The Doctor’s Advocate | Fourth Quarter 2020
An Ounce of Prevention

The Role of Pretest Probability in the Evaluation of Suspected Venous Thromboembolism

Howard Marcus, MD, FACP

The overall incidence of venous thromboembolism (VTE)—including both deep vein thrombosis (DVT) and pulmonary embolism (PE)—is unknown, but, according to the CDC, it could be as many as one to two cases per 1,000 each year in the United States. Estimates suggest that 60,000 to 100,000 people in the U.S. die of VTE every year. Rudolf Virchow, a 19th-century German physician, described risk factors for VTE. Virchow’s Triad of hypercoagulability, stasis, and endothelial injury continues to provide us with a model for identifying patients at risk for VTE.

The presenting signs and symptoms of VTE are often vague and nonspecific, and early diagnosis—often crucial to the patient’s outcome—may be challenging. For example, in patients with a DVT, the finding of leg edema is 97 percent sensitive (the vast majority of patients with a DVT have leg edema), but only 33 percent specific (most patients with leg edema do not have a DVT). Pain is 86 percent sensitive, but only 19 percent specific, and warmth is 72 percent sensitive and 48 percent specific. Likewise, symptoms of PE—such as cough, chest pain, and dyspnea—are nonspecific and usually without pathognomonic physical findings.

The following case illustrates the challenges of making an early and accurate diagnosis of VTE:

A 42-year-old female taking an oral contraceptive fractured her right distal fibula and was treated with a walking cast and crutches. A month later she developed symptoms of waxing and waning right upper quadrant pain associated with deep inspiration. She was evaluated at an urgent care center, but no specific diagnosis was made. Two weeks later she presented to her internist with persistent symptoms. A D-dimer blood test was positive, and computed tomographic pulmonary angiogram (CTPA) showed multiple bilateral PE. The patient was treated with rivaroxaban anticoagulant as an outpatient and had a full recovery.

It is clear from this case that recognition of signs, symptoms, physical findings, and risk factors is essential for the early diagnosis of VTE. Risk factors include recent surgery, obesity, previous VTE, malignancy, estrogen therapy (or equivalent), and possibly testosterone therapy related to secondary erythrocytosis.

Accurate and timely diagnosis of VTE can be improved with the use of diagnostic guidelines such as Wells criteria and scoring for both DVT and PE. The Wells criteria not only focus the clinician on relevant history, risk factors, and physical findings, but they also provide a method for calculating the likelihood of a diagnosis—the pretest probability—of VTE leading to a rational approach for cost-effective and accurate diagnosis. Diagnostic tools include high-sensitivity D‑dimer, venous complete duplex ultrasound (CDUS), and CTPA. 

The D-dimer screening test has excellent diagnostic sensitivity for VTE. Unfortunately, however, D-dimer is also highly nonspecific and elevated for many reasons unrelated to VTE, particularly in hospitalized patients. Therefore, its value is primarily in the outpatient setting, where false-positive D-dimer results are less likely.

Because physiologic D-dimer levels increase with age for patients over age 50, diagnostic efficiency may be improved with age-adjusted D-dimer (age multiplied by 10 µg/L). For example, in the case of a 70-year-old patient, an age-adjusted negative D-dimer may be considered below 700 µg/L.

Deep Vein Thrombosis

Venous CDUS from thigh to ankle is the preferred diagnostic test for the diagnosis of suspected DVT. CDUS involves compression of the deep veins from the inguinal ligament to the ankle, including posterior tibial and peroneal veins in the calf and the common femoral vein in the thigh. CDUS is highly accurate in the diagnosis of DVT. The three-month risk of VTE after a negative CDUS is only 0.57 percent. If CDUS is not available, an extended compression ultrasound can be performed, but this test is limited to the thigh and does not evaluate for DVT of the calf. It is important that you know which type of ultrasound test the patient received.

