Bone metabolism dynamics in the early post-transplant period following kidney and liver transplantation

Bone disease contributes to relevant morbidity after solid organ transplantation. Vitamin D has a crucial role for bone metabolism. Activation of vitamin D depends on the endocrine function of both, liver and kidney. Our study assessed key markers of bone metabolism at time of transplantation and 6 months after transplantation among 70 kidney and 70 liver recipients. In 70 kidney recipients 25-OH vitamin D levels did not differ significantly between peri-transplant (median 32.5nmol/l) and 6 months post-transplant (median 41.9nmol/l; P = 0.272). Six months post-transplant median 1, 25-(OH)2 vitamin D levels increased by >300% (from 9.1 to 36.5ng/l; P<0.001) and median intact parathyroid hormone levels decreased by 68.4% (from 208.7 to 66.0 ng/l; P<0.001). Median β-Crosslaps (CTx) and total procollagen type 1 amino-terminal propeptide (P1NP) decreased by 65.1% (from 1.32 to 0.46ng/ml; P<0.001) and 60.6% (from 158.2 to 62.3ng/ml; P<0.001), respectively. Kidney recipients with incident fractures had significantly lower levels of 1, 25-(OH)2 vitamin D at time of transplantation and of intact parathyroid hormone 6 months post-transplant. Among 70 liver recipients, 25-OH vitamin D, 1, 25-(OH)2 vitamin D and intact parathyroid hormone levels were not significantly altered between peri-transplant and 6 months post-transplant. Contrary to kidney recipients, median CTx increased by 60.0% (from 0.45 to 0.72 ng/ml; P = 0.002) and P1NP by 49.3% (from 84.0 to 125.4ng/ml; P = 0.001) in the longitudinal course. Assessed biomarkers didn’t differ between liver recipients with and without fractures. To conclude, the assessed panel of biomarkers proved highly dynamic after liver as well as kidney transplantation in the early post-transplant period. After kidney transplantation a significant gain in 1, 25-(OH)2 vitamin D combined with a decline in iPTH, CTx and P1NP, whereas after liver transplantation an increase in CTx and P1NP were characteristic.

Introduction Solid organ transplantation is an established treatment for patients with end-stage renal failure or liver insufficiency. In the US more than 19'000 kidney and 7'000 liver transplantations were performed in 2016 [1]. Bone disease and resulting fractures are an important co-morbidity in patients with end-stage organ disease [2,3].
In fact, it has been shown that the majority of liver recipients has abnormal bone mineral density (BMD) already at time of transplantation or has suffered from fractures pre-transplant [2,4]. This may be explained by excessive alcohol consumption in some of these patients [5], but hyperbilirubinemia [6], hypogonadism [7] and reduced insulin-like growth factor-1 levels [8] may also contribute to abnormal bone metabolism. Among patients with end-stage renal failure, bone health is impaired with renal osteodystrophy presenting either as osteitis fibrosa, osteomalacia, adynamic bone disease or a mixed type [9,10]. Consequently, both kidney and liver transplant patients have been found to have a high risk of bone loss and fractures [11,12].
Bone mineralization depends on adequate calcium and phosphate levels [13,14]. Two pivotal hormones regulating these minerals are 1, 25-(OH) 2 vitamin D (1, 25-(OH) 2 D) and parathyroid hormone (PTH). The most important source of vitamin D is synthesis in the epidermis via ultraviolet B exposure. Vitamin D undergoes a first hydroxylation at the 25-position in the liver resulting in 25-OH vitamin D (25-OHD). 25-OHD is still a precursor of the active hormone, but due to its long half-life of 2 to 3 weeks it best reflects vitamin D status [15]. The circulating active hormone, 1, 25-(OH) 2 D, emerges from 25-OHD via an additional hydroxylation at the 1-position in the kidney. 1, 25-(OH) 2 D increases calcium levels not only by stimulation of intestinal calcium resorption but also in corroboration with PTH by increased renal resorption and mobilization out of bone tissue. Furthermore, 1, 25-(OH) 2 D increases intestinal phosphate absorption and decreases phosphate excretion via the kidneys. Main drivers for production of 1, 25-(OH) 2 D are PTH or hypophosphatemia, whereas calcium, fibroblast growth factor 23 and the active hormone 1, 25-(OH) 2 D itself are inhibitory [13,16]. PTH excretion is suppressed by 1, 25-(OH) 2 D and a greater calcium intake [17].
Despite the substantial morbidity caused by impaired bone health among transplant patients, detailed data on bone metabolism changes after kidney or liver transplantation-both key organs of vitamin D hydroxylation-is limited. In the current study we used prospectively collected samples for measurements of vitamin D metabolites (25-OHD, 1, 25-(OH) 2 D), intact PTH (iPTH), creatinine and two bone turnover markers, β-Crosslaps (CTx) and total procollagen type 1 amino-terminal propeptide (P1NP), in the same patients at time of transplantation and 6 months after transplantation. In line with recent recommendations CTx was used as parameter for bone resorption and P1NP for bone formation [18,19]. We additionally reviewed all medical records for dual energy x-ray absorptiometry (DXA) scans performed, incident fractures recorded during post-transplant routine care, and supplementation of vitamin D and calcium.

