Peripheral Vascular Surgery Society

Frank J. Caliendo, MD, Vivienne J. Halpern, MD, Corrado P. Marini, MD,

Ira M. Nathan, PhD, and Jon R. Cohen, MD

Long Island Jewish Medical Center, New Hyde Park, NY

Purpose: Vascular surgeons are often consulted about how to manage patients on Warfarin (WA) needing to undergo surgical procedures. The general practice has been to stop WA and use Heparin (HEP) in the perioperative period. However, bleeding complications with this policy are significant. This study was undertaken to see if WA anticoagulation could safely be continued during surgery and in the perioperative period.

Methods: The study consisted of an animal model, followed by a prospective human study of all patients on HEP or WA at the time of surgery. 24 rabbits were randomized to three groups. All rabbits underwent a laparotomy during which a controlled liver injury was created and repaired. Group 1 was anticoagulated with WA to raise the mean international normalized ratio (INR) to 2.5-3.0. Group 2 was anticoagulated with HEP to raise the PTT to 1.5 to 2 times normal range. The HEP was then stopped 6 hours prior to surgery and resumed 6 hours postoperatively without a bolus. Group 3 was not anticoagulated. The hematocrit (HCT) and operative blood loss (OR-loss) were recorded and at 24 hours the postoperative blood loss (POBL) was measured. Significant differences between the groups was accessed with ANOVA with a Bonferroni modification.

Data was collected on 28 patients undergoing 33 operations from October 1996 until December 1997. Statistics were analyzed with t-test analysis and chi-square analysis.

Results: The animal study revealed a significantly increased OR-loss, POBL and change in HCT from base line in HEP group as compared to WA and control groups (see table).

*p<0.05 for Group 2 vs. Group 1 and Group 3

Thirteen operations (op) occurred in patients while on WA and 20 op in patients while on HEP. There were 10 bleeding complications (BLE) (35%) vs. 3 of 13 op in the WA group (16%). This approached significance (p=0.06, odds ratio 4.37). Two of three op (66.6%) with an INR > 3 had BLE vs. 1 of 10 (10%) with an INR <3. This was not statistically significant.

Conclusions: The results of this study reveal that (1) bleeding was significantly less in the WA group in the animal model, (2) bleeding complications occurred less in patients on WA as opposed to those on heparin, and (3) an INR of greater than 3 at the time of surgery may increase the risk of bleeding. We recommend maintaining patients on WA with an INR <3 during surgery and not following the standard policy of converting to HEP.