Stroke Prediction With ABCD Scores
This is a secondary analysis of the Additional Neurological SYmptoms before Surgery of the Carotid Arteries – a Prospective study (ANSYSCAP). To date, the ANSYSCAP study was the largest single-center study of patients with 50–99% carotid stenosis with up-to-date medical prevention. In short, the ANSYSCAP study prospectively included 230 consecutive patients with symptomatic 50–99% carotid stenosis. The patients were included at the Umeå Stroke Center, Sweden, between 2007-08-01 and 2009-12-31. The sample size was derived from setting study start- and stop dates. Most patients (81%) were sent from 11 referring hospitals. We only included patients that were preliminarily eligible for carotid endarterectomy (CEA). This was defined as a patient who underwent a pre-operative evaluation aimed at CEA. Since this was a selected population, the CEA rate was high (80%). Reasons for not performing CEA were low predicted benefit (7%), too high perioperative risk due to co-morbidities or technical reasons (5%), CEA not meaningful (4%) or patient refusal (4%). Over the course of the study, CEA went from being most often scheduled to be performed earlier. Whenever possible, the delay to CEA was kept as short as possible. The observation period for recurrent stroke was the first 90 days after the presenting event in the primary analysis. For the majority of patients that underwent CEA within 90 days, the observation time was limited to up to the CEA. In order to reduce a possible bias caused by non-random censoring at CEA, the current analysis was limited to the first 14 days after the presenting event. The baseline characteristics of patients are shown in Table 1. In the current analysis, we included both patients with a presenting event lasting <24 hours (TIA and amaurosis fugax) and ≥24 hours (stroke and retinal artery occlusion) because our sample size was small.
In the original study, blood pressure was recorded at the time of the pre-operative evaluation. The first blood pressure should be used for the ABCD2 and ABCD3 scores. Therefore, we retrospectively amended our data with the first recorded blood pressure. This was done by a review of medical records. We recorded the patient's first documented blood pressure – most often at the first health care contact. The remaining variables of the ABCD2 and ABCD3 scores were collected in the original study.
We calculated all variables for the ABCD2 and ABCD3 scores as defined elsewhere: (A) Age <60 years 0 points and ≥60 years 1 point; (B) first recorded blood pressure <140/90 0 points and either ≥140 systolic or ≥90 diastolic 1 point; (C) clinical features of the presenting event with focal weakness 2 points, speech impairment without weakness 1 point, and other 0 points; (D1) duration of presenting event <10 minutes 0 points, 10–59 minutes 1 point, and ≥60 minutes 2 points; (D2) diabetes absent 0 points and present 1 point; and (D3) for ABCD3-scores only, dual events with ≥1 additional TIA or amaurosis fugax within 7 days of the presenting event 2 points and no such event 0 points. All patients with stroke or retinal artery occlusion as the presenting event were assigned 2 points for duration.
We used the same endpoint as in the original ANSYSCAP study: Recurrent ipsilateral ischemic stroke. Ipsilateral retinal artery occlusion was also included in the endpoint (and is henceforth included in "ipsilateral ischemic stroke"). We analyzed recurrent ipsilateral stroke within 14 days after the presenting event. If the patient underwent CEA within 14 days, we only analyzed events that occurred before the CEA (excluding all perioperative events). The presenting event was defined as the last ischemic cerebrovascular event (stroke, retinal artery occlusion, TIA, or amaurosis fugax) before the patient sought health care.
We analyzed the risk of the endpoint at 2 days, 7 days, and 14 days. We analyzed the risk of the endpoint for each parameter in the ABCD2 and ABCD3 scores separately and for different total ABCD2 and ABCD3 scores. We analyzed an ABCD2 score of 2, 3, and 4 points as cut-offs for high and low risk. We used the same cut-offs for the ABCD3 score as in a previous study: low (0–3 points), moderate (4–5 points), and high (6–9 points) risk.
The local Ethical Review Board found the study to be completely observational and therefore did not incur any ethical problems and did not require ethics approval.
We calculated the 14-day risk of recurrent ipsilateral ischemic stroke with Kaplan-Meier curves. We used CEA as a censor and, therefore, patients that underwent CEA only contributed risk-time before their CEA and all perioperative events were excluded. The risks at 2, 7, and 14 days were acquired from this survival analysis. We used the log rank test for differences between groups. A p-value <0.05 was defined as statistically significant. We used IBM SPSS 20.0 Statistical software for all calculations.
