The risk of fractures after kidney transplantation is high. Hyperparathyroidism frequently persists after successful kidney transplantation and contributes to bone loss, but its impact on fracture has not been demonstrated. This longitudinal study was designed to evaluate hyperparathyroidism and its associations with mineral disorders and fractures in the 5 posttransplant years. We retrospectively analyzed 143 consecutive patients who underwent kidney transplantation between August 2004 and April 2006. The biochemical parameters were determined at transplantation and at 3, 12 and 60 months posttransplantation, and fractures were recorded. The median intact parathyroid hormone (PTH) level was 334 ng/L (interquartile 151–642) at the time of transplantation and 123 ng/L (interquartile 75–224) at 3 months. Thirty fractures occurred in 22 patients. The receiver operating characteristic (ROC) curve analysis for PTH at 3 months (area under the ROC curve = 0.711, p = 0.002) showed that a good threshold for predicting fractures was 130 ng/L (sensitivity = 81%, specificity = 57%). In a multivariable analysis, independent risk factors for fracture were PTH >130 ng/L at 3 months (adjusted hazard ratio [AHR] = 7.5, 95% CI 2.18–25.50), and pretransplant osteopenia (AHR = 2.7, 95% CI 1.07–7.26). In summary, this study demonstrates for the first time that persistent hyperparathyroidism is an independent risk factor for fractures after kidney transplantation.
1,25-dihydroxyvitamin D or calcitriol
25-hydroxyvitamin D or calcidiol
- 95% CI
95% confidence interval
adjusted hazard ratio
area under the curve
bone alkaline phosphatase
bone mineral density
body mass index
end-stage renal disease
fibroblast growth factor 23
kidney disease: improving global outcomes
month 12 posttransplantation
month 3 posttransplantation
month 60 posttransplantation
modification of diet in renal disease
polycystic kidney disease
receiver operating characteristic
Kidney transplant recipients have a higher incidence of fractures compared to the general population and compared to patients on dialysis [1-3]. In addition to affecting morbidity, fractures have been associated with decreased recipient survival . The high risk of fractures results from a combination of traditional factors, including immunosuppressive therapy, especially corticosteroids [4, 5], diabetes [4, 6, 7], female sex [4, 6, 8] and older age [2, 4, 7]. Moreover, preexisting osteodystrophy may also be involved, as suggested by the association between the duration of dialysis and the incidence of fractures [2, 7, 9]. After kidney transplantation, persistent hyperparathyroidism (pHPT) is common and contributes to bone loss [10-15]. High levels of serum intact parathyroid hormone (PTH) are associated with an increased risk of fracture in dialysis patients and in those with primary hyperparathyroidism [16, 17]. Nevertheless, the impact of pHPT in renal graft recipients on fractures has not yet been demonstrated.
Although the PTH level drops in the first posttransplant months with the reversal of uremia, it remains high in 17–50% of patients at 1 year [18-20]. The prevalence of pHPT varies largely between studies depending on the PTH assay used, the threshold of PTH chosen to define pHPT, the GFR [20, 21] and the 25-hydroxyvitamin D (25OHD) levels. pHPT can induce other mineral disorders such as hypophosphatemia, hypercalcemia [9, 22] and a high bone turnover state . Hypophosphatemia may also potentiate the fracture risk through osteomalacia . pHPT and hypercalcemia contribute to nephrocalcinosis and are associated with chronic allograft dysfunction [24-26]. The prevalence of these disorders many years after transplantation is not well documented. pHPT is due to a low or incomplete regression of parathyroid gland abnormalities induced by end-stage renal disease (ESRD) , and can be promoted by nonoptimal renal graft function and by native vitamin D insufficiency.
In contrast to the management of hyperparathyroidism in ESRD patients, there are no guidelines for optimal PTH levels after transplantation, and data regarding the clinical outcomes are lacking. The aim of this longitudinal study was to evaluate pHPT in kidney transplant recipients and its association with mineral disorders and fractures in the 5 posttransplant years.
Patients and Methods
We retrospectively analyzed a total of 152 consecutive adult patients who underwent kidney transplantation between August 2004 and April 2006 in the Nephrology-Transplantation Department of Strasbourg University Hospital (France). Nine subjects were excluded because of incomplete data (two patients were followed in another center, five died and two experienced graft losses within 3 months). There were 143 patients analyzed at transplantation, 143 at 1 year and 126 at 5 years (nine graft losses; three patients followed in another center and five deaths during the follow-up; Figure 1).
