Venous Thromboembolism in Gynecologic Surgery

Deep venous thrombosis (DVT) and pulmonary embolism (PE), also referred to as venous thromboembolic events (VTE), are two major complications after gynecologic surgery that can result in significant morbidity and mortality.  The prevalence of DVT after gynecologic surgery varies depending upon the method used for diagnosis.  When I¹²⁵ fibrinogen leg scanning is performed the DVT prevalence ranges from 15% to 30% .The diagnosis of DVT is approximately 3% when diagnosed clinically, and fatal PE occur in 0.2–0.9% of patients.¹

Although many DVTs may be asymptomatic, the presence of a DVT is strongly associated with the development of a symptomatic PE.²  Death from a PE occurs rapidly, with most patients succumbing within 30 minutes of the first clinical symptoms.  Because inadequate time exists for therapeutic interventions, strategies to lower the rate of fatal PE must be directed at preventing the occurrence of DVT. Identification of high-risk patients and institution of consistent, effective thromboprophylaxis can reduce the incidence of this common, often preventable cause of postoperative mortality.

Two million Americans will develop a DVT each year, and almost a third will also suffer a PE, resulting in 60,000 deaths each year.³  The incidence of a first VTE is 1–2 per 1000 individuals per year.^{4, 5}  The case-fatality rate for pulmonary embolism is 11–12%, although this percentage is higher in cancer patients and lower in young patients.^{5, 6} When diagnosed by I¹²⁵ fibrinogen leg scanning, the incidence of DVT in gynecologic surgery varies widely depending on the risk factors of the individual patient. About 14% of patients undergoing gynecologic surgery for benign indications develop thromboembolism.⁷  DVT has been observed in 38% of gynecologic oncology patients postoperatively.⁸

A recent study of more than 2000 patients with cancer undergoing surgery found a 2% rate of clinical VTE formation, despite the fact that over 80% of patients received in-hospital prophylaxis.⁹  Total mortality was 1.72% within 35 days of surgery and, despite prophylaxis, 46% of the deaths were caused by VTE.  It is important to identify and treat high-risk patients, as consistent, systematic thromboprophylaxis will help to decrease the risk of a potentially fatal VTE. 

In 1858, Virchow reported that the development of thromboembolism is dependent on three factors: hypercoagulability, venous stasis, and vessel wall injury (venous endothelial damage).¹⁰ Patients undergoing gynecologic surgery are predisposed to thromboembolism because of alteration in one or more of these factors. Perioperative and postoperative immobility can adversely affect the drainage of blood from the lower extremity, promoting the development of a DVT.¹¹  Pelvic masses, a gravid uterus, surgically induced hematomas or lymphocysts also can lead to venous stasis.¹²  Additionally, vessel wall injury can result from surgical dissection or malignant growth of a tumor into vascular tissues. Coagulation can result from decreased fibrinolytic activity associated with an operative procedure.¹³  Elevated coagulation factors (i.e., factors I, V, VIII, IX, X, and XI), activated intermediates (thrombin-antithrombin III complexes), and platelet abnormalities contribute to a hypercoagulable state in the gynecologic oncology patient.¹⁴ Cancer cells also secrete procoagulants (e.g., tissue factor and cancer procoagulant) as well as factors that affect endothelial permeability (e.g., vascular endothelial growth factor) and promote fibrin deposition.¹⁵

Once a thrombus is formed, the risk of PE depends on the location of the clot. In a prospective study of 382 gynecologic oncology patients, 17% of subjects developed DVT. Eighty-five percent of these thrombi were located in the calf veins.¹⁶  Nearly one third of these calf thrombi lysed spontaneously, and 65% did not propagate during postoperative surveillance. Only 4% propagated to the proximal leg veins, and an additional 4% became symptomatic pulmonary emboli. These findings emphasize that calf vein thrombosis, although a frequent event, is of minimal clinical significance. Moreover, 40% of the gynecologic oncology patients, who developed postoperative symptomatic PE, had no evidence of DVT in the legs, emphasizing that pelvic vein thrombi pose a high risk of PE.¹⁶  Prevention of PE secondary to proximal vein thrombosis also may prove difficult because half of these thrombi are clinically silent.¹⁷

