Article Text

Original research
Effectiveness of angiotensin-neprilysin inhibitor treatment versus renin-angiotensin system blockade in older adults with heart failure in clinical care
  1. Rishi J Desai1,
  2. Elisabetta Patorno1,
  3. Muthiah Vaduganathan2,
  4. Mufaddal Mahesri1,
  5. Kristyn Chin1,
  6. Raisa Levin1,
  7. Scott D Solomon2,
  8. Sebastian Schneeweiss1
  1. 1 Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Boston, Massachusetts, USA
  2. 2 Heart and Vascular Center, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA
  1. Correspondence to Dr Rishi J Desai, Brigham and Women's Hospital, Boston, MA 02115, USA; rdesai{at}bwh.harvard.edu

Abstract

Objective To evaluate the effectiveness of angiotensin receptor-neprilysin inhibitor (ARNI) versus renin-angiotensin system (RAS) blockade alone in older adults with heart failure with reduced ejection fraction (HFrEF).

Methods We conducted a cohort study using US Medicare fee-for-service claims data (2014–2017). Patients with HFrEF ≥65 years were identified in two cohorts: (1) initiators of ARNI or RAS blockade alone (ACE inhibitor, ACEI; or angiotensin receptor blocker, ARB) and (2) switchers from an ACEI to either ARNI or ARB. HR with 95% CI from Cox proportional hazard regression and 1-year restricted mean survival time (RMST) difference with 95% CI were calculated for a composite outcome of time to first worsening heart failure event or all-cause mortality after adjustment for 71 pre-exposure characteristics through propensity score fine-stratification weighting. All analyses of initiator and switcher cohorts were conducted separately and then combined using fixed effects.

Results 51 208 patients with a mean age of 76 years were included, with 16 193 in the ARNI group. Adjusted HRs comparing ARNI with RAS blockade alone were 0.92 (95% CI 0.84 to 1.00) among initiators and 0.79 (95% CI 0.74 to 0.85) among switchers, with a combined estimate of 0.84 (95% CI 0.80 to 0.89). Adjusted 1-year RMST difference (95% CI) was 4 days in the initiator cohort (−1 to 9) and 12 days (8 to 17) in the switcher cohort, resulting in a pooled estimate of 9 days (6 to 12) favouring ARNI.

Conclusion ARNI treatment was associated with lower risk of a composite effectiveness endpoint compared with RAS blockade alone in older adults with HFrEF.

  • heart failure

Data availability statement

Data may be obtained from a third party and are not publicly available. Patient-level data are not available for sharing due to restrictions imposed under data use agreement by the Centers for Medicare and Medicaid Services. All aggregate-level data are presented in the manuscript and supplemental content. Additional information is available from the corresponding author at rdesai@bwh.harvard.edu.

Statistics from Altmetric.com

Introduction

An increasing prevalence of heart failure (HF) due to population-level ageing contributes substantially to morbidity and mortality, posing a significant challenge to healthcare systems worldwide.1 2 Important advances have been made in pharmacological treatment of heart failure with reduced ejection fraction (HFrEF) in the last decade, including treatment with an angiotensin receptor-neprilysin inhibitor (ARNI), a combination product of sacubitril and valsartan. In the pivotal randomised controlled trial (RCT) of patients with HFrEF, PARADIGM-HF, treatment with ARNI demonstrated superiority compared with renin-angiotensin system (RAS) blockade alone.3 Clinical practice guidelines recommend treatment with ARNI in patients with HFrEF who can tolerate RAS blockade for further reduction in morbidity and mortality.4

A well-recognised limitation of RCTs of cardiovascular treatments is under-representation of older adults, patients with multimorbidity, women and racial/ethnic minority populations.5 6 In PARADIGM-HF, the average age of included patients was 64 years, only 21% were women and only 5% were black.3 The inclusion of a run-in period (enrolment of only participants who tolerated target doses of both study drugs) in PARADIGM-HF may further limit generalisability to ‘real world’ clinical practice.7 8 Treatment adherence in clinical practice may also differ substantially from that observed in clinical trials due to costs, regimen complexity or limited health literacy.9 10 Finally, as only patients taking RAS inhibitors were considered for participation in PARADIGM-HF, little is known about the use of ARNI among treatment-naïve participants.11

Therefore, rigorous evaluation of the effectiveness of ARNI treatment in clinical practice is critical to complement RCTs. Medicare beneficiaries constitute a population of special interest because majority of patients with HF in the USA receive health insurance through the Medicare programme.12 We designed this cohort study using Medicare claims data to separately examine the effectiveness of two pathways to ARNI introduction in clinical practice: new initiation (among previously RAS inhibitor-naïve patients) or switching from prior RAS inhibitors.