Determine the probability for suspected DVT by calculating the patient’s Wells score for DVT. Points are assigned for criteria such as leg tenderness, leg and calf swelling, a history of recent bedrest or major surgery, orthopedic casting, and active cancer. Patients with a Wells score of 0 points have low probability of DVT, 1 to 2 points indicates moderate probability, and 3 or more points is high probability for DVT (FIGURE 1). Another version of the Wells criteria is the modified Wells score. It classifies patients into only two groups: unlikely (0 or 1 point) and likely (2 points or greater) (FIGURE 2).

Patients with a low probability of DVT should have the D-dimer test performed. A negative result excludes DVT, and no further testing is required because a negative D-dimer is highly reliable in patients in the low-risk category. Patients with a low-probability Wells score and a positive D-dimer need further evaluation and should have a CDUS.

Patients with a high probability of DVT by Wells score should not undergo a D-dimer test but should go directly to CDUS. The logic for this recommendation is that, in patients with a high probability of DVT, a negative screening D-dimer is insufficient to confidently exclude the diagnosis of DVT. If the CDUS is negative, then DVT has been ruled out except for highly unusual circumstances, such as suspected iliac vein thrombosis, upper extremity DVT, in toto propagation of a DVT to the lungs, or significant Pulmonary Embolism Rule-out Criteria (PERC) findings.

In patients with a moderate probability of DVT by Wells score, a negative D-dimer may be sufficient to rule out VTE, but experts disagree on whether moderate-risk patients should be evaluated with CDUS, similar to high-risk patients. To avoid confusion, consider the two-tier modified Wells score that simplifies decision making. 

Pulmonary Embolism

The logic of evaluation for PE is similar to DVT. Symptoms of PE are nonspecific and include dyspnea at rest (73 percent), chest pain (66 percent), cough (37 percent), and orthopnea (28 percent), but usually not hemoptysis (13 percent). Lower extremity findings occur in about 50 percent of patients with PE. Oxygen saturation may be normal, and chest findings on auscultation are most often negative. If present, abnormal EKG findings, such as tachycardia and minor ST-T changes, are usually nonspecific. The classic S1Q3T3 electrocardiographic abnormality occurs in less than 10 percent of cases.

PERC helps identify patients with a very low probability of PE. The eight favorable criteria are age under 50 years, heart rate under 100, oxygen saturation greater than 95 percent, no hemoptysis, no estrogen use, no prior DVT or PE, no unilateral leg swelling, and no surgery/trauma within the prior four weeks. In the case of a hemodynamically stable nonpregnant patient with no risk factors, the diagnosis of PE can be ruled out by PERC in patients who fulfill all eight criteria.

For patients with suspected PE, the Wells criteria for PE provide pretest probability (FIGURE 3). Positive Wells criteria for PE include clinical symptoms of DVT, heart rate over 100, immobilization, previous DVT/PE, hemoptysis, and malignancy.

Patients at low risk for PE by Wells score should undergo D-dimer testing. Negative results indicate that they do not have VTE and do not require further testing. Patients with a positive D-dimer should undergo CTPA.

In patients at high risk for PE by Wells criteria, the screening D-dimer is not indicated, and a CTPA should be performed. Patients receive a ventilation/perfusion (V/Q) lung scan only if the CTPA is contraindicated or cannot be performed for other reasons (e.g., marked obesity). The V/Q scan may also be indicated if the CTPA is inconclusive.

The vagaries of the presentation of VTE and its potential for catastrophic complications create challenges for clinicians. However, the identification of risk factors and the prudent use of pretest probability assessment by Wells criteria with scoring for both DVT and PE help the clinician focus on the important clinical factors and signs of VTE and provide a diagnostic pathway. For patients with low pretest probability of VTE, a negative D-dimer should obviate the need for imaging studies such as leg ultrasound or CTPA. For patients with moderate-to-high pretest probability of VTE, omit the time and expense of the D-dimer and send the patient directly for imaging.

See more issues