Lots of good comments here. As an operator, I think the answer to the question as to whether to intervene in non culprit lesions during STEMI interventions is "sometimes- but not just because you see it." Clinical features and status of the patient as well as criticality and locations of the other lesions all feature in as noted. And no meta analysis of these types of trials as constituted is going to help much with teasing any of that out.
It perhaps should also be considered where in the coronary tree is the “culprit “and “ non- culprit “ lesions. It may be different if the “culprit “ lesion is in the mid circumflex and the “non culprit “ lesion is a high grade proximal LAD lesion or visa versa. The culprit lesion is in the proximal LAD and the non culprit lesion is in the mid circumflex. Perhaps there is room for clinical judgment?
Dr. Foy has, as usual, nailed it here. Meta-analyses are always somewhat suspect. Typically, they are undertaken in an attempt to squeeze some significance out of a group of related studies that, individually, show contradictory results with very small differences between the groups being compared. When composite endpoints (some with questionable definition) are employed, the conclusions should be read with extra caution. When a number of studies show small differences in both negative and positive directions that is, in my opinion, enough to say that the intervention being studied offers no benefit. Dr. Foy is being kind when he refers to the early termination of the FULL REVASC trial. He implies that an uncharitable assessment of their motivation would lead one to question the scientific honesty of the investigators. I interpret that action as evidence of blatant bias and scientific fraud.
Great piece. I learned a lot about the critical appraisal aspects.
I wouldn’t characterize “complete” revasc in a STEMI pt as being simply akin to PCI of stable CAD. The STEMI pt is categorically (and I would argue, pathophysiologically) different from the stable CAD pt who is walking around and going about their day. Otherwise stable plaques in a milieu of recently ruptured plaque (and acute occlusive disease in some location) may not behave the same as without such a milieu. Whereas there is no evidence to support routine PCI in “stable” “stable CAD”, I consider COMPLETE to be sufficient for me to advise complete revasc in the context of STEMI, for the purposes of reducing MI (and recognizing no benefit in CV and all-cause mortality).
Usually, I’m on the lookout with meta-analyses that are dominated by a few large trials that simply overwhelm a bunch of smaller studies wrt the primary endpoint and/or the alternate endpoints of interest. Your example is very interesting where a bunch of small studies generate a hard endpoint benefit that had specifically been null in the largest trials. I would typically have viewed the conclusions of the meta-analysis authors dimly, but thanks for methodically showing why it is a conclusion that neutral data-consumers should have no confidence in.
As Dr Foy analyzes so thoroughly in this essay, the 'Do more' bias is strongly evident. This phenomenon is pervasive across medicine. I'd bet that the five cardiology societies pushing for the broader intervention have an economic conflict of interest, either by funding from device makers, or just higher fees when they perform the more extensive PCI (percutaneous coronary intervention) rather than the more conservative intervention.
Thanks, Dr. Foy. May I point out that the acronym PCI is nowhere defined in your commentary, a real shame for those of us not in cardiology. I did appreciate your points on the potential errors introduced by meta-analysis.
Lots of good comments here. As an operator, I think the answer to the question as to whether to intervene in non culprit lesions during STEMI interventions is "sometimes- but not just because you see it." Clinical features and status of the patient as well as criticality and locations of the other lesions all feature in as noted. And no meta analysis of these types of trials as constituted is going to help much with teasing any of that out.
It perhaps should also be considered where in the coronary tree is the “culprit “and “ non- culprit “ lesions. It may be different if the “culprit “ lesion is in the mid circumflex and the “non culprit “ lesion is a high grade proximal LAD lesion or visa versa. The culprit lesion is in the proximal LAD and the non culprit lesion is in the mid circumflex. Perhaps there is room for clinical judgment?
D. Thomas Combs
I love reading these. As someone who worked on the research agreements for one of these studies, it’s nice to see the outcomes analyzed.
Great review. Thank you so much.
Dr. Foy has, as usual, nailed it here. Meta-analyses are always somewhat suspect. Typically, they are undertaken in an attempt to squeeze some significance out of a group of related studies that, individually, show contradictory results with very small differences between the groups being compared. When composite endpoints (some with questionable definition) are employed, the conclusions should be read with extra caution. When a number of studies show small differences in both negative and positive directions that is, in my opinion, enough to say that the intervention being studied offers no benefit. Dr. Foy is being kind when he refers to the early termination of the FULL REVASC trial. He implies that an uncharitable assessment of their motivation would lead one to question the scientific honesty of the investigators. I interpret that action as evidence of blatant bias and scientific fraud.
A suggestion: “Numerous trials have proven that preventive PCI…” This would better read: “Numerous trials strongly support that preventive PCI…”
My reasoning is that the words ‘prove’, ‘proven’, ‘proved’ only rarely (if ever) can be prudently used in writing about scientific research.
Fair point. That was my edit. Point taken.
Great piece. I learned a lot about the critical appraisal aspects.
I wouldn’t characterize “complete” revasc in a STEMI pt as being simply akin to PCI of stable CAD. The STEMI pt is categorically (and I would argue, pathophysiologically) different from the stable CAD pt who is walking around and going about their day. Otherwise stable plaques in a milieu of recently ruptured plaque (and acute occlusive disease in some location) may not behave the same as without such a milieu. Whereas there is no evidence to support routine PCI in “stable” “stable CAD”, I consider COMPLETE to be sufficient for me to advise complete revasc in the context of STEMI, for the purposes of reducing MI (and recognizing no benefit in CV and all-cause mortality).
Usually, I’m on the lookout with meta-analyses that are dominated by a few large trials that simply overwhelm a bunch of smaller studies wrt the primary endpoint and/or the alternate endpoints of interest. Your example is very interesting where a bunch of small studies generate a hard endpoint benefit that had specifically been null in the largest trials. I would typically have viewed the conclusions of the meta-analysis authors dimly, but thanks for methodically showing why it is a conclusion that neutral data-consumers should have no confidence in.
As Dr Foy analyzes so thoroughly in this essay, the 'Do more' bias is strongly evident. This phenomenon is pervasive across medicine. I'd bet that the five cardiology societies pushing for the broader intervention have an economic conflict of interest, either by funding from device makers, or just higher fees when they perform the more extensive PCI (percutaneous coronary intervention) rather than the more conservative intervention.
Thanks, Dr. Foy. May I point out that the acronym PCI is nowhere defined in your commentary, a real shame for those of us not in cardiology. I did appreciate your points on the potential errors introduced by meta-analysis.