Abstract
Background
We sought to identify biomarkers that indicate low turnover on bone histomorphometry in chronic kidney disease (CKD) patients, and subsequently determined whether this panel identified differential risk for fractures in community-dwelling older adults.
Methods
Among CKD patients who underwent iliac crest bone biopsies and histomorphometry, we evaluated candidate biomarkers to differentiate low turnover from other bone disease. We applied this biomarker panel to 641 participants in the Health Aging and Body Composition Study (Health ABC) study with estimated glomerular filtration rate (eGFR) less than 60 mL/min/1.73 m2 who were followed for fracture. Cox proportional hazards models evaluated the association of bone mineral density (BMD) with fracture risk and determined whether biomarker-defined low bone turnover modified fracture risk at any level of BMD.
Results
In 39 CKD patients age 64 ± 13 years, 85% female, with mean eGFR 37 ± 14 mL/min/1.73 m2 who underwent bone biopsy, lower fibroblast growth factor (FGF)-23, higher ɑ-Klotho, and lower parathyroid hormone (PTH) indicated low bone turnover in accordance with bone histomorphometry parameters (individual area under the curve = 0.62, 0.73, and 0.55 respectively; sensitivity = 22%, specificity = 100%). In Health ABC, 641 participants with CKD were age 75 ± 3 years , 49% female, with mean eGFR 48 ± 10 mL/min/1.73 m2. For every SD lower hip BMD at baseline, there was an 8-fold higher fracture risk in individuals with biomarker-defined low turnover (hazard ratio 8.10 [95% CI, 3.40-19.30]) vs a 2-fold higher risk in the remaining individuals (hazard ratio 2.28 [95% CI, 1.69-3.08]) (Pinteraction = .082).
Conclusions
In CKD patients who underwent bone biopsy, lower FGF-23, higher ɑ-Klotho, and lower PTH together had high specificity for identifying low bone turnover. When applied to older individuals with CKD, BMD was more strongly associated with fracture risk in those with biomarker-defined low turnover.
Keywords: bone mineral density, bone turnover, fracture, parathyroid hormone, α-Klotho, fibroblast growth factor (FGF)-23, chronic kidney disease
Approximately 1 in 7 Americans has chronic kidney disease (CKD) (1). Risk of fractures in individuals with CKD is higher than in those without CKD across all age groups, yet managing fracture risk is complicated in this high-risk population (2, 3). Rather than simply having age-related bone loss, individuals with CKD may also have a variety of CKD-related bone abnormalities that often include inappropriately high or low bone turnover, with or without mineralization defects (4, 5). These pathologic processes have implications for treatment because some treatments lower bone turnover whereas others increase it. Yet, consideration of bone turnover in the assessment and treatment of bone disease has not been incorporated into routine clinical practice.
Low bone mineral density (BMD) has been associated with higher risk of fractures in CKD patients similar to the general population (6), but BMD measurement does not provide information on the underlying bone pathology. Transiliac crest bone biopsy with histomorphometry is considered the gold standard to diagnose the etiology of bone disease in CKD patients because it allows assessment of bone turnover, mineralization, and volume (7, 8). Yet bone biopsy is an invasive procedure and is available at only a few centers across the United States. Thus, there is interest in identifying panels of biomarkers that might classify individuals based on bone turnover status. The most common biomarkers used in the management of osteoporosis therapy include serum C-telopeptide, a marker of bone resorption, and serum propeptide type I collagen, a marker of bone formation. These markers are not useful in the assessment or management of CKD-related bone abnormalities, however, because they are cleared by the kidney. Instead, in the setting of CKD, clinicians most often use parathyroid hormone (PTH) and bone-specific alkaline phosphatase (BSAP) as well as calcium, calcitriol, and phosphorus levels to determine abnormalities in mineral metabolism prior to consideration of osteoporosis therapy. Just let it read” Thus, although bone turnover marekrs are promising for osteoporosis therapy, including serum C and propeptide type I collagen, are promising for osteoporosis therapy in the general population, they do not capture mineral metabolism abnormalities potentially contributing to altered bone turnover in individuals with CKD. Prior studies that evaluated individual or combinations of bone turnover biomarkers were generally from single-center studies, and the majority evaluated fracture risk in dialysis patients on dialysis rather than nondialysis CKD patients (9-11). Thus, the clinical implications of using biomarkers remain uncertain.
We sought to establish a panel of serum biomarkers indicative of low bone turnover from patients with nondialysis CKD who had undergone bone biopsies for clinical indications. Specifically, we sought to define a panel with high specificity that would be useful to distinguish a subset of patients with very high likelihood of low bone turnover that might obviate the need for a bone biopsy.
We then applied the biomarkers and cutoff values indicative of low turnover bone disease to a community-dwelling population of older individuals with CKD in the Health Aging and Body Composition Study (Health ABC). This larger cohort had BMD measures and long-term follow-up for fractures, but did not undergo bone biopsy, which is typical of patients with CKD in the general population. We sought to determine whether the biomarker panel defined by the bone biopsy cohort could identify a subset at unique risk of fracture above and beyond BMD data in Health ABC.
