Pulmonary vascular obstruction index (PVOI) at inclusion and at pulmonary embolism diagnosis
| Total | Recurrent VTE | No recurrent VTE | Hazard ratio (95% CI) | p-value | |
| PVOI at inclusion (n=300) | |||||
| PVOI inclusion <5% | 169 | 24 (14.2) | 145 | Reference | 0.0002 |
| PVOI inclusion ≥5% | 131 | 40 (30.5) | 91 | 2.58 (1.56–4.29) | |
| PVOI at diagnosis (n=265) | |||||
| PVOI diagnosis <40% | 149 | 21 (14.1) | 128 | Reference | 0.0001 |
| PVOI diagnosis ≥40% | 116 | 41 (35.3) | 75 | 2.90 (1.71–4.91) | |
| Combined PVOI at diagnosis and at inclusion in four categories (n=263)# | |||||
| PVOI diagnosis <40% and PVOI inclusion <5% | 96 | 8 (8.3) | 88 | Reference | 0.0003 |
| PVOI diagnosis <40% and PVOI inclusion ≥5% | 51 | 13 (25.4) | 38 | 3.85 (1.59–9.30) | |
| PVOI diagnosis ≥40% and PVOI inclusion <5% | 46 | 17 (37.0) | 29 | 5.35 (2.31–12.40) | |
| PVOI diagnosis ≥40% and PVOI inclusion ≥5% | 70 | 24 (34.3) | 46 | 5.29 (2.37–11.8) | |
| Combined PVOI at diagnosis and at inclusion in two categories (n=263)¶ | |||||
| PVOI diagnosis <40% and PVOI inclusion <5% | 96 | 8 (8.3) | 88 | Reference | <0.0001 |
| PVOI diagnosis ≥40% and/or PVOI inclusion ≥5% | 167 | 54 (32.3) | 100 | 4.87 (2.32–10.03) |
Data are presented as n or n (%), unless otherwise stated. VTE: venous thromboembolic disease. #: in order to evaluate the concomitant impact of PVOI at inclusion and at the time of pulmonary embolism diagnosis, and taking in account the significant correlation and interaction between these two variables, these latter were combined into one variable and four combinations were considered: 1) initial PVOI <40% and PVOI at inclusion <5%, 2) initial PVOI <40% and PVOI at inclusion ≥5%, 3) initial PVOI ≥40% and PVOI at inclusion <5%, and 4) initial PVOI ≥40% and PVOI at inclusion ≥5%. ¶: as the relative risks of recurrent VTE were increased by a similar extent in three categories compared with the reference category (i.e. PVOI at diagnosis <40% and PVOI at inclusion <5%), the combined variable coding for the two PVOIs was dichotomised for multivariate analysis (table 3).