Gender Differences in the Prevalence of Frailty in Heart Failure: A Systematic Review and Meta-Analysis

Abstract

Objectives:

This study quantitatively synthesized literature to identify gender differences in the prevalence of frailty in heart failure (HF).

Background:

The intersection of frailty and HF continues to garner interest. Almost half of patients with HF are frail; however, gender differences in frailty in HF are poorly understood.

Methods:

We performed a literature search to identify studies that reported prevalence of frailty by gender in HF. Random-effects meta-analysis was used to quantify the relative and absolute risk of frailty in women compared with men with HF, overall, and by Physical and Multidimensional Frailty measures. Meta-regression was performed to examine the influence of study age and functional class on relative risk in HF.

Results:

Twenty-nine studies involving 8,854 adults with HF were included. Overall in HF, women had a 26% higher relative risk of being frail compared with men (95% CI = 1.14–1.38, z = 4.69, p<0.001, I² = 76.5%). The overall absolute risk for women compared to men with HF being frail was 10% (, 95% CI = 0.06–0.15, z = 4.41, p < 0.001). The relative risk of frailty was slightly higher among studies that used Physical measures (relative risk = 1.27, p < 0.001) compared with studies that used Multidimensional measures (relative risk = 1.24, p = 0.024). There were no significant relationships between relative risk and either study age or functional class.

Conclusions:

In HF, frailty affects women significantly more than men. Future work should focus on elucidating potential causes of gender differences in frailty in HF.

1. Introduction

The intersection of frailty and heart failure (HF) has garnered increasing interest in research and clinical practice [1,2], especially given the association of frailty with adverse outcomes in HF [3] and recommendations by scientific statements that frailty assessments should be incorporated in the care of adults with HF [4,5]. It has been recognized that frailty is highly prevalent among adults with HF. We previously showed in a meta-analysis of 26 studies that frailty affects nearly half of adults with HF [6]. Given this finding, perhaps there are additional characteristics, such as gender, that would explain this high estimate and help identify at risk groups.

While it is known that frailty is associated with worse outcomes, gender differences in frailty in HF have not been thoroughly explored. A recent meta-analysis describing the male-female health-survival paradox concluded that while women live longer lives and have better mortality rates, they have higher rates of frailty compared with men across the older adult lifespan in the community [7]. Simply put, frailty may impact women, particularly older women, significantly more so than men without a significant increase in mortality; however, this has not been studied in HF. In HF, women are more likely to be diagnosed with HF at an older age, have HFpEF, more comorbidities, and worse symptoms [8–11], which may predispose them to higher rates of frailty. Given that over half of the population of adults with HF are women [12], there is a need to identify gender differences in the prevalence of frailty in HF to inform clinical management of HF. The purpose of this systematic review and meta-analysis was to synthesize published literature to identify potential gender differences in the prevalence of frailty in HF.

2. Methods

2.1. Data sources and study eligibility

This study was a meta-analysis of published data-based studies on frailty in HF. Studies were considered eligible for inclusion if they met the following criteria: 1) observational or interventional research (baseline data); 2) sample or subsample consisted of patients with HF; 3) prevalence of frailty in the sample or subsample of patients with HF was available using any form of frailty assessment; 4) prevalence of frailty was stratified by gender; and 5) published full-text available. We excluded non-English studies. We first extracted data from published studies that were included in our previous meta-analysis (6) that met the above criteria; this included studies up through July 2016. Then, we searched EMBASE, PubMed and CINAHL from July 2016 through October 2020 using the MeSH heading heart failure and either the keyword frail* or MeSH heading “Frail Elderly” or MeSH heading “Geriatric Assessment.” Abstracts were reviewed for the above criteria and reference lists were hand-searched for additional relevant studies not identified in the search engines. Authors were contacted for studies that met all inclusion criteria except gender breakdown to inquire about inclusion in this meta-analysis. In the case of multiple studies using the same cohort, the study was chosen that provided the most detailed information meeting inclusion criteria and included the largest sample size. Full search strategies are presented within the PRISMA [13] flow diagram (Figure 1). Study screening and evaluation for eligibility for this meta-analysis was performed and validated by two members of the research team (M.R.D. and A.C.). A reference list of excluded studies is available upon request.

Figure 1.

PRISMA flow diagram showing study identification, selection, eligibility, and inclusion (11). For more information, visit www.prisma-statement.org.