Materials and Methods
Identification of low bone turnover biomarkers based on bone biopsy in chronic kidney disease patients
The University of Kentucky Metabolic Bone Disease Clinic received consent from 39 patients age 18 years or older with CKD (defined as eGFR < 60 mL/min/1.73 m2) to undergo bone biopsy and blood draw as part of workup for osteopenia or osteoporosis found by dual energy x-ray absorptiometry and for bone histological analyses to identify the type of renal osteodystrophy (5, 12-14). Individuals on dialysis or with a history of renal transplantation, history of parathyroidectomy, current or prior use of anticonvulsant, long-term steroids, or medications known to affect bone metabolism including antiresorptive medications were excluded. Individuals were also excluded if they had documented chronic alcoholism, drug addiction, malabsorption, malignancy, Paget disease, osteogenesis imperfecta, hemiplegia/paraplegia, organic illness with potential influence on bone metabolism, or uncontrolled systemic illness. The study was reviewed and approved by the institutional review board at the University of Kentucky.
Bone samples were obtained by biopsies of the anterior iliac crest under local anesthesia and sedation, and then fixed in ethanol at room temperature, dehydrated, and embedded in methyl methacrylate as described previously (15). Sections were stained with the modified Masson-Goldner trichrome (16), the aurin tricarboxylic acid (17), and solochrome azurine stains (18). Unstained sections were prepared for phase-contrast and fluorescence light microscopy. Bone histomorphometry for static and dynamic parameters of bone structure, formation, and resorption was performed at a magnification of 200× using the Osteoplan II system (19-22).
Bone turnover was assessed based on reference values (15) by activation frequency (normal range, 0.49-0.72/year), bone formation rate/bone surface (normal range, 1.81-3.80 mm3/cm2/year), and osteoblasts surface per bone surface (normal range, 0.2%-3.5%) and osteoclasts surface per bone surface (normal range, 0.01%-1.90%) (23). Low bone turnover was defined by an activation frequency of less than 0.49 per year, below the normal range (15).
Intact fibroblast growth factor (FGF-23) was measured in serum using a Kainos 2-site enzyme-linked immunosorbent assay (ELISA) (Kainos Laboratories) (23). The intra-assay and interassay coefficients of variation (CVs) were 4.7% and 6.4%, respectively. Soluble α-Klotho was measured using a solid-phase sandwich ELISA (Immuno-Biological Laboratories) (23). The intra-assay and interassay CVs for the α-Klotho assay were 3.1% and 6.9%, respectively. Serum BSAP, a marker of bone formation, and tartrate-resistant acid phosphatase 5b (TRAP-5b), a marker of bone resorption, were measured using a Quidel ELISA (23). The intra-assay and interassay CVs for BSAP were 5.0% and 5.9%, and for TRAP-5b were 2.1% and 2.5%, respectively. PTH was measured in EDTA plasma using a 2-site immunoradiometric assay kit (N-tact PTH SP; DiaSorin) (24). The interassay CV was 8.6%. Serum sclerostin and dickkopf-related protein 1 (DKK1) levels were measured using a Biomedica ELISA kit. The intra-assay and interassay CVs for the sclerostin assay were 8.1% and 5.5% and for the DKK1 assay, they were 3.0%.
Assessment of biomarker-defined low bone turnover in the Health, Aging, and Body Composition Study
The Health ABC Study began in 1997 to 1998 as a longitudinal study of 3075 men and women age 70 to 79 years at baseline, who had no difficulty walking one-quarter mile or climbing up 10 steps. Participants were recruited through the University of Pittsburgh and the University of Tennessee Health Science Center in Memphis and were followed through 2011 for outcomes related to frailty and declines in function including fracture. The study was approved by the institutional review board at the participating institutions, including the coordinating center and the National Institute on Aging.
We measured eGFR using the CKD Epidemiology Collaboration creatinine–cystatin C equation in all participants, and identified 766 individuals with an eGFR less than 60 mL/min/1.73 m2(25). Prior studies had measured a subset of the biomarkers evaluated in the Kentucky bone biopsy cohort, namely FGF-23, α-Klotho, and PTH. Among the 766 CKD participants, 83 lacked FGF-23 measurements, 2 lacked PTH measurements, and 5 lacked α-Klotho measurements. Further, we excluded 35 individuals who reported using antiresorptive medications, resulting in a study sample of 641 with complete measures of bone turnover markers.
Measurement of biomarkers fibroblast growth factor-23, α-Klotho, and parathyroid hormone in the Health, Aging, and Body Composition Study
Biomarkers were measured in serum and plasma collected at a follow-up examination approximately 1 year after the Health ABC baseline. FGF-23 was measured using a commercial ELISA that detects the full-length intact peptide (Kainos Laboratories). Serum samples were stored at –70 °C and were not thawed until time of analysis. The assay has a limit of detection of 3 pg/mL. Samples were assayed in batches over 2 months and the between-batch CV was 10.7%. PTH was measured in EDTA plasma using a 2-site immunoradiometric assay kit (N-tact PTH SP; DiaSorin) (24). The interassay CV for this assay was 8.6%. Soluble α-Klotho was measured using a solid-phase sandwich ELISA (Immuno-Biological Laboratories) (26). This assay had a minimum detectible level of 6.15 pg/mL and an interassay CV of 15%. BSAP, TRAP5b, DKK1, and sclerostin were not measured in Health ABC.
Measurement of bone mineral density and incident fracture in the Health, Aging, and Body Composition Study
Total hip, femoral neck, and lumbar spine BMD in grams per centimeters squared were measured using dual energy x-ray absorptiometry (Hologic) at baseline and repeated at years 3, 5, 8, and 10 (27).
