Introduction

Heart failure (HF) is a persistent health challenge affecting 6.7 million Americans [1]. Prevalence of HF is predicted to exceed 8 million Americans by 2030, owing to an aging population and contemporary therapeutic advances that have improved survival [2, 3]. The economic burden of HF is projected to increase to $69.8 billion in 2030 largely due to unplanned hospitalizations [1]. Between 2015 and 2018, HF was the leading cause of hospitalization, morbidity, and mortality in women accounting for 3 million women living with HF; 44,958 deaths among women were attributed to HF in 2020 [1]. Heart failure subtypes include HF with reduced EF (HFrEF, EF ≤ 40%), mildly reduced EF (HFmrEF, 41–49%), and preserved EF (HFpEF, ≥ 50%) [1, 2]. Among 335,735 patients in the American Heart Association (AHA) Get with the Guidelines HF registry (GWTG-HF) across three time periods (2005–2009, 2010–2014, and 2015–2018), 49.3%, 48.5%, and 46.6% of patients, respectively, were women, of which 68% were White and 19% Black [4]. Although the incidence of HF is similar for women and men, there are phenotypic sex differences with more women than men presenting with HFpEF particularly at older ages [5,6,7,8,9]. Moreover, Black postmenopausal women have the worst prognosis after HF develops [1, 10]. Sex differences across the HF trajectory include risk factors, etiology, guideline-based therapies, receipt of secondary prevention services, and clinical outcomes [8, 11].

Exercise-based cardiac rehabilitation (CR) provides comprehensive secondary prevention services and improves outcomes among patients with cardiovascular disease including a small reduction in all-cause mortality, substantial reduction in all-cause hospitalization, and improved quality of life (QOL) with up to 12 months’ follow-up [12]. However, women represent less than 15% of participants in randomized clinical trials (RCTs) of exercise-based CR [12]. Evidence-based guidelines endorse CR and exercise training as safe and effective for medically stable patients, including women with HFrEF, to improve exercise capacity, clinical outcomes, and QOL [2, 13,14,15]. Furthermore, the most functionally impaired patients often achieve the greatest improvements with CR [16, 17].

Despite the undisputed benefits of CR, only 25% of older patients eligible for CR participate [18, 19]; less than 10% of eligible HF patients receive a CR referral [20]. Inequity in access and quality of healthcare in women is well documented in the USA [21]. Referral, participation, and completion rates are very low in older women especially among ethnic minorities and economically constrained subgroups [22,23,24,25,26,27]. Women participate less often than men across all ages but because women have a longer life trajectory than men, these gaps tend to worsen among older adults [28]. Barriers to CR attendance include the lack of understanding among patients and clinicians of the value of CR, prohibitive logistics particularly because many older women cannot drive, unaffordable copayments for CR for patients, and inflexible delivery formats to achieve personalized CR interventions tailored to the varied needs and circumstances of women [29].

Several studies examining the recruitment of women in HF RCTs reported that enrollment of women varied between 21 and 29% [30,31,32]. A review found that only 25% of participants in RCTs of HFrEF were women; women were under-enrolled relative to the population sex distribution of HF in more than 70% of the RCTs, with no significant improvement over time [33]. The underrepresentation of women in HF RCTs limits the ability to analyze sex-specific efficacy and safety of secondary prevention interventions. The aim of this review was to evaluate the effect of exercise-based CR on the outcomes of mortality, hospitalizations, exercise capacity, and QOL among women with HF. We also examined the referral, enrollment, and adherence of women with HF to CR in observational studies. Finally, we offer recommendations for future research.

Methods

Search strategy and study selection

We utilized the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [34] when obtaining relevant literature regarding the health outcomes among women with HF participating in RCTs of exercise-based CR (Fig. 1). Our search strategy included a systematic review of the literature across three databases (PubMed, CINAHL, and Web of Science). Within each database, a string of Medical Subject Heading (MeSH) terms was utilized (Supplemental Material Table SM1). The complete database of retrieved articles was screened initially for relevance by title and abstract. Potentially relevant articles were further screened via full-text review. The resulting search retrieved n = 380 articles for review. After screening for relevant titles and abstracts, n = 82 records were reviewed by full text. After full-text review, 18 studies were included in the final review. A summary of the included studies is provided in Supplemental Material Table SM2 (Table SM2).

