Does Heparin Prophylaxis Reduce the Risk of Venous Thromboembolism in Patients Undergoing Robot-Assisted Prostatectomy?

Introduction

T he value of subcutaneous heparin prophylaxis in patients undergoing minimally invasive urologic surgery is currently unknown. The 2008 American Urological Association Best Practice Statement recommends sequential compression devices (SCDs) for urologic laparoscopic and robot-assisted procedures but provides no clear recommendations regarding the use of pharmacologic prophylaxis. ¹ Although a rare event after urologic procedures, venous thromboembolism (VTE) is a potentially life-threatening complication that necessitates prophylaxis.

The incidence of symptomatic VTE after major urologic surgery has decreased significantly with the administration of SCDs, recognition of the importance of early ambulation, and the introduction of pharmacologic prophylaxis. ^2,3 Recent large series have demonstrated a VTE risk of less than 1% after RP. ^4,5 Heparin administration is not without drawbacks, however. Cost considerations, allergic reactions, increased bleeding risk, and lymphocele formation are potential risks. ⁶ Meanwhile, there is evidence to suggest pharmacologic prophylaxis may not provide additional risk reduction. ⁷

Given the increased use of robot-assisted surgery for the management of localized prostate cancer and the significant morbidity of VTE, it is important to understand the value of perioperative heparin administration. This study compares the rates of postoperative VTE in a group of patients undergoing RP with perioperative heparin prophylaxis vs a group of patients without heparin prophylaxis.

Patients and Methods

An Institutional Review Board approved, prospectively managed database for patients with prostate cancer who were treated by RP at a single institution was analyzed. A total of 1486 consecutive patients underwent RP by two surgeons between August 1, 2007 and December 31, 2011. Demographic, clinical, and pathologic data were prospectively recorded. Symptomatic VTE occurring at any time point postoperatively was also recorded.

Patients were divided into two groups. Group 1 included 922 patients who underwent RP by surgeon A. SCDs were applied before induction and 5000 units of heparin were administered subcutaneously just after induction. SCDs were continued postoperatively, and heparin was administered twice daily until discharge. Group 2 included 564 patients who underwent RP by surgeon B. SCDs were applied before induction and maintained until hospital discharge. Heparin prophylaxis was administered only if the patient had a known history of VTE or hypercoagulable state.

RP in both groups was performed in the standard lithotomy, steep Trendelenberg position using a transperitoneal approach with identical pneumoperitoneum. Lymph node dissection templates included the nodes overlying the external iliac artery and veins up to the bifurcation of the common iliac vessels, as well as the nodes of the obturator fossa. The templates were identical for both surgeons. Extended templates more proximal along the iliac vessels were performed for higher risk patients. All patients were started on an early ambulation protocol the same day of surgery.

Patients presenting with swelling or erythema of the lower extremities were evaluated for deep vein thrombosis with duplex ultrasonography. Clinical suspicion for pulmonary embolus prompted chest CT angiography. Signs and symptoms suspicious for infection, abdominal or pelvic pain, or distention prompted evaluation for a symptomatic lymphocele that necessitated treatment. VTE rate, patient age, body mass index (BMI), operative time, lymphocele rate, estimated blood loss (EBL), length of stay (LOS), Gleason score, and pathologic stage were compared between the two groups. Categorical variables were analyzed with the chi-square test of proportions and continuous variables with t test using SPSS v 14 software.

Results

There were 1486 consecutive patients who underwent RP between August 1, 2007 and December 31, 2011. Of these, 922 patients received heparin/SCDs (group 1) and 564 received SCDs alone (group 2). Clinical and pathologic characteristics were similar between groups (Table 1). Group 2 had a higher incidence of node positive disease. pT stages were similar between both groups.

There were nine and four VTEs in groups 1 and 2, respectively. Despite longer operative times in group 2 (229.3 vs 170.6 min, P<0.01), the incidence of VTE was similar (1.0% vs 0.7%, P=0.78). There was one lymphocele noted in each group. The rate of lymph node dissection was higher in group 1 (59.4% vs 47.7% P=0.01); however, there was a significantly higher rate of lymph node positive disease in group 2 (3.5% vs 1.3%, P=0.01). See Table 2 for perioperative characteristics. Five patients in group 1 received blood transfusions and one patient from group 2 was transfused. Reasons for transfusion were variable, and none of these patients experienced a VTE.

The overall VTE rate in our population was 0.9%. As a single cohort, an analysis of potential independent risk factors was performed. BMI, EBL, and performance of lymph node dissection were not associated with VTE. Also of note, longer operative time was analyzed as an independent risk factor and was not found to be significant (Table 3).

