Prophylactic Bisphosphonate Treatment Prevents Bone Fractures After Liver Transplantation


* Corresponding author: Martin Bodingbauer,


A randomized controlled prospective open-label single center trial was performed. At the time of transplantation patients were randomly assigned to one of two treatment arms: The study group of 47 patients received zoledronic acid (ZOL, 8 infusions at 4 mg during the first 12 months after LT), calcium (1000 mg/d) and vitamin D (800 IE/d). The control group consisted of 49 patients who received calcium and vitamin D at same doses (CON). The incidence of bone fractures or death was predefined as the primary endpoint. Secondary endpoints included bone mineral density (BMD), serum biochemical markers of bone metabolism, parameters of trabecular bone histomorphometry and mineralization density distribution (BMDD). Patients were followed up for 24 months. Analysis was performed on an intention-to-treat basis. The primary endpoint fracture or death was reached in 26% of patients in the ZOL group and 46% in the CON group (p = 0.047, log rank test). Densitometry results were different between the groups at the femoral neck at 6 months after LT (mean+/-SD BMD ZOL: 0.80 ± 0.19 g/cm2 vs. CON: 0.73 ± 0.14 g/cm2, p = 0.036). Mixed linear models of biochemical bone markers showed less increase of osteocalcin in the ZOL group and histomorphometry and BMDD indicated a reduction in bone turnover. Prophylactic treatment with the bisphosphonate zoledronic acid reduces bone turnover and fractures after liver transplantation.


Liver transplantation (LT) has become a successful treatment option for patients with end-stage liver disease. Surgical techniques and clinical follow-up of this patient group have improved dramatically over the past 10 years and have resulted in a year-to-year increase in survival rates. Nevertheless, LT is still associated with substantial morbidity, with metabolic bone disease with its related increase in bone fractures playing a major role (1). Importantly, such fractures are associated not only with high morbidity but also with increased mortality (2). Bone disease is characterized by reduced bone mass and disruption of bone architecture, leading to decreased bone strength and an increase in the risk of fragility and fractures particularly at the wrist, hip and spine. Some of the fractures are asymptomatic; others may cause severe pain and result in spinal deformity, loss of height and significant functional impairment. The etiology of hepatic osteodystrophy remains largely undefined and is probably multifactorial (3,4). The incidence of documented fracture rates after transplantation has been reported in a number of studies to be as high as 25–35% (5, 6). Most of the fractures occur within the first 12 months after transplantation, but fractures have also been observed with substantial frequency thereafter (2).

Antiresorptive agents have been evaluated for the prevention of bone loss following LT only to a limited degree. Many of these studies have been small and most have not been randomized or controlled (7). Beneficial effects of pamidronate on indirect parameters, such as bone mineral density (BMD), have been shown in patients undergoing heart, liver, lung and kidney transplantation (8–12). One investigator suggested that aledronate is efficacious in preventing the natural course of bone loss associated with LT (13). However, measurements of BMD at the lumbar spine do not accurately predict the risk of fractures after LT (14). So far, no randomized, controlled study on the effect of bisphosphonate therapy for preventing fractures in LT recipients after transplantation has been published.

The clinical importance of fractures after LT and the lack of reliable surrogate parameters predicting fracture-risk led us to design a prospective study with the rate of fractures as the primary endpoint. We chose bisphosphonate zoledronic acid for prophylaxis because of its high potency, which was just recently introduced for treating Paget's disease, and because it is being evaluated for intermittently (once yearly) treating postmenopausal osteoporosis (15–19). We investigated whether using zoledronic acid as a prophylactic treatment prevents bone fractures after LT.

Patients and Methods

Study design

A randomized controlled prospective open-label single centre trial was carried out (registration number NCT00302484). We randomly assigned patients to one of two treatment groups at the time of transplantation: the bisphosphonate group (ZOL, 47 patients) received zoledronic acid, (8 infusions of 4 mg, monthly for the first half of the year, then 9 and 12 months after LT), calcium carbonate (1000 mg/d orally) and vitamin D3 (800 IE/d orally). The control group (CON, 49 patients) received calcium and vitamin D with the same dosage as that of the ZOL group. In both groups treatment was administered for the first 12 months after transplantation. The study groups were stratified according to age (>65 years) and sex. The randomization was done by a separate unit within the Department of Surgery, according to a randomization sequence generated by SAS (Cary, NC). Novartis provided the drug, but had no influence on the design, direction, analysis or preparation of the manuscript, and provided no financial support. The investigators analyzed the data.

