Glomerular filtration rate is estimated either based on creatinine (eGFRCr), which tends to be high in people with low muscle mass, or cystatin C (eGFRCys), which is not as affected by muscle mass.1 It is not uncommon to see clinic patients with discrepancies between the two. We hypothesized that prognostic information is embedded in the difference in these markers. We propose that older adults with eGFRCys > eGFRCr will have less prevalent frailty, and will be at lower risk for incident frailty and mortality, compared with those having a minimal difference between the two estimates.
The Cardiovascular Health Study (CHS) assessed community-dwellers aged ≥ 65 years; its design was described previously.2 eGFRDiff was defined as the difference eGFRCys - eGFRCr, using CKD-EPI equations3 at baseline. Pre-frailty and frailty were assessed using the Fried score at baseline and year 3 for the first cohort and year 4 for the African-American cohort. Mortality was reviewed and ascertained by a Mortality Review Committee. We evaluated eGFRDiff continuously, then as follows: (1) eGFRDiff < −15 (“negative group”); (2) −15 ≤ eGFRDiff < +15 (“reference”); and (3) eGFRDiff ≥ +15 (“positive group”). The cross-sectional association of eGFRDiff with prevalent frailty was examined using logistic regression and was tested for an interaction by gender and race. Among those without prevalent frailty at baseline and with follow-up assessment of frailty (n=1955), the associations of eGFRDiff with incident frailty and all-cause mortality were evaluated 3 to 4 years after the baseline visit, using Poisson regression models, and were tested for interactions by age. Detailed methods are available in supplements.
4635 participants (78.7% of CHS) had baseline serum creatinine, cystatin C, and frailty measurements. The average age was 72 (±5) years, 39% were male, 16% were African-American, and 15% had diabetes. Average eGFRCys was 72 (±18), eGFRCr was 73 (±18) and eGFRDiff was −1.4 (±14) mL/min/1.73 m2. The combined prevalence of pre-frailty and frailty at baseline was 53% (2450/4635). When divided into groups based on eGFRDiff, those in the negative group were more likely to be female, have diabetes and be frail at baseline, relative to the reference group.
Each 15-point increment in eGFRDiff was associated with a 27% and 49% lower odds of prevalent pre-frailty (OR=0.73; 95%CI [0.68; 0.79]) and frailty (OR=0.51; 95%CI [0.43; 0.60]) respectively. There was no interaction by gender (p= 0.57 and 0.17 for pre-frailty and frailty respectively), nor race (p= 0.13 and 0.23). Similar results were found when comparing participants across eGFRDiff categories (Table 1).
Table 1.
Association of eGFRDiff groups with prevalent frailty
| Outcome = Pre-frailty | eGFRDiff (per SD) | eGFRDiffGroups | ||
| Sample size | Negative (< −15) | Reference (−15 ≤ to < +15) | Positive (≥ + 15) | |
| n=740 | n=3362 | n=533 | ||
| # of events | 2139 | 407 | 1541 | 191 |
| OR (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) | |
| Unadjusted | 0.71 (0.66; 0.76) | 1.66 (1.40; 1.98) | 1 (Ref) | 0.62 (0.51; 0.75) |
| Model 1 | 0.70 (0.65; 0.76) | 1.72 (1.43; 2.06) | 1 (Ref) | 0.65 (0.54; 0.80) |
| Model 2 | 0.73 (0.68; 0.79) | 1.59 (1.32; 1.91) | 1 (Ref) | 0.70 (0.57; 0.86) |
| Outcome = Frailty | ||||
| # of events | 311 | 78 | 214 | 19 |
| OR (95% CI) | OR (95% CI) | OR (95% CI) | OR (95% CI) | |
| Unadjusted | 0.53 (0.46; 0.60) | 2.30 (1.72; 3.07) | 1 (Ref) | 0.44 (0.27; 0.72) |
| Model 1 | 0.50 (0.43; 0.58) | 2.61 (1.88; 3.62) | 1 (Ref) | 0.56 (0.34; 0.93) |
| Model 2 | 0.51 (0.43; 0.60) | 2.38 (1.70; 3.33) | 1 (Ref) | 0.56 (0.33; 0.95) |
model 1 = adjusted for age (per 5 years), gender, race, CRP, serum albumin and eGFRCr category
model 2 = model 1 + HTN, diabetes, on BP meds at baseline, HDL, cholesterol, smoking, and prevalent CHD
Among the 1955 participants who were not frail at baseline, and had frailty assessed at follow-up, 39% became pre-frail (771/1955), and 4% frail (75/1955). Each 15-point increment in eGFRDiff was significantly associated with 55% and 48% lower incidence rate of frailty and mortality respectively (Table 2).
