The Evaluation of Renal Osteodystrophy with Cortical Quantitative Ultrasound at Various Bone Sites

Abstract

Background: In this study we evaluated bone abnormalities of patients with chronic renal failure (CRF) by cortical quantitative bone ultrasound (QUS) measurements at different bone sites because these abnormalities may be of variable etiology and may start before symptoms or radiological changes are manifested. Methods: Of fifteen patients with chronic renal failure, seven had moderate–severe disease and eight had been on chronic hemodialysis from 5 to 17 years, with renal osteodystrophy (ROD) confirmed by bone biopsies. Twelve normal subjects of similar age and gender volunteered for the control group. Results: Patients and controls differed in creatinine clearance, in serum phosphate levels and in serum total alkaline phosphatase. Mean intact‐PTH levels differed significantly amongst the three groups of subjects. All patients with ROD had intact‐PTH higher than 200 pg/mL. The cortical ultrasound parameter, speed of sound (SOS), was slower in patients with more severe renal failure at all bone sites measured. The group with ROD had significantly lower cortical ultrasound values than the other patients and the control group at all sites. SOS values at the proximal phalanx, distal radius and midtibial sites were positively and significantly correlated. Cortical ultrasound measurements at the radial site correlated with midtibial and phalangeal sites but the correlation between midtibial and phalangeal sites did not reach significance. IPTH levels correlated negatively and significantly with cortical QUS values at all sites being the correlations higher at phalangeal and radial sites than at the midtibial region. Conclusions: The differences in cortical ultrasound observed indicate the potential clinical application of this methodology to evaluate bone abnormalities in chronic renal failure, especially in patients on chronic hemodialysis.

• Chronic hemodialysis
• Chronic renal failure
• Intact‐PTH
• Quantitative ultrasonometer
• Renal osteodystrophy
• Bone biopsy
• Cortical ultrasound
• Metabolic bone disease
• Total alkaline phosphatase

Introduction

Renal osteodystrophy (ROD) is still a common and, unfortunately, debilitating condition of patients with end‐stage renal disease (ESRD), particularly, those maintained on chronic hemodialysis (CHD). It is known that ROD comprises various types of metabolic bone diseases but the most prevalent form is related to high parathormone (PTH) levels.[1], [2] As kidney function declines, the disturbed calcium–phosphate homeostasis increases the PTH levels in an attempt to correct the mineral imbalance.[3] Intact PTH (IPTH) or (1–84) PTH levels have been used to monitor the development of ROD. Levels of IPTH above 200 pg/mL are generally observed in high‐turnover ROD.[1] It is now known that most IPTH assays may cross‐react with non‐ (1–84) PTH fragments that may antagonize the actions of PTH.[3], [4] In addition, IPTH levels do not increase as expected when there is aluminum deposition in the body.[5] Therefore, the use of IPTH values to indicate the presence of ROD in ESRD seems to be limited and assay dependent.

Bone abnormalities of chronic renal failure (CRF) could start long before symptoms or radiological changes are manifested. Non‐decalcified bone biopsy has been the standard diagnostic test for ROD but it involves some discomfort, requires technical expertise and is generally performed late in the course of these diseases.[1] Some studies have tried to predict early ROD by bone mineral density (BMD) measurements,[6], [7], [9] by bone turnover biochemical markers[7], [8] as well as by bone quantitative ultrasonometer (QUS), with limited success.[9], [10], [11] Most of these techniques are expensive, may pose problems in the interpretation, evaluate only few skeletal regions and do not discriminate the type of bone most affected by ROD, i.e., the cortical bone.[1], [6], [7] In this study, we tested a QUS device[16] capable of evaluating cortical bone regions by measuring the speed of sound (SOS) at multiple skeletal sites[12] in a group of patients with CRF of different etiology and severity, and assessed its relation to IPTH levels and its usefulness to detect ROD.

Subjects and Methods

We studied 15 patients followed at the Nephrology Division of the University Hospital—School of Medicine of Ribeirão Preto, University of São Paulo, Brazil. Seven patients had moderate–severe CRF and 8 patients had been on CHD for 5 to 17 years, all with ROD confirmed by bone histomorphometry, none of them had previous parathyroidectomy. Twelve normal healthy subjects of similar age and gender volunteered for the control group (Table 1).