Methods

Data sources

We used Medicare fee-for-service claims data from July 2014 through 2017. Medicare part A (hospitalisations), part B (medical services) and part D (prescription medications) claims are available for research purposes through the Centers for Medicare and Medicaid Services (CMS). A signed data use agreement with the CMS was available.

Study design

We designed a new user cohort study employing an active comparator group.13 14 Two distinct study cohorts were identified: (1) initiator cohort: patients newly initiating ARNI or ACE inhibitor (ACEI)/angiotensin receptor blocker (ARB) without use of any agent from either drug class in the prior 12 months were included in this cohort and the date of the first dispensing of the study drug of interest was defined as the cohort entry date (online supplemental appendix figure 1A); and (2) switcher cohort: patients switching from ACEI to ARNI or to ARB without use of ARNI or ARB in the prior 12 months were included in this cohort and the date of switch to ARNI or ARB was defined as cohort entry date (online supplemental appendix figure 1B). Focusing comparative evaluations among patients newly initiating (new user designs) or switching to different agents from a common baseline treatment (new switcher designs)14 avoids issues commonly encountered in studies of prevalent users, including depletion of susceptible patients and potential for adjustment for causal intermediates affected by treatment history.

Supplemental material

In both cohorts, a 12-month continuous enrolment period in Medicare parts A, B and D immediately preceding the cohort entry date was required to ensure availability of sufficient baseline data for assessment of inclusion criteria and covariates. We restricted the study population to patients with recorded diagnosis of HF within 6 months prior to the cohort entry date, age ≥65 years, no nursing home stay in the 12 months prior to cohort entry, complete recording of gender and no history of heart transplant. Initiation of ARNI was assumed to be a sufficient indicator of HFrEF in the treatment group. Since ACEI and ARB are also commonly used in patients with heart failure with preserved ejection fraction (HFpEF), we additionally applied a claims-based algorithm15 to restrict the reference group to patients with HFrEF, which is reported to have a positive predicted value of 73% (95% CI 68% to 78%)15 and shown to classify patients into HFrEF and HFpEF populations that are more similar to cohorts described in previous epidemiological studies with respect to population characteristics and mortality incidence than an approach that only relies on diagnosis codes.16

Drug exposure and follow-up

Follow-up began on the day following the cohort entry date and continued until the earliest of outcome occurrence, Medicare disenrolment or 31 December 2017. In the primary as-treated follow-up approach, follow-up time was censored on switch between exposure and reference groups or discontinuation of the index treatment. Discontinuation was defined based on no record of a subsequent prescription of the index medication class for 30 days, after accounting for the number of days of exposure provided by the most recent prescription. In a secondary intent-to-treat follow-up approach, follow-up was continued for a maximum of 12 months regardless of subsequent switching or discontinuation.

Outcomes

The primary outcome of interest was a composite endpoint of time to worsening HF (defined as a hospitalisation with HF as the primary diagnosis or treatment with intravenous diuretics (furosemide, bumetanide, torsemide) in outpatient setting)17 or all-cause mortality. Individual components of the primary composite endpoint were evaluated as secondary endpoints.

Covariates

The following sets of covariates were identified: (1) demographics including age, gender, race, proxies of socioeconomic status (SES) (low-income subsidy receipt and a composite index18); (2) other HF treatments (including mineralocorticoid receptor antagonists (MRA), beta-blockers, diuretics, cardiac resynchronisation therapy, implantable cardioverter-defibrillator) and HF hospitalisation history during 12 months of baseline; (3) comorbid conditions; (4) comedications; (5) healthcare utilisation factors (including number of non-HF hospitalisations, number of emergency department visits, access to cardiologists) as markers of overall health and healthcare access; and (6) healthy behaviour markers, including use of screening services and vaccinations. A full list of 71 covariates is provided in online supplemental appendix table 1.

