Dr. Ted Warkentin describes the pathophysiology and clinical presentation of Heparin-induced thrombocytopenia (HIT) and explains the potential complications, and treatment options.
Back to Symposium Page » It is now my pleasure to introduce Dr Ted Workington, professor in the Department of pathology and molecular medicine at McMaster University, who will speak on hit clinical pearls. Dr. Workington is a good friend of this conference who has participated in a conference many times and is always enthusiastically received. Please welcome Dr Ted Workington. If you have questions you would like him to address, please use the ask a question chat function to ask your questions during the presentation. Questions will be addressed during the panel session. Hello. I'm really happy to be able to participate in this year's scripts conference. Even though it's virtual, and I'm happy to speak on the topic of hit clinical pearls. Here are some disclosures not particularly relevant but more relevant. I will speak about some off label treatments for hit, most notably high dose I V. I. G. Well, here's pearl. Number one hit is a distinct disorder. It's really different than anything else you see in medicine. It features a high thrombosis rate. Despite oftentimes mild to moderate. Thrown beside a pina, there's oftentimes a D I. C. State with important treatment implications. It's caused by unusual platelet activating antibodies recognize a cattle ionic protein pf or bound to Paulie an Ionic heparin. But it's more than just a platelet activation syndrome. It's a pan cellular activation syndrome, involving activation of monos sites, neutrophils, end of helium. And increasingly, it's recognized that complement is important not only for the immunization phenomenon but also for helping to mediate the thrombosis. And it's not like classic drug I tp with very severe thrown beside a pina and bleeding, for example, caused by drugs like vancomycin, but rather despite, as I mentioned Malta, moderate thrown beside opinion, many patients at very high risk of thrombosis, life changing thrombosis, fatal pulmonary embolism, adrenal hemorrhagic infarction, et cetera, et cetera, even ischemic limbs. And did you know that 5% of hit patients developed limb loss? So the pathogenesis is distinct involving platelet activating antibodies that recognize PF or heparin complexes. And it's the I G class of antibodies that, when engaged with platelet FC gamma or I G receptors, leads to clustering of these receptors that you get the strong platelet activation stimulus with formation of Proco violent micro particles. But as mentioned also modest side activation and also in the thallium neutrophils, the Pan Cellular cellular activation response, while Pearl number two the immunology, is unusual in that it's a relatively fast I G response, but transient without immunological memory, and this has all sorts of implications. So here the data showing a fairly early i g response. IgG antibodies in red, and you can see it occurs just a few days after the immunizing heparin exposure. And so this leads to two kinds of syndromes that you see. One is typical onset hit shown here at day six, but typically between five and 10 days after the immunizing heparin exposure. But there's also a form called rapid onset hit, where you restart someone on heparin and you see a very rapid platelet counts fall within 24 hours shown here, and in this case, this is occurring about 30 days after the initial immunizing exposure. Now it turns out that rapid onset had generally occurs only in patients who have recently seen heparin within the past 100 days or so, and the reason for this close association with recent rather than remote heparin exposure is this remarkable property of antibody transients. Hit represents one of the most transient immune responses in all of medicine. So the platelet count on set of typical hit occurring 5 to 10 days after the immunizing heparin exposure. That pattern occurs irrespective of your prior heparin exposure, whether it's your first heparin exposure or your second or your 10th. If you're starting from the point of view of not having detectable antibodies, it takes at least 5 to 10 days to get an episode a hit. As mentioned, the platelet Cal can fall rapidly. This is called rapid onset hit, and this occurs in patients who already have circulating hit antibodies. And there's this very close association with recent heparin exposure. So all of this means that in some circumstances it is safe to be able to give heparin in a patient who has a prior remote history of hit, for example, cardiac surgery or vascular surgery. So here, if the patient no longer has platelet activating antibodies, you're able to give the heparin for the surgery because you know it'll take at least five days for the patient to re stimulate antibodies that could potentially cause hit if the patient even makes them the second time, and often they don't And of course, by then you're off heparin, you're on other drugs, So this interesting temporal features of hit has lots of implications regarding diagnosis and treatment. Well, here's the third pearl, Eliza's or enzyme immuno assays aureus. These are interchangeable terms are highly sensitive for the detection of hit antibodies, including, at the very beginning of the hit Associated Platelet Count Fall. And that's shown here a prospective study I did with