A while ago I was called by a physician at an outside hospital regarding a patient with a saddle PE. The patient was a 70-year-old man with no significant medical history. His CT showed a truly impressive clot burden, RV dilation, contrast reflux into the liver, and dilation of the IVC and hepatic veins. He was tachycardic, borderline hypotensive, and (per report) looked quite ill. The transferring physician was very concerned that he might deteriorate and felt that he needed systemic thrombolysis. I agreed. We decided to stop his heparin, give him 50 mg alteplase (10 mg bolus followed by 40 mg over two hours), and transport him to Genius General Hospital.
Upon arrival some hours later he looked fantastic, with normal vital signs. He reported feeling dramatically better. He had finished the 50 mg alteplase during transport and was not on any medication at the moment. His admission coagulation studies:
Most sources recommend re-starting heparin when the PTT is below twice the upper limit of normal, which it was in this case (barely). However, I was nervous to start heparin with his PTT almost twice normal with such a low fibrinogen. We held anticoagulation and repeated the labs in two hours:
Submassive PE patient gets 50mg TPA, improves dramatically. Post-TPA fibrinogen level is 80 ml/dL. Patient is not on any anticoagulation now. PTT normal, INR 1.3. What is your next move?
— josh farkas (@PulmCrit) June 6, 2018
Answers vary, reflecting that there probably isn’t any definite answer. The rest of this post will explore this further, but for now my approach was as follows:
- Just give heparin? A substantial drop in fibrinogen after being bolused with TPA puts this patient at increased risk of bleeding, so some caution is warranted with heparin. You could probably get away with simply starting heparin, but this is a bit dicey in a patient whose fibrinogen is <100 mg/dL and falling.
- Repeat coags again? Fibrinogen levels will take awhile to recover, often about a day. Furthermore, this patient’s fibrinogen is continuing to decrease (possibly due to persistent dysregulation of fibrinolysis as explored below). Thus, it’s unlikely that waiting a few more hours will resolve his hypofibrinogenemia.
- Cryoprecipitate plus heparin? We opted to give him a unit of cryoprecipitate and then initiate therapeutic enoxaparin immediately thereafter. Our goal was to render him therapeutically anticoagulated with maximal safety. He did extremely well with an unremarkable hospital course. Of course, that doesn’t prove that we did the right thing.
This case and some other similar cases have made me wonder about the value of following fibrinogen levels after TPA boluses for PE. This blog will try to evaluate the evidence behind this.
Getting started: the physiology of TPA
Coagulation effects of TPA
TPA has a half-life of only a few minutes, but it has diverse and long-lasting effects on the coagulation cascade. Most notably:
- TPA causes degradation of fibrinogen, which decreases the amount of fibrinogen available to form clots. Besides fibrinogen, other proteins may be degraded, particularly factor V and factor VIII (Kluft 2017).
- Fibrinogen Degradation Coagulopathy: TPA breaks fibrinogen down into smaller, nonfunctional bits. These fibrinogen degradation products compete with normal fibrinogen during clot formation, thereby inhibiting coagulation. They may also stimulate plasminogen to degrade fibrin, thereby mimicking the action of TPA itself. Thus, the fibrinogen degradation products themselves exert anticoagulant and pro-fibrinolytic effects.
- Normally there is a balance of pro- and anti-fibrinolytic proteins. TPA leads to consumption of these proteins (both pro-fibrinolytic proteins such as plasminogen and anti-fibrinolytic proteins such as alpha-2 antiplasmin). This may cause dysregulation of fibrinolysis which persists for 1-2 days (Stangl 1998). It’s possible that persistent fibrinolysis dysregulation explained the ongoing drop in fibrinogen levels observed in the case above, which occurred well after TPA had been discontinued.
Fibrinogen levels may remain low for about a day, whereas the fibrinogen degradation coagulopathy seems to resolve within several hours (Stangl 1998)(1). Synergy of these multiple coagulation abnormalities may cause the patient to remain substantially coagulopathic for some time after the TPA infusion is completed.
Giving TPA isn’t like giving ceftriaxone: the same dose of TPA may have vastly different effects in different patients. Fibrinolysis is enormously complex. Different patients have different fibrinolytic tendencies, making them more or less sensitive to TPA.
Individual variability quickly becomes evident to anyone managing patients undergoing catheter-directed thrombolysis. Such patients are treated with infusions of TPA at 1-2 mg/hour, with serial monitoring of fibrinogen level. In some patients, the fibrinogen level doesn’t budge. In other patients, the fibrinogen level falls rapidly – which often precedes minor bleeding.
Published literature supports this heterogeneity. For example, the figure above shows the change in fibrinogen levels following TPA administration for myocardial infarction. Many other studies describe wide variation in coagulation parameters arising after TPA (e.g. Lee 2014).
