Comprehensive Associations between Acidosis and the Skeleton in Patients with Kidney Disease

Abstract

Significance Statement

Renal osteodystrophy (ROD) contributes substantially to morbidity in CKD, including increased fracture risk. Metabolic acidosis (MA) contributes to the development of ROD, but an up-to-date skeletal phenotype in CKD-associated acidosis has not been described. We comprehensively studied associations between acidosis and bone in patients with CKD using advanced methods to image the skeleton and analyze bone-tissue, along with biochemical testing. Cross-sectionally, acidosis was associated with higher markers of bone remodeling and female-specific impairments in cortical and trabecular bone quality. Prospectively, acidosis was associated with cortical expansion and trabecular microarchitectural deterioration. At the bone-tissue level, acidosis was associated with deficits in bone mineral content. Future work investigating acidosis correction on bone quality is warranted.

Background:

Renal osteodystrophy is a state of impaired bone quality and strength. Metabolic acidosis (MA) is associated with alterations in bone quality including remodeling, microarchitecture, and mineralization. No studies in patients with CKD have provided a comprehensive multimodal skeletal phenotype of MA. We aim to describe the structure and makeup of bone in patients with MA in the setting of CKD using biochemistry, noninvasive imaging, and histomorphometry.

Methods:

The retrospective cross-sectional analyses included 180 patients with CKD. MA was defined as bicarbonate ≤22 mEq/L. We evaluated circulating bone turnover markers and skeletal imaging with dual energy x-ray absorptiometry and high-resolution peripheral computed tomography. A subset of 54 participants had follow-up. We assessed associations between baseline and change in bicarbonate with change in bone outcomes. Histomorphometry, microCT, and quantitative backscatter electron microscopy assessed bone biopsy outcomes in 22 participants.

Results:

The mean age was 68±10 years, 54% of participants were male, and 55% were White. At baseline, acidotic subjects had higher markers of bone turnover, lower areal bone mineral density at the radius by dual energy x-ray absorptiometry, and lower cortical and trabecular volumetric bone mineral density and impaired trabecular microarchitecture. Over time, acidosis was associated with opposing cortical and trabecular effects: cortical expansion but trabecular deterioration. Bone-tissue analyses showed reduced tissue mineral density with increased heterogeneity of calcium distribution in acidotic participants.

Conclusions:

MA is associated with multiple impairments in bone quality. Future work should examine whether correction of acidosis improves bone quality and strength in patients with CKD.

Introduction

CKD is a state of poor skeletal health. CKD mineral bone disease encompasses a wide spectrum of derangements, including metabolic (e.g., hyperphosphatemia and hypocalcemia) as well as ROD. ROD, a disorder of bone turnover and mineralization, is present in most patients as kidney function declines.^1,2 Osteoporosis, defined as a T-score ≤ −2.5 by dual energy x-ray absorptiometry (DXA), is two-fold more common in patients with than without CKD.³ These impairments in bone quality and strength are associated with a 2-fold to 14-fold greater fracture risk in CKD and a parallel increase in risk with worsening CKD severity.^4–8 After fracturing, patients with CKD have a doubling of mortality risk.^8,9 Currently, there are no fracture prevention therapies developed explicitly for patients with CKD. Fracture prevention therapies used in the general population may have both skeletal and kidney toxic effects in patients with CKD,^10–14 which limits their acceptability for use in this high-risk population. Therefore, therapies are needed that safely mitigate the underlying pathogenesis of CKD-associated bone disease to lower fracture risk and improve patient-level outcomes.

MA is an important complication of CKD that is associated with altered bone quality and strength. Bone is the primary buffer for acid.^15–17 In animal models, incubation of mouse calvariae at reduced pH suggested uncoupling of bone turnover, favoring osteoclast-mediated bone resorption.¹⁸ In vivo models of acidosis in rats demonstrated suppression of circulating markers of formation, increases in circulating markers of resorption, and an association with decreased cortical mass, cortical thinning, decreased trabecular bone mass, and abnormal trabecular microarchitecture (i.e., decreased number and increased spacing).¹⁹ Exposure of mice to 14 days of diet-induced acidosis resulted in decreased femur ultimate strength, a surrogate marker of fracture risk.²⁰ Using the National Health and Examination Survey, Chen et al. reported that in the general US population, low serum bicarbonate levels were associated with lower areal bone mineral density (BMD) by DXA.²¹ In humans with CKD, there have been few bone biopsy studies, and none in the nonaluminum phosphate binder era, to assess bone-tissue–level effects of MA. In the presence of aluminum-based phosphate binders, MA was associated with osteomalacia²² and correction of acidosis in patients on hemodialysis correlated with improved bone turnover in patients with both low and high turnover disease.²³ However, notwithstanding the putative benefit of alkali therapy on bone density and quality, a recent clinical trial in patients with CKD stages 3 and 4 failed to demonstrate an effect of 24 months of sodium bicarbonate administration on femoral neck areal BMD.²⁴ This is in stark contrast to the beneficial effects of alkali supplementation on bone density and quality in kidney healthy older adults without low serum bicarbonate.^25,26 It is not clear whether these discrepant findings are due to differences in the skeletal effects of acidosis in patients without exposure to aluminum, whether metabolic complications of CKD attenuate those of alkali on the skeleton, or whether the DXA end point selected by Melamed et al.²⁴ was not correct. To determine relationships between acidosis and bone quality in the nonaluminum phosphate binder era and to evaluate potential bone imaging targets for use in future alkali trials, we conducted a comprehensive assessment of the skeletal phenotype in patients with CKD, with and without acidosis.

Methods

In a cross-sectional cohort of patients with CKD with a subset who underwent longitudinal follow-up and bone biopsy, we quantified relationships between bicarbonate and bone density and quality. To assess areal BMD, we used DXA at the spine, hip, and forearm. To assess bone quality, we used several state-of-the-art methods including (1) high-resolution peripheral computed tomography (HR-pQCT) at the radius and tibia to measure cortical and trabecular geometry, volumetric BMD, and microarchitecture and (2) bone biopsy with quantitative histomorphometry to measure dynamics, microCT (resolution approximately 5 µm) to measure trabecular microarchitecture and tissue mineral density, and quantitative back-scatter electron microscopy (qBEI) to measure calcium content and distribution.²⁷

Study Sample

One-hundred eighty patients with CKD stages 2–5D were enrolled into a cross-sectional study with longitudinal follow-up of relationships between kidney function and bone quality and strength. From the parent cohort, 54 subjects agreed to participate in the longitudinal substudy and 22 subjects agreed to participate in the bone biopsy substudy.

