Thrombolysis vs delayed PCI for STEMI – the STREAM trial (+ 1 case study)


Published in the NEJM last April was a study titled “Fibrinolysis or Primary PCI in ST-Elevation Myocardial Infarction”… or “the STREAM trial” to its friends. This is a largish (n = 939 + 943) and fairly well-designed randomised interventional trial comparing two treatment strategies for all-comers with STEMI who presented within 3 hours of symptom onset but could not get to primary PCI within one hour. The treatment arm was given tPA (tenecteplase) as soon as possible and then had delayed (6-24 hrs) PCI. The control arm got no tPA but got to primary PCI as soon as possible. The primary outcome was a composite of death, shock, CCF, or re-infarction at/within 30 days.

Interestingly, 36% of the tPA group required urgent rescue angioplasty due to failure of reperfusion with fibrinolysis (median time to PCI = 2.2 hrs). The patients who were given tPA and did reperfuse had a median time to PCI = 17 hrs. The primary PCI group had a median time to PCI = 178 mins, or close enough to 3 hrs.

They found no significant difference in primary outcome between the two groups. Rates of intracerebral haemorrhage were 1.0% vs 0.2% for the tPA and primary PCI groups respectively, and the need for CABG was 4.7% vs 2.1% for tPA vs PCI as well. Both differences were statistically significant given the power of the study.

This trial was funded by Boehringer Ingelheim, the manufacturer of tenecteplase, and while maintaining quite rigorous objectivity, the authors summary conclusion was that:

“…a strategic alignment of prehospital or early fibrinolysis and contemporary antithrombotic cotherapy coupled with timely coronary angiography resulted in effective reperfusion in patients with STEMI who presented within 3 hours after symptom onset and who could not undergo PCI within 1 hour after the first medical contact. However, early fibrinolysis was associated with a slightly increased risk of intracranial bleeding”.


Not an entirely unreasonable precis of the results, but I think I would probably portray the findings a little differently in light of the spectre of the “discussions” with my cardiologically inclined colleagues that will undoubtedly ensue when decisions have to be made about urgent transfer for PCI for a 3am STEMI patient versus early tPA at a smaller centre and delayed PCI.

A couple of points are worth noting before everyone jumps on the “tPAyyy and de-layyy” bandwagon (complete with satisfied smiles and holding of hands):

  • There was no difference in primary (cardiac) outcome for the punters in each group.
  • 36% of the tPA/delay group required urgent rescue PCI to achieve this equality.
  • The tPA/delay group had a 5-fold increased risk of ICH.
  • The tPA/delay group had a 2-fold increased requirement for CABG.

For me, the practical upshot of all of this is:

My remote STEMI patient has a 74% chance of the optimal outcome (minimum cardiac badness at 30 days), 5 times the risk of ICH, and twice the risk of needing a CABG if I choose to give then tPA and transfer them in a leisurely manner for PCI, rather than having them haul arse to the cath. lab. as soon as possible.

I am but a simple emergency doctor, a wandering vagrant, an illegal alien, if you will, in the ivory, monitored-bed-lined  tower of academic cardiology. But given the information above, call me demanding, but I’m going to be advocating transfer for primary PCI every time.

Interestingly, the median time to urgent primary PCI in this trial was 178 minutes, suggesting that the current consensus international guidelines for choosing early tPA over transport for PCI if the delay to PCI is > 90 minutes might be a little too conservative; the subjects in the urgent PCI arm of this trial had equivalent cardiac outcomes and a lesser adverse event rate (ICH, need for CABG) with a median PCI delay of 3 hours. The tricky question is still what to do with the STEMI patient who is 1-2 hours from your cath. lab… do you give them tPA pre-hospital (or pre-transfer from a smaller centre) aiming for a chance (significantly < 100%) at earlier reperfusion at the expense of a higher risk of bleeding, or transfer them urgently without thrombolysis, as this study suggests that we can expect equivalent benefit from just scooping and running to the cath. lab at least out to 3 hours.

In summary, one can still debate the merits of tPA vs no tPA if you can get to PCI in under 2-3 hours, but in light of the STREAM trial results, I think it is reasonably clear that even if you do opt for tPA, there is no evidential basis for intentionally delaying transfer for PCI as soon as possible.

The full text of the paper can be found here:

The PCI/tPA decision-making process discussed above was relevant to two real cases I was involved with, and might be worth briefly discussing:

Case 1 – “This…. is lucky Phil…”    … or …    “Things that go splat in the night”

‘Twas a dark and brooding night. The pale, second-hand light of a cold and indifferent moon struggled to seep through the low clouds over the misty moors of the Southern Highlands. From a distance, the horrendously brain-jarring sound of possums shagging pierced the silence of the night… and who knows, maybe that’s why two unrelated truck drivers sleeping at the same truck stop were both awake at the same time in the wee hours of the morning? In any case, our intrepid hero, let’s call him Phil, was walking back to his truck after performing some necessary ablutions, when he was seen by another truck driver to wobble a bit, grunt slightly, and then faceplant with all the grace and poise of a sack of potatoes, onto the ground. A long, silent moment passed. Even the possums reined it in a bit. It became increasingly obvious to our serendipitous bystander that Phil was seemingly not getting back up… nor moving… and indeed, on closer inspection, didn’t seem to be bothering to do much in the way of breathing, either. This situation elicited some concern from our good samaritan, rapidly followed by a phone call to 000 for an ambulance, and then the immediate commencement of what, with the aid of the retrospectoscope, must have been absolutely top-shelf bystander CPR.

