Peripheral bypass-induced pulmonary and coronary vascular injury. Association with increased levels of tumor necrosis factor.

BACKGROUND Although cardiopulmonary bypass is associated with systemic complement activation and neutrophil sequestration, it is unclear whether bypass-induced vascular injury is localized and dependent on organ ischemia. We hypothesized that other factors perhaps related to placement of a bypass circuit or to blood perfusion of a pump-oxygenator system may produce vascular injury caused by systemically circulating mediators. In dogs, we determined whether application of a systemic venoarterial bypass circuit with pump-oxygenator perfusion but without pulmonary or cardiac flow diversion (peripheral bypass) leads to vascular injury. Since several features of the postperfusion syndrome after bypass resemble sequelae of endotoxin exposure, we also measured circulating endotoxin and tumor necrosis factor levels. METHODS AND RESULTS Anesthetized dogs underwent 2 hours of exposure to a pump-oxygenator with peripheral venoarterial bypass. We used a double indicator measurement of pulmonary and coronary vascular permeability (protein leak index [PLI]) as indexes of vascular injury. Compared with controls (n = 7), the pulmonary PLI of dogs undergoing bypass (n = 11) increased more than threefold (18.8 +/- 2.3 vs 63.3 +/- 7.6 x 10(-4) min-1; P < .05) and the coronary PLI increased more than twofold (P < .05). The rate of disappearance of intravascular radiolabeled protein increased threefold after bypass (disappearance t1/2, 241 +/- 35 vs 84 +/- 15 minutes, control vs bypass; P < .05), suggesting a generalized increase in vascular permeability. Circulating endotoxin was detectable in blood samples from 8 of 8 bypass animals (range, 0.24 to 4.56 ng/mL) compared with 2 of 5 controls (P < .05). Tumor necrosis factor levels increased significantly with bypass (6.7 +/- 3.8 vs 146.7 +/- 33.6 U/mL, baseline vs bypass; P < .05) and were only slightly and nonsignificantly increased in controls (7.0 +/- 4.4 vs 18.2 +/- 5.9 U/mL; P = NS). Peak tumor necrosis factor but not peak endotoxin levels correlated with pulmonary and with coronary protein leak. As expected, circulating complement (CH50) levels decreased significantly during bypass, reflecting systemic complement activation. However, the levels correlated poorly with the severity of vascular injury. CONCLUSIONS We conclude that peripheral pump-oxygenator bypass causes coronary and pulmonary vascular injury that is independent of blood flow diversion and is associated with the appearance of circulating levels of endotoxin and tumor necrosis factor, which may play a role in bypass-induced vascular injury.