couple patients receiving positive pressure ventilation experienced marked gaseous abdominal distension.

couple patients receiving positive pressure ventilation experienced marked gaseous abdominal distension. Analysis of gases from the stomach, ventilator, and play air suggested that the gastric gases came from the ventilator in the same patient. The diagnosis of tracheosophageal fistula was confirmed by the agency of esophagoscopy. Analysis of gases in the other patient did not support the suspicion of tracheoesophageal fistula, and no fistula was build at autopsy. The technique of gastric air analysis is at handed as a simple supporting tool for the clinical diagnosis of tracheoesophageal fistula in patients reveiving positive press ventilation.

Tracheoesophageal fistula usually is demonstrated through radiologic contrast study or on direct endoscopic visualization.[1-3] However, these studies may be cumbersome in unstable patients who are intubated for positive constraining force ventilation. In this report, a novel bedside technique to support the diagnosis of tracheoesophageal fistula is described.

CASE REPORTS

CASE 1

A 75-year-old woman with COPD experienced protracted respiratory failure following cholecystectomy. Several attempts at weaning from mechanical ventilatory support failed, and repeated endotracheal intubation was required. Tracheostomy was performed upon the 24th hospital day.

In the following week, she bring to maturityed pneumonia. Methylene blue tinge instilled inthe nasogastric tube appeared in the endotracheal tube. She was transferred to our hospital in succession the 32nd day for evaluation and management of a presum acquired tracheoesophageal fistula.[4,5] Flexible fiberoptic bronchoscopy demonstrated the fistula at the posterior trachea, beginning at the even of the sternal notch and extending 35 to 4 cm to approximately 2 cm above the carina. Surgical repair was accomplished by way of a transcervical approach.

Postoperatively, respiratory failure to be ascribed to severe COPD persisted. Numerous attempts at weaning from the ventilator riseed in only brief periods of independent ventilation. Esophagoscopy at sum of two units and five months after the surgical repair showed no evidence of returning fistula. However, six month following the fistula repair, persistent abdominal pain and bloating cause to growed and chest and abdominal radiographs demonstrated marked gaseous distension of the stomach and intestines (Fig 1) Barium contrast studies and colonoscopy exclud obstruction. These observations glance ated a recurrence of the tracheoeosphageal fistula,[6] which allowed air to leak unde positive constraining force (peak inspiratory pressure was 28 to 43 cm [Hsub]O]) from the trachea into the esophagus and stomach.[7,8]

To support the suspicion of intermittent tracheoesophageal fistula, specimens of gas from the ventilator, the patient's stomach (via gastrostomy tube), and the space air adjacent to the patient's orifice were obtained simultaneously and analyzed for [Osub2] and [COsub2] The tracheostomy box pressure was maintained at 20 mm Hg during sampling to minimize buffet leak and swallowing of respiratory gases. The inspired gas was analylzed with an MMC Horizon analyzer (Sensor Medics, Anaheim, Calif) in line with the inspiratory conduit of the ventilator circuit. Gastric and stead air samples were analyzed for [Osub2] and COsub2] in succession a Beckman Infrared Analyzer OM 11 (Fullerton, Calif).

The fraction of [Osub2] in gastric air was 038 and the fraction in inspired air was 039 (Table 1) The fraction of [Osub2] in the swing air adjacent to the patient's aperture was 0.21. The fraction of [Osub2] in the range air adjacent to the patient's inlet was 0.21. The [PaO.sub.2] was [tilda]92 mm Hg and [PaCO.sub.2] [tilda]56 mm Hg These data glance ated that the source of gastric air was the ventilator rather than swallowed apartment air. Recurrent tracheoesophageal fistula was confirmed according to esophagoscopy, at the time of which the tracheostomy beat was visualized protruding into the esophagus 2 to 3 cm below the upper esophageal sphincter.

CASE 2

A 47-year-old man with AIDS expanded adult respiratory distress syndrome complicating Pneumocystis carinii pneumonia, cytomegalovirus pneuomonia, and Salmonella and Candida septicemia. Respiratory failure necessitated protracted mechanical ventilation, which was initiated onward hospital day 3. Candida esophagitis was suspected because of the demeanor of oral thrush and Candida septicemia. Abdominal distension evolveed in hospital day 5, and tracheoesophageal fistula was suspected. Specimens of gas from the ventilator, patient's stomach (via nasogastric tube), and the scope air adjacent to the patient's cavity between the jaws were analyzed, just as in case 1 Tracheal blow pressure was 16 cm [Hsub2]O The fraction of [Osub2] in gastric air was 017 and the fraction in the inspired air was 070 (Table 1) The fraction of [Osub2] in air adjacent to the patient's opening was 0.20. The patient's [PaO.sub.2] was 73 cm Hg and the [PacO.sub.2] was 38 mm Hg These data did not support the suspicion of tracheoesophageal fistula. More invasive diagnostic practices were precluded by hemodynamic instability. Abdominal distension was attributed to ileus. Progressive multisystem failure l to death. Absence of a tracheoesophageal fistula was confirmed at autopsy.

...