Table 1. Group Characteristics and Biochemical and Bone Quantitative Ultrasound Values (Mean ± SEM)

	Control	CRF	ROD	P
Gender	6M/6F	4M/3F	5M/3F	NS
Age (years)	36.2 ± 2.7	48.9 ± 4.2	36.2 ± 4.4	NS
Creat. Cl.^1 (mL/min)	106.3 ± 4.57	19.5 ± 3.1	NA
s. Ca ^2 + (mmol/L)	1.23 ± 0.01	1.16 ± 0.01	1.17 ± 0.04	NS
s. PO (mg/dL)	3.48 ± 0.15	4.94 ± 0.41	5.86 ± 0.55	**
s. Alk. Phos^1 (U/L)	63.2 ± 5.2	116.2 ± 10.9	165.9 ± 20.22	**#
IPTH^1 (pg/mL)	38.67 ± 4.9	309.7 ± 87.1	792.8 ± 122.2	**
SOS‐TIBIA (m/s)	3941.8 ± 30.2	3945.6 ± 47.8	3596.6 ± 88.5	**#
SOS‐1/3 RAD (m/s)	4074.5 ± 37.5	4065.0 ± 36.3	3664.1 ± 121.9	**#
SOS‐PHAL.X (m/s)	3812.0 ± 35.3	3720.9 ± 52.2	3520.9 ± 55.2	**#

ANOVA with Bonferroni's test or ¹Kruskal‐Wallis (Dunn's) test.

^*(Control vs. CRF).

^**(Control vs. ROD).

^#(CRF vs. ROD) for p < 0.05.

NA—not applicable, NS—not significant.

The CRF group had the following diagnosis: diabetes mellitus in 4, essential arterial hypertension in 2 and gout in 1. They were managed conservatively for renal failure. The ROD group consisted of 5 patients with chronic glomerulonephritis and chronic pyelonephritis, bilateral cortical necrosis, and hypertensive nephrosclerosis, one of each. These patients had typical clinical and radiographic features of ROD. Osteitis fibrosa (OF) was present in 4, adynamic osteodystrophy (AO) in 3 and mixed osteodystrophy in one. Except for 2 with OF, all the others had some aluminum (Al) deposit in the bone biopsies.

Patients and volunteers gave consent for all tests performed in this study, which was approved by the School of Medicine Ethics Committee. Patients were maintained on their usual diet or medications.

Fasting blood samples were collected from 08:00 h to 10:00 h. Biochemical parameters were evaluated by standard laboratory methods. Creatinine clearance values were normalized for 1.73 m² of body surface. Serum‐ionized calcium was determined by the ion‐selective electrode method, with a normal range of 1.15–1.29 mmol/L (Radiometer ICA II, Copenhagen, Denmark) and serum IPTH by immunochemoluminescense (ICMA) (Diagnostic Products Corporation, Los Angeles, CA, USA). This normal range of the assay is 12 to 72 pg/mL and the results show good linearity up to 16‐fold dilutions.

Quantitative bone ultrasound (QUS) determinations of the SOS, in m/s, were performed by one observer (JASC) with the Sunlight Omnisense device (Sunlight Ultrasound Technologies Ltd., Israel) without knowledge of the underlying diagnosis, although all patients in CHD had AV fistulae. The QUS measurements were made with the patient in the sitting position at the one‐third distal radius (RAD) in the non‐dominant forearm, except in CHD patients with an AV fistula in this forearm region. QUS were also measured in the proximal phalanx (PHAL) of the third finger of the same upper limb and in the midtibial (TIB) region of the nondominant leg.

QUS of the three sites were done on the same day and each site QUS value was reported as the mean of three repeated measurements in sequence. The large probe was used for RAD and TIB sites and the small probe for PHAL. Quality control was performed daily for both probes with a reference bar phantom as well as in the same control subject measured at RAD and PHAL sites. Percent coefficient of variation (CV) for the bar phantom ranged from 0.13% to 0.17% for the large and the small probe, respectively. CV for the control subject was 0.7% at the RAD and 1.2% at the PHAL sites.

Statistical comparisons between groups were made by ANOVA and Bonferroni tests. Nonparametric comparisons were performed by Kruskal‐Wallis and Dunn tests. Spearman's test was applied to determine the associations of QUS with IPTH and total alkaline phosphatase values. Statistical significance was accepted for P less than 0.05.

Results

As expected, the 3 groups differed in creatinine clearance, in serum phosphate (s. P.) and in serum total alkaline phosphatase (s. Alk. Phos) levels. Mean serum‐ionized calcium (s. Ca 2 +) was not significantly different (Table 1). Mean IPTH levels differed significantly among the three groups of subjects, ranging from 42.6 to 739.0 pg/mL, in CRF group, IPTH values were below 200 pg/mL except in 3 patients. In the ROD group, all patients had IPTH higher than 200 pg/mL.

