The authors state that they have no conflicts of interest.
Incidence and Predictors of Fractures in Children After Solid Organ Transplantation: A 5-Year Prospective, Population-Based Study†
Article first published online: 21 NOV 2005
Copyright © 2006 ASBMR
Journal of Bone and Mineral Research
Volume 21, Issue 3, pages 380–387, March 2006
How to Cite
Helenius, I., Remes, V., Salminen, S., Valta, H., Mäkitie, O., Holmberg, C., Palmu, P., Tervahartiala, P., Sarna, S., Helenius, M., Peltonen, J. and Jalanko, H. (2006), Incidence and Predictors of Fractures in Children After Solid Organ Transplantation: A 5-Year Prospective, Population-Based Study. J Bone Miner Res, 21: 380–387. doi: 10.1359/JBMR.051107
- Issue published online: 4 DEC 2009
- Article first published online: 21 NOV 2005
- Manuscript Accepted: 21 NOV 2005
- Manuscript Revised: 1 NOV 2005
- Manuscript Received: 12 AUG 2005
- vertebral fractures;
- solid organ transplantation;
- bone disorder
In this population-based prospective follow-up study, children undergoing solid organ transplantation had a highly elevated risk for fractures: The incidence of all fractures was 6-fold higher (92 versus 14 fractures/1000 persons/year; p < 0.001) and vertebral fractures was 160-fold higher (57 versus 0.35 fractures/1000 persons/year; p < 0.001) in the study group compared with the control population. Thus, screening of vertebral fractures at regular intervals is recommended, and preventive strategies should be studied.
Introduction: The incidence and predictors of fractures after solid organ transplantation are not well documented in the pediatric age group.
Materials and Methods: A total of 196 children, which is 93% of patients surviving kidney, liver, and heart transplantation in our country, participated in a retrospective chart review at enrollment followed by a 5-year prospective follow-up study between January 1999 and December 2004. Hospital and medical records were reviewed. All children underwent clinical examinations and answered questionnaires concerning fracture history at the beginning and at the end of the prospective follow-up. Radiographs of the thoracic and lumbar spine were obtained. The fracture incidence was compared with data obtained from public health registries.
Results: Seventy-five (38%) of the transplant patients suffered from a total of 166 fractures after organ transplantation. The incidence of all fractures was 6-fold higher (92 versus 14 fractures/1000 persons/year; p < 0.001) and vertebral fractures was 160-fold higher (57 versus 0.35 fractures/1000 persons/year; p < 0.001) in the study group compared with the control population. The age- and sex-adjusted hazard ratios (95% CI) were 61.3 (40.7-92.4) for vertebral, 17.9 (8.96-35.8) for symptomatic vertebral, 0.99 (0.65-1.50) for nonvertebral, and 2.90 (2.25-3.73) for all fractures in the patients compared with the control population. In a multivariate analysis, older age (hazard ratio [95% CI]; 2.02 [1.07-3.83]), male sex (2.15 [1.22-3.81]), liver transplantation (1.78 [1.01-3.14]), and fractures before transplantation (2.02 [0.92-4.47]) were the most significant independent risk factors.
Conclusions: Children undergoing solid organ transplantation have a highly elevated risk for fractures. Screening of vertebral fractures at regular intervals is recommended, and preventive strategies should be studied.
KIDNEY, LIVER, AND heart transplantations have become a successful mode of therapy for children with terminal organ failure. Long-term patient survivals of >80% are reported in large registries, and the quality of life of most patients is good.(1-3) The transplant children, however, need life-long immunosuppressive medication, usually consisting of three drugs: a calcineurin inhibitor (cyclosporin A or tacrolimus), cytotoxic agent (azathioprine or mycophenolate mofetil), and corticosteroid (prednisone or methylprednisolone). Proper dosing of these drugs is important to avoid rejections and to minimize side effects. This is especially important in pediatric patients who have an expected life span of several decades.
