Last updated on
Clot Chronicles: Predicting and Preventing Cancer-Associated Thrombosis
Hi. My name is Alok Khorana. I’m a medical oncologist and Director of the GI Cancer Program* at the Cleveland Clinic in Cleveland, Ohio. On this episode of Clot Chronicles, we will be discussing our new review of the prediction and prevention of cancer-associated thromboembolism (CAT).
Many of you who take care of patients with cancer know that cancer patients are at high risk for venous thromboembolism (VTE), and that typically includes deep vein thrombosis (DVT) and pulmonary embolism (PE), as well as arterial thromboembolism, which is not as well understood but includes a greater risk of stroke and myocardial infarction, at the very least.
The rates of these illnesses—venous events and arterial events—vary across cancer patients; they vary even in the same cancer patients over a period of time. And what we’ve learned from the last 10 or 20 years of researching this area is that there is a high degree of risk variation within cancer patients, but that we can successfully predict the risk VTE in this population using some risk assessment tools.
In addition, we can focus on the high-risk population and consider primary prevention, and there has been a great deal of discussion about optimal primary prevention strategies in this field. I’m going to try today to cover both of those aspects, risk assessment and primary prevention, in the specific setting of CAT.
Let’s start with risk assessment first. There are certainly rudimentary forms of risk assessment. As I mentioned, I direct the GI Cancer Program. Anybody who takes care of patients with pancreatic cancer knows, for instance, that there is a clearly high risk of VTE, much higher than most other solid tumors. Similarly, people who take care of patients with primary brain tumors also know that there is a very high risk of DVT or PE in that population. So that’s one form of risk assessment where you recognize certain types of cancers are strongly associated with the risk of VTE.
A more robust form of risk assessment involves a risk assessment tool that incorporates the type of cancer but also four other variables. This is a tool that my research group and I initially published in Blood in 2008 using a national database of about 3,000 cancer patients. This risk tool, now known as the Khorana score, has been validated in over 45,000 patients worldwide across multiple different settings and studies.
It depends on five simple variables. One, as I mentioned, is the type of cancer. So if you have pancreas cancer, stomach cancer, you get 2 points; if you have lung cancer or lymphoma, you get 1 point; kidney cancer, you get 1 point; most other types of solid tumors—breast cancer, colorectal cancer—you get 0 points for those types of cancers. You can also get points for a high white cell count or a leukocyte count of more than 11,000; a high platelet count of more than 350,000; a high BMI of more than 35 kg/m2.
So, there are different ways to get to a high-risk category, such as just the type of cancer or a combination of a high leukocyte count. Low hemoglobin is another variable in there, so hemoglobin of less than 10. So, you can have a low hemoglobin, a high white cell count, and a high platelet count and that gets you 3 points, which puts you at a high risk.
Over the years, the definition of “high risk” has evolved, as well. In our original paper, we had proposed a risk score of 3 or higher as being high risk. But recent studies have shown that the score of 2 or higher is associated with roughly 10% risk of VTE at approximately 6 months, so that’s been the definition more recently at which prophylaxis is being considered.
So, as I mentioned, there’s variation in risk. Risk can be identified using rudimentary techniques such as the type of cancer; it can be better quantified using a validated risk assessment tool. And the risk definition used to be 3 or higher, but is now evolving towards 2 or higher.
Now, how do we use risk assessment to identify patients at high risk that would benefit from primary prevention? Again, the last one to two decades have seen a lot of research in this area. It’s important to note that VTE in the cancer population can occur in both the inpatient setting and in the outpatient setting.
We’ve known about the inpatient risk for a long period of time because we do inpatient prophylaxis for most cancer patients that are admitted to the hospital and certainly for almost all cancer patients that undergo major abdominal or pelvic surgery. And this is based on data that goes back several years looking at inpatient prophylaxis and postsurgical prophylaxis.
But the reality is that most cancer patients now receive their treatment in the outpatient setting. And so most clots, 70-80% of clots, now occur in the outpatient setting, not in the inpatient setting. So, we could be perfect with inpatient and postsurgical prophylaxis and still not prevent the majority of clots that occur in cancer patients. And recent trials, therefore, have focused on outpatient prophylaxis.