In the routine pretransplant evaluation, bone mineral density (BMD) was measured using a Hologic Delphi densitometer (Hologic®, Inc., Waltham, MA) at the lumbar spine (L2–L4), femoral neck and total hip in 133 of 143 patients. According to the World Health Organization criteria , osteopenia was defined as a T-score between −1 and −2.5, and osteoporosis was defined as a T-score <−2.5. In patients with pretransplant osteoporosis, our local practices recommended a preventive treatment with bisphosphonates consisting of administration of two intravenous infusions of pamidronate (30 mg each) 15 and 45 days after transplantation.
During the 5 years posttransplant, clinical fractures were registered in the medical records. All of the collected fractures were documented by X-rays or other imaging methods such as bone scintigraphy, bone scan or MRI if needed after clinical rheumatologic assessment. Because spine X-rays were not performed systematically, only clinical vertebral fractures were diagnosed. Deaths, graft losses and immunosuppressive treatments, including corticosteroids and bone-specific therapies during the follow-up period, were recorded.
Considering immunosuppression, patients received an induction therapy (thymoglobulins or basiliximab) associated with a calcineurin inhibitor, mycophenolate mofetil and corticosteroids. At induction, corticosteroids boluses (2 mg × 250 mg) were administered for all but four patients (three diabetics and one patient with porphyria). Prednisone was dosed at 1 mg/kg/day for the first week and then gradually tapered and discontinued between the third and sixth months, except in patients with high immunologic risk or acute rejection (maintained at 0.1 mg/kg/day).
The bone and mineral metabolism biochemical parameters were prospectively determined before the surgical transplant procedure and at 3, 12 and 60 months after transplantation (M3, M12 and M60). After transplantation, all blood samples were collected in the morning in the fasting state. The serum intact PTH level was measured by a radioimmunometric assay (Elecsys®; Roche, Basel, Switzerland; reference range: 15–65 ng/L). Calcemia, phosphatemia, albuminemia and creatininemia were quantified by an autoanalyzer. The corrected calcemia (CCa) was calculated according to the formula: CCa = calcemia + (40 − albuminemia)/40. A 24-h urine collection was analyzed for volume, creatinine, calcium and phosphate. The GFR was estimated by the four-variable equation derived from the Modification of Diet in Renal Disease Study and measured by Iohexol clearance according to the Gaspari method . Radioimmunoassays were used to determine the serum 25OHD (Diasorin, Stillwater, MN) and 1,25-dihydroxyvitamin D levels (1,25OH2D) (Nichols Institute, San Juan Capistrano, CA; reference range: 18–60 ng/L). According to the KDIGO guidelines, vitamin D deficiency was defined as 25OHD <15 µg/L, and insufficiency was defined as 25OHD between 15 and 30 µg/L . Bone turnover was assessed by evaluating the bone alkaline phosphatase (BAP) serum level by an immunoassay (Access Ostase; Beckman, Chaska, MN; reference range: 2.2–14.5 µg/L). Similarly, the C-telopeptide serum level was determined using an immunoassay (Elecsys; reference: <0.54 µg/L). pHPT was defined as a PTH higher than twice the upper normal limit of the assay (>130 ng/L), hypercalcemia was defined as calcemia ≥2.60 mmol/L (≥10.4 mg/dL) and hypophosphatemia was defined as phosphatemia <0.80 mmol/L (<2.5 mg/dL).
The analyses were performed using the SPSS 11.0 software (SPSS, Inc., Chicago, IL). The data are given as the mean ± standard deviation, the median [interquartile range] or the percentages for parametric, nonparametric parameters and categorical variables, respectively. Correlations were assessed by Spearman's test. Variables between groups were compared using Student's t-test, Mann–Whitney U-test, and chi-square test, as appropriate. Receiver operating characteristic (ROC) curves were plotted for D0 and M3 PTH levels and fractures occurrence, censored for fractures occurring before 3 months. Kaplan–Meier analyses were used to evaluate the time to fracture and to identify the factors associated with fractures using the log-rank test. The date of fracture onset was assumed to be the first date of a documented fracture with patients censored at death, loss to follow up, dialysis or the end of the follow-up period. Potential risk factors for fractures that showed an association with a p ≤ 0.20 in the univariate analysis were examined as covariates in multivariable Cox regression models (covariates: female sex, pretransplant osteopenia, albuminemia <40 g/L at M3, corticosteroids use at M12, and M3 PTH >130 ng/L or M3 PTH >300 ng/L) for patients with an available pretransplant BMD result (n = 133). Two-tailed p < 0.05 were considered statistically significant.