Fig. 1
figure 1

PRISMA flow diagram

Full size image

Inclusion criteria included English language publications between 2009 and 2022, specifically addressing RCTs that evaluated outcomes or adherence either in the primary outcomes or in subgroup analyses of women diagnosed with all HF subtypes in exercise training and exercise-based CR. Outcomes evaluated included mortality (all-cause and cardiovascular), hospitalizations (all-cause and cardiovascular), exercise capacity, and QOL. We excluded studies that did not include women with HF, pilot studies, case reports, review articles, and expert opinion editorials. We sought to exclude studies without sex-specific primary or subgroup analyses on the outcome variables reported, but there were so few such studies that we retained all studies that included women.

Results

Study characteristics

Of the 18 studies included in this review, 6 were conducted in the USA [35,36,37,38,39,40]; 3 in Australia [41,42,43]; 2 each in Germany [44, 45], Italy [46, 47], and the UK [48, 49]; and 1 each in Norway [50], Greece [51], and Spain [52]. Nine studies were multi-center trials [35, 39, 40, 42, 44, 45, 47, 49, 50]. Follow-up of outcomes ranged from upon completion of the intervention at 3 months to up to 10 years.

Participant characteristics

The 18 RCTs enrolled 4917 patients, of which 1714 were women. The number of participants in each of the studies ranged from 25 to 2331 while the number of women ranged from 15 to 661. Women comprised between 21% (n = 15) [51] and 81% (n = 81) [38] of study participants. One large multi-center RCT, Heart Failure: A Controlled Trial Investigating Outcomes of exercise Training (HF-ACTION), recruited not only the largest proportion of HF participants (51%) but also the greatest number of women (n = 661) [35]. In seven studies, ≥ 50% of the participants were women [36,37,38, 40, 44, 45, 52]. Only 2 trials conducted pre-specified subgroup analyses of women [40, 53]. Participant age ranged from about 59 to 83 years. Nine studies included patients with HFrEF [35, 39, 41, 46,47,48,49,50,51] and twelve studies included patients with HFpEF [36,37,38,39,40,41,42,43,44,45, 47, 52]. In general, enrollment of women in HFpEF trials was greater than that in HFrEF trials reflecting the preponderance of HFpEF among women.

Interventions

Eight studies described testing a comprehensive multidisciplinary CR intervention [38,39,40,41,42, 46, 48, 49] and all studies implemented various exercise training protocols (see Table SM2 for details). Six studies implemented both aerobic and resistance exercise training [39,40,41,42, 44, 48]. Several interventions included a home-based CR component [35, 39,40,41, 47] and two studies implemented entirely home-based interventions [48, 49]. Davidson et al. incorporated a structured HF-specific disease management component into CR [41]. Similarly, Mudge et al. provided study participants in both groups with a HF disease management program and a home exercise program whereas the CR participants (experimental group) were offered an additional 24 weeks of supervised center-based exercise training [42]. Three studies implemented high-intensity interval training (HITT) [45, 50, 51] while Palau et al. implemented inspiratory muscle training over 3 months [52]. Kitzman et al. examined the independent effects of a calorie restricted diet or aerobic exercise training as well as the combined effect of a diet and exercise intervention [38]. Kitzman and colleagues subsequently implemented an early, transitional, tailored progressive CR intervention focused on multiple physical function domains for older patients who were hospitalized for acute decompensated HF [40]. The dose of aerobic exercise training, and thus, the total volume of exercise, varied widely across the studies with single session durations of 30–105 min, session frequencies of 2–7 sessions per week, and total program durations of 12–54 weeks.

Outcomes

Composite mortality and hospitalization

Four studies reported data on the composite outcome of all-cause mortality and hospitalization, all of which were not statistically significantly different between groups (Table 1): O’Connor et al. [35] (HR 0.93, 95% CI 0.84–1.02, p = 0.13), Jolly et al. [48] (HR 1.45, 95% CI 0.43–4.86), Mudge et al. [42] (OR 0.80 95% CI 0.49–1.30), and Kitzman et al. [40] (rate ratio 0.93, 95% CI 0.77–1.12). Only one study reported a HR for the primary endpoint of all-cause mortality and hospitalization by sex: women (HR 0.83, 95% CI 0.68–1.00), men (HR 0.97, 95% CI 0.871.09), with a nonsignificant treatment-by-sex interaction (p = 0.17) [35]. After adjustment for significant covariates, women in the exercise training group had a significant 26% reduction in all-cause mortality or all-cause hospitalization (HR 0.74, 95% CI 0.59–0.91) compared to no reduction in men (HR 0.99, 95% CI 0.86–1.13) with a significant treatment-by-sex interaction (p = 0.027) [53].