Information regarding incident fractures was obtained through alternating annual clinic visits and telephone interviews every 6 months (6). Beginning at the visit of biomarker measurement and followed forward, fractures were identified via self-report and subsequently validated by radiographic reports (6, 27). Vertebral fractures in this study were limited to clinical vertebral fractures, and pathological fractures were excluded (27).
Risk factors and covariates
Weight was measured on balance beam scales to the nearest 0.1 kg, and height was measured using a stadiometer to the nearest 0.1 cm. Body mass index (BMI) was calculated as weight in kilograms divided by height in centimeters squared (kg/cm2). Questionnaires collected information on age, race (non-Hispanic white or African American), alcohol use (average number of drinks per week), physical activity (kilocalorie per kilogram per week), smoking status (current, former, or never), and use of medications including antihypertensive medications, diuretics, and bisphosphonates. Systolic and diastolic blood pressures were measured 2 times using a conventional mercury sphygmomanometer in the seated position after 5 minutes of rest. The average of 2 measurements was used in analyses. Urine albumin was measured using a particle-enhanced turbidimetric inhibition immunoassay for direct quantification of albumin (Siemens), and urine creatinine was measured using a modified Jaffe method on a Siemens clinical chemistry analyzer (28, 29). Interassay CVs for albumin and creatinine were 6% and 1%, respectively. Serum 25-hydroxyvitamin D was measured using a 2-step radioimmunoassay (25-Hydroxyvitamin D 125I RIA kit; DiaSorin) in a laboratory that met the criteria for the Vitamin D External Quality Assessment Scheme (24). The CV for this assay was 6.8%.
Statistical analysis
In the Kentucky bone biopsy samples, we created receiver operating characteristic (ROC) curves to determine appropriate cutoffs for low turnover vs non–low bone turnover. Cutoff values for determination of low vs non–low bone turnover were obtained using the Youden J statistic, which gives the maximum potential effectiveness of a biomarker using the summary of the ROC analysis defined as J = maxc [sensitivity + specificity – 1] (30). These calculations were made using SPSS version 24 (SPSS, Inc). Area under the ROC curves (AUC) greater than or equal to 0.60 was used as a criteria for moving biomarkers forward for all except PTH, for which the AUC equaled 0.55. We elected to force PTH as a marker to move forward because it is the only marker currently used in clinical practice to define bone turnover in CKD patients. Because we found that the AUCs were 0.60 or greater for FGF-23 and α-Klotho, we used a biomarker panel that consisted of FGF-23, α-Klotho, and PTH to assess biomarker-defined low bone turnover in Health ABC. Biomarker measurements were performed separately in the Kentucky samples and in Health ABC, so we based our cutoffs on the Youden J index. These Youden J index cutoff values were all at or close to the median of the entire Kentucky population (47th, 49th, and 52nd percentiles for intact FGF-23, soluble ɑ-Klotho, and PTH, respectively). Thus, we used the median cut point of each assay in Health ABC, rather than the measured concentration cut points, to constitute biomarker-defined “low turnover” in Health ABC. Therefore, biomarkers indicative of low bone turnover in Health ABC included an intact FGF-23 measurement of less than 55pg/mL, α-Klotho greater than or equal to 589 pg/mL or greater, and PTH less than 39 pg/mL. To construct a biomarker panel with high specificity so individuals classified as having biomarker-defined low turnover would be very likely to have low turnover if they underwent bone biopsy, we combined all 3 variables to define low bone turnover, in which individuals simultaneously had below-median intact FGF-23, above-median α-Klotho, and below-median PTH values.
Using median cutoff values for FGF-23, α-Klotho, and PTH to determine biomarker-defined low bone turnover, we evaluated descriptive characteristics by individual bone turnover markers as well as biomarker defined low vs non–low bone turnover using means and proportions.
To determine associations between BMD and biomarker-defined low bone turnover with incident fracture, we used Cox proportional hazards models over a mean of 8.4 years of follow-up. We included low bone turnover vs non–low bone turnover as an interaction term with baseline hip and spine BMD for the outcome of incident fracture to determine whether low bone turnover modifies the association between baseline BMD and fracture risk in individuals with CKD.
In each analysis, we constructed a set of multivariable models. Model 1 adjusted for age, sex, race, and clinical site. Model 2 additionally adjusted for BMI, systolic blood pressure, eGFR, albumin/creatinine ratio, serum 25-hydroxyvitamin D, smoking status, alcohol use, diabetes, physical activity, and use of diuretics. We were limited by power to explore sex interaction in all final models or main effects stratified by sex. Analyses were performed using R, and P less than .05 was considered statistically significant.
Results
Identification of a biomarker panel for low bone turnover
Among 39 individuals in the University of Kentucky CKD cohort who had undergone anterior iliac crest bone biopsy for workup for bone disease, the average age was 64 ± 13 years, 85% were female, 18% were postmenopausal, and 97% were Caucasian. Mean eGFR was 37 ± 14 mL/min/1.73 m2. The mean serum concentrations of intact FGF-23, soluble ɑ-Klotho, and PTH were 148 ± 132 pg/mL, 713 ± 257 pg/mL, and 36 ± 32 pg/mL, respectively. Sixty-nine percent (N = 27) were found to have low bone turnover on biopsy (Table 1).
Table 1.