Statistical analysis

We used propensity score (PS)-based fine-stratification and weighting for confounding adjustment.19 20 PS was calculated as the predicted probability of initiating ARNI treatment conditional on pre-exposure covariates described above in logistic regression models. After trimming the non-overlapping portions of the PS distributions to exclude non-comparable patients, 50 strata were created based on the distribution of PS in the ARNI group, and reference patients were weighted proportional to the distribution of ARNI patients in their respective stratum.19 To demonstrate balance achieved in each individual covariate after weighting, % standardised mean differences were reported and postweighting c-statistics were reported as a measure of overall balance.21 22

In the weighted sample, cumulative incidence estimates stratified by treatment group for the primary composite endpoint and secondary all-cause mortality endpoint were calculated using the Kaplan-Meier method. For the secondary outcome of worsening HF, cumulative incidence functions were used to appropriately handle competing risk of all-cause mortality.23 Weighted Cox proportional hazard regression models provided HR with 95% CI calculated based on robust variance estimators to account for weighting.20 Additionally, we calculated 1-year restricted mean survival time (RMST) differences with 95% CI calculated using 10 000 non-parametric bootstraps.24 All-cause mortality was treated as a competing event when modelling worsening HF as secondary endpoint.23 25 All analyses were conducted separately for initiator and switcher cohorts. We used a fixed effects meta-analysis to combine estimates.26 27

Subgroup and sensitivity analyses

The following subgroups were prespecified: age (65–74, 75–84 and ≥85 years), gender, race, baseline diabetes, baseline atrial fibrillation, baseline chronic kidney disease, baseline MRA treatment and prior HF hospitalisation within the previous 12 months. In each subgroup, we re-estimated the PS and conducted stratification and weighting to address confounding because the PS calculated in the whole cohort may be inadequate to ensure covariate balance between treatment and reference groups within each subgroup stratum.28

In a sensitivity analysis to address the possibility of temporal confounding due to changing practice patterns, we calculated calendar quarter-specific PS29 and conducted 1:1 PS matching of patients initiating treatment with ARNI or reference treatment within the same calendar quarter. Second, we implemented a high-dimensional PS adjustment algorithm to adjust for 200 empirically identified confounding variables prioritised based on the Bross bias formula in addition to all prespecified covariates,30 which has been shown to further reduce confounding through proxy variable adjustment.31 Third, in patients treated with ACEI/ARB, we used the standard approach of identifying HFrEF based on diagnosis codes for systolic HF32 33 as an alternative to the model-based approach used for primary analysis. Fourth, we symmetrically applied the claims-based algorithm,15 which was only applied to the reference group of ACEI/ARB treated patients to identify HFrEF in the main analysis, to both exposure groups. Further, we tested the impact of residual confounding by SES, which is not directly measured in Medicare records, on our estimates using the ‘array’ approach (fully described in online supplemental appendix figure 6).34

Patient involvement

No patients were involved in setting the research question or the outcome measures, nor were they involved in developing plans for design or implementation of the study.

Results

Study population

A total of 51 208 patients qualified for the analysis, of whom 29 196 were included in the initiator cohort and 22 012 in the switcher cohort (online supplemental appendix figure 2). The average age (SD) was 76 (7) years among initiators and 75 (7) among switchers. Women comprised 33.5% and 31.5% and black patients comprised 11.3% and 10.9% of the initiator and switcher cohorts, respectively. A total of 16 913 patients on ARNI treatment were included across the two cohorts (4310 initiators and 12 603 switchers). Table 1 summarises the key patient characteristics stratified by treatment in both cohorts (online supplemental appendix table 1 contains all characteristics). Online supplemental appendix table 2 provides the study medication doses for the included patients.

Table 1

Select patient characteristics in patients with heart failure with reduced ejection fraction treated with angiotensin receptor-neprilysin inhibitor or renin-angiotensin system blockade alone

After PS weighting, balance was achieved across treatment groups in all patient characteristics, as demonstrated by the average absolute % standardised mean difference close to 0 (0.5 and 1.1 in the initiator and switcher cohorts, respectively) and postweighting c-statistics close to 0.5 (0.52 and 0.53 in the initiator and switcher cohorts, respectively). Online supplemental appendix figure 3 contains the PS distribution before and after weighting, and online supplemental appendix figure 4 demonstrates the distribution of PS weights.

