More women than men currently live with heart failure (HF) in the US (1); this imbalance is projected to persist over the next decade (2). Sex and gender influence biological drivers, etiologies, background epidemiology, markers of risk, and care quality in HF. To bolster enrollment of women in clinical trials, the United States Food and Drug Administration (FDA) has put forth a series of targeted guidance documents and regulations since the early 1990s. The FDA now requires all sponsors to include efficacy and safety data in women in applications for regulatory approval of drugs and devices and disclose any meaningful sex-based differences in product labelling (Online Ref. 1). The National Institutes of Health (NIH) have similarly supported adequate enrollment and reporting of sex-based subgroup information in all NIH-funded phase 3 clinical trials (Online Ref. 2).
Heart Failure Collaboratory: Improving Representation in Contemporary Trials
Since the introduction of these policies, the representation of women in contemporary clinical trials of HF has slowly improved over time (3). However, little is known about the completeness of reported sex/gender subgroup information or whether sex/gender influence treatment effects to investigational therapies. As such, we conducted a systematic review focused on sex/gender subgroups in all published clinical trials of HF from 2001 to 2016. The work is embedded within a larger multi-stakeholder effort, the “Heart Failure Collaboratory”, which includes patients, clinical trialists, industry sponsors, payers, and representatives from the US FDA, NIH, and Centers for Medicare and Medicaid Services.
Systematic Query of Sex-Specific Reporting
We identified published randomized clinical trials of therapeutic interventions in HF from 2001-2016 using a systematic search via PubMed/MEDLINE (search terms: publication year, “heart failure”, “trial*”, and “randomized”) and ClincialTrials.gov (search limits: adult, interventional, phase II-IV, “heart failure”). We only included primary trial publications of phase 2-4 trials of adult populations. We set a sample size threshold (>400 patients) for the total enrolled population to identify trials more likely to inform clinical practice. All procedures and reporting were by PRISMA guidelines. Two independent reviewers screened all published reports for relevance and eligibility.
Key data extracted included HF categories by ejection fraction, acute vs. chronic HF, trial phase, intervention, primary endpoint, funding source, and region of enrollment. Funding source was based on ClinicalTrials.gov entry and categorized as (1) industry; (2) government; or (3) university or other non-profit or non-federal organizations. We also separately identified studies specifically supported by the NIH. Enrollment rates were estimated based on study duration and expressed as patients per site per month.
Sex/gender-specific subgroup information was searched in all primary publications and associated supplementary materials. The proportion of total trials reporting sex/gender-specific analyses were evaluated across 4 four-year periods based on publication date (2001-2004, 2005-2008, 2009-2012, and 2013-2016). We separately assessed temporal trends in phase 3 or 4 trials (when trial phase was known/reported). In studies with reported subgroup analyses by sex/gender, trials were identified that had significant, unadjusted sex/gender interaction analyses for the primary endpoint (P<0.05). All computations were performed with IBM SPSS 23 (IBM Corporation).
Current State of Sex-Specific Reporting
In total, we screened 5,488 studies and 118 published trials met inclusion criteria and were included. Overall, 81 (69%) trials reported outcome data by sex/gender. The proportion of trials reporting sex/gender-specific analyses increased from 57.1% in 2001-2004 to 71.9% in 2013-2016; Table 1.
Table 1.
