Sex differences in coronary artery plaque composition detected by coronary computed tomography: quantitative and qualitative analysis

All patients with suspected CAD referred for CCTA between 2006 and 2016 were prospectively included in a local registry. The study was approved by our local institutional review board; the requirement for patient informed consent forms was waived. Exclusion criteria for this study were: known CAD or prior revascularisation (including percutaneous coronary intervention, PCI, or coronary artery bypass graft, CABG), prior myocardial infarction, severe renal dysfunction (glomerular filtration rate <45 ml/min/1.73 m²), chest radiation or chemotherapy, non-diagnostic image quality (involving proximal or mid-segments). A flow chart of patient recruitment is provided in Fig. 1.

Follow-up was performed for all patients by reviewing medical records from the local system and referring hospitals were contacted for adverse events. Our centre is a tertiary-care, university hospital and is the only clinic providing invasive coronary angiography (ICA) and cardiac surgery in the province (covering specialised care for approx. 800,000 inhabitants). Additionally, the health care information system is connected to the central secondary care hospitals, expanding access to patient information. Data on ICA procedures and results, consecutive coronary revascularisation procedures (either PCI or CABG) were collected. The primary endpoint was a major adverse cardiac event (MACE) defined as a composite of ACS (including unstable angina, non-ST-elevation myocardial infarct (NSTEMI) and STEMI) and cardiac death by in-hospital documentation and/or post-mortem histology report. The secondary endpoint was revascularisation (PCI or CABG).

Two different CT scanners were used: a 64-slice scanner (Somatom Sensation 64, Siemens) with a detector collimation of 64 × 0.75 mm and a rotation time of 0.33 s until 2009; thereafter CCTA examinations (40.6%) were performed using a first-generation, 128-slice, dual-source CT scanner (Somatom Definition FLASH, Siemens).

An initial non-contrast ECG-gated calcium score (CCS) CT scan was performed and the Agatston score was calculated followed by CCTA. In patients examined with a 64-slice CT scanner, retrospective ECG gating was applied. Patients examined with the 128-DSCT equipment were scanned in either prospective ECG-triggered mode if they had a regular heart rate <65 bpm (70% of RR interval) and >65 bpm (40% of RR interval) or with retrospective ECG gating in patients with an irregular heart rate.

A contrast agent with an iodine concentration of 370 mg/ml was injected (flow rate 4–6 ml/s followed by 40 cc saline bolus); scans were triggered into arterial phase using a bolus tracking technique at a threshold of 100 Hounsfield units (HU) in the ascending aorta. Contrast volume varied between 65 and 120 ml in accordance with the individual patient’s body mass index and scan characteristics. All images were reconstructed with 0.75 mm slice width.

Coronary CT images were evaluated on a dedicated 3D post-processing software by two independent observers (2 years and 8 years of experience) and a consensus reading obtained in case of unclear findings. Only data sets without relevant artefacts leading to any non-diagnostic segment were included.

A standardised CCS was calculated for all patients. Contrast-enhanced CCTA evaluation included coronary stenosis divided into no stenosis, non-obstructive <50%, intermediate 50–70% and obstructive >70%. All plaques were documented and classified according to their morphology (1 calcified; 2 predominantly calcified; 3 predominantly non-calcified; 4 non-calcified).

Total plaque load was quantified using the segment involvement score (SIS, total number of segments with any plaque) according to the 16-segment American Heart Association classification and the non-calcified SIS (ncSIS), including non-calcified plaques only (type 4).

All lesions were evaluated regarding predefined high-risk features of vulnerable plaques [18].

Follow-up was performed using patient chart check-ups. Information on clinical outcome was documented, including ACS (unstable angina, NSTEMI and STEMI), all-cause and cardiac mortality as well as revascularisation (PCI and CABG).

All diagnostic and therapeutic coronary angiographic procedures were performed at our centre. Obstructive CAD was defined as stenosis >70% or fractional flow reserve (FFR) <0.8 in cases of unclear visual grading. In cases of borderline stenosis without FFR validation, additional non-invasive ischaemia testing is required according to the local standard operating procedure in non-high-risk stenosis on CCTA before elective revascularisation.