Table 1 Summary of mortality/hospitalization of women with HF after exercise-based cardiac rehabilitation
Full size table

All-cause mortality

Five studies provided data for all-cause mortality with follow-up periods from 6 to 120 months [35, 40,41,42, 49]. After 30 months, the HF-ACTION trial reported nonsignificant reductions in mortality in the CR group versus the UC group (HR 0.96, 95% CI 0.79–1.17, p = 0.70) [35]. Dahal et al. reported no between-group differences in mortality (HR 1.01, 95% CI 0.26–3.92) [49] nor did Kitzman and colleagues (rate ratio, 1.17; 95% CI 0.61–2.27, p = 0.63). Conversely, Mudge et al. reported significantly fewer deaths at 12 months in the CR group compared with the UC group (p = 0.04) [42]. Moreover, Davidson and colleagues [41] reported that 79% of the UC group and 93% of the CR group were alive at 12 months (OR 3.85, 95% CI 1.03–14.21, p = 0.004) [41].

Cardiovascular mortality

Two studies provided data for cardiovascular mortality with no sex-specific analyses. After 120 months of follow-up, Belardinelli et al. reported that cardiovascular mortality was higher in the UC group compared with the CR group (10 vs 4, HR: 0.68, 95% CI 0.30–0.82, p < 0.001) [46]. Moreover, gender did not predict the outcome. The HF-ACTION trial reported a nonsignificant reduction in the combined endpoint of cardiovascular mortality or cardiovascular hospitalizations in the CR group versus the UC group (HR 0.92; 95% CI 0.83–1.03, p = 0.14). There was also a nonsignificant reduction in cardiovascular mortality or HF hospitalization in the CR group (HR 0.87, 95% CI 0.75–1.00. p = 0.06) [35].

Hospitalizations

All-cause hospitalizations were reported in eight studies [35, 40,41,42, 46,47,48,49]; three studies reported fewer hospitalizations with exercise-based CR [41, 46, 47]. Belardinelli et al. reported 8 hospitalizations in the CR group and 25 in the UC group (HR 0.64, 95% CI 0.34–0.81, p < 0.001) [46]. Another study reported that among participants aged 76.9 ± 5.67, over 6 months, all-cause hospitalizations occurred in 25 (15.2%) in the CR group versus 60 (36.8%) in the UC group (HR 2.91, 95% CI 1.70–4.97, p < 0.001). After adjustment for covariates, CR reduced all-cause hospitalizations by 44.2% (β = 0.558, 95% CI 0.326–0.954, p = 0.033). Participant sex had no effect on hospitalization (β = 0.129, HR 1.138, 95% CI 0.652–1.985, p = 0.650) [47]. Davidson and colleagues reported that at 12 months, compared with the UC group, fewer participants in the CR group were hospitalized (44% vs 69%, p = 0.01). Those in the CR group had lower odds of having any hospital admission (OR 0.20, 95% CI 0.07–0.58, p = 0.01) or having a cardiovascular hospitalization over 12 months (OR 0.12, 95% CI 0.04–0.37) [41]. Dalal and colleagues reported no significant between-group differences in hospitalizations during 12 months of follow-up (OR 0.72, 95% CI 0.35–1.51, p = 0.386) [49]. Mudge and colleagues also found no significant between-group differences at 12 months in all-cause readmissions or HF readmissions [42]. Finally, Kitzman et al. reported that the rate of all-cause rehospitalization at 6 months was not different between groups (rate ratio, 0.93; 95% CI 0.66–1.19) [40].

