3.1 Plaque XCI skewing is highly prevalent in female atherosclerotic patients
We selected all female atherosclerotic patients who underwent carotid endarterectomy (CEA) and had DNA stored for XCI determination (n=154). Of these 155 women, a subset of 55 women had DNA stored from blood samples. Baseline characteristics are displayed in Table 1. In short, these women were 66 years old on average and had a median BMI of 26 at baseline. At the day of the hospitalization, 32% of the patients had a previous history of peripheral artery occlusive disease (PAOD), while 26% had a previous history of coronary artery disease (CAD). Most of the patients suffered from hypertension (88%) and/or hypercholesterolemia (70%) and were on antiplatelet (94%) and/or lipid-lowering (77%) medications. Also, 71 patients (47%) reported to smoke and 19% had diabetes.
XCI skewing was determined in 154 carotid plaque and 55 blood-matched DNA samples by performing a HUMARA assay followed by PAGE gel quantification. We used these quantitative data to determine XCI skewing status using a computed cut-off value of 63.9% (Fig. S1). XCI skewing occurred in 76 of 154 plaques (49.4%), while in blood, it was higher with 67.2% (37/55) (Fig. 1 A). Blood skewing was strongly correlated with plaque skewing (P = 0.0001) (Fig. 1 B). Age was not associated with blood (10-year β = -0.14; P = 0.89, Fig. S2 A – B) nor with plaque skewing (10-year β = 0.54; P = 0.43, Fig. S2 C – D). The baseline characteristics showed no significant difference between the populations with skewed and non-skewed plaque (Table 1). Indeed, the patients presenting skewed plaques were on average 67 years old and had a median BMI of 25, while the ones presenting with non-skewed plaques were on average 66 years old and had a median BMI of 26. Also, 37% of the skewed group had a history of PAOD and 29% of CAD while in the non-skewed group, 28% had history PAOD and 23% of CAD. Smokers were reported to be 41% in the skewed and 53% in the non-skewed group and the presence of diabetes was 18% and 19%, respectively.
To assess whether the extent of skewing (i.e., how skewed a plaque is) was important in relation to baseline characteristics, we created a categorical binned variable where X-skewing was classified in 4 levels (No, Low, Mid, High). Baseline characteristics of these 4 levels are shown in Table S1.
Table 1. Baseline characteristics of patients with and without XCI skewing in plaque.
XCI Skewing:
|
Overall
|
No Skewed
|
Skewed
|
P Value
|
N (%)
|
154 (100%)
|
78 (50.6)
|
76 (49.4)
|
-
|
Age in years (mean, SD)
|
66 (9)
|
66 (9)
|
67 (9)
|
0.40
|
BMI (median, IQR)
|
26 [23, 28]
|
26 [23, 28]
|
25 [23, 27]
|
0.26
|
Current Smoker, yes (%)
|
71 (47)
|
40 (53)
|
31 (41)
|
0.22
|
Diabetes Mellitus, yes (%)
|
29 (19)
|
15 (19)
|
14 (18)
|
1.00
|
Hypertension, yes (%)
|
135 (88)
|
71 (91)
|
64 (84)
|
0.30
|
Hypercholesterolemia, yes (%)
|
105 (70)
|
51 (67)
|
54 (72)
|
0.63
|
History of coronary artery disease (%)
|
40 (26)
|
18 (23)
|
22 (29)
|
0.52
|
History of PAOD (%)
|
50 (32)
|
22 (28)
|
28 (37)
|
0.33
|
Use of antiplatelet therapy (%)
|
144 (94)
|
71 (91)
|
73 (96)
|
0.35
|
Use of lipid-lowering drugs (%)
|
119 (77)
|
58 (74)
|
61 (80)
|
0.50
|
GFR (MDRD) mL/min per 1.73 m2 (mean, SD)
|
70 (20)
|
69 (20)
|
71 (19)
|
0.41
|
LDL in mg/dL (median, IQR)
|
111 [82, 145]
|
109 [80, 144]
|
111 [82, 144]
|
0.90
|
HDL in mg/dL (median, IQR)
|
46 [37, 55]
|
45 [37, 55]
|
46 [37, 60]
|
0.80
|
Total cholesterol in mg/dL (median, IQR)
|
189 [147, 228]
|
185 [147, 226]
|
203 [147, 230]
|
0.78
|
Triglyceride levels in mg/dL (median, IQR)
|
127 [94, 189]
|
127 [96, 187]
|
129 [89, 197]
|
0.76
|
Presenting symptoms (%)
|
0.25
|
Asymptomatic
|
25 (18)
|
15 (21)
|
10 (16)
|
|
Transient Ischemic Attack (TIA)
|
77 (57)
|
36 (50)
|
41 (64)
|
|
Stroke
|
34 (25)
|
21 (29)
|
13 (20)
|
|
3.2 XCI skewing in different plaque phenotypes
To investigate whether XCI skewing in plaque was more common in a specific plaque phenotype, we assessed the relation between plaque skewing and histological plaque characteristics. For this, fat content (>10% and >40%), amount of calcification, amount of collagen, presence of plaque hemorrhage, macrophage, smooth muscle cell, neo-vessel and glycophorin content within the plaque (Table 2) was studied30. Univariate analysis showed a linear association between plaque skewing (Continuous variable) and plaque hemorrhage (OR: 1.44 [1.06 - 1.98]; P = 0.02), but not when using the dichotomous variable (Table 2). Data on the binned plaque skewing levels are showed in Supplementary Table S2.