The initial set of these trials were not risk-adapted and focused on using low-molecular-weight heparins (LMWHs). And they were successful in terms of meeting their primary end points, and these were trials using LMWHs such as nadroparin. PROTECHT was one trial. SAVE-ONCO was another trial using an ultra-LMWH called semuloparin.
And both these studies took a similar approach. They took a bunch of solid-tumor patients; half of them got outpatient prophylaxis with a daily injectable. The other half didn’t. Rates were low. They were about 3.5, 4% down to about 1.5, 2%. And so statistically significant, met the primary endpoint, but because the rates were so low—because patients were not risk-adapted—those recommendations of outpatient prophylaxis were not widely made nor utilized.
What’s changed in the past couple of years has been this focus on risk adaption that I just mentioned, as well as the advent of direct oral anticoagulants or DOACs, which can be used orally and don’t require a lot of monitoring.
Two large, randomized trials focused on the high-risk population. These two trials were AVERT, using apixaban, and CASSINI, using rivaroxaban. Both these trials focused on a population of patients with a Khorana score of 2 or higher—mostly solid-tumor patients and lymphoma patients, the work also included some myeloma and brain tumor patients—and randomized to either apixaban or rivaroxaban versus placebo. And consistently, these trials found a reduction in the risk of VTE, a slight increase in absolute risk of major bleeding. The number needed to treat, depending on the trial and depending on the meta-analyses, ranges between 15-25. The number needed to harm, to have 1 major bleed, is about 100 for both these trials.
Based on the results of these trials, guidelines from American Society of Clinical Oncology (ASCO), National Comprehensive Cancer Network (NCCN), International Initiative on Thrombosis and Cancer (ITAC), and the International Society on Thrombosis and Haemostasis (ISTH) have all been updated, and all guidelines now recommend consideration of patient prophylaxis for cancer patients with a score of 2 or higher who don’t have a major bleeding risk.
The most recent data and research projects have now focused on how we implement these trial and guideline recommendations into clinical practice. Because as many of you who take care of cancer patients know, cancer patients have so many ongoing issues, so many different competing priorities—the cancer, obviously, the treatment decisions, the side effects of treatment, etc.
So, who is going to take ownership of thromboprophylaxis? There are different models for managing this, which we discuss in our review. One model is to introduce an electronic medical record alert. That’s what we’re doing at the Cleveland Clinic and at certain other institutions, where the score is automatically calculated, and it’s presented to the clinician, and the clinician’s alerted before they go in to see the patient, and then the clinician can have a discussion with the patient about whether prophylaxis makes sense in that specific setting or not.
A second model is the Vermont model that actually creates a multidisciplinary pharmacist and hematology physician-based service. High-risk patients are identified and then they are referred to a hematologist, and then the hematologist has the discussion with them about whether they want to go on prophylaxis or not. This model was recently published in the Journal of Oncology Practice (JOP), and they found close to 95% compliance with prophylaxis for patients that were referred to the hematology service. So different models are currently being used to study, essentially, implementation science, how we can implement the findings of clinical trials and guidelines.
I talked mostly about VTE, but briefly want to discuss arterial events in one of the two DOAC trials, CASSINI. Arterial events were recorded and were rare, but were also reduced in numerical terms with thromboprophylaxis. And so there may be benefits beyond VTE prevention into arterial thromboembolism prevention in this population with outpatient thromboprophylaxis as well.
So, to summarize, there is lots of exciting new data, including validated risk assessment tools to identify high-risk patients, and trials with both LMWH and DOACs in outpatient thromboprophylaxis. There is definitely well-known data for inpatient and postsurgical prophylaxis—those recommendations don’t change—and then there are new implementation science approaches to finding the right model in terms of delivering and ensuring compliance with outpatient thromboprophylaxis.
Thank you very much for taking the time to listen to me today. References and citations for all that I’ve discussed today can be found in our review that was just published in The Oncologist. Thank you again for listening.
*Your patients can find more information about GI cancer here.