Next, I'm going to ask Dr. Dornbos to come up, please, and present his paper, Predicting Venous Thrombobolic Complication Following Neurological Surgery Procedures. Thank you. Yes, I'm David Dornbos. I'm a six-year resident at The Ohio State University. I'll be discussing a project that kind of actually spawned off of a QI initiative that we started following a pretty significant uptick in our DBT and PE rates, and just us trying to start to figure out some of the risk factors and some of the underlying reasons why that happened. Just to start off with, venous thrombobolic events happen approximately 900,000 times per year in the United States, and they carry a significant mortality risk of 33% within 30 days of diagnosis. This accounts for about 10% of mortality amongst all hospitalized patients, and this is due to Virchow's triad, which is pretty well known, but it's this venous stasis, vascular injury, and hypercoagulability, which we see frequently in neurosurgery patients. Within neurosurgery specifically, some risk factors that have already been defined as being significantly correlative of DBTs, intracranial versus spine surgery carries a higher risk, so do malignant versus benign tumors. Overall length of surgery, the presence of leg weakness, whether preoperative or postoperative, and the elderly population all carry increased DBT risk. The overall rates vary quite a bit across neurosurgery studies. DBTs can be seen in almost up to a third of patients in certain studies, and PEs are seen usually in around usually about 2% to 3%. Craniotomy for brain tumor presents the highest risk surgery subtype, with 1.4% of patients developing a DBT within six weeks. Prophylactic strategies are pretty standard for the most part. Obviously mechanical prophylaxis is used in pretty much everybody. The addition of chemical prophylaxis has been shown to significantly reduce the risk of DBTs from 34 down to about 6% in one pretty solid study. Oh, sorry. The safety profile obviously needs to be weighed, and that's the biggest thing in this study that we were looking at. But numerous prospective studies have shown no significant increase in ICH with prophylactic heparin. In spontaneous ICH, use of heparin starting on hospital day two doesn't increase risk of hemorrhagic complications, but does significantly decrease your risk of DBT complications. Due to this, the American College of Chest Physicians puts out antithrombotic guidelines and puts them on the front lines of PE treatment. For neurosurgery specifically, they recommend mechanical prophylaxis with heparin. Whether it's low molecular weight or unfractionated, they kind of left up to the dealer. A recent study actually looked at compliance across numerous subspecialties, and neurosurgery was one of the least compliant in terms of following the chest physician's guidelines. This was largely thought to be due to the potential risk of hemorrhagic complications and the severe nature when they occur in our patient population. To better characterize the thrombotic risk in neurosurgery patients, we turned to the Caprini risk assessment model. It's a model that was initially developed by a single general surgeon based essentially on just his gestalt. He did not develop this model in any statistical way. He took any risk factor that potentially would lead to development of a DBT or PE and assigned it either one, two, three, or five points. The one category is obviously pretty large. It includes some very obvious things, acute MI, presence of sepsis, current presence of swollen legs. It also includes some very kind of things we don't see all the time, but recent pregnancy, history of spontaneous abortions, and things of that nature. Things that would get two points would be something as simple as a central line or as severe as the presence of a malignancy. The most significant three-point risk factor that we encountered was a history of a DBT or PE. That category also includes a lot of kind of familial disorders that just predispose you to having prothrombotic complications. And then the five-point category, the big ones, are often seen in nursery patients. So stroke within a month, polytrauma, and also acute spinal cord injury. Based on the sum total from these checkboxes, he developed kind of a stratification of how he felt that anticoagulation and simple STDs should be passed out. One important thing to note here is the relatively low sum totals of his risk category. So his highest risk only had to have a caprinia of five or more. So any stroke patient, anybody with acute spinal cord injury, immediately falls into the highest risk category. His model was not actually validated for well over a decade after it was first passed out. The first validation of it actually showed that it did stratify patient risk pretty well across those four categories that he had developed. That being said, the lion's share of the DVTs was found in the highest risk category. When that was further subdivided out into kind of subcategories in the highest risk population, it again showed a pretty good stratification of patient risk for DVT and PE development. Again, the greater than nine category carries the large majority of DVTs and PEs. So to validate this in a neurosurgery population, we retrospectively looked through an entire year's worth of surgical patients that came in for any neurosurgical procedure, whether it was endovascular, functional, craniotomy, or anything along the spectrum of spine. We collected baseline demographics and the various components of the Caprini score. Our primary outcome was DVT or PE development within 30 days. What we found was of the 180 patients that developed a DVT or a PE, their average Caprini score was 12, whereas in the patients that were DVT, PE negative, their average score was less than nine. This was significant, and the overall incidence of DVTs, PEs in our population was 6.4%. Looking at the various predictors, at least on univariate analysis and the different components of the Caprini score, the Caprini score was predictive of developing a DVT, PE in our neurosurgery population. The various components of that score, BMI, acute MI, the presence of swollen legs, prolonged bed rest, which had to be greater than three days, sepsis and pneumonia, malignancy, a history of DVT, PE, recent stroke, and length of surgery greater than three hours, but not less than three hours, were all predictive of developing a DVT or PE. Now a lot of these obviously confound each other, bed rest and stroke due to some degree, sepsis and pneumonia due to some degree. So to start to control for some of these confounders, we did a multivariate regression analysis and found these to be the independent predictors of DVT and PE development. Significant ones were BMI, the presence of swollen legs, which is pretty self-explanatory, prolonged bed rest, pneumonia, malignancy, and an acute MI. Interestingly, a history of DVT, PE did not actually bear out when controlling for confounders. To assess the predictability of the Caprini score to predict the development of a DVT or PE in the perioperative period, we found the test to be okay, not great. The area under the curve was 0.75, which interestingly is exactly the same as it's been found in the general surgery population through some validation studies that they've done as well. Using the Caprini score of five as the very high risk, which is the initial recommendation, gave a good sensitivity, which you'd expect because you're going to capture pretty much all of your DVTs and PEs. But the specificity definitely leaves something to be desired. Shifting this up to a Caprini score of nine, the sensitivity still holds decently at 89%, with an increased specificity of 60%, which is important when you also have to kind of worry about the potential serious sequelae of a hemorrhagic complication. We also just briefly looked at kind of stratifying this risk over Caprini's original subgroups. And essentially the initial subgroups are kind of useless for neurosurgery patients. The low to high risk only comprises about 8% of neurosurgery patients with zero DVTs or PEs in that group. When we actually subcategorized that very high risk group, you started to see a little bit better stratification of those patients. Whereas patients that have a Caprini score of greater than nine or equal to nine, while it encompasses a large portion of the population, it starts to stratify out that risk of DVT and PE development a bit more robustly. So in conclusion, there's a lot of different variables in that group, or in that initial Caprini score that don't really apply to a neurosurgery population. It definitely doesn't take into account cranial, spine, functional, the different types of subspecialty procedures. And the treatment recommendations are not all that helpful. But the risk assessment model does provide an adequate predictive model that moderately stratifies risk when subcategorizing that very high risk group. So thank you. Are you using it in your clinical setting? Not yet. We're actually in the process of developing a new risk assessment model for neurosurgery patients specifically, looking at the different factors that are independently predictive in our population, kind of reweighting them, and then doing some modeling of our own to hopefully use that in the future. Thank you very much, Mayor Apdekul. Thanks.

venous thromboembolic complications

neurological surgery procedures

mortality risk

prophylactic strategies

Caprini risk assessment model