Cysteine-Altering NOTCH3 Variants Are a Risk Factor for Stroke in the Elderly Population

Abstract

Background and purpose: Cysteine altering NOTCH3 variants, which have previously been exclusively associated with the rare hereditary small vessel disease cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy, have a population frequency of 1:300 worldwide. Using a large population database, and taking genotype as a starting point, we aimed to determine whether individuals harboring a NOTCH3 cysteine altering variant have a higher load of small vessel disease markers on brain magnetic resonance imaging than controls, as well as a higher risk of stroke and cognitive impairment.

Methods: A cross-sectional study using integrated clinical, neuroimaging, and whole-exome sequencing data of 92 456 participants from the Geisinger DiscovEHR initiative cohort. The case group consisted of individuals harboring a NOTCH3 cysteine altering variant (n=118). The control group consisted of randomly selected age- and sex-matched individuals who did not have any nonsynonymous variants in NOTCH3 (n=184). Medical records including brain magnetic resonance imagings were evaluated for clinical and neuroimaging findings associated with small vessel disease. Group comparisons were done using Fisher exact test and ordinal logistic regression models. Risk of stroke was assessed using Cox regression.

Results: Of the 118 cases, 39.0% were men, mean age 58.1±16.9 years; 12.6% had a history of stroke, compared with 4.9% of controls. The risk of stroke was significantly increased after age 65 years (hazard ratio, 6.0 [95% CI, 1.4-26.3]). Dementia, mild cognitive impairment, migraine with aura and depression were equally prevalent in cases and controls. Twenty-nine cases (25%) and 45 controls (24%) had an available brain magnetic resonance imaging. After age 65 years, cases had a higher white matter lesion burden and more lacunes. A severe small vessel disease phenotype compatible with cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy was rarely seen.

Conclusions: Cysteine altering NOTCH3 variants are an important contributor to the risk of stroke, lacunes, and white matter hyperintensities in the elderly population.

Keywords: control group; cysteine; neuroimaging; phenotype; white matter.

Figure 1.

Stroke incidence in cases with a NOTCH3^cys variant vs controls. Kaplan-Meier plot showing the proportion free of stroke in cases (n=118) and controls (n=184). Cases had a significantly shorter stroke-free survival compared with controls (P=0.02). The table under the graph shows the number of cases and controls at risk and the number of individuals with a first stroke per 10-y interval.

Figure 2.

White matter hyperintensity lesion load in cases with a NOTCH3^cys variant vs controls. Proportional bar charts showing (A) deep white matter (DWM) Fazekas scores and (B) periventricular white matter (PVWM) Fazekas scores, showing a higher WMH lesion load in cases vs controls. C and D, Scatterplots showing the age distribution per Fazekas score in cases vs controls. The cases with Fazekas DWM 3 did not have an alternative cause for their confluent deep white matter hyperintensities besides the NOTCH3^cys variant. Of the 3 controls with Fazekas DWM 3 or PVWM 3, one had been treated with cranial radiotherapy and one had a high vascular risk factor burden. Horizontal black lines represent mean ages.