Follow-up data from the PEITHO trial shows that thrombolytics don't affect long-term morbidity. This simplifies management substantially.
PEITHO trial & long-term follow up
The PEITHO trial was a multi-center RCT investigating the effect of thrombolysis in submassive PE. Tenecteplase caused an increase in intracranial hemorrhage and a reduction in hemodynamic collapse. Overall there was a non-significant trend towards reduced all-cause mortality with thrombolysis.
This trial is discussed in further detail here. A major limitation of this study is that patients in the thrombolysis group were treated simultaneously with a loading bolus of heparin plus a full-dose bolus of tenecteplase. This is a very aggressive regimen, which produced bleeding complications substantially greater than other regimens that have been used for submassive pulmonary embolism. An unnecessarily aggressive thrombolytic regimen likely increased hemorrhagic complications, masking benefits from thrombolysis.
Recently, a follow-up study demonstrated that thrombolysis caused no long-term improvement in functional status or pulmonary hypertension (table above). This is strong evidence that thrombolysis doesn't affect long-term morbidity. The trial is explored in further detail by Rory Speigel here.
Revised goal of thrombolysis & how to achieve this safely
Goals of thrombolysis
Until recently, the goals of thrombolysis were essentially two-fold: reduction of morbidity related to chronic pulmonary hypertension and avoidance of PEA arrest. PEITHO simplifies this by eliminating concerns regarding long-term morbidity. This leaves the primary goal of thrombolysis as reducing the risk of PEA arrest (1).
Avoiding PEA arrest is a modest goal, which may be easier to accomplish. We don't need to normalize the pulmonary pressure, nor do we have to clear the pulmonary vasculature of all clot burden. All we need to do is relieve the pulmonary pressure enough so that patients don't drop dead. This is probably achievable with lower doses of thrombolytic than have been used historically.
What is the ideal dose of thrombolytic?
Two RCTs comparing full-dose alteplase (100 mg) versus reduced-dose alteplase both found equal efficacy from either dose (Goldhaber 1994, Wang 2010). Lower doses do cause less major bleeding (figure below). Thus, available evidence indicates that 100 mg alteplase is too much (Sharifi 2016).
The ideal dose of thrombolytic remains murky. The MOPETT trial utilized 0.5 mg/kg up to a maximal dose of 50 mg (giving many patients a dose of 30-40 mg). Aykan et al. used a regimen of 25 mg alteplase over 8 hours in patients with massive PE, causing an average 23 mm reduction in pulmonary artery systolic pressure (7). Studies from catheter-directed thrombolysis suggest that a slow infusion of 25 mg alteplase at 1 mg/hr can reduce the pulmonary artery systolic pressure by about 15 mm (2). The hemodynamic effects of 25 mg alteplase appear grossly similar to the effects of 50-100 mg alteplase (which reduce pulmonary pressures by ~15 mm; figure above). Although large prospective RCTs are lacking, this suggests that a slow infusion of 25 mg alteplase probably improves hemodynamics enough to avoid PEA arrest.
Risks: How dangerous is alteplase, really?
Conventional wisdom is that heparin is safe, whereas alteplase is dangerous. This dogma is propagated by circular logic:
Available evidence suggests that reduced-dose alteplase isn't actually more dangerous than heparin. Two small RCTs comparing heparin vs. heparin plus ~30-50 mg alteplase found exactly equal rates of major bleeding (Zhang 2014). Combining several studies of reduced-dose alteplase (table below), the rate of intracranial hemorrhage was 1/453 (0.2 %). This equals the rate of intracranial hemorrhage from heparin alone (1/500), suggesting that a few patients will develop ICH regardless of how they are anticoagulated (3,4).
Most attention is generally focused on the total dose of thrombolytic. However, the details of how this thrombolytic is administered are likely critical as well. Two factors in particular might improve safety:
1. Avoid combining thrombolytic with unfractionated heparin
Much of the risk which we have traditionally assigned to thrombolysis is probably due to clumsy combinations of alteplase plus unfractionated heparin infusions (heparin infusions being a veritable roller-coaster of anticoagulation, with wide swings between sub-therapeutic and supra-therapeutic anticoagulation). For example, the 2% rate of intracranial hemorrhage in PEITHO reflected a combination of full-dose tenecteplase plus a simultaneous loading bolus of heparin (a formula for disaster). Recent studies involving alteplase have minimized risks by giving alteplase more carefully, for example (5):
- Before and during alteplase infusion, the heparin infusion may be stopped or reduced substantially. After alteplase, heparin may be re-initiated several hours later without a bolus.
- Alteplase may be combined with low molecular-weight heparin (rather than a heparin infusion).
2. Slow infusion of thrombolytic with monitoring of fibrinogen levels
Fibrinogen levels are generally ignored during thrombolysis, blinding us to some very interesting dynamics. The response of fibrinogen levels to thrombolysis varies enormously between patients, reflecting variability in numerous proteins regulating fibrinolysis. If alteplase is given very slowly (e.g. 24 mg infused over 24 hours), this allows monitoring of fibrinogen levels during thrombolysis. If the fibrinogen level falls too rapidly, thrombolysis may be stopped before bleeding occurs. This strategy of controlled thrombolysis might be the safest approach to thrombolysis (extrapolating from extensive data about catheter-directed thrombolysis of PE and DVT).
Consideration of patient-specific risk factors for hemorrhage is also essential (e.g. using a contraindication checklist). The following factors in particular seem to predispose patients to develop intracranial hemorrhage (Chatterjee 2017):
- History of CVA or intracranial pathology
- Known vascular disease (prior myocardial infarction or peripheral artery disease)
Who benefits from thrombolysis? Identifying patients at risk of dying from PE
Physiology of death due to PEA arrest
- Clot-throwing death: This is due to throwing additional emboli to the lungs. These patients look fine initially, with no hemodynamic instability. They may do great for a while, but then they suddenly have a PEA arrest.
- Teetering death: These patients present to the hospital with a very large clot burden in their lungs and severe RV strain. They are teetering on the brink of cardiac arrest. Any hemodynamic stress (e.g. fluid shifts, autonomic fluctuation, vasovagal episode, additional emboli) may push them into a death spiral, with progressive RV dilation and PEA arrest:
These are your stereotypical high-risk submassive PE patients. They aren't hypotensive, but they have clinical signs of hemodynamic instability. The most notable examples of this are:
- Tachycardia, elevated shock index
- Syncope or presyncope (indicating lack of cardiac reserve)
- Patients being kept alive by a huge outpouring of endogenous catecholamines. Clinically, this may be manifested by one or more of the following:
- (a) Lactic acidosis (produced via aerobic metabolism stimulated by endogenous epinephrine)
- (b) Patients who look ill (endogenous epinephrine causes diaphoresis and pallor, makes people look and feel terrible)
Following PE diagnosis, these patients can be identified based on history and physical examination. Bedside echocardiogram should be consistent with right ventricular failure (showing both RV and IVC dilation), otherwise an alternative explanation should be sought for their instability (e.g. small pulmonary embolism plus volume depletion).
Identifying patients at risk of clot-throwing death
Proximal DVT is an independent risk factor for poor outcomes (Jimenez 2014). Meta-analysis confirms that DVT doubles 30-day mortality (Becattini 2016, above). Proximal DVT confers a threat of additional emboli to the lungs, causing hemodynamic collapse. Patients who are at the greatest risk of sudden death due to future embolic events meet the following description (Alviar 2016):
- Already have a submassive PE (e.g. RV dilation and troponin elevation).
- Significant residual clot burden in the legs (e.g. proximal DVT)
The combination of a submassive PE plus a proximal DVT is a formula for disaster. The RV is already strained, so it may be unable to tolerate additional emboli.
Time to stop performing “delayed” thrombolysis
Imagine that a man presents with dyspnea of one week's duration, which has been stable. He is found to have right ventricular strain and an elevated troponin. He is well-appearing and hemodynamically stable. Should he be lysed?
Previously I might have considered thrombolysis with a goal of reducing his long-term pulmonary morbidity. However, based on newer evidence, he shouldn't be lysed:
- He has been stable for several days, so the likelihood that he will suddenly have a PEA arrest is very low (he has already survived an extended trial of “observational therapy” at home).
- Over time, clot may mature and become less susceptible to thrombolytics.
The precise time frame during which thrombolysis may be beneficial remains unclear. Patients with accelerating instability might benefit from thrombolysis, even if they started experiencing symptoms some days earlier. However, patients with stable symptoms over several days probably don't derive much benefit from thrombolysis.
Pre-emptive vs. rescue thrombolysis
After identifying patients at risk of death from PE, two strategies are possible:
- Pre-emptive thrombolysis: Provide thrombolysis immediately and in a controlled fashion, with a goal of preventing decompensation.
- Rescue thrombolysis: Admit patients to the ICU, treat with heparin, and observe. If the patient deteriorates, then treat them with thrombolysis.
The best strategy may depend on the risks vs. benefits of thrombolysis for any individual patient. My general preference is for pre-emptive thrombolysis, for the following reasons:
- Even in the ICU, if the patient has a PEA arrest this is difficult to salvage. These patients often do survive, but may still suffer from anoxic brain injury.
- Observing patients in the ICU for a while doesn't protect them from subsequent deterioration. I've managed some patients who were admitted to the ICU, did great for 24-36 hours, felt better, and later died on the ward.