In adults, osteoporosis-related fractures are common after transplantation.(4-8) Increasing age, low lumbar BMD, and vertebral fractures before transplantation have been predictors of fractures.(6) Several factors, such as immobilization, metabolic bone disease, and immunosuppressive drugs, may compromise the quality of bone also in children undergoing transplantation.(9-11) In contrast to adults, decreased BMD has been reported only in a small proportion of pediatric transplant patients.(12-14) The relationship between low BMD and fracture risk has not been established in children. In addition, 2-D DXA may overestimate the occurrence of decreased BMD if only age- and gender-specific reference values are used without volumetric or height adjustment.(15,16)
The data on fracture incidence in pediatric transplant patients are very limited.(10,11) In this patient group, only two small retrospective studies on the occurrence of fractures have been reported. In a retrospective study of 117 liver transplant children, at least one fracture was observed in 19 (16%) children,(10) and in a cross-sectional study of BMD in 36 patients after cardiac, liver, or bone marrow transplantation, 3 children (8%) had radiographic evidence of spinal fracture.(13)
We studied the incidence and predictors of bone fractures in a population-based cohort of pediatric kidney, liver, and heart transplant patients and compared the findings with an age-matched control population. This is the first reported prospective study to investigate the incidence and predictors of fractures after pediatric solid organ transplantation.
MATERIALS AND METHODS
All pediatric organ transplantations in our country are performed at the Hospital for Children and Adolescents, Helsinki University Hospital. The postoperative care is also centralized so that all patients have follow-up visits at our hospital at 1- to 6-month intervals during the first 2 years and annually thereafter. During the period of 1983-2002, 211 subjects <18 years of age received kidney (n = 135), liver (n = 46), or heart (n = 30) transplants. The indications for transplantation were typical for the pediatric age group.(1-3) All 211 subjects were invited to participate in this study, and 196 (93%) agreed (Table 1). Fifteen patients (7%) did not participate because of poor health (six patients), social reasons (seven patients), or lack of interest (two patients). Five (2.6%) patients died during the prospective follow-up period. Twenty-four primary transplantations were performed during the first half of the follow-up period. Fifteen patients had, altogether, 18 fractures before transplantation (Table 1), all of which were nonvertebral. Of them, six were kidney, six were liver, and three were heart transplant patients with typical diagnoses leading to transplantation. Their mean age at transplantation was 10.9 years (range, 1.9-18.0 years). Their pretransplant medication or medical history was not different from the rest of the patients.
The patients received triple drug immunosuppression with cyclosporine A, azathioprine, and methylprednisolone.(17) The protocol was very similar for all organs (Table 1). Cyclosporine was started perioperatively, and the dose was adjusted to obtain trough blood concentrations of 300-400 μg/liter during the first weeks. The doses were slowly reduced to attain levels of ∼100 μg/liter at 1 year and thereafter (for heart transplant patients, the target was 120-140 μg/liter). Methylprednisolone was given 1-3 mg/kg/day during the first postoperative days and was rapidly tapered to 0.25 mg/kg at 2 weeks. Every other day dosing (0.3 mg/kg) of methylprednisolone was commenced at 3 months in kidney transplant patients and at 6 months in liver and heart recipients. After the first year, methylprednisolone dose was not increased with growth, and the mean dose was 0.12 mg/kg/day (Table 1). Azathioprine was first given at 2 mg/kg/day and was reduced to 1 mg/kg/day after 2 weeks and increased to 1.3-1.5 mg/kg/day at 3 (kidney transplantation) or 6 months (liver and heart transplantation). Only a few patients were switched to tacrolimus or mycophenolate mofetil during the study period. Acute rejections were mostly treated with a 5-day course of methylprednisolone (3 mg/kg/day). The patients were recommended to use calcium supplementation (500 mg/day), and all children under 4 years of age also had vitamin D (400 IU/day) recommended. No one received bisphosphonates or calcitonin before or during the follow-up period.
