Anesth Analg 2002;94:1534-1536
© 2002 International Anesthesia Research Society
TECHNOLOGY, COMPUTING, AND SIMULATION
An Unusual Presentation of End-Tidal Carbon Dioxide After Esophageal Intubation
Paul B. Bigeleisen, MD
Department of Anesthesiology, University of Rochester School of Medicine and Dentistry, Strong Memorial Hospital, Rochester, New York
Address correspondence to Paul E. Biegeleisen, MD, Department of Anesthesiology, University of Rochester School of Medicine and Dentistry, Strong Memorial Hospital, 601 Elmwood Avenue, Rochester, NY 14642. Address e-mail to Pau1_Bigeleisenurmc.rochester.edu.
 |
Abstract
|
|---|
IMPLICATIONS: This article discusses the inherent danger of general anesthesia and the need for a variety of tools to safely manage the airway.
|
Introduction
|
|---|
I present the case of a 52-yr-old man who likely died from an unrecognized esophageal intubation after the induction of general anesthesia. This case was notable for a normal end-tidal carbon dioxide tracing after intubation. A possible source of this anomalous carbon dioxide is discussed and the physiological conditions producing this condition in an artificial stomach are described. The importance of having fiberoptic bronchoscopy immediately available is also stressed.
The patient was a morbidly obese male (143 kg, 170 cm in height) who was scheduled for knee arthroscopy. His medical history included depression, coronary artery disease, hypertension, hiatal hernia, and gastric reflux. His surgical history included three knee arthroscopies and a left nephrectomy. Examination before anesthesia showed adequate neck extension and a Mallampati class II airway. Lung fields were distant but clear to auscultation, and his heart sounds were regular and without murmur. He took Prilosec (AstraZeneca, Wayne, PA), verapamil, and Prozac (Eli Lilly, Indianapolis, IN) daily, including the morning of surgery, with a sip of water.
Before the induction of anesthesia, the patient was administered 30 mL Bicitra (Alza Pharmaceuticals, San Bruno, CA) by mouth and metoclopramide 10 mg IV in the holding area. In the operating room he was denitrogenated with 100% oxygen for 3 min. Fentanyl 100 µg, propofol 200 mg, and succinylcholine 100 mg were then administered in rapid sequence. The first attempt at intubation by a nurse anesthetist resulted in an esophageal intubation, which was recognized by the absence of end-tidal carbon dioxide. The endotracheal tube was removed.
Immediately, a second laryngoscopy, which resulted in a grade IV view, was performed by the attending anesthesiologist. Finally, a third direct laryngoscopy was performed by the attending with visualization of the vocal cords. After placement of the endotracheal tube, an end-tidal carbon dioxide recording of 35 mm Hg, with a normal square wave tracing was observed on the Datex monitor (Datex, Helsinki, Finland). From the time of the esophageal extubation until the next intubation, the patient was not ventilated by mask. Breath sounds were distant after the second intubation. Over the ensuing minute, the patient’s oxygen saturation declined from 93% to 82% and the end-tidal carbon dioxide decreased to 10 mm Hg. During this period, approximately 10 carbon dioxide square waves were observed. The patient was difficult to ventilate, with peak airway pressures exceeding 40 mm Hg. A diagnosis of bronchospasm was made and the patient was administered isoflurane 5% and albuterol via the endotracheal tube. Despite the above measures, the patient’s oxygenation continued to decline and he became bradycardic. Atropine 1 mg and then epinephrine 1 mg were administered. Next, the patient had a cardiac arrest and advanced cardiac life support was begun. An ear, nose, and throat surgeon, called in to assist with resuscitation, listened to the breath sounds and concluded that the endotracheal tube was in the trachea. The patient could not be resuscitated and was pronounced dead after 45 min of advanced life support, after which the department chairman reviewed the case. He passed a fiberoptic bronchoscope through the endotracheal tube. The patient’s esophagus was visualized, as were the rugae of his stomach. Moreover, the cuff of the endotracheal tube was noted to be inflated within the lumen of the esophagus. These findings were independently confirmed by the head and neck surgeon who assisted at the patient’s resuscitation. The patient’s family declined to have an autopsy performed.
