Note: This is not an official protocol of Genius General Hospital, but rather an idea for discussion.
Patients undergoing catheter-directed thrombolysis for deep vein thrombosis receive ~1 mg/hour alteplase infusions with monitoring of their fibrinogen levels. Usually, their fibrinogen level remains fairly stable and patients do well. However, occasional patients will experience rapid drops fibrinogen, requiring discontinuation of the thrombolysis. Such patients seem to experience minor bleeding complications (e.g. oozing around intravenous catheters, mucosal bleeding).
Our approach to treating PE with systemic lytics is totally different. We choose an alteplase dose, cross our fingers, and hope for the best. We have no way of predicting how the alteplase will affect the patient's coagulation system. Usually the alteplase works, sometimes its does nothing, and occasionally the patients bleed.
Rapid thrombolysis is essential in some situations (e.g. STEMI, massive PE). However, stable patients with submassive PE allow us the time to perform thrombolysis similarly to our approach to DVT: slowly and carefully. Instead of giving alteplase over two hours, we could provide alteplase as an infusion over 24 hours. If the coagulation tests started plummeting or the patient developed minor bleeding, we could stop – before the patient progressed to develop severe hemorrhage.
Basic science behind controlled thrombolysis
Clot breakdown is tightly regulated. A fine balance is needed to avoid either clotting or bleeding.
The fibrinolytic balance varies between patients. For example, patients with high circulating levels of plasminogen activator inhibitor may fail to respond to alteplase (Nicholls 2003). This may explain why patients react differently to identical doses of alteplase:
Serial fibrinogen levels may measure a patient’s fibrinolytic balance
It would be useful to monitor the true biological activity of a patient’s fibrinolytic system during therapy. A common approach is measuring serial fibrinogen levels (1). The amount of systemic fibrinolysis may be indirectly measured based on the speed at which the fibrinogen level falls.
Among patients undergoing catheter-directed thrombolysis for DVT, some studies have found that low fibrinogen levels are a better predictor of bleeding compared to the alteplase dose (Skeik 2013, STILE study). This makes sense, because changes in fibrinogen may reflect the actual biological activity of fibrinolysis (whereas alteplase dose is merely one factor modulating fibrinolysis).
Proposal for controlled thrombolysis
A possible approach to submassive PE might be an infusion of 24 mg alteplase over 24 hours via a peripheral vein. This is based on protocols for catheter-directed thrombolysis of deep vein thrombosis, which have been utilized for over a decade.
Efficacy of low-dose alteplase infusions
Safety aside, there are some reasons to believe that slow infusions of low-dose alteplase could be very effective.
Theory: Dissolving takes time
A popular conception of thrombolysis is that the “clot-busting” drug causes the thrombus to explode on contact. Reality is probably less dramatic. In general, the process of dissolving solid material takes time, being limited by the exposed surface area. Slower infusion of thrombolytic might favor more complete clot dissolution, as progressively deeper layers of the thrombus become exposed to thrombolytic over time.
Theory: Excess alteplase impairs fibrinolysis
Numerous experimental systems show that excess alteplase paradoxically impairs clot breakdown. Possible explanations are:
- Plasminogen steal: Alteplase facilitates the conversion of plasminogen to plasmin, the enzyme that actually dissolves clot. Ideally this occurs selectively on the surface of the clot. However, large doses of alteplase may cause disseminated activation of plasminogen, depleting plasminogen stores (thus “stealing” plasminogen from the clot; Torr 1992).
- Direct inhibition of fibrinolysis: Wu 1995 found that high concentrations of alteplase impaired fibrinolysis, even when plasminogen was supplemented to maintain normal levels. These authors hypothesized that high amounts of alteplase may bind fibrinogen extensively, blocking the ability of plasmin to cleave fibrinogen.
This creates an unusual dose-response curve, with a plateau followed by diminishing efficacy at higher concentrations (Fischer 1997). Some en vitro experiments suggest that alteplase concentrations achieved by full-dose thrombolysis might be unnecessarily high (Wu 1995)(2). This could explain why RCTs comparing full-dose vs. reduced-dose alteplase in PE has found similar efficacy with either regimen (Wang 2010, Goldhaber 1994, Sors 1994).
Evidence regarding the efficacy of gradual, low-dose alteplase for submassive PE was explored previously here (see #3).(3)
Safety of alteplase infusions: Evidence
Overview of alteplase safety in PE
Accepted dogma is that alteplase is extremely dangerous, whereas heparin infusions are safe. This isn't supported by the best evidence: RCTs haven't found much increase in bleeding with the use of alteplase (tables above). Unfortunately, this dogma is maintained by a self-perpetuating loop of circular logic.