Population characteristics and treatments
The patient characteristics and treatments modalities at the time of transplantation are described in Tables 1 and 2. Three of the study participants received multiple transplants. Most of the patients were Caucasian (n = 136). At 1 year, 42% of patients received prednisone (n = 61; mean dose of 6.4 mg per day), 55% received cyclosporine A (n = 79), 36% received tacrolimus (n = 52) and 8% received mTOR inhibitors (n = 12). Bisphosphonates were administered to 14 subjects (9.8%) during the first year, and 18 subjects (15%) at 5 years. Five patients required parathyroidectomy after transplantation for severe HPT with hypercalcemia. Cinacalcet was administered to 13 patients (9%) in the first year and seven patients at 5 years. Native vitamin D was administered to 54 subjects (38%) in the first year, and 92 (74%) at 5 years. Calcium supplements were prescribed for 26 patients (18%) at 1 year and 27 (21%) at 5 years. Phosphate supplementation was given to 117 patients (82%) during the first month, 13 (16%) at 1 year and 12 (9.5%) at 5 years.
|Mean ± SD or %|
|Age at transplantation (years)||47.4 ± 12.2 (22–74)|
|Female sex (%)||41|
|BMI (kg/m2)||24.9 ± 4.4 (16–38)|
|Transplantation rank ≥2 (%)||18.2|
|Deceased donor (%)||89|
|HD/PD/preemptive transplantation (%)||79/17/4|
|Dialysis duration (months)||24 [14–44]|
|Primary kidney disease (%)|
|Interstitial nephropathy (including PCKD)||28.7|
|Other or unknown||6.3|
|Pretransplant BMD, n = 133 (%)|
|Previous corticotherapy (%)||31.5|
|Pretransplant parathyroidectomy (%)||12.1|
|PTH (ng/L)||334 [151–642]|
|25OH-vitamin D, n = 139 (µg/L)||14.3 ± 11.6 (<5–69)|
|25OH-vitamin D <15 µg/L, n = 139 (%)||64.0|
|N (% patients)|
|Sevelamer, n = 104||62 (59.6%)|
|Calcium binders, n = 104||52 (50%)|
|Active vitamin D, n = 104||29 (27.9%)|
|Native vitamin D, n = 104||11 (10.6%)|
|Cinacalcet, n = 138||6 (4.3%)|
Mineral and bone disorders in the 5-year posttransplant period
The prevalence rates of pHPT, hypercalcemia and hypophosphatemia during the follow-up period are displayed in Table 3. pHPT (PTH >130 ng/L) was present in 48% of patients at M3, 41% at M12 and 34% at M60. Moreover, 74% of the patients who showed pHPT at M60 were already diagnosed with pHPT at M3. In patients with GFR >60 mL/min at M60 (n = 33), 15 patients (42%) had a PTH level above the upper normal limit (>65 ng/L), and 4 (12%) had PTH >130 ng/L. Among these 15 recipients, one was 25OHD deficient, and five were 25OHD insufficient. Vitamin D deficiency dropped from 56% at M12 to 15% at M60. The levels of bone formation and resorption markers were high in the first year and then decreased (Table 3). PTH was positively correlated to BAP at M3 (r = 0.58, p < 0.001), M12 (r = 0.47, p < 0.001), and M60 (r = 0.29; p = 0.001), as well as to C-telopeptide levels at M12 (r = 0.43, p < 0.001) and M60 (r = 0.41, p < 0.001).
