Acute abdomen after coronary artery bypass surgery masked by thoracic epidural analgesia

EDITOR: We report an unusual case of acute abdomen after coronary artery bypass surgery that was detected late due to atypical presentation caused by thoracic epidural analgesia. A 64-yr-old male, 100 kg body weight, ASA IV, was admitted to hospital 5 months after an acute anterolateral myocardial infarction for elective coronary artery bypass surgery. He had triple vessel coronary artery disease and a left ventricular ejection fraction of 40%. He was also suffering from chronic bronchitis, gout, myasthenia gravis and left internal carotid artery stenosis. Preoperatively, he was receiving only vasodilator and antihypertensive therapy: isosorbide dinitrate, atenolol and captopril. The patient was premedicated with atropine 0.5 mg, midazolam 5 mg and merperidine (pethidine) 50 mg. A thoracic epidural catheter was inserted at the T4-5 interspace. Anaesthesia was then provided with propofol (1.5 mg kg−1), remifentanil (1 μg kg−1), and atracurium (1 mg kg−1) for muscle relaxation. A bolus dose of bupivacaine 0.125% 20 mL (together with fentanyl 25 μg) was given followed by a continuous infusion of bupivacaine at a rate of 10-15 mL h−1. Anaesthesia was maintained by a continuous infusion of propofol (100 μg kg−1 min−1) and remifentanil (0.25 μg kg−1 min−1). A triple aortocoronary bypass and a thymectomy were performed. The early postoperative course was uneventful and satisfactory epidural analgesia through the catheter was maintained. The patient was hypertensive - up to 170/100 mmHg - and was treated with glyceryl trinitrate (nitroglycerin) in doses of up to 5 μg kg−1 min−1. The trachea was extubated 7 h after the surgical procedure. The patient felt well, was free of pain and was haemodynamically stable; the thoracic drain was removed the next day. One hour after the removal of the drain, while the patient was still receiving continuous epidural analgesia, he complained of sudden and intense pain under the right costal margin and in the right shoulder. The pain was characterized as a visual analogue score (VAS) of 9, but the patient could not localize it clearly. He was sweating heavily and felt cold. His arterial pressure was 120/80 mmHg and CVP 4 mmHg. White blood cell count was 15.3 × 109 L−1. We suspected an acute abdomen: the continuous infusion of bupivacaine was halted and a general surgical opinion requested. Abdominal palpation and percussion did not reveal any abdominal tenderness, muscle spasm or ascites. However, the abdomen was tense and slight peristalsis was present. A nasogastric tube was inserted and 200 mL clear fluid obtained. The patient continued to complain of pain. Repeated palpation revealed again no abdominal tenderness and no muscle spasm; body temperature was 34.6°C. The patient could not stand to permit a radiographic abdominal examination; chest radiography in the sitting position and echocardiography findings were unremarkable. Treatment consisted of: volume replacement (mainly with colloid solutions, up to 1500 mL h−1) inotropic support (dopamine up to 15 μg kg−1 min−1 and then epinephrine up to 0.5 μg kg−1 min−1) and warmth (warming of the i.v. infusions and warm covers). The patient responded favourably and his condition improved. The pain intensity decreased to VAS 3 and body temperature rose to 36.7°C. Because the hypotension worsened and large quantities of fluids had been used, epinephrine was introduced for inotropic support in doses up to 21 μg kg−1 min−1(Table 1). Metabolic acidosis gradually developed and was treated with bicarbonate 50 mL h−1(Table 2). Dopamine was given in a dose of 3 μg kg−1 min−1 to promote urine output. Over the next 12 h, 8 L of liquid were given and the urine output was 2 L. Arterial pressure continued to be low (ranging between 74/52 and 94/50 mmHg) and CVP rose to 13 and 14 mmHg.Table 1: Course of hypotension according to the time from the first onset of pain.Table 2: Acid-base status.The following day, an abdominal radiograph was taken with the patient in the erect position, but it did not show any signs of intestinal perforation or ileus. A repeat chest radiograph in the sitting position was unremarkable. The patient's condition slightly improved, the pain diminished, the blood pressure was normal with the inotropic support (dopamine 10 μg kg−1 min−1), diuresis was maintained, and peristalsis was audible. The patient was subfebrile (37-38°C), and white blood cell count was 18.7 × 109 L−1. There were slight increase of urea, creatinine and some liver function tests. Serum lactate concentration was within normal range. On the third day after the onset of pain, although there were no clinical signs of an acute abdomen, an Hippocratic facies was apparent. It was decided to perform an explorative laparotomy: a perforating duodenal ulcer and acute cholecystitis were found with a large volume of pus in the peritoneal cavity. The ulcer was oversewn and pyloroplasty and cholecystectomy were performed. The postoperative course proceeded without complications. Epidural analgesia for cardiac surgery has many benefits [1]. Gastrointestinal complications are relatively rare after cardiac surgery, and the problem lies in the fact that such possibilities are rarely considered, particularly in centres that specialize only for cardiovascular diseases [2]. If the patient's preoperative history suggests gastrointestinal problems or the patient has been treated previously for some gastrointestinal disease, more attention would have been paid and major complications would be less likely to be missed. However, our patient did not indicate that he was suffering from an ulcer (although a chronic duodenal ulcer with perforation was discovered at operation). He underwent an extensive surgical procedure and, most importantly, had an epidural catheter with continuous epidural analgesia. The tip of the catheter was at level T3, and therefore the level of analgesia did not reach the C5 and C6 dermatomes in the cranial direction, and this explains the pain in the shoulder [3] and why it was not possible to localize the usual severe pain of ulcer perforation or acute cholecystitis. Local anaesthetic in the epidural space blocks the afferent and efferent reflex responses and thus prevents spasm of the skeletal muscles; this may explain the absence of muscle spasm that misled making a correct diagnosis. Importantly, at the time of the gut perforation, a continuous epidural infusion of bupivacaine had already been in progress for 20 h, so that its cessation did not immediately remove the analgesic effect; the analgesic effect of epidural bupivacaine lasts for 180-300 min. That was the time when the patient's pain was at its most intense. The relative hypotension (120/80 mmHg in a patient who was markedly hypertensive and had previously maintained an arterial pressure of 160/120 mmHg), which developed at the start of our patient's pain, was initially a misleading sign and was interpreted as a consequence of the epidural use of both local anaesthetic and fentanyl. Although the concentration of the local anaesthetic was small and did not normally cause hypotension, bupivacaine with fentanyl together can have a hypotensive effect [4]. In our patient, only indirect signs pointed to something happening in the abdomen: hypotension, hypothermia, sweating in the beginning and metabolic acidosis later indicated a loss of fluid into a third space. Thoracic epidural analgesia induced sympathetic and motor blockade completely masked the signs of an acute abdomen. We should emphasize the use of computed tomographic (CT) scanning in the investigation of a patient with such clinical deterioration. Considering that epidural blocks may mask symptoms, early imaging such as CT would prove an invaluable tool. In conclusion, the typical signs and symptoms of an acute abdomen may be masked or modified during the early postoperative period due to thoracic epidural analgesia. Z. Jankovic Leeds, UK D. Radojevic V. Neskovic D.