Safety of alteplase infusions
The safety of slow infusions of alteplase is supported by substantial evidence. For example, nine trials below describe the use of catheter-directed thrombolysis in PE. These studies incorporated an infusion of ~24 mg alteplase. Although some bleeding complications occurred (often access-site hematomas), there were no lethal or intracranial hemorrhages among 537 patients (4).
Data from catheter-directed thrombolysis for DVT is similar. Below are twelve studies of alteplase infusions for deep vein thrombosis, including no lethal hemorrhages and only a single intracranial hemorrhage among 598 patients (5). The intracranial hemorrhage occurred in 60-year-old woman with diabetes and hypertension who was treated with 126 mg of alteplase, causing her fibrinogen to decrease to 14 mg/dL (this wouldn't occur with modern protocols).
Patel 2015 surveyed a national database to identify 352 patients who underwent catheter-directed thrombolysis for PE between 2010-2012. They identified only a single instance of intracranial hemorrhage in 2010, with none occurring in 2011-2012. Unfortunately the details surrounding this case are unavailable.
Combining all of the above evidence yields an intracranial hemorrhage rate of 2/1487 (0.13%) with no lethal hemorrhages. For comparison, the rate of fatal or intracranial hemorrhage among patients with pulmonary embolism treated with heparin infusion is higher, at 0.26% (table below)(6).
Although an intracranial hemorrhage rate of 0.13% is low, even lower rates may be achievable with added safeguards.
Safeguard #1: Use of a fixed low-dose heparin infusion
Whether to use heparin during thrombolysis is controversial. Thrombolysis itself may suppress enzymatic coagulation, so it is unclear whether heparin is needed. Recent studies of catheter-directed thrombolysis show wide variation in practice (tables above).
A reasonable compromise may be the use of a fixed low-dose heparin infusion without any bolus (e.g. 5 units/kg/hr, up to a maximal rate of 500 units/hr). This is similar to the regimen used in the PERFECT registry, which describes current practice by interventional leaders at Stanford and Cornell (7).
Safeguard #2: Protocoled monitoring of fibrinogen levels
Fibrinogen is a critical component of coagulation, so monitoring it seems rational (8). Fibrinogen monitoring may identify rare patients with unusually rapid fibrinolysis who are at increased risk of bleeding. Choosing 150 mg/dL below which to stop thrombolysis is a conservative threshold designed to maximize safety (9).
Safeguard #3: Slow alteplase infusion to avoid fibrinogen degradation coagulopathy
Fibrinogen degradation coagulopathy occurs when excessive fibrinolysis generates a high level of fibrinogen degradation products which impair coagulation (fibrin degradation products inhibit enzymatic coagulation, reduce fibrin polymerization, and inhibit platelets). Studies of thrombolysis of myocardial infarction and ischemic stroke suggest that this is a significant risk factor for intracranial hemorrhage (Matosevic 2013).
Levels of fibrinogen degradation products fall over several hours following thrombolysis (Stangl 1998). This suggests that a very slow alteplase infusion might promote continuous clearance of fibrinogen degradation products, avoiding dangerous spikes in fibrinogen degradation products.
Controlled thrombolysis & overall treatment strategy
NOTE: The strategy designed below is no longer up-to-date. For a current strategy please see this post. -@PulmCrit
The following schema illustrates situations where controlled thrombolysis might be useful. Controlled peripheral thrombolysis may share a niche with catheter-directed thrombolysis (CDT):
Controlled thrombolysis vs. half-dose thrombolysis
At Genius General Hospital, we have had considerable success using half-dose alteplase, following some initial problems encountered when combining it with heparin (10). How does controlled thrombolysis compare to half-dose thrombolysis?
There is more evidence to support the efficacy of half-dose alteplase (e.g. Sharifi 2013, 2014). Alternatively, there is greater evidence to support the safety of slow, quarter-dose alteplase infusions. Available data on half-dose alteplase shows an intracranial hemorrhage rate of 0/293 patients. This is impressive, but possibly less generalizable than data supporting the safety of 1 mg/hr alteplase infusions.
Half-dose alteplase over two hours might be superior for a high-risk submassive patient who needs prompt stabilization. Alternatively, a slow infusion of 25 mg alteplase might be appropriate for patients with low-risk submassive PE, who are clinically stable and desire a minimal-risk lytic option.