|Variables||At transplantation (n = 143)||M3 (n = 143)||M12 (n = 143)||M60 (n = 126)|
|PTH (ng/L)||334 [151–642]||123 [75–224]||117 [71–181]||85 [52–152]|
|PTH >130 ng/L (%)||77.5||48.2||41.3||34.1|
|GFR, Iohexol clearance (mL/min)||nd||56.4 ± 14.8 (n = 125)||56.6 ± 17.3 (n = 124)||50.9 ± 16.7 (n = 112)|
|Patients with GFR >60 mL/min||n = 48||n = 57||n = 33|
|Patients with GFR >60 mL/min and PTH >130 ng/L||19 (40%)||22 (39%)||4 (12%)|
|Patients with GFR >60 mL/min and PTH >65 ng/L||39 (81%)||41 (72%)||15 (42%)|
|Calcemia (mmol/L)||2.28 ± 0.23||2.41 ± 0.18||2.35 ± 0.18||2.36 ± 0.17|
|Corrected calcemia (mmol/L)||2.27 ± 0.25||2.34 ± 0.19||2.26 ± 0.18||2.29 ± 0.14|
|Calcemia ≥2.60 mmol/L (%)||11.2||13.3||8.4||7.9|
|Phosphatemia (mmol/L)||1.62 ± 0.51||0.93 ± 0.21||1.01 ± 0.22||1.09 ± 0.28|
|Phosphatemia <0.80 mmol/L (%)||1.4||22.4||16.0||14.3|
|25OH-vitamin D (µg/L)||14.3 ± 11.6 (n = 139)||17.3 ± 12.5 (n = 122)||15.7 ± 9.5 (n = 141)||28.8 ± 12.4 (n = 124)|
|25OH-vitamin D < 15 µg/L (%)||64.3||53.6||55.6||15.0|
|1,25OH2-vitamin D (ng/L)||9.3 ± 8.9 (n = 135)||37.1 ± 23.1 (n = 123)||43.1 ± 26.2 (n = 135)||54.4 ± 21.3 (n = 123)|
|Bone alkaline phosphatase (µg/L)||nd||21 [13–40] (n = 62)||23 [14–35] (n = 139)||11 [9–16] (n = 125)|
|C-telopeptide (µg/L)||nd||nd||1.0 [0.6–1.7] (n = 139)||0.6 [0.3–0.9] (n = 125)|
Characteristics of patients with pHPT
Forty-eight percent of renal transplant recipients displayed pHPT at 3 months (Table 4, Figure 2). This group had a higher body mass index (BMI), a longer dialysis duration, higher calcemia, lower phosphatemia, higher bone turnover marker levels, a lower 25OHD level, and a lower GFR. There was a nonsignificant trend for this group to be older. No differences were observed between pHPT and non-pHPT patients regarding gender, pretransplant diabetes, pretransplant BMD, albuminemia or corticosteroid use at M1, M12 and M60. Survival was decreased in pHPT patients (5 deaths vs. 0, p = 0.006), and graft survival did not differ significantly (5 graft losses vs. 4).
|M3 PTH > 130 ng/L (n = 69)||M3 PTH ≤ 130 ng/L (n = 74)||p|
|Age (years)||49.3 ± 12.0||45.6 ± 12.2||0.069|
|BMI (kg/m2)||25.9 ± 4.4||23.9 ± 4.1||0.004|
|Dialysis duration (months)||30 [18–48]||18 [10–38]||0.001|
|Pretransplant osteoporosis||13 (20.6%)||17 (24.3%)||0.465|
|Pretransplant osteopenia||29 (46.0%)||29 (41.4%)||0.643|
|D0 PTH (ng/L)||577 [353–800]||177 [79–332]||<0.001|
|D0 calcemia (mmol/L)||2.31 ± 0.23||2.26 ± 0.23||0.279|
|D0 phosphatemia (mmol/L)||1.69 ± 0.51||1.55 ± 0.50||0.110|
|D0 25OH-vitamin D (µg/L)||13.5 ± 10.4||15.1 ± 12.7||0.422|
|D0 25OH-vitamin D <30 µg/L||95.7%||82.4%||0.032|
|M3 calcemia (mmol/L)||2.46 ± 0.20||2.36 ± 0.15||<0.001|
|M3 phosphatemia (mmol/L)||0.90 ± 0.24||0.96 ± 0.18||0.097|
|M3 calciuria (mmol/24 h, N = 2.5–7.5)||3.46 ± 3.19||2.48 ± 1.54||0.028|
|M3 phosphaturia (mmol/24 h, N = 13–42)||38.6 ± 19.0||31.5 ± 16.0||0.027|
|M3 25OH-vitamin D (µg/L)||14.9 ± 10.5||19.4 ± 13.7||0.047|
|M3 25OH-vitamin D < 30 µg/L||76.8%||68.9%||0.024|
|M3 1,25OH2-vitamin D||43 ± 28||32 ± 16||0.008|
|M3 bone alkaline phosphatase (µg/L)||34 [21–60]||15 [11–19]||<0.001|
