Low ankle-brachial index and cognitive function after stroke—the PROSpective with Incident Stroke Berlin (PROSCIS-B)

Introduction Low ankle-brachial index (ABI) ≤0. 9 is a marker for generalized atherosclerosis and a risk factor for cognitive decline in the general population. Objective To evaluate the impact of ABI ≤0.9 on cognitive function up to 3 years after first-ever ischemic stroke. Methods Data was used from the “PROspective Cohort with Incident Stroke-Berlin” (PROSCIS-B; NCT01363856). ABI was measured at baseline and categorized into normal (1.4–0.9) vs. low (≤0.9). Cognitive function was assessed with the Montreal Cognitive Assessment (MoCA) and the Mini-Mental-State-Examination (MMSE) at baseline and with the Telephone Interview for Cognitive Status-modified (TICS-m) at 1–3 years of follow-up. We performed confounder adjusted generalized linear models (GLM) to calculate relative risks (RR) for cognitive impairment at baseline (MMSE≤26; MoCA≤25) and linear mixed models (LMM) to estimate the impact of low ABI on TICS-m over time. Results We included 325 patients [mean age: 66 (SD = 13); 38% female, median NIHSS = 2 (IQR = 1–4), ABI≤0.9: 59 (18%)]. Patients with low ABI were at increased risk of cognitive impairment at baseline (adjusted RR for MoCA≤25 = 1.98; 95%-CI:1.24 to 3.16). TICS-m scores were consistently lower over time in patients with low ABI (adjusted ß = −1.96; 95%-CI:−3.55 to −0.37). Independent of ABI, cognitive function did not decline over time (adjusted ß:0.29; 95%-CI:−0.06 to 0.64). Conclusion In patients with mild to moderate first-ever ischemic stroke, low ABI is associated with reduced cognitive function over a 3-year follow-up. Study Registration https://clinicaltrials.gov; Unique identifier: NCT01363856.

Introduction: Low ankle-brachial index (ABI) ≤ . is a marker for generalized atherosclerosis and a risk factor for cognitive decline in the general population.
Objective: To evaluate the impact of ABI≤ . on cognitive function up to years after first-ever ischemic stroke. Conclusion: In patients with mild to moderate first-ever ischemic stroke, low ABI is associated with reduced cognitive function over a -year follow-up.

Introduction
The ankle-brachial index (ABI) is a reliable, non-invasive, and inexpensive screening tool for peripheral artery disease (PAD) (1, 2). Low ABI ≤ 0.9 is a sensitive marker for generalized atherosclerosis and independently associated with cognitive impairment and decline in the general population (3,4).
However, observational studies investigating the impact of low ABI on cognition after stroke are limited, particularly regarding long-term cognitive outcome (5). Stroke occurrence in itself is associated with ensuing cognitive impairment and decline (6). Between 20 and 80% of stroke survivors suffer from cognitive deficits, whereby the severity and recurrence of stroke are related to subsequent cognitive decline (6,7). Low ABI could be a useful marker to detect cognitive impairment as a stroke-related disability and monitor secondary prevention therapies. A cross-sectional study of 103 hospitalized patients with acute lacunar stroke found that low ABI is associated with cognitive impairment in the acute stroke phase using the Montreal Cognitive Assessment (MoCA; cutoff score at ≤23) (5). Nonetheless, data on low ABI and cognitive function over time after ischemic stroke are lacking.

Aim
We aimed to investigate whether low ABI ≤ 0.9 is associated with reduced cognitive function at baseline and subsequent decline over a three-year follow-up after first-ever ischemic stroke.

Study design and population
This study is part of the Prospective Cohort with Incident Stroke Berlin (PROSICS-B), an observational, hospitalbased prospective cohort study, described in detail elsewhere (8). Between January 2010 and May 2013 stroke patients were recruited at three stroke units of the Charité -Universitätsmedizin Berlin within 7 days after stroke onset (see flow chart in Supplementary Figure 1). Inclusion criteria were patients with (1) first-ever ischemic stroke, primary intracranial hemorrhage, or cerebral venous sinus thrombosis, (2) aged 18 and older, (3) written informed consent by patient or legal guardian prior to study participation. Exclusion criteria were (1) prior stroke (definition according to WHO criteria), (2) patients with brain tumor or brain metastasis, (3) participation in an interventional trial.
Patients were followed up annually with telephone-based interviews for up to 3 years. Patients who had suffered a prior stroke, presented brain tumors, brain metastases or participated in an intervention study were excluded. Only patients with mild to moderate ischemic stroke events, defined by a score of < 16 in the National Institutes of Health Stroke Scale (NIHSS), were included in the analysis. The assessment of ABI and MoCA were introduced to the study as an amendment in January 2011, ∼1 year after patient enrollment had started according to the requirements of the ethical committee.

Patient characteristics
Baseline characteristics were assessed upon admission, including sociodemographic characteristics (e.g., age, sex and education, living situation and lifestyle habits), laboratory blood measures, stroke severity, and cardiovascular risk factors (BMI, current smoking and alcohol consumption, history of hypertension, diabetes mellitus and coronary heart disease).

ABI measurement
ABI was measured using a manual cuff and a portable Doppler ultrasound device. Participants were assessed in supine position following a 5-min resting period. The systolic blood pressure was determined for the brachial arteries, posterior tibial and dorsal pedis arteries bilaterally. Measurements were assessed at all three stroke centers following a pre-established standard operating procedure.
In accordance with the guidelines issued by the American Heart Association, ABI was calculated on each side by dividing the higher value of the posterior tibial or dorsal pedis artery systolic blood pressure by the higher systolic blood pressure of both brachial arteries (9). We used the lower ABI value. Low ABI was defined as ≤ 0.9. Participants with abnormally high ABI ≥ 1.4, indicative of non-compressible blood vessel calcification, were excluded (9).

