3 | RESULTS
3.1 | Anatomical variations
Studies that matched the required criteria and contained both complete and relevant data were included in the anatomical variations part of this meta-analysis. Basic characteristics of all 73 included studies are presented in Table 1. Anatomical variations were collected using a two-stage compilation method, established by authors of this study, based on the most common variations described in the literature. In stage one, data from qualified studies were divided into two groups: subjects who (1) had classical arrangement of the PV confluence or (2) had any other arrangement than classical type. As classical type (Type 1) authors consider two main PVs on the right side and two on the left side: RSPV, RIPV, LSPV and LIPV. In the second stage, authors established 9 subtypes of arrangements of the PVs, that differed from the classical one: Type 2 - arrangement contains RSPV, RIPV and one left common pulmonary vein (LCPV); Type 3 - arrangement contains LSPV, LIPV and on right common pulmonary vein (RCPV); Type 4 - arrangement contains RSPV, RIPV, LSPV, LIPV and an additional right middle pulmonary vein (RMPV); Type 5 - arrangement contains RSPV, RIPV, LSPV, LIPV and both additional RMPV and LMPV; Type 6 - arrangement contains LSPV, LIPV, LMPV and one RCPV; Type 7 - arrangement contains RSPV, RIPV, LSPV, LIPV and an additional left middle pulmonary vein (LMPV); Type 8 - arrangement contains RSPV, RIPV, RMPV and one LCP; Type 9 - arrangement contains one LCPV and one RCPV; Type 10 - includes all other arrangements. The analysis of the 9 types, different than the classical one, was performed using only those studies that clearly differentiated those arrangements. Imaged arrangements according to authors compilation method are shown in Figure 2.
Overall, the anatomical variations of PV ostia were examined in 73 studies including 13811 subjects . The pooled prevalence of classical type was 70.25% (95% CI: 66.05% - 74.28%). The pooled prevalence of any other type was 29.50% (95% CI: 25.20% - 33.99%). It may be concluded that data regarding those two types may represent considerable heterogeneity as the I2 was established appropriately 96.28 for the classical type and 96.77 for other types. The pooled prevalence of 9 other types were established based on 8504 subjects and were found to be 17.37% (95% CI: 14.67% - 20.25%) for type 2, 8.78% (95% CI: 6.03% - 11.98%) for type 3, 8.38% (95% CI: 6.27% - 10.76%) for type 4, 5.11% (95% CI: 3.87% - 6.51%) for type 5, 1.24% (95% CI: 0.72% - 1.89%) for type 6, 1.19% (95% CI: 0.73% - 1.75%) for type 7, 1.09% (95% CI: 0.64% - 1.66%) for type 8, 0.20% (95% CI: 0.12% - 0.31%) for type 9 and 5.29% (95% CI: 4.08% - 6.65%) for type 10 (Table 2). The data regarding types 2, 3, 4, 5, 6, 8 and 10 may represent considerable heterogeneity as the established I2 were included in the compartment from 75 to 100 (Table 2).
The studies that were carried out on patients with AF were subjected to additional analysis. The pooled prevalence in patients with AF was calculated based on 11103 subjects and was found to be 70.61% (95% CI: 65.55% - 75.42%) for classical arrangement of PVs and 29.05% (95% CI: 23.92% - 34.45%) for any other type (Table 2). It may be concluded that data regarding those types may represent considerable heterogeneity as the established I2 were 96.81 for classical type and 97.23 for any other than classical type. The most common variation regarding all other arrangements than the classical one in patients with AF was found to be Type 2 (RSPV, RIPV and LCPV) and its pooled prevalence was 18.16% (95% CI: 15.30 – 21.21%) (Table 2). When comparing PV ostia variations between general population and AF patients, no statistically significant differences were found in the distribution of the classical type (p=0.92). More thorough analyses show no significant differences in subtypes distribution with the types with either additional middle PV ostium on left or right or both sides (Types 4-8) patients with and without AF (p > 0.05). For details, please see Tables 2 and 3.
