Results
A total of 4734 CAs were analyzed. Baseline clinical and procedural characteristics (per procedure assessment) including gender differences are shown in Table 1. There were no meaningful differences in the medication between males and females. Overall, 188 (4.0%) major complications were detected: 86/1512 (5.7%) in females and 102/3222 (3.2%) in males (P < 0.0001). Of them, 59.6% were classified as MVCs. A total of 112 (2.4%) MVCs were unevenly distributed between genders: 54/1512 (3.5%) in females and 58/3222 (1.8%) in males (P < 0.0001). Thrombin injection, blood transfusion or surgical revision was required in 45% of all MVCs. There was a significant trend in the reduction of MVCs during the study period (Figure 1) with an average year-to-year drop by 0.13% (P = 0.048), which produced an overall drop by 1.7% during the study span of almost 13 years. Several clinical and procedural characteristics differed between those with and without MVCs as shown separately for males and females in Table 2. Details on gender differences in the frequency of individual MVCs are provided in Table 3. The female gender was associated with relative risk (RR) of 3.2 (95% confidence interval (CI): 1.9–5.4, P < 0.0001) of developing groin hematoma, RR of 3.9 (95% CI: 1.9–7.9, P < 0.0001) of significant drop of hemoglobin level, RR of 2.6 (95% CI: 1.1–6.7, P = 0.02) of blood transfusion, and RR of 2.3 (95% CI: 1.1–5.1, P = 0.02) of surgical revision.
Predictors of MVCs
Among “modifiable” factors, the strategy of non-interrupted anticoagulation and “Z”-stitch for hemostasis did not appear to have a significant impact on the reduction of the rate of MVCs (Table 2). On the contrary, US-guided venipuncture was associated with 1.44% drop of MVCs in males but not in females.
Longer procedures were associated with a higher rate of MVCs with an increase of 0.21% and 0.39% per every 30 minutes of procedure duration in males and females, respectively. The rate of MVCs slowly decreased by 0.15% and 0.12% per year in males and females, respectively, during the study period. Both relationships were statistically significant only in males (Table 4).
The risk of MVCs almost linearly rose with age in both males and females (Figure 2). On univariate regression analysis, older age, and lower body height were associated with MVCs in females (Table 4). Particularly, low-stature women (< 170 cm), were at the highest risk of MVCs (Figure 3). In males, older age, smaller body surface area (BSA), persistent AF, and longer procedure time was associated with MVCs (Table 4). Interestingly, the body weight was almost unrelated to MVCs in both genders (Table 4, Figure 3). The combined impact of body height and weight is illustrated separately for females and males in Figure 4.
On multivariate regression analysis in males (Table 4), age and US-guided venipuncture were independently associated with MVC that increased by 0.64% per every decade of age and, on the other hand, decreased by 1.55% when US-guided venous access was employed. The same analysis in females revealed that only body height was independently associated with MVCs that increased by 1.3% per 5-cm decrease of body height.
When the multivariate analysis was performed in the total cohort, age (P = 0.002), body height (P = 0.0002), and year of procedure (P = 0.01) were associated with the rate of MVCs, while gender itself was not an independent risk factor (P = 0.54).