Kline 2016 provided a beautiful illustration of individual variability to thrombolysis among PE patients. Thromboelastography was used to generate clots in vitro and test how susceptible they were to TPA. Patients with greater in vitro susceptibility to thrombolysis were more likely to bleed when treated with tenecteplase. Alternatively, patients with reduced in vitro susceptibility to thrombolysis were more likely to have persistent pulmonary symptoms following lytic therapy. This implies that rather than a one-size-fits-all approach, sophisticated in vitro testing could be used to determine a personalized dose of thrombolytic.
Do fibrinogen levels predict bleeding?
Fibrinogen & catheter-directed thrombolysis
This is quite controversial, so let’s take a look at the primary literature involving TPA infusions:
STILE 1994: This was a prospective RCT comparing surgery to thrombolysis for peripheral artery disease of the leg. One weakness of this study is that among patients receiving thrombolysis (n=249), half were treated with urokinase and half with TPA. Lower fibrinogen level correlated with hemorrhage risk:
Arepally 2002: This was a retrospective study describing 35 patients who underwent catheter-directed thrombolysis for peripheral arterial disease using various doses of TPA. No relationship was found between hemorrhage and, well, anything at all:
Skeik 2013: This was a retrospective description of 69 catheter-directed thrombolytic procedures for acute lower extremity thrombosis (arterial or venous). Rates of major bleeding were considerably higher among patients who developed hypofibrinogenemia:
Lee 2015: This was a retrospective description of 42 patients who underwent extremity arterial or venous thrombolysis with TPA. All patients who developed major bleeding also experienced low fibrinogen levels (defined as <150 mg/dL; red box below). However, due to the small sample size this didn’t reach statistical significance. The study concluded that “A fibrinogen level <1.5 g/dL during thrombolysis was not associated with an increased bleeding risk” – although technically correct, this is meaningless because the study is underpowered.
Summary of catheter-directed thrombolysis data: Bleeding is a rare adverse event, so a large study is needed to investigate it with adequate statistical power. Larger studies detected a correlation between hypofibrinogenemia and bleeding. Predictably, smaller studies did not. Overall, the best evidence with TPA shows a correlation between hypofibrinogenmia and bleeding (2). This is consistent with the practice of most centers, which is to trend fibrinogen levels, reduce TPA infusion rate if fibrinogen falls below 150 mg/dL, and stop thrombolysis entirely if fibrinogen drops below 100 mg/dL (Kaufman 2018).
Fibrinogen & hemorrhage in thrombolysis for MI
Classic studies of thrombolysis in MI found a relationship between bleeding, low fibrinogen levels, and elevated levels of fibrinogen degradation products:
- TIMI-1 evaluated the effect of infusing 80 mg TPA over three hours among 124 patients (3). The hemorrhage rate was 60% among patients whose fibrinogen level fell to below 150 mg/dL, compared to 22% among patients whose fibrinogen remained above 150 mg/dL.
- TAMI evaluated the effect of infusing 150 mg of TPA over 6-8 hours among 386 patients. The nadir fibrinogen correlated with bleeding, as shown below. Two patients experienced intracranial hemorrhage – in both cases there was exuberant fibrinolysis with large drops in fibrinogen and commensurate elevations in fibrinogen degradation products:
The most useful data about fibrinogen and hemorrhage arguably comes from stroke patients:
- Let’s be honest: the thing we’re really worried about is intracranial hemorrhage (ICH). Most other forms of bleeding are treatable without significant morbidity.
- Different types of coagulopathy tend to manifest with bleeding in different places (e.g. hemophilia causes hemarthrosis, thrombocytopenia tends to cause mucosal bleeding). If we’re worried about ICH, the best way to study this is to evaluate ICH rates directly. The only source of data that allows for this is stroke patients, because they have a high ICH rate (thus providing enough data to study it accurately)(4).
Trouillas 2004: This study described 157 stroke patients treated with 0.8 mg/kg TPA infused over 90 minutes. They found a strong correlation between fibrinogen degradation products (FDP) and intracranial hemorrhage. The relationship between ICH and absolute hypofibrinogenemia and ICH was weaker (only 11 patients developed ICH, so power was limited).
Matosevic 2013: This study involved 547 stroke patients treated with TPA. These authors found that the decrease in fibrinogen was the strongest predictor of ICH, which implicates fibrinogen degradation coagulopathy (similar to Trouillas above). As shown below, a reduction of fibrinogen levels by more than 200 mg/dL was strongly predictive of ICH. Interestingly, a higher baseline fibrinogen level correlated with increased bleeding risk, presumably because this led to the generation of more fibrinogen degradation products.
Sun 2015: This study involved 72 stroke patients treated with TPA. These authors found that absolute fibrinogen levels two hours after thrombolysis was most strongly predictive of ICH (more than the change in fibrinogen during thrombolysis). The strongest single predictor of ICH was a post-TPA fibrinogen level <200 mg/dL:
Vandelli 2015: This study involved 104 stroke patients treated with TPA. Logistic modeling found that the best predictor for ICH was a >25% decrease in fibrinogen levels during thrombolysis:
Summary of stroke data: Absolute hypofibrinogenemia and drops in fibrinogen both appear to be powerful predictors of ICH. Sorting out the precise contribution from each factor is difficult, because they are closely inter-related variables.