Immobilization is a major risk factor for developing a VTE, with a 9-fold increase seen in patients undergoing bed rest.  Hospitalization and surgery are also associated with an increased thrombosis risk, with odds ratios of 11.1 and 5.9, respectively.¹⁸

A prospective study of 411 women undergoing major gynecologic surgery for both benign and malignant conditions identified the following clinical findings as independent risk factors based on a multivariate analysis: prior history of VTE, cancer, increasing age, African American, ankle edema or varicose veins, prolonged surgical time and prior radiation therapy.  Patients undergoing pelvic exenteration or radical vulvectomy with inguinal-femoral lymphadenectomy were at increased risk as well.

Patients should be classified preoperatively into one of four risk categories to determine the appropriate thromboprophylaxis regimen.  VTE risk is determined based on procedure type and duration, age, and presence of other risk factors (Table 1). Not only do patients have different risk factors, but also not all prophylactic regimens are appropriate or effective in certain risk groups.  The proper risk classification is therefore important in order to prescribe the best prophylaxis regimen.  Additionally, there are certain risk factors that are associated with an especially high risk of developing VTE.  A retrospective review of over 1800 patients identified age greater than 60 years, presence of cancer, and history of DVT as being closely associated with postoperative VTE, despite the use of intermittent pneumatic compression prophylaxis.¹⁹  Women with two or three of these risk factors had a 3.2% incidence of VTE, compared with an incidence of 0.6% in women who had none or one risk factor.  Consideration for more intense prophylaxis is warranted in this highest-risk population. Recommendations for VTE prevention are described in Table 2. 

Table 1. Venous thromboembolism risk factors

Geerts WH, Bergqvist D, Pineo GR, Heit JA, Samama CM, et al. Prevention of venous thromboembolism. The Eighth ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 2008;133;381S–453S.

Table 2. Risk classification and recommended thromboprophylaxis

LDUH, low dose unfractionated heparin; LMWH, low molecular-weight heparin; GCS, graduated compression stockings; IPC, intermittent pneumatic compression; VKA, vitamin K antagonist; INR, international normalized ratio.

*Risk factors listed in Table 1.
**In patients at high risk for bleeding, use IPC or GCS, consider switch to anticoagulant thromboprophylaxis when bleeding risk decreases.

Modified from Geerts WH, Bergqvist D, Pineo GR, Heit JA, Samama CM, et al. Prevention of venous thromboembolism. The Eighth ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 2008;133;381S–453S.

 
Many environmental, inherited, and acquired risk factors influence coagulability.  Most inherited factors do not result in VTE formation until a precipitating event such as pregnancy, surgery, or exogenous hormone use occurs.²⁰  The most common mutations found in patients with a VTE are factor V Leiden mutation and prothrombin gene mutation G20210A.  Presence of one of these conditions in the setting of pregnancy or major surgery confers an elevated VTE risk and may place a patient into the highest risk category.

Identified in 1993 as the major cause of activated protein C (APC) resistance, factor V Leiden is the most common inherited thrombophilia, and is carried by 5% of Caucasians.^{21, 22}  Half of patients with thrombophilia and 20% of patients with VTE carry this mutation.  Heterozygotes have a 3–8-fold increase of VTE, while homozygotes are more severely affected, with a 50–80-fold increase in risk.²³  Prothrombin 20210A mutation is another common mutation found almost exclusively in Caucasians and in 6% of patients with VTE.  This mutation causes an abnormally elevated prothrombin level, which results in a VTE rate three times higher than baseline.²⁴  Factor V Leiden mutation and prothrombin mutation may be diagnosed by DNA analysis; factor V Leiden mutation can also be detected in an abnormal APC resistance assay.