All participants were recruited from the general medicine and nephrology clinics of Columbia University Irving Medical Center (CUIMC) between August 2006 and September 2010. All patients referred to the nephrology clinics and meeting study inclusion criteria were eligible. The CUIMC nephrology clinics serve as a referral center for patients with CKD from the northern Manhattan, Bronx, Queens, southern New York State and Connecticut, and northern and central New Jersey areas. We did not exclude patients on the basis of CKD etiologies, which were broadly represented and included diabetes, hypertension, glomerular diseases, and tubulointerstitial etiologies, including polycystic kidney disease, lithium toxicity, nephrolithiasis, urinary tract abnormalities, and renal artery stenosis. eGFR was determined by the race-free Chronic Kidney Disease Epidemiology Collaboration estimating equation.^28,29 Patients on hemodialysis had to have been on hemodialysis for at least 6 months. Patients with a history of malignancy, bilateral lower extremity amputations, residing in a nursing home, requiring a wheelchair, and those taking antiresorptives, anabolics, gonadal steroids, aromatase inhibitors, and anticonvulsants that induce hepatic cytochrome P450 enzymes were excluded. Therefore, the cohort included in this investigation represents the wide spectrum of ambulatory patients with CKD and is generalizable to those patients typically treated in general nephrology and dialysis clinics. Demographic variables collected included age, self-identified sex, self-identified race, and self-identified Hispanic ethnicity. Race and ethnicity data were collected because these variables have previously been associated with differences in bone and mineral metabolism.³⁰

Ethics

This study was conducted in keeping with the Declaration of Helsinki and the Declaration of Istanbul. The CUIMC Institutional Review Board approved this study, and all subjects provided written informed consent under Institutional Review Board approval AAAB9295.

Laboratory Measurements

Routine laboratory parameters were measured by Quest diagnostics. Serum creatinine was determined by the Jaffe reaction, and serum calcium, phosphorus, and bicarbonate were measured by spectrophotometry. Calciotropic hormones and bone turnover markers (BTMs) were measured at CUIMC in a specialized research laboratory. Intact parathyroid hormone, serum total 25-hydroxyvitmain D, bone-specific alkaline phosphatase, N-Mid osteocalcin, procollagen of type 1 N-terminal propeptide, tartrate-resistant acid phosphatase 5b (TRAP5b), and C-terminal telopeptides of type I collagen were measured by using Roche Elecsys 2010 analyzer (Roche Diagnostics, Indianapolis, IN). Details of the reference ranges and precision of bone biomarkers are listed in Supplemental Table 1.

Exposure

Patients were classified as acidotic if they had serum bicarbonate ≤22 mEq/dl at baseline, in keeping with current literature.^21,31,32

Measurement of Areal BMD by DXA

Areal BMD by DXA was measured at the total lumbar spine (L1–L4), total hip, femoral neck, and nondominant 1/3 and ultradistal radius using a Hologic QDR 4500 densitometer (Hologic, Inc., Waltham, MA) in the array (fan beam) mode. In our laboratory, short-term, in vivo precision is 0.68% for the spine, 1.36% for the femoral neck, and 0.70% for the radius. Z-scores compared subjects with data from young-normal populations of the same race, age, and sex provided by the manufacturer (spine and forearm) and by the National Health and Nutrition Examination Survey III (total hip and femoral neck).

HR-pQCT Imaging of the Radius and Tibia

All subjects were scanned with HR-pQCT (XtremeCT resolution 82 µm³; SCANCO Medical, Brüttisellen, Switzerland) at the nondominant forearm and leg unless there was previous fracture or an arteriovenous fistula or graft at the desired site in which case the opposite limb was scanned. All scan acquisition was performed in our laboratory by a single dedicated research densitometrist according to the standard manufacturer's protocols described previously.^3,33 The arm or leg was positioned in the scanner, and a 9.02-mm region of interest was defined on a scout film by manual placement of a reference line at the endplate of the radius or tibia, with the first slice 9.5 and 22.5 mm proximal to the reference line at the radius and tibia, respectively. Attenuation data were converted to equivalent hydroxyapatite (HA) densities. A phantom was scanned daily for quality control. To analyze the same region in the longitudinal scans, the manufacturer's software was used to find the overlapping regions between the baseline and follow-up scans.³⁴ This is performed by matching the cross-sectional areas of the individual slices to find the common region between the two scans. A single technician performed all image analysis using the standard manufacturer's software (SCANCO Medical AG). From this standard analysis trabecular, BMD is defined as the average bone density within the trabecular volume of interest and the ratio of bone volume to total volume (BV/TV, %) is derived from trabecular density assuming that the density of fully mineralized bone is 1.2 g HA/cm³ (BV/TV=100 × Dtrab/1200 mg HA/cm³). Because direct measurements of trabecular microstructure are limited by the resolution of HR-pQCT, which approximates the width of individual trabeculae, trabecular structure is assessed using semiderived algorithms.^35,36 Trabecular number is defined as the inverse of the mean spacing of the midaxes. Trabecular separation is derived from BV/TV and trabecular number using formulas from traditional quantitative histomorphometry: trabecular thickness = (BV/TV)/trabecular number and trabecular separation = (1 − BV/TV)/trabecular number. The intraindividual distribution of separation (µm), a parameter that reflects the heterogeneity of the trabecular network, is also measured. In addition to the standard analysis, a validated autosegmentation method³⁷ was applied to segment the cortical and trabecular compartments to measure cortical porosity (%), direct cortical thickness (mm), and cortical BMD (mg HA/cm³).^38,39 Cortical porosity was calculated as the number of void voxels in the cortex using Image Processing Language (version 5.08b, SCANCO Medical AG). Cortical thickness was measured directly using a distance transform,⁴⁰ and cortical BMD was defined as the average mineral density in the autosegmentation cortical bone mask. In vivo precision of HR-pQCT measurements have been reported to be <1% for density measurements and <4.5% for morphologic measurements⁴¹ and in our laboratory, density measurements <1.06% and morphologic measurements <5.22% (unpublished data).

Transiliac Bone Biopsies

Biopsies were performed for research purposes, not as part of routine clinical care. Transiliac bone biopsy was performed with a 7.5-mm trephine. Tetracycline double labeling occurred according to our previously published work.⁴² Undecalcified bone fragments were subjected to histomorphometric analysis after standard processing. Histologic interpretation of bone biopsies was performed at Helen Hayes Hospital, West Haverstraw, NY. Bone histomorphometry was analyzed using a semiautomatic technique in Osteomeasure software (Osteometrics, Atlanta, GA, USA). Static histomorphometric parameters used for this analysis were trabecular bone volume (BV/TV, %), trabecular width (Tb.Wi, μm), number (Tb.N, /mm) and separation (Tb.Sp, μm), cortical width (Ct.Wi, μm), and porosity (Ct.Po, %). Dynamic parameters included bone formation rate/bone surface (μm³/μm² per day), mineralizing surface/bone surface (%), mineral apposition rate (μm/day), and mineralization lag time (days).

