Predicting risk of recurrent stroke based on changes in risk factors: the VISP study

Background: Recurrent stroke carries a greater risk of disabling or fatal aftereffects (1,2,3) and poor quality of life compared to first-ever stroke 1. Secondary stroke prevention has been successful in controlled trials but less so in the general population 4. Improvements in secondary stroke prevention may be realized through optimization of the management of risk factors following a stroke (5,6,7). The objective of this study was to examine whether recurrent stroke risk can be predicted from risk factor profile changes and to determine whether improvements in care guideline adherence resulted in a reduction of recurrent stroke risk. Methods: We analyzed data from the VISP trial. VISP's methods and results have been described previously 8, 9. Demographic, clinical and laboratory data were collected at baseline and at six-month intervals. Participants on heparin (n= 36), from the Glasgow clinic (n= 46), or with irremediable missing data (n=274) were excluded, leaving 3324 participants with 14142 observations for use in the first analysis. For the second analysis, the dataset was further restricted to those participants with a baseline and a twelve-month follow-up visit (N=2750). Results: We developed and calibrated two recurrent stroke risk models using repeat measurements of risk factors. Whites with untreated Type 2 diabetes or high homocysteine levels, nonwhites with strokes prior to the VISP eligibility stroke, and individuals with any dependency or any disability had an increased recurrent stroke risk; aspirin use showed a protective effect in nonwhites. Care guideline adherence post-index stroke was heterogeneous, with changes in adherence nearly exactly balanced between adherence and non-adherence. Participants not using aspirin were at a significantly higher recurrent stroke risk if after their stroke they did not use anti-diabetic medications (within the first six months ) or started using diabetic medications or maintained a BMI of 25 kg/m2 or greater during the first twelve months. Conclusion: We have established that it is possible to predict recurrent stroke risk using repeat measurements of risk factors and that lack of adherence to medication adherence guidelines results in a higher risk of recurrent stroke. Psychoneurological factors appear to influence recurrent stroke risk.