Rates of VTE after gynecologic surgery are similar to those reported in the general surgery literature, and average around 16% in an untreated population.²⁵  Graded compression stockings (GCS), intermittent pneumatic compression (IPC) devices, low-dose unfractionated heparin (LDUH), and low molecular weight heparin (LMWH) have each been shown to effectively reduce VTE development.  Two randomized trials and a large retrospective series found the incidence of VTE to be 1–6.5% in a gynecologic oncology patient population treated with one of the above modalities.^{19, 26, 27}  A combined regimen of medical and mechanical prophylaxis may improve efficacy, especially in the highest-risk patients.  Although limited data exist to support this approach in gynecology patients, studies from the general surgical and neurosurgical literature suggest significant benefit from a combined regimen.^{28, 29} Until more evidence is accumulated, patients undergoing laparoscopic surgery should be stratified by risk category, and provided prophylaxis, similar to patients undergoing laparotomy.

Prophylaxis options

DVT formation can be reduced by a number of prophylactic methods. There are two types of prophylactic options: mechanical and pharmacologic methods.   Mechanical methods reduce venous stasis and may promote endogenous fibrinolysis.  Pharmacologic methods prevent clot formation by affecting different points on the clotting cascade.  Cost, benefit, risk, and feasibility of each method must be weighed in determining the appropriate prophylaxis for an individual patient.

Graduated compression stockings

Most postoperative thrombi occur in the capacitance veins of the calf, and develop within 24 hours of surgery.  In addition to early postoperative ambulation and elevating the foot of the bed, graduated compression stockings (GCS) prevent pooling of blood in the calves.  A Cochrane review of randomized, controlled trials found a 50% reduction in DVT formation with GCS, although they were more effective when combined with a second prophylactic method.³⁰ Low cost and simplicity are the main advantages of using GCS.  Correct fit is essential, as tight or improperly fitted stockings can cause an increase in venous stasis by acting as a tourniquet at the knee or mid-thigh.³¹  Knee-length GCS are as effective as thigh-length and should be preferentially used.³²

Pneumatic compression

Intermittent pneumatic compression (IPC) devices regularly compress the calf with an inflatable pneumatic sleeve, thereby reducing venous stasis.  When used during and after major gynecologic surgery, IPC devices are as effective as LDUH and LMWH in reducing DVT incidence.^{26, 27, 33}  Most studies have included a small number of patients and are underpowered to prove efficacy in lowering PE incidence or mortality.  The benefits of IPC have been postulated to include an increase in systemic fibrinolysis.^{34, 35}  However, larger series have failed to confirm this finding.^{36, 37}  IPC devices should be used continuously until ambulation and preferably discontinued at the time of hospital discharge.¹  In a study of patients with gynecologic malignancies undergoing surgery, IPC devices were placed intraoperatively and continued for 5 days.³⁸  IPC use was associated with a three-fold reduction in VTE. 

Low-dose unfractionated heparin

Low-dose unfractionated heparin (LDUH) is the most extensively studied method of thromboprophylaxis.  When administered subcutaneously starting 2 hours before surgery and continued every 8–12 hours postoperatively, numerous controlled trials have found LDUH effective in preventing DVT.¹  Two large meta-analyses of randomized clinical trials in general surgery patients showed a two-thirds reduction in fatal PE with the use of LDUH every 8 hours over placebo or no prophylaxis.^{17, 39}

Patients undergoing major gynecologic surgery for benign indications also benefit from LDUH given in a preoperative dose and postoperatively at 12-hour intervals.¹  This dosing schedule was found to be ineffective in patients with gynecologic cancer.⁴⁰ However, the administration of 5000 U of heparin beginning 2 hours preoperatively and continued every 8 hours postoperatively does provide effective VTE prophylaxis in women at high risk with gynecologic malignancies.⁴¹