Professor K. Elton. We looked at the timing of the various events and hit in a series of patients where we had prospectively collected blood samples. And what we found was that the first day of antibody detection well, current on average, a couple of days before the beginning of the hit associate fall. That was a couple days before the fall, reaching a 50% or greater magnitude and then the thrombosis event, generally a day or two later. So this was the timing of hidden. So the remarkable feature of Hit is that the patient already had antibodies at the very beginning of the hit associate play the Count fall, and this is shown here where you can see that the Eliza reactivity is already quite positive at the very beginning the first day. The platelet count started to go down and was just a little bit higher in optical density reactivity a couple days later, when the patient reached the 50% or greater decline in platelet count. So this is an important feature of Hit that the antibodies are readily detectable early in the in the onset of hits. So what does this mean? Well, it means the Eliza is a very sensitive test, at least 99% sensitivity. We know that the prospective studies show that the antibodies detectable even just before the platelet count begins to fall. And so what it means from a clinical point of view. If you have a negative Eliza and you and you ordered it because of, say, a throne beside a pina, you can be sure that that throne beside a pina is not explainable by hit. So it means you should not automatically repeat a hit test because you say worry that maybe you ordered the test too early. In fact, if you automatically repeat it, you're more likely to detect subclinical serial conversion over the next few days, which, as I'll say later, occurs quite commonly and you should only repeat the test if something changes, there's a new fall or a progressive fall or a robotic event or something changes that makes you think, Well, that wasn't hit that I investigated a few days earlier. But could this now be hit? That's the only time you ought to repeat the test. Alright, The fourth, uh pearl, is that most antibodies that are triggered by heparin exposure are not pathogenic, are not hit causing, and this is shown by the iceberg model. The concept here is that the totality of the iceberg represents the totality of the immune response, but the portion of the iceberg above the water line that represents the subset of patients who actually get clinical hit. And this subset is caused by patients who have generally high, tighter egg class antibodies that are platelet activating in a platelet activation as a such as the serotonin release Sassy. So it means that if you're relying on an Eliza for the diagnosis, you have a high chance of a a false positive diagnosis. If you're relying just on Eliza now, is there a way you can enhance your diagnostic power using an immuno assay and The answer is that there is so the more strongly positive the immuno assay, the more likely it predicts for a platelet activating antibody that is a positive S r A. And this is shown not only for the Eliza but also for these newer, rapid essays such as the Clear, the Kemah Luminescence Essay or the Leah the Latex Enhanced Asset. If you're lucky enough to work at a center where you get one of these rapid essays back within about an hour or so of sample acquisition, but I'll show the data for the more common as, say, the Eliza. So as you can see here as you go to the right, that is higher optical density levels. You see that the proportion that are platelet activating that is blue, which indicates positive S r A increases compared to the yellow, which is the negative. Sorry samples. And you can see it's a very striking relationship, which I've summarized here. So, for example, if you have a negative ELISA less than 0.4 optical density units, your chance of a positive Esrey's less than 1% effectively. You've ruled out it on the other extreme. If you're ELISA's over 3.0. It's essentially hit over 99% probability, and you can see as your Eliza increases over that range, the proportions gradually increasing. So for the most part, if the patient has an Eliza reactivity under too, you know you're not that sure it's hit. And you know you should consider performing in many of those patients a serotonin release asset to try to ascertain the presence of platelet. Activating anybody's alright, Well, here's the fifth Pearl. Be aware of this syndrome that's now called Auto Immune hit. What do I mean by this? Well, first of all, it's an important subset that comprises about 1/5 or 1/6 of hit patients. And these patients have both heparin dependent antibodies but also highly pathogenic heparin, independent platelet activating antibodies. And when we perform the serotonin release ass in Hamilton, we can see the two kinds of hits. So this is the pattern of a completely heparin dependent type of hit serum, and you can see that in pharmacologic heparin concentrations, you see strong platelet activation. But contrast that to this reaction pattern. This is also a hit reaction pattern, but here the patient has strong serum induced platelet activating properties even in the absence of heparin. And so patients who events this pattern, they have particularly severe head a high frequency of D I C. High frequency of bad outcomes, for example, ischemic limb losses. And there are various