Characteristics of 118 contemporary heart failure trials by reporting status of sex/gender-based subgroup data
| Trials without Sex/Gender-Based Subgroup Reporting (n=37, 31%) | Trials with Sex/Gender-Based Subgroup Reporting (n=81, 69%) | P-value | ||
|---|---|---|---|---|
| Year | 2001-2004 | 9 (42.9) | 12 (57.1) | 0.66 |
| 2005-2008 | 9 (29) | 22 (71) | ||
| 2009-2012 | 10 (29.4) | 24 (70.6) | ||
| 2013-2016 | 9 (28.1) | 23 (71.9) | ||
| Trial Characteristics | Participants per trial | 512 (458-921) | 1325 (809-2426) | <0.001 |
| Proportion of women per trial | 27 (21.3-32) | 26.5 (20.7-34) | 0.75 | |
| Sites per trial | 50 (9-102) | 98 (42-244) | <0.001 | |
| Duration of trial (years) | 2.3 (1.3-3.4) | 2.9 (2-4) | 0.014 | |
| Number of countries per trial | 1 (1-3) | 9 (2-20) | <0.001 | |
| Enrollment rate (patients/site/month) | 0.8 (0.4-2.2) | 0.4 (0.2-0.7) | 0.004 | |
| Trial Phase | Phase 2 | 4 (44) | 5 (56) | 0.09 |
| Phase 3 or 4 | 13 (24) | 41 (76) | ||
| Not Reported | 20 (36) | 35 (64) | ||
| Heart Failure Type | Heart failure with reduced ejection fraction | 29 (30.9) | 65 (69.1) | 0.28 |
| Heart failure with preserved ejection fraction | 0 (0) | 4(100) | ||
| Heart failure (unspecified ejection fraction) | 8 (40) | 12 (60) | ||
| Primary Sponsor/Funding Body | Government | 7 (30.4) | 16 (69.6) | |
| University/Non-for-profit Organizations | 8 (42.1) | 11 (57.9) | 0.44 | |
| Industry | 20 (27) | 54 (73) | ||
| NIH-Based Funding | NIH | 2 (16.7) | 10 (83.3) | 0.25 |
| Non-NIH | 33 (32.7) | 68 (67.3) | ||
| Intervention | Invasive | 8 (27.6) | 21 (72.4) | 0.18 |
| Medication | 17 (27) | 46 (73) | ||
| Other | 12 (46.2) | 14 (53.8) | ||
| Primary Endpoint | Mortality | 16 (19.3) | 67 (80.7) | <0.001 |
| Intermediate | 13 (59.1) | 9 (40.9) | ||
| Surrogate | 8 (61.5) | 5 (38.5) | ||
| Primary Region of Enrollment | Multiregional | 9 (15.8) | 48 (84.2) | 0.001 |
| North America | 16 (55.2) | 13 (44.8) | ||
| Western Europe | 9 (47.4) | 10 (52.6) | ||
| Outside North America or Western Europe | 2 (33.3) | 4 (66.7) | ||
Abbreviations: NIH = National Institutes of Health
The proportion of women enrolled was 27% irrespective of whether the trial reported or did not report subgroup information (P=0.75). Trials with reported subgroup information were on average larger (median 1,325 [interquartile range 809-2,426] vs 512 [458-921] participants; P<0.001) with a greater number of enrolling sites (median 98 [42-244] vs. 50 [9-102]; P<0.001). Trials reporting subgroup data were commonly global experiences with primary endpoints inclusive of mortality (P≤0.001 for both). Among phase 3 or 4 trials, 76% reported sex/gender-specific subgroup data. The proportion of sex/gender subgroup reporting in phase 3 or 4 trials increased significantly from 50% in 2001-2004 to 82.4% in 2013-2016. Among NIH-sponsored studies (n=12), 10 (83%) reported sex/gender subgroup data.
Treatment Effects by Sex/Gender
Of the 81 trials that reported subgroup analyses, 77 (95%) reported no significant treatment differences by sex/gender with respect to the primary endpoint. Of the 4 trials (5%) (Online Refs. 3–6) which did report significant treatment differences by sex/gender, 3 evaluated endpoints which included all-cause mortality; Table 2. Sex/gender-based subgroup analysis was among multiple subgroups tested, and pre-specified in 3 of the 4 trials. Only 1 of the 4 trials suggesting sex/gender treatment differences had a primary endpoint suggesting benefit. When treatment differences by sex/gender were identified, most reported more favorable therapeutic responses in women. One trial with a neutral overall trial result demonstrated a signal for harm with the investigational therapy vs. control in women, but not men (Online Ref. 3).
Table 2.
Four of 81 trials reporting differential treatment response to investigational therapies in randomized clinical trials of heart failure from 2001 to 2016
| Trial | Ref | ClinicalTrials.gov Identifier | Year of Publication | Total Randomized | # of Women | Intervention | Primary Endpoint | # of Tested Subgroups | Testing by Sex/Gender | Overall Treatment Result | Treatment Response Favored |
|---|---|---|---|---|---|---|---|---|---|---|---|
| INNOVATE-HF | Online Ref. 3 | NCT01303718 | 2016 | 707 | 149 | Vagal nerve stimulation vs. Control | All-cause mortality + Worsening HF | 6 | Pre-specified | Neutral | Men |
| SMART-AV | Online Ref. 4 | NCT00677014 | 2010 | 980 | 314 | CRT AV Delay Methods | LV end-systolic volume | 5 | Post-hoc | Neutral | Women |
| Tele-HF | Online Ref. 5 | NCT00303212 | 2010 | 1,653 | 695 | Telemonitoring vs. Usual care | All-cause mortality + All-cause readmission | 6 | Pre-specified | Neutral | Women |
| MADIT-CRT | Online Ref. 6 | NCT00180271 | 2009 | 1,820 | 453 | CRT-ICD vs. ICD alone | All-cause mortality + Non-fatal HF | 7 | Pre-specified | Benefit | Women |
Only trials with reported subgroup analyses by sex/gender were screened. These 4 trials were selected based on significant, unadjusted sex/gender interaction analyses for the primary endpoint (P<0.05).
