4. DISCUSSION
The present study evaluated whether patient data from the JADER database
could be used for survival analysis to compare the long-term efficacies
of PAH-specific vasodilators targeting the same biological pathway in
PAH patients. Our study findings indicate that continuous intravenous
epoprostenol and oral ERAs were associated with long-term survival in
PAH patients. In contrast, sildenafil, a phosphodiesterase 5 inhibitor
(PDE5i), did not associate with improvement in PAH mortality in this
study. To the best of our knowledge, this is the first report to
comprehensively compare the effects of PAH-specific vasodilators on
long-term survival using a spontaneous reporting database. Our results
suggest that spontaneous reporting databases partially provide useful
information about associations between medications and all-cause
mortality in limited situations such as analysis for rare diseases.
The demographics of the PAH patients reported in the JADER database were
characterized by a high proportion of females, and most of the patients
were idiopathic PAH. Similar demographics of PAH patients were reported
in studies using RWD, such as patient registries or medical
records.16–19 Ogawa et al. conducted a retrospective
medical chart review assessing 141 Japanese PAH patients and reported
that 40 patients died during the study period, and 21 patients died from
heart failure.16 Although the JADER database is a
spontaneous adverse event reporting database, the data of PAH patients
contain causes of disease-related death, including death from heart
failure, as well as just side effects of drugs. In our study population,
the percentage of patients whose clinical outcomes were reported as
death and the treatment distributions of respective pulmonary
vasodilators were similar to the rate of all-cause death and treatments
distributions in PAH patients registered in the Japan Pulmonary
Hypertension Registry (JAPHR), the largest network of pulmonary
hypertension hospitals in Japan.20,21 In the JAPHR,
approximately 20% of the registered patients resulted in death from any
causes; the most frequently prescribed drugs in each major drug class
were epoprostenol, selexipag, tadalafil, and macitentan, in consistency
with our results.20,21 The high similarity of our
study population to other RWD often used for cohort studies indicates
the reliability of our results. In limited situations where
representative patient population data in rare diseases, such as PAH,
are unavailable, patient data from a spontaneous reporting database,
with confirmed similarity of the demographics to the whole patient
population, may be alternative data for cohort studies.
The present study showed that among drugs targeting the prostacyclin
pathway, epoprostenol was most associated with long-term survival in the
PAH patients reported in the JADER database. The 2022 European Society
of Cardiology/European Respiratory Society (ESC/ERS) guidelines
recommend the addition of epoprostenol or treprostinil to the
combination of an ERA and a PDE5i for high-risk PAH
patients.12 The effect of epoprostenol on long-term
survival has been confirmed in clinical trials;22–24however, there are no clinical trials directly comparing epoprostenol
and treprostinil regarding long-term survival. Epoprostenol requires
continuous administration via a central venous catheter owing to its
short half-life.25 In contrast, treprostinil is
chemically stable, and several delivery systems, including implantable
pumps, have been developed.26,27 Although the
long-term efficacy of treprostinil was not fully confirmed in the
present study, treprostinil may be an alternative for PAH patients
unsuitable for epoprostenol therapy.
The associations of the use of
iloprost, beraprost, and selexipag with improved survival were not
observed in the present study. The effects of iloprost and beraprost on
improvement in PAH mortality have not been fully elucidated in clinical
trials.28,29 The 2022 ESC/ERS guidelines contain no
specific recommendations on the use of iloprost and beraprost in the
treatment algorithm for PAH.12 The guidelines
recommend add-on therapy of selexipag for PAH patients at intermediate
risk despite receiving the initial combination therapy of an ERA and a
PDE5i.10,12 The JADER database cannot distinguish
whether reported drugs are administered as initial or second-line add-on
therapy, while 71.1% of the selexipag users were also ERA and PDE5i
users in our study (data not shown). The PAH patients reported as
selexipag users in the JADER database may include many patients
resistant to initial therapy, which might influence our results.