2.2. Data extraction

Variables extracted for analysis are listed in Table 1. If clarification on extracted findings was required, the corresponding author was contacted via electronic mail to request this information and to query about any pending published work on frailty in HF. Extracted data were independently verified by a second member of the study team (A.C.) In the case of a discrepancy between two researchers (M.R.D. and A.C.), a third researcher (Q.D.) retrieved respective data and led the consensus discussion. The authors conducted this meta-analysis in concordance with PRISMA standards of quality for reporting meta-analyses [13] and the guidelines for Meta-Analyses and Systematic Reviews of Observational Studies [14]. The quality of research reported in studies was evaluated using the Strengthening the Reporting of Observational Studies in Epidemiology statement [15](Table S1).

Table 1.

Studies examined for systematic review and meta-analysis.

Author (year)	Country	HF sample (n)	Women with HF n (%)	Total Frail n (%)	Women Frail n (%)	Men Frail n (%)	Mean Age (years) ±SD	Proportion of NYHA III/IV %	Frailty Measure	HF Sample Characteristics Included in study	STROBE requirements met (n/22)
Cacciatore (2005) [20]	Italy	120	72 (60.0)	18 (15.0)	10 (13.9)	8 (16.7)	75.9 ± 6.7	Not reported	MDF: Frailty Staging System	Age > 65; random sample from community with HF; CCI	(22/22)
Chung (2014) [21]	USA	72	8 (11.0)	16 (22.2)	4 (50.0)	12 (18.8)	59 ± 2	99	PF: Handgrip Strength	Advanced HF before LVAD implant; Mean EF	(22/22)
Cooper (2017) [22]	USA	320	59 (18.4)	146 (45.6)	20 (33.9)	135 (51.7)	not reported	98.4	PF: Gait Speed	Pre-op Destination Therapy LVAD; “multimorbidity”	(22/22)
Denfeld (2018) [23]	USA	49	16 (32.7)	24 (49.0)	10 (62.5)	14 (42.4)	57.4 ±9 0.7	92	PF: Fried’s Frailty Phenotype	In & Out-patient scheduled for RHC;Mean EF; CCI	(22/22)
Dominguez-Rodriguez (2015) [24]	Spain	102	48 (47.1)	29 (28.4)	15 (31.3)	14 (25.9)	73 ± 3	100	PF: Fried’s Frailty Phenotype	Non-ischemic HF; undergoing CRT-D implantation; Mean EF; CCI	(21/22)
Dunlay (2014) [25]	USA	99	18 (18.2)	34 (34.3)	7 (38.9)	27 (33.3)	65 ± 9.4	Not reported	MDF: Frailty Index	Pre-op Destination Therapy LVAD; Mean EF; “Comorbidity Deficits” listed	(22/22)
Ferguson (2017) [26]	Australia	92	27 (29.3)	58 (63.0)	20 (74.1)	38 (58.5)	72 ± 16	62	MDF: Frailty Index	Inpatient with concomitant HF & Atrial Fibrillation; % EF < 45; CCI	(22/22)
Gastelurrutia (2014) [27]	Spain	1314	364 (27.7)	582 (44.2)	240 (65.9)	342 (36.0)	67 ±12.4	32.9	MDF: multiple geriatric scales/tests	Ambulatory HF Clinic; Mean EF; “Comorbidities”	(22/22)
Hornsby (2019) [28]	USA	114	66 (57.9)	39 (34.2)	25 (37.9)	14 (29.2)	68 ± 13	81.6	PF: SPPB	HFpEF Clinic; Comorbid Conditions listed	(22/22)
Jha (2016) [29]	Australia	156	47 (30.1)	51 (32.7)	20 (42.6)	31 (28.4)	53 ± 13	100	PF: Fried’s Frailty Phenotype	Advanced HF referred for transplant; Mean EF	(22/22)
Joseph (2017) [30]	USA	75	19 (25.3)	44 (58.7)	13 (68.4)	31 (55.6)	58 ± 12	100	PF: Fried’s Frailty Phenotype	Inpatient scheduled for LVAD; List of Comorbid Conditions	(22/22)