Implications for fibrinogen monitoring in PE patients treated with half- or full-dose thrombolysis.
Evidence reviewed above indicates that fibrinogen consumption correlates with hemorrhage. This doesn’t prove that hypofibrinogenemia itself is causing bleeding (in fact, it seems more likely that fibrinogen degradation products are the major culprit). Therefore, this data doesn’t prove that giving cryoprecipitate to increase the fibrinogen level will reduce bleeding risk.
No high-quality evidence exists on fibrinogen monitoring in PE (5). Most practitioners don’t check fibrinogen levels for patients underlying thrombolysis with TPA for PE. This is a bit paradoxical, because fibrinogen is usually monitored in patients receiving catheter-directed thrombolysis – a procedure involving lower doses of TPA with a markedly lower risk of intracranial hemorrhage.
Checking fibrinogen before and after the completion of a two-hour infusion of 50 or 100 mg TPA might offer the following benefits:
- Risk-stratify which patients are at the greatest risk of bleeding.
- Knowing the fibrinogen level could be useful if the patient subsequently bleeds, to guide prompt and rational use of blood products. As shown in the figure above by Trouillas et al., many patients may have normal or even elevated levels of fibrinogen despite receiving TPA – so blindly giving cryoprecipitate to every post-TPA hemorrhage is illogical.
- For patients who have received 50 mg TPA and remain unstable, understanding their fibrinogen dynamics could help predict the risk of giving additional TPA.
- If marked hypofibrinogenemia occurs, correction with cryoprecipitate may be considered. However, this should be undertaken cautiously (with the understanding that we don’t know what an ideal target fibrinogen level is here)(6).
- Initiation of heparin anticoagulation might be delayed for a few hours in patients with unusually severe coagulopathy. Some authors have suggested delaying heparin re-initiation if fibrinogen levels are <100 mg/dL (Ceresetto 2017). However, given that fibrinogen may take a day to increase, an alternative approach is to give cryoglobulin followed by re-initiation of heparin anticoagulation.
- Although TPA itself has a short half-life, it effects on coagulation last several hours.
- Different patients may have markedly different responses to the same dose of TPA, depending on their baseline fibrinolytic tendencies.
- Monitoring fibrinogen may help understand the effects of TPA on an individual patient’s coagulation physiology.
- Extrapolating from data other situations (especially ischemic stroke), the risk of intracranial hemorrhage may be greatest among patients who drop their fibrinogen substantially (>200 mg/dL decrease) and/or develop absolute hypofibrinogenemia.
- There is no high-quality evidence regarding whether to monitor fibrinogen levels in patients with PE treated with full- or half-dose thrombolysis. It may be reasonable to follow this pre- and post-TPA (noting that we will be checking other coagulation tests at those timepoints anyway).
Acknowledgement: Thanks to Dr. Gilman Allen for thoughtful comments on this post.
- Fibrinogen degradation coagulopathy is likely one cause of elevated PTT which is commonly observed following fibrinolysis. This will generally resolve within a few hours. The traditional practice of waiting for PTT to resolve after fibrinolysis before starting heparin may allow clinicians to avoid anticoagulating a patient with active fibrinogen degradation coagulopathy.
- This relationship may not hold up when using urokinase, which affects the hemostatic system in a different manner (greater extent of systemic fibrinolysis). The relationship between fibrinolysis, fibrinogen levels, and bleeding seems to be different depending on which lytic agent is used.
- TIMI-1 also evaluated streptokinase, but I’m ignoring that arm of the trial for our purposes here.
- Of course, nothing is perfect. Some patients with stroke will develop ICH simply as the natural evolution of stroke (hemorrhagic conversion). And ICH due to acute stroke might be different from an ICH in a PE patient due to an old infarction. But overall, as a model for the relationship between TPA boluses, fibrinogen and ICH in humans, this is probably the best data available.
- Many earlier studies on fibrinolytics in PE did evaluate the magnitude of changes in fibrinogen and other coagulation parameters.The precise significance of these studies is difficult to interpret in a modern context, since it seems that different formulations of TPA have been utilized over the years which had different effects on coagulation (Mueller 1987). These studies were often smaller than their cardiology counterparts, with limited ability to define the relationship between coagulation parameters and bleeding. The ideal study would have a prospective design, randomize patients into two groups regarding whether or not fibrinogen is monitored, and then measure a clinically-relevant outcome such as hemorrhage.
- Physiologically, fibrinogen is required for clotting. We probably need someminimal amount of fibrinogen to avoid spontaneous hemorrhage. The exact minimal amount is unclear, however, especially within the context of a post-TPA patient with numerous other hematologic derangements.