Incidence of HF outcomes

The 1-year cumulative incidence of the composite endpoint under as-treated follow-up was 27.2% in the ARNI group and 26.9% in the reference group among initiators and 23.8% and 29.7% among switchers after PS weighting (figure 1). Among individual components (figure 2), the 1-year cumulative incidence of worsening HF was 21.0% in the ARNI group and 19.7% in the reference group in the initiator cohort and 19.6% and 23.8%, respectively, in the switcher cohort. For all-cause mortality, the 1-year cumulative incidence was 9.3% in the ARNI group and 11.0% in the reference group in the initiator cohort and 7.6% and 11.5%, respectively, in the switcher cohort.

Figure 1

Cumulative incidence of composite endpoint of worsening heart failure and all-cause mortality after propensity score weighting in patients treated with angiotensin receptor-neprilysin inhibitor (ARNI) or renin-angiotensin system blockade alone. Numbers in the table below the plot represent populations at risk under as-treated follow-up. ARB, angiotensin receptor blocker.

Figure 2

Cumulative incidence of individual endpoints of worsening heart failure (HF) and all-cause mortality after propensity score weighting in patients treated with angiotensin receptor-neprilysin inhibitor (ARNI) or renin-angiotensin system blockade alone. Numbers in the table below the plot represent populations at risk under as-treated follow-up. ARB, angiotensin receptor blocker.

Comparative effectiveness estimates

Table 2 presents the results from the primary as-treated analysis. In the initiator cohort, PS-weighted HRs (95% CI) comparing ARNI with RAS blockade alone were 0.92 (0.84 to 1.00) for the composite endpoint, 0.95 (0.86 to 1.05) for worsening HF and 0.81 (0.69 to 0.94) for all-cause mortality. In the switcher cohort, PS-weighted HRs (95% CI) comparing ARNI with ARB were 0.79 (0.74 to 0.85) for the composite endpoint, 0.81 (0.75 to 0.88) for worsening HF and 0.69 (0.62 to 0.78) for all-cause mortality. Combined HRs (95% CI) across the two cohorts were 0.84 (0.80 to 0.89) for the composite endpoint, 0.86 (0.81 to 0.91) for worsening HF and 0.73 (0.67 to 0.80) for all-cause mortality. RMST at 1 year was 4 days higher in the ARNI group in the initiator cohort (95% CI 1 day lower to 9 days higher) and 12 days higher in the switcher cohort (95% CI 8 to 17 days higher), resulting in a pooled estimate of 9-day higher (95% 6 to 12 days higher). The results were consistent when using an intent-to-treat follow-up approach: combined HR 0.90 (95% CI 0.86 to 0.94) and pooled 1-year RMST 7 days higher (95% CI 4 to 10 days higher) for the composite endpoint (online supplemental appendix table 3).

Table 2

Comparative outcomes in patients with heart failure with reduced ejection fraction treated with angiotensin receptor-neprilysin inhibitor or renin-angiotensin system blockade alone under as-treated follow-up

Subgroup and sensitivity analyses

We did not find any evidence of treatment effect heterogeneity across prespecified subgroups; however, due to the small sample size, the upper bound of the CI included the null value for certain subgroups, including patients older than 85 years and black patients (figure 3 and online supplemental appendix figure 5). The results from all sensitivity analyses where we varied our study assumptions or analytic approach were consistent with the primary analysis (figure 4). In the sensitivity analysis for residual confounding by SES, we observed that this could have altered our findings if the proportion of lower SES patients was more than twofold higher in the reference group (online supplemental appendix figure 6).

Figure 3

Subgroup analyses. Plotted are pooled estimates under as-treated follow-up across initiator and switcher cohorts after propensity score weighting. Numbers indicate total cohort size included for estimation of treatment effect within each subgroup across initiator and switcher cohorts. Numbers do not add up to total because propensity scores were constructed within each subgroup and patients in non-overlapping regions of the propensity scores were trimmed to exclude non-comparable patients within each subgroup. ARB, angiotensin receptor blocker; ARNI, angiotensin receptor-neprilysin inhibitor; HF, heart failure; MRA, mineralocorticoid receptor antagonist.

Figure 4

Sensitivity analyses. Plotted are propensity score-weighted estimates under as-treated follow-up. In SA4, model-based HFrEF identification was applied to both ARNI and ACEI/ARB treated patients. ARB, angiotensin receptor blocker; ARNI, angiotensin receptor-neprilysin inhibitor; HF, heart failure; HFrEF, heart failure with reduced ejection fraction; PS, propensity score; SA, sensitivity analysis.