Table 2. Association of plaque skewing with plaque characteristics.
|
XCI Skewing
|
Odds Ratio [95% CI]
|
β [95% CI]
|
P Value
|
Fat content
(>10%)
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
-
|
Dichotomous (Skewed)
|
0.87 [0.45 to 1.71]
|
-
|
0.69
|
*Continuously
|
0.96 [0.71 to 1.31]
|
-
|
0.82
|
Fat Content
(>40%)
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
-
|
Dichotomous (Skewed)
|
1.41 [0.64 to 3.22]
|
-
|
0.39
|
*Continuously
|
1.08 [0.74 to 1.54]
|
-
|
0.67
|
Calcification
(major)
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
-
|
Dichotomous (Skewed)
|
1.37 [0.71 to 2.68]
|
-
|
0.35
|
*Continuously
|
1.08 [0.80 to 1.48]
|
-
|
0.60
|
Collagen
(major)
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
-
|
Dichotomous (Skewed)
|
0.80 [0.34 to 1.88]
|
-
|
0.61
|
*Continuously
|
0.96 [0.66 to 1.42]
|
-
|
0.82
|
Plaque Hemorrhage
(major)
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
-
|
Dichotomous (Skewed)
|
1.50 [0.79 to 2.87]
|
-
|
0.22
|
*Continuously
|
1.44 [1.06 to 1.98]
|
-
|
0.02
|
Macrophage
(major)
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
-
|
Dichotomous (Skewed)
|
0.77 [0.41 to 1.45]
|
-
|
0.42
|
*Continuously
|
1.03 [0.77 to 1.39]
|
-
|
0.80
|
Smooth Muscle Cells (major)
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
-
|
Dichotomous (Skewed)
|
0.98 [0.46 to 2.08]
|
-
|
0.96
|
*Continuously
|
0.92 [0.66 to 1.30]
|
-
|
0.66
|
**Neo-vessels
(major)
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
-
|
Dichotomous (Skewed)
|
0.88 [0.44 to 1.77]
|
-
|
0.72
|
*Continuously
|
0.93 [0.68 to 1.28]
|
-
|
0.62
|
***Glycophorin
(increase of plaque area)
|
Dichotomous (Non-skewed)
|
-
|
(1.0) ref
|
-
|
Dichotomous (Skewed)
|
-
|
0.09 [-0.26 to 0.45]
|
0.60
|
*Continuously
|
-
|
0.02 [-0.13 to 0.18]
|
0.77
|
*Calculated for 10 points percentage of XCI skewing. Data transformed with bestNormalize package in R: **Binarize technique; *** The Ordered Quantile Normalization technique.
3.3 Risk factors do not influence the relation between plaque skewing and plaque hemorrhage
Although risk factors did not show any significant differences between skewed and non-skewed plaques, we studied whether they modulate the association between plaque skewing and plaque hemorrhage (OR [95% CI]: 1.44 [1.06 - 1.98]; P = 0.02). We used classical cardiovascular risk factors, such as age, body mass index (BMI), smoking, diabetes mellitus and glomerular filtration rate to correct the model for one risk factor at the time. We found that BMI slightly attenuated the association of plaque skewing with plaque hemorrhage (OR [95% CI]: 1.37 [0.99 - 1.94]; P = 0.07). Also, smoking marginally increased this association (OR [95% CI]: 1.50 [1.10 - 2.09]; P = 0.01) (Table 3). Supplementary Table S3 shows the data on the binned plaque skewing levels.