Joyce (2018) [31]	USA	56	15 (26.8)	33 (58.9)	6 (40.0)	27 (65.9)	77 ± 7	35	PF: Handgrip Strength	Age > 65; inpatient; prior to discharge; List of Cormorbid Conditions	(21/22)
Kawashima (2019) [32]	Japan	92	38 (41.3)	63 (38.4)	30 (78.9)	33 (61.1)	82 ± 6.6	Not Reported	MDF: Kihon Checklist	Age > 65; inpatient for HF, stable; Mean EF	(22/22)
Madan (2016) [33]	USA	40	23 (57.5)	26 (65.0)	17 (73.9)	9 (52.9)	75 ± 6.5	100	PF: Fried’s Frailty Phenotype	Age ≥ 65, outpatient; LVEF <35%; Mean EF; CCI	(22/22)
Martin-Sanchez (2017) [34]	Spain	465	283 (60.9)	169 (36.3)	125 (44.2)	44 (24.2)	82 ± 7	23	PF: Fried’s Frailty Phenotype	Age ≥ 65; Emergency Dept for acute HF; % EF < 45; CCI	(22/22)
McDonagh (2020) [35]	Australia	131	32 (24.4)	71 (54.2)	23 (71.9)	48 (48.5)	54 ± 14	Not Reported	PF: Fried’s Frailty Phenotype	Inpatient cardiology and Outpatient HF; Mean EF; List of Conditions	22/22
McNallan (2013) [36]	USA	448	191 (42.6)	84 (18.8)	39 (20.4)	45 (17.5)	73 ± 13	Not Reported	PF: Fried’s Frailty Phenotype	Community sample with HF; Mean EF; CCI	(22/22)
Mlynarska (2018) [37]	Poland	156	27 (17.3)	118 (75.6)	19 (70.4)	99 (76.7)	74 ± 6.8	Not Reported	MDF: Tillburg Frailty Scale	Inpatient scheduled for CRT device; Mean EF; List of Conditions	(20/22)
Nishiguchi (2016) [38]	Japan	206	63 (30.6)	34 (16.5)	11 (17.5)	23 (16.1)	74 ± 7.3	9.3	PF: Fried’s Frailty Phenotype	Age ≥ 60; outpatient previously hospitalized for HF; List of Conditions	(21/22)
Pandey (2019) [39]	USA	202	110 (54.5)	101 (50.0)	65 (59.1)	36 (39.1)	72 ± 7.5	92	PF: Fried’s Frailty Phenotype	Age ≥ 60; inpatient for HF;% < 45; CCI	(22/22)
Pilotto (2010) [40]	Italy	376	213 (56.6)	67 (17.8)	48 (22.5)	19 (11.7)	80 ± 7.3	Not Reported	MDF: Multidimensional Prognostic Index	Age ≥ 65; inpatient; Mean EF; List of Conditions	(22/22)
Pulignano (2016) [41]	Italy	331	140 (42.6)	88 (26.6)	31 (22.1)	57 (29.8)	78 ± 6	51.4	PF: Gait Speed	Age ≥ 70; community based; Mean EF; List of Conditions	(22/22)
Rodriquez-Pascual (2017) [42]	Spain	497	303 (61.6)	286 (57.5)	193 (63.7)	93 (47.9)	85 ± 7.3	27.9	PF: Fried’s Frailty Phenotype	Age ≥ 75; ambulatory discharged from hospital with HF in previous 12 months; %EF < 45; CCI	(22/22)
Sanders (2018) [43]	USA Canada Brazil Argentina	1767	882 (49.9)	1285 (72.7)	668 (75.7)	617 (69.7)	71 ± 9	35.1	MDF: Sum of Deficits Model	HFpEF; Americas cohort of trial; Mean EF; List of Conditions	(22/22)
Son (2018) [44]	Korea	190	51 (26.8)	117 (61.6)	40 (78.4)	77 (55.4)	70 ± 7.7	10	PF: K-FRAIL	Age ≥ 60; outpatient clinic; %EF < 40; List of Conditions	(22/22)
Sze (2017) [45]	UK	265	100 (37.7)	139 (52.5)	57 (57.0)	82 (49.7)	80 ± 6	74	MDF: Clinical Frailty Scale	Inpatient; LVEF<40%; CCI	(21/22)
Tanaka (2018) [46]	Japan	603	225 (37.3)	339 (56.2)	144 (64.0)	195 (51.6)	75 ± 6	34.9	PF: Gait Speed	Age ≥ 65; inpatient for acute HF; % “Reduced EF”; CCI	(22/22)
Uchmanowicz (2015) [47]	Poland	100	47 (47)	89 (89.0)	44 (93.6)	45 (84.9)	65 ± 8.5	37	MDF: Tillburg Frailty Indicator	Age ≥ 60;	(19/22)
Vidan (2016) [48]	Spain	416	206 (49.5)	316 (76)	177 (85.9)	139 (51.1)	80 ± 6	25.5	PF: Fried’s Frailty Phenotype	Age ≥ 70; inpatient with HF; Categories of EF; List of Conditions	(21/22)