Abbreviations: AV = atrioventricular; CRT = Cardiac Resynchronization Therapy; HF = heart failure; ICD = implantable cardioverter-defibrillator; INNOVATE-HF = Increase in Vagal Tone in CHF; MADIT-CRT = Multicenter Automatic Defibrillator Implantation Trial-CRT; SMART-AV = SmartDelay Determined AV Optimization: A Comparison to Other AV Delay Methods Used in CRT; Tele-HF = Telemonitoring to Improve Heart Failure Outcomes
A Compelling Need for Sex/Gender-Specific Data
This comprehensive systematic review demonstrates that a quarter of large, phase 2-4 contemporary HF trials do not adequately report the potential variation in treatment response by sex/gender. Reassuringly, the fraction without subgroup data has gradually decreased over time (<20% in trials published 2013-2016) and over 80% of NIH-sponsored studies reported sex/gender-specific analyses. A minority (5%) of trials identified significant variation in treatment responses to investigational therapies in HF between men and women; most of these trials reported more favorable therapeutic responses in women.
The ongoing collection and complete reporting of sex/gender-specific data are important to identify safe and effective therapies in women for several reasons. First, the population of women with HF is growing in the US and worldwide (2). Second, although there have been slow declines in overall rates of hospitalization for HF in older US adults over the last decade, these favorable trends appear less apparent in certain sex- and race-specific groups (4). Third, sex-based differences may present biologically plausible reasons for differential treatment responses in HF. For instance, angioedema related to angiotensin-converting enzyme inhibitors, albeit rare, may occur more frequently in women compared with men due to sex-dependent variation in a genetic polymorphism that influences bradykinin (5). Finally, while other key demographic characteristics, such as race/ethnicity, influence guideline recommendations for specific HF therapies (e.g. isosorbide dinitrate/hydralazine), sex/gender-specific guidance for individual therapies have not been put forth.
We identified few trials that reported significant sex/gender treatment interactions with respect to primary efficacy endpoints. Three of the 4 trials found that therapies were more beneficial in women compared with men, while 1 trial suggested that women may have worse outcomes with vagal nerve stimulation compared with control (Online Ref. 3). The investigators, however, believed that this sex-based variation in treatment response was likely a function of differences in baseline clinical profiles between men and women in the trial. Importantly, none of the evaluated trials were specifically powered to assess treatment responses by sex/gender.
Study Limitations
There are several notable limitations of this study. Although we conducted a comprehensive systematic review of the literature using 2 different search strategies, the study is subject to publication bias, such that published reports may inflate the observed reporting of sex/gender-specific subgroup data. It is possible we missed reported subgroup data that were not disclosed in the primary publication or supplementary materials. Due to heterogeneity in collection of data and reporting, we limited our systematic review to primary efficacy endpoints, and not safety events. Finally, we were not able to dissect designations of “sex” (a term reflecting biology of reproductive organs and secondary sex characteristics) and “gender” (referring to a person’s self-identity), although the latter is typically collected in clinical trials.
A Path Forward
We observed encouraging trends in reporting of sex/gender-based subgroup data in contemporary HF clinical trials, which may be related to dedicated efforts and requirements by regulatory agencies, funding bodies, and journals. In parallel with increasing representation of women in future clinical trials, reporting of treatment responses by sex/gender should be complete and encompass efficacy and safety endpoints (Table 3). Clinical trials must work to adapt current binary reporting structures as gender identity classification becomes broader and more complex.
Table 3.