Study design, population and patient-related data
This study was a nested project within the Swiss Transplant Cohort Study (STCS, www.stcs. ch). Since May 2008 data on all solid organ transplants carried out in Switzerland have been prospectively collected in the STCS database [20]. All Swiss transplant centers, i.e. Basel, Bern, Geneva, St. Gallen, Lausanne and Zurich, contribute to data acquisition. The STCS was approved by the Ethic Committees of all participating institutions, i.e. Ethikkommission Nordwest-und Zentralschweiz EKNZ, Ethikkommission Bern, Ethikkommission Genf, Ethikkommission Ostschweiz EKOS, Ethikkommission Zürich. None of the transplant donors were from a vulnerable population and all donors or next of kin provided written informed consent that was freely given". Information about vitamin D supplementation, calcium supplementation, BMD measurements and incidence of fractures were retrospectively collected by patient chart review, whereas all other data derived from prospective records. DXA scans to determine BMD were performed at the discretion of the treating physician. All fractures were radiographically confirmed. 70 kidney and 70 liver recipients were analyzed. Kidney transplantations and liver transplantations were performed between 05.05.2008 to 28.09.2009 and between 16.05.2008 to 20.12.2009, respectively. Median follow up was 5.6 years (IQR 5.5-5.8 years) in kidney recipients and 4.9 years (IQR 4.6-5.4 years) in liver graft recipients. For the subgroup analysis of female transplant recipients a simplified age-based approach was chosen to define menopausal status [21].

Laboratory analysis
For laboratory analyses prospectively collected citrate plasma samples, drawn at the time of transplant and 6 months post-transplant, which were stored at -80˚C in the STCS biobank. These samples were retrieved from the STCS biobank for centralized, uniform measurement at the Institute of Clinical Chemistry of the University Hospital Zurich. 25-OHD measurement was performed with Roche Diagnostics Vitamin D total assay on Cobas 8000 (Roche Diagnostics, Mannheim, Germany). Vitamin D status was categorized as follows: 25-OH vitamin D < 25nmol/l severe deficiency, ! 25 and < 50nmol/l deficiency and ! 50nmol/l no deficiency.
Creatinine was determined with a kinetic color test based on Jaffe Method from Roche Diagnostics (Mannheim, Germany) running on Cobas c701 system. Estimated glomerular filtration rate (eGFR) was calculated according to the CKD-EPI method [22], as this method was shown to provide more reliable results after liver transplantation [23].

Statistical analysis
All statistical analyses were performed with R (version 3.2.3). Continuous variables were reported as median and interquartile range (IQR), categorical variables as absolute numbers and frequencies (%). Statistical testing was performed with two-sided tests, p-values < 0.05 were considered significant. Wilcoxon rank-sum test was used for comparison of continuous variables between two groups, whereas Wilcoxon matched-pairs signed-rank test was applied for pairwise comparisons. Categorical variables were compared with Fisher's exact test. For investigation of linear relationships between two variables linear regression was used.

Patients' characteristics
A total of 140 consecutive patients, 70 kidney and 70 liver transplant recipients, were included in this study. Baseline characteristics are shown in Table 1.

Incident fractures and bone densitometry
Kidney recipients. Overall 7 fractures occurred in a total of 7 kidney recipients (corresponding to an incidence of 16 fractures per 1000 person-years); 4 (57.1%) fractures affected the lower extremities, 2 (28.6%) the upper extremities and 1 (14.3%) the spine. The median time interval from transplantation to fracture was 866 days (IQR 352-1368). Kidney recipients with incident fractures had lower 1, 25-(OH) 2 D levels peri-transplant (P = 0.008) ( Table 3). Likely, 6 months post-transplant 1, 25-(OH) 2 D tended to be higher in kidney recipients without incident fractures (P = 0.089). Patients suffering from incident fractures showed a marginally significant better eGFR (P = 0.049) at time of transplantation, lower iPTH levels (P = 0.008) and a trend of lower CTx levels (P = 0.064) 6 months after transplantation. BMI tended to be higher in patients without incident fractures (P = 0.054). In 36 of 70 kidney recipients DXA scans were available. As bone densitometry was performed by the treating physician's indication, a large variety in timespan from transplant to bone densitometry was observed (median 242 days, IQR 57-742). Osteoporosis, defined by a T-score of less than -2.5 at any location, was present in 9 (25%) and osteopenia (T-score ! − 2.5 and < − 1) in 20 (55.6%) individuals, whereas only 7 (19.4%) kidney transplant recipients had normal BMD. Notably, an increasing timespan until DXA scan was associated with a linear decrease in the T-score of the femoral neck (R 2 = 0.14, P = 0.027) (Fig 3). In 5 individuals with incident fractures bone densitometry was available. Fractures were recorded in 1 patient with osteoporosis, 3 patients with osteopenia and 1 patient with normal BMD.
Liver recipients. In 9 liver recipients fractures were observed, equaling an incidence of 23 fractures per 1000 person-years. Vertebral fractures were most common (n = 6, 66.7%), followed by 2 (22.2%) fractures of the upper extremities and 1 (11.1%) fracture of the lower extremities. Fractures occurred after a median of 274 days (IQR 171-661) following liver transplantation. Liver recipients with incident fractures were older (P = 0.004), but no significant differences were detected in BMI, diabetes mellitus, use of steroids for immunosuppression, phosphate levels and the assessed panel of biomarkers (Table 3). In 18 liver recipients BMD measurements were performed after transplantation. Normal BMD was detected in 5 (27.8%) liver transplant recipients, whereas in 11 (61.1%) and 2 (11.1%) patients osteopenia and osteoporosis were diagnosed, respectively. In 3 patients with incident fractures DXA scans were performed. 2 patients suffering from vertebral fractures had osteoporosis and 1 patient with vertebral fracture osteopenia.