Advantages of LDUH include well-studied efficacy and low cost.  A major concern with perioperative LDUH use is increased intraoperative and postoperative bleeding complications.  While surgical blood loss does not seem to be affected by preoperative LDUH administration, an increase in postoperative bleeding has been noted, specifically in wound hematoma formation.  Additionally, use for more than 4 days warrants monitoring of platelet counts, as 6% of patients will experience heparin-induced thrombocytopenia.^{17, 42}

Low molecular weight heparin

Advantages of low molecular weight heparin (LMWH) include greater bioavailability and once-daily dosing.  These benefits result from a longer half-life, more predictable pharmacokinetics, and equivalent efficacy when compared with prophylactic dosing of LDUH.⁴³  LMWH has more anti-factor Xa and less anti-thrombin activity than LDUH, which may decrease medical bleeding and wound hematoma formation.  However, LMWH is more expensive than LDUH. Heparin-induced thrombocytopenia is very rarely observed with LMWH, and screening for this is not necessary.⁴⁴

Since initial reports in 1985,⁴⁵ multiple well-designed trials have shown LMWH to be a reliable method of thromboprophylaxis.  A Cochrane review of randomized, controlled trials in gynecologic patients undergoing major surgery found LMWH and LDUH equally useful in preventing DVT.⁴⁶  Effective VTE prophylaxis was also shown in patients undergoing surgery for gynecologic malignancies.  Equivalent risk reductions were seen with the use of preoperative and daily postoperative LMWH when compared with IPC devices.²⁷  A major prospective trial including 2373 patients showed a 2% incidence of clinical VTE in patients undergoing general, urologic, and gynecologic surgery for cancer who received LMWH prophylaxis.⁹  A retrospective analysis of more than 3500 patients showed a statistically significant reduction in DVT and fatal PE in patients receiving LMWH prophylaxis compared to no LMWH, although the authors did not control for use of mechanical methods.⁴⁷

Duration of prophylaxis varies depending on risk factors.  Major risk factors for the development of a clinical VTE include age greater than 60, cancer, prior VTE, and prolonged surgery or bed rest.^{9, 19}  Forty percent of patients with cancer who develop a VTE will do so more than 21 days after surgery.⁹  A placebo-controlled trial of LMWH administered for 1 week versus 4 weeks postoperatively showed a 60% reduction in VTE with 4 weeks of treatment and no increase in bleeding or thrombocytopenia.⁴⁸  Patients at the highest risk for VTE may benefit from prolonged LMWH prophylaxis.

Dual prophylaxis

The combined use of two prophylactic methods has been examined in the general surgical literature, specifically in patients undergoing colorectal surgery.  A Cochrane review of 19 studies showed that LDUH combined with GCS was four times more effective at VTE prevention than LDUH alone.²⁹  A randomized trial of 307 patients undergoing neurosurgical procedures showed a significant reduction in VTE with LMWH and GCS combined over GCS alone.²⁸  A decision analysis in high-risk gynecologic oncology patients determined that combined IPC and LMWH use is cost-effective.⁴⁹

No data exist in the gynecologic literature on the benefits of using a combination of mechanical and pharmacologic prophylaxis.  However, possessing two of three identified risk factors for failing IPC (age >60, cancer, prior VTE) places patients in the highest risk category for the development of VTE.¹⁹  As a result, the use of a combined approach possesses inherent appeal, as it may reduce both hypercoagulability and venous stasis in highest-risk surgical patients.  Though data from randomized trials in gynecologic patients are lacking, a combined approach seems appropriate in the highest-risk patients, and this practice is endorsed by the Eighth ACCP Consensus Conference.¹

More than 80% of patients with a VTE have one or more identifiable risk factors. The signs and symptoms of DVT of the lower extremities include pain, edema, erythema, and prominent vascular pattern of the superficial veins.  On exam they may have calf tenderness, especially with dorsiflexion (Homan’s sign) and occasionally a palpable venous “cord”.  Rarely, a patient may present with massive edema, cyanosis, and ischemia of the lower extremity.  This is called phlegmasia cerulea dolens, and is considered to be a limb-threatening medical emergency.  These signs and symptoms are relatively nonspecific; 50–80% of patients with these symptoms will not actually have DVT.⁵⁰  Conversely, approximately 80% of patients with symptomatic PE have no signs or symptoms of thrombosis in the lower extremities.^{51, 52}  Because of the lack of specificity when signs and symptoms are recognized, additional diagnostic tests should be performed to establish the diagnosis of DVT.

