Stroke Scale Cutoff Might Not Be Ideal Guide for Ordering CTA and Detecting Large Vessel Occlusions

Aren't you glad your doctor is guessing about what to do when you present to the ER as a stroke patient?

Stroke Scale Cutoff Might Not Be Ideal Guide for Ordering CTA and Detecting Large Vessel Occlusions

BOSTON — In emergency department stroke consultations, the National Institute of Health Stroke Scale (NIHSS) alone does not appear to be a reliable guide for ordering diagnostic tests for a large vessel occlusion (LVO), according to large body of data presented at the 2023 annual meeting of the American Academy of Neurology.

If the goal is not to miss any LVOs, there is no NIHSS score below which these do not occur, according to Theresa Sevilis, DO, regional medical director, TeleSpecialists, Fort Myers, Fla.

For example, her evaluation of a large and nationally representative dataset shows that more than 10% of the LVOs eventually identified and accepted for intervention would be missed with a cutoff of NIHSS score of 6 or higher. Moving the cutoff NIHSS score to 4 or greater, 6% of LVOs among the 23,166 strokes evaluated would have gone undetected.

"The current guidelines do not address low NIHSS score largely due to a paucity of data," according to Dr. Sevilis, who showed data indicating that there is great variation among institutions in regard to ordering computed tomography angiography (CTA). She indicated that CTA is the current imaging standard for detecting LVO.

Large prospective dataset

The data for this study were derived from the TeleCare database, which captures acute stroke consultations in the emergency departments in 227 facilities in 27 states. Stroke consultations over a 6-month period from July through December 2021 were evaluated. The prospectively collected data were subjected to a multivariate analysis to determine the odds ratio for a CTA performed and LVO found at each NIHSS score of 0 to 5. Scores 6 or above served as the reference.

"Only consults performed within 24 hours [of presentation] were included," Dr. Sevilis said.

After excluding cases in which no NIHSS score was captured, which represented less than 1% of cases, more than 10,500 cases underwent CTA, providing a rate of 45.5%. The rate of CTA for the whole dataset was 45.5%. Of the study population, 24.6% had a NIHSS score of 6 or above.

"When you are discussing when to perform CTA in patients with a low NIHSS score, you are discussing the majority of patients," Dr. Sevilis said.

Of those with a NIHSS stroke of 6 or below, 28.2% had a score of 0. Not surprisingly, these were the least likely to have a CTA performed on the basis of an odds ratio of 0.14 and the least likely to have a LVO detected (OR, 0.1). With the exception of a NIHSS stroke score of 1, the likelihood of CTA and LVO climbed incrementally with higher stroke scores. These odds ratios were, respectively, 0.16 and 0.09 for a score of 1; 0.27 and 0.16 for a score of 2; 0.33 and 0.14 for a score of 3; 0.49 and 0.24 for a score of 4; and 0.71 and 0.27 for a score of 5.

In the group with NIHSS score of 6 or above, 24.1% were found to have an LVO. Of these, the proportion accepted for a mechanical thrombectomy was less than half. The intervention acceptance rate for mechanical intervention among LVOs in patients with lower NIHSS scores again fell incrementally by score. The acceptance rate was about 35% among LVO patients with a NIHSS score of 3 or 4 and 25% for those with a score of 0-2.

The interpretation of these data "depends on goals," Dr. Sevilis said. "If the goal is to not miss a single LVO, then it is important to consider the balance between benefits and risks."

No consistent cutoff

In participating facilities, the protocol for considering CTA to detect and treat LVOs ranges from neurologist choice to cutoffs of NIHSS scores of 2, 4, and 6, according to Dr. Sevilis. Where the data suggest that a cutoff of 4 or above might be reasonable, she said that NIHSS scoring is not a useful tool for those "who do not want to miss any LVOs."

These data are based on emergency room stroke consultations and not on confirmed strokes," Dr. Sevilis emphasized. Indeed, she noted that the final discharge diagnosis was not available. Recognizing that the analysis was not performed on a population with confirmed strokes is particularly important for understanding the limited rate of CTAs performed even in those with relatively high NIHSS scores. She noted this could be explained by many different reasons, including suspicion of hemorrhage or clinical features that took the workup in a different direction.

Reconsidering protocols

Based on the large sample size, Dr. Sevilis contended that it is likely that these data are representative, but she considers this study a first step toward considering protocols and developing guidelines for addressing stroke alerts in the emergency department.

A more important step will be ongoing trials designed specifically to generate data to answer this question. Pascal Jabbour, MD, chief of the division of neurovascular and endovascular neurosurgery, Thomas Jefferson University Hospitals, Philadelphia, is participating in one of these trials. He agreed with the premise that better evidence-based criteria are needed when evaluating acute stroke patients with a potential LVO.

The trial in which he is a coinvestigator, called ENDOLOW, is testing the hypothesis that outcomes will be better if acute stroke patients with a LVO and a low baseline NIHSS score (< 5) are treated with immediate thrombectomy rather than medical management. If this hypothesis is confirmed in the randomized ENDOLOW, it will provide an evidence basis for an approach already being practiced at some centers.

"There should be a very low threshold for CTA," said Dr. Jabbour in an interview. This imaging "takes less than 2 minutes and it can provide the basis for a life-saving endovascular thrombectomy if a LVO is found."

It is already well known that LVO is not restricted only to patients with an elevated NIHSS score, he said.

For determining whether to order a CTA, "I do not agree with NIHSS score of 6 or above. There is no absolute number below which risk of missing a LVO is eliminated," Dr. Jabbour said. He also argued against relying on NIHSS score without considering other clinical features, particularly cortical signs, which should raise suspicion of a LVO regardless of NIHSS score.

One problem is that NIHSS scores are not static. Decompensation can be rapid with the NIHSS score quickly climbing. When this happens, the delay in treatment might lead to a preventable adverse outcome.

"There is a change in the paradigm now that we have more evidence of a benefit from aggressive treatment in the right candidates," according to Dr. Jabbour, referring to the recently published SELECT2 trial. In that trial, on which Dr. Jabbour served as a coauthor, patients with LVO and large territory infarct were randomized to thrombectomy or medical care within 24 hours of a stroke. It was stopped early for efficacy because of the increased functional independence (20% vs. 7%) in the surgical intervention group.

If the ongoing trials establish better criteria for ruling in or out the presence of LVO in patients with acute stroke, Dr. Jabbour predicted that guidelines will be written to standardize practice.(What useless crapola! GUIDELINES NOT PROTOCOLS!)

Dr. Sevilis reports no potential conflicts of interest. Dr. Jabbour has financial relationships with Cerenovus, Medtronic, and Microvention.

This article originally appeared on MDedge.com, part of the Medscape Professional Network.