|
Discussion
|
|---|
The case was referred to me for postmortem analysis. It was confusing for several reasons. The first attempted intubation was recognized as an esophageal placement without end-tidal carbon dioxide. The second intubation showed a normal end-tidal carbon dioxide wave form that declined over a minute, accompanied by decreasing oxygen saturation. Finally, the endotracheal tube was found to be in the esophagus at postmortem. The attending anesthesiologist’s explanation was that the tube had been properly placed in the trachea but had become displaced into the esophagus during resuscitation.
The scenario above, however, suggested an unrecognized esophageal intubation. At postmortem, the patient’s endotracheal tube was still securely taped in place, 22 cm at the teeth and the cuff was still inflated. It seems unlikely that an endotracheal tube with the cuff inflated could be displaced from the trachea and passed into the esophagus with the cuff inflated under the circumstances described. Moreover, carbon dioxide can be found in the stomach under certain circumstances. A difficult masked ventilation may force air as well as carbon dioxide into the stomach. The ingestion of carbonated beverages or carbonated antacids can also result in carbon dioxide in the stomach. Either of these could mimic a normal end-tidal tracing after esophageal intubation. The patient, however, did not have masked ventilation, and I dismissed carbonated beverages as a source of carbon dioxide because the patient stated he had been NPO and because there was no carbon dioxide after the first esophageal intubation.
Several of my colleagues suggested that the Bicitra may have played a role in the production of carbon dioxide in the patient’s stomach. Bicitra is a mixture of sodium citrate and citric acid with a pH of approximately 4. In vivo, citrate can be metabolized to succinyl CoA and carbon dioxide, but the reaction requires aconitase in the cell at physiological temperature. In the absence of enzymes, citrate is stable to more than 100°C in both acidic and alkaline media. Citric acid, however, is a triprotic acid. Each packet of Bicitra contains enough citric acid to donate 0.03 moles of hydrogen ion in an alkaline environment. Citric acid can then react with endogenous bicarbonate to form carbon dioxide via the following reactions:
equation
equation
In response to food or acid from the stomach, or an exogenous secretin challenge, the pancreas can dump 20 mEq/h of bicarbonate into the duodenum (Fig. 1) (1,2). Reflux of duodenal contents into the stomach is common. Typical volumes that occur passively are on the order of 8 mL (3). This may occur normally and is significantly increased in the morbidly obese and in patients taking proton pump inhibitors (4,5). In addition, the patient had three direct laryngoscopies. Given the modest doses of propofol and succinylcholine administered to this large man, any one of the laryngoscopies could have elicited coughing, bucking, or gagging. Any one of these active reflexes could significantly increase the volume of duodenal reflux relative to passive reflux.
In the normal physiological state, the parietal cells of the stomach secrete hydrochloric acid continuously. Larger amounts of acid are secreted after eating (6). This hydrochloric acid washes downstream and is slowly neutralized by bicarbonate released into the duodenum from the pancreas. Small amounts of carbon dioxide created in this manner are absorbed into the blood stream, or passed as flatus or eructation. Thus, no large accumulation of bicarbonate ever accumulates in the duodenum. This patient, however, was taking a proton inhibitor (Prilosec). In such patients, the pH in the stomach is nearly 7. Thus, there is no hydrochloric acid to react with bicarbonate in the duodenum. In these patients, a relatively large reservoir of bicarbonate may accumulate in the duodenum.
Mixing the amount of pancreatic bicarbonate discussed above (20 mEq) with the Bicitra the patient ingested would yield approximately 0.02 moles of carbon dioxide if the reaction went to completion. Carbon dioxide is a gas at 37°C that is soluble in water at 0.04 mol/L. At 37°C, it forms a supersaturated solution. As the solution is agitated, more carbon dioxide is released. If all of the carbon dioxide (0.02 moles) generated in the scenario above were released into the atmosphere in the stomach, it would create 0.5 L of carbon dioxide gas. Based on the calculations above, and the properties of dissolved carbon dioxide, it is possible that the citric acid that the patient ingested reacted with bicarbonate that refluxed from the duodenum into the stomach. This may have occurred shortly after the first esophageal intubation.