Abbreviations: CRT-D=Cardiac Resynchronization Therapy Defibrillator; HFpEF=Heart Failure with Preserved Ejection Fraction; LVEF=Left Ventricular Ejection Fraction; LVAD=Left Ventricular Assist Device; MDF = Multidimensional Frailty; PF= Physical Frailty; RHC = Right Heart Catheterization; SPPB = Short Physical Performance Battery; VAD = Ventricular Assist Device CCI = Charlson Comorbidity Index.

2.2.1. Data Analysis and Coding

Data analyses were performed using STATA version 16.0 (Stata, College Station, TX). Studies that used primarily physical frailty assessments such as the full Frailty Phenotype measure, portions of the Frailty Phenotype measure (e.g. handgrip strength, gait speed), or the Short Physical Performance Battery were classified as “Physical Frailty”. Studies that used multidimensional frailty assessments, such as the Frailty Index, the Tillburg Frailty Indicator, or geriatric assessment with multiple exams were classified as “Multidimensional Frailty”. We performed a random-effects meta-analysis to quantify the relative and absolute risk of frailty in women compared to men with HF overall, and the relative risk was divided into Physical and Multidimensional Frailty measures. Random effects meta-analysis is appropriate for studies with variation in characteristics [16], such as with clinically heterogeneous populations like HF. Relative risk for each study was calculated by dividing the percent of total women who were frail in the study by the percent of total men who were frail in the study. Significance tests for the weighted average effect size among studies were estimated and reported using the 95% confidence interval (CI), z-scores (weighted estimate divided by the standard error of the weighted estimate), and associated p values [16]. We examined the heterogeneity Q statistic, which is the weighted sum of squared differences between individual study effects and across studies pooled effect, and I², which is the variation across studies ranging from entirely spurious (0%) heterogeneity to entirely real (100%) heterogeneity [17]. We also reported the 95% prediction interval, which provides the expected range of true effects in future settings for clinical interpretation of the heterogeneity [18]. We performed subgroup analyses of Physical vs Multidimensional Frailty Measures and Inpatient/Advance HF settings vs Ouptatient/Mixed settings. We performed meta-regression to examine if theaverage study age, or proportion of NYHA III/IV patients explained any variance in the results. Data was examined for publication bias visually and with Egger’s test [19].

3. Results

3.1. Study characteristics

Results of study identification, screening, eligibility, and inclusion are outlined in the PRISMA flow diagram (Figure 1). Twenty-nine published studies [20–48] involving a total of 8854 patients, including 3693 (41.7%) women, with HF and representing 9 countries were considered eligible and included in the systematic review and meta-analysis (Table 1). The average study age of participants ranged 53–85 years. Nineteen studies measured Physical Frailty [21–24,28–31,33–36,38,39,41,42,44,46,48] while 10 measured Multi-dimensional Frailty [20,25–27,32,37,40,43,45,47]. Table 1 summarizes sample characteristics of the studies.

A majority of studies were inpatient (44.8%) [21,22,24–26,29–32,37,39,40,45,46,48] and included samples of only older adults at least 60 years of age (48.3%) [20,23,31–34,36,38–42,44,45]. Eight studies (27.6%) included only patients diagnosed with HFrEF [21,22,24,25,29,30,33,45] and 2 were limited to patients with HFpEF [28,43]. The remaining studies were combined or undefined phenotype. Five studies (17.2%) were limited to advanced HF patients referred to transplant [29] or ventricular assist device placement [21,22,25,30]. Eight studies’ (27.6%) samples consisted of at least 50% women [20,28,33,34,39,40,42,43] and 10 (34.5%) included less than 30% women [21,22,25–27,30,31,35,37,44,45]. Most studies met all 22 of the STROBE [15] guidelines for quality reporting (Tables 1 and S1). Of those who did not meet all 22 guidelines, most failed to report potential sources of bias [27,31,37,48] or funding [24,38,45,47].