The study design was a retrospective chart review at enrollment followed by a 5-year prospective study from January 1999 to December 2004. Clinical data including age, sex, height and weight, underlying condition leading to transplantation, rejections, immunosuppressive medications, and diagnosed fractures were recorded from the hospital charts. Daily and cumulative methylprednisolone (mg/kg), azathioprine, and cyclosporine A doses were calculated. In the beginning and at the end of the prospective study, all patients underwent a detailed clinical examination of the spine and extremities and filled out a questionnaire on fracture history. In addition, fracture history was systematically recorded at yearly follow-up visits during the 5-year prospective study period. For each reported fracture, the localization, type, mechanism of injury (low/high energy; e.g., falling >3 m or traffic accident), and treatment were recorded. If a fracture was clinically suspected, additional radiographs of the injured or painful area were obtained.
BMD was measured at the time of transplantation and yearly thereafter starting in 1999 in patients ≥4 years of age (n = 54). A posteroanterior and lateral radiograph of the thoracic and lumbar spine was obtained at the final follow-up visit in those participants who had a history of vertebral fracture or back pain; increased kyphosis, rib hump, or other deformity in the clinical examination (n = 53). In the other patients, these images were obtained using a Hologic Discovery A scanner (n = 132). A lateral and posteroanterior image of the thoracic and lumbar spine at the final follow-up was available in 185 (94%) of the 196 patients.
The study was approved by The Ethics Research Board, Helsinki University Hospital; a written informed consent was obtained from the patient/parents.
Lumbar BMD measurement
BMD was measured at the lumbar spine (L1-L4) by DXA using a Hologic 4500A scanner (Waltham, MA, USA) in 1999-2002 and a Hologic Discovery A scanner in 2003-2004. The equipment was calibrated with the European Spine Phantom. To minimize the effect of bone size on BMD values, a mathematical model for correction of areal BMD values for the anteroposterior depth of the vertebrae was used to obtain volumetric BMD, which was calculated as follows: volumetric BMD (L2-L4) = total areal BMC (L2-L4)/[(L2-L4 area)1.5].(18) Reference values were derived from previously published data.(19,20) In addition to volumetric Z score values, absolute lumbar BMD values (g/cm2) are given.
A lateral and posteroanterior image of the spine (from T4 to L5) was obtained by the Hologic Discovery A scanner or by radiography. Each vertebra was morphologically graded as normal, wedged, concave, or compressed; characteristic findings of Scheuermann's disease and osteophytes were recorded. If a new fracture was suspected in the Instant Vertebral Assessment images obtained with the DXA scanner, standing posteroanterior and lateral plain radiographs were also obtained. Vertebrae with wedge, concavity, or compression deformity were recorded.(21) Only obvious vertebral height reductions (a decrease of the vertebral height >20% for anterior, middle, or posterior part of the vertebral body) or fracture lines were classified as fractures, and the fractures were further graded as previously published (grades 2a, 2b, 3a, or 3b).(21) All radiographs were graded independently by two orthopedic surgeons (IH, SS). The final grades were based on consensus decision; borderline cases were classified as normal.
More than 99% of acute fractures occurring in children and young adults are treated at primary health care centers or university hospitals in the city of Helsinki.(22) All visits are recorded in detailed registers. The patients' age, sex, and number, localization, type, and mechanism of injury (low or high energy; e.g. falling >3 m or traffic accident) of fractures were recorded from these registers for 2001 for patients 0-30 years of age. This population was used as a control population for this study and comprised 207,103 individuals. Because only 1 year of data were available, we estimated the number of normal population remaining free of fractures (Kaplan-Meier method) during a similar follow-up period than in the patient group.
The probability of patients remaining free of all, vertebral, and nonvertebral fractures were estimated using the method of Kaplan-Meier. Stratified Kaplan-Meier curves were used to assess effect risk factors (age, sex, body mass index, underlying disease, lumbar BMD, transplanted organ, retransplantation) for fractures after transplantation. Survival data obtained in the Kaplan-Meier analysis were compared by the log-rank test (univariate analyses). A Cox model with time-dependent variables was used to assess the effect of cumulative methylprednisolone or cyclosporine dose and the number of rejections on fracture incidence. A multicovariate, Cox's proportional hazard regression model was used to calculate age- and sex-adjusted hazard ratios (HRs) and 95% CIs for all, vertebral, symptomatic vertebral, and nonvertebral fractures in the patient group compared with the control population. A multicovariate, Cox's proportional hazard regression model was used to assess the effect of multiple risk factors for developing any fracture in the patient group. This model included the following covariates: age, gender, type of transplantation (kidney, heart, liver), and fractures before transplantation.
