann Surg 2016; 263: 808-813
• 5 Rigo GP, Camellini L, Azzolini F. et al. What is the utility of selected clinical and endoscopic parameters in predicting the risk of death after caustic ingestion?. Endoscopy 2002; 34: 304-310
• 6 Tosca J, Villagrasa R, Sanahuja A. et al. Caustic ingestion: development and validation of a prognostic score. Endoscopy 2021; 53: 784-791
• 7 Kluger Y, Ishay OB, Sartelli M. et al. Caustic ingestion management: World Society of Emergency Surgery preliminary survey of expert opinion. World J Emerg Surg 2015; 10: 48 Erratum in: World J Emerg Surg 2015; 10: 56