- Reduce-dose alteplase probably doesn't increase the risk of intracranial hemorrhage compared to a heparin infusion, if given very carefully (Sharifi 2016). For example, an infusion of 24 mg alteplase over 24 hours while monitoring fibrinogen levels is very safe. If a patient is so sick that they require ICU admission, it seems illogical not to treat them with a small, safe dose of alteplase.
Evidence: is there a mortality benefit to pre-emptive thrombolysis in submassive PE?
The largest meta-analysis by Chatterjee 2014 detected a 1.5% mortality benefit from pre-emptive thrombolysis, compared to anticoagulation with rescue thrombolysis as needed (p=0.03, above). However, a subsequent meta-analysis which excluded some studies due to concern about bias found a non-significant 1.4% mortality benefit (Nakamura 2014). The fact that these studies found any mortality reduction is impressive, given the following considerations:
- Most studies included all submassive PE patients (including low-risk submassive patients who are less likely to benefit).
- The meta-analysis found that thrombolysis caused a 1.3% increase in intracranial hemorrhage, which is much higher than rates obtained utilizing reduced-dose alteplase (0.2%). This high hemorrhage rate may reflect that many studies combined full-dose thrombolysis with simultaneous full-dose heparin infusion (e.g. PEITHO, TIPES). High hemorrhage rates would tend to mask any benefit from thrombolysis.
- Many studies included patients who had been symptomatic for several days (e.g. PEITHO included patients with symptoms for up to 15 days). Such patients are very unlikely to obtain any mortality benefit.
We may never have proof of a mortality benefit, since that this would require an improbably large clinical trial. However, extrapolation suggests that a meaningful mortality benefit could be obtained in a subgroup of submassive PE patients by:
- Restricting thrombolysis to patients at the greatest risk of death (high-risk submassive PE).
- Restricting thrombolysis for patients with acute PE, not patients who have had stable symptoms for several days.
- Mitigating the risk of hemorrhage by using a reduced dose of thrombolytic (25-50 mg alteplase) and avoiding the simultaneous use of thrombolytic and therapeutic heparin infusions.
Submassive PE tends to defy algorithms, because it seems that each patient is somehow unique. Nonetheless, having a general schema is useful. Previously, the following approach was proposed:
- Long-term outcomes of patients in the PEITHO trial indicate that thrombolysis doesn't affect chronic pulmonary morbidity. This simplifies the approach to submassive PE, by making the primary goal of thrombolysis to avoid PEA arrest.
- Patients with low-risk submassive PE who are less likely to arrest probably don't benefit from thrombolysis. However, high-risk submassive patients may still benefit from thrombolysis.
- If our goal from thrombolysis is merely to avoid PEA arrest, all we need to do is reduce the pulmonary pressures somewhat (not normalize them). This may be achievable with lower doses of alteplase than have been used historically, with a superior safety profile.
- PEITHO Followup study (EMNerd)
- 2016 series on quarter-dose TPA: Part I & Part II (PulmCrit)
- Submassive PE: are we treating it backwards? (PulmCrit)
- PERT team & risk sub-stratification (EMCrit)
- Sharifi M. Systemic full dose, half dose, and catheter directed thrombolysis for pulmonary embolism. When to use and how to choose? Curr Treat Options Cardio Med 2016; 18: 31.
- Some other arguments do exist in favor of thrombolysis (e.g. more rapid clinical improvement and reduced hospital length-of-stay). However, compared to concerns regarding PEA arrest and intracranial hemorrhage, these concerns are generally less pressing.
- It remains unknown whether this dose must be delivered to the pulmonary arteries, or whether it would be equally effective if administered via a peripheral vein. Given that 100% of the venous return goes directly to the lungs, location of infusion might not matter. This is explored further here.
- Statistic from the CHEST 2016 guidelines (Kearson 2016), Table 8.
- However, the precise rate of intracranial hemorrhage from reduced-dose thrombolysis remains unknown. Since the rate is so low, to precisely define this would require many more patients (e.g. ~5000 patients). Please also note that in real life, the rate of bleeding is often greater than in randomized controlled trials. However, the relative risk of bleeding would be expected to be similar (i.e. comparative danger of heparin vs. alteplase should be the same).
- Note that unlike alteplase, heparin has no proven immediate benefit during the acute phase of PE management. Therefore, interrupting or decreasing heparin infusion to give alteplase is a rational strategy.
- PE typically doesn't cause death due to respiratory failure, but this can occur in patients with chronic respiratory illness. In these cases, patients may manifest with respiratory failure (e.g. hypoxemia, tachypnea, exhaustion). These patients will obviously be ill and require aggressive therapy.
- This study currently is only available in abstract format: Aykan AC et al. Low dose prolonged infusion of tissue type plasminogen activator therapy in massive pulmonary embolism. European Heart J 2014; 35(Suppl 1): 69.