Search

Introduction: To evaluate the prevalence, risk factors and morbidity associated with gastrointestinal (GI) complications after cardiac surgery, with and without cardiopulmonary bypass, we carried out a retrospective cohort study at a university teaching hospital., Methods: We divided the 11,405 eligible adult patients into 2 groups: group A (operated on between January 1992 and June 1996) (4657 patients) and group B (operated on between July 1996 and December 2000) (6748 patients)., Results: We found 147 GI complications in 134 (1.2%) patients. The incidence of GI morbidity was similar for the 2 groups of patients (group A, n = 59/4657 [1.2%]; group B, n = 75/6748 [1.1%]. Patients from group B were older, obese, diabetic and presented with more peripheral and cerebrovascular disease. Bleeding, gastritis and ulcer with perforation, the most common of these GI events, were associated with the esophagus and stomach (67/147 [45.5%]). Other events that we documented included cholecystitis 10 (6.8%), pancreatitis 13 (8.8%), episodes of small and large bowel ischemia 17 (11.6%), pseudomembranous colitis 12 (8.3%) and diverticulitis 5 (3.4%). Mesenteric ischemia was responsible for 11 (37.9%) of the 29 deaths. Two hundred and ninety-three patients were revascularized without extracorporeal circulation during this study. In this group, we were able to pinpoint 5 (1.7%) GI complications with 3 cases of mesenteric ischemia. Multivariate analysis identified renal insufficiency, prolonged intubation and sepsis as significant, predictive variables of GI complications for the 2 groups of patients whereas the Parsonnet score and stroke were predictive for the second group., Conclusions: Although cardiac surgery is now being performed on older patients with significant comorbidity, we could not demonstrate a significant increase of GI complications after cardiac surgery. Off-pump coronary artery bypass does not seem to protect patients from these complications.

This study was designed in order to validate the respiratory mechanical variables measured by the Puritan-Bennett 7200a ventilator equipped with the 30/40 module. Two ventilators were connected to a lung model and submitted to several breathing patterns by modifying the respiratory rate, the tidal volume, the inspiratory flow-rate and the model resistance. The inspiratory flow-rate (V), tidal volume (VT), peak inspiratory pressure (Pmax), plateau pressure (Pplat) and PEEP measured by the ventilators were compared with the same variables measured at the connection between the breathing circuit and the lung model. The compliance (C30/40) and the resistance (R30/40) calculated by the 30/40 module were compared with those calculated by using the variables measured by the reference equipment. Both ventilators made a constant underestimation of V by 2.8 and 3.7 L.min-1, respectively. The VT was measured with a mean error of less than 10 ml but did not reflect the preselected values in the presence of an intrinsic PEEP. The Pplat was overestimated by 7 and 10%, respectively. The same calibration error was observed with Pmax which was also affected by a pressure gradient due to the resistance of the breathing circuit. Even in the absence of intrinsic PEEP, C30/40 presented an error due to the combination of the measurement errors on VT, Pplat and PEEP. Finally, R30/40 presented a high percentage of error due to the combination of the measurement errors on V, Pmax, and Pplat, and to a sporadic aberrant selection of V. Due to these numerous sources of error, the two ventilators studied did not give reliable estimates of resistance and compliance.(ABSTRACT TRUNCATED AT 250 WORDS)