Occasionally in science we encounter a truly bizarre result. Our natural inclination is to ignore it. It’s uncomfortable. It creates cognitive dissonance, challenging our understanding of the world. However, struggling to understand the bizarre result can reset our perspective. Often it is the bizarre, unexpected result that changes everything.
Four bizarre studies: OPTALYSE PE, Wang 2010, Sors 1994, Goldhaber 1994
OPTALYSE PE trial: Four-arm dose-finding study for catheter-directed thrombolysis
This is a fresh multi-center RCT which randomized one hundred patients with submassive PE to receive one of the four treatment regimens via an EKOS vibrational pulmonary catheter:1
- 4 mg tPA per lung over 2 hours
- 4 mg tPA per lung over 4 hours
- 6 mg tPA per lung over 6 hours
- 12 mg tPA per lung over 6 hours
In nearly all cases, patients received bilateral catheters to treat bilateral clots (for a total dose of tPA ranging from 8 mg to 24 mg). The primary endpoint was RV/LV ratio based on CT scan after 48 hours. There was no difference between any of the groups:
We are not certain why the RV-to-LV diameter ratio improved similarly across all arms
More bleeding complications were seen in patients receiving 24 mg alteplase over six hours, causing that arm to be stopped prematurely.
Wang 2010: 50 mg vs 100 mg tPA
OPTALYSE is reminiscent of Wang 2010, a prospective RCT comparing 50 mg vs. 100 mg tPA infused peripherally over two hours among 118 patients with submassive or massive PE.2 That study found a similarly inexplicable result: no difference in hemodynamics between the two doses:
There was an increased risk of hemorrhage with 100 mg tPA, particularly among patients weighing below 65 kg. These results mirror OPTALYSE: higher doses increase bleeding, without increasing efficacy.
Sors 1994 & Goldhaber 1994
Both of these studies compared a regimen of 100 mg tPA over two hours peripherally versus reduced-dose tPA (0.6 mg/kg up to a max dose of 50 mg) over 15 minutes.3 4 These studies are a bit murky because they are simultaneously comparing different doses and different infusion times. Neither study detected any real difference in efficacy.
Heparin dosing vs. tPA dosing
Let’s take a step back and consider how we dose tPA in comparison to unfractionated heparin infusion:
Our approach to tPA dosing is fairly simplistic. We give the same dose to all comers, regardless of weight or physiologic effect. Let’s imagine what would happen if we dosed heparin infusions in the same way (e.g. just throw everyone on an infusion of 1200 units/hour without any lab monitoring). We might see the following results:
- Some patients would wind up with subtherapeutic heparin levels, causing lack of efficacy.
- Other patients would have supratherapeutic heparin levels, leading to hemorrhage.
- Overall, the heparin would seem to be a terrible drug.
Let’s take this a step further and imagine an experiment where we compare four fixed-dose heparin infusion regimens. My guess is that results would show something along the following lines:
- Risk of bleeding would be greatest among patients getting the highest heparin dose.
- There wouldn’t be measurable differences in efficacy, because a lot of patients in all the groups would wind up receiving a reasonably therapeutic dose of heparin. Demonstrating an efficacy difference is harder than demonstrating a toxicity difference, because toxicity (hemorrhage) is more obvious than efficacy (hemodynamic improvement).
In short, a study with blind-fixed dosing of heparin infusions might yield results resembling what we’re seeing from OPTALYSE.
Fixed-dose tPA is a flawed strategy
OPTALYSE illuminates a fundamental flaw in the way we administer tPA to patients with PE: using a fixed dose for all comers works poorly. This strategy gives some patients too much drug, while giving other patients an inadequate dose. Variable effects of a fixed-dose strategy on different patients will make it impossible to find the “optimal dose” of tPA in RCTs – because there is no such thing as a population-wide “optimal dose.” Any RCT embarking to find the optimal dose of TPA is fundamentally flawed from inception and likely doomed to failure.
How should tPA be dosed?
1) tPA dosing in PE should be weight-based
This isn’t a revolutionary concept. tPA dosing in ischemic stroke is weight-based. Obesity is a risk factor for PE, so PE patients tend to span a roughly four-fold range of weight (e.g. from ~50 kg to ~200 kg). Giving the same dose of tPA to someone who weighs 50 kg and someone who weighs 200 kg just doesn’t make sense.
2) tPA should be titrated to physiologic effect
Every patient has a complex and dynamic balance between fibrin generation and fibrinolysis. Some patients tend to degrade fibrin (e.g. hyperfibrinolysis) whereas other patients are resistant to the degradation of fibrin.5 Thus, the same exact dose of tPA to two patients of equal weight may have dramatically different physiologic effects.
The concept of monitoring fibrinogen levels during fibrinolysis has been explored before. If you start paying attention to these levels, it will become clear that the effect of tPA on any specific patient varies widely in an unpredictable fashion. This has been demonstrated already in several studies, but seeing it in practice makes it scary and harder to ignore.6
We wouldn’t imagine giving a heparin infusion without lab monitoring. tPA infusions administered by interventional radiology are generally monitored with serial fibrinogen levels to avoid excessive fibrinolysis. However, the concept of titrating tPA to biologic effect for pulmonary embolism remains an underutilized concept.