The primary study endpoint was defined as the rate of the patients' first bone fracture within 24 months after LT. In addition, we analyzed a combined endpoint, the, rate of the first bone fracture and death among the patients, in order to uncover a potential negative effect of treatment on survival. On the basis of the previous data, we assumed an expected fracture rate of approximately 50% within 12 months after LT in the CON group (5). Thus this trial was powered at 80% to detect a risk reduction of at least 50% given a two-sided p value < 0.05. Secondary endpoints included: BMD of the femoral neck and the lumbar spine evaluated by dual energy X-ray absorptiometry and serum biochemical markers of bone metabolism (determined before and 3, 6 and 12 months after LT). Additionally, histomorphometry according to Parfitt and bone mineralization density (BMDD) in transilical bone biopsies were measured by quantitative backscattered electron imaging (d and 6 months after LT) (20). Patients were followed up for 24 months, and no patient was lost to follow-up. Analysis was done on an intention-to-treat basis.

The study was approved by the local institutional review board and written informed consent was provided by all patients.


Between January 2002 and December 2005, 269 patients underwent orthotopic liver transplantation in our centre (Vienna General Hospital). Only adults (>18 years) receiving a first transplant were eligible. There was no selection. The reasons why not all patients receiving a liver transplant participated in the our study were as follows: participation in other clinical trials in our centre, refusal of informed consent, predefined exclusion criteria (second liver transplantation, acute liver failure as indication for LT, age under 18 years) and logistical problems related to inclusion during off-hours. Of the 269 patients who underwent LT, 96 patients were randomly selected into two groups: 47 in the study group (ZOL) and 49 in the control group (CON) (Figure 1).

Figure 1.

Study profile due to the ITT analysis; patients who died after withdrawal (w+d) were counted twice in this study flow chart.


All patients received induction therapy with rabbit ATG (thymoglobuline), 2.5 mg/kg body weight for 3 days. Initially we used cyclosporine A-microemulsion (Neoral) for immunosuppression, and only in the case of relevant side effects (hypercholesterin, hypertriglyceridaemia, bad resorption of cyclosporine A), the patients were given tacrolimus instead. For maintenance immunosuppression cyclosporine was administered to 61 patients (63.6%) and tacrolimus to 35 patients (36.4%). There was no difference between the two groups in the numbers (ZOL 17/47 and CON 18/49) of patients who switched drugs (p = 0.95). All patients received corticosteroids, which were gradually tapered (40 mg of dexamethasone on the day of transplantation, tapered to 4 mg at day 5 and then substituted with 20 mg of oral methylprednisolone per day), which were discontinued within 3 months after LT in most of the patients.


All randomized patients were evaluated with X-ray for bone fractures (hip, pelvis and thoracic and lumbar part of the vertebral column) before LT and subsequently at 6 and 12 months after LT. Bone fractures were counted if either clinically or radiologically evident or if a decrease in any vertebral height of >20% (or ≥4 mm) occurred within 12 months after transplantation (21). The evaluation of fractures was done by a radiologist blinded to the treatment regimen.

Bone mineral density

All randomized patients were evaluated with BMD before LT and subsequently at 6 and 12 months after LT. Bone density was measured with dual-energy X-ray absorptiometry and with a QDR-4500 scanner (Hologic, Waltham, MA), using the manufacturer's recommended standard procedures for the posterior–anterior lumbar spine at L1-L4, and for the proximal femur at the femoral neck, trochanter, intertrochanterica region, total region, and Ward's triangle. The projectional BMD values were given (in g/cm2).

Biochemical measurements

Blood samples for biochemical bone markers were taken before LT and at 1 week, 3, 6, 9 and 12 months after. Osteoprotegerin (OPG) was measured by a custom sandwich enzyme-linked immunoassay. Microtiter plates were coated with monoclonal antibody against OPG, and the detection antibody was an affinity-purified goat anti-OPG conjugated to biotin (both antibodies from R&D Systems, Minneapolis, MN). The standard recombinant OPG that we used was from RDI (Flanders, NJ). Bone alkaline phosphatase (ALP) was measured by immunosorbent enzyme-linked assay; osteocalcin (OC), intact parathyroid hormone (iPTH) and type I collagen peptides cross-laps (CTx), by electrochemiluminescence (Elecsys; Roche, Reinach, Switzerland); and 1.25 (OH)2 D3 by a radioimmunoassay (Diasorin, Stillwater, MN). All tests were carried out according to the manufacturer's instructions in a routine laboratory.