Table 3. Effect of classical cardiovascular risk factors on the association between plaque skewing and plaque hemorrhage.
|
Plaque Hemorrhage (major)
|
XCI Skewing
|
Odds Ratio [95% CI]
|
P Value
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
Dichotomous (Skewed)
|
1.50 [0.79 to 2.87]
|
0.22
|
Continuously*
|
1.44 [1.06 to 1.98]
|
0.02
|
|
Adjusted for: Age
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
Dichotomous (Skewed)
|
1.49 [0.78 to 2.86]
|
0.22
|
Continuously*
|
1.43 [1.06 to 1.98]
|
0.02
|
|
Adjusted for: BMI
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
Dichotomous (Skewed)
|
1.49 [0.76 to 2.97]
|
0.25
|
Continuously*
|
1.37 [0.99 to 1.94]
|
0.07
|
|
Adjusted for: Smoking
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
Dichotomous (Skewed)
|
1.62 [0.84 to 3.14]
|
0.15
|
Continuously*
|
1.50 [1.10 to 2.09]
|
0.01
|
|
Adjusted for: Diabetes Mellitus
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
Dichotomous (Skewed)
|
1.50 [0.79 to 2.87]
|
0.22
|
Continuously*
|
1.44 [1.06 to 1.98]
|
0.02
|
|
Adjusted for: GFR (MDRD)
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
Dichotomous (Skewed)
|
1.46 [0.77 to 2.80]
|
0.25
|
Continuously*
|
1.44 [1.06 to 1.99]
|
0.02
|
|
Adjusted for: BMI + Smoking
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
Dichotomous (Skewed)
|
1.53 [0.77 to 3.05]
|
0.23
|
Continuously*
|
1.38 [0.99 to 1.96]
|
0.06
|
*Calculated for 10 points percentage of XCI skewing.
3.4 Association with cardiovascular endpoints
We investigated if XCI skewing in plaque was associated with the incidence of secondary cardiovascular events in a 3-year follow-up after endarterectomy. For this, we used survival analysis, Kaplan-Meier (Fig. S3) and cox proportional hazard model, corrected for BMI and current smoking (Fig. 2), to assess the association of plaque skewing with composite (41 events, 27%), major adverse cardiovascular (MACE; 13 events, 9%) and peripheral artery events (28 events, 19%) during 3-years follow-up.
Women with skewed plaques had no more MACE (Fig. 2 B) compared to women without skewing (HR [95%CI]: 0.89 [0.30 - 2.65]; P = 0.84), also when correcting for BMI and current smoking (HR [95%CI]: 0.80 [0.25 - 2.53]; P = 0.70) (Table 3). During 3 years of follow-up, plaque skewing was also not significantly associated with composite endpoints (Table 3; Fig 2 A). However, a trend was observed in the association of dichotomous plaque skewing with composite events (HR [95%CI]: 1.70 [0.91 - 3.19]; P = 0.09) which increased slightly in strength when correcting the model for BMI and current smoking (HR [95%CI]: 1.82 [0.96 - 3.45]; P = 0.07). We found that women with skewed plaques had significantly more peripheral artery events during 3 years follow up (Table 3; Fig. 2 C) when using both continuous (HR [95%CI]: 3.14 [1.38 - 7.18]; P = 0.01) and dichotomous XCI (HR [95%CI]: 1.46 [1.09 - 1.97]; P = 0.007) in multivariate analyses. Also adjusting for plaque hemorrhage did not change the relation between XCI and outcome (data not shown). We performed similar analysis using the binned plaque skewing levels which are showed in Supplementary Table S4, Fig. S4 and Fig. S5.
Table 3. Association of plaque skewing with secondary cardiovascular endpoints during 3-years follow-up.
|
Crude analysis
|
Adjusted for BMI and Smoking
|
XCI Skewing
|
Hazard Ratio [95% CI]
|
P Value
|
Hazard Ratio [95% CI]
|
P Value
|
|
Composite Event
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
(1.0) ref
|
-
|
Dichotomous (Skewed)
|
1.70 [0.91 to 3.19]
|
0.1
|
1.82 [0.96 to 3.45]
|
0.07
|
Continuously*
|
1.24 [0.96 to 1.62]
|
0.1
|
1.21 [0.93 to 1.57]
|
0.15
|
|
Major Cardiovascular Event (MACE)
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
(1.0) ref
|
-
|
Dichotomous (Skewed)
|
0.89 [0.30 to 2.65]
|
0.84
|
0.80 [0.25 to 2.53]
|
0.7
|
Continuously*
|
0.98 [0.59 to 1.62]
|
0.94
|
0.81 [0.47 to 1.41]
|
0.46
|
|
Peripheral Artery Event
|
Dichotomous (Non-skewed)
|
(1.0) ref
|
-
|
(1.0) ref
|
-
|
Dichotomous (Skewed)
|
2.85 [1.25 to 6.47]
|
0.012
|
3.14 [1.38 to 7.18]
|
0.007
|
Continuously*
|
1.45 [1.07 to 1.97]
|
0.018
|
1.46 [1.09 to 1.97]
|
0.011
|