Population characteristics of 39 chronic kidney disease patients with transiliac bone biopsies from the University of Kentucky
| Age ± SD, y | 64 ± 13 |
|---|---|
| Women, n (%) | 33 (85) |
| Postmenopausal women, n (%) | 6 (18) |
| Black race, n (%) | 1 (3) |
| eGFR ± SD, mL/min/1.73 m2 | 37 ± 14 |
| Diabetes, n (%) | 10 (26) |
| Hypertension, n (%) | 17 (44) |
| CKD stage, n (%) | |
| IIIa | 15 (38) |
| IIIb | 10 (26) |
| IV | 9 (23) |
| V (nondialysis) | 5 (13) |
| Cause of kidney disease, n (%) | |
| Diabetes | 1 (3) |
| Hypertension | 2 (5) |
| Glomerulonephritis | 1 (3) |
| ADPKD | 1 (3) |
| Unspecified | 17 (44) |
| FGF-23 ± SD, pg/mL | 148 ± 132 |
| PTH ± SD, pg/mL | 36 ± 32 |
| α-Klotho ± SD, pg/mL | 713 ± 257 |
| BSAP ± SD, U/L | 26 ± 13 |
| TRAP5b ± SD, U/L | 3.17 ± 1.61 |
| DKK1 ± SD, pmol/L | 36 ± 16 |
| Sclerostin ± SD, pg/mL | 1323 ± 455 |
| Bone histomorphometry | |
| BFR/BS | 0.80 ± 0.91 |
| ACF | 0.27 ± 0.33 |
| OS/BS, % | 11.59 ± 12.18 |
| Ob.S/BS, % | 0.83 ± 1.33 |
| Oc.S/BS, % | 0.76 ± 0.89 |
Abbreviations: ACF, activation frequency; ADPKD, autosomal dominant polycystic kidney disease; BFR/BS, bone formation rate/bone surface; CKD, chronic kidney disease; DKK1, dickkopf-related protein 1; eGFR, estimated glomerular filtration rate; FGF, fibroblast growth factor; Oc.S/BS, osteoclast surface/bone surface; Ob.S/BS, osteoblast surface/bone surface; OS/BS, osteoid surface/bone surface; PTH, parathyroid hormone; TRAP5b, tartrate-resistant acid phosphatase 5b.
When evaluated individually, intact FGF-23 (AUC = 0.62), soluble ɑ-Klotho (AUC = 0.73), DKK1 (AUC = 0.67), BSAP (AUC = 0.62), TRAP-5b (AUC = 0.61), and sclerostin (AUC = 0.62) had AUCs of 0.60 or greater for low bone turnover status. PTH had an AUC equal to 0.55. Because DKK1, BSAP, TRAP-5b, and sclerostin were not measured in Health ABC, we limited our biomarker panel to FGF-23 and soluble α-Klotho, and included PTH because it is the only one of these markers used in contemporary clinical practice to assess bone turnover. Youden Index values corresponded to cutoff values 78.9 pg/mL for intact FGF-23, 663.30 pg/mL for soluble ɑ-Klotho, and 27.95 pg/mL for PTH. These cutoff values were all at or close to the median of the entire Kentucky study population(47th, 49th, and 52nd percentile for intact FGF-23, soluble ɑ-Klotho, and PTH, respectively). The combination of intact FGF-23 below the median, soluble ɑ-Klotho above the median, and PTH below the median was observed in 6 individuals (15%) in the bone biopsy cohort. All these individuals had low bone turnover on biopsy. Thus, the specificity of having all 3 biomarkers suggesting low bone turnover was 100% (95% CI, 74%-100%) and the sensitivity was 22% (95% CI, 7%-42%).
Health, Aging, and Body Composition Study participant characteristics
Characteristics of the 641 participants in the Health ABC cohort with an eGFR less than 60 mL/min/1.73 m2 at baseline are shown in Table 2. Overall, the mean age was 75 ± 3 years, 49% were female, and 37% were African American. Mean eGFR was 48 ± 10 mL/min/1.73 m2, 44% had diabetes, and 84% had hypertension. The mean total hip and lumbar spine BMD were 0.90 ± 0.16 and 0.90 ± 0.16 g/cm2, respectively. Median (25th, 75th percentile) values for FGF-23, ɑ-Klotho, and PTH were 55 pg/mL (42, 73), 589 pg/mL (457, 756), and 39 pg/mL (29, 56), respectively.
Table 2.
Baseline characteristics of 641 Health ABC participants with chronic kidney disease
| All | Bone turnover | ||
|---|---|---|---|
| Non-low | Lowa | ||
| No. | 641 | 538 | 103 |
| Age, mean (SD), y | 75 (3) | 75 (3) | 76 (3) |
| Female, n (%) | 313 (49) | 266 (49) | 47 (46) |
| Race | |||
| White, n (%) | 402 (63) | 332 (62) | 70 (68) |
| Black, n (%) | 239 (37) | 206 (38) | 33 (32) |
| Site | |||
| Memphis, n( %) | 319 (50) | 269 (50) | 50 (49) |
| Pittsburgh, n (%) | 322 (50) | 269 (50) | 53 (52) |
| Education | |||
| < High school, n (%) | 150 (24) | 126 (24) | 24 (23) |
| High school graduate, n (%) | 212 (33) | 176 (33) | 36 (35) |
| Postsecondary, n (%) | 277 (43) | 234 (44) | 43 (42) |
| SBP, mean (SD), mm Hg | 136 (23) | 137 (24) | 133 (19) |
| DBP, mean (SD), mm Hg | 70 (12) | 71 (12) | 68 (10) |
| Current smoker, n (%) | 74 (12) | 58 (10) | 18 (18) |
| BMI, mean (SD), kg/m2 | 28.1 (4.9) | 28.3 (5.0) | 27.0 (4.4) |
| Current drinker, n (%) | 293 (46) | 240 (45) | 53 (52) |
| Physical activity, median (25th, 75th percentile), kcal/kg/wk | 0 (0, 5) | 0 (0, 5) | 0 (0, 7.5) |
| Diabetes, n (%) | 283 (44) | 239 (44) | 44 (43) |
| Hypertension, n (%) | 538 (84) | 460 (86) | 78 (76) |
| Antihypertensive use, n (%) | 444 (69) | 385 (72) | 59 (57) |
| Vitamin D supplement, n (%) | 53 (8) | 46 (9) | 7 (7) |
| Any osteoporosis drugs, n (%) | 0 | 0 | 0 |
| Diuretics use, n (%) | 260 (41) | 231 (43) | 29 (28) |
| eGFR, mean (SD), mL/min/1.73 m2 | 48 (10) | 47 (10) | 53 (6) |
| UACR, median (25th, 75th percentile), mg/g | 11 (5, 37) | 12 (5, 38) | 9 (5, 35) |
| Prevalent osteoporosis | |||
| Definite, n (%) | 16 (3) | 13 (3) | 3 (3) |
| Possible, n (%) | 27 (4) | 19 (4) | 8 (8) |
| Hip BMD score < –2.5, n (%) | 37 (6) | 31 (6) | 6 (6) |
| 25-hydroxyvitamin D, mean (SD) | 25.7 (12.4) | 25.4 (12.8) | 27.1 (10.5) |
| FGF-23, median (25th, 75th percentile) | 55 (42, 73) | 60 (45, 77) | 42 (35, 49) |
| α-Klotho, median (25th, 75th percentile) | 589 (457, 756) | 544 (432, 714) | 756 (663, 873) |
| PTH, median (25th, 75th percentile) | 39 (29, 56) | 43 (32, 61) | 27 (21, 33) |
Abbreviations: BMD, bone mineral density; BMI, body mass index; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; FGF, fibroblast growth factor; Health ABC, Health Aging and Body Composition Study; PTH, parathyroid hormone; SBP, systolic blood pressure; UACR, urine albumin-to-creatinine ratio.
aLow bone turnover = low FGF-23, high α-Klotho, and low PTH.
We defined low bone turnover based on biomarkers in participants who simultaneously had FGF-23 below the median, soluble ɑ-Klotho above the median, and PTH below the median according to the bone biopsy cohort findings. We observed that 103 individuals (16%) met this classification. This proportion was relatively similar to the 15% meeting this definition in the Kentucky bone biopsy cohort. We next compared characteristics of those with biomarker defined low bone turnover vs all others in Health ABC. Individuals with biomarker-defined low bone turnover were more likely to smoke, had a lower BMI, were less likely to be hypertensive, and had a higher eGFR than the remainder of CKD participants in Health ABC (Table 2), but BMD was similar.
Bone mineral density and incident fractures stratified by biomarker-defined bone turnover status
Among the 641 Health ABC participants, there were 134 incident fractures during follow-up. Among the 103 individuals with biomarker-defined low bone turnover, there were 23 incident fractures (22%); and among the remaining 538 individuals with non–low bone turnover, there were 111 incident fractures (21%). Lower total hip and spine BMD both were associated with risk of incident fracture (HR = 2.55 [1.94-3.34] and 1.59 [1.25-2.02], respectively) during follow-up (Table 3). However, these associations differed among individuals characterized as having biomarker-defined low bone turnover (Pinteractions = .082 and 0.081 for hip and spine BMD, respectively). In stratified analyses, for each SD lower hip BMD, the risk of incident fractures was nearly 8-fold higher in participants with biomarker-defined low bone turnover (HR = 8.10, 95% CI, 3.40-19.30) compared to an approximately 2-fold higher risk in others (HR = 2.28, 95% CI, 1.69-3.08). Similarly, among those with biomarker-defined low bone turnover, each SD lower (SD = 0.16 g/cm2) spine BMD was associated with a 3-fold higher risk of fracture (HR = 3.53; 95% CI, 1.39-8.96) compared to a 1.45 (95% CI, 1.12-1.88) risk of fracture among all others. (Table 3) When the bone turnover biomarkers were evaluated individually, none of the interactions were statistically significant, yet the point estimate for fracture risk was consistently higher in the group contributing individuals to the low bone turnover subset (31). Similarly, when we defined low turnover using only FGF-23 and α- Klotho, results were qualitatively similar to the main analyses, but P values for interaction were higher.
Table 3.
