2.5 Survival analysis
Figure 1 shows the flow chart of survival analysis in this study. We
performed survival analyses using the Kaplan-Meier method to evaluate
all-cause mortality in the PAH patients reported in the JADER database.
Survival times of the PAH patients with at least one adverse event that
resulted in death were calculated from the number of days between the
start date of any medication and the first-reported date of death from
adverse events. Survival times of the PAH patients without adverse
events that resulted in death were censored at the last-reported date of
adverse events that resulted in recovery, remission, non-recovery, or
sequelae. In the Kaplan-Meier analysis, continuous intravenous
epoprostenol or treprostinil, which is recommended for high-risk PAH
patients as permanent therapy, was analyzed separately from other drugs
targeting the prostacyclin pathway.12 Patients with
missing information about the date of adverse events in the JADER
database were excluded from the survival analyses. The log-rank test was
used to compare survival times between the PAH patients treated with
each pulmonary vasodilator or supportive therapy. Hazard ratios (HRs) of
respective therapies evaluated in this study for all-cause mortality in
the PAH patients were estimated using Cox proportional hazards models.
Based on the results of HRs and Kaplan-Meier curves, the covariates for
multivariate analysis were identified by selecting the clinically known
variables that may affect all-cause mortality in PAH patients. We also
employed a variable of warfarin use for the multivariate analysis, which
has been reported as a poor prognostic factor in epoprostenol
users.15,16 A two-sided p value of <
0.05 was considered statistically significant. All statistics were
analyzed using R version 4.2.0 (R Foundation for Statistical Computing,
Vienna, Austria).