qBEI

For qBEI analysis, the entire sectioned surface of the residual sample block of the undecalcified, polymethylmethacrylate-embedded transiliac biopsy was imaged in the scanning electron microscope (DSM 962, Zeiss, Oberkochen, Germany, using a four-quadrant semiconductor backscatter electron detector). Generally, in calibrated qBEI images, the pixel gray levels reflect the local calcium concentrations in the bone material. Thus, the assessment of gray-level histograms (denoted Bone Mineralization Density Distributions, BMDD) provides information about the frequency distribution of these different calcium concentrations. The gray-level calibration procedure which is required for BMDD measurement was described together with other details of the method extensively in a previous work.⁴³ Reference materials of known atomic numbers (aluminium [Z=13] and carbon [Z=6]) were used for this purpose. The DSM was operated at following microscope settings: accelerating voltage 20 kV, probe current 110±0.4 pA, working distance 15 mm, and a scan speed of 100 seconds per frame. The entire available cancellous and cortical tissue areas of the biopsy samples were recorded in a series of images with each 50× nominal magnification (pixel resolution of 12.4 µm²) for the evaluation of cancellous and cortical BMDD separately (histogram bin width was 0.17 wt% Ca). Five parameters were obtained from the BMDD curves²⁷ which are explained in the legend of Figure 2.

Figure 2.

Box plots showing BMDD measurements in trabecular (A–E) and cortical (F–J) bone for acidotic and nonacidotic participants including (A, F) CaMean which is the weighted mean Ca-concentration of the bone area; (B, G) CaPeak which is the histogram peak position (indicating the most frequently measured calcium concentration); (C, H) CaWidth which is the full width at half maximum of the distribution (describing the variation in mineralization density); (D, I) CaLow which is the percentage of low mineralized bone (i.e., bone areas having a calcium content of <5th percentile of our reference BMDD); and (E, J) CaHigh which is the percentage of highly mineralized bone areas (i.e., bone areas having a calcium content of >95th percentile of our reference BMDD). The center line represents the median, with the box outlining the interquartile range and whiskers showing minimum and maximum. The overlying gray boxes show the trabecular reference ranges for an adult population without CKD published previously.⁵⁸ There are no reference data for cortical bone. BMDD, bone mineralization density distribution. * p < 0.05

MicroCT

The microtomographic imaging system (µCT 40, Scanco Medical AG, Brüttisellen, Switzerland) is equipped with a 5-µm focal spot x-ray tube as a source. A two-dimensional charge-coupled device, coupled to a thin scintillator as a detector, permits acquisition of 210 tomographic images in parallel. The long axis of the intact biopsy is oriented along the rotation axis of the scanner. The x-ray tube is operated at 50 kVp and 160 µA with an integration time set to 200 ms, and all projection frames are recorded six times and then averaged. Scans (high-resolution mode) are performed at an isotropic nominal resolution of 8 µm (nominal spatial resolution/voxel size rather than true spatial resolution). For each subject, the intact biopsy was scanned. Biopsies varied in length ranging from 7 up to 15 mm, which resulted in measurement times of between 7 and 15 hours. A cylindrical volume of interest was then placed in the digital image data to select the trabecular bone compartment. The mineralized tissue was segmented from soft tissue by a global thresholding procedure,⁴⁴ with a threshold value set to 34% of the maximum grayscale value. Morphometric indices were determined for the trabecular bone compartment using a direct 3D approach³⁶ and include bone volume density (BV/TV), trabecular thickness, separation, and number.

Statistical Methods

Continuous variables were summarized using medians and interquartile ranges or means and standard deviation as appropriate, whereas categorical variables were summarized as frequencies and percentages. Normality was assessed visually by histogram and formally tested by using Shapiro-Wilk test. For DXA BMD, Z-scores were used in all analyses to limit bias because of age, sex, and race. Between-group comparisons were performed by using t test without or with Satterthwaite correction for unequal variances, Wilcoxon rank-sum test for non-normally distributed continuous variables, and chi-squared test or Fisher exact test for categorical variables. Spearman correlations, adjusted for eGFR, were determined between biochemistries and DXA, HR-pQCT, histomorphometry, and tissue mineral density. For HR-pQCT, sex-specific Z-scores in CKD participants were generated on the basis of data from the healthy reference population. Bayes estimates of BTMs, histomorphometry dynamic and structural indices, and microCT differences between CKD participants with and without acidosis were calculated using Markov Chain Monte Carlo simulation with normal distribution priors for intercept and slope, inverse gamma distribution for variance, 10,000 iterations with a 2000 iteration burn-in: The posterior estimate of the difference and highest posterior density credible 95% interval are shown. The Bayes approach to estimating differences between CKD participants with and without acidosis was repeated for male and female patients for HR-pQCT radius and tibia as separate models. In cross-sectional analyses, sex-stratified generalized linear models were used to assess acidosis status (0/1) and imaging outcomes with and without adjustment for age, height, weight, the height-weight interaction, and kidney function. We assessed bone change rate differences between nonhemodialysis and hemodialysis participants as a function of their changing CO₂ concentrations using a general linear models test of the heterogeneity of slopes: i.e., the interaction of CO₂ concentration with the hemodialysis category with the hemodialysis class variable and baseline CO₂ level and CO₂ change as continuous variables. In prospective analyses, multiple regression models with semipartial and partial correlation were used to assess relationships between the absolute change in imaging Z-score with the change in bicarbonate, adjusting for the baseline bicarbonate and bone imaging Z-score with and without the baseline and change in kidney function. Analysis of the biopsy cohort repeated the analyses of the cross-sectional cohort and added analyses of histomorphometry and tissue mineral density indices. Finally, given the known associations between diabetes mellitus and bone health,⁴⁵ a sensitivity analysis excluding diabetic subjects was conducted, which did not materially alter the findings (data not shown). All statistical analyses were conducted using SAS STAT 13.2 (SAS Institute, Cary, NC).

Results

Cross-Sectional Cohort

Demographics and Biochemistries

In a cross-sectional analysis, we included all 180 individuals with CKD to measure associations of acidosis on the skeleton. Baseline demographics; medical history; skeletal imaging measures by DXA at the spine, hip, and forearm; and biochemical data are presented in Table 1. The mean age was 68±10 years, 54% were male, and 55% were White. At baseline, one patient was using topical glucocorticoids and another had used oral prednisone more than 1 year before the baseline visit. At follow-up, one patient was using topical glucocorticoids, one patient was taking prednisone 5 mg daily for treatment of microscopic polyangiitis, and one patient was using an inhaled glucocorticoid. Of those subjects on dialysis, the mean dialysis vintage was 2.75±2.26 years. There were differences in etiology of CKD and in use of medications used to manage CKD mineral bone disease. The use of sevelamer did not differ between participants with and without acidosis. Patients with acidosis compared with without acidosis had lower levels of mean eGFR, 25-hydroxy vitamin D, the Wnt signaling antagonist sclerostin, and FGF-23 and higher levels of phosphorus, intact parathyroid hormone, markers of bone formation (procollagen of type 1 N-terminal propeptide, osteocalcin), and resorption (C-terminal telopeptides of type I collagen, TRAP5b) (Table 1).