Association between bone mineral densitya and incident fractures stratified by low bone turnover vs non–low bone turnover in 641 Health ABC participants with chronic kidney disease
| Incident fractures per SD lower hip BMD | |||||
|---|---|---|---|---|---|
| No. events/No. at risk | Unadjusted | Age-, sex-, race-, and site-adjusted | Fully adjustedb | P interaction | |
| HR (95% CI) | HR (95% CI) | HR (95% CI) | |||
| Low turnover | 23/103 | 3.26 (2.08-5.11) | 3.97 (2.26-6.97) | 8.10 (3.40-19.30) | .082 |
| Non–low turnover | 111/538 | 1.79 (1.44-2.23) | 1.93 (1.49-2.49) | 2.28 (1.69-3.08) | |
| Incident fractures per SD lower spine BMD | |||||
| No. events/No. at risk | Unadjusted | Age-, sex-, race-, and site-adjusted | Fully adjusted b | P interaction | |
| HR (95% CI) | HR (95% CI) | HR (95% CI) | |||
| Low turnover | 23/103 | 2.90 (1.48-5.71) | 2.90 (1.41-5.94) | 3.53 (1.39-8.96) | .081 |
| Non–low turnover | 111/538 | 1.49 (1.21-1.83) | 1.39 (1.10-1.76) | 1.45 (1.12-1.88) | |
Abbreviations: BMD, bone mineral density; BMI, body mass index; eGFR, estimated glomerular filtration rate; Health ABC, Health Aging and Body Composition Study; HR, hazard ratio; HTN, hypertension; SBP, systolic blood pressure; UACR, urine albumin-to-creatinine ratio.
aPer SD lower BMD, which was 0.16 g/cm2 in the hip and 0.16 g/cm2 in the spine.
bFully adjusted for age, sex, race, site, BMI, SBP, HTN medications, eGFR, UACR, vitD25(OH)2, smoking status, alcohol use, diabetes, physical activity, and diuretics.
Participants with low bone turnover had equivalent fracture risk to individuals without, based on the biomarkers alone. Specifically, comparing those with low bone turnover to those with non–low bone turnover based on biomarkers, there was a similar risk of incident fracture (HR = 0.96; 95% CI, 0.60-1.53 adjusted for hip BMD; and HR = 1.00; 95% CI, 0.63-1.60, adjusted for spine BMD) (Table 4).
Table 4.
Association between low/non–low bone turnover and incident fractures in 641 Health ABC participants with chronic kidney disease
| Incident fracture, adjusted for hip BMD | Incident fracture, adjusted for spine BMD | |
|---|---|---|
| HR (95% CI) | HR (95% CI) | |
| Non–low bone turnover | 1.00 (Reference) | 1.00 (Reference) |
| Low bone turnover | ||
| Unadjusted | 0.96 (0.61-1.50) | 0.97 (0.62-1.53) |
| Age-, sex-, race-, and site-adjusted | 0.89 (0.54-1.40) | 0.96 (0.61-1.50) |
| Fully adjusted | 0.96 (0.60-1.53) | 1.00 (0.63-1.60) |
Fully adjusted for age, sex, race, site, BMI, SBP, HTN medications, eGFR, UACR, vitD25(OH)2, smoking status, alcohol use, diabetes, physical activity, and diuretics.
Abbreviations: BMD, bone mineral density; BMI, body mass index; eGFR, estimated glomerular filtration rate; Health ABC, Health Aging and Body Composition Study; HR, hazard ratio; HTN, hypertension; SBP, systolic blood pressure; UACR, urine albumin-to-creatinine ratio.
Discussion
Low BMD is associated with a higher risk of fracture in CKD patients, but does not provide information on underlying bone pathology critical for appropriate treatment decisions (6). We sought to determine whether a biomarker panel could provide information on bone turnover in CKD, which might be combined with BMD data in CKD patients. We found that a biomarker panel of lower FGF-23, higher α-Klotho, and lower PTH had excellent specificity for low bone turnover in CKD patients based on anterior iliac crest bone biopsy and histomorphometry. When applied to a separate cohort of older community-dwelling individuals with CKD, this panel of biomarkers identified a subset of individuals among whom lower BMD was much more strongly associated with subsequent fracture risk.
The panel of 3 biomarkers was highly specific for low bone turnover. Although we know of no other study that has evaluated these 3 markers concurrently, our findings regarding differential risk of fracture among individuals likely to have low bone turnover are supported by prior studies. For example, in a study among hemodialysis patients from Japan, Iimori and colleagues reported that the association of BMD with fracture risk was stronger in dialysis patients with low PTH concentrations (32). Similarly, Coco and Rush reported that dialysis patients in the United States with lower serum PTH concentrations were more likely to sustain hip fractures than those with higher PTH levels (10). Although limited to end-stage renal disease patients, and PTH levels per se, these findings suggest that biomarkers identify individuals with low bone turnover who may be at particularly high risk for future fractures, findings that are again supported by our data using a biomarker panel with very high specificity for low bone turnover.
We know of no randomized trials comparing different pharmacological treatments for prevention of fracture risk using biopsy data to guide therapy. Nonetheless, current understanding of bone pathology would suggest benefits of using different therapies based on bone turnover status in CKD patients. For example, bisphosphonates inhibit osteoclasts, which can diminish bone turnover (33). Receptor activator of nuclear factor κβ ligand (RANKL) inhibitors have similar effects on bone turnover. These drugs may not be the best option in individuals known to have low bone turnover before treatment; such individuals might benefit from therapies that would stimulate rather than diminish bone turnover. Examples of potential treatments in such individuals might include discontinuation of use of calcium and vitamin D to allow endogenous PTH to rise, and potentially the use of PTH analogues to stimulate higher bone turnover. On the other hand, individuals with high bone turnover may benefit from bisphosphonates or other antiresorptive medications. Because these decisions are typically determined based on bone biopsy and histomorphometry, an approach of defining therapy based on biopsy results would be available only at select centers. Our findings are a first step only, and need confirmation, but suggest that biomarker panels may ultimately have clinical utility to identify low bone turnover in CKD patients without a bone biopsy. Future studies are needed to refine biomarker panels, and ultimately to test whether using such panels can guide therapies that could translate into lower fracture rates for CKD patients.
