A retrospective analysis of close to 1,000 patients published in World Neurosurgery has revealed that left-sided anterior-circulation large vessel occlusion (LVO) strokes treated via thrombectomy may be associated with worse outcomes compared to their right-sided equivalents—a phenomenon seemingly driven by a multitude of factors including baseline clinical severity and puncture-to-recanalisation (PTR) times. Here, corresponding author Daniel Sahlein (Goodman Campbell Brain and Spine, Indianapolis, USA) discusses the minutiae of these findings with NeuroNews.
What were the most interesting findings within this single-centre study?
We wanted to see whether or not patients with anterior-distribution LVO stroke have different outcomes based on laterality, because there are surprisingly little literature or data on this. As language centres are located in the left side of the brain, it would make sense that patients with left-sided stroke do worse clinically—and we did find that these patients had worse outcomes. That was partly because they had higher National Institutes of Health stroke scale (NIHSS) scores at presentation, which is one of the major determinants of stroke thrombectomy outcomes.
However, something that really surprised us was the fact that our PTR time was greater on the left than on the right. In fact, when we did a multivariate analysis including PTR time but also things like age and degree of recanalisation, we were able to confirm that discrepancies in time between the two sides seem to make a difference, which was unexpected. There was something even more interesting about the histograms of those times as well: the two sides look very similar up to about 15 minutes, at which point there’s a peak in PTR time. But, while the histogram on the left side falls off pretty quickly after that, it’s actually bimodal on the right side—it goes down, and then there’s another peak 10 minutes later before it starts to decrease again.
What do you think are the underlying reasons for these findings?
When we sat down to try to explain it, we concluded that it’s really to do with aortic arch anatomy. In some cases, the left and right sides look pretty similar, and are both fairly straightforward to catheterise, meaning you end up with a very short PTR time. However, on the left side, you have this fairly common variant anatomy called a bovine arch, where the common carotid artery comes off the subclavian artery and then takes a very tight turn back along the aortic arch. And, ultimately, to catheterise that type of anatomy, you need a catheter with a back bend—called a Simmons 2—in order to make that turn. Using a standard Berenstein-tip catheter is going to fail, and then you will eventually switch to a Simmons 2, which essentially means the catheterisation is going to take you an extra 10 minutes.
“A key learning from this is just to pay really close attention to your stroke outcomes and try to audit every step of the stroke continuum carefully”
We don’t have second-by-second granularity to see exactly what was happening in the left-sided versus right-sided procedures, but this hypothesis seems very plausible and rational. Our analysis also showed that the degree of recanalisation and number of passes were the same on both sides, meaning we were not technically more successful on one side versus the other. In other words, once we got up into the neurovasculature, the procedures went very similarly, so the fact the earlier part of the procedure is not as consistent may account for those differences by laterality. I don’t think that time is the only determinant of the difference in outcomes—there’s also the idea of hemispheric dominance, and so forth—but time was a determinant, and it was more of a determinant than we were expecting.
Why is there so little published clinical research on stroke laterality?
I don’t know, but I do find it a bit perplexing. I think people feel that a lot of this is covered by NIHSS scores; essentially, you might say that, if you have an NIHSS score of 20, why does it matter which side it’s on? I find that argument to be a little challenging, because NIHSS is both non-linear and non-equivalent. If you’re running a clinical trial, it might work grossly because you need to have some line in the sand in terms of stroke severity, but one NIHSS score of 20 is not necessarily the same as another NIHSS score of 20. They’re composed of a totally different sum of functional deficits, and the likelihood of recovery is not the same even between identical NIHSS scores. The fact it’s non-linear also means the steps between 5 versus 6, and 12 versus 13, are not the same. That speaks to the overall challenge of coming up with any sort of quantitative scale that accurately captures the true essence of stroke severity—and that’s why it’s important to break these things down into other meaningful and more nuanced categories.
What are the most important messages you would convey based on these findings?
I think one of them is the importance of access and access planning. These data went from 2011–2020, but we have a pretty large amount of more recent data, and that bimodal peak actually disappeared and became unimodal in our more recent dataset. That shows our group has improved over time and become better at predicting these more challenging anatomies based on hundreds of thrombectomy cases. But, for groups that are not doing this many procedures on a per-physician basis, I think there are two things that can help. One of them is getting a better three-dimensional sense of the aortic arch and the anatomic architecture that you’re going to need to traverse, and that’s where I think automated 3D reconstruction technologies [like Lumina 3D (RapidAI)] offer a really tremendous opportunity because of the confidence they give you in your catheterisation. The second is simply catheter construction and selection, with respect to picking a slip catheter to help you get in position, and a wire that’s stiff enough, and all these kinds of things.
Another key learning from this is just to pay really close attention to your stroke outcomes and try to audit every step of the stroke continuum carefully, with really robust data, and take nothing for granted. We did not expect to see this finding, and we only saw it because the quality of our datasets is extraordinary. So, it starts with the data, and it also starts with being cognisant of your blind spots and having an open mind as to what you might find, because that’s how we end up unearthing some of our most interesting discoveries.
