Creatinine to Cystatin C Ratio, a Biomarker of Sarcopenia Measures and Falls Risk in Community-Dwelling Older Women

Abstract

Background

The ratio of creatinine to cystatin C (Cr:Cyc) has been proposed as a biomarker of sarcopenia, as greater Cr:Cyc is typically associated with greater muscle mass. We examined the relationship between Cr:Cyc with individual sarcopenia measures, 5-year self-reported falls, and 12-year fall-related hospitalizations in a prospective cohort study of 1 118 community-dwelling older women (mean age 75.2 ± 2.7 years).

Methods

Serum Cr:Cyc, hand grip strength, and timed-up-and-go performance were assessed at baseline (1998), while dual-energy x-ray absorptiometry-derived appendicular lean mass (ALM)/height² (m) was obtained in a subset of women at baseline and 1 year (n = 334). Incident 5-year self-reported falls and 12-year falls-related hospitalizations were considered.

Results

In a multivariable-adjusted model, women with the lowest Cr:Cyc (Quartile [Q] 1) had 5% (1.0 kg) weaker grip strength, as well as 3.7% (0.22 kg/m²) and 5.5% (0.031) lower ALM adjusted for height² or body mass index, respectively, compared to women in Q4 (all p < .05). 329 women reported an incident fall over 5 years, and 326 fall-related hospitalizations were recorded over 12 years. Women in Q1 of Cr:Cyc had a greater relative hazard for a fall over 5 years (hazard ratio [HR] 1.50; 95% confidence interval [CI] 1.11–2.01) and fall-related hospitalization over 12 years (HR 1.53; 95% CI 1.13–2.07) compared to Q4 in the multivariable-adjusted model.

Conclusion

These findings support further investigation into the use of Cr:Cyc as a muscle biomarker to help clinicians identify individuals at risk of falls for early inclusion into evidence-based primary prevention programs targeting improvements to diet and exercise.

Common manifestations of aging include a loss of muscle mass, strength, and physical function, with a higher deterioration of these measures collectively termed “sarcopenia” (1). These components of sarcopenia have been associated with poorer long-term health outcomes in older adults such as falls (2), frailty and loss of independence (3), disability (4), and mortality (5). As such, preventing sarcopenia is essential to facilitate improvement in quality of life and other health measures (6). However, while the importance of muscle function (comprising muscle strength and physical function) for the aforementioned outcomes is well recognized, the clinical importance of lean mass assessment still remains unclear (2,5,7).

Appendicular lean mass (ALM) is often assessed using dual-energy x-ray absorptiometry (DXA) or bioelectrical impedance analysis (BIA) (8). However, it has been proposed that DXA and BIA-derived ALM may be inadequate to quantify muscle mass. Specifically, when considering a direct measure of muscle mass, deuterated creatine (D3Cr) has been reported to be consistently associated with muscle function and disability (9). Despite these benefits, the D3Cr method is more time-consuming as it requires the ingestion of a stable isotope measured over 3–6 days. Hence, identifying clinically relevant biomarkers capable of reflecting components of sarcopenia (muscle function and muscle mass) that are also indicative of clinical outcomes such as falls would be of interest. Biomarkers of these manifestations may also provide insight into the pathophysiology and temporal processes during aging (10), which might help identify susceptible individuals with high falls risk for targeted primary prevention exercise and nutrition programs.

Creatine phosphate is produced and stored predominantly by muscle cells and excreted as creatinine via the kidneys in urine (10). As such, serum creatinine’s role as a biomarker of either muscle and kidney function may be influenced by the other (11). Cystatin C is a protein ubiquitously expressed in all tissues and as such is not directly related to muscle mass. Cystatin C is also freely filtered by the kidneys (12). Given that the generation and clearance of creatinine and cystatin C are relatively constant, the serum creatinine (derived predominantly from muscles) to cystatin C (ubiquitously expressed) ratio has been proposed as a biomarker of relative muscle mass and muscle function in numerous smaller studies within various clinical populations (eg, diabetics, vascular, and chronic kidney disease) (13–17) and healthy older adults (18–21). However, the relationship between the serum creatinine to cystatin C ratio (Cr:Cyc) and lean mass assessed via DXA remains unclear (15), while the association of this ratio with clinical outcomes primarily linked with sarcopenia such as falls remains unknown. For example, falls were also considered by the Sarcopenia Definitions and Outcomes Consortium as one of the outcomes when validating their sarcopenia definition (7). Therefore, this study aimed to determine the associations between Cr:Cyc and measures of sarcopenia, long-term risks of self-reported falls, and falls-related hospitalizations in a large prospective cohort of older community-dwelling Australian women.