|M3 MDRD GFR (mL/min/1.73 m2)||46.8 ± 14||52.4 ± 14||0.018|
|M12 PTH (ng/L)||172 [119–271]||75 [58–117]||<0.001|
|M12 calcemia (mmol/L)||2.39 ± 0.21||2.31 ± 0.14||0.013|
|M12 phosphatemia (mmol/L)||0.99 ± 0.23||1.03 ± 0.20||0.314|
|M12 bone alkaline phosphatase (µg/L)||26 [18–40]||18 [12–29]||0.001|
|M12 C-telopeptide (µg/L)||1.3 [0.8–1.9]||0.8 [0.5–1.5]||0.004|
|M12 MDRD GFR (mL/min/1.73 m2)||45.4 ± 14.8||52.9 ± 15.2||0.003|
|M60 PTH (ng/L)||134 [80–211]||65 [42–97]||<0.001|
|M60 calcemia (mmol/L)||2.39 ± 0.24||2.33 ± 0.16||0.078|
|M60 phosphatemia (mmol/L)||1.07 ± 0.38||1.10 ± 0.23||0.452|
|M60 bone alkaline phosphatase (µg/L)||13 [9–18]||10 [8–14]||0.058|
|M60 C-telopeptide (µg/L)||0.7 [0.3–1.0]||0.6 [0.3–0.8]||0.205|
|M60 MDRD GFR (mL/min/1.73 m2)||44.9 ± 16.9||53.6 ± 20.7||0.014|
Fractures and risk factors
Thirty clinical fractures were registered in 22 patients during the 5-year posttransplant period. Six patients presented several fractures. The cumulative incidence of fractures was 7.7% at 1 year, 12.0% at 2 years and 14.9% at 5 years. The mean delay to the first fracture was 16.4 ± 15.4 months (range: 1–58 months). Twelve fractures required hospitalization. Twenty-four were nontraumatic. The fracture sites included the spine (10 fractures), foot (6), forearm (6), leg (4), rib (2), hand (1) and hip (1). Patients with fractures were characterized by higher PTH levels, as well as a higher rates of hypercalcemia and hypophosphatemia (Table 5, Figure 3). Estimated and measured GFR were equivalent at M3, M12 and M60. The patient and graft survival rates did not differ significantly (5 deaths in the fracture group vs. 0 and 6 graft losses vs. 3). Corticosteroids treatment did not differ significantly between groups (Table 5). During the follow-up, 14 patients with fractures received biphosphonates (given both before and after the occurrence of the fracture in three patients and only after the fracture in 11 patients). Three patients with fractures had been treated by parathyroidectomy (performed after the occurrence of the fracture). These three patients had severe HPT complicated by hypercalcemia and were previously treated with cinacalcet. Two other patients with fractures had been treated with cinacalcet after the occurrence of the fracture.
|Fracture (n = 22)||No fracture (n = 121)||p|
|Age (years)||49 ± 12||47 ± 12||0.648|
|BMI (kg/m2)||25.2 ± 4.1||24.8 ± 4.4||0.700|
|Female gender (%)||54.5||38.8||0.172|
|Dialysis duration (months)||35 [17–60]||22 [13–38]||0.137|
|Pretransplant osteoporosis (n, %)||4 (19.0%)||26 (23.2%)||0.183|
|Pretransplant osteopenia (n, %)||13 (61.9%)||45 (40.1%)||0.162|
|Pretransplant diabetes (n, %)||5 (23%)||17 (14%)||0.321|
|Pretransplant steroids use (n, %)||8 (36.4%)||38 (31.4%)||0.530|
|Corticosteroids use at M1 and M3 (n, %)||22 (100%)||117 (96.7%)||0.244|
|M1 steroids dose (mg per day)||18.9 ± 4||17.4 ± 6||0.238|
|Corticosteroids use at M12 (n, %)||12 (54.5%)||49 (40.5%)||0.196|
|Corticosteroids use at M60 (n, %)||6/18 (33.3%)||34/108 (31.5%)||0.816|
|Tacrolimus at M12 (n, %)||11 (50%)||41 (34%)||0.155|
