PLUMBER Study (Prevalence of Large Vessel Occlusion Stroke in Mecklenburg County Emergency Response).
Authors: Dozois A, Hampton L, Kingston CW, et al.
Published in: Stroke, 2017 Dec; 48(12): 3,397–9.
This month we discuss a very interesting manuscript that examined the prevalence of large-vessel occlusion strokes in a single North Carolina county. Large-vessel occlusion (LVO) strokes are a hot topic in our field. The American Heart Association/American Stroke Association (AHA/ASA) have recommended endovascular thrombectomy for LVO strokes. Rather than administering tPA, the treatment team removes the clot.
If you haven’t watched any of the videos of this procedure online, you are missing out on an amazing sight. There is practically immediate revascularization. The procedure is being called a breakthrough. Research on this topic has consistently shown that using the word breakthrough is not an overstatement. This will likely change our field in much the way initiating transport straight to a cath lab has for STEMI patients.
While this is a new and exciting therapy that will improve outcomes for our acute ischemic stroke patients, its success relies, in part, on our assessment and the time since last known well. Endovascular thrombectomy must be initiated within six hours of symptom onset. We also have to correctly identify our stroke patients. Research has shown that many of the stroke screens we use in the field are not that great at helping identify strokes. Further, if we’re going to consider transporting patients directly to stroke centers with endovascular capability, it is important to understand how large of a patient population this will impact.
The PLUMBER Study
Hence discussion of the PLUMBER study in this month’s Trip Report. PLUMBER is another acronym. We’ve discussed before how researchers like acronyms as much as EMS professionals. PLUMBER stands, with some license, for Prevalence of Large-vessel occlUsion strokes in MecklenBurg County Emergency Response. The authors broke the unwritten acronym rule of using only the first letter of each word, but it’s a great title!
The authors’ objective was to determine the prevalence of LVO strokes in Mecklenburg County, N.C. To do this they conducted a cross-sectional study of patients transported by Mecklenburg County EMS from December 2015 to October 2016. A cross-sectional study is similar to taking a picture: It will tell you about the moment the picture was taken but will not give you any information about what came before or after. Research reporting survey results is another example of a cross-sectional study.
While the study period here was less than a year, the authors performed a power analysis to determine how many patients they would need to appropriately address their study outcome. It is important to start research with a power analysis. There’s nothing worse than collecting a bunch of data and then finding out you didn’t collect enough to answer your question. However, to do a power analysis, you have to make some assumptions up front. The authors assumed the prevalence of LVO strokes in their county would be 5%. They input this number into the power calculation formula and determined they needed data for 2,000 patients to appropriately address their study outcome. Since Mecklenburg is a large county, it took less than a year to transport 2,000 stroke patients.
Patients were included in the study if they had a dispatch complaint of, or the EMS provider’s primary impression was consistent with, stroke or TIA. One of the most impressive parts of this study is that the authors were able to obtain and link EMS and hospital records from one EMS agency and seven hospitals. That is a lot harder than it sounds, but it’s a testament to the commitment to research and improving patient care in Mecklenburg County.
Study leaders received approval from multiple institutional review boards (an IRB is basically a board of individuals that reviews your research idea before it’s conducted to make sure it’s ethical and won’t place undue risk on research subjects) to use patient identifiers (name, date of birth, social security number, etc.) for data linkage. This is important because using identifiers to link reduces the likelihood of linking records that aren’t a true match.
Overall there were 2,402 patients included: 1,165 had a dispatch complaint of stroke or TIA, 467 had an EMS provider’s primary impression of stroke or TIA, and 770 had both. Among all patients in the study, only 4.87% were found to have had an actual LVO stroke (n=117; 95% CI, 4.05%–5.81%). Of the 2,402 patients, 67.7% ended up having a non-stroke related diagnosis, 8% had a TIA, and over 4% had a hemorrhage. There were 485 patients who had an ischemic stroke diagnosis. Of those, 45 did not have a CTA or MRI performed, so there were really only 440 patients who had the possibility of LVO diagnosis. Of those 117 (26.6%; 95% CI, 22.6%–31.0%) actually had an LVO stroke.
Overall this is a very interesting and well-conducted study. The authors linked EMS data over multiple hospitals and performed appropriate statistical assessments. This study has limitations, just like every study; only using data from one county surely limits the generalizability of these results. However, we are in the very early stages of sharing and linking data. Currently there is likely no way to perform this study nationwide or even statewide.
The authors’ conclusion focused on the low prevalence of LVO strokes overall, and this needs to be taken into account when developing triage plans to route patients directly to centers with endovascular thrombectomy capabilities. This is certainly an appropriate conclusion given their finding that 68% of patients were incorrectly identified prehospitally as stroke patients. This study surely adds to the knowledge on LVO strokes, but it also adds to the growing literature suggesting we need a new and better way to identify strokes in the field.
As always, I hope you have an opportunity to read the manuscript. There are some results we didn’t have space to discuss here regarding the location of the occlusions and some supplemental tables that help put the results into context.
Thanks for your continued interest in EMS research. I hope to review one of your studies in a Trip Report soon!
Antonio R. Fernandez, PhD, NRP, FAHA, is research director at the EMS Performance Improvement Center and an assistant professor in the Department of Emergency Medicine at the University of North Carolina–Chapel Hill. He has been a nationally certified paramedic since 2005 and completed the EMS Research Fellowship at the National Registry of Emergency Medical Technicians.