Cardiopulmonary exercise capacity

Ten studies reported data on cardiopulmonary exercise capacity (VO2peak ml/min/kg) [35,36,37,38, 43,44,45,46, 50, 52]. Most of these (7/10) studies showed an increased exercise capacity resulting from exercise training (Table 2). One study (19% women) showed no benefit in VO2peak of HIIT compared to moderate continuous exercise (MCT); however, 51% of the HIIT group exercised below the prescribed intensity whereas 80% in the MCT group trained at a higher intensity [50]. Moreover, at baseline, the UC group had the highest mean VO2peak. In a second study (50% women), after 4 months of exercise training, VO2peak increased 24.6% (p = 0.02) in HFpEF patients compared with no change in the UC group (5.1%, p = 0.19). However, there were no significant between-group differences (p = 0.06) [43]. Similarly, Mueller and colleagues found no significant change in VO2peak in patients with HFpEF (66.5% women) at 3 or 12 months between the HIIT and MCT groups and neither group met the minimal clinically important difference compared with the UC group [45].

Table 2 Effects of cardiac rehabilitation on exercise capacity among women with heart failure
Full size table

Twelve studies provided data on the 6-min walk test (6MWT) [35,36,37,38,39,40,41,42, 44, 47, 51, 52, 57]. Two studies measuring exercise capacity with the incremental shuttle walk test (ISWT) found no significant between-group differences at 3 or 12 months after CR [48, 49]. Of the 12 studies measuring the 6MWT, five studies found significant improvement with CR over UC at 3 months [35, 44, 47, 51, 52], two studies at 4 months [36, 37], one study at 5 months [38], and one study at 6 months [47]. Three studies that measured the 6MWT at 12 months found no significant between-group differences [35, 39, 41, 57]. Pozehl and colleagues found significant improvement in the 6MWT among the HFpEF subgroup but not the HFrEF subgroup and only at the 18-month follow-up; there were no between-group differences in this subgroup between CR and UC at 6 and 12 months [39, 57].

Quality of life

Most studies measured QOL with either the Kansas City Cardiomyopathy Questionnaire (KCCQ) which measures 5 domains (physical limitation, symptoms, self-efficacy, QOL, social limitations) and generates 3 summary scores (overall summary, clinical summary, total symptom score) [54] or the Minnesota Living with Heart Failure Questionnaire (MLHFQ) [55]. Of the twelve studies measuring QOL with the MLHFQ, six assessed QOL at 3 months [41, 43, 44, 47, 51, 52] and two of these studies found no significant between-group differences [43, 47] (Table 3). Three studies reported MLHFQ scores at 4 months [36,37,38] and only one study found a significant improvement in MLHFQ physical score and no change in the mental score or the total score between groups [36]. Of the two studies that administered the MLHFQ at 6 months, one found significant between-group differences [47] and one did not [48]. Three studies administered the MLHFQ at 12 months, one of which found significant between-group differences [49] and two did not [41, 48]. One Italian study measured QOL using the MLHFQ over 120 months and found significant between-group differences which were sustained over the study period [46].

Table 3 Effects of cardiac rehabilitation on quality of life among women with heart failure 
Full size table

Six studies measured QOL with the KCCQ, four of which measured this outcome at 3 months [40, 45, 50, 56] and two of which found significant between-group differences [40, 56]. The HF-ACTION trial administered the KCCQ every 3 months for the first year and then annually thereafter. They found significant between-group differences at 3 months in the exercise training group that was sustained throughout the study [56]. One study administered the KCCQ at 4 months and found that only the diet intervention improved scores [38]. Finally, one study administered the KCCQ at 6, 12, and 18 months and found significant between-group differences over time in the HFpEF subgroup only [57].