All statistical analyses were performed using SSPS software (V21.0, SPSS Inc.). Quantitative variables are expressed as means ± standard deviation, categorical variables as absolute values and percentages. A p-value of less than 0.05 was considered significant. Mean differences between parametric data were tested with Student’s t-test for two samples or one-way ANOVA for more, chi-square test or Fisher’s exact test for categorical data. Odds ratios were calculated using binary logistic regression.

Matching factors were determined a priori and all variables forced into propensity scoring using logistic regression. Factors entered included age, male sex, hypertension, dyslipidaemia, diabetes mellitus, family history of premature cardiac death, smoking and body mass index. A nearest-neighbour approach using 1:1 matching was performed on site.

Of 6,050 screened patients, 1,050 propensity score matched patients were selected for inclusion in the analysis. The mode of recruitment is depicted in Fig. 1; the demographics are outlined in Tab. 1.

Male patients had significantly worse CCTA findings than females. After adjusting for coronary risk factors, there was a significant difference in CCS (median 11.6 (0–174.4) vs 0 (0–12.5), p < 0.0001), prevalence of any CAD (66.0 vs 34.1%, p < 0.0001) and high-grade stenosis (27.5 vs 11.9%, p < 0.0001).

Plaque analysis showed a significantly higher proportion of overall affected segments in males (861 vs 752, p < 0.0001) with a higher proportion of calcified plaques (62% vs 42.4%, p < 0.0001) compared to females. Interestingly, females had significantly more mixed and non-calcified plaques (33.5% vs 24.4%, p = 0.006 and 24.1% vs 13.6%, p = 0.22, respectively). The prevalence of high-risk plaque features was higher for male patients (mean 41.1% vs 23.8%, p < 0.0001).

Sub-analysis of individual high-risk plaque features also showed a distinctly higher prevalence of every feature in male patients (Tab. 2, Fig. 2).

The total rate of all-cause mortality was 3.7% with significant differences in cardiac mortality of 4.8% for males versus 1.9% for females. MACE rate was higher in male than in female patients (5.3% vs 1.9%).

Patient follow-up was performed up to 11.5 years. Relevant information was available for at least 1 year in 975 patients (92.9%). After a mean period of 5.6 years, 31.1% of male patients were referred for invasive angiography (14.1% within 1 month after CCTA), while only 16.4% of female patients (5.0% within 1 month after CCTA) underwent ICA. After ICA, almost 50% of male patients underwent revascularisation therapy (80/163, 49.1%), but only approximately one third of female patients underwent PCI or CABG (28/86, 32.6%). FFR verification of haemodynamic relevance in obstructive stenosis was performed in 30 patients (12.1% of total ICA examinations), 51 patients additionally underwent nuclear stress imaging before ICA, and 46 patients (particularly earlier cases before 2008) underwent stress treadmill testing with significant ECG alterations.

The odds ratio for conventional CCTA findings and each high-risk plaque feature was significant in predicting revascularisation, MACE and cardiac death. While the event rate was not high enough to compare inter-sex differences, a trend towards a higher odds ratio for females could be observed. Positive remodelling was in fact significantly higher for female patients than for males (OR 27.3 (5.8–128.8) vs 2.2 (1.0–4.7)) (Fig. 3, Tab. 3).

In a sub-analysis of the 42 MACE events, female patients had a lower CCS (63.3 AU vs 265.2 AU, p < 0.05) and all-grade stenosis (40% vs 57.1%, p < 0.05); however, females showed a trend towards a relatively higher prevalence of non-calcified plaques (33.3% vs 14.3%, p = 0.22) and presence of any high-risk plaque feature (80% vs 71.4%, p = 0.35) (Fig. 4).

Our derivation cohort consisted of a relatively young population (mean 58 years) with a low prevalence of CAD. All patients were referred with a clinical suspicion of cardiovascular heart disease, the majority with a low-to-intermediate risk of CAD, which may have resulted in a selection bias, considering different clinical symptoms and indications in men and women.

Although long-term clinical follow-up was performed (for up to 11 years), a low overall mortality (3.5%) and MACE rate (3.6%) was observed. This resulted in a non-significant difference in MACE odds ratio in the sub-analysis between the sexes.