Method

Study Population

The population included participants from the Perth Longitudinal Study of Aging in Women (PLSAW). Women were recruited in 1998 to a 5-year, double-blind, randomized placebo-controlled trial of daily calcium supplementation to prevent fracture, the Calcium Intake Fracture Outcome Study (CAIFOS) (22). As previously described, the women were included based on an expected survival beyond 5 years and not receiving any medications (including hormone replacement therapy) known to affect bone metabolism (22). The women (n = 1 460) were recruited from the Western Australian general population of women aged 70 years or older by mail using the electoral roll, which is a requirement of Australian citizenship. After completion of the 5-year trial, women were invited to participate in 2 follow-up observational studies. The total follow-up was 14.5 years. The current study then included 1 118 women (Supplementary Figure 1). All participants provided written informed consent. Ethics approval was granted by the Human Ethics Committee of the University of Western Australia. Both studies were retrospectively registered on the Australian New Zealand Clinical Trials Registry (CAIFOS trial registration number #ACTRN12615000750583 and PLSAW trial registration number #ACTRN12617000640303) and complied with the Declaration of Helsinki. Human ethics approval for the use of linked data was provided by the Human Research Ethics Committee of the Western Australian Department of Health (project number #2009/24).

Baseline Characteristic Assessment

Questionnaires relating to both recreational and physical activities were undertaken for all participants in the 3 months prior to their baseline visit in 1998 to determine structured activity levels. This has previously been described in detail within this cohort (23,24). The amount of structured activity (kcal per day) was then calculated using a validated method applying the type of activity, time engaged in the activity, and the participant’s body weight (kg) (25), described in more detail within the Supplementary Material. Smoking status was coded as nonsmoker or smoked ever (if they had smoked more than one cigarette per day for more than 3 months at any time in their life or were a current smoker). Body weight was measured using digital scales to the nearest 0.1 kg and height was assessed using a wall-mounted stadiometer to the nearest 0.1 cm, both while participants were wearing light clothes and without socks and shoes. Body mass index (BMI) (kg/m²) was then calculated. Treatment (placebo or calcium) over the 5 years of the CAIFOS trial was included as a covariate. Diabetes status was determined by the use of antidiabetic medications (oral hypoglycemic agents or insulin). Medications were verified by participants’ general practitioner where possible and were coded (T89001–T90009) using the International Classification of Primary Care-Plus (ICPC-Plus) method, which allows aggregation of different terms for similar pathologic entities as defined by the International Classification of Diseases (ICD)-10 coding system (26). Prevalent atherosclerotic vascular disease (ASVD) was obtained from primary discharge diagnoses from hospital records (1980–1998) as described previously (27). Prevalent falls were determined by asking participants at their initial clinical visit if they experienced a fall (yes/no) in the previous 3 months.

Serum Creatinine and Cystatin C

Fasting blood samples for creatinine and cystatin C were collected at baseline (1998) (28). Serum creatinine was analyzed using an isotope dilution mass spectrometry traceable Jaffe kinetic assay for creatinine on a Hitachi 917 analyzer (Roche Diagnostics GmbH, Mannheim Germany) (11). Serum cystatin C was measured on the Siemens Dade Behring Nephelometer, traceable to the International Federation of Clinical Chemistry Working Group for Standardization of Serum Cystatin C and the Institute for Reference Materials and Measurements certified reference materials (11). The Cr:Cyc was calculated as creatinine (mg/dL)/cystatin C (mg/L). Higher Cr:Cyc would be indicative of greater relative muscle mass (19).

Measures of Sarcopenia

Muscle strength (hand grip strength) and physical function (Timed-Up-and-Go [TUG]) were assessed in 1 112 women at baseline (1998). Grip strength was assessed using a dynamometer (Jamar Hand Dynamometer; Lafayette Instrument Company) and TUG as the time taken for an individual to rise from a chair, walk 3 m, turn around, and return to sit on the chair. The interobserver coefficient of variation error was 7% for hand grip strength and 6% for TUG in our laboratory as assessed on a random sample of 30 women. ALM was evaluated in a subset of 334 women at baseline and 1 year (1998, 1999) using a Hologic Acclaim QDR4500A DXA machine (software version: 8.20; Hologic Corp., Waltham, MA). ALM was then adjusted for height (ALM/height²) based on recommendations by the revised European Working Group for Sarcopenia in Older People (EWGSOP2) (8).

