Hemorrhage increases periop Q wave MI and stroke following noncardiac surgery

A new article in Circulation reviews hundreds of thousdands of patients in the NSQIP database.  They excluded cardiac, transplant, trauma, and neurosurgery patients. They find that approximately 1% of patients have >4 U RBC transfusion.  These patients have a 3-4 fold increase in rate of periop Q wave MI and/or stroke (approx 1% each).  No surprises there.  Patients undergoing vascular surgery were significantly over-represented in the hemorrhage category.

This retrospective study does not address the relationship of preexisting anticoagulant or antiplatelet drug use in these patients.  The reporting did not include routine ROMI (rule-out MI), suggesting under-reporting of PMI; we know that troponin increases, even in the absence of Q waves, are associated with poor long-term outcome after vascular surgery.


The authors and the editorial below discuss that perioperative strokes are usually ischemic, not hemorrhagic; and that hypotension is probably more important that anemia (low oxygen carrying capacity) perse in contributing to periop stroke.

In an accompanying editorial, Lee Goldman points out a number of limitations and suggestions for future research:

Most studies suggest a ~ 5:1 ratio of PMI:stroke, not the 1:1 ratio seen here.

In the Carson study, the ratio was 5.5 postoperative MIs for each postoperative stroke. So, NSQIP, which may well have missed some small postoperative strokes, undoubtedly missed 80% of postoperative MIs by modern diagnostic criteria. As a result, Kamel and colleagues clearly have substantially underestimated the absolute risk of postoperative MI attributable to perioperative hemorrhage and may also have substantially underestimated the relative risk resulting from postoperative hemorrhage.

He then goes on to criticize the present evidence base and guidelines for perioperative management of antiplatelet and anticoagulant therapy.

The American College of Chest Physicians publishes evidenced-based clinical practice guidelines for the perioperative management of antithrombotic therapy, the last of which was published in 2012.12 Although each iteration of these guidelines is logical and consensus-driven, each is based on weak data, essentially none from randomized trials…

As noted by Kamel and colleagues, this problem cries out for better data, which can be obtained only from placebo-controlled randomized trials. And the problem is getting even more important… we really do not know at all how best to manage these patients through noncardiac surgery…  It simply is not appropriate for us to keep tap dancing with recommendations whose scientific basis is no more sound than a variety of treatment guidelines…

He suggests three large RCTs to address the following questions:

  1. How best to manage patients who are on aggressive antiplatelet therapy after coronary stenting
  2. How best to manage patients who are treated long-term with aspirin or other antiplatelet agents for the secondary prevention of myocardial infarction or ischemic cerebrovascular events
  3. How best to manage patients who have atrial fibrillation and are receiving prophylactic anticoagulation to prevent embolic stroke.


Comments are closed.