Mean SOS values at all sites measured were slower in the patients. The ROD group had significantly lower QUS values at all sites than CRF and Control. The QUS values of the CRF group were not significantly different from control (Table 1).

SOS values of the 3 skeletal sites were positively and significantly correlated with each other, except between the midtibial and phalangeal sites (rs=0.384) (Table 2). IPTH and s. Alk. Phos levels strongly correlated with each other and also were negatively and significantly correlated with cortical QUS values at all sites. IPTH was more strongly correlated with SOS measurements at phalangeal and radial sites than in the mid‐tibial region.

Table 2. Spearman's correlation coefficients between SOS (m/s) in the midtibia, third distal radius and third proximal phalanx and IPTH (pg/mL) and serum alkaline phosphatase (U/L)

	IPTH	s. Alk. Phos	TIBIA	RAD	PHAL.X
IPTH	–	0.824	(−) 0.398#	(−) 0.660	(−) 0.665
s. Alk. Phos	0.824	–	(−) 0.393#	(−) 0.452#	(−) 0.540
TIBIA	(−) 0.398#	(−) 0.393#	–	0.600	0.384
RAD	(−) 0.660	(−) 0.452#	0.600	–	0.583#
PHAL.X	(−) 0.665	(−) 0.540	0.384	0.583#	–

^*p < 0.001.

^#p < 0.05.

Discussion

In this study we observed a reduction of cortical QUS values in various skeletal sites as the kidney function deteriorated in the groups evaluated. At the radius and at the phalanx, the SOS values reduced gradually as the kidney function decreased, but at the reduction in midtibial, SOS was observed only in patients with histomorphometrically confirmed ROD. In the ROD group, SOS values were significantly slower than Control or CRF groups at all three sites. Mean SOS values at the radial and phalangeal sites were lower in CRF patients than in Controls but the difference was not significant. Previous studies have shown reductions of heel, phalangeal, as well as of midtibial QUS measurements values in ESRD patients mainly in relation to the time on dialysis.[9], [10], [11] In the study by Foldes et al.,[11] the patients on long‐tern dialysis presented a significant and inverse relation of midtibial SOS and IPTH with rs = − 0.39, similar to that found in the present study for the same bone region. We observed that SOS values were negatively and significantly correlated with serum IPTH values at all three skeletal sites. Moreover, when only the patients from the CRF and ROD groups were considered, the correlations between IPTH and cortical QUS measurements were even higher and remained significant (data not shown). For all subjects in this study, the correlation of SOS and IPTH at the radius and at proximal phalanx were, rs = − 0.66 (p < =0.001) higher than at the midtibial site, indicating the better sensitivity of QUS measurements in the upper extremities to estimate the bone abnormalities of impaired kidney function. Serum IPTH levels, determined by an assay that does not discriminate large PTH fragments, were significantly different in the 3 groups. None of the ROD patients had values below 200 pg/mL but 3 of 7 with CRF had IPTH values above 200 pg/mL. Also, in 3 patients from the ROD group with adynamic osteodystrophy, IPTH levels were elevated. These elevations may be caused by a different pathophysiologic mechanism or by the lack of specificity of the IPTH assay employed.

QUS parameters of bone are considered good estimates of bone resistance to fragility fractures.[13] The combination of QUS measurements at multiple sites seems to improve the prediction of osteoporotic fractures compared to a single bone region determination.[12] Recently, reference data have been published for American women for the same type QUS device we applied in this study.[14] Our subjects, in all groups, included man and women. The cortical QUS measurements at the radius and midtibia of our control subjects were within the range for the American women but were lower at the phalanx. We have no plausible explanation for this variation since all our measurements had similar precision errors,[14] except the ethnic difference.[15] Besides the resistance to fractures, QUS may evaluate other bone material properties such as mineralization and microstructure that coexist in many metabolic bone diseases.[13], [16] We also observed that the QUS values for these groups were negatively, significantly related to total alkaline phosphatase values but not as strong as with IPTH levels. This suggests that cortical QUS could be reflecting a different mechanism for bone affection, not PTH related. Another possible explanation could be that previous treatment with calcitriol in ESRD patients may have affected differently the levels of total alkaline phosphatase and IPTH or yet, that an unidentified hepatic disorder could be present in some of our patients.

Our patient sample is small, variable in gender and in the diseases causing CRF. This could be seen as a limitation of our data but the differences we observed indicate the potential clinical application of this QUS methodology to evaluate the bone affection of CRF. In conclusion, we believe that cortical QUS evaluation at multiple bone sites could have an important role in the management of patients on dialysis, or may at least help identify candidates to more aggressive evaluation and/or treatment to prevent ROD.