Discussion

In this population-based cohort study of 51 208 older adults with HFrEF, treatment with ARNI was associated with a reduction in the risk of a composite effectiveness endpoint consisting of worsening HF and all-cause mortality compared with RAS blockade alone. Observed treatment effects were consistent across a range of subgroups and robust across multiple sensitivity analyses.

We observed treatment effect estimates favouring ARNI versus RAS blockade among initiators with no prior RAS blockade use, which is a population likely corresponding to recent-onset HFrEF, as well as among switchers from prior ACEI treatment, which is a population likely corresponding to long-standing HFrEF. The estimated treatment benefit had lower magnitude among initiators (HR ARNI vs ACEI/ARB 0.92, 95% CI 0.84 to 1.00) than switchers (HR ARNI vs ARB 0.79, 95% CI 0.74 to 0.85) for the composite endpoint of worsening HF and all-cause mortality. A possible explanation of this apparent smaller effect size could be that patients with recent-onset HFrEF may be more responsive to any decongestion therapy, as recently suggested in a subanalysis of the PIONEER-HF trial based on greater reduction in NT-terminal pro B-type natriuretic peptide concentration in patients with de novo HFrEF compared with patients with worsening chronic HFrEF regardless of treatment assignment.11 Alternatively, limitations related to our study design features may also partly explain this observation. For instance, since clinical practice guidelines have traditionally only recommended treatment with ARNI in patients who can tolerate RAS blockade (consistent with the design of PARADIGM-HF),4 patients selected for ARNI initiation without any RAS blockade history in routine practice may represent a unique group of patients with HF who may not be directly comparable with RAS blockade initiators with respect to unmeasured factors such as severity of HF, duration of HF or frailty/comorbid disease burden. Among the currently available RCTs evaluating comparative effectiveness of ARNI,3 11 PARADIGM-HF did not include patients without history of RAS blockade, while PIONEER-HF included such patients but had important limitations, including a limited sample size (881 total patients, 459 without history of RAS blockade) and a short follow-up time of 8 weeks. Given the nuances surrounding this clinical question and limited knowledge from currently available studies, future clinical investigations focused on comparing the effectiveness of de novo ARNI initiation based on prospective RCTs or other observational registries would be useful to confirm observations from our study.

In a population of patients with a history of RAS blockade (switcher cohort), which is similar to the population enrolled in the PARADIGM trial, we observed very similar results reflecting the effectiveness of ARNI in the overall cohort (HR of 0.80 (95% CI 0.73 to 0.87) vs enalapril for a composite efficacy endpoint consisting of HF hospitalisation or cardiovascular mortality in PARADIGM vs HR of 0.79 (95% CI 0.74 to 0.85) vs ARBs here), with no substantial heterogeneity by age or race.3 35 36 It is important to note that in this study we compared patients switching from an ACEI to ARB or ARNI in a ‘new switcher design’ to avoid biases related to prevalent user design, including depletion of susceptible patients and potential for adjustment for causal intermediates affected by treatment history that threaten the validity of non-randomised investigations.13 14 It is plausible that switching from ACEI to ARB may occur in situations of adverse drug reaction or intolerance (such as ACEI-related cough), while switching from an ACEI to ARNI may more frequently occur due to progression of disease or ongoing symptoms due to treatment failure on maximally tolerated ACEI. To the extent that the differences in underlying reasoning for treatment switch from ACEI to ARB or ARNI are not captured by a broad range of variables we adjusted for, including history of HF hospitalisations and other HF treatments, our results may be subject to confounding bias. Our estimates combining initiator and switcher cohorts are consistent with a previous cohort study by Tan et al.32

Despite the evidence of superior clinical efficacy of ARNI over RAS blockade in RCTs, utilisation has remained low, with previous studies reporting fewer than 15% of eligible patients initiating treatment with ARNI in routine care.9 37 38 A potential barrier to widespread use may be the high out-of-pocket cost faced by patients for ARNI versus ACEI/ARB, which was estimated to be in excess of $1400 annually in a recent study.39 Other barriers may include inconsistent formulary coverage and prior authorisation requirements,40 which may be reflective of uncertainty regarding the cost-effectiveness of ARNI treatment for routine care patient populations from a payer perspective due to generalisability concerns with PARADIGM-HF results.41 Our results from a cohort of patients with substantially older average age (76 years vs 64 years in PARADIGM-HF) and notably higher representation of women (33% vs 21%) and black patients (11% vs 5%) strengthen the evidentiary base supporting the broad effectiveness of ARNI in routine care patient populations. Further, our results may also address concerns regarding limited generalisability of PARADIGM-HF results owing to exclusion of patients who could not tolerate target doses of study medications during the run-in period.7 42