Clinical Outcomes: Self-Reported Incident Falls and Fall-Related Hospitalization

Self-reported falls were obtained primarily via telephone interviews by a team member every 4 months over 5 years. However, if the 4-month follow-up coincided with a prescheduled clinic visit as part of the CAIFOS trial, this data was obtained in person. Falls were defined as unintentionally coming to rest on the ground, floor, or other lower level (29). This method was adopted to limit recall bias while enabling “time-to-event” to be calculated. Data relating to fall-related hospitalizations over 14.5 years were extracted from the Western Australia Hospital Morbidity Data Collection (HMDC), via the Western Australian Data Linkage System (Department of Health Western Australia, East Perth, Australia). HMDC records were obtained for each of the study participants from the date of their clinical visit in 1998. We had 14.5 years of follow-up for fall-related hospitalizations and deaths (all-cause) for all women that remained in Western Australia over the study period. These falls were considered injurious as it was serious enough to require hospitalization. Falls-related hospitalizations were identified using the international classification of external causes of injury codes and the ICD-coded discharge data pertaining to all public and private inpatient admissions in Western Australia. Ascertainment of hospitalizations, independent of self-report, avoids potential limitations of patient self‐reporting. As described previously (2), falls from standing height or less, not resulting from external force, were included (ICD‐10 codes W01, W05–W08, W10, W18, and W19). Where ICD-10-coded death data were not yet available, we used Parts 1 and 2 of the death certificates or all diagnosis text fields from the death certificate.

Statistical Analysis

Statistical analysis was performed using IBM SPSS Statistics for Windows, version 25.0 (IBM Corp., Armonk, NY), Stata software, version 14 (StataCorp LLC, College Station, TX), and R software (version 3.4.2; R Foundation for Statistical Computing, Vienna, Austria) (30). Generalized linear models were used to examine the association between Cr:Cyc (exposure) with grip strength, TUG, ALM/height², and ALM/BMI. To investigate the potential nonlinearity of the relationships, exposures were modeled using restricted cubic splines. p Values for the overall effect of the exposure on the outcomes (false discovery rate corrected) and for a test of nonlinearity were obtained using likelihood ratio tests to compare appropriate nested models. Associations are presented graphically using the “effects” R package (31). To determine where significant differences between quartiles (Q) of Cr:Cyc exist, ratios of means and 95% confidence intervals (CIs) were obtained from the model with the exposure fitted as a continuous variable and are reported for the median Cr:Cyc in each Q relative to the median Cr:Cyc in Q4. Cox-proportional hazards modeling was used to investigate the relationship between Cr:Cyc and falls outcomes including (a) self-reported falls and (b) fall-related hospitalizations. Global tests (estat phtest) indicated proportional hazards assumptions were not violated when considering the relationship between Cr:Cyc and 5-year self-reported falls (p > .05). However, for 14.5-year fall-related hospitalizations, the proportional hazards assumption was violated (p = .033). We therefore truncated the data to 12 years of follow-up; no violations of the assumptions were detected (p = .299). The 12-year analysis for fall-related hospitalizations is now presented as a primary outcome, while analysis for 14.5 years is reported within Supplementary Material. Kaplan–Maier survival curves for fall-related hospitalizations are presented in Supplementary Figure 2. Hazard ratios (HRs) and 95% CIs were obtained from the model with Cr:Cyc fitted as a continuous variable through a restricted cubic spline using the “rms” R package (32). HR estimates were graphed and calculated relative to a reference value being the median Cr:Cyc of Q4, while being plotted against the outcome variable, with 95% confidence bands provided. p Values for hazard ratios were obtained using Wald tests. In all graphs presented, the x-axis was truncated at 3 SD above the mean for visual simplicity only. Two models of adjustment were used for all regression analysis, (a) minimally adjusted: age, treatment code (placebo/calcium) and BMI and (b) multivariable-adjusted: minimally adjusted model plus smoking history (yes/no), diabetes status (yes/no), prevalent ASVD (yes/no), physical activity (kcal/day), and prevalent falls (yes/no).