- The ideal approach to dosing alteplase for thrombolysis of PE remains unclear.
- Patients' response to alteplase varies depending on their balance of pro- vs. anti-fibrinolytic modulators.
- A slow peripheral alteplase infusion (e.g. 1 mg/hour infusion) could allow monitoring of the patient's response to thrombolysis throughout the procedure, with discontinuation if there were excessive fibrinolysis or bleeding.
- Available data suggests that quarter-dose alteplase is effective. Based on data from >1,000 patients undergoing catheter-directed thrombolysis of PE or DVT, slow infusions of alteplase are safe.
- Controlled thrombolysis using a slow 25-mg alteplase infusion with protocoled monitoring might offer patients the benefit of lytic therapy with an extremely low risk of severe hemorrhage.
- Prequel to this post: Deconstructing catheter-directed thrombolysis (PulmCrit)
- Choosing your poison: thrombolysis vs. heparin
- Submassive PE: Are we treating it backwards? (PulmCrit)
- PE treatment options and the PEAC Team with Oren Friedman (EMCrit)
- (Classification of submassive PE shown above is based on this post)
- Fibrinolysis in pulmonary embolism (EMCrit)
- Resus: The crashing PE patient
- Other stuff
- D-dimer and fibrinogen degradation products could be a more direct assay of fibrinolysis, but it is less clear how to clinically utilize this data in PE. Some investigators have used TEG to determine the balance of thrombolysis and titrate thrombolytic medications (Ploppa 2010). However, at this point in time fibrinogen is the most commonly used clinical marker of fibrinolysis and fibrinogen reserves. It's not difficult to imagine more accurate approaches being utilized in the future.
- There is disagreement between different en vitro assays of alteplase efficacy, which makes it difficult to know exactly what the ideal alteplase concentration in vivo would be. Additionally, there is probably substantial variation between different patients, so there may be no universal ideal en vivo alteplase concentration.
- Additionally Dr. Ahmet Aykan recently shared with me his manuscript under review describing the use of 25 mg alteplase infusions over six hours in massive PE, with impressive results (both based on serial echocardiography and CT angiography).
- References for this table: ULTIMA, SEATTLE II, PERFECT, Engelhardt 2011, Kennedy 2013, Dumantepe 2014, McCabe 2015, Engelberger 2015, Bagla 2015.
- References for this table: Engelberger 2014, Dumantepe 2013, Enden 2012, Raabe 2010, Baekgaard 2010, Martinez 2008, Laiho 2004, Grunwald 2004, Sugiomoto 2003, Verhaeghe 1997, Bounameaux 1992, Goldhaber 1990.
- Studies listed in this table include all studies including >500 patients which were listed within the Cochrane Review of heparin vs. LMWH for pulmonary embolism (Erkens 2010).
- Monitoring the PTT may be overly cautious, because it would be unusual for a heparin infusion at this dose to elevate the PTT substantially. If the PTT is markedly elevated, this is more likely to be a reflection of aberrancy elsewhere in the coagulation cascade (barring a medication dose error with the heparin infusion). Nonetheless, if unexpected abnormalities in PTT are observed, it may be safest to halt thrombolysis.
- It remains controversial whether fibrinogen needs to be monitored during catheter-directed thrombolysis. A correlation between low fibrinogen levels and bleeding is often, but not always, detected in studies on catheter-directed thrombolysis.
- Selecting a fibrinogen target to halt or reduce the alteplase infusion is a playoff between efficacy and safety. Many protocols for catheter-directed thrombolysis for DVT recommend continuing alteplase at 0.5 mg/hour if the fibrinogen level falls between 100-150 mg/dL, or suspending alteplase until the fibrinogen level increases. However, some studies have correlated fibrinogen levels below 150 mg/dL with increased rates of bleeding.
- Patients treated with half-dose alteplase may hemorrhage when treated with heparin boluses, even if the heparin boluses are given several hours after the alteplase. With strict avoidance of heparin boluses following thrombolysis, we have found half-dose alteplase to be safe and effective in carefully-selected patients. However, our experience does hint that the safety of half-dose alteplase in real-world conditions may not be quite as impressive as suggested in the literature.
Latest posts by Josh Farkas (see all)
- PulmCrit-.050 shades of grey in p-value cutoffs - April 9, 2018
- PulmCrit- APROCCHSS vs. ADRENAL:Are we asking the right question? - March 26, 2018
- PulmCrit Wee- Secondary endpoints:Can we separate the wheat from the chaff? - March 21, 2018