Table 1.

Baseline demographic data on 180 participants by acidosis status

Variables	Nonacidotic (HCO3>22 mEq/dl) (n=108)	Acidotic (HCO3≤22 mEq/dl) (n=72)	P Value
Demographics
Age (mean±SD, y)	68.9±9.5	66.9±10.4	0.2
Male sex (%)	52 (48%)	30 (42%)	0.4
Race (%)			0.1
American Indian or Alaska Native	4 (4%)	0 (0%)
Asian or Pacific Islander	1 (1%)	3 (4%)
Black	14 (13%)	16 (22%)
White	60 (56%)	39 (54%)
Other	29 (27%)	1 (19%)
Hispanic ethnicity	60 (56%)	29 (40%)	0.05
Body mass index (kg/m2)	29.0±5.4	28.4±5.4	0.4
Medical history (%)
CKD etiology			<0.001
Hypertension	47 (44%)	32 (44%)	0.9
Diabetes mellitus	32 (30%)	13 (18%)	0.08
Nephritis/Nephrosis	4 (4%)	10 (14%)	0.01
Post-transplant	0 (0%)	2 (3%)	0.08
Tubulointerstitiala	25 (23%)	15 (21%)	0.7
Hemodialysis	28 (26%)	19 (26%)	0.9
Fracture	31 (29%)	24 (33%)	0.5
Medication use (%)
Bicarbonate supplement	15 (14%)	12 (17%)	0.6
Calcium supplement	32 (30%)	17 (24%)	0.4
Phosphate binder	34 (19%)	26 (14%)	0.5
Calcium acetate	6 (3%)	6 (8%)	0.3
Calcium carbonate	23 (13%)	7 (9%)	0.5
Sevelamer (carbonate and hydrochloride)
Sevelamer carbonate	2 (2.5%)	4 (7.8%)	0.2
Sevelamer hydrochloride	7 (8.8%)	5 (9.4%)	>0.99
Lanthanum carbonate	5 (3%)	3 (4%)	0.9
Aluminum hydroxide	0 (0%)	0 (0%)	n/a
Nutritional vitamin D	39 (36%)	35 (49%)
Vitamin D receptor activators	36 (33%)	31 (43%)	0.2
Cinacalcet (n=254)	8 (7%)	7 (8%)	0.6
Multivitamin	66 (37%)	36 (33%)	0.6
Serum biochemistry (median [Interquartile range])
eGFR (ml/min per 1.73 m2)b	39.8 (23.8–54.1)	23.1 (15.5–32.5)	<0.001
Bicarbonate (mEq/L)	25 (24–27)	21 (19–22)	<0.001
Calcium (mg/dl)c	9.4 (9.0–9.7)	9.3 (8.8–9.7)	0.35
Phosphorus (mg/dl)d	3.6 (3.2–4.3)	4.1 (3.4–4.7)	0.02
25-OH vitamin D (ng/ml)e (n=120)	25.3 (19.2–39.5)	19.9 (13.0–36.4)	0.09
1,25-di-OH vitamin D (pg/ml)f (n=119)	31.7 (19.6–42.9)	26.1 (16.6–39.6)	0.19
Intact PTH (pg/ml) (n=120)	60.2 (34.2–118.3)	89.0 (55.4–183.8)	0.006
Bone-specific alkaline phosphatase (unit/L) (n=128)	31.0 (25.2–43.7)	35.0 (27.4–52.3)	0.26
Procollagen 1 Intact N-Terminal propeptide (mcg/L) (n=125)	75.2 (47.3–187.6)	108.3 (64.6–242.1)	0.09
Osteocalcin (ng/ml) (n=129)	39.8 (25.3–99.9)	75.6 (32.8–175.0)	0.003
Tartrate-resistant acid phosphatase 5b (unit/L) (n=125)	3.49 (2.81–4.49)	3.87 (2.96–5.76)	0.19
C-telopeptide (pg/ml) (n=129)	0.673 (0.376–1.02)	0.945 (0.571–1.565)	0.003
Fibroblast growth factor 23 (pg/ml) (n=77)	219.2 (129.7–415.3)	181.2 (135.6–389.1)	0.88
Sclerostin (ng/ml) (n=124)	1.70 (1.11–2.27)	1.62 (1.09–2.16)	0.82
DXA Z-scores (mean±SD)
Lumbar spine	0.83±1.78	0.80±1.89	0.7
Femoral neck	−0.10±1.02	−0.25±1.02	0.2
Total hip	0.11±1.20	−0.06±1.15	0.3
1/3 radius	0.52±1.66	0.10±1.75	0.05
Ultradistal radius	0.04±1.31	−0.28±1.63	0.04

PTH, parathyroid hormone; DXA, dual-energy x-ray absorptiometry.

^aIncluding polycystic kidney disease, lithium toxicity, nephrolithiasis, urinary tract abnormalities, and renal artery stenosis.

^bSubjects not on dialysis.

^cSI conversion: to convert to mmol/L, multiply by 0.25.

^dSI conversion: to convert to mmol/L, multiply by 0.323.

^eSI conversion: to convert to nmol/L, multiply by 2.496.

^fSI conversion: to convert to pmol/L, multiply by 2.4.

Imaging

In the 180 participants, we evaluated differences in age-adjusted, sex-adjusted, and race-adjusted Z-scores for areal BMD by DXA and cortical and trabecular volumetric BMD, geometry, and microarchitecture by HR-pQCT between patients with and without acidosis. For DXA, patients with acidosis had Z-scores that were approximately 0.5 standard deviations lower at the one third (approximately 85% cortical bone) and ultradistal (approximately 80% trabecular bone) radius (Table 1). Analyses of the peripheral skeleton by HR-pQCT demonstrated that acidosis was associated with abnormalities in bone geometry and density in women but not in men (Table 2). At the radius, women with acidosis had lower trabecular density and number and greater trabecular spacing and heterogeneity. At the tibia, women with acidosis compared with women without acidosis had lower total density, thinner cortices, lower trabecular number, and greater trabecular spacing and heterogeneity.

Table 2.