There was a significant correlation between LOS and VTE rate (5.8 days vs 1.4 days, P=0.01), largely because of a prolonged hospital stay of 40 days by one patient in the VTE group. There were 807 (54.8%) patients who underwent lymph node dissection with a VTE incidence of 1.2%. The remaining 666 patients who did not undergo a lymph node dissection (45.2%) had a VTE incidence of 0.4%, P=0.09 (Table 4).

Discussion

VTE is a potentially life threatening postoperative complication that has decreased significantly by the use of SCDs and early ambulation. The value of pharmacologic prophylaxis is unclear. The American College of Chest Physicians guidelines support its use, ⁸ but the evidence is from nonurologic surgery and may not be applicable to minimally invasive urologic procedures. The evidence suggests there is approximately a 4.1-fold increased risk of VTE in patients with malignancy with an even greater risk postchemotherapy. ⁹ In general, the odds ratio for development of deep venous thrombosis in a cancer patient is 2.2 vs patients without malignancy. ⁹ Pharmacologic prophylaxis has been used empirically to decrease this risk. This evidence may not be applicable to urologic procedures and RP, however.

A review of the literature confirms that VTE is a relatively rare event in robotic and laparoscopic RP. Hu and associates ⁴ reported a rate of 0.6% in 322 patients who underwent RP and 0% in 358 patients who underwent laparoscopic prostatectomy. Similarly, Secin and colleagues ⁵ reported a 0.5% rate of VTE in 5951 patients who underwent minimally invasive radical prostatectomy. Our VTE incidence of 0.9% is consistent with these large series and supports the evidence that the risk of VTE in patients undergoing RP is low. Our series suggests that the risk is not significantly reduced with the administration of prophylactic heparin/SCDs compared with SCDs alone. A similar study by Patel and coworkers ⁷ also questions the value of heparin administration for patients with prostate cancer undergoing RP.

Longer operative time has been proposed as a potential risk factor for VTE, ⁹ but our analysis did not support this as significant. Pneumoperitoneum needed for laparoscopy causes venous compression and altered intravascular hemodynamics; longer operative times theoretically increase this phenomenon. In our study, group 2 (SCDs alone) had longer operative time but an equal VTE rate to that of group 1. Furthermore, we found that operative time was not an independent predictor of VTE.

Although there was a significant difference in the rate of lymph node dissection between the two groups, this did not correlate with the risk of VTE. Only one lymphocele was noted in each group making the incidence of this variable too small to meaningfully analyze. While a higher rate of VTE was found in those patients who had a node dissection performed, this did not prove to be statistically significant (1.2% vs 0.4%, P=0.09).

Limitations of this study include the lack of randomization. However, if one assumes a 0.9% VTE rate, more than 4000 patients would have to be randomized to show a 1% difference in VTE rates at 90% power. These numbers make such a study highly impractical.

Another limitation is the inherent difference in surgical technique by two different surgeons. This is reflected in the difference in operative times as well as the decision to perform lymph node dissection. Although a comparison of lymph node yield between groups was not performed, similar dissection technique and templates suggest yields would not be significantly different. Variation between individual pathologists, changes in the institution's analysis practice and packet submission practices all confound the interpretation of mean lymph node yield, however. Group 2 did have a higher positive lymph node rate. One explanation for the difference in node positivity is the practice of the group 1 surgeon to abort the procedure in cases of macrometastasis (Epstein definition of >5 mm).

Like most studies, we only accounted for symptomatic VTEs, so there may be a difference in asymptomatic VTE rates between the two groups. We also note the lower rate of lymph node dissection in group 2. It is possible that the rate of VTE in group 2 may have been higher if the rates of lymph node dissection were the same. When we evaluated the VTE rate from both groups with or without lymph node dissection, however, there was not a statistical difference. Of all the patients who underwent a node dissection, eight (1.5%) patients from group 1 experienced a VTE compared with two (0.7%) patients from group 2, P=0.31. This suggests that the difference in lymph node dissection rate between the two groups is unlikely to be a significant confounder.

Last, we were unable to specify exactly how many patients had preexisting clotting disorders or a history of VTE in each group from our database. Because these data were not prospectively recorded, we could not evaluate this as an independent risk factor and we do not know the number of patients in group 2 who received prophylaxis. Clearly, this would be a group at risk for VTEs and would be reasonable patients to consider expanding prophylaxis precautions.

Conclusions

Overall, the incidence of VTE in patients undergoing robot-assisted prostatectomy is low. Our data suggest that the perioperative administration of heparin for VTE prophylaxis to those patients already receiving SCDs may be unnecessary for the average risk patient population.