3.2. Meta-Analysis

Across all studies, women with HF had a 26% higher risk of being frail relative to men with HF (Figure 2) (95% CI.1.14–1.38; z = 4.69; p < 0.001). Women had an absolute risk increase of 10% for being frail compared with men (95% CI. 0.06 – 0.15; z = 4.41, p < 0.001) (Figure 3). The relative risk of being frail for women was slightly higher at 27% among studies using Physical Frailty measures (95% CI. 1.15–1.14; z = 4.61; p < 0.001) and slightly lower at 24% among studies using Multidimensional Frailty measures (95% CI. 1.03–1.50; z = 2.26; p = 0.024) (Figure S1). Heterogeneity statistics (Q = 118.92; p < 0.001; I² = 76.5%) for the overall model indicated substantial variation in the prevalence of frailty. The 95% prediction interval was 0.83–1.89. Studies that used Physical Frailty measures showed less heterogeneity (Q = 36.3; p < 0.001; I² = 50.4%) than those that used Multidimensional Frailty measures (Q = 80.8; p = 0.02; I² = 88.9%). The estimated relative risk for women being frail in studies that included inpatient and advanced HF studies (RR= 24%,95% CI 1.08–1.43, z = 3.03, p = 0.002) versus the estimated relative risk for women being frail in studies that included outpatients or both inpatient and outpatients (RR = 25%, 95% CI 1.12– 1.42, z = 3.75, p<0.001) was not substantially different Figure S4). Effect sizes reported in published studies were distributed symmetrically in the visual funnel plot (Figure S2). There was no significant bias from small studies (Egger’s test p = 0.396). Removing one study [21] with the most extreme effect size resulted in little difference in the estimate.

Figure 2.

Relative risk for frailty in women compared with men with HF - Random effects meta-analysis of the relative risk of prevalence of frailty in HF in women vs. men showing that women have 26% higher risk of frailty in HF (z = 4.69, p < 0.001) <RR = Relative Risk. ♦ = RR of individual study ⸺ = 95% CI of individual study>

Figure 3.

Risk difference for frailty in women compared with men with HF - Random effects meta-analysis of the absolute risk of prevalence of frailty in HF in women vs. men (z=4.17, p < 0.001 <RD = Risk Difference (i.e. absolute risk) ♦ = RD of individual study ⸺ = 95% CI of individual study>

3.3. Meta-Regression

Given the significant heterogeneity, we used meta-regression to examine average age and NYHA classification as study-level factors that might influence gender differences in frailty. There were no significant relationships between observed gender differences in the relative risk of frailty in HF and either age (β = −.01 ± 0.01; p = 0.21) (Figure S3a) or proportion of NYHA III/IV patients (β = 0.0002 ± 0.003; p = 0.95) (Figure S3b) across studies, indicating that neither age nor severity of HF explained the heterogeneity observed.

4. Discussion

This is the first known meta-analysis to examine the prevalence of frailty in HF among women compared with men. Our previous meta-analysis showed that nearly half of patients with HF are frail [6]. In this analysis of 29 studies, we expanded on these findings to demonstrate that among patients with HF, women have a significantly higher risk of being frail than men. This finding is relevant to practitioners because approximately half of all HF diagnoses are among women [12]. Given the implications of frailty, including higher mortality and hospitalization rates [3,20], increased risk for falls and disability [3,20,49] and decreased quality of life including institutionalization [23,47] it will be important to incorporate assessments of frailty into clinical practice.

Our findings contribute to the recognition of the male-female health survival paradox, where women report poorer health and quality of life but live longer than men [7]. The 2016 meta-analysis by Gordon et al. found that across older adult populations, women have higher frailty scores compared with men at all ages, but men have a higher risk of death even when controlling for frailty and age [7]. Additionally, Collard et al. showed in a systematic review that among community dwelling adults older than 64 years, women had higher rates of frailty (9.6%) compared with men (5.2%) [50]. In our study, we have similarly shown that the gender difference in prevalence of frailty is also found in HF. Notably the overall unweighted percentage of women who were frail across studies was 57%; this is significantly higher than the prevalence of frailty among community dwelling women across many other studies [51, 52].