|Cyclosporine at M12 (n, %)||10 (45%)||69 (57%)||0.317|
|D0 PTH (ng/L)||480 [271–800]||311 [129–605]||0.026|
|D0 calcemia (mmol/L)||2.32 ± 0.21||2.28 ± 0.24||0.400|
|D0 phosphatemia (mmol/L)||1.81 ± 0.54||1.58 ± 0.48||0.048|
|D0 25OH-vitamin D (µg/L)||16.0 ± 15||14.0 ± 11||0.474|
|M3 PTH (ng/L)||199 [153–451]||110 [70–189]||0.001|
|M3 calcemia (mmol/L)||2.47 ± 0.23||2.40 ± 0.17||0.070|
|M3 calcemia ≥2.6 mmol/L||6 (27.3%)||13 (10.7%)||0.054|
|M3 phosphatemia (mmol/L)||0.83 ± 0.19||0.95 ± 0.21||0.014|
|M3 phosphatemia <0.8 mmol/L||11 (50%)||27 (22.3%)||0.010|
|M3 25OH-vitamin D (µg/L)||19.1 ± 14.8||17.0 ± 12.2||0.539|
|M3 bone alkaline phosphatase (µg/L)||16 [12–62]||21 [14–37]||1|
|M3 albuminemia (g/L)||41.8 ± 2.3||43.0 ± 1.10||0.056|
|M3 MDRD GFR (mL/min/1.73 m2)||51 ± 13||49 ± 15||0.576|
|M12 PTH (ng/L)||142 [104–221]||103 [64–171]||0.011|
|M12 calcemia (mmol/L)||2.46 ± 0.17||2.33 ± 0.17||0.001|
|M12 phosphatemia (mmol/L)||1.00 ± 0.21||1.01 ± 0.22||0.783|
|M12 bone alkaline phosphatase (µg/L)||24 [18–31]||23 [14–35]||0.808|
|M12 C-telopeptide (µg/L)||1.3 [0.7–2.0]||1.0 [0.6–1.6]||0.578|
|M60 PTH (ng/L)||127 [75–236]||84 [47–142]||0.045|
|M60 calcemia (mmol/L)||2.46 ± 0.19||2.34 ± 0.19||0.004|
|M60 phosphatemia (mmol/L)||1.02 ± 0.36||1.10 ± 0.26||0.227|
To document the relationship between PTH level and fractures, we constructed ROC curves (Figure 4). At transplantation, the area under the curve (AUC) was 0.65 (p = 0.026), and the best PTH threshold for predicting fracture occurrence was 285 ng/L (sensitivity = 73%, specificity = 53%). At 3 months, AUC was 0.711 (p = 0.002), and the best PTH threshold was 130 ng/L (sensitivity 81%, specificity 57%). At this time, a threshold of 300 ng/L showed a high specificity (86%) but a lower sensitivity of 38%.
Kaplan–Meier analyses showed that the factors significantly associated with the occurrence of fractures were: PTH > 130 ng/L at M3 (p < 0.001, Figure 5) and at M12 (p = 0.017); PTH > 300 ng/L at M3 (p = 0.001); hypercalcemia at M3 (p = 0.021), at M12 (p = 0.056) and at M60 (p < 0.001); hypophosphatemia at M3 (p = 0.006) and at M60 (p = 0.042); and albuminemia <40 g/L at M3 (p = 0.042). There was a nonsignificant trend for female sex (p = 0.194), dialysis duration >3 years (p = 0.097), pretransplant osteopenia (p = 0.144), PTH >285 ng/L at transplantation (p = 0.076), and the use of corticosteroids (p = 0.200) and tacrolimus at M12 (p = 0.16). Fracture occurrence was not statistically associated with pretransplant diabetes (p = 0.219), age, primary kidney disease, transplantation rank, donor type, corticosteroid use before transplantation, recipient BMI, pretransplant osteoporosis, 25OHD deficiency at transplantation, M3 and M12, and bone marker levels at M3 and M12.
In multivariable analyses, a higher risk of fracture was independently associated with pHPT at M3 (adjusted hazard ratio [AHR] = 7.46, 95% CI 2.18–25.51) and pretransplant osteopenia (AHR = 2.79, 95% CI 1.07–7.26; Table 6). Patients with PTH > 300 ng/L at M3 had a fourfold greater risk of fractures compared to patients with PTH < 300 ng/L (95% CI 1.6–9.1; Table 6).
