Published in the February 2015 issue of Today’s Hospitalist
WHEN YOU THINK of moving targets in clinical medicine, few topics come close to venous thromboembolism (VTE). That’s because doctors have to make a slew of clinical calls to assess patients’ risk, figure out how long (or if) patients should be hospitalized, and decide how long to continue anticoagulation.
“People are uncertain because recommendations have changed so much over the years,” said Tracy Minichiello, MD, director of anticoagulation and thrombosis services at the San Francisco VA Hospital.
Speaking at last fall’s management of the hospitalized patient conference at the University of California, San Francisco, Dr. Minichiello led hospitalists through the evidence and strategies she relies on to diagnose and manage VTE.
The first step in assessing DVT is determining patients’ pre-test probability. Most doctors are familiar with the Wells scoring system, which assigns points for DVT symptoms and subtracts them if an alternative diagnosis is more likely.
“The Wells criteria accurately stratify patients to low, moderate or high pre-test probability,” Dr. Minichiello reminded the audience. “For patients with low pretest probability, start with a D-dimer. If it’s negative, you’re done.”
To rule out DVT in patients with moderate pre-test probability, start with either a D-dimer or ultrasound. If you start with a sensitive D-dimer that proves to be negative, DVT is ruled out. If you get a whole leg ultrasound and it’s negative, “you’re done.” And if a patient has a negative D-dimer and a negative proximal ultrasound, you would not need further testing.
But if you start with a proximal ultrasound and it is negative but no D-dimer is drawn, “that’s not enough to rule out DVT in patients with moderate pre-test probability,” she noted. “About 15% can have an isolated calf vein thrombosis, and a certain percentage of those will progress.” In such a case, the recommendation is to get a follow-up ultrasound in a week. If, however, you get a whole leg ultrasound that is negative, you won’t need either D-dimer or serial testing.
For patients with high pre-test probability, “go right to the ultrasound,” Dr. Minichiello said. “If you get only a proximal ultrasound without a D-dimer and it is negative, the patient will need to return for another ultrasound in a week. However, if you get a highly-sensitive D-dimer that is negative in addition, serial ultrasound will not be necessary.”
Repeat ultrasound is also recommended for patients with extensive swelling and a high pre-test probability who have a negative ultrasound and either a positive D-dimer or no D-dimer.
“These patients may have more proximal DVT like iliac vein thrombosis, which you can miss on ultrasound,” said Dr. Minichiello. “Pursuing CT or MR venography might be the way to go.”
As for patients with suspected recurrent DVT, “a highly sensitive D-dimer is preferred,” she said. You can rule out DVT in patients who have a history of DVT with a negative ultrasound and a negative D-dimer. “But if the initial ultrasound is negative and the D-dimer is positive, follow-up ultrasound is recommended.”
Guidelines from the American College of Chest Physicians published in the February 2012 issue of CHEST “seem to prefer proximal leg ultrasound, but they recommend whole leg ultrasound in patients unlikely to return for imaging,” she pointed out. And while guidelines may recommend a number of follow-up ultrasounds, “those are hard to get in many institutions. If you can get one in the next week, you’re lucky, so this is very institutionally based.”
To determine patients’ pre-test probability for PE, doctors can again use the Wells criteria. There’s also the Geneva scoring system, which bases assessment on only clinical factors.
But the revised and simplified Geneva criteria have been validated for only outpatients, Dr. Minichiello explained. And while a few studies have examined the performance of the Wells score in the hospital, it is not currently recommended in this setting.
Then there’s clinical gestalt, which researchers evaluated in a study published in the August 2013 Annals of Emergency Medicine. “In the study, that looked as good as clinical prediction scores,” she noted. “But the CHEST guidelines still recommend that you use some kind of scoring system.”
For patients with low pre-test probability, “if the D-dimer is negative, you can consider PE ruled out,” Dr. Minichiello said. “If it’s positive, go ahead and get a CT angiography.”
At the same time, “we know that D-dimer goes up as patients age,” she pointed out. Retrospective studies have looked at the traditional D-dimer cutoff of 500 mcg/L and found that, if adjusted for age, “we could potentially safely rule out PE without pursuing CT angiography.”
Those findings were confirmed with the ADJUST-PE trial in the March 19, 2014, Journal of the American Medical Association (JAMA). Researchers used the traditional cutoff only for patients under age 50; for older individuals, they used a cutoff that equaled 10 times the patients’ age and then followed them for three months.
“The thromboembolism rate of patients with a D-dimer above 500 but less than their age-adjusted cutoff was very low: 0.3%,” said Dr. Minichiello. At the same time, the percentage of patients over 75 for whom doctors could rule out PE without CT increased from 6% to 30%.
“That ruled out PE in one in every three-and-a-half patients, as opposed to one in every 16,” she pointed out. “This can definitely make a big difference in throughput in the ED.” Because the study evaluated six different D-dimer assays, “it’s important to know which assay you use to see if you can extrapolate results.”