Mean±SD sex-specific Z-scores for high-resolution peripheral computed tomography parameters in 180 participants with CKD by acidosis status

Bone Parameters	Nonacidotic (HCO3>22 mEq/dl)	Acidotic (HCO3≤ 22mEq/dl)	Bayes Posterior Difference (95% HPD Interval)
Male radius	n=55	n=42
Total area, mm2	−0.061±0.962	0.056±0.934	0.119 (−0.312 to 0.515)
Total density, mg HA/cm3	−0.423±1.075	−0.350±1.447	0.081 (−0.467 to 0.607)
Cortical area, mm2	−0.445±1.303	−0.432±1.534	0.027 (−0.534 to 0.613)
Cortical density, mg HA/cm3	−0.547±1.313	−0.516±1.482	0.041 (−0.531 to 0.594)
Cortical thickness, mm	−0.398±1.128	−0.414±1.319	−0.015 (−0.541 to 0.478)
Cortical perimeter, mm	0.016±0.946	0.100±0.850	0.096 (−0.259 to 0.477)
Cortical porosity	0.599±1.401	0.181±1.180	−0.399 (−0.930 to 0.158)
Trabecular area, mm2	0.044±0.945	0.156±0.977	0.126 (−0.265 to 0.545)
Trabecular density, mg HA/cm3	−0.471±1.511	−0.180±2.097	0.296 (−0.402 to 1.016)
Trabecular number, 1/mm	−0.478±1.846	−0.392±1.411	0.098 (−0.587 to 0.759)
Trabecular thickness, mm	−0.289±1.163	−0.085±1.950	0.205 (−0.412 to 0.886)
Trabecular spacing, mm	1.622±5.517	0.640±2.396	−0.965 (−3.333 to 1.091)
Trabecular heterogeneity, mm	1.461±6.544	0.376±1.579	−1.046 (−3.137 to 0.944)
Male tibia
Total area, mm2	−0.373±1.075	−0.154±1.122	0.234 (−0.218 to 0.689)
Total density, mg HA/cm3	0.378±1.549	−0.108±1.902	0.296 (−0.405 to 0.952)
Cortical area, mm2	0.400±1.263	−0.230±1.674	0.182 (−0.445 to 0.769)
Cortical density, mg HA/cm3	−0.472±1.230	−0.574±1.627	−0.098 (−0.650 to 0.508)
Cortical thickness, mm	−0.272±1.201	−0.204±1.442	0.080 (−0.442 to 0.653)
Cortical perimeter, mm	−0.229±1.072	0.028±1.074	0.265 (−0.187 to 0.747)
Cortical porosity	0.632±1.257	0.680±1.502	0.058 (−0.517 to 0.596)
Trabecular area, mm2	−0.270±1.032	−0.099±1.103	0.184 (−0.271 to 0.589)
Trabecular density, mg HA/cm3	−0.445±1.474	0.021±2.023	0.474 (−0.196 to 1.228)
Trabecular number, 1/mm	−0.258±1.250	0.091±1.372	0.360 (−0.173 to 0.893)
Trabecular thickness, mm	−0.360±1.094	−0.222±1.414	0.130 (−0.380 to 0.672)
Trabecular spacing, mm	0.557±2.546	0.143±1.799	−0.414 (−0.134 to 0.477)
Trabecular heterogeneity, mm	0.836±4.413	0.311±2.513	−0.482 (−1.983 to 1.084)
Female radius	n=52	n=30
Total area, mm2	0.104±0.847	0.142±1.237	0.042 (−0.438 to 0.510)
Total density, mg HA/cm3	−0.593±0.876	−0.910±0.738	−0.309 (−0.742 to 0.061)
Cortical area, mm2	−0.879±1.018	−1.057±0.802	−0.175 (−0.630 to 0.266)
Cortical density, mm	−0.959±1.230	−1.076±0.942	−0.108 (−0.646 to 0.441)
Cortical thickness mm	−0.817±1.006	−1.010±0.687	−0.190 (−0.629 to 0.242)
Cortical perimeter mm	0.123±0.933	0.203±1.285	0.082 (−0.399 to 0.599)
Cortical porosity	0.740±1.372	0.909±1.371	0.176 (−0.468 to 0.776)
Trabecular area, mm2	0.325±0.921	0.408±1.200	0.097 (−0.385 to 0.600)
Trabecular density, mg HA/cm3	−0.173±0.875a	−0.633±0.785a	−0.455 (−0.877 to −0.052)a
Trabecular number, 1/mm	0.038±0.890a	−0.623±1.149a	−0.647 (−1.106 to −0.165)a
Trabecular thickness, mm	−0.326±0.898	−0.255±0.932	0.076 (−0.365 to 0.523)
Trabecular spacing, mm	−0.063±0.801a	1.047±2.464a	1.119 (0.397 to 1.874)a
Trabecular heterogeneity, mm	−0.040±0.799a	1.097±2.750a	1.127 (0.273 to 1.898)a
Female tibia
Total area, mm2	0.144±0.973	0.410±0.999	0.278 (−0.195 to 0.749)
Total density, mg HA/cm3	−0.523±1.041a	−1.161±0.751a	−0.631 (−1.069 to −0.209)a
Cortical area, mm2	−0.664±1.788	−1.358±0.967	−0.676 (−1.424 to 0.058)
Cortical density, mg HA/cm3	−0.954±1.024	−1.310±0.913	−0.356 (−0.815 to 0.106)
Cortical thickness, mm	0.688±1.427a	−1.331±0.814a	−0.652 (−1.247 to −0.092)a
Cortical perimeter, mm	0.321±1.017	0.650±0.960	0.339 (−0.170 to 0.782)
Cortical porosity	0.678±0.885	0.850±0.855	0.172 (−0.235 to 0.598)
Trabecular area, mm2	0.264±0.942	0.648±0.916	0.234 (−0.071 to 0.833)
Trabecular density, mg HA/cm3	−0.130±0.894	−0.625±0.810	−0.493 (−0.895 to −0.088)
Trabecular number, 1/mm	0.030±0.918a	−0.707±0.794a	−0.725 (−1.126 to −0.321)a
Trabecular thickness, mm	−0.244±0.927	0.004±1.187	0.269 (−0.219 to 0.772)
Trabecular spacing, mm	−0.068±0.783a	0.776±1.669a	0.841 (0.222 to 1.353)a
Trabecular heterogeneity, mm	0.001±0.802a	1.029±2.540a	1.037 (0.299 to 1.873)a

HPD, highest posterior density; HA, hydroxyapatite.

^aStatistically significant.

Prospective Cohort

Fifty-four study participants returned for a follow-up visit at a median (min–max) of 1.54 (0.9–4.3) years from the baseline visit. Demographic details of this group are presented in Supplemental Table 2. The mean difference in bicarbonate for all participants over the follow-up period was +1.1±3.9 mEq/dl (P < 0.05), and levels of 25-dihydroxy vitamin D (+4.5±10.4, P < 0.01) and sclerostin (+0.223±0.376, P < 0.001) increased over time. Baseline bicarbonate levels were associated with changes in serum phosphorus (ρ=0.45, P = 0.04), and there were no associations between changes in bicarbonate and changes in BTMs.