Additional Analysis

We also undertook competing risks analyses based on Fine and Gray’s proportional subhazards model to account for the competing risks of mortality due to the advanced age of our participants. Considering Cr:Cyc was positively associated with grip strength, additional analysis was performed with its inclusion into the multivariable model for both fall outcomes. When considering 5-year self-reported falls and 12-year fall-related hospitalizations, the Youden index was used to determine an optimal cutoff for low Cr:Cyc. Subsequently, we examined the relationship between this low Cr:Cyc cut-point with falls outcomes using multivariable-adjusted Cox-proportional hazards models. As physical activity is known to influence muscle function and mass, we examined whether women who performed no structured physical activity had different Cr:Cyc in comparison to women who performed any structured physical activity. Finally, logistic regression was used to examine the relationship between baseline (1998) Cr:Cyc and the presence of sarcopenia as defined by the revised EWGSOP2 definition (8) (ALM/height² <5.5 kg/m² with grip strength <16 kg) at 5 years in 669 women who also had DXA-derived ALM and grip strength assessed in 2003.

Results

Baseline characteristics of the 1 118 participants according to quartiles of Cr:Cyc are presented in Table 1. Participants’ mean (±SD) age was 75.2 ± 2.7 years. Compared to women with the lowest Cr:Cyc (Q1), those with the highest Cr:Cyc (Q4) had a lower cystatin C consistent with a lower tissue mass (as identified by a lower BMI of 1.8 kg/m²) and weight, but a relatively greater muscle mass as identified by a higher creatinine. The proportion of individuals who reported a prevalent fall at baseline was also highest (15.4%) for women with the lowest Cr:Cyc (Q1) compared to all other quartiles.

Relationship to Muscle Function and DXA-Derived Lean Mass

Graphic representations of the multivariable-adjusted relationship between Cr:Cyc with grip strength, TUG, and ALM adjusted for height² and BMI are presented in Figure 1 (minimally adjusted is presented in Supplementary Figure 3), with estimated means and 95% CI presented in Table 2. Cr:Cyc was linearly related to grip strength, ALM/height², and ALM/BMI (p for nonlinearity = .07, .82, and .91, respectively). No relationship was observed for TUG. In the multivariable-adjusted model, for grip strength, women with the highest Cr:Cyc (Q4) were 5% (1.0 kg) stronger compared to individuals in Q1 (Table 2). Women with the highest Cr:Cyc (Q4) had a 3.6% (0.21 kg/m²) higher ALM/height² and 5.6% (0.031) higher ALM/BMI compared to individuals with the lowest Cr:Cyc (Q1) in the multivariable-adjusted model (Table 2).

Self-Reported Falls

Over 5 years (4 328 person-years) of follow-up (mean ± SD; 3.9 ± 1.6 years), 28.5% (319 of 1 118) of women experienced a self-reported fall. The linear relationship between Cr:Cyc and self-reported falls is presented in Figure 2 (p for nonlinearity = .907). Each SD decrease in Cr:Cyc was associated with higher hazards for a self-reported fall in the minimally and multivariable-adjusted model (HR 1.20; 95% CI 1.06–1.34, p = .003, for both). The proportion of women who reported an incident fall in Q1 of Cr:Cyc (36.1%) was higher compared to Q4 (22.9%). Women in Q1 and Q2 of Cr:Cyc also had 150% and 134% greater hazard, respectively, for a self-reported fall than those in Q4 in the multivariable-adjusted model (Table 3).

Fall-Related Hospitalizations Over 12 Years

Over 12 years (10 888 person-years) of follow-up (mean ± SD; 9.7 ± 3.2 years), 29.2% (326/1 118) of women experienced a fall-related hospitalization. Each SD decrease in Cr:Cyc was associated with higher hazards for a fall-related hospitalization in the minimally (HR 1.20; 95% CI 1.07–1.35, p = .002) and multivariable-adjusted model (HR 1.19; 95% CI 1.06–1.34, p = .003). The relationship between Cr:Cyc and fall-related hospitalization is presented in Figure 2. The proportion of women who experienced a fall-related hospitalization in Q1 of Cr:Cyc (33.5%) was higher compared to Q3 and Q4 (26.5% and 23.6%, respectively; Table 2). In the minimally adjusted model, compared to Q4 of Cr:Cyc, women in Q3, Q2, and Q1 had 122%, 132%, and 144% greater hazard for a fall-related hospitalization (Table 2). The aforementioned hazards in Q3, Q2, and Q1 increased to 124%, 142%, and 153%, respectively, in the multivariable-adjusted model. We have also presented a detailed description and a diagrammatic presentation (Supplementary Figure 4) of the relationship between Cr:Cyc and fall-related hospitalizations over 14.5 years in Supplementary Material. Briefly, compared to Q4, women in Q3, Q2, and Q1 had 112%, 133%, and 148% greater hazard for a fall-related hospitalization over 14.5 years (Supplementary Table 1).

