Agreement between upper endoscopy and esophagography in the diagnosis of megaesophagus in Chagas disease

Introduction: The diagnosis and classification of megaesophagus can be challenging in patients with Chagas disease. The present study aimed to evaluate the agreement between upper endoscopies and esophagographies for the diagnosis and classification of megaesophagus in Chagas disease. Methods: A cross-sectional study of 50 patients with Chagas disease with upper digestive symptoms was undertaken. Esophagography and upper endoscopy exams were performed to compare diagnoses. Statistical analysis included sensitivity and specificity used to evaluate the diagnostic accuracy of upper endoscopies, and measures of agreement: linearly weighted Kappa (κw) and Cohen`s classical Kappa (κ) coefficients with 95% confidence intervals (95% CI). Results: Twenty-three patients (46%) were diagnosed with megaesophagus by esophagography. The upper endoscopy sensitivity and specificity for megaesophagus diagnosis were 100% and 33.3%, respectively. Regarding megaesophagus classifications, there was a substantial agreement between the two exams (κw = 0.622; 95% CI: 0.498 to 0.746). Within megaesophagus groups, agreement for group I was slight (κ = 0.096; 95% CI: 0.000 to 0.403); for group II, substantial (κ = 0.703; 95% CI: 0.456 to 0.950); and for groups III and IV, inconclusive (κ = 0.457; 95% CI: 0.000 to 0.967; κ = 0.540; 95% CI: 0.035 to 1.000, respectively). Conclusions: Upper endoscopy has a high sensitivity, but a low specificity to diagnose megaesophagus. Agreement between the two exams varies depending on the megaesophagus grade. Thus, upper endoscopy can be used in the diagnostic workup of a possible Chagas disease megaesophagus, but all identified cases should undergo esophagography.


INTRODUCTION
Digestive involvement in Chagas disease may occur in up to 15 -21% of patients 1,2 .However, megaesophagus is often not diagnosed due to several reasons such as a lack of information regarding the patient, health professionals missing the symptoms of Chagas disease, nonspecific symptoms, and a lack of adequate diagnostic exams in health units.This diagnostic delay directly affects the post-surgical outcome since long-term outcomes worsen with the degree of stenosis and dilatation 3 .
Esophageal involvement in Chagas disease can be determined by esophagography with good accuracy according to the classification by Rezende 4 .The changes that may elucidate the diagnosis and degree of esophageal involvement include altered motility, delayed emptying velocity, changes in caliber, presence of air-fluid levels, and thinning of the esophagogastric junction.Therefore, the initial diagnostic workup of patients with a possibility of Chagas disease must include an esophagography, but in practice it is usually replaced by an upper endoscopy, which has some limitations especially for inexperienced practitioners 5 .Furthermore, the diagnostic accuracy of an upper endoscopy for megaesophagus is still unknown.However, an upper endoscopy is useful in the diagnosis of associated diseases such as candidiasis and to exclude the presence of malignancies.
The objectives of this study were to evaluate the accuracy of upper endoscopies for the diagnosis of megaesophagus and to evaluate the agreement between upper endoscopies and esophagographies for the classifi cation of megaesophagus in patients with Chagas disease.

Study population
This was a cross-sectional study of patients with Chagas disease who were followed at the Evandro Chagas National Institute of Infectious Diseases from the Oswaldo Cruz Foundation (INI/Fiocruz) and presented with continuous or intermittent dysphagia for liquids or solids.The diagnosis of Chagas disease was confi rmed by two simultaneous serological tests (an enzymelinked immunosorbent assay and an indirect immunofl uorescence assay).The study was approved by the Research Ethics Committee at the INI/Fiocruz under number 832.303.
Participants included in the study underwent a clinical evaluation performed by a single surgeon who assessed symptoms compatible with megaesophagus during medical appointments.Afterwards, patients underwent endoscopy, by a single endoscopist, followed by esophagography (6 months maximum interval between exams).
To date, there are no defined endoscopic criteria and parameters for the classifi cation of megaesophagus published in the literature.Therefore, to evaluate the degree of involvement, the reports of the Endoscopy Service were subjectively based on the following fi ndings: 1. Group I -Greater salivary retention, presence of reduced peristaltic waves and no dilatation; 2. Group II -Mild dilation with greater retention of liquids and possibly food residue; 3. Group III -Upstream dilatation with diffi culty of passage of the endoscope through the cardia; and 4. Group IV -Clear esophageal tortuosity mainly in the distal third of the esophagus.
The parameters used for assessment and classifi cation of megaesophagus during esophagography were determined according to the classifi cation of Rezende 4 as follows (Figure 1): 1. Group I -Esophagus apparently with a normal caliber to the radiological examination.Slow traffi c.Small retention on radiography taken one minute after ingestion of barium sulfate; 2. Group II -Esophagus with small to moderate increase in caliber.Appreciable contrast retention.F requent presence of tertiary waves, with or without association with hypertonia of the esophagus; 3. Group III -Esophagus with a great enlargement of diameter.Reduced motor activity.Hypotonia of the lower esophagus.Great contrast retention; and 4. Group IV -Dolicomegaesophagus. Esophagus with great retention capacity, stretched, bending over the diaphragmatic.