subtypes of so called autoimmune hit. For example, the syndrome Professor Kelton I described many years ago what we call delayed onset hit that is hit that either begins or worsens after stopping heparin. This is now recognized as being an autoimmune hit syndrome or persisting or refractory hit. These are patients who the heparin stopped in. The play count remains low for several days or even weeks and occasionally even a few months, or patients whose hit has been triggered only by heparin. Flush is this is now recognized as being an autoimmune hit syndrome, or the remarkable syndrome of spontaneous hit syndrome that is a patient who has an illness identical to hit. In fact, it is a kind of hit, clinical picture of hit, serological picture of hit yet no proximate preceding immunizing exposure to heparin to explain the episode of Hit, and this is increasingly recognized and spontaneous hit. There are two variants. There's the post orthopedic surgery type, which is actually almost invariably knee replacement surgery, and the other form, which is after an infection, the medical or the post infectious version. So let's talk about some of these Ottoman hit syndromes while the classic delayed onset hit. There's an example here where the patient in this case got a single peri operative effort. Exposure went home and about a week later came back to hospital with a stroke severe thrown beside a pina, and you can see a falling fibrinogen. So a great example of how you can get over D compensated D. I. C. In some patients with Hit or this patient that we diagnosed a year or two ago. This was a young man, previously perfectly healthy, never been in hospital, never seen happen before. He had a viral illness bad enough over the viral illness, but then a couple days later started getting a headache, and when he came to hospital, was recognized as having a cerebral Venus dural Sinus thrombosis, rare kind of Venus infarct headache. He was found to have a severe thrown beside a pina with a platelet and a deer of 24 he had over t compensated D I. C with low fibrinogen, greatly elevated d dime er and, uh, somewhat elevated Einar. And so I had the idea. What could this be? Spontaneous hit syndrome? It looks like it. And sure enough, when we got an early blood sample, we were able to prove the presence of hit antibodies both by Eliza and by serotonin release, including almost 100% serotonin release in the presence of heparin and also, in the absence of heparin, the the serological hallmark in the S. R. A of spontaneous hit syndrome. And so you ought to consider spontaneous hit syndrome in a patient who has otherwise unexplained thrown beside a pina and thrombosis, particularly if you're seeing the patient a couple of weeks after knee replacement surgery or after a viral illness. All right, Pearl number six. This is a pretty important one. Did you know that the ash hit treatment guidelines are agnostic with respect to which non heparin anticoagulant to give? It's not just the FDA approved anticoagulant Argotron band, but also agents like by Valor Rudin or D N A. Polaroid. If you happen to live in Canada or Europe or find a para Knicks or a direct, uh, sorry. A direct oral, an equivalent such as a pixel ban or river rocks Urban Or perhaps the big A trend. Now I use this everyday in my practice. I I use the agent that I feel is is most appropriate for the situation. And I actually tend not to use direct thrombin inhibitors such as our gotcha band. And one of the reasons is that if you see a very severe hip patient who has hit D. I. C. The problem is that if you use a PTT adjusted therapies such as Argotron Band, you have trouble getting the dose right and you are at high risk of causing erroneous dose adjustments. And this is a phenomenon I call PTT confounding. And it leads to an important paradox that the worst hit cases the ones with overt, be compensated D I. C. Can be essentially untreatable with PTT adjusted therapies such as Argotron Band and by Valerie Rudin. And here's an example of a case we published a number of years ago. This is a patient who had a severe form of really autoimmune kind of hit. You can see the platelet count going down. You can see it continuing to go down even after the heparin has been stopped. So clearly this is not immune kind of hit. This patient was treated with Argotron Band starting at one mic tequila per minute and look at the PTT pre Argha Trabant. It's a little bit elevated at about 40 seconds. Doesn't sound like a big deal, but it is a big deal because when the patient was treated with our gastric band, the PTT rose to very high levels. The physician thought the patient was over any coagulated. So the doses cut and you can see how the patients basically getting hardly any Argotron band, even though the hit is progressing and the patients at their worst hyper questionably right here at that time, they're getting essentially Noorda Trabant in this unfortunate patient, developed brain death micro thrombosis throughout the cerebral, so an unfortunate, fatal outcome of hit. So it's an example, and there's several in the literature of PTT, confounding one of the reasons why I'm nervous about using direct thrombin inhibitors in patients who have this kind of severe hit. So my own practice is to use factor 10 a inhibitors, and there are several, uh, Donna Polaroid. But in terms of drugs available in United States, fond