Additional Analysis

When considering the subdistribution hazard ratios that accounted for the competing risk of mortality, each SD decrease in Cr:Cyc was associated with a 20% greater risk for a fall-related hospitalization in both the minimal and multivariable-adjusted model (sHR 1.20; 95% CI 1.08–1.34, p = .001). The inclusion of grip strength to the multivariable-adjusted model did not alter the relationship between Cr:Cyc and fall-related hospitalizations (per SD decrease, HR 1.18; 95% CI 1.06–1.33, p = .004). Based on the Youden index, a cut-point for low Cr:Cyc of <0.75 and <0.85 for 5-year self-reported falls and 12-year fall-related hospitalizations was obtained, respectively. Women with low Cr:Cyc had greater risk for a self-reported fall (HR 1.47; 95% CI 1.16–1.86, p = .002) and fall-related hospitalization (HR 1.43; 95% CI 1.13–1.79, p = .002), compared to women with higher Cr:Cyc. The results remained similar when considering fall-related hospitalizations over 14.5 years (Supplementary Material). The Cr:Cyc in women who did not perform any structured physical activity (n = 274, median [interquartile range {IQR}] 0.80 [0.72–0.89]) was significantly lower (p < .001) compared to women who performed any structured physical activity (n = 844, median [IQR] 0.84 [0.75–0.98]). At 5 years, 8.5% (57 of 669) of women presented with EWGSOP2-defined sarcopenia. The nonlinear relationship (p for nonlinearity = .030) between Cr:Cyc and sarcopenia is presented in Supplementary Figure 5. Specifically, women in Q1, Q2, and Q3 had 379%, 283%, and 148% higher odds for sarcopenia at 5 years compared to those in Q4 of Cr:Cyc in the multivariable-adjusted analysis (Supplementary Table 2).

Discussion

The main findings from this study are that Cr:Cyc was inversely associated with the risk for long-term incident self-reported falls and fall-related hospitalizations in community-dwelling older women. Specifically, compared to women with the highest Cr:Cyc (Q4), those with the lowest Cr:Cyc (Q1) had 150% and 153% greater risk for a self-reported fall over 5 years and a fall-related hospitalization over 12 years, respectively. We also report a positive relationship between Cr:Cyc with ALM (height or BMI adjusted) and grip strength, as well as reduced likelihood for sarcopenia 5 years later. Collectively, our findings demonstrate the potential for Cr:Cyc to be considered a biomarker for measures of sarcopenia and falls in older women. Specifically, Cr:Cyc may be a useful measure of muscle that makes an independent contribution to falls risk, in addition to strength (2).

To our knowledge, this is the first study to demonstrate that lower Cr:Cyc was associated with a greater risk for self-reported falls and injurious falls requiring hospitalization. These findings are especially relevant because the underlying causes that may be related to neuromuscular and/or cognitive impairments may differ between self-reported and injurious falls (33). Hence, these falls types should be distinguished and not considered interchangeably. Previously, it has been reported that creatinine (34) and cystatin C (35), when examined individually, may be considered a risk factor for self-reported falls. Specifically, in a study of 186 community-dwelling older adults (>70 years, ~52% female), low creatinine clearance (<65 mL/min) was associated with up to a four-fold higher incidence for self-reported falls over 9 months (34). However, caution should be exercised when trying to interpret creatinine levels, particularly in individuals with compromised glomerular filtration, as this can increase circulating levels (36). Alternatively, in a small study of 74 institutionalized older individuals (mean age 84 years, 80% women), higher cystatin C especially in women was identified as a falls risk factor over 20 months (35). Despite demonstrating a link to falls, when such biomarkers are considered individually, they are unable to provide an estimate of relative muscle mass. To obtain a measure of relative muscle mass, it is important to acknowledge that creatinine is produced by myocytes and cystatin C from all nucleated cells. Therefore, a combination of creatinine and cystatin C, expressed as a ratio, may serve as an indicator of relative muscle mass, which we have also reported to be associated with muscle strength.