Bone biopsies

Transiliacal bone biopsies were obtained during and 6 months after liver transplantation, as described by Bordier et al. (22). Bone biopsies were done after 6 months and after six infusions of zoledronic acid where given to the study group patients to detect changes in early bone turnover (inhibition) in the most vulnerable phase for fractures after transplantation. Histomorphometry was done on biopsies from 26 CON and 19 ZOL patients and BMDD measurements were done on biopsies from 26 CON and 21 ZOL patients. The biopsies were fixed in 70% ethanol, dehydrated in ethanol prior to embedding them in polymethylmethacrylate (PMMA). Two-dimensional histomorphometric analysis was done on 3-μm thick Goldner's Trichrome stained sections according to the guidelines of the ASBMR nomenclature committee (20). The following parameters were evaluated: trabecular bone volume (BV/TV), trabecular thickness (Tb.Th.), trabecular number (Tb.N.) osteoid volume (OV/BV), osteoid thickness (O.Th.), osteoid surface (OS/BS), osteoblast surface (Ob,S/BS), osteoclast surface, (Oc.S/BS) and eroded surface (ES/BS). Mineralizing surface (MS/BS) or mineral apposition rate (MAR), the primary data of dynamic bone formation could not be evaluated because the patients were not given any tetracycline labelling. The BMDD was determined in the trabecular bone by quantitative backscattered electron imaging as described previously (qBEI) (23).

Adverse events

At each visit we documented the medication used, side effects of the study drugs, transplantation procedure and any other adverse events. If the serum calcium level was >2.7 mmo/L, calcium supplementation was reduced to half of the dose until the hypercalcemia normalized. Before zoledronic acid treatment, the serum creatinine level of the patients were checked. If the serum creatinine concentration was >2.5 mg/dL, the bisphosphonate infusion was delayed until the creatinine concentration decreased below 2.0 mg/dL after appropriate medical management. This was the case for the first post-transplant infusion in four patients.

Statistical analysis

Differences in the primary end point were evaluated by the Kaplan-Meier method and the log-rank test respectively. Comparisons of continuous variables were done by using the Student's t-test or the Wilxocon test respectively; categorical variables were compared by the chi-square test or Fisher's exact test when appropriate. Evaluation of differences in longitudinal measurements of biochemical bone markers between groups was done by using mixed linear models. A p-value of < 0.05 was considered statistically significant. Data are given as means ± SD or median and interquartile range (IQR) when appropriate. The analysis was conducted with SAS version 9.1.3 (Cary, NC).


Study population

A total of 96 patients underwent randomization, 47 in the study (ZOL) and 49 patients in the control group (CON) (Figure 1). The ZOL and CON groups did not differ significantly in major baseline characteristics (Table 1).

Table 1.  Patient baseline characteristics
 CON group (49)ZOL group (47)p-Value
  1. Data are presented as means±SD or as counts when appropriate.

  2. BMI = body mass index (kg/m2); MELD = model of end-stage liver disease score.

  3. *Pre-liver transplantation data (bone mineral density: Z-score, T-score).

  4. **Child–Turgotte–Pugh  Score.

Z-score femur*−0.33−0.290.87
T-score femur*−1.41−1.230.47
Z-score (L1-L4)*−0.67−0.910.47
T-score (L1-L4)*−1.12−1.290.61
BMI (BW kg/m2)26±4.026±4.10.73
Child-Pugh A/B/C**7/26/163/25/190.42
Indications for LT
 Choletstatic 5 41.0 
 Other causes12 70.31

The mean age of the patients was 54 years and the majority of the patients were male (75%). There was no difference in sex distribution between the ZOL (12 women and 35 men) and the CON group (12 women and 37 men; p = 0.91). Nutritional status as estimated by BMI (body mass index) was not significantly different between the ZOL (25.8 kg/m2) and CON groups (26.1 kg/m2; p = 0.73). No differences were found in age (p = 0.198) and in the numbers of postmenopausal women (p = 0.54) between the two groups (Table 1).

According to the Child-Turcotte-Pugh classification, the degree of liver disease was as follows: 3 patients presented with class A cirrhosis in the ZOL group and 7 patients in the CON group (most of these patients had hepatocellular carcinoma), 25 patients presented with Child class B in the ZOL group versus 26 patients in the CON group, and 19 patients presented in class C in the ZOL group vs. 16 patients in CON group (p = 0.42). Both groups had a similar model for end-stage liver disease (MELD) score (ZOL 17.0, CON 15.4; p = 0.54) at the time of LT (24).