Stratifying PE patients
For patients with PE, doctors need to risk-stratify them to guide triage and treatment. Subgroups include low risk (nonmassive), intermediate risk (submassive) or high risk (massive), with corresponding 30-day mortality risk as 1%, up to 15% or 30-plus% respectively.
Patients deemed to have a massive PE “are persistently hemodynamically unstable, which is not subtle,” Dr. Minichiello said. “These patients typically receive thrombolysis and are cared for in the ICU.”
But patients at low or intermediate risk “are difficult to differentiate as they’re both normotensive,” she explained, adding that it’s critical to make that distinction because disposition and treatment can be very different. Experts believe that the higher mortality rate in intermediate-risk patients may be due in part to right ventricular dysfunction (RVD), “which you cannot appreciate at the bedside,” she noted. ECHO has been the gold standard to diagnose RVD, but that “requires additional testing that is expensive and not always available.”
In differentiating low- from intermediate-risk patients, the Pulmonary Embolism Severity Index (PESI) scoring system, which assigns points for various clinical features, has been validated in several studies. A simplified PESI score has also been evaluated using only five factors to determine whether patients are at low or high risk.
One study looking to identify intermediate-risk patients was published by Thorax in January 2011. “Researchers looked at three prognostic variables in normotensive patients: ECHO, troponin and ultrasound of the lower extremities,” said Dr. Minichiello. Researchers included ultrasound results presumably because “mortality and recurrence rates are higher if you present with PE and also have acute DVT.”
Patients who had two out of three abnormal prognosticators had higher mortality rates than those with just one out of three. “ECHO plus ultrasound had the greatest positive predictive value for mortality,” she noted.
Another report, published online by Thorax in February 2011, provided “a reasonable starting protocol,” said Dr. Minichiello. Patients with confirmed PE but not a massive PE were evaluated with both a PESI score and a troponin.
According to this protocol, patients who are troponin-negative and have a low PESI score “are considered candidates for outpatient treatment,” she said. If, however, the troponin is high, “they proceed to ECHO or their CT RV/LV ratio is analyzed to look for evidence of right ventricular dysfunction. Patients with RV dysfunction and those deemed to be at intermediate- or high-risk by PESI would be admitted.”
Sending low-risk patients home?
For patients with nonmassive PEs, should you admit them for anticoagulation and monitoring? Or send them home with anticoagulation and arrange for close follow-up?
“If I was giving this talk six or 10 years ago and I asked if doctors should admit a low-risk patient with PE, most audience members would answer ‘yes,’ ” Dr. Minichiello pointed out. Sending low-risk PE patients home is a newer trend in the U.S., but one with a long track record in Europe and Canada. “Now, we are getting more comfortable with this.”
A trail looking at the outpatient treatment of acute PE was published in the July 2, 2011, issue of Lancet. Researchers randomized low-risk patients with a PESI score of 1 or 2 to either outpatient or inpatient treatment.
They found no difference in recurrent PE or mortality rates. The trial did exclude patients with renal insufficiency and those at high risk of bleeding or weight extremes. Researchers also ensured very robust follow-up.
“Obviously, you have to be really selective, and not just rely entirely on the low PESI score,” Dr. Minichiello advised. “Make sure patients are going to take their medications, follow up well and don’t have other reasons to be admitted. Patients sent home with PE really need a good safety net.”
As for hospitalized patients at intermediate risk, how long do they need to stay?
What is known as the PESI 48 may help you identify patients who need only an abbreviated hospital stay, said Dr. Minichiello. Researchers published in the February 2013 issue of the European Respiratory Journal admitted patients with a PESI score of 3 or more (indicating intermediate-risk PE), then recalculated patients’ scores 48 hours later.
“After 48 hours of inpatient care, patients found to have a PESI score of 1 or 2 “indicating low-risk PE ” had a very low mortality rate of 1.2% in the next 30 days,” she said. Those patients could presumably have been discharged. However, for patients who maintained the same score as on admission or higher, “their mortality rate was high, about 11%, so expedited discharge may not be appropriate.”
Thrombolytics and submassive PE
Another management decision to make for patients hospitalized with submassive PEs: Should you treat them with your heparin of choice, or with thrombolytics and heparin? While thrombolytics are routinely used in patients with massive PE, Dr. Minichiello noted that the 2012 CHEST guidelines recommend against thrombolysis in normotensive patients. But doctors should consider some studies published since then.
One is the Pulmonary Embolism Thrombolysis trail (PEITHO) in the April 10, 2014, issue of the New England Journal of Medicine. Researchers recruited normotensive patients with submassive PE and RV dysfunction as evidenced by ECHO and troponin. Patients were then randomized to either thrombolysis and heparin or placebo and heparin.