The PAH patients treated with sildenafil were associated with a poor
prognosis compared with the patients treated with tadalafil or the
non-PDE5i users in the present study. The 2022 ESC/ERS guidelines
recommend initial combination therapy with tadalafil and an ERA for PAH
patients at low or intermediate risk based on the evidence from a
randomized controlled trial that compared the combination of tadalafil
and ambrisentan with each monotherapy.11,12 There is
little evidence for the superiority of tadalafil over sildenafil.
Tadalafil is a long-acting PDE5i with a half-life of 17.5 hours compared
with 4 hours for sildenafil.30 The longer half-life of
tadalafil contributes to stable hemodynamics in PAH patients, which may
be associated with our results. The superiority of riociguat, a soluble
guanylate cyclase stimulator, over PDE5is was not fully confirmed in the
present study. Although the effect of switching to riociguat from PDE5is
has been evaluated,31 initial use of riociguat in
newly diagnosed PAH patients should be assessed in future studies.
The PAH patients treated with ERAs, especially macitentan, had
significantly better survival than the non-user patients in the present
study. There is little information about the comparison of the
efficacies of ERAs, while different safety profiles have been observed
in clinical trials of each ERA.9,32,33 The use of
bosentan was related to dose-dependent increases in liver transaminases
and is not strongly recommended in the 2022 ESC/ERS guidelines compared
with other ERAs.12,33 Ambrisentan and macitentan had
low liver toxicity; however, the clinical trials of ambrisentan or
macitentan demonstrated an increased incidence of peripheral edema or
anemia, respectively.9,32 Although our data suggest
that macitentan was superior to other ERAs for long-term survival, their
safety profiles and baseline characteristics of PAH patients should be
considered in ERA selection.
Oxygen therapy for PAH patients was not associated with long-term
survival in our results. Ulrich et al. reported that oxygen therapy
improved exercise capacity in PAH patients.34 Oxygen
therapy is effective for PAH patients as a symptomatic treatment. In the
present study, the effect of warfarin on improvement in all-cause
mortality was observed in only non-epoprostenol users. Similar
tendencies were reported by Ogawa et al.16 These
findings suggest that PAH patients with epoprostenol therapy do not
benefit from anticoagulant therapy with warfarin. The relationship
between anticoagulant therapy with DOAC and epoprostenol was not
assessed in the present study because of the small sample size of DOAC
users. The use of tolvaptan was associated with a poor prognosis in our
study. Tolvaptan is used for the treatment of right ventricular failure
due to PAH in combination with traditional diuretics such as
furosemide.35 Thus, in the present study, PAH patients
using tolvaptan potentially have had a poor prognosis owing to the
comorbidity of right ventricular failures.
Our study has several limitations. Firstly, high censoring rates were
observed in the survival analyses. The JADER database does not include
follow-up data for each patient. We used the last-reported date of
adverse events as a censoring date to estimate the survival times of the
non-death PAH patients. This approach enables large-scale survival
analysis using a spontaneous adverse event reporting database, while
careful interpretation of the results is needed owing to the tendency to
consider non-death patients as censored. Secondly, our study did not
assess the effect of combination therapy of PAH-specific vasodilators.
We extracted all drug information related to the treatment of PAH from
the DRUG table; however, our analyses did not distinguish whether these
drugs were administered as mono or combination therapy. The present
study performed multivariate analyses to eliminate the influence of
covariates including concomitant use of other pulmonary vasodilators.
Our results potentially reflect the inherent efficacies of PAH-specific
vasodilators as monotherapy. Future clinical trials are required to
confirm the effect of combination therapy of PAH-specific vasodilators.
Thirdly, the JADER database contains several biases inherent to RWD such
as the potential of missing or wrongly entered data and the duplications
of patient data. Additionally, the JADER database is a passive reporting
system and is influenced by under-reporting which is the problem that
not all occurrences of adverse events are reported to the database.
Careful application of our data to clinical decision making is needed.