Intraoperative insertion of Greenfield filters: lessons learned in a personal series of 152 cases

The objective of this study was to define outcomes of 151 patients who underwent insertion of 152 Greenfield filters in the operating room by general and vascular surgery residents with supervision by one attending vascular surgeon. Each patient was taken to the operating room for inferior vena cava (IVC) interruption immediately after a vena cavagram was performed. One patient required a subsequent return to the operating room after developing paradoxical arterial embolism from a large venous thromboembolism which was trapped by and spanned both sides of the first IVC filter. In this case a second suprarenal filter was placed at the time of arterial embolectomy. In each of these 152 cases intraoperative venacavography was performed using a mobile C-arm. Complications such as hemothorax, filter misplacement, and vena cava perforation were identified. Late survival was defined using the Social Security Death Index. Of the 151 patients undergoing intraoperative insertion of Greenfield filters there was one hemothorax from attempts at acquiring venous access via percutaneous puncture of the internal jugular vein. This required transfusion but not thoracotomy, and IVC interruption was achieved. A separate patient had insertion of a Greenfield filter into a gonadal vein which required placement of a second filter into the IVC. There was one IVC perforation from a transfemoral filter insertion which required placement of a second filter above this perforation and laparotomy to retrieve the filter and repair the IVC. In one more patient the IVC filter initially failed to open, and a second filter was placed above the first filter. In this experience the misplacement rate was 0.7 per cent and the serious complication rate was 1.3 per cent. None of the patients was adversely affected per se by transfer to the operating room for Greenfield filter insertion. No patient died from filter insertion, but in two cases serious associated complications contributed to the adverse outcomes in these already terminally ill patients. Overall 30-day mortality rate was 6.6 per cent. Late survival was defined as follows: survival at one year after filter insertion was 75 per cent, at 2 years 63 per cent, at 3 years 60 per cent, at 4 years 57 per cent, and at 5 years 54 per cent. Mean survival after filter placement was 4.96 years. We conclude that Greenfield filters can be inserted in the operating room by general and vascular surgery residents with attending supervision with reasonable safety and with a low rate of filter misplacement. The caval perforation and gonadal vein filter misplacement could both have been avoided by use of an over-the-wire filter deployment system, which at the time of these specific complications was not available. Vena cava filter insertion should remain within the scope of practice of surgeons and can be done with reasonable safety under C-arm guidance in the operating room. Use of over-the-wire systems could have helped reduce the likelihood of all but one of the filter-related complications experienced in this series.