Patients receiving thrombolytics didn’t have better mortality, “but they had a marked reduction in their rate of hemodynamic collapse,” Dr. Minichiello said. That advantage, however, came at a cost. “The risk of major bleeding jumped from 2% in the placebo group to 11% in those with thrombolysis.” The thrombolytic group also had an increased risk of intracranial hemorrhage (2% vs. 0.2%).
PEITHO results were included in a meta-analysis published in the June 18, 2014, JAMA. According to that analysis, thrombolytics decreased patients’ mortality rate from almost 4% to 2% and significantly cut the risk of recurrent PE.
But again, “thrombolytics markedly increased the risk of major bleeding, up to 10% from 3%,” she pointed out. Of note, the meta-analysis did not find any greater risk of major bleeding with thrombolytics in patients under age 65.
“That’s where we are with thrombolysis for submassive PE, and there is still a fair amount of debate,” said Dr. Minichiello. She considers thrombolytics for intermediate-risk patients as long as they don’t have contraindications. But typically, she prefers close observation with a plan to intervene more aggressively at any early sign of instability.
Clinicians should consider thrombolysis in a hemodynamically stable patient with submassive PE on a case-by-case basis. “I would obtain as much data as possible,” she said, “including an ECHO, troponin and perhaps a lower-extremity ultrasound to carefully evaluate for bleeding risk and help determine candidacy.”
Finally, doctors have to consider how long VTE patients should be on anticoagulation. The minimum effective duration is three months. After three months, doctors must determine if the benefits of ongoing anticoagulation outweigh the risks.
Studies have shown, said Dr. Minichiello, that patients with calf vein thrombosis, whether it’s provoked or not, have very low rates of recurrence. In such cases, a defined course of anticoagulation is appropriate.
And patients with DVT or PE related to a major transient risk factor like trauma or orthopedic surgery also have a low risk of recurrence “3% “the first year off therapy. They run about a 10% risk over the next five years.
“That comes out to 2% a year,” Dr. Minichiello pointed out. “Again, indefinite anticoagulation isn’t warranted.” For such patients, guidelines recommend only three months of therapy.
But patients with recurrent VTE or even a single unprovoked VTE “have high recurrence rates when you take them off anticoagulation,” she explained: about 10% the first year and 30% over the next five. “This warrants consideration of indefinite therapy.”
For patients with a single unprovoked episode, CHEST guidelines recommend conducting a careful risk-benefit analysis of ongoing anticoagulation after the initial three months.
“Have patients tolerated anticoagulation or had any bleeding complications? Do they have significant risk factors for future bleeding like end-stage renal disease, advanced age or prior GI bleeding?” Dr. Minichiello pointed out. “If patients can tolerate it and are not at high risk for future bleeding complications, the recommendation is to continue anticoagulation or at least strongly consider it.”
Proximal DVT comes with a much higher risk of recurrence than calf vein, “so if it’s unprovoked, we think about treating indefinitely.” And “PE tends to beget PE,” she said. Patients with initial PE “have a much higher chance of their next event being a PE than a DVT,” with a higher case fatality rate for PE than DVT patients.
That makes her recommendation for indefinite anticoagulation much stronger for PE patients than for DVT. She is more inclined to entertain a trial period off anticoagulation followed by individual risk stratification for a patient with DVT than PE. But therapy must be balanced with bleeding risk.
“For patients over 75, we tend to recommend just a defined course,” said Dr. Minichiello, “or if they have renal insufficiency, have had a GI bleed or if they’ve bled with anticoagulation in the past.”
Phyllis Maguire is Executive Editor of Today’s Hospitalist.
Ordering a thrombophilia work-up?
SAY YOU’RE TREATING A PATIENT with a large and unprovoked PE. Should you order a thrombophilia workup?
“We used to do a thrombophilia workup to guide duration of therapy and see if someone should stay on anticoagulation,” said Tracy Minichiello, MD, director of anticoagulation and thrombosis services at the San Francisco VA Hospital, during a presentation on VTE at a hospital medicine conference last fall at the University of California, San Francisco. “But we now know that the real predictor of recurrence is whether an event is provoked or unprovoked.”
A thrombophilia workup, she explained, “doesn’t add to the predictive value of provoked vs. unprovoked, and it’s not very reliable in acute settings.” Different anticoagulants “including the more recent target specific oral anticoagulants “can affect findings and produce false positives, as can acute thrombosis.
Instead, “we reserve a thrombophilia workup for patients for whom we feel it will influence management,” said Dr. Minichiello. That list includes pregnant patients, those with suspected antiphospholipid antibody syndrome, patients with a pronounced history of multiple thrombosis episodes, and those with recurrent VTE despite anticoagulation.
And when testing is warranted, “we tend to do it after the acute phase and after three months of anticoagulation,” she said. (Ideally, patients being tested are no longer being anticoagulated.) “We rarely recommend a workup in the inpatient setting.”