Adherence to interventions in randomized controlled trials

The dose–response relationship between exercise-based CR participation and outcomes has been well established [58, 59]. Adherence to the CR interventions in the RCTs described in Table SM2, where reported, is difficult to synthesize given the wide range of methods of measuring the dose of exercise received or the number of sessions attended. Understanding the dose of the intervention received is important for interpreting the results and for understanding the acceptability of the intervention by participants. In the HF-ACTION trial, participants exercised a median of 76 min per week at 3 months with the goal of 90 min per week. By 12 months, exercise time decreased to about 74 min per week [35]. Moreover, 50% of the UC group were dissatisfied with their group assignment with the possibility of crossover to exercise. The Heart Camp (HC) trial reported that the intervention group had significantly greater adherence to exercise at 12 months (42%) and 18 months (35%) compared with the enhanced UC group (28% and 19%, respectively) with no significant between-group differences found at 6 months [39]. Mudge et al. reported that there was no association between mortality and rehospitalizations and exercise training attendance (< 8 sessions 21/30 [70%], 8–15 sessions 28/43 [65%], 16–23 sessions 10/30 [49%], ≥ 24 sessions 16/28 [57%]; p = 0.27) [42]. Moreover, they reported that center-based CR adherence was poor at 43%. Conversely, nine studies reported over 80% adherence to their interventions [37, 38, 40, 43,44,45,46, 48, 50].

Cardiac rehabilitation referral and participation in observational studies

Alongside the RCTs reviewed, we identified additional observational studies that conducted sex-specific subgroup analyses of referral to and participation in CR. Among 105,619 participants in the GWG-HF registry, there was no significant difference in referral between the sexes after adjusting for clinically relevant covariates [20]. Conversely, a study from Denmark found that among 33,257 participants, men with HF were significantly less likely to be referred to CR than women with HF (OR 0.85, 95% CI 0.80–0.89) [60]. These findings are consistent with prior findings that Italian women were significantly more likely to be enrolled in CR than men following decompensated HF [61]. In a more contemporary Australian study, women and older patients were significantly less likely to be referred to CR than men (OR: 0.65, 95% CI 0.42–1.02, p = 0.06) [62]. Notably, only women with HFrEF were significantly less likely to be referred to CR than men (OR 0.47, 95% CI 0.21–1.04, p = 0.06), and not women with HFpEF (p = 0.21). Similarly, an investigation found that among 243,208 Medicare beneficiaries, women with HF were less likely to participate in CR than men (1.8% vs 3.7%, p < 0.001) but there was no sex difference among 66,710 veterans [63]. Among Medicare beneficiaries with HF from 2015 to 2016, only 611 (4.3%) of 11,696 HF patients eligible for CR participated within 6 months of HF hospitalization, with lower participation in women versus men (3.3% vs 5%; p < 0.001) and Black versus White patients (2.9% vs 3.7%; p < 0.001) [26]. This sex disparity in CR referral, enrollment, and adherence is consistent with trends established over the past decade [64, 65]. While these contrasting results provide additional insight into sex-specific differences in referral to and participation in CR, the synthesis of these studies should be taken together with the RCTs reviewed herein when drawing generalizations.

Discussion

We reviewed 18 RCTs that included women as participants and that examined the effects of exercise-based cardiac rehabilitation on mortality, hospitalizations, exercise capacity, and QOL among patients with HFpEF and HFrEF. A systematic review was not feasible because only two studies provided any type of sex-specific analyses. Despite this, we found limited evidence for mortality benefit for men or women participating in exercise-based CR. Plausible explanations for limited prognostic effects include underpowered studies with small sample sizes and less than optimal adherence to the exercise training protocol. In the only study reporting sex-specific data, after adjustment for significant covariates, women in the exercise training group in the HF-ACTION trial had a significant 26% reduction of all-cause death or all-cause hospitalization compared with no reduction in men. Our review of RCTs found that three of eight studies reported reductions in rehospitalizations among the intervention group. Generally, exercise training improved cardiopulmonary exercise capacity as measured by VO2peak and the 6MWT. The influence of exercise-based CR on QOL was mixed with most studies favoring CR at 3 months but generally not at 4, 6, and 12 months which is typical of the fading effects when interventions are withdrawn. Moreover, generally, the physical dimension of QOL but not the mental dimensions improved.

