Table 3.
Multivariate competing-risks regression analysis of cardiovascular death in PCNSLs.
| Variable | Before PSM | After PSM | ||
|---|---|---|---|---|
| HR (95% CI) | P Value | HR (95% CI) | P Value | |
| Unadjusted HR | ||||
| Chemotherapy | 0.476 (0.262–0.862) | 0.014 | 0.445 (0.240–0.824) | 0.010 |
| No evidence | Reference | Reference | ||
| Model 1 a | ||||
| Chemotherapy | 0.523 (0.290–0.943) | 0.031 | 0.469 (0.255–0.862) | 0.015 |
| No evidence | Reference | Reference | ||
| Model 2 b | ||||
| Chemotherapy | 0.519 (0.288–0.933) | 0.028 | 0.473 (0.257–0.870) | 0.016 |
| No evidence | Reference | Reference | ||
| Model 3 c | ||||
| Chemotherapy | 0.496 (0.274–0.898) | 0.021 | 0.446 (0.241–0.824) | 0.010 |
| No evidence | Reference | Reference | ||
a
In model 1, hazard ratios were adjusted for statistically significant factors according to univariate analysis (age at diagnosis and year of diagnosis).
b
In model 2, hazard ratios were adjusted for all factors in model 1, plus the potential confounders included in propensity score matching (sex, tumor location, histological type and surgery).
c
In model 3, hazard ratios were adjusted for all factors in model 2, plus the other demographic characteristics (race and marital status).
HR, hazard ratios; PSM, propensity score matching; 95% CI, 95% confidence interval.
The bold values mean P value < 0.05.