The most common underlying liver diseases was alcohol-related cirrhosis in 48 patients (ZOL 26 vs. 22 patients; p = 0.41) followed by chronic viral hepatitis in 20 (ZOL and CON each 10 patients; p = 1.0), primary biliary cirrhosis and primary sclerosing cholangitis in 9 (ZOL 4 vs. CON 5; p = 1.0) patients. The remaining 19 patients had different and rare liver diseases, such as autoimmune hepatitis, hepatocellular carcinoma, hereditary hemochromatosis, cryptogenic cirrhosis and Wilson disease (ZOL 7 vs. CON 12 patients; p = 0.31) (Table 1).


We did not find any significant difference in the frequency of biopsy-proven acute rejection episodes between the ZOL group (9 patients) and the CON group (6 patients) (p = 0.35). The cumulative doses and trough levels of cyclosporine A were not different between ZOL and CON groups (dose: p = 0.70; trough level: p = 0.71). The cumulative steroid dose (prednisolone-equivalent) within the first 12 months was 1863 ± 945 mg in the ZOL and 1773 ± 615 mg in the CON group (including bolus therapy for treatment of acute rejection episodes) and did not differ significantly among the groups (p = 0.56) (Table 2).

Table 2.  Adverse events
Adverse eventZOL group n = 47 (%)CON group n = 49 (%)p-Value
  1. Data are given as counts and percentages.

  2. GvHD = graft versus host disease; PTLD = post-transplant lymphoproliferative disease; GI = gastrointestinal.

Death 8 (17)12 (24)0.45
 Septicaemia4 (8) 8 (16)0.52
 PTLD1 (2)00.49
 GvHD1 (2)00.49
 GI bleeding02 (4)0.50
 Cardiac arrest1 (2)00.49
 Epidermoid cancer lingua01 (2)0.49
 Thrombosis A. hepatica1 (2)00.49
 Hepatocellular carcinoma recurrence01 (2)0.49
Discontinuation 7 (15) 7 (14)0.49
 Strong acute phase reaction to study drug1 (2)00.49
 Liver or kidney dysfunction320.67
 Intracranial Haemorrhage111.0 
 GI bleeding100.49
 Bisphosphonate therapy011.0 
 Biliary tract complications011.0 
 Pontine myelinolysis100.49
Infection 9 (18)10 (20)1.0 
Nausea 8 (17)4 (8)0.23
Pyrexia 5 (11)0 0.025
Musculoskeletal pain 8 (17)2 (4) 0.049
Hypocalcemia (serum Ca < 2.1 mmol/L)7 (15)2 (4) 0.088
Hypercalcemia (serum Ca >2.65 mmol/L)2 (4)3 (6)1.0 

Main outcome

The primary endpoint of bone fracture was reached in four patients (8.5%) in the ZOL group and in 11 patients (22.5%) in the CON group (p = 0.050, log-rank test) (Figure 2). All fractures were vertebral fractures. Seventy-five percent event-free time of survival was 360 days in the ZOL group and 200 days in the CON group.

Figure 2.

Primary study endpoint: fracture (log-rank test, p = 0.050).

Secondary endpoints

The combined endpoint of fracture and death was reached in 26% of the patients in the ZOL group and 46% in the CON group (p = 0.047, log-rank test) (Figure 3). The densitometry measurements before LT were the same between the ZOL and CON groups (lumbar spine (L1-L4): Z-score: p = 0.47; T-score: p = 0.61; femur: Z-score: p = 0.87, T-score: p = 0.47) (Table 1).

Figure 3.

Composite endpoint: fracture and death (log-rank test, p = 0.047).

Densitometry results of femoral neck density were different between the groups at 6 months after transplantation (mean±SD BMD ZOL: 0.80±0.19 g/cm2 vs. CON: 0.73 ± 0.14 g/cm2; p = 0.036). No significant differences were found in the BMD of the femoral neck at 12 months (p = 0.11) and at lumbar part of spinal column after 6 (p = 0.07) and 12 months (p = 0.52) after LT.

Mixed linear models for biochemical bone markers showed a lower increase in osteocalcin in the zoledronic acid group. All other serological markers (OPG, bone ALP, intact parathyroid hormone, type I collagen peptides cross-laps, 1.25 (OH)2 D3) remained the same between the two groups over time.

Histomorphometry showed lower values for OV/BV (−66.3%, p = 0.006), O.Th. (−30.4%, p = 0.009), OS/BS (−59.2%, p = 0.029), Ob.S/BS (−85.4, p = 0.022) in the ZOL versus the CON group.

The parameters of BMDD like the mean bone matrix mineral content CaMean (21.85 ± 0.60/21.37 ± 1.21 p = 0.16 ZOL/CON) as well as CaPeak and CaWidth were not different between the groups. However, CaLow (−42.9% p = 0.019), the proportion of newly formed bone, significantly decreased in the ZOL group compared with the CON.

