Giant duodenal ulcers after neurosurgery for brainstem tumors that required reoperation for gastric disconnection: a report of two cases

In these two cases, the brainstem tumors might have been related to duodenal ulcer perforation that progressed to septicemia. In 1841, Rokitansky suggested for the first time that ulcerative processes of the stomach might involve dysfunction of nervous mechanisms [3]. In 1932, Cushing reported gastroduodenal ulcers produced by elevated intracranial pressure caused by an intracranial tumor, head injury, or other space-occupying lesion. He suggested that such ulcerative processes might be related to diencephalic or brainstem disorders affecting the parasympathetic nervous system. Since then, ulcers of this type have been called Cushing’s ulcers [1].

The mechanism of ulceration appears to involve three routes from the central nervous system to the stomach: 1) anterior hypothalamus – vagus nerve; 2) posterior hypothalamus – sympathetic nerve; and 3) posterior hypothalamus – anterior pituitary gland – adrenal cortex. Through these three routes, factors that aggravate the stomach are increased or protective factors are decreased. The sympathetic and parasympathetic nervous systems usually maintain a balance of the blood supply, gastric secretion, and gastric motility. Dysfunction of the central nervous system stimulates the hypothalamus, which then stimulates the sympathetic and parasympathetic nervous systems. Stimulation of sympathetic nerves decreases blood supply to the stomach, and stimulation of parasympathetic nerves increases gastric secretion. Moreover, adrenal cortical hormones through the anterior pituitary gland decrease gastric mucus secretion [4‐10]. These factors then contribute to the development of gastroduodenal ulcers.

In general, factors such as advanced age, concomitant disease, preoperative shock, large size of the perforation, and delays in presentation and operation have been identified as risk factors for mortality in duodenal ulcer perforation [2]. Based on these factors, several scoring systems have been used to evaluate the condition of the patient with duodenal ulcer perforation, such as the Boey score [11], Mannheim Peritonitis Index [12‐14], APACHE II score [15] and Jabalpur score [16]. In particular, the perforation-operation interval seems to represent an important factor for mortality. Mishra et al. reported that the mortality rate is 3 % within 24 h, 57 % from 25 to 72 h, and 80 % over 120 h after duodenal ulcer perforation [16]. Many reports have stated that an interval to operation larger than 24 h increases the mortality rate [17‐19], because heavier bacterial contamination occurs in patients with delayed treatment [20]. In the present two cases, decreased level of consciousness was the major cause of delayed diagnosis in both patients. Although gauging the interval since ulcer perforation is difficult, at least 48 h may have elapsed in both cases, given the presence of septic shock. Of the above risk factors, our two cases showed large perforations, delayed diagnosis, concomitant disease, and preoperative shock, as well as advanced age in Case 2. Operations in such cases are generally difficult. Nonetheless, antrectomy with Billroth II reconstruction was performed for Case 1 and omental patching was performed for Case 2. Because gastric disconnection requires a second operation for digestive reconstruction, we hesitated to perform this procedure, but gastric disconnection was unavoidable at the first emergency surgery.

Most duodenal ulcer perforations are less than 1 cm in length, and can be successfully treated with one-layer closure plus a pedicled omental patch (Cellan-Jones technique) or an omental patch repair (Graham technique) [21‐23]. On the other hand, giant duodenal ulcers are uncommon, with duodenal ulcer perforation more than 3 cm in length reportedly accounting for about 1.23 % of cases [2]. Giant duodenal ulcers are difficult to manage and are associated with high rates of both morbidity (20–70 %) and mortality (15–40 %) because of the extensive duodenal tissue loss and surrounding tissue inflammation [24]. The Cellan-Jones and Graham techniques often fail to achieve closure of the perforation, resulting in postoperative leakage or gastric outlet obstruction.

Several reports have described surgical procedures for giant ulcers, including partial gastrectomy, jejunal serosal patch [25], free omental plug [26], and jejunal pedicle graft [27]. Lal et al. reported the efficacy of triple-tube-ostomy (tube gastrostomy, retrograde tube duodenostomy, and feeding jejunostomy) with repair of the perforation for large duodenal ulcer perforations [28]. Cranford et al. advocated gastric disconnection with truncal vagotomy, antrectomy, and triple-tube-ostomy [29]. This surgical approach is considered the most appropriate procedure for giant duodenal ulcer perforation in cases with poor general conditions owing to late diagnosis. Because one of the present cases showed bleeding and leakage from the repaired duodenal ulcer, antrectomy including the ulcerative portion was thought to be necessary for giant duodenal ulcer. In cases with poor general conditions owing to late diagnosis, digestive tract reconstruction is hazardous, and gastric disconnection might be needed. This approach necessitates a second elective operation for digestive reconstruction, but is thought to represent the safest procedure given the high mortality rate of this condition. Moreover, cholecystectomy with insertion of a bile drainage tube from the cystic duct might also be necessary in preparation for duodenal stump leakage.