Hypoxemia following long bone or pelvic fracture (LBPF) is often attributed to fat embolism syndrome (FES), but the true incidence and etiology of postfracture pulmonary shunt (Qsp) are not well defined. Over 12 months, 92 patients with LBPF admitted to a Level I trauma center were prospectively evaluated. Arterial blood gases, Hct, platelet count, serum fibrinogen, serum lipase, and urinary fat bodies (UFB) were determined serially from admission through the fifth hospital day. Patients were evaluated daily by chest x-ray, vital signs, mental status, and presence of petechiae. Four patient groups were established: No Qsp, Qsp with pulmonary injury (Qsp + PI), Qsp without pulmonary injury or petechaie (FES - P), and Qsp without pulmonary injury and with petechiae (FES + P). Qsp indicated by an alveolar/arterial PO2 gradient greater than 100 torr developed in 49 (53%) of the patients. Pulmonary injury was present in 39 (81%) of those 49 and was at least partially responsible for the shunt. The remaining ten patients were diagnosed as having FES; four had petechiae (FES + P) and six were without petechiae (FES - P). The minimum incidence of FES in LBPF is therefore 11%.