In a recent study published in the JAMA Network Open, researchers analyzed individual participant data of 982 adult stroke survivors aged ≥18 years from four cohort studies conducted in the United States between 1971 and 2019.
The study aimed to assess whether their higher systolic blood pressure (SBP), glucose, and low-density lipoprotein (LDL) cholesterol levels post-stroke accelerated their cognitive decline.
Study: Associations Between Vascular Risk Factor Levels and Cognitive Decline Among Stroke Survivors. Image Credit: SewCreamStudio/Shutterstock.com
Background
Nearly 35% of stroke survivors develop dementia within one year due to persistent cognitive decline. In 53% of cases, stroke increases the risk of dementia; thus, deferring cognitive decline or associated dementia could improve outcomes in stroke survivors, including survival and quality of life.
Stroke-free adults are also at increased dementia risk due to high SBP, glucose, and LDL cholesterol. All these modifiable vascular risk factors (VRFs) are associated with accelerated cognitive decline.
However, it is unclear whether post-stroke levels of these modifiable VRFs accelerate cognitive decline, independent of pre-stroke VRF levels.
Experimental evidence of this association could help clinicians personalize treatment and help researchers identify interventional targets to preserve cognitive function post-stroke.
Previous studies in this research domain have so far fetched limited results owing to a lack of longitudinal cognitive and VRF measurements pre- and post-hospitalization and small sample sizes.
About the study
In the present study, 50 participants in each cohort suffered from physician-adjudicated ischemic or hemorrhagic stroke and, subsequently, provided objective measures of BP and cognition, i.e., at least one measurement before and after stroke, taken using standard protocols and equipment.
First, the team made cognitive function assessments like global cognition, executive, and memory function using standard tests over the phone.
Next, they cocalibrated all available cognitive test items, shared and unique, across cohorts into three domains representing memory, global cognition, and executive function using confirmatory factor analysis and item response theory methods.
The team summarized all measurements of interest as the time-dependent aggregate averages at or prior to each poststroke cognitive assessment.
The primary and secondary study outcomes were changes in global cognition and, executive function & memory, respectively. The study covariates varied with cohort measures; for instance, the team selected age as a covariate at the stroke time.
Likewise, the team measured poststroke depression using the Center for Epidemiologic Studies Depression Scale and summarized them as time-independent arithmetic averages.
Further, the team used regression models to estimate cognitive factor scores, having the same meaning across cohorts, and set them to a t-score metric (mean standard deviation {SD}). A difference of one point represented a 0.1-SD difference in the cognition distribution across all four cohorts. Notably, higher cognitive scores indicated better performance.
Three linear mixed-effects models, M1a, M1b, and M1c, helped the researchers estimate the relationships of poststroke SBP, LDL cholesterol, and glucose levels with cognitive outcomes separately, whereas they used a model M2 to estimate the combined associations of all three with cognitive decline.
In this way, researchers performed a complete case analysis alongside within-cohort analyses and estimated longitudinal changes in each continuous cognitive outcome. Finally, the researchers determined the robustness of their findings via five sensitivity analyses.
The study analyses lasted almost two years, beginning in August 2021, it culminated in March 2023.
Results and conclusion
The authors identified 1,120 eligible dementia-free individuals with incident stroke, of which only 982 had available covariate data. The median age of these 982 people at incident stroke was 74.6 years, and 48.9% were females.
Of aggregate average poststroke SBP, LDL cholesterol, and glucose levels, only glucose levels showed a significant association with a faster decline in global cognition, i.e., −0.04 points/year faster for each 10mg/dL increase after adjusting for aggregate average poststroke SBP and LDL cholesterol levels.
The authors found no evidence of any association of aggregate average poststroke SBP and LDL cholesterol levels with any cognitive outcomes, including an executive function or memory and global cognition.
The authors quoted multiple possible explanations for these observations. First, these measures are unrelated to any cognitive outcome evaluated in this study. Secondly, the older study sample might have resulted in a null finding.
Evidence accumulated by previous studies for older people is unclear, while studies have shown a strong association of high LDL cholesterol levels and SBP with cognitive decline in stroke-free adults in their middle age.
Another possible explanation is that the executive function and memory measures fail to detect the poststroke glucose–cognitive decline associations. Nonetheless, the authors found strong evidence of an association between female sex and faster poststroke cognitive decline.
The authors also analyzed a subset of 798 individuals whose apolipoprotein E4 (APOE4) data was available.
After adjusting for APOE4, they observed that higher cumulative mean poststroke glucose level was still associated with a faster decline in global cognition.
This result extends prior evidence that a higher aggregate average glucose level after stroke-related hospitalization was independently associated with faster global cognitive decline, irrespective of pre-stroke diabetes status.
Also, APOE4 did not modify the effect of poststroke glucose levels on global cognition.
In this cohort study, higher aggregate average poststroke glucose levels resulted in a rapid decrease in global cognition, but not memory or executive function deficits, when accounting for poststroke LDL cholesterol and SBP levels.