Included studies differed in a way of imaging the number and the arrangement of the pulmonary veins. Computed tomography, echocardiography, magnetic resonance imaging, magnetic resonance angiography, intraoperative imaging and cadaver dissection methods were used in analyzed studies to establish PVs arrangements. Hence, an additional statistical analysis according to imaging technique was performed (no significant differences, all p > 0.05). Also, a further analysis according to geographical location of each study was performed. For details, please see Table 2.
3.2 | PVs ostia diameters
Mean maximal diameters of PVs ostia were extracted from 86 studies that matched the inclusion criteria . Baseline characteristics of each included study are gathered in Table 4. Data regarding mean maximal diameters of 6 most commonly occurring PV ostia: RSPV, RIPV, LSPV, LIPV, RMPV and LCPV are shown in Table 5. All data concerning diameters were either presented or have been converted by authors of this study to millimeters [mm]. Mean maximal diameter of RSPV was established based on 12073 subjects and was found to be 18.87 mm (95% CI: 18.35 – 19.40 mm) and mean maximal diameter of RIPV (n=11980) was 17.17 mm (95% CI: 16.65 – 17.69 mm). Mean maximal diameter of LSPV was established based on 11865 subjects and was found to be 18.12 mm (95% CI: 17.63 – 18.61 mm), while the mean maximal diameter of LIPV (n=11849) was 16.25 mm (95% CI: 15.82 – 16.69 mm). The comparison between individual PV ostia shows that RSPV ostium is significantly larger than RIPV (Table 5). The same was observed for left-sided PV ostia (superior larger than inferior) (Table 5). Left-sided PV ostia were observed to have smaller diameters than corresponding right-sided PVs; however, the difference was not statistically significant (Table 5). Mean maximal diameter of RMPV was established based on 5030 subjects and was found to be 9.43 mm (95% CI: 8.89 – 9.96 mm) and was the smallest among analyzed ostia. Mean maximal diameter of LCPV was established based on 4966 subjects and was found to be 29.01 mm (95% CI: 28.07 – 29.94 mm) (Table 5), which was the largest value among all analyzed veins.
Included studies significantly differed in methodology of measuring the maximal diameter of each PV ostium (angiography, computed tomography, echocardiography, magnetic resonance imaging and magnetic resonance angiography methods). Hence, additional statistical analyses of data according to imaging technique and geographical location were performed (Table 5).
3.3 | Ostial area
The data on ostial area were extracted from 18 studies that matched the inclusion criteria . Baseline characteristics of each analyzed study are gathered in Table 6. Data regarding mean ostial area of 6 most common PV ostia: RSPV, RIPV, LSPV, LIPV, RMPV and LCPV are shown in Table 7. All data regarding ostial area were either presented or have been converted by authors of this study to square millimeters [mm2]. Mean ostial area of RSPV was established based on 2697 subjects and was found to be 280.12 mm2(95% CI: 253.25 – 307.01 mm2) and mean ostial area of RIPV (n=2697) was 232.75 mm2 (95% CI: 214.45– 251.05 mm2). Mean ostial area of LSPV was established based on 2624 subjects and was found to be 228.27 mm2(95% CI: 210.53 – 246.01 mm2) and mean ostial area of LIPV (n=2464) was 175.99 mm2 (95% CI: 162.94 – 189.05 mm2). The comparison between individual PV ostia indicates that RSPV ostium area is significantly larger than RIPV (Table 7). The same was observed for left-sided PVs ostial area (superior greater than inferior) (Table 7). Moreover, the difference between left-sided and right-sided veins was observed (right ostial areas larger than left) (Table 7). Mean ostial area of RMPV was established based on 1285 subjects and was found to be 68.01 mm2 (95% CI: 54.11 – 81.91 mm2), which was the smallest area among all analyzed vein ostia. Mean ostial area of LCPV was established based on 1269 subjects and was found to be 451.92 mm2 (95% CI: 388.50 – 515.35 mm2) (Table 7), which was the largest observed area (statistically significant).
Included studies differed in a way of imaging the ostial area of each pulmonary vein. Computed tomography, magnetic resonance imaging and magnetic resonance angiography methods were used to establish those areas. Hence, additional statistical analyses of data according to imaging technique and location were performed. All results mentioned above and more detailed statistics are presented in Table 7.