Association between Bicarbonate and Changes in Skeletal Imaging

We used linear regression to determine prospective relationships between bicarbonate and Z-score measures of bone density and quality. The primary exposure variables for all models included the baseline bone measure and baseline and absolute change in bicarbonate and eGFR. For DXA, there were no relationships between changes in bone outcomes and either the baseline or change in bicarbonate. By contrast (Table 3), for HR-pQCT at the radius, more severe acidosis at baseline and worsening acidosis over time were both related to decreases in trabecular number, while more severe acidosis at baseline was related to increases in trabecular separation and network heterogeneity. For HR-pQCT at the tibia, more severe acidosis at baseline was related to decreases in heterogeneity of the trabecular network, and worsening acidosis over time was related to expansion of the cortex (increases in cortical area and perimeter) and increases in cortical density (Figure 1). We examined the contribution of the hemodialysis state on the longitudinal change in DXA and HRpQCT variables as a function of change in CO₂ level. Tests of the heterogeneity of slope differences between groups defined by hemodialysis status did not differ in their bone response to changing CO₂ level (Supplemental Table 3), with the exception of trabecular thickness at the radius. At the radius, for every mEq decrease in bicarbonate from baseline, trabecular thickness Z-score fell 0.094 more in the dialysis group than the nondialysis group.

Table 3.

Regression models of association between absolute change in serum bicarbonate with change in bone imaging Z-score outcomes

Bone Parameters	Model 1a		Model 2b
	Baseline CO2	ΔCO2	Baseline CO2	ΔCO2
Radius
Total area	−0.0057	−0.021	−0.0061	−0.021
Total density	0.014	0.012	0.0097	0.012
Cortical area	0.010	0.0035	0.0036	0.0054
Cortical density	0.0081	−0.024	−0.0045	−0.022
Cortical perimeter	−0.0032	−0.019	−0.0040	−0.018
Cortical thickness	0.016	0.011	0.0087	0.013
Cortical porosity	0.015	0.021	0.054	0.011
Trabecular area	−0.0094	−0.020	−0.0097	−0.020
Trabecular density	0.0055	0.0032	0.012	0.00076
Trabecular number	0.036	0.052c	0.051c	0.047c
Trabecular thickness	−0.033	−0.022	−0.045	−0.023
Trabecular spacing	−0.083	−0.037	−0.12d	−0.029
Trabecular heterogeneity	−0.055	−0.038	−0.082c	−0.032
Tibia
Total area	0.00499	−0.0037	0.0043	−0.0032
Total density	−0.0092	−0.022c	−0.020	−0.017
Cortical area	−0.014	−0.033c	−0.028	−0.034c
Cortical density	−0.018	−0.042d	−0.030	−0.027c
Cortical perimeter	0.0037	−0.025e	0.0024	−0.023d
Cortical thickness	−0.010	−0.032c	−0.022	−0.027
Cortical porosity	0.0014	0.021	0.011	0.015
Trabecular area	0.0077	0.0028	0.0094	0.0023
Trabecular density	0.0038	−0.0044	−0.0019	−0.0028
Trabecular number	0.0047	0.0051	−0.0071	0.0052
Trabecular thickness	0.0076	−0.0039	0.010	−0.00040
Trabecular spacing	0.017	0.0089	0.040	0.0094
Trabecular heterogeneity	0.044	0.010	0.090c	0.015

^aAdjusted for baseline bone outcome.

^bAdditional adjustment for baseline and change in eGFR.

^cP < 0.05.

^dP < 0.01.

^eP < 0.001.

Figure 1.

Representative images from high-resolution peripheral computed tomography from a 71-year-old man with acidosis and with eGFR 26 ml/min and PTH 72 pg/ml (A) and an 86-year-old man without acidosis and with an eGFR 32 ml/min and PTH 92 pg/ml (B). PTH, parathyroid hormone.

Biopsy

Twenty-two participants with CKD underwent tetracycline double-labeled transiliac crest bone biopsy; demographic details are presented in Supplemental Table 4. We evaluated continuous and categorical relationships between severity of acidosis and bone turnover, mineralization and volume from histomorphometry, 3-dimensional trabecular microarchitecture and tissue mineral density from microCT, and the content and distribution of calcium in bone-tissue by qBEI (Table 4). Bayesian estimates demonstrated that more severe acidosis was associated with lower tissue mineral density (ρ=0.60, P = 0.004) by microCT. By qBEI, there was a shift of calcium content distributions to lower levels and a broader distribution (i.e., more heterogeneous calcium content) in trabecular and cortical bone (Figure 2; Supplemental Table 5). In exploratory analyses, lower tissue mineral density and lower and more heterogeneous calcium content and distribution were related to larger cortical circumference (i.e., perimeter) and larger cortical but smaller trabecular areas (P < 0.05 for all relationships, data not shown).

Table 4.

Histomorphometric features by acidosis status in 22 participants

Bone Parameters	Total	Nonacidotic (n=12)	Acidotic (n=10)	Bayes Difference (95% HPD Interval)
Turnover
BRF (µm3/µm2 per day)	0.019±0.031	0.008±0.007	0.031±0.042	0.015 (−0.458 to 0.557)
Eroded surface/BS	7.52±3.58	6.58±2.98	8.64±4.06	2.01 (−1.47 to 4.83)
Adjusted apposition rate	0.299±0.451	0.166±0.114	0.458±0.639	0.278 (−0.317 to 0.961)
Mineralization
MS/BS (%)	7.67±10.17	3.85±3.42	12.25±13.57	8.41 (−0.84 to 16.02)
MLT (d)	40.4±38.7	52.7±47.1	25.2±17.6	−27.9 (−62.1 to 8.8)
MAR (µm/d)	0.498±0.265	0.492±0.247	0.506±0.298	0.002 (−0.610 to 0.508)
Volume and structure
BV/TV (%)	18.0±5.4	17.7±2.8	18.5±7.6	0.85 (−4.0 to 5.9)
BS/TV (%)	3.26±0.69	3.38±0.45	3.11±0.90	−0.25 (−1.08 to 0.50)
Trabecular width (µm)	111±21	106±17	117±24	11.1 (−5.9 to 30.4)
Trabecular number (mm−1)	1.63±0.34	1.69±0.22	1.56±0.45	−0.14 (−0.69 to 0.49)
Trabecular spacing (µm)	633±154	592±82	682±205	90 (−33 to 222)
Cortical width (µm)	531±266	486±223	585±314	105 (−121 to 328)
Cortical area (mm2)	15.3±8.9	13.0±7.6	18.1±10.0	5.0 (−4 to 12)
Cortical porosity (%)	10.2±5.1	9.5±4.8	11.0±5.5	1.4 (−2.8 to 6.3)
Cortical porosity number	7.1±2.8	7.5±3.6	6.6±1.5	−0.9 (−3.7 to 1.4)
microCT
Trabecular number	1.57±0.35	1.58±0.36	1.56±0.34	−0.01 (−0.60 to 0.61)
Trabecular thickness	0.144±0.024	0.145±0.021	0.143±0.029	−0.010 (−0.483 to 0.532)
Tissue mineral density	995±44	1012±43	974±36	−38 (−72 to −2)a
BV/TV	0.192±0.069	0.185±0.051	0.200±0.088	0.017 (−0.491 to 0.515)
Connectivity density	12.7±21.1	15.7±28.5	9.1±4.2	−6.8 (−24.7 to 12.2)
Structure model index	0.518±0.935	0.823±0.693	0.151±1.086	−0.682 (−1.560 to 0.241)
Trabecular spacing	0.729±0.168	0.716±0.154	0.744±0.190	0.019 (−0.474 to 0.584)
Apparent mineral density	216±80	212±66	220±97	7 (−66 to 74)
Degree of anisotropy	1.40±0.19	1.37±0.16	1.43±0.23	0.05 (−0.53 to 0.54)