Comparison of the primary design and results of the current systematic review with previously published work on representation in contemporary heart failure clinical trials
| Prior Work | New Work | |
|---|---|---|
| Title | Enrollment of Older Patients, Women, and Racial and Ethnic Minorities in Contemporary Heart Failure Clinical Trials | Do Women and Men Respond Similarly to Therapies in Contemporary Heart Failure Clinical Trials? |
| Reference | Reference 3: JAMA Cardiol. 2018;3(10):1011-1019. doi: 10.1001/jamacardio.2018.2559 | Current Manuscript |
| Key Question(s) | How representative are contemporary heart failure trials with respect to age, sex, and race/ethnicity? | How complete is reporting of treatment effects by sex / gender in heart failure trials? Do treatment effects to investigational therapies differ significant by sex/gender? |
| Design | Systematic Review Large (n>400), phase 2-4 randomized clinical trials 2001-2016 | Systematic Review Large (n>400), phase 2-4 randomized clinical trials 2001-2016 |
| Key Data Elements | Proportion of older adults, women, and racial/ethnic minorities | Subgroup analyses with respect to primary efficacy endpoints by sex/gender |
| Key Findings |
|
|
Supplementary Material
ACKNOWLEDGEMENTS
We would like to thank all the participants and staff involved with the multi-stakeholder “Heart Failure Collaboratory” for their work in promoting representation in contemporary HF trials.
FUNDING
None
DISCLOSURES
This paper reflects the views of the authors and should not be construed to represent the FDA’s views or policies.
Dr. Muthiah Vaduganathan is supported by the KL2/Catalyst Medical Research Investigator Training award from Harvard Catalyst (NIH/NCATS Award UL 1TR002541), and serves on advisory boards for AstraZeneca, Bayer AG, and Baxter Healthcare.
Dr. Ayman Samman Tahhan is supported by the Abraham J. & Phyllis Katz Foundation (Atlanta, GA) and NIH/NIA grant AG051633.
Dr. Stephen J. Greene is supported by the National Heart Lung and Blood Institute T32 postdoctoral training grant (T32HL069749-14) and a Heart Failure Society of America/ Emergency Medicine Foundation Acute Heart Failure Young Investigator Award funded by Novartis, and has received research support from Novartis and Amgen.
Dr. Gregg C. Fonarow reports significant consulting for Novartis, and modest consulting for Abbot, Amgen, Bayer, Janssen, and Medtronic.
Dr. Orly Vardeny received research support from AstraZeneca and Novartis and consultant support from Novartis and Amgen.
Dr. JoAnn Lindenfeld received grant research support from Novartis and consultant support/other from St. Jude, Abbott, Relypsa, RESMED, Cardiokinetix, and CVRx.
Dr. Mona Fiuzat reports receiving grant support from the NIH and Roche Diagnostics.
Dr. Christopher O’Connor reports receiving grant support from the NIH and Roche Diagnostics.
Dr. Javed Butler has received research support from the NIH and European Union; and has been a consultant for Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, CVRx, Janssen, Luitpold Pharmaceuticals, Medtronic, Merck, Novartis, Relypsa, Vifor Pharma, and ZS Pharma.
All other authors report no disclosures to this manuscript.
REFERENCES
- 1.Benjamin EJ, Virani SS, Callaway CW et al. Heart Disease and Stroke Statistics-2018 Update: A Report From the American Heart Association. Circulation 2018;137:e67–e492. [DOI] [PubMed] [Google Scholar]
- 2.Heidenreich PA, Albert NM, Allen LA et al. Forecasting the impact of heart failure in the United States: a policy statement from the American Heart Association. Circ Heart Fail 2013;6:606–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Tahhan AS, Vaduganathan M, Greene SJ et al. Enrollment of Older Patients, Women, and Racial and Ethnic Minorities in Contemporary Heart Failure Clinical Trials: A Systematic Review. JAMA Cardiol 2018;3:1011–1019. [DOI] [PubMed] [Google Scholar]
- 4.Chen J, Normand SL, Wang Y, Krumholz HM. National and regional trends in heart failure hospitalization and mortality rates for Medicare beneficiaries, 1998-2008. JAMA 2011;306:1669–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Woodard-Grice AV, Lucisano AC, Byrd JB, Stone ER, Simmons WH, Brown NJ. Sex-dependent and race-dependent association of XPNPEP2 C-2399A polymorphism with angiotensin-converting enzyme inhibitor-associated angioedema. Pharmacogenet Genomics 2010;20:532–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.