Many of the signs and symptoms of PE are associated with other, more commonly occurring pulmonary complications following surgery. The classic findings of pleuritic chest pain, hemoptysis, shortness of breath, tachycardia, and tachypnea should alert the physician to the possibility of PE. Many times, however, the signs are much more subtle and may be suggested only by a persistent tachycardia, a slight elevation in the respiratory rate or a decrease in pulse oxymetry measurements.  Auscultation often reveals normal breath sounds.  Clinicians should maintain a high level of suspicion for this common postoperative complication and pursue further diagnostic testing as indicated.

Laboratory findings

Laboratory testing when DVT or PE is suspected includes D-dimer levels and an arterial blood gas (ABG).  A D-dimer assay measures circulating amounts of fibrin mesh breakdown products, and an elevated level implies the presence of a significant clot.  This test has poor specificity as it can be markedly elevated in other circumstances, including recent surgery.  However, it has good negative predictive value, and a normal D-dimer is predictive of a low likelihood of VTE.  An ABG will help diagnose hypoxia and acidosis, which is especially helpful in the presence of normal peripheral oxygen saturation.  In addition, calculation of the alveolar-arterial (A-a) oxygen gradient from an ABG can diagnose ventilation-perfusion mismatch, when a significant volume of lung tissue is hypoperfused due to obstruction of the pulmonary arterial system by clot.  Expected A-a gradient can be estimated by the formula (4 + patient age)/4.

Radiologic findings

DEEP VENOUS THROMBOSIS  

Radiologic testing for DVT includes B-mode duplex Doppler ultrasound, contrast venography, and magnetic resonance venography (MRV).  Doppler ultrasound is currently the most common technique for the diagnosis of symptomatic DVT.  With duplex Doppler imaging, the femoral vein can be visualized and clots may be seen directly.  Compression of the vein with the ultrasound probe tip allows assessment of venous collapsibility; the presence of a thrombus diminishes vein wall collapsibility.  Doppler ultrasound has a greater than 90% sensitivity to detect a clot in the proximal extremity, but is less sensitive for calf or pelvic thrombosis.^{53, 54}

Although contrast venography has been the “gold standard” for diagnosis of DVT, other diagnostic studies are accurate when performed by a skilled technologist and, in most patients, may replace the need for routine contrast venography. Venography is moderately uncomfortable, requires the injection of a contrast material that may cause allergic reaction or renal injury, and may result in phlebitis in approximately 5% of patients.⁵⁵  However, if noninvasive imaging is normal or inconclusive and the clinician remains concerned given clinical symptoms, venography should be obtained to obtain a definitive answer.  MRV has a sensitivity and specificity comparable to venography.⁵⁶  In addition, MRV may detect thrombi in vena cava, pelvic, or ovarian veins that are not imaged by venography. The primary drawback to MRV is the time involved in examining the lower extremity and pelvis, as well as the expense of this technology.

PULMONARY EMBOLISM         

PE may be diagnosed by a number of radiologic methods, including chest radiograph (CXR), ventilation-perfusion (V-Q) scan, helical/spiral chest computed tomography (CT), and pulmonary angiography.  CXR is typically normal in patients with PE, but occasionally may show a wedge-shaped pulmonary infarct with an adjacent effusion.  A V-Q scan detects areas in the lung with a ventilation-perfusion mismatch.  Its sensitivity is affected by the pre-test probability of PE, and in clinical practice it has an accuracy of 15–86%.  A normal V-Q scan, however, virtually excludes PE.  V-Q scans are also helpful in patients with renal dysfunction or other contraindications to intravenous (IV) contrast.  Chest CT is the most frequently used diagnostic method in patients with a suspected PE and no contraindication to IV contrast.  The accuracy can vary with the experience of the interpreting physician.  With the addition of venous-phase imaging the sensitivity and specificity approach 90 and 95 percent, respectively.⁵⁷  Pulmonary angiography is considered the gold standard in diagnosing PE, but is highly invasive and only slightly more accurate than chest CT; thus, it is rarely used.

