Thread by @TeamCaptainJohn

Inspired by @BenMazer and the topic of #platelet #transfusion in #cerebralhemorrhage patients previously taking anti-platelet therapy, I want to take a few minutes to talk about the EVIDENCE in platelet transfusion indications, thresholds, and efficacy. #blooducation

To recap, the PATCH trial (Lancet 2016) concluded that in patients taking anti-platelet medications (aspirin, plavix, etc) who had spontaneous cerebral hemorrhage, platelet transfusion was assc with WORSE outcomes than standard medical therapy.

This is shocking? Right? Here's the classic logic:
-ICH is bad
-ICH is worse if your platelets are iatrogenically broken
-Non-broken platelets are good
-Transfusion of non-broken platelets helps stop the bad bleed

Do you feel lied to by modern medicine? I do. Vote if you agree.

Like any transfusion, platelets, are *potentially* life saving. Debating when that *potential* applies is point is not the point here. Instead, lets talk about what the efficacy evidence we actually have. And lets us the context of transfusion guidelines as a framework.

So lets turn to the @AABB platelet transfusion guidelines, as published in the ANNALS OF I.M. which provide guidlines for a variety of clinical scenarios:
-Prophylactic in chemo patients
-CVC catheters
-Lumbar punctures
-Major surgery
-Cardiac Surgery
-ICH on anti-plt

This is a robust clinical guidline paper, authored by many famous people with far more brain power, experience, and letters after their name than me. Feel free to peruse their methods to learn about their lit reviews. Spoiler alert: their lit search spanned 113 years. Legit.

So lets take a look at each one of these guideline scenarios, and what evidence they have to support the position, starting with Iatrogenic Thrombocytopenia as a result from chemotherapy (10k/uL threshold).

Recommendation 1: The AABB recommends that platelets should be transfused prophylactically to reduce the risk for spontaneous bleeding in adult patients with therapy-induced hypoproliferative thrombocytopenia.

Quality of evidence: moderate; strength of recommendation: strong.

So its pretty clear that prophylactic transfusions prevent WHO Grade 2 bleeds, based on three RCTs, with over 1000 patients. Based on historical perspectives (most pt who died during induction chemo died because of ICH bleeding)

When it comes to threshold, 4 RCTs with >650 patients found that tx at 10k/uL was not inferior to tx at 20 or 30k/uL. Not surprisingly, the lower threshold did have more days bleeding (less than WHO grade 2), but used less products and fewer reactions. Ok, that all checks out.

Ok so regarding dose in these patients, standard dose (4-6 pools or 1 apheresis unit) are not inferior to a "double dose"...why give two isn't just for pRBCs. But heres the thing: a low dose (~3 pools) is equally as effective in preventing higher grade WHO bleeds!

So it all seems reasonable to give 1 adult dose for prophylaxis in chemo patients with <10K/uL platelets. And while its reasonable to include this to ALL patients with plt <10K/uL (as AABB implies and we all do), we really don't have much (or any) good evidence to support this!

Dont forget, patients on active chemotherapy also have pretty beat up microvasculature, putting them at risk for bleeds far beyond patients who are profoundly thrombocytopenic for other reasons. Can we look to other scenarios to consider this issue: What about ITP patients?

Well, the 2011 ASH guidelines recommend withholding plt transfusion in ITP patients except in cases of life threatening bleeding, and (as far as i have ever been able to tell) do not use any plt count thresholds OTHER than the 30k/uL, solely to initiate medical therapy.

So while it certainly is reasonable to consider plt transfusion any patient with a plt <10k/uL, theres not exactly a mountain of evidence to support that. But at least in iatrogenic pancytopenia, the evidence is pretty good. Don't worry, it all goes downhill from here...

Ok so moving on to part two: Minor procedures including CVC placement and LP procedures

Recommendation 2: The AABB suggests prophylactic platelet transfusion for patients having elective central venous catheter placement with a platelet count less than 20 × 109 cells/L.
Quality of evidence: low; strength of recommendation: weak.

Ok so only 8 observational studies, and it took a multivariate analysis to show that (already rare) bleeding complications were higher risk in plt <20k/uL. Mind you, all of these complications were controlled with local pressure. But thats still a bleed.

So it seems reasonable to NOT place a CVC with a platelet count below 20K/uL, especially in someone deemed clinically high risk. Obviously, the risks have to be weighed, but even when a bleed occurs, its overwhelmingly likely to be solved with some compression.

But heres the thing: where is the evidence that platelet transfusion actually do anything in this population?

Recommendation 3: The AABB suggests prophylactic platelet transfusion for patients having elective diagnostic lumbar puncture with a platelet count less than 50 × 109 cells/L.
Quality of evidence: very low; strength of recommendation: weak.

Ok so for LP, we're again limited to observational studies. In the largest study with children, plt count didn't matter, as they had no bleeding even with plt <20K/uL. We do get a look at efficacy of prophy tx...but no one bled anyways, so we don't know what any of that meant.

Its important to note that the central nervous system is no place to play around. Having a low tolerance for bleeding risk is acceptable, but clinical judgement > treating a number. Next time you get a call for an LP and a plt count of 47, ask them if they know about the CBC's CV

We'll wrap up talking about major surgery in a new thread, as we've already touched on the issue of intracranial hemorrhage on anti-platelets (rec. 7), discussion of plt tx in cardiac surgery is a whole other issue (rec. 6), and this is getting exhausting....

So in the end, we shouldn’t be shocked when some RCTs show lack of efficacy (or even inferiority) in certain, *specific* situations with platelet transfusion (like in @BenMazer original post), because we simply haven’t asked the questions enough!