Femoral Nailing in a Porcine Model Causes Bone Marrow Emboli in the Lungs and Systemic Emboli in the Heart and Brain

Abstract

Background: Shaft fractures of the femur are commonly treated with intramedullary nailing, which can release bone marrow emboli into the bloodstream. Emboli can travel to the lungs, impairing gas exchange and causing inflammation. Occasionally, emboli traverse from the pulmonary to the systemic circulation, hindering perfusion and resulting in injuries such as heart and brain infarctions, known as fat embolism syndrome. We studied the extent of systemic bone marrow embolization in a pig model.

Methods: Twelve anesthetized pigs underwent bilateral intramedullary nailing of the femur, while 3 animals served as sham controls. Monitoring included transesophageal echocardiography (TEE), pulse oximetry, electrocardiography, arterial blood pressure measurement, and blood gas and troponin-I analysis. After surgery, animals were monitored for 240 minutes before euthanasia. Post mortem, the heart, lungs, and brain were biopsied.

Results: Bone marrow emboli were found in the heart and lungs of all 12 of the pigs that underwent intramedullary nailing and in the brains of 11 of them. No emboli were found in the sham group. The pigs subjected to intramedullary nailing exhibited significant hypoxia (PaO₂/FiO₂ ratio, 410 mm Hg [95% confidence interval (CI), 310 to 510) compared with the sham group (594 mm Hg [95% CI, 528 to 660]). The nailing group exhibited ST-segment alterations consistent with myocardial ischemia and a significant increase in the troponin-I level compared with the sham group (1,580 ng/L [95% CI, 0 to 3,456] versus 241 ng/L [95% CI, 0 to 625] at the 240-minute time point; p = 0.005). TEE detected emboli in the right ventricular outflow tract, but not systemically, in the nailing group.

Conclusions: Bilateral intramedullary nailing caused bone marrow emboli in the lungs and systemic emboli in the heart and brain in this pig model. The observed clinical manifestations were consistent with coronary and pulmonary emboli. TEE detected pulmonary but not systemic embolization.

Clinical relevance: Femoral intramedullary nailing in humans is likely to result in embolization as described in our pig model. Focused monitoring is necessary for detection of fat embolism syndrome. Absence of visual emboli in the left ventricle on TEE does not exclude the occurrence of systemic bone marrow emboli.

Fig. 1

Allocation of pigs and distribution of emboli. Fifteen pigs were included: 12 allocated to intramedullary reaming and 3 allocated to the sham group.

Fig. 2

Overview of instrumentation in the pigs undergoing intramedullary reaming and nailing.

Fig. 3

Histopathologic examinations of tissue prepared in O.C.T. compound and frozen on dry ice revealed bone marrow emboli in cardiac (Fig. 3-A), cerebral (Fig. 3-B), and pulmonary cryosections (Fig. 3-C) with ×10, ×40, and ×10 magnification, respectively. Bone marrow emboli were stained red by the oil red O.

Fig. 4

Transesophageal echocardiography with an M-mode image showing the left and right ventricular outlet tracts (LVOT and RVOT, respectively) before (left image) and immediately after (right image) femoral intramedullary nailing. The x axis shows time in seconds, and the y axis shows depth in centimeters. Dense, hyperechogenic material filled the RVOT as intramedullary nailing commenced. No emboli were visible in the LVOT.

Fig. 5

Changes in the ST segment and troponin-I levels during monitoring. ST-segment changes at the J point, shown as millimeters of deviation from the J point, for the 12 animals undergoing bilateral femoral nailing revealed substantial ischemia, which was most pronounced during actual reaming and nailing. The blue line shows averaged ST-segment changes in ECG leads II, aVR, and aVL. The orange line shows averaged ST-segment changes in ECG leads aVF and III. The shaded areas cover ±1 SD. The mean troponin-I level (red line), shown in ng/L, increased throughout the experiment in all pigs undergoing femoral intramedullary nailing but varied considerably between the pigs. Error bars span ±1 SD. If a measurement minus 1 SD was negative, the lower error bar was bounded at zero.