A previous review of 44 trials of exercise-based CR for HF patients (n = 5783) reported that 21 trials (n = 2182) demonstrated reduction in overall hospital admissions at 12 months of follow-up (CR 180/1093 [16.5%] versus UC 258/1089 [23.7%]; RR 0.70, 95% CI 0.60–0.83) but no difference in all-cause mortality [66]. Long and colleagues also reported that 14 trials (n = 1114) demonstrated reductions in HF-specific hospitalization (CR 40/562 [7.1%] versus UC 61/552 [11.1%], RR 0.59, 95% CI 0.42–0.84). Improved QOL (MLFHQ) was reported in 17 trials (n = 1995) (mean difference (MD) − 7.11 points, 95% CI − 10.49 to − 3.73) [66]. Another review of 25 RCTs (n = 4481) that included only HFrEF patients reported no benefit for hospitalizations or mortality with exercise-based CR [67]. One systematic review found that exercise training improved physical performance and cardiorespiratory outcomes in patients with HFpEF [68]. Similarly, a meta-analysis of 8 RCTs with 436 HFpEF patients found that exercise training improved VO2peak (weighted MD, 95% CI, 1.660 [0.973–2.348] ml/kg/min), 6MWD (33.883 [12.384–55.381] m), and MLHFQ total score (9.059 [3.083–15.035] points) compared with UC [69]. They also observed that exercise training did not significantly improve mental or emotional dimensions of QOL in HFpEF patients. Without exception, these reviews were unable to report on sex-specific outcomes.

The most glaring deficiency of the studies reviewed was the lack of sex-specific data analyses. Despite the stipulation of the National Institutes of Health Revitalization Act to include women and men in RCTs proportionate to the sex-specific prevalence of the disease, women are under-represented in HF RCTs [8, 11, 70]. Morgan and colleagues identified 146 HF RCTs including 248,620 patients, of which only 25.8% were women. Women were least likely to be enrolled in trials recruiting patients with ischemic cardiomyopathy (17.9%) and severe systolic dysfunction (EF < 35%) (21.4%), and those involving an invasive procedure (21.1%). The highest proportion of women was enrolled in RCTs assessing HFpEF (51.6%) and trials including older participants (40.5%) [70]. This underrepresentation has been partly attributed to age biases because HF is predominately present in older women. Others point to the numerous barriers women encounter when approached to participate in RCTs including transportation, caregiving responsibilities, lack of perceived benefit, and poor health literacy about HF trials [71]. Notably, RCTs led by women are more likely to enroll representative samples of women [72, 73]. Given the paucity of data on the effects of exercise-based CR among women with HF, we offer recommendations for future research (Table 4).

Table 4 Recommendations for research
Full size table

Recommendations for research

Under-enrollment of women with HF in secondary prevention RCTs deprives them of the benefits of trial participation, including guideline-directed medical therapy, close clinical follow-up, and better health outcomes. Women should be included in RCTs in proportion to the sex-specific distribution of HF. Strategies for including more women in RCTs include removing upper and lower age limits in exclusion criteria for enrollment in trials, training clinicians and research personnel on the importance of enrolling women, targeting important stakeholders in the community to increase awareness and importance of enrolling women, and removing barriers to enrollment to make participation convenient [8, 71]. Additionally, given that racially and ethnically diverse women present with unique biological and social determinants of health, studies that evaluate the intersection between race and sex warrant further investigation as well as the possible role of both implicit and explicit biases in CR enrollment and participation [8, 71]. The health literacy and cognitive function of participants should be evaluated during recruitment and intervention implementation so that research materials and instructions can be tailored to their needs [17]. Medical education and training on sex and gender diverse cardiovascular needs will better prepare our future scientists. The sex-specific manifestations of HF and response to treatments among cis- and gender diverse individuals warrant additional research [74]. Finally, study participants should be included as equal partners and stakeholders in designing the recruitment process and messaging, the informed consenting process, the research questions, the interventions, and the selection of meaningful and relevant patient-reported outcomes.

Many of the RCTs reviewed herein primarily examined the effects of the exercise training aspect of secondary prevention. Cardiac rehabilitation for patients with HF is a complex, multidisciplinary intervention for which a standardized dose or delivery format may not be feasible or acceptable given the unique needs, skills, capacities, competencies, exercise capacity, motivation, and preferences of women compared to men [29, 75]. Given that many individuals with HF are elderly and frail and have multiple comorbidities such as diabetes, depression, and cognitive impairment, and limited or no social support, there is a need for more holistic approaches to designing rehabilitation interventions that address these issues among women for improved uptake and adherence. The mechanisms by which varied exercise-based secondary prevention interventions work for whom, when, and why are unclear, and more nuanced approaches for women are needed for this to be elucidated [76]. Research on effective multifaceted interventions to address the sex and gender gaps in HF outcomes is needed. Mixed methods research should be used not only to measure intervention outcomes but also to understand the complexities of what intervention components work for whom, when, and why.

Studies of secondary prevention interventions for women with HF should be co-designed with diverse patients so that they are tailored to their needs in terms of essential components, mode of delivery, frequency, and setting, and then rigorously tested in RCTs. While exercise training is central to secondary prevention for patients with HF, other components are equally essential including education, diet and nutrition counseling, and psychosocial counseling [77]. Theory-based behavior change techniques, which consider the multiple factors that influence health behaviors, are important for improving adherence to interventions. The perceptions, motivations, skills, and the social environment are key influences on the behavior of participants in our studies. Knowledge of behavior change theory and the processes by which behavior change occurs helps us design interventions that lead to meaningful improvement in health behaviors. Delivering comprehensive CR (e.g., education, counseling, support, skill building) has been found to be superior to providing only selected core components such as exercise training [78]. A systematic review found that exercise training, psychosocial management, and risk factor modification each contributed directly to the effectiveness of CR [79]. Women-focused CR interventions may better engage women [80], might result in better psychosocial outcomes and QOL than traditional programs [29, 81,82,83], and should be tested in RCTS with women with HF. Ghisi and colleagues provide a clinical practice guideline that offers 15 recommendations for improving referral of women to CR, for adjustments to the mode of exercise training delivery in CR, and for women-focused education and counseling. This guideline specifically recommends that women are provided with greater choice of their preferences for modes of exercise training and the setting in which it is delivered [81]. Additional research is needed to examine the efficacy of such interventions for women with HF.

Many examples of new delivery models of CR are emerging; some of which might be most beneficial for women with HF. Dalal and colleagues tested a novel, tailored, home-based HF self-management program for patients with HF and their caregivers [49]. However, only 22% of the participants were women and it would be beneficial to examine such an intervention in a study with a greater proportion of women using sex-specific analyses. Home-based and technology-enhanced, patient-centered models of CR with appropriate quality assurance as an alternative or adjunct to traditional, center-based programs are needed to improve access to CR [27, 84]. A home-based strategy to CR may overcome various barriers contributing to underutilization [85] especially for women with HF. A Cochrane review of 23 trials of 2890 participants in CR demonstrated that home-based programs had similar efficacy as center-based programs with no significant group differences in mortality, exercise capacity, or QOL [86]. Increasingly, people 65 years of age and older are using smartphones which may enhance CR with features such as messaging and two-way video capabilities as well as accelerometers to track steps [27]. However, caution is warranted because individuals with poor health and technology literacy and who are non-English speaking, women, and older adults especially in rural areas may have difficulty engaging in technology-enhanced CR [87].

For research findings to be maximally interpretable and generalizable, sex-specific analyses in research and RCT design should be the standard, not the exception. Future HF clinical trials should be powered to investigate the substantial heterogeneity of the female HFpEF population in terms of age, ejection fraction, comorbidities, and frailty. For example, in a study of 295 patients with HFpEF (59% women), women had worse peripheral oxygen extraction with exercise, worse right ventricular and left ventricular systolic reserve, and worse diastolic reserve compared with men [88]. Scientific journals could also require authors to address sex and gender differences to publish manuscripts.

Limitations

We intended to conduct a systematic review of the effects of exercise-based cardiac rehabilitation on relevant health outcomes among women with HF. Due to the paucity of sex-specific data in the studies reviewed, this goal was not realized. Moreover, most of the studies included were small. We also included studies of patients with all subtypes of HF and with heterogenous characteristics, comorbidities, and cardiorespiratory fitness. Finally, study interventions were diverse with variable adherence by participants which likely influenced the outcomes evaluated.

Conclusion

The review of RCTs examining the effects of exercise-based cardiac rehabilitation of patients with HF provided very little sex-specific information for which to conclude exercise-based CR is effective for women with HF. There remain significant gaps in our knowledge about the best strategies for recruiting more women to clinical trials, the appropriate intervention components most effective for women with HF, and how the health outcomes of women differ from men after these interventions. More creative, flexible, and individualized, preference-based interventions that exploit the appropriate use of digital technologies may fill the existing gaps and lead to sex-specific clinical practice guidelines. Significantly more research is warranted.