This research followed the guidelines established by the Declaration of Helsinki. The study protocol was approved by the Ethics Committee of the University Hospital Center in which the study was implemented and all participants gave their informed consent to the anonymous use of data for research purposes.

The sample was described according to the participants’ age categories (<65 and ≥65 years). The distribution of quantitative variables was assessed using the Shapiro-Wilk test. Quantitative data were associated using parametric and non-parametric tests and categorical data were compared using the Chi-square test or Fisher's exact test as adequate. Results are presented as frequencies (percent) for categorical variables, mean (standard deviation) for normally-distributed continuous variables and median (lower and upper quartiles) for continuous variables with skewed distribution.

Binary logistic regressions were carried out and the crude and adjusted odds ratios (OR) and respective 95% confidence intervals (95% CI) were calculated as measures of association in models with age categories as dependent variables: <65 years old and ≥65 years old. Three models, which included both frail and pre-frail participants, were constructed regarding the overall sample and sex. Predictors in the models included: measures of physical functionality, using quantitative HGS and GSAH; surrogate measures of body composition, using BMI categorized as obese (≥30 kg m−2) and non-obese (<30 kg m−2, reference), and quantitative MAMC; the presence of symptoms and tolerance for physical exercise, using NYHA functional classes categorized as symptomatic (NYHA classes II+III) and not symptomatic (NYHA class I, reference), and the presence or absence (reference) of type 2 diabetes mellitus (T2DM). Missing values for NYHA classes (n=2) were included in the models on the reference group. The level of statistical significance for all tests was defined by p<0.050.

All analyses were performed with IBM SPSS Statistics 27 (SPSS, Inc., an IBM Company, USA).

Skeletal muscle function is an important component of aging and disease. Evidence toward the use of HGS as a biomarker for aging has been growing,22,23 such as in its use as an outcome predictor of HF, being associated with incidence, hospitalization, readmission and mortality.13,24–26

Hand grip strength was the most consistent predictor for age differences in this sample. Also, low HGS as defined by Fried at al., was already present in a third of the younger participants. These results show that loss of muscle function with age is observable in younger HF patients and that HGS could be used to establish its onset.

The gradual loss of muscle strength with age in pre-frail and frail HF patients should be a focus for further longitudinal studies with larger samples. Age-specific HGS cut-offs are needed for the entire HF population, as HGS seems to be a good measure of global myopathy, thus reflecting critical changes in HF progression that need rapid attention and personalized interventions.

Slow gait is independently associated with HF mortality, all-cause mortality and hospitalization in older (≥70 years) HF patients, as shown by the IMAGE-HF Study.27 During hospitalization, short term increments in GS are associated with reduced risks of death and readmission in older acute HF patients.28 In the present study, lower GS was associated with older age, which was expected. However, only 15.2% of the participants were classified as being slow, according to the cut-offs established by Fried et al.,15 which we attributed to the fact that this sample was comprised of community-dwelling individuals and not hospitalized patients. Moreover, there were no significant differences between younger and older patients in respect to this defining FP criterion.

The association of lower functionality with older age, consubstantiated by our results regarding HGS and GSAH, was not accompanied by lower estimated muscle mass using MUAC. This means that either our method for estimating muscle mass could not predict age differences, or that functional decline could precede muscular decline in this sample. While these associations deserve further research, it is interesting to verify that functional assessment has been gaining traction in contrast with body composition: a recent position statement from the Sarcopenia Definition and Outcomes Consortium recommends the use of HGS to define and assess sarcopenia, with many consortium members agreeing that dual-energy X-ray absorptiometry (DEXA) derived lean mass measures were not good predictors of health-related outcomes of sarcopenia.29 Until recently, DEXA was the preferred method for accessing lean mass by the European Working Group on Sarcopenia in Older People, while the use of other methods, such as bioelectric impedance and anthropometric assessment, was being discouraged.30

Obesity coexisted with FP in this sample, with 43.4% of the individuals having a BMI ≥30 kg m−2. There was, however, no association between obesity and age categories. Yet, 42% of the obese participants had concomitant low HGS, which raises important questions regarding the presence of sarcopenic obesity in this sample.

According to the Cardiovascular Health Study, the prevalence of T2DM is 18.8% in individuals without FP, 24.5% in pre-frail and 32.4% in frail individuals.31 T2DM can also reduce the likelihood of improving pre-frailty.32 Sarcopenia may play an important role in the pathophysiological mechanisms linking FP and T2DM, as muscle deterioration is often associated with insulin resistance, low-grade inflammation and mitochondrial alterations, typical of diabetes.33

Type 2 diabetes is an independent predictor for the development of HF.34 Diabetes is highly prevalent in HF and patients with both conditions have a higher risk of mortality compared with patients with only HF or T2DM.35

The frequency of T2DM in this sample was 34.3%, being notably higher in pre-frail participants compared with the frail (37.0% vs. 23.8%, respectively). Our multivariate analysis has shown T2DM to be associated with older age in frail and pre-frail patients. The associations between aging, diabetes, frailty and HF deserve further study. Nonetheless, our results might indicate a need for better glycemic control on younger pre-frail and frail HF patients in order to prevent or mitigate T2DM at older ages.

In our multivariate analysis, NYHA functional classes, which were recoded as indicators of the presence vs. absence of symptoms and tolerance to physical exercise, were not a significant predictor of age. This shows that younger and older pre-frail and frail HF patients tended to be similar in terms of the severity of symptoms and the level of tolerance to physical activity.

The frailty phenotype is associated with both advanced age and HF. Goldwater et al. discuss the possibility that secondary FP, related with HF, is a separate entity to that of primary FP, associated with advanced age, although both share the propensity to a catabolic state fueled by inflammation, mitochondrial dysfunction, oxidative stress and hormonal dysregulations. Identifiable differences between both conditions are subtle at most. Muscle loss in HF seems to be associated with a higher percentage of fast-twitch glycolytic type II fibers, and HF frail patients seem to have a better response to physical therapy compared with those with primary frailty.2 However, it is possible that age-related primary FP and HF-related secondary FP are not mutually exclusive, and the onset of FP occurs in younger ages in HF individuals than in the general population. Whether or not FP is a separate entity in HF or in aging, frailty seems to be a process of accelerated aging associated with chronic diseases such as HF,36 thus disputing the utility of taking an HF patient's chronological age into consideration when addressing assessment methods or therapeutic strategies to prevent or revert FP.

Some limitations should be acknowledged. First, this is a cross-sectional study, therefore it does not allow for causal associations. Secondly, the sample is rather small, which precluded its stratification in more than two age intervals; this stratification would eventually allow for a more thorough analysis. The effect of the small sample size was also felt in the frail (n=21) individuals, hampering the possibility of conducting a separate multivariate analysis on this group. For the same reason, the 37 participants who were excluded for not having FP could not be used to contrast our findings, as only six were 65 years or older. Also, body composition was assessed solely on the basis of anthropometric measurements, namely MAMC and BMI, as the conditions of the study setting, as well as the research logistics, did not allow for easy access to other assessment methods. However, the anthropometric measurements that led to the definition of MAMC and BMI were conducted by the same experienced registered nutritionist to avoid inter-observer errors. Finally, only the Fried phenotype was used to assess frailty. It would be interesting to study the effect of other frailty constructs, such as the multidimensional frailty classifications, as it is known that different measures can affect the proportion of HF individuals classified as frail and pre-frail.1

Despite the described limitations, this is, to our knowledge, the first study comparing older and younger individuals with concomitant FP and HF, and we hope it can contribute to broadening the view of FP in HF beyond the classical aspects of a geriatric syndrome.