The importance of muscle strength in falls prevention is widely recognized (37). We previously demonstrated that greater muscle strength, assessed via hand grip strength (per SD increase of 5 kg), was associated with a 27% lower risk for a fall-related hospitalization in this cohort (2). Although hand grip strength can be performed relatively quickly with a dynamometer, it is rarely assessed in clinical practice as part of routine medical screening (38). Therefore, investigating commonly assessed biomarkers as part of standard clinical assessment to highlight deficiencies in muscle strength is a viable addition (and/or alternative). Our results demonstrating a positive relationship between Cr:Cyc and grip strength is consistent with previous research (20,21). For example, in a Chinese cohort of community-dwelling middle-aged and older adults (aged ≥40 years, 1 098 men and 1 241 women), hand grip strength was positively associated with Cr:Cyc among men (r = 0.38) and women (r = 0.41) (21). A separate cohort of 796 community-dwelling Japanese women aged 60 years or older found that compared to individuals with the lowest Cr:Cyc (Q1), women with the highest Cr:Cyc (Q4) had ~15% stronger grip strength (19.9 vs 22.9 kg) (20). Although larger in magnitude, these findings are comparable to our findings where women with the highest Cr:Cyc (Q4) had 5% higher grip strength than individuals with the lowest Cr:Cyc (Q1). While such differences may appear small, every kilogram increase in grip strength was associated with a 4% lower relative hazard for a fall-related hospitalization over 12 years in this cohort (data not shown). The results highlight potential clinical significance of these findings. In contrast, no such relationship was observed in our study between Cr:Cyc and physical function assessed by the TUG test. Noteworthy, Kusunoki et al. (19) recorded a weak correlation (r = 0.18, p < .05) between Cr:Cyc and normal gait speed in 464 older community-dwelling Japanese women (mean age 72 years). However, the differing assessment of physical function may limit direct comparison between studies. Furthermore, the lack of association between Cr:Cyc and physical function may relate to the TUG test requiring coordination of various functional qualities such as muscle power, gait, and dynamic balance that may not be adequately captured by Cr:Cyc. Given Cr:Cyc may be considered a marker of relative muscle mass, perhaps it is unsurprising that it was associated with muscle strength (which is generally shown to be strongly associated with lean mass) but not with TUG performance, especially in older women (39), that involves other physical components beyond just muscle strength. Noteworthy, when adopting a direct measure of whole-body muscle mass such as the D3Cr method, higher muscle mass has been associated with better physical performance (Short Physical Performance Battery and RAND-36) in a comparable cohort of community-dwelling older women (n = 74, mean age ~82 years). However, no differences in hand grip strength were observed in women with higher versus lower muscle mass (40). Despite contrasting results to the current investigation, both studies support a link between muscle and physical attributes that warrant further investigation.

In the current investigation, we have also demonstrated a positive relationship between Cr:Cyc and height and BMI-adjusted ALM. Although previous work has reported similar findings, these studies have typically used BIA that provides an estimate of fat-free mass (FFM), whereas DXA is able to provide a better indirect estimate of lean mass (18,19). Previous studies have reported that Cr:Cyc was significantly associated with BIA-derived appendicular FFM (ALM/height²) in 213 men (β = 0.23) and 464 women (β = 0.18) in multiple-regression analysis (19). Similarly, a study in 796 healthy community-dwelling individuals (≥60 years) found that women with the lowest Cr:Cyc (Q1) had 271% higher odds for presenting with low femoral muscle cross-sectional area (assessed by CT) compared to women with the highest Cr:Cyc (Q4) (20). Although we have previously reported that greater ALM is not associated with fall-related hospitalizations (2), a higher ALM in this study (adjusted for height or BMI; 3.7%–5.0%) may be associated with other favorable outcomes, such as improvements in cardiometabolic health (41).