Several important limitations of the current investigation deserve mention. First, assessment of HFrEF as well as all comorbid conditions was solely based on administrative claims; therefore, misclassification is possible. Similarly, we did not have access to cause of death information and used administratively recorded all-cause mortality as our endpoint. Second, residual confounding due to unmeasured factors is always an important concern in observational studies. Finally, comparative effectiveness is one dimension to consider while making a treatment decision; an equally important dimension is comparative safety, which we did not study. Future research focused on comparative safety of ARNI would be important.

In conclusion, we noted robust effectiveness of ARNI compared with RAS blockade alone in a cohort of older patients with HFrEF in routine care. Taken together with evidence from other studies, our results support integration of ARNI as a cornerstone therapy for a broad spectrum of patients with HFrEF.

Key messages

What is already known on this subject?

  • In the pivotal randomised controlled trial of patients with heart failure with reduced ejection fraction, PARADIGM-HF, treatment with angiotensin receptor-neprilysin inhibitor (ARNI) demonstrated superiority compared with renin-angiotensin system (RAS) blockade alone.

  • Due to under-representation of older patients, women and racial minority, as well as exclusion of patients who could not tolerate target doses of study medications during the run-in period, questions remain regarding the generalisability of the findings from PARADIGM-HF.

What might this study add?

  • Among 51 208 patients from the US Medicare programme included in the study with an average age of 76 years, treatment with ARNI was associated with a 16% lower risk of a composite effectiveness endpoint consisting of worsening heart failure and all-cause mortality compared with RAS blockade alone.

How might this impact on clinical practice?

  • Taken together with evidence from other studies, our results support integration of ARNI as a cornerstone therapy for a broad spectrum of patients with heart failure with reduced ejection fraction seen in routine care.

Data availability statement

Data may be obtained from a third party and are not publicly available. Patient-level data are not available for sharing due to restrictions imposed under data use agreement by the Centers for Medicare and Medicaid Services. All aggregate-level data are presented in the manuscript and supplemental content. Additional information is available from the corresponding author at rdesai@bwh.harvard.edu.

Ethics statements

Ethics approval

The Brigham and Women’s Hospital’s Institutional Review Board approved this study.

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • Twitter @Rishidesai11

  • Contributors Planning: RJD, EP, SS. Conduct: RJD, RL, MM, KC. Interpretation: all authors. Writing: all authors. RJD takes the responsibility for the overall content as guarantor of this article.

  • Funding This study was funded by internal sources of the Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School.

  • Competing interests RJD reports serving as principal investigator on research grants to Brigham and Women’s Hospital from Vertex, Novartis and Bayer. EP was supported by a career development grant (K08AG055670) from the National Institute on Aging. She is investigator of an investigator-initiated grant to the Brigham and Women’s Hospital from Boehringer Ingelheim, not related to the topic of the submitted work. MV is supported by a KL2/Catalyst Medical Research Investigator Training award from Harvard Catalyst (NIH/NCATS Award UL 1TR002541), receives research grant support from Amgen, serves on advisory boards for Amgen, AstraZeneca, Baxter Healthcare, Bayer, Boehringer Ingelheim, Cytokinetics and Relypsa, and serves on the clinical endpoints committees of studies sponsored by Galmed, Novartis and the NIH. SDS has received research grants from Alnylam, Amgen, AstraZeneca, Bellerophon, Celladon, Gilead, GlaxoSmithKline, Ionis Pharmaceuticals, Lone Star Heart, Mesoblast, MyoKardia, NIH/NHLBI, Novartis, Sanofi Pasteur and Theracos, and has consulted for Alnylam, Amgen, AstraZeneca, Bayer, Bristol-Myers Squibb, Corvia, Gilead, GlaxoSmithKline, Ironwood, Merck, Novartis, Pfizer, Takeda and Theracos. SS is co-principal investigator of investigator-initiated grants to the Brigham and Women’s Hospital from Boehringer Ingelheim, unrelated to the topic of this study. He is a consultant to Aetion, a software manufacturer of which he owns equity. His interests were declared, reviewed and approved by the Brigham and Women’s Hospital and Partners HealthCare System in accordance with their institutional compliance policies.

  • Provenance and peer review Not commissioned; internally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles