Vertebral Tortuosity Is Associated With Increased Rate of Cardiovascular Events in Vascular Ehlers-Danlos Syndrome

Abstract

Background Arterial tortuosity is associated with adverse events in Marfan and Loeys-Dietz syndromes but remains understudied in Vascular Ehlers-Danlos syndrome. Methods and Results Subjects with a pathogenic COL3A1 variant diagnosed at age <50 years were included from 2 institutions and the GenTAC Registry (National Registry of Genetically Triggered Thoracic Aortic Aneurysms and Cardiovascular Conditions). Height-adjusted vertebral artery tortuosity index (VTI-h) using magnetic resonance or computed tomography angiography was calculated. Associations between VTI-h and outcomes of (1) cardiovascular events (arterial dissection/rupture, aneurysm requiring intervention, stroke), or (2) hollow organ collapse/rupture at age <50 years were evaluated using receiver operator curve analysis (using outcome by age 30 years) and mixed-effects Poisson regression for incidence rate ratios. Of 65 subjects (54% male), median VTI-h was 12 (interquartile range, 8-16). Variants were missense in 46%, splice site in 31%, and null/gene deletion in 14%. Thirty-two subjects (49%) had 59 events, including 28 dissections, 5 arterial ruptures, 4 aneurysms requiring intervention, 4 strokes, 11 hollow organ ruptures, and 7 pneumothoraces. Receiver operator curve analysis suggested optimal discrimination at VTI-h ≥15.5 for cardiovascular events (sensitivity 70%, specificity 76%) and no association with noncardiovascular events (area under the curve, 0.49 [95% CI, 0.22-0.78]). By multivariable analysis, older age was associated with increased cardiovascular event rate while VTI-h ≥15.5 was not (incidence rate ratios, 1.79 [95% CI, 0.76-4.24], P=0.185). However, VTI-h ≥15.5 was associated with events among those with high-risk variants <40 years (incidence rate ratios, 4.14 [95% CI, 1.13-15.10], P=0.032), suggesting effect modification by genotype and age. Conclusions Increased arterial tortuosity is associated with a higher incidence rate of cardiovascular events in Vascular Ehlers-Danlos syndrome. Vertebral tortuosity index may be a useful biomarker for prognosis when evaluated in conjunction with genotype and age.

Keywords: VEDS; arterial rupture; cardiovascular; dissection; genetics.

Figure 1. ROC for height‐adjusted vertebral tortuosity index and event before age 30 years.

A, ROC of any event; B, ROC of cardiovascular events; C, ROC of organ events. ROC analysis of height‐adjusted vertebral tortuosity index and having an event before age 30 years demonstrated mild discriminatory value with an area under the curve of 0.64. However, predictive ability increased when limiting to cardiovascular events before age 30 years (area under the curve=0.71), comparatively higher than the ROC for organ events before age 30 (area under the curve=0.49). Based on ROC analysis of cardiovascular events alone, an optimal cut point was selected at 15.5 (sensitivity 70.0%, specificity 76.4%). This optimal cut point is denoted by the red arrow. AUC indicates area under the curve; and ROC, receiver operator curve.

Figure 2. Freedom from cardiovascular event by characteristics.

A, By dichotomized VTI‐h; B, By ethnicity; C, By risk‐stratified genotype; D, By sex. All Kaplan–Meier graphs present freedom from event (1‐hazard of event) on its y axis. KM indicates Kaplan–Meier; VTI‐h, height‐adjusted vertebral tortuosity index.

Figure 3. Freedom from cardiovascular event by dichotomized height‐adjusted vertebral tortuosity index and risk‐stratified COL3A1 genotype.

A, Limited to subjects with high‐risk variants (n=40); B, Limited to subjects with low‐risk variants (n=18). All Kaplan–Meier graphs present freedom from event (1‐hazard of event) on its y axis. KM indicates Kaplan–Meier; and VTI‐h, height‐adjusted vertebral tortuosity index.

Figure 4. Incidence rate of events by dichotomized height‐adjusted vertebral tortuosity index and person‐decade.

Incidence rates per 100 person‐years and corresponding 95% CIs are presented for each decade of life. P values denote whether or not the incidence rate ratio computed between the rate of any event by each person‐decade and event rate within the first decade of life, within each respective group stratified by height‐adjusted vertebral tortuosity index, are significantly different. VTI‐h indicates height‐adjusted vertebral tortuosity index.