Statistical methods
Generalized linear models (GLM) were used to calculate crude and confounder adjusted relative risks (RR) for cognitive impairment at baseline (MoCA ≤ 25; MMSE ≤ 26) in patients with low versus normal ABI and corresponding 95% confidence intervals (CI). We performed linear mixed models (LMM) to calculate effect sizes (ß) and corresponding 95% CI for the crude and confounder adjusted association between low vs. normal ABI and cognitive function over a three-year follow-up, using the annual TICS-m score.
To explore whether the severity of low ABI had an effect on cognitive function in the sense of a dose-response relationship, we further categorized low ABI into mildly low (0.75-0.9) and moderately to severely low ABI (<0.75) and recalculated our main analyses with these subgroups compared to normal ABI.
Multiple models were used to control for possible sources of confounding. Potential confounding factors were selected a priori according to their presumable impact on both the exposure (ABI) and the outcome (cognitive function). Model 1 adjusted for the sociodemographic variables age (continuous), sex and years of education received (≤10 years of schooling, > 10 years of schooling). Model 2 additionally adjusted for the cardiovascular risk factors BMI (kg/m 2 ), current smoking (yes/no), history of diabetes mellitus (yes/no) and hypercholesterolemia (yes/no), and history of coronary heart disease (yes/no) and atrial fibrillation (yes/no). Model 3 moreover adjusted for stroke severity as defined by the NIHSS (in 2 categories: 0-4 and 5-15).
We conducted different sensitivity analyses, described in Methods I of the Supplementary material.
Data were prepared using IBM Statistics for Windows, version 25 (IBM Corp, Armonk, NY). All descriptive and statistical analyses were calculated using Stata version 14.1 (Stata Corp, College Station, TX). Data visualizations were performed in the R project (R 4.0.0) using the ggplot2 package.

Ethics approval
Patients or their legal guardians gave written informed consent prior to study participation. The study was approved by the Charité -Universitätsmedizin Berlin ethics committee (EA1/218/09) and was conducted according to the declaration of Helsinki.

Study population
This analysis included 325 patients with mild to moderate ischemic stroke (NIHSS score < 16) and data on ABI (for detailed information on patient inclusion and exclusion see the flow chart in Supplemental Figure 1).

Cognitive function at baseline
Patients with low ABI had a higher risk of baseline cognitive impairment than patients with normal ABI (MoCA adjusted RR in

Discussion
The main finding of this study is that low ABI (≤0.9), measured after an index stroke, is associated with reduced cognitive function, both at baseline and over a 3-year followup compared to normal ABI. This effect is independent of age, sex, educational level and several cardiovascular risk factors and most pronounced when comparing participants .
with moderately to severely low ABI (< 0.75) to those with normal ABI. Our findings expand on recent observations that low ABI is linked to cognitive impairment immediately post-stroke by demonstrating detrimental effects on cognitive function also in the long-term (5). Our results support atherosclerosis as an underlying pathomechanism for cognitive disfunction and as a contributing factor in the development of dementia (e.g., Alzheimer's disease) (14). These findings expand the use of ABI as a sensitive screening mechanism, controlling for the progression of atherosclerotic blood vessel calcification, to a biomarker for cognitive dysfunction after stroke. Research has shown that statin use helps reduce the risk of post-stroke cognitive impairment, underlining the impact of atherosclerosis as a modifiable risk factor on unfavorable stroke outcomes (15). While low ABI as a surrogate marker can be improved through protective interventions, further research is needed to examine whether modifying ABI also modifies the risk for future cognitive impairment (16).

Strengths and limitations
Strengths of the study include its prospective design with a homogenous cohort, long follow-up period of 3 years with annual screenings and the assessment of cognitive function both at baseline and during follow-up. However, several limitations need to be addressed. First, we only included patients with mild to moderate ischemic stroke. This impedes the generalizability of our results to severe stroke patients and patients with other than ischemic stroke subtypes. However, patients who suffered severe stroke events are more likely to have low ABI at baseline and suffer from more severe cognitive impairment over time (7,17). Secondly, predominantly male patients (62.3%) and a relatively high proportion of "strokes of undetermined cause" (35%) were included into the study. This should be taken into account when interpreting the data. Furthermore, we only assessed ABI at baseline, not during annual follow-ups. Further studies are needed to examine whether ABI is a modifiable risk factor for cognitive impairment. Thirdly, while our results showed that cognitive function was reduced both at baseline and over time in participants with low vs. normal ABI, cognitive performance did not significantly decline over time in either group. This may indicate that post-stroke cognition over time is predicted by baseline cognitive function following stroke occurrence (18). Then again, participants had only suffered mild-to-moderate ischemic strokes and may thus have remained mostly cognitively unimpaired. Moreover, the serial application of the TICS-m might have led to practice effects, masking cognitive decline (19). In addition, the cognitive tests we administered assessed global cognition, so no conclusions can be drawn as to whether the observed associations are more prominent in certain cognitive domains.

Conclusions
In conclusion, our 3-year prospective cohort study including mild-to-moderate first-ever ischemic stroke patients showed that low ABI at stroke occurrence was associated with reduced cognitive function both at baseline and over a 3-year followup period. The associations were strongest in participants with moderately and severely low ABI.

Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics statement
The studies involving human participants were reviewed and approved by Ethics Committee Charité. The patients/participants provided their written informed consent to participate in this study.