Adverse events

We did not find any significant differences between the ZOL and the CON groups in the rates of transplantation-related or drug-related adverse events in our analysis except for the side effects after the first infusion of zoledronic acid, with more myalgia, athralgia, pyrexia (ZOL 5 and CON 0 patients), most of which were rated as mild. One patient in the ZOL group withdrew from the study because of a strong reaction of high fever over 3 days and severe athralgia and myalgia. One patient in the control group withdrew from the study for taking bisphosphonates. Only pyrexia was used more frequently in the ZOL group. In all five patients fever was directly related to the bisphosphonate infusion and stopped within 24 hours.

Hypocalcaemia (15%, 7 patients) developed in more patients in the ZOL group. Serum creatinine averaged 1.19 ± 0.78 mg/dL before the first use of zoledronic acid in the ZOL and 1.12 ± 0.73 mg/dL in the CON groups respectively (p = 0.65). Four patients in the ZOL group had creatinine levels higher than 2.5 mg/dL 1 month after transplantation. These patients received the regular bisphosphonate infusion a week later, after medical therapy. One year after transplantation, mean serum creatinine values were 1.37 ± 0.42 mg/dL in the ZOL and 1.25 ± 0.35 mg/dL in the CON groups (p = 0.18).


In this study, we evaluated the use of zoledronic acid for the prevention of bone fractures after LT. Fewer patients receiving zoledronic acid treatment reached the primary end point of first fracture. To the best of our knowledge this is the first study identifying a treatment effectively reducing atraumatic fractures after LT.

Several other studies on the prevention of bone loss after organ transplantation have been published previously, but no trial has shown an effective treatment to prevent fractures after LT. Previous studies have investigated the effects of bisphosphonates on BMD after cardiac and renal transplantation (11,25–27,31). Most have suggested a beneficial effect; however, these studies were small, some nonrandomized, and used historical controls (11, 25). Shane et al. showed minimal bone loss after heart transplantation treated with either calcitriol or alendronate without a detectable advantage for either substance (28). Another investigator of an uncontrolled trial suggested that alendronate was efficacious in preventing bone loss associated with LT (13). No significant differences could be shown in their trial and Crawford et al. pointed out in a randomized controlled setting, that zoledronic acid can prevent bone loss within the first year after liver transplantation, but his trial was not powered to assess fractures (29). So far no effective prevention for fractures after organ transplantation exists, but bone fractures have a severe negative impact on the patients' quality of life (2). BMD is a weak predictor for the risk of fractures in the normal population and it has been extensively evaluated for its diagnostic utility in women with postmenopausal osteoporosis (30). Its value for the assessment of fracture risk in patients after organ transplantation suffering from pathophysiologically complex metabolic bone diseases remains unclear. However, many studies have shown that particularly during the first month after transplantation, dramatic bone loss can occur due to the catabolic disequilibrium of bone metabolism (29).

At the time we initiated this study we therefore decided to use a high dosage of zoledronic acid (total 24 g over 1 year) to ensure a strong anticatabolic effect within the initial post-transplantation period. Although we gave the patients a high dosage of bisphosphonate, the side effects were mild. We saw a mild acute phase reaction of mylagia, athralgia and pyrexia in some patients only during the first application of the study drug. Only one patient withdrew from the study for having a strong reaction of high fever lasting for more than 3 days.

The results of osteodensitometry showed significant differences in the femoral neck density 6 months after transplantation. Analyzing BMD as a study endpoint seems problematic because of the high variability and lack of correlation between the risk of fracture and BMD. Mixed linear models of biochemical bone markers showed a lower increase in osteocalcin in the ZOL group, indicating suppression of bone turnover. All other serological bone markers were the same between the groups over time.

Bone biopsies showed that the static parameters of bone formation and the amount of newly formed bone (CaLow) were reduced in the ZOL group compared with the CON group. These changes in CaLow and histomorphometric parameters are consistent with the anticatabolic action of zoledronic acid on bone tissue. The lack of changes in bone mineral content (CaMean) indicates that the treatment with high doses of zoledronic acid did not result in any kind of hypo- or hypermineralization of the bone matrix within the 6 months of observation.

In conclusion, the main finding of this prospective study was that prophylactic treatment with bisphosphonate zoledronic acid reduced fractures associated with a reduction in bone turnover after liver transplantation. This provides a major benefit for the transplant recipients.


This study was supported partly by the Austrian Science Fund FWF (Grant P-18325) and the Austrian Academy of Science (Grant EST 370/04) to R.O. and FWF (Grant P-16880) to K.K.