The Impact of Vascular Risk Factors on Cerebral Amyloid Angiopathy: A Cohort Study in Hereditary Cerebral Amyloid Angiopathy and a Systemic Review in Sporadic Cerebral Amyloid Angiopathy

Abstract

Introduction: Cerebral amyloid angiopathy (CAA) has a remarkably variable disease course, even in monogenetic hereditary forms. Our aim was to investigate the prevalence of vascular risk factors and their effect on disease onset and course in Dutch-type hereditary (D-)CAA and sporadic CAA.

Methods: We performed a cohort study in D-CAA to investigate the association between vascular risk factors (hypertension, hypercholesterolemia, smoking, and alcohol use) and age of intracerebral hemorrhage (ICH) onset and time of ICH recurrence with survival analyses. In addition, we performed a systematic review to assess the prevalence of vascular risk factors and their effect on clinical outcome in sporadic CAA. We searched PubMed, Embase, Web of Science, and Cochrane Library from 1987 to 2022 and included cohorts with ≥10 patients. We created forest plots, calculated pooled estimates, and reported variability (heterogeneity plus sampling variability) and risk of bias.

Results: We included 70 participants with D-CAA (47% women, mean age 53 years). Sixteen (23%) had hypertension, 15 (21%) had hypercholesterolemia, 45 (64%) were smokers, and 61 (87%) used alcohol. We found no clear effect of vascular risk factors on age of first ICH (log-rank test hypertension: p = 0.35, hypercholesterolemia: p = 0.41, smoking: p = 0.61, and alcohol use: p = 0.55) or time until ICH recurrence (log-rank test hypertension: p = 0.71, hypercholesterolemia: p = 0.20, and smoking: p = 0.71). We identified 25 out of 1,234 screened papers that assessed the prevalence of risk factors in CAA and 6 that reported clinical outcomes. The pooled prevalence estimates of hypertension was 62% (95% CI: 55-69%), diabetes was 17% (95% CI: 14-20%), dyslipidemia was 32% (95% CI: 23-41%), and tobacco use was 27% (95% CI: 18-36%). One study reported study diabetes and hypertension to be associated with a lower risk of recurrent ICH, whereas another study reported hypertension to be associated with an increased risk. All other studies showed no association between vascular risk factors and clinical outcome. High-quality studies focusing on vascular risk factors were lacking.

Conclusion: In patients with D-CAA and sporadic CAA, the prevalence of vascular risk factors is high. Although this suggests an opportunity for prevention, there is no clear association between these risk factors and CAA-related ICH onset and recurrence.

Keywords: Cerebral amyloid angiopathy; Intracerebral hemorrhage; Risk factors; Systematic review.

Fig. 1.

Survival analysis for risk factors combined and time to first ICH in years (n = 70). We categorized the participants into two groups: with presence of 0 or 1 risk factor and presence of 2 or more risk factors. Log-rank test: p = 0.76.

Fig. 2.

Survival analysis for risk factors and time until first ICH in years (n = 70). We explored the effect of the risk factors separately. Log-rank test hypertension (p = 0.35), hypercholesterolemia (p = 0.41), smoking (p = 0.61), and alcohol use (p = 0.55). Diabetes was excluded from this preliminary analysis due to the low prevalence in this cohort (n = 3 participants).

Fig. 3.

Survival analysis for risk factors combined and ICH recurrence in months (n = 34). We categorized the participants into two groups: with presence of 0 or 1 risk factor and presence of 2 or more risk factors. Log-rank test: p = 0.63.

Fig. 4.

Survival analysis for risk factors and ICH recurrence in months (n = 34). We preliminarily assessed the risk factors separately. Log-rank test hypertension (p = 0.71), hypercholesterolemia (p = 0.20), and smoking (p = 0.71). Diabetes and alcohol use were excluded from this preliminary analysis due to the low prevalence in this cohort (n = 3 participants with diabetes and n = 6 participants who did not use alcohol).

Fig. 5.

Study selection.

Fig. 6.

Forest plot with pooled estimate of hypertension in n = 25 studies.

Fig. 7.

Forest plot with pooled estimate of diabetes in n = 19 studies.

Fig. 8.

Forest plot with pooled estimate of dyslipidemia in n = 17 studies.

Fig. 9.

Forest plot with pooled estimate of tobacco use in n = 12 studies.

Fig. 10.

Risk of bias assessment.