Table 2.
Association of baseline eGFRDiff (per 15-point increment) and eGFRDiff groups with incident frailty and mortality at follow-up time point.
| eGFRDiff | eGFRDiffGroups | |||
|---|---|---|---|---|
| Per 15 point increment | Negative (< −15) | Reference (−15 ≤ to < +15) | Positive (≥ + 15) | |
| Incidence Rate (95% CI) | Incidence Rate (95% CI) | Incidence Rate (95% CI) | Incidence Rate (95% CI) | |
| Pre-frailty | n = 1821 | n = 181 | n = 1355 | n = 285 |
| # events = 771 | # events = 85 | # events = 586 | # events = 100 | |
| Unadjusted | 0.82 (0.76; 0.89) | 1.16 (0.93; 1.46) | 1 (Ref) | 0.68 (0.55; 0.84) |
| Fully adjusted * | 0.89 (0.81; 0.97) | 1.06 (0.83; 1.35) | 1 (Ref) | 0.81 (0.65; 1.01) |
| Frailty | n = 1125 | n = 123 | n = 809 | n = 193 |
| # events = 75 | # events = 27 | # events = 40 | # events = 8 | |
| Unadjusted | 0.48 (0.38; 0.61) | 5.32 (3.27; 8.68) | 1 (Ref) | 0.76 (0.36: 1.63) |
| Fully adjusted * | 0.45 (0.34; 0.61) | 6.97 (3.89; 12.49) | 1 (Ref) | 0.88 (0.40; 1.94) |
| Mortality | n = 1109 | n = 111 | n = 807 | n = 191 |
| # events = 59 | # events = 15 | # events = 38 | # events = 6 | |
| Unadjusted | 0.66 (0.50; 0.88) | 3.20 (1.76; 5.82) | 1 (Ref) | 0.60 (0.25; 1.42) |
| Fully adjusted * | 0.52 (0.37; 0.74) | 6.57 (3.27; 13.19) | 1 (Ref) | 0.59 (0.24; 1.44) |
adjusted for age (per 5 years), gender, race, CRP, serum albumin, eGFRCr category, HTN, diabetes, on BP meds at baseline, HDL, cholesterol, smoking, and prevalent CHD
CHS participants are community dwelling, hence more independent for activities of daily living than an unselected group of older adults. Yet, only 47% were neither pre-frail nor frail at baseline. We demonstrate that lower eGFRDiff is strongly associated with prevalent and incident frailty and all-cause mortality, in fully adjusted models. These findings, along with those from SPRINT, support the idea that eGFRDiff is an early marker of frailty, poor outcomes and mortality.
Several studies4–6 found that lower eGFRCys is associated with higher risk of frailty, whereas lower eGFRCr is not. Cystatin C has also been shown7 to be a better predictor for mortality than creatinine. One unique feature of the current analysis is examining the eGFRDiff within individual patients, rather than patterns of differences across the population. Our data suggest that individual patients with a lower eGFRCys than eGFRCr, i.e. a negative eGFRDiff, are at higher risk of frailty.
Those in the negative group were found to have higher BMI compared to both the reference and positive groups. This is possibly due to high BMI being associated with greater adipose tissue, contributing to higher levels of cystatin C, but not correlating well with muscle mass.8 Future studies including body composition measures are needed to better understand the determinants of eGFRDiff.