Despite such findings, the clinical utility of Cr:Cyc to predict “low ALM” (according to various sarcopenia definitions) was poor (AUC range 0.505–0.558) in 371 older adults (mean age 71 years) (18). Although low ALM presents as a key characteristic of most sarcopenia definitions (8,42), it does not appear to be a risk factor for clinical outcomes such as injurious falls and mortality in community-dwelling older women (2,5). Nevertheless, when considering women with low height-adjusted ALM in conjunction with weak grip strength (EWGSOP2 defined sarcopenia), our subgroup analysis in 669 women showed that lower Cr:Cyc was associated with increased odds of sarcopenia 5 years later. While such findings contrast 2 previous cross-sectional studies (15,18), this may also be related to the assessment method of ALM (DXA vs BIA), the inclusion of men, the relatively small sample sizes (n = 100–371), and/or the definition of sarcopenia adopted. While Cr:Cyc may serve as an indirect measure of relative muscle mass, future work should consider examining its relationship with direct measures of muscle mass (eg, D3Cr method) that have been associated with physical performance (40).

Limitations of the current study include its observational nature, which increases the possibility of bias due to residual confounding. Therefore, causal links cannot be established. Nonetheless, we considered a range of falls risk factors (eg, falls history, prevalent diseases, physical activity) in our multivariable-adjusted model. Furthermore, despite self-reported falls assessed every 4 months via telephone interviews, it is possible that these falls may be affected by recall bias. This study also consisted primarily of community-dwelling Caucasian women, so results may not be generalized to other populations such as older men or younger cohorts. However, research in older women is especially pertinent because they represent a high-risk population for falls and resulting injuries such as fractures. Noteworthy, despite cut-points for low Cr:Cyc (obtained from the Youden index) being associated with greater relative hazards for both long-term self-reported falls and fall-related hospitalizations, further research into its clinical utility and external validation is warranted. Furthermore, due to lack of agreement on how to define sarcopenia, we focused on examining the relationship between Cr:Cyc with common measures of sarcopenia (muscle strength, physical function, and ALM), while considering a highly relevant clinical outcome such as falls (7). Strengths of this study include the prospective design and population-based setting with ascertainment of injurious falls-related hospitalizations (in conjunction with self-reported falls over 5 years) with almost no loss to follow-up. Our data provide a unique opportunity to examine different types of falls including injurious fall hospitalization, which are rarely reported in the literature. The seriousness of injurious falls is demonstrated by only including falls where the injuries were deemed severe enough (eg, large lacerations and major contusions, loss of consciousness) upon presentation to hospital emergency departments that the patient required admission. Finally, as per sarcopenia criteria (7,8,42), we assessed muscle strength and physical function using well-established tests such as hand grip strength and TUG using standardized assessment, as well as lean mass using DXA.

In conclusion, a lower Cr:Cyc may be considered a risk factor for falls in community-dwelling older women. Similarly, we report that higher Cr:Cyc was associated with greater muscle strength, DXA-derived ALM (height-adjusted), and reduced likelihood for sarcopenia 5 years later. Collectively, such findings support further investigation into the use of Cr:Cyc as a muscle biomarker to help clinicians identify individuals at risk of falls (and sarcopenia) for early inclusion into evidence-based primary prevention programs targeting improvements to diet and exercise.

Acknowledgments

The authors wish to thank the staff at the Western Australia Data Linkage Branch, Hospital Morbidity Data Collection and Registry of Births, Deaths and Marriages for their work on providing the data for this study.

Funding

The Perth Longitudinal Study of Ageing in Women (PLSAW) was funded by Healthway, the Western Australian Health Promotion Foundation and by project grants 254627, 303169, and 572604 from the National Health and Medical Research Council (NHMRC) of Australia. M.S. is supported by a Royal Perth Hospital Career Advancement Fellowship (ID: CAF 130/2020). D.S. is supported by a NHMRC of Australia Emerging Leader (Level 2) Fellowship (ID: GNT1174886). J.M.H. is supported by an NHMRC of Australia Senior Research Fellowship (ID: 1116973). J.R.L. is supported by a National Heart Foundation of Australia Future Leader Fellowship (ID: 102817). None of these funding agencies had any role in the conduct of the study; collection, management, analysis, or interpretation of the data; or preparation, review, or approval of the manuscript.

Conflict of Interest

None declared.

Author Contributions

Conception and study design: M.S., J.D.V., D.S., R.M.D., G.D., and J.R.L. Study conduct and data collection: K.Z., W.H.L., J.R.L., and R.L.P. Data interpretation and analysis: M.S., J.M.H., and J.R.L. Drafting of the manuscript: M.S., J.D.V., D.S., R.M.D., and J.R.L. All authors reviewed the manuscript and approve the final version. M.S. and J.R.L. take responsibility for the integrity of the data analysis.

References