|PTH at 3 months >130 ng/L (reference: PTH ≤130 ng/L)||7.462||2.183–25.509||0.001|
|Albuminemia at 3 months <40 g/L||2.398||0.933–6.166||0.070|
|Corticosteroids use at 12 months||0.443||0.173–1.134||0.089|
|PTH at 3 months >300 ng/L (reference: PTH ≤300 ng/L)||3.804||1.592–9.090||0.03|
|Albuminemia at 3 months <40 g/L||2.398||0.910–6.319||0.077|
|Corticosteroids use at 12 months||0.535||0.218–1.315||0.068|
Our series is the first longitudinal study to evaluate the impact of pHPT and mineral disorders on fractures after kidney transplantation. pHPT is common after kidney transplantation with 1-year rates varying among studies from 17% to 50% [18-20]. Its definition differs by author, and the optimal PTH level is not yet defined in renal graft recipients. In our study, pHPT was defined as an intact PTH level higher than twice the upper normal limit of the assay (130 ng/L). Our results consolidate this definition by the associated clinical outcome of fracture occurrence. pHPT was present in almost half of our patients at 3 months and in 41% at 1 year. In a recent study, the prevalence of pHPT with the same definition was lower at 1 year, at 24% . This difference could be explained by the higher PTH levels in our patients at the time of transplantation, as our population was not selected and was therefore more representative of a patient population in a clinical setting. The use of phosphate supplementations may also have led to an increase of PTH levels . However, the available evidence in the field is controversial; indeed, a study has suggested a potential muscular benefit of this supplementation without affecting the evolution of PTH levels . As previously shown, we observed that pHPT was associated with a longer dialysis duration , a high PTH level before transplantation [20, 32, 33], renal insufficiency [20, 21, 32, 33] and calcidiol deficiency [21, 34]. Patients with pHPT were also more likely to have a high BMI, which is consistent with recent studies on renal graft recipients , dialysis patients  and obese patients . Additionally, obesity may stimulate PTH secretion via leptin . After more than 1 year posttransplant, the persistence of HPT is less documented. In our study, one-third of patients presented a PTH level higher than twice the normal level at 5 years. Some authors speculate that the long-term PTH levels in patients with normal renal function should approach normal values to restore normal bone turnover . In our study, 42% of patients with normal renal function had PTH levels higher than normal (>65 ng/L) at 5 years posttransplant, although the majority of them had a calcidiol level in the target range. These data suggest an incomplete recovery from the parathyroid gland abnormalities induced by ESRD over the long term in a large subset of patients.
Mineral abnormalities induced by pHPT, such as hypercalcemia and hypophosphatemia, were also associated with fracture occurrence. Hypophosphatemia due to a reduced tubular reabsorption of phosphate was still observed in 14% of patients 5 years after transplantation. Renal phosphate wasting may also be a consequence of the persistence of excessive fibroblast growth factor 23 (FGF23) production by bone cells . In addition to affecting bone via hypophosphatemia, elevated FGF23 secretion may directly impact bone and inhibit mineralization .
Fractures occurred in 8% of our cohort in the first year and in 15% in 5 years, which is consistent with recent studies reporting 5-year fracture incidences of 12%  and 22% . We observed that patients with pHPT at 3 months had a greater risk of fracture in the 5 years posttransplant. Although several studies found that pHPT negatively impacts bone loss [10-15], fracture risk has been rarely evaluated. In a cross-sectional study of 125 patients on average 4 years after kidney transplantation, Giannini et al  showed that an elevated PTH level was related to an increased risk of vertebral fractures detected by systematic radiography. The contribution of hyperparathyroidism to fracture risk was also recently described in a large cohort of patients on hemodialysis , where PTH levels above 900 pg/mL were associated with an increased risk of any new hip fracture. This association was also reported in primary hyperparathyroidism . By contrast, hypoparathyroidism has been also associated with bone loss after renal transplantation [39, 40] as well as with fractures in patients on dialysis . In our study, no fracture occurred in subjects with PTH <130 ng/L at the time of transplantation; however, this result should be interpreted with caution because of the small sample size of this subgroup.
The impact of pHPT on bone may be mediated by the stimulation of bone turnover by PTH, as illustrated by the positive correlation between posttransplant PTH and bone turnover markers levels in our cohort and in other studies [15, 23]. Nevertheless, we did not find that fracture occurrence was associated with higher bone marker levels. Bone abnormalities induced by pHPT may be explained by the removal of skeletal resistance to PTH after kidney transplantation. In chronic kidney disease, bone becomes hyporesponsive to PTH. Many factors, such as phosphate loading, decreased calcitriol, antagonistic PTH fragments, the downregulation of PTH receptors and decreased pulsatility of PTH, lead to bone resistance to PTH . After transplantation, most of these factors decrease, most likely inducing an important bone response to PTH. In addition, in vitro studies have demonstrated that the bone response to PTH is enhanced by glucocorticoid treatment . The use of steroids in the presence of high PTH levels, particularly in the immediate posttransplant period, appears to have a cumulative deleterious effect on bone and may explain the early bone loss described in renal graft recipients.