Discussion
In both, kidney and liver recipients, 25-OH vitamin D remained on similar levels within the first 6 months post-transplant, whereas 1, 25-(OH) 2 D was significantly higher 6 months posttransplant in kidney recipients, but not in liver recipients. Levels of iPTH dropped in kidney recipients, but stayed unchanged in liver recipients. The bone turnover markers CTx and P1NP showed a significant decrease in kidney transplant recipients and on the contrary a significant increase in liver transplant recipients within the first 6 months after transplantation.
Parallel to the excretory renal function an improvement of the hormonal capacity of the transplant was observed after kidney transplantation, indicated by the increased ratio of 1, 25-(OH) 2 D to 25-OHD 6 months post-transplant. The low level of 1, 25-(OH) 2 D might be also partly due to the assumed premenopausal status of most kidney transplant recipients shown in a subgroup analysis (S2 Table). The high peri-transplant iPTH levels, reflecting most likely secondary hyperparathyroidism in the majority of kidney recipients, decreased to significantly lower levels 6 months post-transplant. In kidney transplant recipients, elevated peri-transplant levels of CTx and P1NP indicated simultaneous ongoing bone formation and bone resorption.  These contrarious processes were occurring simultaneously as indicated by the significant correlation between CTx and P1NP. The parallel elevation of these bone turnover markers at time of transplantation might reflect a state of high bone turnover, such as osteitis fibrosa, or might be, at least partly, caused by a diminished renal excretion of CTx and accumulation of the monomeric form of P1NP [24,25]. In agreement with our data, a linear relationship between CTx and P1NP was described by Ueda et al for hemodialysis patients [26]. The considerable increase in both, CTx and P1NP, at 6 months post-transplant in liver recipients is indicative of a high bone turnover state. The reasons for these changes in bone metabolism are unclear. One explanation may be a combined effect of vitamin D deficiency, triggering an overall, albeit not significant, increase in iPTH, and deterioration of kidney function, possibly due to calcineurin-inhibitor toxicity, thus promoting incipient renal osteodystrophy. This hypothesis would be supported by the significant decrease in eGFR of liver recipients. Alternatively, the change in eGFR might be influenced by a gain in muscle mass post-transplant.
The longitudinal development of the assessed biomarkers showed major differences between kidney and liver transplant recipients in 1, 25-(OH) 2 vitamin D, iPTH, CTx, P1NP, eGFR and phosphate levels. In kidney transplant recipients the increase in 1-position hydroxylation of vitamin D was to be expected as the decrease in iPTH due to hormonal activity of the graft. Likely, the improvement in eGFR and decrease in phosphate reflects exocrine renal function of the kidney graft. With regard to the hormonal function of the liver an increase in 25-OH vitamin D would be assumed. The avoidance of sun exposure resulting in a shortage of the precursor hormone vitamin D might be a possible explanation for the unchanged 25-OH vitamin D levels after liver transplantation. The observed decline in eGFR after liver transplantation has been reported before [27] and might be multifactorial as discussed above.
Bone turnover markers are rarely measured in clinical routine. Determination of CTx and P1NP has been encouraged as a monitoring tool in patients treated for various bone disorders [28,29]. Previous studies have demonstrated a correlation of CTx with biopsy-proven bone resorption [30]. A further application of bone turnover markers includes cancer patients, where CTx has been shown to be a marker of bone metastases [31] and of bone involvement in multiple myeloma [32]. Beyond determination of vitamin D metabolites, iPTH and markers of bone metabolism our study adds a translational aspect characterizing the association of these markers with the most relevant clinical endpoint, i.e. incidence of fractures.
All reported fractures in our study were diagnosed by radiographs prompted by symptoms. 12.9% of liver recipients suffered from fractures. Previously published data on fracture incidence after liver transplantation show a wide variability [11,[33][34][35][36][37][38]. This imprecision might be caused by differences in diagnostic assessment (radiographs per protocol vs. radiographs triggered by symptoms), different lengths of patient follow up as well as temporal changes in immunosuppression with recent strategies favoring early steroid withdrawal. In our study liver graft recipients suffering from fractures were significantly older than recipients without fractures. This finding is in line with previous studies [11,36].
Similar to liver transplantation, literature on fracture rates after kidney transplantation is characterized by a large variety [39]. Studies with longer observation periods indicated overall a higher fracture rate [40,41]. We recorded fractures in 10% of kidney recipients over a median follow-up of 5.6 years. Consistent with literature most fractures in kidney recipients affected the appendicular skeleton and, in particular, the lower extremities [12,42]. In kidney graft recipients with incident fractures 1, 25-(OH) 2 D levels were significantly lower peri-transplant, this observation waned to a trend at 6 months post-transplant. This finding highlights the importance of the biologically active vitamin D metabolite for bone health. Higher levels of iPTH and a trend of higher CTx were found in kidney recipients without incident fractures 6 months post-transplant, but not at time of transplantation. This observation might indicate that a certain level of bone resorption needs to be maintained for bone health. A longer timespan between kidney transplantation and BMD measurement showed a negative linear relationship with the T-score of the femoral neck. Time-dependent deleterious effects of immunosuppression might be at the cause of this observation. Immunosuppressive agents have been associated with abnormal bone composition. This harmful effect is best established for corticosteroids which have been linked to both, osteoporosis and fractures [43,44]. Of note, in our study all kidney transplant recipients with incident fractures received a steroidcontaining immunosuppressive regimen at least until 6 months post-transplant (vs. 87.3% of patients without incident fractures). The impact of other immunosuppressive agents on bone metabolism is less clear, e. g. conflicting results have been reported for cyclosporine A [45][46][47].
One main strength of our study lies in the longitudinal, uniform measurement of key variables involved in bone metabolism, including vitamin D metabolites, iPTH and the bone turnover markers CTx and P1NP, utilizing samples derived from time of transplant and 6 months after transplantation. The multicenter cohort design with participation of all Swiss transplant centers minimizes the impact of center-specific differences in post-transplant patient care.
Our study has several limitations. First, data on DXA scans and incident fractures were retrospectively collected. Second, BMD measurements were performed at the discretion of the treating physician and not according to a protocol. All analyzed DXA scans were performed post-transplant, different devices were used and scans were done at varying time points post-transplant, but timespan since transplantation is likely a crucial factor for bone loss. Baseline BMD at time of transplantation was not assessed, thus we might have missed preexisting abnormal BMD. Third, radiographs were exclusively taken, if clinically indicated, i.e. suspicion of a fracture. This approach inevitably results in missing asymptomatic fractures.
In conclusion, our data illustrates substantial alterations of bone metabolism markers in the first 6 months after transplantation. Hallmarks were a significant gain in 1, 25-(OH) 2 D combined with a decrease in iPTH, CTx and P1NP after kidney transplantation and an increase in CTx and P1NP after liver transplantation.
Future studies with longitudinal measurement of a comprehensive bone metabolism panel, including CTx and P1NP, combined with serial DXA scans could help to gain a more precise insight into the role of bone turnover markers in the transplant population.
Supporting information S1 Table. Reference values and coefficients of variation for used laboratory assays.