Feasibility of the mNIHSS in Surgical Patients

Overview

This prospective observational study will be assessing whether it is practical and feasible to use a stroke screening tool (the modified NHISS, mNIHSS) in an adult surgical population. Perioperative stroke is an uncommon but devastating complication of surgery and previous research suggests that stroke symptoms may be missed in this population. This stroke screening tool has been used extensively in non-surgical patients but has not be used in non-cardiac surgical patients. Surgical patients have additional factors that may affect their ability to complete this assessment such as pain medications and sedatives, pain, and use of epidurals and nerve blocks. This study will involve a total of five testing sessions. The first will occur at the time of consent to screen for cognitive impairment via the Montreal Cognitive Assessment (MoCA), as part of our exclusion criteria. Then over the course of four subsequent visits, participants will be assessed on the mNIHSS to track for any changes before and after their procedure, as well as evaluate the real-life usability of the mNIHSS as a potential tool to screen for stroke in the surgical population.

Purpose \& Justification: Perioperative stroke is a potentially devastating complication that is currently poorly characterized with limited clinical tools available to detect and prevent its occurrence. To date only two stroke assessment scales have been applied to the perioperative setting, though none have been tested in non-cardiac, non-neurologic surgeries. One of those assessments, the National Institutes of Health Stroke Scale (NIHSS), is lengthy and may be impractical in the surgical setting. This is further complicated by several factors potentially existing in surgical patients that may alter the assessment, such as pain and drowsiness from medications. However, there is a modified simpler version of the NIHSS (mNIHSS) that has been shown to have improved reliability than the full version and reduced time requirement, thus making it preferable for perioperative stroke screening. To overcome the current gaps in knowledge and screening for risk of perioperative stroke, this prospective observational study will be assessing the feasibility of applying the mNIHSS in the surgical setting. Primary Objective: To determine the change in mNIHSS score from baseline to immediately after surgery in the absence of diagnosed stroke in patients after surgery. Secondary Objectives: 1. To determine the components of the mNIHSS most likely to change following surgery; 2. To determine the feasibility of administering the mNIHSS in surgical patients (e.g. acceptability, resources, length of time to complete, completeness). 3. To determine the relationship between baseline MoCA score and ability to complete the mNIHSS screening after surgery Research design: This is a prospective observational study involving five testing sessions. A cognitive assessment (MoCA) will be conducted at the time of consent to screen for any cognitive impairment. There will be 4 additional visits during which the mNIHSS will be administered to surgical patients pre and post-operatively. Statistical Analysis: Sample Size Calculation There is no specific data available on the application of the mNIHSS in our study population. We estimate that we will require a sample size of 21 patients, assuming a mean difference of 2 (standard deviation 3) points between pre- and post-operatively, an alpha error of 0.05 and power of 80%. Anticipating a 20% drop out rate, we will recruit 25 patients. Statistical Analysis We will describe the patient population using mean (SD), median (IQR) and percentage as appropriate. The application of the mNIHSS will be described using percent completed for each question, and the time taken to complete the test at each time point (mean, SD). The change in mNIHSS on postoperative days 0, 1 and 2 will be compared to baseline using a Wilcoxon Signed Ranks test for matched pairs. All tests will be two-sided with a p-value less than 0.05 considered statistically significant. All statistical analysis will be done using STATA 12.1 (StataCorp, Texas, USA).