To test this hypothesis, I created an artificial stomach using a 1000 mL pediatric reservoir bag. In to this artificial stomach, I placed 30 mL Bicitra and 20 mEq of sodium bicarbonate in a total of 100 mL water. The artificial stomach was attached to the Y of the circle and ventilated with tidal volumes of 500 mL at 10 breaths per minute. The resulting carbon dioxide wave form was recorded on the strip chart of our Ohmeda Rascal monitor (Fig. 2). The resulting wave form looks strikingly like an ordinary endotracheal intubation, except for the following subtle factors: 1) the inspiratory phase is very short reflecting the high compliance of the artificial stomach; and 2) the end-tidal carbon dioxide decreases in time and then suddenly increases. This decrease in carbon dioxide reflects the exhaustion and elimination of carbon dioxide liberated in Equations 1 and 2. The sudden increase in carbon dioxide comes from shaking the artificial stomach and liberating more carbon dioxide from the supersaturated solution. All of these features could easily be missed during the stress of a difficult intubation or resuscitation.
In summary, I present the death of a patient that I believe was caused by an unrecognized esophageal intubation. I believe that several factors contributed to the patient’s death. The patient was obese and difficult to intubate, and auscultation of his lungs was unreliable because of his obesity. The combination of a proton inhibitor, Bicitra per os and possible bicarbonate reflux from the duodenum may have created a reservoir of carbon dioxide in his stomach shortly after his anesthetic induction. This carbon dioxide reservoir may have been large enough that it mimicked an endotracheal intubation for at least 60 s. Although a normal, sustained, end-tidal carbon dioxide is diagnostic for endotracheal intubation, any abnormalities must suggest the possibility of an esophageal intubation. Other possibilities for a declining wave form would include decreased cardiac output and severe bronchospasm.
The American Society of Anesthesiologists (ASA) has issued guidelines for the safe management of the difficult airway (7). These guidelines stress awake intubation when a difficult airway is anticipated and a variety of adjuvants when a difficult airway is encountered after the induction of anesthesia. The ASA algorithm emphasizes the presence of end-tidal carbon dioxide for the determination of proper endotracheal intubation. In contrast, this case report emphasizes the need for fiberoptic bronchoscopy in the management of the airway and its immediate use when endotracheal placement is uncertain. Although the putative conditions in this case are extremely rare, the ASA may wish to revise its algorithm to make it clearer that end-tidal carbon dioxide alone is not always the appropriate standard for determining the proper placement of an endotracheal tube. Finally, the case illustrates the need to keep the entire differential for failed intubation in mind, even if the initial carbon dioxide wave form appears normal.
|
References
|
|---|
-
McNaught AB, Callender R. Illustrated physiology. London: Churchil Livingstone, 1975: 62–4.
-
Sabiston DC. Textbook of surgery. 12th ed. Philadelphia: WB Saunders, 1981: 1287–8.
-
Hausken T, Li XN, Goldman B, et al. Quantification of gastric emptying and duodeno-gastric reflux stroke volumes using three-dimensional guided color Doppler imaging. Eur J Ultrasound 2001; 13: 205–13.[Medline]
-
Salava LM, Viribev LP, Shestakov VA. Duodenogastric reflux in patients with alimentary obesity. Lik Sprava 1997; 2: 44–6.
-
Cai J, Jia BQ. Clinical characteristics of bile reflux gastritis [in Chinese]. Zhonghua Nei Ke Za Zhi 1989; 28: 89–92.[Medline]
-
Sabiston DC. Textbook of surgery. 12th ed. Philadelphia: WB Saunders, 1981: 896–904.
-
Caplan RA, Benumof JL, Berry FA, et al. Practice guidelines for management of the difficult airway. Anesthesiology 1993; 78: 597–601.[ISI][Medline]
Accepted for publication January 28, 2002.
Top
Abstract
Introduction