There are roughly two ways that tPA could be titrated to physiologic effect:
- Give tPA in small boluses (e.g. 12-24 mg). Measure fibrinogen levels before and after the tPA bolus, as well as the patient’s hemodynamic response. Additional doses of tPA could be administered if the patient remains unstable and fibrinogen levels remain relatively unchanged.
- Give tPA as a continuous slow infusion. Monitor hemodynamics and serial fibrinogen levels. Stop the infusion if the fibrinogen level falls too low, or after the patient improves hemodynamically. This concept is borrowed from the interventional radiology literature, where an extensive body of evidence supports its safety (more on this strategy here).
Overall, it’s important to bear in mind that the body will eventually break down the clot. The goal of tPA is merely to accelerate clot lysis enough to remove the patient from immediate danger (figure above). In order to achieve this humbler goal, much less tPA may be required than has traditionally been used (e.g. perhaps 12-24 mg rather than 50-100 mg).
Caveats with the OPTALYSE trial
The OPTALYSE PE trial was funded by the manufacturer of the EKOS catheter. The study lacks several important control groups, so interpreting it feels like riding a unicycle in a hurricane. The lack of control groups may be strategic: including more controls would threaten to disprove the efficacy of the EKOS catheter. The following controls are missing:
- There is no control group involving catheter-directed tPA without ultrasonic energy. This makes it impossible to determine whether the vibrational energy of the EKOS catheter is doing anything.
- There is no control group involving peripheral administration of tPA. This control would allow us to determine whether catheter-directed administration of tPA is more effective than systemic tPA.
- The primary endpoint is measured after 48 hours, more than a day after completion of the tPA/EKOS treatment. Choosing a late time-point means that we are observing a combination of the results of the procedure plus two days of endogenous thrombolysis (even with only heparin therapy, some thrombolysis will naturally occur over time).7 This will make tPA/EKOS appear more effective than it actually is. Evaluating efficacy at an earlier timepoint would control for this.
- There is no heparin-only control group.
The only way to make sense of this trial in the absence of control groups is to make assumptions. I have made the following assumptions. Although these seem reasonable based on available evidence, they remain my own assumptions which many folks will surely disagree with:
- Assumption #1: Ultrasonic energy provided by EKOS has no efficacy.8 9
- Assumption #2: Peripheral tPA infusion has similar efficacy compared to catheter-directed tPA infusion.10
- Assumption #3: The majority of improvement over the first 48 hours was due to the tPA therapy (not endogenous thrombolysis).7
- Assumption #4: Hemodynamics wouldn’t improve very rapidly in a heparin-only group.
Now here is where things get a bit sketchy. The study used a 48-hour timepoint, which makes it impossible to tell how much improvement is due to tPA/EKOS versus endogenous clot lysis over time. Nowhere in the manuscript is there evidence about timepoints before 48 hours. However, a promotional brochure by the company marketing EKOS catheters contains the following figure:
This shows that most of the improvement in RV/LV ratio does indeed occur over the first four hours of treatment, supporting assumptions #3 and #4 above. That’s good. However, it’s odd that important data is being used for promotional material, without being included in the manuscript.
Parting shot: Correct tPA dose for massive PE?
This has been a largely theoretical post, but let’s end with a practical point. Imagine the following scenario:
- 83-year-old woman presents with shock due to massive PE. She is hypotensive, but stabilizes with 8 mcg/min norepinephrine infusion.
- Past medical history of myocardial infarction, currently on aspirin and clopidogrel.
- Echo shows IVC dilation, RV dilation, and impaired RV systolic function.
- Clinically she looks OK but is unequivocally dependent on norepinephrine.
- She is thin, weighing 55 kg.
This patient has a massive PE and should be treated with tPA. But what dose? The traditional dosing would be 100 mg tPA for a massive PE. However, as reviewed above, several studies haven't found any benefit to 100 mg versus 50 mg. This patient isn't crashing, and she is at increased risk of intracranial hemorrhage (due to age and anti-platelet agents). My approach here would be to start with 50 mg tPA over two hours, follow fibrinogen levels and hemodynamics, and then consider an additional 50 mg tPA if needed. My experience (consistent with Wang 20102) is that 50 mg is often adequate to stabilize this sort of patient.
- The optimal dose of tPA for PE remains unknown. No RCT has been able to demonstrate any difference in efficacy when comparing different doses. This is perhaps most notable within the recent OPTALYSE trial, which found equivalent results when using either ~24 mg, ~12 mg, or ~8 mg tPA.
- Most studies use fixed-dose tPA dosing. This fails to account for differences in patient weight or the patient’s balance between fibrinolysis versus fibrin generation. Failure to account for these variables could make fixed-dose tPA a shot in the dark: some patients will receive an excessive dose while other patients receive an inadequate dose.
- tPA should arguably be provided using weight-based dosing with titration to physiologic effect. It’s possible that dosing tPA in a precise fashion could allow for optimization of the risk/benefit ratio for each individual patient (rather than guessing a dose which will work OK across a diverse population of patients).
- The optimal dose of tPA for massive PE is unknown. 100 mg is the traditional dose, but 50 mg may be reasonable for patients at increased risk of bleeding who aren't acutely unstable.
- Should we monitor fibrinogen during full- and half-dose PE thrombolysis?
- Controlled thrombolysis for submassive PE
- Deconstructing ultrasound-assisted thrombolysis for PE: What is actually doing the work?
- Submassive PE 2017: Getting 'em off the cliff
Image credit: pin the tail on the donkey