A key strength of this study is its evaluation of a population referred for bone biopsy, and application of the results to a community-dwelling population with CKD but without bone biopsies, all within one study. We recognize that these populations are inherently different because individuals with CKD referred for bone biopsy represent a select cohort because of indication, and differ from community-dwelling elders with CKD. However, historically, bone biopsy studies uniformly report relationships of biomarkers or other parameters in small, single-center samples, making comparisons across populations difficult. Thus, a novel and important contribution of our study is expanding the findings to a community-dwelling population and demonstrating the importance of the biomarker measurements for predicting fracture risk in the wider community-dwelling CKD population, even though the findings were driven from a single-center cohort whose members underwent bone biopsy. In this way, we believe this manuscript meaningfully moves forward the field of bone and mineral research with improved external generalizability.
Other strengths of this manuscript include the prospective nature of the Health ABC cohort wherein kidney function, BMD, fractures, and other risk factors are all characterized. Limitations of our study include a lack of availability of other potential biomarkers of bone turnover, such as DKK1, BSAP, TRAP5b, and sclerostin in Health ABC, and limited power to determine whether relationships of interest differ by sex. We also recognize that spine BMD was not as predictive of fracture. This may be due to contributions of surrounding calcification of the joint and overlying aortic calcification, which is common in CKD patients.
In conclusion, in community-dwelling individuals with CKD, those with low BMD who concurrently had a biomarker panel specific for low bone turnover had a higher fracture risk than individuals with low BMD alone. These findings suggest that individuals with low bone turnover comprise a subset of CKD patients who can be identified without a bone biopsy and may ultimately benefit from different therapeutic strategies. Future studies should determine whether serum biomarker panels may allow determination of low bone turnover with high precision. If our findings are confirmed, future clinical trials could determine whether choosing bone therapy based on biomarker panels marking turnover might lead to lower fracture risk in CKD patients.
Acknowledgments
The authors would like to thank the patients, staff, and investigators at the University of Kentucky and in the Health, Aging, and Body Composition Study for their invaluable contributions.
Financial Support: This research was supported by the National Heart Lung and Blood Institute (Grants R03 R03 HL146875 and K01 HL122394); the National Institute of Diabetes and Digestive and Kidney Diseases (Grants K24DK110427 and R01 DK080770); the National Institute on Aging (NIA) (Contracts N01-AG-6-2101, N01-AG-6-2103, and N01-AG-6-2106) and NIA Grants R01 AG065876, R01 AG028050 and R01 AG027002); the National Institute of Nursing Research (Grant R01-NR012459); and in part by the Intramural Research Program of the National Institutes of Health, NIA, and the Kentucky Nephrology Research Trust.
Glossary
Abbreviations
- CKD
chronic kidney disease
- CV
coefficient of variation
- BMD
bone mineral density
- BMI
body mass index
- BSAP
bone-specific alkaline phosphatase
- DKK1
dickkopf-related protein 1
- eGFR
estimated glomerular filtration rate
- Health ABC
Health Aging and Body Composition Study
- HR
hazard ratio
- ROC
receiver operating characteristic
- TRAP-5b
tartrate-resistant acid phosphatase 5b
Additional Information
Disclosure Summary: The authors have nothing to disclose.
Data Availability: The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.
References
- 1. National Chronic Kidney Disease Fact Sheet. Centers for Disease Control and Prevention. 2017. https://www.cdc.gov/diabetes/pubs/pdf/kidney_factsheet.pdf. Accessed November 9, 2018. [Google Scholar]
- 2. Kazama JJ. Chronic kidney disease and fragility fracture. Clin Exp Nephrol. 2017;21(Suppl 1):46-52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Kim SM, Long J, Montez-Rath M, Leonard M, Chertow GM. Hip fracture in patients with non-dialysis-requiring chronic kidney disease. J Bone Miner Res. 2016;31(10):1803-1809. [DOI] [PubMed] [Google Scholar]
- 4. Jørgensen HS, Winther S, Bøttcher M, et al. . Bone turnover markers are associated with bone density, but not with fracture in end stage kidney disease: a cross-sectional study. BMC Nephrol. 2017;18(1):284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Malluche HH, Mawad HW, Monier-Faugere MC. Renal osteodystrophy in the first decade of the new millennium: analysis of 630 bone biopsies in black and white patients. J Bone Miner Res. 2011;26(6):1368-1376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Yenchek RH, Ix JH, Shlipak MG, et al. ; Health, Aging, and Body Composition Study Bone mineral density and fracture risk in older individuals with CKD. Clin J Am Soc Nephrol. 2012;7(7):1130-1136. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Malluche HH, Porter DS, Pienkowski D. Evaluating bone quality in patients with chronic kidney disease. Nat Rev Nephrol. 2013;9(11):671-680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis. 2002;39(2 Suppl 1):S1-266. [PubMed] [Google Scholar]