Data reported as mean±SD unless otherwise noted. HPD, highest posterior density; BRF, bone formation rate; BS, bone surface; MS, mineralizing surface; MLT, mineralization lag time; MAR, mineral apposition rate; BV, bone volume; TV, total volume.

^aStatistically significant.

Discussion

This investigation provides a comprehensive assessment on the skeletal associations of acidosis with bone imaging, biomarkers of turnover, and biopsy. Our study is unique from prior investigations on the role of acidosis in CKD-related bone disease for three critical reasons: (1) we provide a comprehensive skeletal assessment in patients not taking aluminum-based phosphate binders; (2) for the first time, we apply state-of-the-art methods to assess associations of acidosis on bone quality—including HR-pQCT of the peripheral skeleton and microCT and qBEI of bone-tissue; and (3) we present analyses adjusted for kidney function, enabling determination of independent skeletal correlations with acidosis. We detected multiple skeletal abnormalities associated with acidosis. Cross-sectionally, participants with MA had higher levels of BTMs. By bone imaging, subjects with acidosis had lower areal BMD at the one-third radius by DXA. In women with acidosis, trabecular density and microarchitecture were impaired and cortices were thinner by HR-pQCT. Longitudinally, acidosis was associated with broken trabeculae (i.e., worsening heterogeneity of the trabecular network) and cortical expansion. At the bone-tissue level, more severe acidosis was associated with decreased tissue mineral density and lower and more heterogenous calcium content. In exploratory analyses, lower and more heterogenous tissue mineral content was related to cortical geometric alterations.

In vitro and in vivo animal data have greatly informed our understanding of the effects of MA on bone-tissue. Acidosis has been shown to induce the expression of osteoblast receptor activator of nuclear factor κ B ligand, which stimulates osteoclast activity and promotes bone resorption.⁴⁶ Indeed, incubation of mouse calvariae at reduced pH suggested uncoupling of bone turnover favoring resorption, with suppression of collagen synthesis and alkaline phosphatase (i.e., suppressed markers of osteoblast function) in conjunction with a net increase in calcium efflux that was associated with increased osteoclast activity.¹⁸ Furthermore, in an in vivo rat model of MA, BTMs demonstrated uncoupling, with decreases in the formation marker osteocalcin and increases in the resorption marker Trap5b. We found similar relationships in humans with CKD. with acidosis, markers of both bone formation and resorption were increased, suggesting increased remodeling. Using standard histomorphometry and state-of-the-art methods to assess bone-tissue–level microstructure and mineral content and distribution, acidosis was associated with deficits in tissue mineral density and calcium content and distribution. These findings support but do not establish a causal role for acidosis in altering the bone-tissue mineral characteristics, independent of other metabolic complications that occur contemporaneously with CKD.

In both cross-sectional and prospective analyses, important findings associated with acidosis that have broad implications for bone strength included impairments in bone density, geometry, and microarchitecture. In rat models of acidosis, changes in bone geometry and microarchitecture measured by pQCT and microCT have been reported.¹⁹ In rats with acidosis compared with without acidosis, cortical mass decreased because of changes in cortical geometry that included cortical thinning and expansion of trabecular area (i.e., endocortical trabecularization), without changes in cortical perimeter. Furthermore, trabecula dropout occurred, resulting in decreased trabecular number and increased spacing between trabeculae.¹⁹ Similar to animal studies, cross-sectionally, we reported that subjects with acidosis had lower areal BMD at the mainly cortical one-third radius by DXA. For the first time, with high-resolution imaging by HR-pQCT, we demonstrated that multiple microarchitectural impairments to bone quality occurred more often in acidotic women. At the non–weight-bearing radius, defects in trabecular bone quality included lower trabecular density and number with wider spacing between trabecula. By contrast, at the tibia, both trabecular and cortical bone were impaired, including lower trabecular number and wider spacing, thinned cortices, and lower total bone density. These impairments detected by DXA and HR-pQCT reduce resistance to fracture. In the prospective cohort, more severe acidosis at baseline and worsening acidosis over time were associated with changes in bone, independent of sex and kidney function. Examination of the interaction between bicarbonate change and dialysis status showed statistically significant slope change for only one parameter, within the expected false positive finding rate for the number of tests performed. At the radius, more severe acidosis at baseline and decreases in bicarbonate over time were associated with trabecula dropout and increases in trabecular network heterogeneity. By contrast, at the tibia, more severe baseline acidosis was associated with decreased trabecular network heterogeneity, and decreases in bicarbonate over time were associated with increased cortical density, area, and circumference. It is interesting to note that while the changes in trabecular bone at the radius predispose to fractures, the changes in cortical geometry and trabecular microarchitecture at the tibia protect against fractures. Cortical bone comprises more than 85% of the skeleton and is essential to supporting axial loads.^3,33,47–57 In simulated models of bone atrophy, reductions in cortical thickness had greater negative effect on whole bone strength than reductions in trabecular number or thickness.⁵⁶ Although the mechanisms underlying acidosis-associated cortical expansion and decreased trabecular heterogeneity cannot be determined from this study design, we hypothesize that this is a compensatory protective mechanism, perhaps mediated by weight bearing, to preserve overall bone strength in the setting of acidosis-associated defects in trabecular microarchitecture and bone-tissue–level mineralization, both of which impair resistance to fracture. Indeed, in exploratory analyses, lower tissue mineral density and lower and more heterogenous bone-tissue calcium content were associated with larger cortical perimeter and area. Although we expected decreased cortical density to correlate with worsening acidosis, cortical expansion correlated moderately with density (r=0.58, P = 0.006, data not shown) and likely explains in part this paradoxical finding.