HF and frailty are thought to share some common mechanisms, although little is understood about the pathophysiological intersection of the two [5,52,53]. It has been long recognized that frailty is associated with older age and increased comorbidities [53]. Similarly, HF in women tends to intersect with older age and comorbidities [54–56]. Taken together, these common multifactorial causes of HF and frailty may predispose women to higher rates of frailty in HF. Moreover, women are more likely to have HFpEF [55,56], which also intersects with older age and comorbidities [58]. One pathophysiological mechanism that underlies all of these is inflammation. Inflammation is associated with frailty, older aging (i.e. “inflammaging” [59]), and HFpEF. Given that women, especially post-menopausal women, are more likely to have inflammatory illnesses [57], they may be at higher risk for developing concurrent inflammatory syndromes such as frailty and HF. Understanding the underlying pathophysiology driving frailty among women with HF would be valuable in developing strategies to prevent and/or mitigate the effects of frailty among women with HF.

As with previous meta-analyses of frailty, and likely because of the multitude of frailty measures used across studies [59] our findings showed significant heterogeneity. Moreover, study-level factors do not explain this variation. Even within similar frailty measures, we found varying heterogeneity statistics. When examined by physical versus multidimensional frailty, however, studies that utilized physical frailty measures tended to cluster around the relative risk. Notably, the 12 studies that utilized Fried’s Frailty Phenotype [1] are closest to the relative risk of 1.26. Standardization of frailty measures in HF would allow for meaningful comparison across studies [44,60]. Continued research to refine definitions and frailty measures in HF is important, including distinguishing between physical frailty and multidimensional frailty as well as the role of cognitive function and depression in HF. Research identifying biomarkers of frailty in HF may provide additional clinical tools to identify high-risk patients.

It is important to note that the percentage of women in the included studies varied dramatically and, in some studies, men show more frailty than women. These findings may be the result of study design or participant characteristics. The percentage of women across all studies in this meta-analysis averaged about 42% and only one-third of included studies comprised 50% or more women. The lack of gender balance across HF studies has been noted as a recent analysis showed that women comprise only 28% of all participants in HF trials [61] despite the fact that the population of HF patients is about 50% women [12,58]. Thus, women are still under-enrolled and understudied in HF research, and the reasons for this are unclear. Most of the participants in the studies included in this meta-analysis were recruited from HF clinics or hospitalization, and only 3 were community-based studies [20,21,35]. Women are less likely to be referred to HF specialists [62]; this may explain why women are not equally represented across all studies. Future studies should focus recruitment efforts on community settings [58,62].

While this is the first known meta-analysis to examine gender differences in the prevalence of frailty in HF, we acknowledge limitations. First, while we utilized multiple databases and sources to locate relevant studies, it is possible that we inadvertently missed published research in this area. Second, heterogeneity in HF and frailty measures, along with the relatively low numbers of women enrolled in HF studies, may contribute to an incomplete understanding of gender differences in frailty in HF. Finally, our study examined only a few study-level factors that were readily available and consistently reported across the majority of studies, but other variables (e.g. comorbidities) could play a role. Additionally, ecological bias (i.e. bias due to aggregating data) must still be considered when interpreting these findings [16]. Because we were limited to aggregate data in these studies, we must temper our conclusion that neither age nor functional class do not influence gender differences in the prevalence of frailty in HF. Pooling patient-level data may demonstrate more of a relationship between gender, age, NYHA, and frailty, as well as other factors (e.g. HFpEF).

5. Conclusions

Frailty is highly prevalent among all HF patients, but more so in women than in men. In this first known meta-analysis of gender differences in the prevalence of frailty in HF, we found that women have a 26% higher risk of being frail in HF compared with men. Women with HF are less likely to be included in research trials than men, despite comprising about half of the HF population. Future research, using gender-balanced studies, should focus efforts on understanding why women are more likely to be frail compared with men in HF and describing key phenotypic gender differences related to frailty in HF. Routinely measuring frailty in all HF patients should be a part of the clinical assessment and may alert clinicians to increased risk for morbidity and mortality. Comprehensive patient education about the importance of physical activity and diet embedded in the context of HF self-care (risk for falling and frailty) may help mitigate frailty and improve quality of life.

Highlights.

• Frailty is highly prevalent in heart failure.

• Women have a 26% higher relative risk than men to be frail in heart failure.

• More gender-balanced studies of frailty in heart failure should be pursued.