Regional Effects of Modifiable Risk Factors for Acute Stroke

Study Questions:

What is the impact of modifiable risk factors for stroke on stroke subtypes, in different world regions, and among different demographic groups?


INTERSTROKE is an international case-control study that took place between January 2007 and August 2015 at 142 centers in 32 countries. Cases were patients with their first stroke within 5 days of symptom onset and 72 hours of hospitalization. Nonstroke controls were identified from hospitals or the community. There was no standard approach to identifying controls. Controls were matched for sex and age. Risk factors were identified using structured questionnaires and standard examinations. The risk factors used in this study included: abdominal obesity (waist-to-hip ratio), hypertension, physical activity, apolipoprotein (Apo)B/ApoA1 ratio, diabetes, diet, smoking, alcohol intake, cardiac disease (including atrial fibrillation), and psychosocial factors. Logistic regression was used to identify associations between risk factors and stroke type (all stroke, ischemic, and intracerebral hemorrhage [ICH]) by region, ethnicity, and age. Population-attributable risk (PAR) was calculated for each risk factor and the combination of risk factors.


There were 13,447 stroke cases (73% ischemic) and 13,472 controls. All stroke was associated with hypertension, smoking, physical activity, ApoB/ApoA1 ratio, diet, abdominal obesity, psychosocial factors, current smoking, cardiac disease, alcohol intake, and diabetes. The same associations were seen with ischemic stroke. When ICH was examined, there were associations with hypertension, physical activity, diet, abdominal obesity, psychosocial factors, cardiac disease, and alcohol intake. While hypertension was associated with both stroke types, the association was stronger for ICH.

When all 10 risk factors were combined, the PAR was 90.7% for all stroke, 91.5% for ischemic stroke, and 87.1% for ICH. There was regional variation in the all-stroke PAR for all risk factors, ranging from 97.4% in Southeast Asia to 82.7% in Africa. The composite all-stroke PAR was the same for men and women, 90.6%. When specific risk factors were examined for their association with all stroke, hypertension had the largest PAR (45.2% in men and 52.3% in women), whereas diabetes had the lowest (3.7% in men and 4.1% in women). There was variation in the all-stroke PAR between men and women for current smoking (men 16.6%, women 5.3%), abdominal obesity (men 12.7%, women 25.8%), and alcohol intake (men 10.0%, women -0.7%).


The authors concluded that targeting these modifiable risk factors could reduce the incidence of stroke worldwide.


This work builds on the first INTERSTROKE study by increasing the number of subjects, expanding the analysis to developing countries, and including stroke type. While there are some regional and sex variations in the impact of risk factors on stroke, overall, there is tremendous commonality across regions to suggest that addressing these risk factors will have a beneficial impact across the globe. Policymakers can identify the interventions that have the highest PAR in their region and then tailor interventions so they have the highest impact. Limitations of this study relate to its case-control design. For instance, acute hypertension could be the result of a stroke in evolution and could be misclassified as a chronic risk factor. Additionally, the selection of controls was not standardized, which could introduce bias. That said, the size and global reach of this study are significant strengths.

< Back to Listings