The Incidence of Venous Thromboembolism in Breast Reconstruction and the Efficacy of Current Prediction Models

Introduction

The incidence of venous thromboembolism (VTE) in methods of breast reconstruction has not been well described.  Many models have been developed to determine VTE risk including the Davidson risk score, the 2005 Caprini risk score, and the 2010 Caprini risk score.  Recently, the 2005 Caprini risk score has been validated in plastic surgery patients¹ and has demonstrated superior predictive value than the 2010 Caprini risk score².

Methods

We performed a retrospective chart review of breast reconstructions by a single surgeon.  One hundred consecutive TRAM patients, one hundred consecutive implant patients, and fifty consecutive latissimus dorsi patients were identified.  All patients were followed for a minimum of 30 days.  Each chart was thoroughly reviewed to collect patient factors and to examine for evidence of VTE in the perioperative period.  Using the collected patient factors, 2005 Caprini risk scores and Davidson risk scores were calculated for each patient. 

Results

The TRAM reconstruction group had a significantly higher VTE rate (6%) than the implant (0%) and latissimus (0%) reconstruction groups (p<0.01).  The 2005 Carpini risk score stratified patients into “high risk” (Caprini Score>5) more frequently than the Davidson risk score (p<0.01). However, the positive predictive values of the Davidson score (3.09%) and the 2005 Caprini score (2.80%) were not statistically different (p>0.1).  

The rate of VTE among TRAM patients stratified as “high risk” by 2005 Caprini score was 6.81% (6/88). Similarly, the rate of VTE among TRAM patients stratified as “high risk" by Davidson score was 7.93% (5/63) (p>0.1).  None of the patients in the Implant or Latissimus groups who were stratified as “high risk” (127 by Caprini and 99 by Davidson) developed a VTE.

Conclusion

Overall, TRAM reconstruction appears to have a significantly higher risk of VTE than either implant or latissimus reconstruction. While the current VTE models had some predictive value in TRAM patients, both were poorly predictive of VTE in the Implant and Latissimus reconstruction groups. The 2005 Caprini risk score and the Davidson risk score did not significantly differ in their risk stratification of breast reconstruction patients. VTE risk assessment models should take the surgical procedure into account to more accurately determine VTE risk.