Strengths of this study included the use of a previously validated frailty scale,4,9 and centrally measured creatinine and cystatin C. The length of follow-up allowed us to capture a substantial percentage of outcomes. One important limitation was that incident frailty was not assessed until the follow-up visits, preventing the use of proportional hazards regression models.
In conclusion, one in four participants of the CHS cohort had a difference between eGFRCys and eGFRCr of ≥15 mL/min/1.73m2. Elders in the negative group were more likely to have prevalent frailty, slower walking speed, and more comorbidities. Even those in this subgroup who were without frailty at baseline were at much higher risk for both incident frailty and mortality during follow-up. Considering eGFRDiff as a marker of patients’ functional status may be helpful to clinicians as an indicator of poor prognosis.
Supplementary Material
Table S1. Baseline Characteristics of CHS participants by eGFRDiff (eGFRCys - eGFRCr)
Table S2. Association of baseline eGFRDiff (per 15-point increment) with incident pre-frailty, frailty, and mortality at follow-up, stratified by eGFRCr, and eGFRCys.
Figure S1. Study flow diagram
Acknowledgements:
Data acquisition was performed by the CHS.
Support: OAP was supported by grant 5T32DK104717–02, by the American Kidney Fund Clinical Scientist in Nephrology Fellow Program, and Akebia Therapeutics, Inc. DER was supported by grant K23 DK09152 and Veterans’ Affairs Merit award H160180 (Health Services Research and Development). The funders did not have a role in study design, data collection, analysis, reporting, or the decision to submit for publication.
Footnotes
Financial Disclosure: The authors declare that they have no relevant financial interests.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Ferguson TW, Komenda P, Tangri N. Cystatin C as a biomarker for estimating glomerular filtration rate. Curr Opin Nephrol Hypertens. 2015;24(3):295–300. doi: 10.1097/MNH.0000000000000115 [DOI] [PubMed] [Google Scholar]
- 2.Fried LP, Borhani NO, Enright P, et al. The cardiovascular health study: Design and rationale. Ann Epidemiol. 1991;1(3):263–276. doi: 10.1016/1047-2797(91)90005-W [DOI] [PubMed] [Google Scholar]
- 3.Inker LA, Schmid CH, Tighiouart H, et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med. 2012;367(1):20–29. doi: 10.1056/NEJMoa1114248 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dalrymple LS, Katz R, Rifkin DE, et al. Kidney Function and Prevalent and Incident Frailty. Clin J Am Soc Nephrol CJASN. 2013;8(12):2091–2099. doi: 10.2215/CJN.02870313 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hart A, Blackwell TL, Paudel ML, et al. Cystatin C and the Risk of Frailty and Mortality in Older Men. J Gerontol A Biol Sci Med Sci. 2017;72(7):965–970. doi: 10.1093/gerona/glw223 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Canney M, Sexton DJ, O’Connell MDL, Kenny RA, Little MA, O’Seaghdha CM. Kidney Function Estimated From Cystatin C, But Not Creatinine, Is Related to Objective Tests of Physical Performance in Community-Dwelling Older Adults. J Gerontol A Biol Sci Med Sci. 2017;72(11):1554–1560. doi: 10.1093/gerona/glx039 [DOI] [PubMed] [Google Scholar]
- 7.Krolewski AS, Warram JH, Forsblom C, et al. Serum Concentration of Cystatin C and Risk of End-Stage Renal Disease in Diabetes. Diabetes Care. 2012;35(11):2311–2316. doi: 10.2337/dc11-2220 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Reinders I, Murphy RA, Martin KR, et al. Body Mass Index Trajectories in Relation to Change in Lean Mass and Physical Function: The Health, Aging and Body Composition Study. J Am Geriatr Soc. 2015;63(8):1615–1621. doi: 10.1111/jgs.13524 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56(3):M146–156. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Baseline Characteristics of CHS participants by eGFRDiff (eGFRCys - eGFRCr)
Table S2. Association of baseline eGFRDiff (per 15-point increment) with incident pre-frailty, frailty, and mortality at follow-up, stratified by eGFRCr, and eGFRCys.
Figure S1. Study flow diagram