Physiopathological mechanism of respiratory distress syndrome associated with acute pancreatitis remains unclear. We observed a case of drowning-like hypoxemia due to a pancreaticobronchial fistula; this patient, with a necrotizing isthmic pancreatitis, had only two Ranson’s criteria, and showed a clearly isolated mechanical reason for respiratory failure. Then, the retrospective study of 40 patients having an objective surgical recognition of necrotizing pancreatitis (NP) was performed. Twelve patients received ventilatory support. The prevalence of isolated sepsis-related respiratory failure was 9 out of 12 cases (75%). Two patients had a secondary infection of necrosis also associated with a pancreaticobronchial fistula, and in 1 case this isolated mechanism of respiratory distress was obvious. Such a prevalence of 25% of fistula-related respiratory failure has never been observed before. The onset of severe hypoxemia in the course of NP appears to be a sensitive and pathognomonic sign of evolving complication of pancreatitis, leading to endoscopic retrograde pancreatography, and to prompt surgery. In the opposite way than that suggested by the enzymatic hypothesis, we were not able to show any reason for applying mechanical ventilation to patients suffering from pancreatitis, in the absence of intraabdominal sepsis, and/or pancreaticobronchial fistul