This protocol was registered with PROSPERO: CRD42020154661. We combined a meta-analysis of current literature and a retrospective database study from the University of Ottawa Heart Institute (UOHI) Pulmonary Hypertension clinic located in Ottawa, Ontario, Canada. The retrospective database study was approved by the Research Ethics Board at the Ottawa Hospital Research Institute.
Databases searched included PubMed, EMBASE, the Web of Science and EBSCO from Jan 2013 - Oct 2019. All English language studies were screened. See Appendix: eAppendix 1 for search terms. For the systematic review, all adult patients (≥ 18 years) diagnosed with PAH on a PDE5i, who were subsequently transitioned to riociguat were included. Any patients with prior treatment on riociguat was excluded. For the retrospective database study, all group 1 PAH patients, over the age of 18 who failed therapy with PDE5i (which is defined as any of: cardiac index (CI) ≤ 2.5, WHO FC ≥ III or 6MWTD less than predicted) and subsequently transitioned to riociguat were included. To be included, patients must have had a right heart catherization (RHC) with PVR of ≥ 300dyn-sec-cm− 5 or mean pulmonary arterial pressure (mPAP) of ≥ 25mmHg or pulmonary wedge pressure (PCWP) of ≤ 15mmHg. The patient’s clinic charts were reviewed from the periods between January 30 2013 to August 31, 2017 and analyzed.
We included randomized controlled trials and longitudinal observational studies (prospective or retrospective cohort studies and case control studies). We excluded any non-English articles, review articles, case reports, letters to the editor, conference abstracts and any irretrievable studies. For the systematic review: two authors (K.S. and T. K.) independently screened in a stepwise manner, the titles, then abstracts, and finally the full article of studies retrieved using the pre-specified search strategy (Appendix S1) to identify studies that met the inclusion criteria. Disagreements were resolved through discussion and if required, a senior author (G.C.) was consulted. A standardized form was used to extract data from the included studies. The data extracted included specific details about the study, study setting (e.g., clinical vs. community-based study), populations (participant demographics and baseline characteristics), details of the definition(s) of pulmonary hypertension, specific PDE5i and dosage prior to transition, the reason for transition, results post transition and any mortality or morbidity data.
For the retrospective database study, clinical data was collected on each eligible patient from a detailed chart review. These included: age, gender, medical record number, hemodynamic measurements obtained from RHC, echocardiogram, current and previous relevant co-morbidities (ie pulmonary embolism, rheumatological diseases, surgical pulmonary endarterectomies etc.). Furthermore, medication history, especially use of any calcium channel blocker use, diuretics, anticoagulation, endothelin receptor agonists, PDE5 inhibitor, prostaglandin analogues, and soluble guanylate cyclase was recorded.
The primary outcome was the change in the 6MWTD post-transition from PDE5i to riociguat. The secondary outcomes were the post transition changes in: (1) WHO FC; (2) Hemodynamic data: CI, mPAP, PCWP and B-type natriuretic peptide levels (BNP); (3) time to clinical worsening defined by hospitalizations and lung transplantation, any observed adverse side effects and mortality.
Study results were synthesized through tabulation and qualitative description. Data regarding the stated outcomes of the review (standardized mean difference (SMD) or mean difference (MD) for continuous) were extracted. Pooled results were presented as forest plots. We anticipated a limited scope for meta-analysis due to high methodological, clinical and statistical heterogeneity between studies. Thus, results were combined using a random effect model. Heterogeneity was assessed using the Cochran’s Q-test of (residual) heterogeneity; a p-value of less than 0.10 was considered to represent evidence of heterogeneity. To quantify inconsistency, we used an I² statistic estimate (percent), which assesses how much of the total variability in the effect size estimates can be attributed to heterogeneity between studies. We considered an I² value greater than 50% to be indicative of substantial heterogeneity. The R software was used for data analysis. For the retrospective database study, descriptive statistics were used to characterize the sample of interest. The Wilcoxon signed-rank test was used to compare continuous outcomes; the McNemar's Chi-squared test with continuity correction was used for categorical outcomes. Generalized linear mixed regression models were used to adjust for confounders. Given the small sample size, only age was considered in the statistical model as a potential confounder.
We planned for one reviewer (K.S.) to assess for the risk of bias for each study using either the Cochrane Handbook for Systematic Reviews for randomized controlled trials or the Newcastle-Ottawa scale (NOS) for observational trials [10, 11]. Results were corroborated by a second reviewer (T. K.). There were no randomized control trials and thus only the Newcastle-Ottawa scale was used.
The NOS uses a 'star system' to evaluate studies on three domains: the selection of the study groups; the comparability of the groups; and the ascertainment of either the exposure or outcome of interest for case-control or cohort studies, respectively [11]. The scale evaluates the domains of selection, comparability and outcome or exposure [11]. Although there is no formalized process of grading the quality of studies per the NOS, previous studies have employed a score of 7 or higher to be considered high quality studies [12].