Statistical analysis
Endoscopic validity for detecting megaesophagus was evaluated through sensitivity and specifi city measurements, considering esophagography as the gold standard 6 .The analysis of agreement between esophagography and endoscopy was performed using the linearly weighted Kappa (κ w ) 7 , since they have an ordinal scale with 5 levels (normal and groups I, II, III, and IV).When there are 5 categories, the weights in the linear set used to calculate κ w , are: 1, 0.75, 0.50, 0.25, and 0 when there is a difference of 0 (total agreement) or 1, 2, 3, and 4 categories between esophagography and endoscopy diagnostics, respectively.Cohen´s classical Kappa (κ ) coeffi cient was used for concordance analysis, by collapsing the original 5 x 5 contingency table into fi ve 2 x 2 tables where each category is compared with all others 8 .We used the scale presented in the work of Landis and Koch for interpretation of all Kappa coeffi cients where 0 to 0.20 is slight, 0.21 to 0.40 is fair, 0.41 to 0.60 is moderate, 0.61 to 0.80 is substantial, and 0.81 to 1.00 is perfect 9 .The 95% confidence intervals (95% CI) for κ w and κ coefficients were also presented.The analyses were performed using the R language, the vcd package, and the Kappa and confint functions.

RESULTS
Altogether 50 patients (33 women and 17 men) with a mean age of 49 years with Chagas disease and dysphagia symptoms were analyzed.Twenty-three patients (46%) were diagnosed with megaesophagus by esophagography.Since all patients with megaesophagus were detected by both exams, the upper endoscopy sensitivity for megaesophagus diagnosis was exactly 100%.On the other hand, only 9 patients (out of twenty-seven) were considered negative for megaesophagus by both methods, resulting in 33.3% specificity.Among 18 normal subjects by esophagography, 17 were misdiagnosed with grade I megaesophagus and 1 subject was misdiagnosed with grade II megaesophagus by upper endoscopy (Table 1).

DISCUSSION
Chagas disease is a chronic and progressive disease and serious lesions can affect the digestive tract 10 .However, most health professionals are not appropriately trained in the clinical management, diagnosis, and treatment of this condition as it is a neglected disease.Thus, delay or even non-diagnosis can allow irreversible digestive tract injuries requiring more invasive operations as the disease progresses.
Chagas megaesophagus is the most common form of gastrointestinal involvement, affecting any age, sex, and stage of the disease.Initial symptoms may be quite nonspecific such as nocturnal coughing, hypersalivation, coughing sensation after eating, and weight loss which further complicates the diagnosis 11 .More classic signs and symptoms include dysphagia for liquids or solids, weight loss, odynophagia, choking, and even recurrent pneumonia making a diagnostic workup mandatory in such cases 12 .
The diagnostic assessment usually begins with an upper endoscopy due to the availability of this exam in most hospitals, but this is an operator-dependent examination without a standardized, validated grading scale for megaesophagus.However, symptomatic Chagas disease patients must undergo this exam since diseases associated with or mimicking achalasia, such as neoplasms, infections, and other mucosal changes, need to be excluded.Although the use of endoscopy, along with high-resolution manometry for the diagnosis and classification of achalasia has been proposed to evaluate patient candidates for peroral endoscopic myotomy 13 , there is no endoscopic classification to assess the degree of megaesophagus due to Chagas disease by endoscopic methods.
In our study, we found that upper endoscopy has a high sensitivity for the diagnosis of chagasic megaesophagus, but a very low specificity.This may be due to the criteria we used to diagnose and estimate the classification of megaesophagus by endoscopy which includes the absence of peristalsis during the examination, non-relaxation of the lower esophageal sphincter, non-passage of the endoscope to the stomach, salivary or food retention, and dilation.During the endoscopy, a moderate amount of air is injected so that there is sufficient space for the examination, which will slightly distend the organ.In addition, the image generated by the device is slightly increased and can result in an inaccurate diameter estimate of the megaesophagus.The endoscopy device itself is thicker than the contrast used to do the esophagography and therefore more difficult to get in the stomach.All of these factors might justify the high number of false positives on endoscopy seen in megaesophagus and the low concordance between an upper endoscopy and an esophagography for megaesophagus grade I classifications.Therefore, whenever Chagas disease megaesophagus, especially grade I, is diagnosed using endoscopy, the diagnostic workup should include an esophagography in order to avoid misdiagnoses.Moreover, manometry should also be included in the diagnostic workup as it has the sensitivity to identify the problem according to its pathophysiology (nerve and muscle injury of the lower esophageal sphincter) and is considered a diagnostic gold standard 14 .Manometry is essential for pre-operative evaluations since the type of operation is defined depending on the function of the esophagus thus avoiding very tight anti-flow valves that lead to dysphagia or incompetent valves that will lead to gastroesophageal reflux after myotomy 15 .However, many health institutions do use manometry, which delays a correct diagnosis.
Based on our findings, esophagography should still be the exam chosen to classify megaesophagus, as the concordance between exams for megaesophagus classification varies from slight to substantial depending on the megaesophagus group.The dynamics and the number of variables evaluated by esophagography provide more information than just an upper endoscopy.With real-time radioscopy or recorded movies, esophagography allows visualization of important changes such as motility disorders, the presence of tertiary waves, delayed emptying of the esophagus, changes in caliber, the presence of barium and liquid levels, and the presence of the "bird's beak" signal, which is a distal and symmetrical tapering contrast sign 16 .Nonetheless, in cases of megaesophagus groups III or IV, the endoscopist might have a higher percentage of correct grading responses as some parameters are better identified using endoscopy such as lack of peristalsis in the esophageal body, tortuosity, large amounts of retained food, and the non-passage of the endoscope to the stomach.

CONCLUSION
Upper endoscopy has a high sensitivity, but a low specificity for the diagnosis of megaesophagus due to Chagas disease in comparison to esophagography.Moreover, the agreement between endoscopy and esophagography for megaesophagus classifications has a wide variation depending on the megaesophagus grade.Therefore, an upper endoscopy can be used in the diagnostic workup of Chagas disease patients with possible megaesophagus, but confirmatory exams are needed whenever a case is identified, especially, if the case is classified as a grade I or II by endoscopy.Therefore, after a possible megaesophagus is identified using endoscopy, esophagography should also be performed to confirm the diagnosis and grade.