An elevated risk of fracture was also observed for patients with pretransplant osteopenia. This finding is in accordance with a recent study showing that low BMD conferred an increased risk of fracture after kidney transplantation . Nevertheless, we were surprised to find no association between fractures and osteoporosis. This could be explained by different management protocols for recipients with osteoporosis, including the limited use of steroids, the use of calcium and vitamin D and preventive treatment with bisphosphonates. Indeed, some trials have shown a protective effect of bisphosphonates on early bone loss after transplantation [45, 46]. This study was not specifically designed to address the impact of pretransplant BMD; nevertheless, we found interesting to assess it in the routine pretransplant evaluation, to select patients in whom preventive treatment in the immediate posttransplant weeks can be beneficial. Larger randomized studies are required to evaluate the usefulness of bisphosphonates for the prevention of bone loss and fractures in selected patients, particularly in those with bone loss, pHPT and high levels of bone turnover markers. The KDIGO guidelines recommend caution when using bisphosphonates because fracture data are limited and an associated risk of adynamic disease has been described.
The present results also suggest that hypoalbuminemia was a risk factor for fractures. Hypoalbuminemia was associated with bone loss in a study of renal graft recipients  and with hip fracture in dialysis patients , suggesting that malnutrition and sarcopenia play roles in the pathogenesis of fracture. We also noted that female sex and pretransplant diabetes were associated with fractures, but these relationships were not significant, although this was most likely due to the small sample size. Other studies have not confirmed traditional risk factors, such as older age [2, 3, 6, 7] or diabetes . The use of corticosteroids is a recognized risk factor for fractures . Patients treated with prednisone at 1 year were more likely to develop fractures. This association was not significant, possibly because the dose of prednisone was not sufficiently different between subjects with or without steroids at 1 year, particularly in the first posttransplant months. In addition, all deaths occurred in the group of patients with pHPT (p = 0.006). This finding can be explained by the lower GFR, the longer dialysis duration and the trends of older age in this group. Although dialysis patients with hyperparathyroidism  and renal transplant recipients with hypercalcemia  have a greater risk of mortality, it is not clear whether hyperparathyroidism can be linked to an increased risk of death for transplant recipients. This result must be further evaluated in larger studies.
This study has some limitations due to its observational design. The timing between pretransplant BMD assessment and transplantation may have varied among the study participants (from several months to 3 years). Bone-specific therapies were administered in some patients. Corticosteroid sparing was most likely performed in some subjects, particularly in diabetic and osteoporotic patients. Another caveat is the lack of information on the cumulative corticosteroid dose for each patient; however, steroid exposure at specific time points did not differ significantly between groups. In addition, we may have missed some asymptomatic fractures particularly vertebral fractures. Nonetheless, this study provides new and longitudinal data on PTH levels after kidney transplantation, which are typically missing from large registry studies on fractures. A large and multicenter study controlling for treatments influencing mineral and bone disorders is warranted to corroborate our data.
With the aim to reduce fracture risk after kidney transplantation, this study identified a new risk factor represented by pHPT, which may better select recipients who could benefit from treatment. Better control of hyperparathyroidism after kidney transplantation should be considered in some patients using efficient treatments, such as calcium supplementation , native and active vitamin D, cinacalcet and parathyroidectomy. Previous studies suggest a bone protective effect of cinacalcet in dialysis patients. An analysis of four randomized studies showed that cinacalcet reduced the fracture risk in dialysis patients with intact PTH >300 ng/L . However, the recent EVOLVE trial did not support the protective effect of cinacalcet against the development of fractures . In any case, it should be noted that the intention-to-treat analysis of this trial had significant limitations (e.g. the results were unadjusted for baseline characteristics such as age and there was a high rate of treatment crossover). Moreover, it is noteworthy that in the prespecified analysis with lag censoring, there was a significant reduction in the risk of fractures among cinacalcet-treated patients . In two small studies of kidney transplantation, treatment with cinacalcet improved BMD in subjects with hypercalcemia secondary to pHPT [51, 52].
In conclusion, this study is the first to demonstrate that pHPT is an independent risk factor for fractures after kidney transplantation. This new risk factor has a high prevalence, particularly in the first year when the fracture rate is the highest. Better control of hyperparathyroidism before and after kidney transplantation should be beneficial. Randomized studies are needed to determine whether cinacalcet reduces the occurrence of fractures in recipients with pHPT.
We would like to thank Raphael Porcher for his help in statistical methods.
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.