- 9. Atsumi K, Kushida K, Yamazaki K, Shimizu S, Ohmura A, Inoue T. Risk factors for vertebral fractures in renal osteodystrophy. Am J Kidney Dis. 1999;33(2):287-293. [DOI] [PubMed] [Google Scholar]
- 10. Coco M, Rush H. Increased incidence of hip fractures in dialysis patients with low serum parathyroid hormone. Am J Kidney Dis. 2000;36(6):1115-1121. [DOI] [PubMed] [Google Scholar]
- 11. Sista SK, Arum SM. Management of adynamic bone disease in chronic kidney disease: a brief review. J Clin Transl Endocrinol. 2016;5:32-35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Malluche HH, Porter DS, Monier-Faugere MC, Mawad H, Pienkowski D. Differences in bone quality in low- and high-turnover renal osteodystrophy. J Am Soc Nephrol. 2012;23(3):525-532. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Moe S, Drüeke T, Cunningham J, et al. ; Kidney Disease: Improving Global Outcomes (KDIGO) Definition, evaluation, and classification of renal osteodystrophy: a position statement from Kidney Disease: Improving Global Outcomes (KDIGO). Kidney Int. 2006;69(11):1945-1953. [DOI] [PubMed] [Google Scholar]
- 14. Malluche HH, Monier-Faugere MC. Renal osteodystrophy: what’s in a name? Presentation of a clinically useful new model to interpret bone histologic findings. Clin Nephrol. 2006;65(4):235-242. [DOI] [PubMed] [Google Scholar]
- 15. Malluche HH, Faugere MC.. Atlas of Mineralized Bone Histology. San Diego, CA: Academic Press; 1986. [Google Scholar]
- 16. Goldner J. A modification of the Masson trichrome technique for routine laboratory purposes. Am J Pathol. 1938;14(2): 237-243. [PMC free article] [PubMed] [Google Scholar]
- 17. Lillie PD, Fullmer HM. Histopathologic Technique and Practical Histochemistry. 4th ed. New York: McGraw-Hill; 1976:534. [Google Scholar]
- 18. Denton J, Freemont AJ, Ball J. Detection and distribution of aluminium in bone. J Clin Pathol. 1984;37(2):136-142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Malluche HH, Sherman D, Meyer W, Massry SG. A new semiautomatic method for quantitative static and dynamic bone histology. Calcif Tissue Int. 1982;34(5):439-448. [DOI] [PubMed] [Google Scholar]
- 20. Manaka RC, Malluche HH. A program package for quantitative analysis of histologic structure and remodeling dynamics of bone. Comput Programs Biomed. 1981;13(3-4):191-201. [DOI] [PubMed] [Google Scholar]
- 21. Dempster DW, Compston JE, Drezner MK, et al. . Standardized nomenclature, symbols, and units for bone histomorphometry: a 2012 update of the report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res. 2013;28:2-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Parfitt AM, Drezner MK, Glorieux FH, et al. . Bone histomorphometry: standardization of nomenclature, symbols, and units. Report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res. 1987;2:595-610. [DOI] [PubMed] [Google Scholar]
- 23. Lima F, El-Husseini A, Monier-Faugere MC, et al. . FGF-23 serum levels and bone histomorphometric results in adult patients with chronic kidney disease on dialysis. Clin Nephrol. 2014;82:287-295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Kritchevsky SB, Tooze JA, Neiberg RH, et al. ; Health ABC Study 25-Hydroxyvitamin D, parathyroid hormone, and mortality in black and white older adults: the Health ABC Study. J Clin Endocrinol Metab. 2012;97(11):4156-4165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Inker LA, Schmid CH, Tighiouart H, et al. ; CKD-EPI Investigators Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med. 2012;367(1):20-29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Yamazaki Y, Imura A, Urakawa I, et al. . Establishment of sandwich ELISA for soluble alpha-Klotho measurement: age-dependent change of soluble alpha-Klotho levels in healthy subjects. Biochem Biophys Res Commun. 2010;398(3):513-518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Chalhoub D, Marques E, Meirelles O, et al. ; Health ABC Study Association of serum Klotho with loss of bone mineral density and fracture risk in older adults. J Am Geriatr Soc. 2016;64(12):e304-e308. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Larsen K. Creatinine assay by a reaction-kinetic principle. Clin Chim Acta. 1972;41:209-217. [DOI] [PubMed] [Google Scholar]
- 29. Goldenstein L, Driver TH, Fried LF, et al. ; Health ABC Study Investigators Serum bicarbonate concentrations and kidney disease progression in community-living elders: the Health, Aging, and Body Composition (Health ABC) Study. Am J Kidney Dis. 2014;64:542-549. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Habibzadeh F, Habibzadeh P, Yadollahie M. On determining the most appropriate test cut-off value: the case of tests with continuous results. Biochem Med (Zagreb). 2016;26(3):297-307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Hughes-Austin JM, Katz R, Semba RD, et al. . Biomarkers of bone turnover identify subsets of chronic kidney disease patients at high risk for fracture. Figshare Repositories. Deposited 3 March 2020. 10.35092/yhjc.11921697.v1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Iimori S, Mori Y, Akita W, et al. . Diagnostic usefulness of bone mineral density and biochemical markers of bone turnover in predicting fracture in CKD stage 5D patients—a single-center cohort study. Nephrol Dial Transplant. 2012;27:345-351. [DOI] [PubMed] [Google Scholar]
- 33. Amerling R, Harbord NB, Pullman J, Feinfeld DA. Bisphosphonate use in chronic kidney disease: association with adynamic bone disease in a bone histology series. Blood Purif. 2010;29:293-299. [DOI] [PubMed] [Google Scholar]