Methods
Patients
This is a secondary analysis of the Additional Neurological SYmptoms before Surgery of the Carotid Arteries – a Prospective study (ANSYSCAP). To date, the ANSYSCAP study was the largest single-center study of patients with 50–99% carotid stenosis with up-to-date medical prevention. In short, the ANSYSCAP study prospectively included 230 consecutive patients with symptomatic 50–99% carotid stenosis. The patients were included at the Umeå Stroke Center, Sweden, between 2007-08-01 and 2009-12-31. The sample size was derived from setting study start- and stop dates. Most patients (81%) were sent from 11 referring hospitals. We only included patients that were preliminarily eligible for carotid endarterectomy (CEA). This was defined as a patient who underwent a pre-operative evaluation aimed at CEA. Since this was a selected population, the CEA rate was high (80%). Reasons for not performing CEA were low predicted benefit (7%), too high perioperative risk due to co-morbidities or technical reasons (5%), CEA not meaningful (4%) or patient refusal (4%). Over the course of the study, CEA went from being most often scheduled to be performed earlier. Whenever possible, the delay to CEA was kept as short as possible. The observation period for recurrent stroke was the first 90 days after the presenting event in the primary analysis. For the majority of patients that underwent CEA within 90 days, the observation time was limited to up to the CEA. In order to reduce a possible bias caused by non-random censoring at CEA, the current analysis was limited to the first 14 days after the presenting event. The baseline characteristics of patients are shown in Table 1. In the current analysis, we included both patients with a presenting event lasting <24 hours (TIA and amaurosis fugax) and ≥24 hours (stroke and retinal artery occlusion) because our sample size was small.
Data Acquisition
In the original study, blood pressure was recorded at the time of the pre-operative evaluation. The first blood pressure should be used for the ABCD2 and ABCD3 scores. Therefore, we retrospectively amended our data with the first recorded blood pressure. This was done by a review of medical records. We recorded the patient's first documented blood pressure – most often at the first health care contact. The remaining variables of the ABCD2 and ABCD3 scores were collected in the original study.
Calculation of ABCD2 and ABCD3 Scores
We calculated all variables for the ABCD2 and ABCD3 scores as defined elsewhere: (A) Age <60 years 0 points and ≥60 years 1 point; (B) first recorded blood pressure <140/90 0 points and either ≥140 systolic or ≥90 diastolic 1 point; (C) clinical features of the presenting event with focal weakness 2 points, speech impairment without weakness 1 point, and other 0 points; (D1) duration of presenting event <10 minutes 0 points, 10–59 minutes 1 point, and ≥60 minutes 2 points; (D2) diabetes absent 0 points and present 1 point; and (D3) for ABCD3-scores only, dual events with ≥1 additional TIA or amaurosis fugax within 7 days of the presenting event 2 points and no such event 0 points. All patients with stroke or retinal artery occlusion as the presenting event were assigned 2 points for duration.
Endpoint & Analyses
We used the same endpoint as in the original ANSYSCAP study: Recurrent ipsilateral ischemic stroke. Ipsilateral retinal artery occlusion was also included in the endpoint (and is henceforth included in "ipsilateral ischemic stroke"). We analyzed recurrent ipsilateral stroke within 14 days after the presenting event. If the patient underwent CEA within 14 days, we only analyzed events that occurred before the CEA (excluding all perioperative events). The presenting event was defined as the last ischemic cerebrovascular event (stroke, retinal artery occlusion, TIA, or amaurosis fugax) before the patient sought health care.
We analyzed the risk of the endpoint at 2 days, 7 days, and 14 days. We analyzed the risk of the endpoint for each parameter in the ABCD2 and ABCD3 scores separately and for different total ABCD2 and ABCD3 scores. We analyzed an ABCD2 score of 2, 3, and 4 points as cut-offs for high and low risk. We used the same cut-offs for the ABCD3 score as in a previous study: low (0–3 points), moderate (4–5 points), and high (6–9 points) risk.
Ethics
The local Ethical Review Board found the study to be completely observational and therefore did not incur any ethical problems and did not require ethics approval.
Statistics
We calculated the 14-day risk of recurrent ipsilateral ischemic stroke with Kaplan-Meier curves. We used CEA as a censor and, therefore, patients that underwent CEA only contributed risk-time before their CEA and all perioperative events were excluded. The risks at 2, 7, and 14 days were acquired from this survival analysis. We used the log rank test for differences between groups. A p-value <0.05 was defined as statistically significant. We used IBM SPSS 20.0 Statistical software for all calculations.
Source...