NSW Ambulance responded swiftly, and arrived to find Phil unresponsive and in cardiac arrest, with VF as the rhythm de jour. They promptly delivered a single DC shock. Shortly afterwards they phoned the The Big Hospital’s Admitting Officer to report that they currently had a 52 year old man post collapse and VF arrest, in whom they had achieved ROSC, and who was now in sinus rhythm, with a decent blood pressure, but was looking persistently grey and clammy, clutching his thorax in a disconcerting manner, and insisting that despite all visual evidence and spatial inconsistency to the contrary,  he was quite sure he had an elephant sitting on his chest. A pre-hospital 12-lead ECG was obtained, and looked a bit like this:


“Oh, poo…”, paraphrased the Admitting Officer, and politely asked where the crew currently were, and what their transport options were.

“We can get him to the Little Rural Hospital in 30 minutes, or to The Big Hospital in 60 minutes.” offered the Ambulance crew.

— Pause for reflection —

  • Little Rural Hospital is closer, and has tPA on hand.
  • The Big Hospital has a cardiac cath lab. with an average activation to ready-to-rock’n’roll time of 20-30 minutes.
  • You are James T.Kirk…. Captain of the USS Acute Flight Deck at 3am…
  • Do you tell them to go to The Little Hospital, get him thrombolysed, then transfer for PCI later sometime?
  • Do you tell them come hard to port and punch it at warp speed to The Big Hospital for PCI as soon as possible?

Whatcha gonna do…. Whatcha donna do-oooo?   Whatcha donna do-when-there’s-a-STEMI for youuu ? (Must be sung to the theme from “COPS”)


And thus the Admitting Officer did ponder and consider their options for a nanosecond or two, and spake unto the stolid Ambulance crew: “Heparin. Prasugrel. The Big Hospital as fast as you can. I’ll have the cath lab. ready for the patient by the time you get here.” And the Ambulance crew did crack on at a clipping pace, and transported the patient at best speed to The Big Hospital. The Admitting officer also briefly pondered the order in which the two tasks “Activate cath lab” and “Talk to cardiologist” should best occur in order to secure the optimal outcome for the patient. At this point it is unclear whether it was innate cynicism, prior knowledge of the course of similar conversations, the creeping nihilism of the I-am-just-so-very-over-it end of a night shift, an unusually high midichlorian count, or merely sensory input from somewhere in the myenteric plexus, but the Admitting Officer opted to activate the catheter laboratory first, and then fielded the phone call from the interventional cardiologist a few minutes later.

A polite but pointed conversation followed, in which the cardiologist was appraised of the patient’s situation and the logistic/treatment options, expressed their displeasure at the fact that the crew had not been directed to Goulburn for thrombolysis, and conveyed their opinion of the Admitting Officer’s (and, by extension, ED clinicians’) decision-making in such circumstances. The terms “guidelines” and “best evidence” may have been mentioned, but seemed to have little effect.

— Pause for reflection —

Even prior to the STREAM trial (April 2013), the American (AHA), European (ESC) and Australian (NHF) guidelines for tPA vs PCI for STEMI were pretty clear:


The cath. lab. team arrived at The Big Hospital prior to the patient, and as per protocol he bypassed ED and was delivered directly to the the suite for immediate PCI. Angiography revealed the expected totally occlusive lesion, balloons were inflated, stents were placed, and a much happier patient was admitted to CCU with good systolic function, a GCS of 15, and a clinically insignificantly different number of viable brain cells compared to the previous day. As intimated earlier, this patient’s excellent outcome is testament to both his exceptional luck, and what must have been superlatively effective bystander CPR.

Careful perusal of the guidelines linked to above will reveal that it is recommended that if you’re more than 30 minutes from a hospital, then pre-hospital tPA is a perfectly reasonable thing to do. When you’re within 90 minutes of a PCI-capable centre, however, it is questionable whether the additional risks of thrombolysis (especially while not in a hospital) are balanced by any significant time-based benefit for the patient. Sure, you might re-perfuse some or all of their ischaemic myocardium a little sooner, but you might not (rescue PCI was necessary in 34% of the STREAM trial patients, and in 19% of cases during the pilot implementation of the NSW rural/remote pre-hospital STEMI thrombolysis protocol: ) and you definitely introduce a significant risk of major and possibly catastrophic bleeding either pre-hospital, or while on the cath. lab. table. The STREAM trial suggests that one might consider stretching the window of avoid-tPA-and-go-fast-to-PCI out to 3 hours, but current guidelines still maintain consensus on the 90 minute timeframe for PCI over tPA, so please keep this in mind when making clinical decisions, defending yourself in court, or discussing options with your cardiology colleagues.

The other major take home point in this case is the issue of the inevitable delay introduced, no matter what the pathology involved, when you take a patient from the scene to an intermediate facility for “stabilisation” or any other non-definitive treatment, prior to getting them to a centre that can provide definitive care for their condition. This is a huge bug-bear of retrieval medicine, and I should take pains to point out that while sometimes it is absolutely, totally and unequivocally appropriate to sprint to a smaller centre first, in the majority of cases it is not appropriate and, without exception, the delay so introduced will be longer than anyone involved anticipates. The practical upshot is a significant and usually deleterious delay to definitive care. This should be strenuously avoided, and unless there is a very clear indication for making a pitstop, you should be aiming to do everything you can to minimise the “therapeutic vacuum”.

[ Second case study may follow after some editing / redacting of some identifying / incriminating details…]


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s