“Even though coronary, cerebrovascular, and venous thromboses are the leading killers of adults in ICUs, clinicians often have greater apprehension about bleeding than clotting.” Todd Rice and Art Wheeler wrote this first sentence in a review of coagulopathy in critically ill adults in 2009 (1). I include it to start this post for two reasons: 1. If you did not get to know Art when he was alive, this is a classic Wheeler statement that completely puts a massive critical care topic (coagulopathy) into clear perspective and 2. This is the perfect way to frame the discussion of the PATCH Trial.
The PATCH Trial (Platelet transfusion versus standard care after acute stroke due to spontaneous cerebral haemorrhage associated with antiplatelet therapy: a randomized, open-label, phase 3 trial)(2) beautifully illustrates Todd and Art’s point. Antiplatelet medications, such as aspirin, have provided marked benefit in primary and secondary prevention of the most common cause of death (cardiovascular disease) with 120 more lives saved among every 10,000 patients compared to not taking aspirin; yet, there remains a focus on the 2.6% risk of hemorrhagic stroke with aspirin users compared to 1.2% in non-users (3). Additionally, if you bleed into your head while on an antiplatelet medication, there is an associated increased risk of death than if you bleed into your head and are free of antiplatelet medications. Newer antiplatelets (clopidogrel, dipyridamole) may have an associated risk of intracerebral hemorrhage as high as 4%. We critical care physicians don’t get to see the millions of healthy patients who benefit from primary and secondary prevention with these medications, so it shouldn’t surprise us, as Todd and Art wrote, that apprehension remains about the bleeding risks of these medications manifesting as a constant quest for “reversal strategies” to calm our fears.
The PATCH trial was an unblinded, randomized trial of platelet transfusion to reverse the antiplatelet effects of three potential medications: aspirin, clopidogrel, and dipyridamole. In 60 hospitals in the Netherlands, UK, and France over 6 years (2009-2015), adults presenting to the ED with supratentorial intraparenchymal hemorrhage and who had been taking one or more of the above medications for at least 7 days were included. Patients were randomized to usual care or platelet transfusion within 90 minutes of identifying the bleed on imaging. In patients randomized to platelet transfusion, patients on aspirin got one platelet concentrate versus two platelet concentrates for patients on clopidogrel. The primary outcome was the modified Rankin score at 3 months, a typical neurocritical care outcome as describing life with good or bad functional outcome is accepted as a more patient-centered outcome in brain injury than just dead or alive.
This is basically a trial of patients on aspirin, who presented looking pretty good as far as ICH goes (ICH score of 1), low volume bleeds, and median GCS on presentation of 14 (Table 1). There were some baseline imbalances between groups that may have influenced the results (more aspirin only and intraventricular extension in patients in usual care, higher ICH volume in the platelet group). In almost every analysis they performed including the primary analysis, patients randomized to platelet transfusion caused more death and poor neurologic outcomes at 3 months. Additionally, there were significantly more serious adverse events due to intracerebral hemorrhage in the transfusion group, such as ICH enlargement, edema, herniation, intraventricular extension and hydrocephalus.
Although not mentioned by the authors, there is ample evidence that this was a tough trial to conduct. They took 6 years and used 60 hospitals in 3 countries to enroll 190 patients. To put this in perspective, the original study that validated the ICH score enrolled a similar number of patients with ICH at 1 hospital over 2 years (4). The trial did not collect screening or eligible but not enrolled data, so we have no idea how representative the trial patients are of the general ICH patient population. The data presented suggests that the trial patients are very different (very low ICH severity compared to the historical ICH population, only a total of 5 patients on clopidogrel enrolled) which raises the likely possibility that there isn’t equipoise on this intervention and physicians just didn’t allow patients with more severe ICH or those on clopidogrel to be enrolled in this trial. This doesn’t affect the internal validity of the trial, but the external validity is definitely in question. Further supporting this being a difficult trial to conduct, 19% of the patients randomized should never have been enrolled in the trial as they had at least one exclusion criterion. It was reassuring that the result didn’t change when these patients were excluded from a sensitivity analysis.
In patients on antiplatelet medications (predominantly aspirin) who have a bleeding brain, why would platelet transfusion cause death and worse neurologic outcome? One potential explanation is that platelet transfusion is not harmful and the result is just a function of baseline imbalances between groups. I think this explanation is a stretch as the imbalances went for and against transfusion. The other potential, and more worrisome, explanation brings us back to Todd and Art’s point: maybe the trial transfused platelets to patients with one of the most common causes of death, thrombosis. Hemorrhagic conversion of an ischemic stroke is common in the patient population they enrolled (mean age 74 yrs) and possibly suggested by the 54 patients who had lobar ICH rather than deep ICH.
My conclusion: platelet transfusion to reverse the antiplatelet effects of aspirin in patients with relatively mild ICH cannot be recommended. I’m not saying we now have the definitive evidence to support this statement, just that we have gone from zero evidence for or against this practice to a randomized trial showing harm. I don’t think we can generalize these data to patients on non-aspirin antiplatelet medications. It may be reasonable to argue that we can’t generalize these data to more severe ICH; however if transfusions didn’t help early I’m not sure why they would then help late. Finally, I am still more worried about diseases of thrombosis than I am of bleeding.
- Rice TW, Wheeler AP. Coagulopathy in critically ill patients: part 1: platelet disorders. Chest 2009;136:1622–1630.
- Baharoglu MI, Cordonnier C, Salman RA-S, de Gans K, Koopman MM, Brand A, Majoie CB, Beenen LF, Marquering HA, Vermeulen M, Nederkoorn PJ, de Haan RJ, Roos YB, PATCH Investigators. Platelet transfusion versus standard care after acute stroke due to spontaneous cerebral haemorrhage associated with antiplatelet therapy (PATCH): a randomised, open-label, phase 3 trial. Lancet 2016;0:.
- He J, Whelton PK, Vu B, Klag MJ. Aspirin and risk of hemorrhagic stroke: a meta-analysis of randomized controlled trials. JAMA 1998;280:1930–1935.
- Hemphill JC, Bonovich DC, Besmertis L, Manley GT, Johnston SC. The ICH score: a simple, reliable grading scale for intracerebral hemorrhage. Stroke 2001;32:891–897.