The most common findings on an electrocardiogram (ECG) in patients with PE are tachycardia with nonspecific ST-segment and T-wave changes.  Classic, but infrequent, findings include atrial fibrillation, S1Q3T3 pattern, right ventricular strain, and new, incomplete right bundle branch block.  Patients with a suspected PE who are too unstable for helical chest CT may undergo bedside echocardiogram; a massive PE is generally associated with increased right ventricular (RV) size, decreased RV function, and tricuspid regurgitation.

Patients suspected of PE should be evaluated initially by CXR, ECG, and ABG assessment.  Any evidence of abnormality should be further evaluated by V-Q lung scan, or a spiral CT scan of the chest.  Unfortunately, a modest percentage of lung scans may be interpreted as “indeterminate”. In this setting, careful clinical evaluation and judgment are required to decide whether pulmonary arteriography should be obtained to document or exclude the presence of a pulmonary embolism.

Differential diagnosis

The differential diagnosis for DVT includes cellulitis, lymphedema, extrinsic venous compression due to mass or lymphadenopathy, muscular or ligamentous injury, compartment syndrome, and synovial cyst (Baker’s cyst).  DVT may also occur in isolation or as a causative factor or sequelae of one of these conditions.  DVT can be misdiagnosed as an infectious process, given the findings of a red, warm, swollen extremity.  The presence of fever or an elevated white blood cell count is helpful in identifying an infection; however, Doppler ultrasound may be the most efficient method to exclude a primary or secondary DVT.

PE can present with relatively nonspecific findings and may imitate any of the following diagnoses:

Pneumonia
Pulmonary edema
Pneumothorax/pneumomediastinum
Asthma
Emphysema/COPD
Pleuritis
Airway obstruction
Epiglottitis
Foreign body
Myocardial infarction
Angina
Pericarditis
Aortic dissection
Esophageal spasm or rupture
Musculoskeletal/costochondritis
Anxiety

A thorough history and physical exam in addition to a CXR and ECG can help to narrow the wide range of possible diagnoses.

The treatment of postoperative DVT requires the immediate institution of anticoagulant therapy.  Treatment may be with either unfractionated heparin or low molecular weight heparins, followed by 6 months of oral anticoagulant therapy with warfarin. Prolonged anticoagulation (life-time) is recommended for women who continue to have active cancer (i.e., those not in remission after treatment) as they remain at very high risk to re-thrombose.

Unfractionated heparin

After VTE is diagnosed, UFH should be initiated to prevent proximal propagation of the thrombus and allow physiological thrombolytic pathways to dissolve the clot.  An initial IV heparin bolus followed by a continuous infusion is commonly administered according to a weight-based heparin nomogram (Table 3).  Heparin dosage is adjusted to maintain activated partial thromboplastin time (APTT) levels at a therapeutic level 1.5–2.5 times the control value.  Initial APTT should be measured after 6 hours of heparin administration and the dose adjusted as necessary.  Patients having subtherapeutic APTT levels in the first 24 hours may have a risk of recurrent thromboembolism up to 6 times the risk of patients with appropriate levels.⁵⁸  Patients, therefore, should be managed aggressively using IV heparin to achieve prompt anticoagulation.  Oral anticoagulation (warfarin) should be started on the first day of the heparin infusion.  International normalized ratio (INR) should be monitored daily until a therapeutic level is achieved (INR 2.0–3.0).  The change in the INR resulting from warfarin administration often precedes the anticoagulant effect by approximately 2 days, during which time low protein C levels are associated with a transient hypercoagulable state.  Therefore, heparin should be administered until the INR has been maintained in a therapeutic range for at least 2 days, confirming proper warfarin dose.  IV heparin may be discontinued in 5 days if an adequate INR level has been established.

