Extended Thromboprophylaxis in Medically Ill Patients - Episode 4
C. Michael Gibson, MS, MD: Deepak, I wanted to return to what you asked about new clot versus existing clot. And we now have some data from some of the trials showing that if you have a clot there on that ultrasound at say 30 days, we used to think, oh, that’s kind of a soft end point. What we found in one of the studies is that poses a risk of death in the time after 30 days. The more clot, the higher the risk of death. The other big predictor of having that clot there was being D-dimer—positive. We do have a marker, even if you don’t get an ultrasound, of the likelihood not just to the presence but to the extent of the clot. And greater clot, greater risk.
Deepak Bhatt, MD, MPH: So the audience is clear, you’re referring to asymptomatic clot.
C. Michael Gibson, MS, MD: Asymptomatic clot. The point is, asymptomatic clot can be a killer.
Gary Raskob, PhD: If I can add to that, because we have very similar information. The guidelines back in 2012 kind of took a departure where the group said the important events are only the symptomatic events. If you think about this condition, the primary reason to give prophylaxis is that fatal pulmonary embolism in many patients presents as sudden death, so there’s no chance to treat the patient. If a patient survives a pulmonary embolus, most of them, the prognosis is excellent with treatment. The thrombi that kill are the ones that are silent and there in the thigh and not known about. These asymptomatic thrombi are not surrogate end points from the point of view of other trial terminology. It’s the disease before it has the last manifestation that we’re trying to prevent, which is fatal embolism. And so proximal thrombosis, even if it’s asymptomatic, is a sinister and serious condition.
Deepak Bhatt, MD, MPH: I think that’s a really important point because a lot of trials have included that as an end point. Some people say, well, if it’s asymptomatic, is it a real end point? And the point that you’re making, in fact, is it can be an end point, but unfortunately it becomes a hard end point. It’s the horse is out of the barn.
Alex C. Spyropoulos, MD, FACP, FCCP, FRCPC: Deepak, I think Gary’s point is crucial in my view because when you look at the epidemiologic data in the acute medical patient as opposed to the surgical patient, they tend to present with much more severe forms of VTE [venous thromboembolism]. They tend to present with more proximal forms of DVT [deep vein thrombosis]. They tend to present with much more severe central PE [pulmonary embolism] compared to surgical patients. They tend to die more often from VTE-related causes. So when you get a patient who’s elderly, who’s comorbid, who has limited cardiopulmonary reserve, it doesn’t take a large clot to put them over the edge in terms of cardiopulmonary collapse. It’s crucial and we have to remind our audience that the first manifestation of a VTE in this population may be sudden death.
Deepak Bhatt, MD, MPH: It’s really a provocative thought when you realize there’s trouble lurking there. I’m going to ask a question; I’m not trying to generate unnecessary health care costs. Would there be any utility in that sort of medically ill patient, they’ve been in a hospital for a week with their pneumonia, or their influenza, or heart failure, something like that, to scan their legs? We talk about machine learning, and AI [artificial intelligence], and Skynet Healthcare Technologies products and that sort of stuff. Someday, machine learning is going to be able to interpret images pretty well. Just get somebody to scan the leg with a hand-held ultrasound, the AI will point out clot or no clot. Is there any merit to doing that other than cost and logistics?
C. Michael Gibson, MS, MD: I do think perhaps machine learning and AI could then direct you to who is the person at greater risk of this subsequent clot. What just astounds me, having grown up on the arterial side, is that people think when you’re discharged from the hospital you’re OK on the venous side. You’re not. Many, if not most of the events will happen after discharge. We just keep focusing on the inpatient care when in fact the biggest threat is in the outpatient setting.
Deepak Bhatt, MD, MPH: Do we have any sense of what those numbers are in terms of what is the risk exactly? You leave the hospital, what is the risk if you’ve been medically ill, of VTE?
Alex C. Spyropoulos, MD, FACP, FCCP, FRCPC: In the generalized patient population of medically ill, the general medically ill patient, it’s about a 1.5% symptomatic VTE event rate at about 90 days or so.
Deepak Bhatt, MD, MPH: So it’s a symptomatic DVT or a pulmonary embolism.
Alex C. Spyropoulos, MD, FACP, FCCP, FRCPC: Correct. So it really underestimates the fatal pulmonary embolic events. Now, when we look at certain high-risk groups however they’re defined, that maybe jumps up to 3% or 3.5%, 4%. Again, there’s an underestimate of what we call the fatal pulmonary embolic events within those numbers.
Deepak Bhatt, MD, MPH: It’s interesting, Mike. Stent thrombosis rates these days are much lower, but we obsess so much about dual antiplatelet therapy, what’s the right duration.
C. Michael Gibson, MS, MD: Right. This is so strange to me that on that side we’re worried about a year, should we extend beyond a year, and here we’re thinking, well, just 5, 10 days, maybe that’s enough. It’s very odd. It’s a big disconnect between the timing of the events.
Gregory Piazza, MD, MS: That’s another major point is that the mortality rates for a lot of these manifestations of venous disease are actually higher now than MI [myocardial infarction]. Like in-hospital and 90-day mortality for PE is higher than STEMI [ST-elevation myocardial infarction], probably because our therapies are so advanced for STEMI but it’s still true. If you talk to house staff, they’re not really worried about PE or DVT. There’s a very laissez-faire attitude about it when actually it’s a very important comorbidity and complication.
Gary Raskob, PhD: To add to that, I think cardiology has done such a good job of reducing the outcomes, and improving the outcome of patients with coronary disease, and so on.
Deepak Bhatt, MD, MPH: Well, thank you.
Gary Raskob, PhD: But there are a lot of people living with finicky hearts and other things, and so we don’t need a huge embolus to push these people over. I think that’s really important. Just the numbers again. Alex mentioned the symptomatic rates, right? If you look at the asymptomatic proximal thrombosis rates that we talked about, those are going to be like 8% to 10% at about a 30-day period. That’s one thing. Secondly, we vastly underestimate the contribution of pulmonary embolism to death in these cardiorespiratory patients because we get autopsies in very few. A patient who dies with 4 or 5 reasons to have shortness of breath and other conditions, pulmonary embolism is way down the thinking. And to wind it up with your question about ultrasound….
Deepak Bhatt, MD, MPH: I was just thinking no one’s actually answered the question.
Gary Raskob, PhD: Right. At the moment, the yield is not high enough to apply that to the broad population. I would say it may be a useful thing to do if prophylaxis was contraindicated, and you couldn’t give prophylaxis to a patient. Then maybe that’s OK. I think if machine learning, as Mike talked about, can help us pick out people with much higher risks and get those yields up, then it may be useful. You may not need a full leg type of ultrasound. What you’re trying to find are big thrombi in the thigh and popliteal.
Deepak Bhatt, MD, MPH: That’s what I meant, just the hand-held device, scan the thigh.
Gary Raskob, PhD: Right, the thrombi that are going to kill somebody, and that’s what you may be able to do. I wouldn’t dismiss the case finding strategy, but because we’ve really improved the safety of prophylaxis in recent years, if we can give primary prevention, that’s a better approach.
Transcript edited for clarity.