of paramedics or the 10 a wax such as picks abandoned River Rock, Saban and what I like about these drugs is they have long half lives, which means you have stable anti coagulation in the severe, hyper quick ability. State of hit. There's no interference with protein C activation because after all, protein C is activated by thrombin. There's no risk of PTT confounding because you're not making dose adjustments with Pts. If you want to get a drug level, you can measure these with an anti 10 a level, and if you discharge the patient, you can discharge them on a dock. Well, that avoids transitioning to warfarin, and that's really important. Why, because of this complication, warfarin induced Venus limb gangrene. Presumably you've heard about this is important for hematologist to be aware that warfarin is an important cause of micro thrombosis and hit an ischemic limb injury, and it reflects the hit associated hyper quick ability that's not controlled by warfarin. In fact, warfarin leads to protein C depletion, and so the picture is that of a DVT with a cruel or distal extremity, micro thrombosis and ischemic limb losses, even though the patient has X marks, the spot palpable or Doppler identifiable pulses. And so why does warfarin fail? Well, number one, it doesn't block thrombin generation. It's not, after all, an inhibitory type of anticoagulant, but worse, it impairs the protein C system's ability to control hyper kogel ability. So it's worse than just neutral. It actually makes the situation worse. Warfarin does. Thirdly, it raises the PTT, of course. So if you try to rescue the situation with a PTT adjusted agent, such as a direct thrombin inhibitor, it's liable to fail. And then finally, warfarin has a long half life, so the effect the adverse effect will persist for a while. So it's, of course, important to give vitamin K when you recognize such a patient, as well as to give an effective non heparin anticoagulant if this is a hit situation. So how does one control severe, hyper quick ability of hit? Besides giving an anticoagulant that hopefully will work well an emerging treatment which I am increasingly using as high dose IVF big, particularly in patients who I think have severe hit or autoimmune hit. And why is this? Well, we know from in vitro studies that high dose IVF is very good. High dose I V i. G. Is very good at inhibiting the heparin independent platelet activating properties of these autoimmune hits here. Such patients generally have a very rapid ascent in their platelet count after commencing high dose IVF egg. And we believe it's inhibiting the hit associated Hyper quick ability state based on the literature, which suggests very good outcomes. And so the dose is easy to remember. It's just like I tp management a gram per kilogram on two consecutive days, and sometimes patients will need a third either partial or full 1 g per kilo to fully achieve rapid platelet count increase. And as I mentioned, there's favorable literature arguing against this, uh, increasing the hyper quick ability state. Rather, it seems to be rapidly deescalating the hyper quick ability state. And just to show an example of a of a case that I personally treated with high dose i v i g. This patient, presented with a stroke then had ischemic limbs due to large artery thrombosis, had numerous other from bought a complications including a adrenal failure from adrenal vein thrombosis bilaterally and and actually at risk for adrenal failure. And so I treated this patient with intravenous Dana paranoid to allow for immediate anti coagulation for surgical amble ectomy of the limb artery thrombosis. But then immediately, post operative, I started high dose IVF G. I gave it on two consecutive days and you can see this very rapid increase in the platelet count from about uh 61 to 320 within just a little over two days, a remarkable increase in platelet count. And also in this case, look at this. A very abrupt decrease in the serum independence or the heparin independent serum induced platelet activating properties dramatically. Deescalated. Now the heparin dependent platelet activating properties remained, but so what? The patient was now off heparin, but the heparin independent reactivity the one that can make the patient continue to activate platelets for several days. That very dangerous, effective hit that was abruptly interrupted by the high dose IVF egg. So to summarize the seven clinical pearls before we get the very last part of my lecture, number one hit is a distinct disorder. Number two, it has funny timing. Features The classic day 5 to 10 picture of typical onset hit the picture of rapid onset hit related to recent Hepburn exposure. The fact you can re exposed patients under special circumstances. Because of this timing, these timing features the high sensitivity of Eliza's, including, at the very onset of the platelet count fall the iceberg model, which implies that you you ought to get, if you can, the serotonin release asset to be able to confirm a diagnosis. But to realize that Eliza degree of reactivity is quite predictive for the likelihood of having a positive S r A. To be aware of the various autoimmune hit syndromes such as delayed onset hit, persisting hit heparin, flush it and remarkably spontaneous hit syndrome. The post orthopedic uh knee replacement variant, as well as the day, post viral or post infectious variant. The