Table 3. Heparin administration for treatment of deep vein thrombosis or pulmonary embolism: weight-based nomogram

U, units; kg, kilogram; h, hour; APTT, activated partial thromboplastin time.

Raschke RA, Reilly BM, Guidry JR, Fontana JR, Srinivas S. The weight-based heparin dosing nomogram compared with a “standard care” nomogram. Ann Intern Med 1993;119: 874–881

Low molecular weight heparin 

Low molecular weight heparins (enoxaparin and dalteparin) have been shown to be effective in the treatment of VTE and have a cost-effective advantage over IV heparin in that they may be administered in the outpatient setting.  The dosages used in treatment of thromboembolism are unique and weight adjusted according to each LMWH preparation.  Because LMWH have a minimal effect on activated APTT, serial laboratory monitoring of APTT levels is not necessary.  Similarly, monitoring of anti-Xa activity, except in difficult cases or those with renal impairment, has not been shown to be of significant benefit in a dose adjustment of LMWH.  The increased bioavailability associated with LMWH allows for once- or twice-daily dosing, potentially making outpatient management for a subset of patients an option.  A meta-analysis involving nearly 2000 patients from 12 trials suggests that LMWH is as safe and effective as unfractionated heparin in preventing recurrent thromboembolism.⁵⁹

Oral anticoagulants: warfarin

In most cases, the conversion from IV heparin or LMWH to oral warfarin may start on the initial day of therapy.  Both heparin and warfarin are given and the heparin is discontinued when the warfarin has reached a therapeutic INR of 2–3 for 2 consecutive days.  Initially, the INR should be monitored daily in order to appropriately adjust the warfarin dose.  Once a stable warfarin dose is established, the INR may be checked less frequently.  Warfarin may be a difficult drug to administer to some patients, especially if their nutrition is inadequate, their oral intake is variable or if they require prolonged use of antibiotics or other drugs which might alter the metabolism of warfarin.  This is particularly common in women with advanced ovarian cancer. Given the wide variation in the INR in many of these patients who are then predisposed to either bleeding complications or re-thrombosis, we have found that it is safer to use subcutaneous LMWH (at therapeutic doses) for prolonged therapy.

Treatment of pulmonary embolism  

The treatment of PE is as follows:

1. Immediate anticoagulant therapy, identical to that outlined for the treatment of DVT, should be initiated.
2. Respiratory support, including oxygen and bronchodilators and an intensive care setting, may be necessary.
3. Although massive pulmonary emboli are usually quickly fatal, pulmonary embolectomy has been performed successfully on rare occasions.
4. Pulmonary artery catheterization with the administration of thrombolytic agents bears further evaluation and may be important in patients with massive pulmonary embolism.

Vena cava interruption may be necessary in situations in which anticoagulant therapy is ineffective in the prevention of re-thrombosis and repeated embolization from the lower extremities or pelvis. A vena cava umbrella or filter may be inserted percutaneously above the level of the thrombosis and just below the renal veins.

In most cases, however, anticoagulant therapy is sufficient to prevent repeat thrombosis and embolism and to allow the patient’s own endogenous thrombolytic mechanisms to lyse the pulmonary embolus.

DVT and subsequent PE are a significant source of morbidity and mortality in gynecologic surgery. Most patients experiencing fatal thromboembolism are diagnosed at autopsy. Prophylaxis against DVT therefore should be used in an effort to decrease the incidence of PE. Each patient should be assessed for thromboembolic risk. Women at low risk benefit from GCS. Moderate and high-risk patients should be treated with LDH, LMWH, GCS, or IPC. Highest-risk patients (i.e., older than 60 years, history of DVT or PE, presence of neoplasm) may benefit from combined and prolonged prophylaxis.