It is noteworthy that in cross-sectional analyses, alterations in bone geometry, density, and microarchitecture were detected only in women. Prior investigations on the skeletal effects of acidosis reported similar sex-specific findings. In a rat model of acidosis, the combination of ovariectomy with acidosis accelerated bone loss compared with that in rats with ovariectomy without acidosis and in rats without ovariectomy regardless of acidosis status.¹⁹ In humans, Chen et al.²¹ used National Health and Nutrition Examination Survey to study the effects of acidosis on both total body and lumbar spine areal BMD by DXA in members of the general population older than 20 years. Acidosis was associated with lower total body and lumbar spine BMD in postmenopausal women. While in men, there was no association between bicarbonate and total body BMD, and the association between bicarbonate and lumbar spine BMD was not apparent until bicarbonate levels exceeded 27 mEq/L. Similarly, in premenopausal women, there was no association with total body BMD and only a weak association between bicarbonate and lumbar spine BMD. These findings suggest (but do not establish) that acidosis effect modifies bone loss because of estrogen withdrawal and may have implications on the use of alkali supplementation to protect skeletal health in patients with and without CKD.

This study has several limitations. This is a secondary analysis of an investigation designed to describe the skeleton in subjects with kidney disease. However, acidosis is a common complication of CKD and falls within the scope of the parent study. These data are observational and hypothesis-generating. However, our use of cross-sectional, prospective, and bone biopsy cohorts provides a comprehensive description of skeletal changes over time in the setting of MA at macrostructural, microstructural, and nanostructural levels. Follow-up visit duration was not standardized across the cohort and ranged from 0.9 to 4.3 years. We used bicarbonate and not serum pH as a measure of acidosis; however, this is similar to most data on the associations of acidosis on bone outcomes in human subjects' research. We did not create discovery and validation datasets and instead relied on Bayes estimates and robust credible intervals to report acidosis group differences. Although we did not statistically adjust for multiple testing, the consistency of associations seen, especially between acidosis and cortical expansion at the tibia, make false positives less plausible. We detected sex-specific differences on associations between acidosis and bone quality. This study was not designed to assess sex-specific differences of acidosis on the skeleton, and these findings need further investigation. The bone biopsy findings would be strengthened by a larger sample size, although this is difficult to attain given its invasiveness. Yet, this remains one of the largest bone biopsy studies on acidosis.

In conclusion, to the best of our knowledge, this is the first study assessing the comprehensive skeletal associations of MA in patients with CKD in the nonaluminum binder era. In this cohort of 180 patients, we demonstrated that MA, independent of kidney function, was associated with decreased tissue mineral density, lower and more heterogenous calcium content, and defects in cortical and trabecular bone quality in women. These data provide a comprehensive skeletal phenotype of MA in CKD and indicate several bone targets to assess efficacy in future alkali trials. Future work should investigate the effects of correction of MA on these skeletal abnormalities, including prospective observational studies and randomized controlled trials. Such trials would benefit from incorporating state-of-the-art imaging and histomorphometric assessment, as performed here. They should strive to assess patients across the spectrum of CKD including patients with early stage disease in which frank hyperparathyroidism may not yet be evident as well as those receiving kidney replacement therapy. Finally, any analysis should include prespecified sex stratification to account for hormonal differences in bone health.

Supplementary Material

jasn-34-668-s001.pdf ^{(304.4KB, pdf)}

Disclosures

D.J. Mcmahon reports Other Interests or Relationships: provides statistical consulting services to Hospital for Special Surgery and New York University, both located in New York City. D. Dempster reports Consultancy: Amgen Inc, Radius Health; Honoraria: Amgen, Inc, Radius Health; Advisory or Leadership Role: Amgen, Inc., Radius Health; and Speakers Bureau: Amgen, Inc., and Radius Health. and K. Reidy reports Research Funding: Advicienne, Travere; Advisory or Leadership Role: AAP Executive committee, Frontiers, Neonatal Kidney Collaborative; and Other Interests or Relationships: NIH funding, Preeclampsia Foundation funding, Ruth Gottscho Kidney Foundation. J. Kumar reports Honoraria: Reata Pharmaceuticals, one time advisory board participation. M. Fusaro reports Consultancy: Vifor-Pharma; and Honoraria: ABIOGEN, and AMGEN. M.L. Melamed reports Advisory or Leadership Role: American Board of Internal Medicine Nephrology Exam Committee; and Other Interests or Relationships: New York Society of Nephrology. T.L. Nickolas reports Consultancy: Alnylam, Pharmacosmos; Research Funding: Amgen; Honoraria: Alnylam, Pharmacosmos; Patents or Royalties: Columbia University has licensed patents on NGAL to Abbott Diagnostics and Alere; and Advisory or Leadership Role: Amgen, and Pharmacosmos. All remaining authors have nothing to disclose.

Funding

R.V. Levy is sponsored by National Institutes of Health grant NIDDK T32-DK007110. T.L. Nickolas is sponsored by National Institutes of Health grant K23 DK080139.

Author Contributions

Conceptualization: Donald J. McMahon, Thomas L. Nickolas

Data curation: Sanchita Agarwal, Maria Alejandra Aponte, Denver D. Brown, David Dempster, X. E. Guo, Mafo Kamanda-Kosseh, Juhi Kumar, Donald J. McMahon, Michal L. Melamed, Barbara M. Misof, Thomas L. Nickolas, Hua Zhou

Formal analysis: Maria Fusaro, X. E. Guo, Rebecca V. Levy, Donald J. McMahon, Barbara M. Misof, Thomas L. Nickolas, Kimberly Reidy

Funding acquisition: Thomas L. Nickolas

Investigation: Thomas L. Nickolas

Methodology: Thomas L. Nickolas

Project administration: Mafo Kamanda-Kosseh

Supervision: Thomas L. Nickolas

Writing – original draft: Rebecca V. Levy, Donald J. McMahon

Writing – review & editing: Sanchita Agarwal, Denver D. Brown, David Dempster, Maria Fusaro, X. E. Guo, Juhi Kumar, Rebecca V. Levy, Donald J. McMahon, Michal L. Melamed, Barbara M. Misof, Thomas L. Nickolas, Kimberly Reidy, Hua Zhou

Data Sharing Statement

Individual participant data that underlie the results reported in this article will be available after deidentification to investigators whose proposed use of the data has been approved by an independent review committee for any purpose. Data will be available from three months after publication until five years after publication. Proposals should be directed to the corresponding author. The study protocol, statistical analysis plan, and analytic code will all be available to anyone who wishes to access them immediately following publication for five years after publication. Requests should be directed to the corresponding author.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/JSN/D734.

Supplemental Table 1. Baseline demographic data on 54 longitudinal participants by acidosis status.

Supplemental Table 2. Baseline demographic data on 22 biopsy participants by acidosis status.

Supplemental Table 3. Interaction of ΔCO₂ with dialysis status on change in imaging parameters.

Supplemental Table 4. Baseline demographic data on 22 biopsy participants by acidosis status.

Supplemental Table 5. Spearman correlations between baseline serum bicarbonate and bone biopsy measures in 22 participants with adjustment for baseline eGFR.