important treatment consideration that you can use one of several non heparin anticoagulants, and I mentioned my own personal favorite with today and inhibitory drugs and then this, uh, the rapid realization or the recent realization by by workers in the field that high dose I V. I. G. Is a good way to interrupt hit antibody effect on platelet activation, particularly the autoimmune variant. Now I think I have a few minutes to cover the topic of microvascular thrombosis in critically ill patients. And here I want to make the point that hit PathoGenesis has informed this to some extent, so to get back to warfarin induced micro thrombosis, remember, the concept is a profound disturbance and pro coagulant anticoagulant balance. And so what about the intensive care unit patients who develop microvascular losses? They have d I see. They have shock. They have April symmetrical limb loss. We call this SPG symmetrical peripheral gangrene. Well, why is it that only some patients develop this? Well, we think the role of shock liver has has until recently not really been appreciated. And so a clinical pearl is that shock liver is a kind of warfarin equivalent that explains the risk of SPG and critically ill patients. So, just like warfarin induced, micro thrombosis occurs 2 to 5 days after the initiation of warfare in the median of three days. Well, that's the same with, uh, the critically ill patients who develop SPG. It typically occurs beginning 2 to 5 days with a median of three days after the onset of shock liver and the first case that suggested this was this one published about eight years ago. When this critically ill patient, you can see the severe thrown beside a penny of D. I. C. Had a cut ischemic hepatitis or shock liver. Three days later, hematology got the consult for ischemic limb injury. Notice the three day gap there. And so we measured pro coagulant markers, tat complexes, fiber monomer, fiber indeed, Eimer Sky High and we measured to natural anticoagulants, protein, C n anti thrombin, and both were severely depleted. So you can see this marked imbalance in the pro coagulant, an equivalent markers. And so to test the idea well, if the liver is not working, the factors should go down in relation to their half lives. We plotted out the coagulation factor activity levels of several factors in relation to their half lives. As as you can see, there's a very nice correlation that supports this hypothesis of of acute liver injury and not making these coagulation factors, helping to explain why a few days left later, the levels would be profoundly reduced, particularly protein C, but also to some extent anti thrombin. So this is the concept. Natural inequality levels are severely reduced in patients with SPG and purple phone ins. And here's a paper for literature that supports that this study in in in about 20 patients who developed what they call purple phone ins that can be used anonymously with symmetrical peripheral gangrene, they measured three natural anticoagulants. You can see how profoundly reduced these were in the in the patients who had purple foam, AnAnd's versus the controls who were similarly ill or who had d i c but who did not have ischemic limb losses. If you're interested in reading about this, there's a nice review article that talks about these various Hemet to logic syndromes that underlie Venus limb gangrene as well as symmetrical peripheral gangrene and just as quickly summarize SPG or purple foam. AnAnd's. These are associated with septic shock and cardiogenic shock. These patients, invariably if they get the syndrome, have tissue losses in their lower extremities and about a third of the time in their upper extremities and the important role of the preceding shock liver and a schematic. The patients have this crashing platelet count shown in red they have the rising iron are shown in blue, and they have this three day on average gap between the onset of the shock liver and three days later, the onset of the symmetric peripheral gangrene implicating protein C depletion as well as impairment of the anti thrombin anticoagulant pathway. So this is the concept profound disturbance and pro coagulant, anti coagulant balance. The combination of D. I. C and Shock liver and, I guess, ongoing shock that is the try A that leads to ischemic limb injury. And it's mediated by ongoing D. I C with profoundly reduced protein C anti thrombin, and we believe in some patients protein s as well. So the clinical Pearl is that when you think about SPG and purple, Foreman ends in critical illness. Think of the Triad. Ongoing shock, ongoing dicey liver dysfunction, uh, manifest as a shock liver state. That is the combination that can lead to ischemic limb injury in the setting of profound natural anticoagulant depletion. And I think it's a myth that this syndrome is caused by vessel presses. A lot of clinicians believe that vase oppressors do this, and the literature actually suggests that it's not vase oppressors on a causal basis. They're often being given. Not always there, often being given because of the shock state. But the real pathogenesis is the combination of ongoing shock D I C and shock liver and associated natural anticoagulant depletion. So I hope this wasn't too many, uh, pearls. But I enjoyed giving this talk, and I look forward to participating in the panel discussion during the meeting. Thank you very much.