The corresponding author (IH) had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
The incidence and type of fractures were assessed in a national cohort of 123 kidney, 44 liver, and 29 heart transplant children both in retrospective and prospective manner. The subjects were transplanted at the mean age of 6.5 years (range, 0.4-18.1 years) and were followed on average for 9.2 years (range, 2.4-20.5 years) with a total follow-up time of 1795 person-years (Table 1). Their mean age at the time of last follow-up was 15.7 years (range, 3.7-30.4 years).
A total of 166 fractures in 75 (38%) patients were recorded after the transplantation (Table 2). Forty-seven of these were diagnosed in 21 children (11%) during the retrospective phase, and 119 occurred in 54 subjects (28%) during the prospective 5-year period. Sixty-one percent (102 of 166) of the fractures were vertebral (in 37 patients) and the rest (39%, 64 of 166) were nonvertebral (Table 2). One-half (51 of 102) of the vertebral fractures were asymptomatic. Back pain currently was reported often or very often at rest by 42 (21%) of the patients. Eight (11%) patients needed operative fracture treatment. The fractures occurred at any time after transplantation (Fig. 1), and the mean age at the time of fracture was 13.9 years. The mean body mass index in patients with fractures was slightly higher at the time of fracture (19.8 kg/m2; range, 12.6-28.9 kg/m2) compared with those without fractures (mean, 18.0 kg/m2; range, 13.6-32.8 kg/m2; p = 0.0002). In the control group, a total of 2924 fractures were observed during 1 year. The majority of all fractures in the patients (151 of 166; 91%) and in the controls (2636 of 2924; 90%) were caused by low-energy trauma.
The incidence of all fractures was 92 fractures per 1000 person-years for the transplant patients, which was 6-fold higher than that observed for the controls (14 fractures per 1000 person-years; p < 0.001; Table 2). Similarly, the incidence of vertebral (57 versus 0.35 fractures per 1000 person-years, p < 0.001), symptomatic vertebral (28 versus 0.35 fractures per 1000 person-years, p < 0.001), and nonvertebral fractures (38 versus 14 fractures per 1000 person years, p = 0.0025) were 160-, 81-, and 3-fold higher in the study group compared with the control population, respectively.
As shown in Fig. 1, the proportion (Kaplan-Meier estimate) of transplant patients having at least one vertebral, symptomatic vertebral, nonvertebral, or any fracture was 18%, 11%, 27%, and 40% after 10-year follow-up and 0.3% (relative risk [95%CI]; 5.14 [2.65-9.97]), 0.3% (6.41 [2.23-18.4]), 13% (1.44 [1.01-2.05]), and 14% (2.33 [1.64-3.29]) in the control population, respectively.
The age- and sex-adjusted HRs (95% CI) were 61.3 (40.7-92.4) for vertebral, 17.9 (8.96-35.8) for symptomatic vertebral, 0.99 (0.65-1.50) for nonvertebral, and 2.90 (2.25-3.73) for all fractures in the patients compared with the control population.
Reduced lumbar spine BMD (Z score < −2.0 after volumetric correction) was observed in 9 (17%) of the 54 patients studied at the time of transplantation. The same finding was recorded in 10 of these patients (19%) at the final follow-up. The corresponding mean Z scores were −0.37 (range, −4.8 to 2.8) and −1.1 (-3.4 to 3.5), respectively, and absolute areal BMD values were 0.67 (0.37-1.09 g/cm2) and 0.77 g/cm2 (0.45-1.13 g/cm2), respectively. In the entire study group, a Z score < −2.0 after volumetric correction was found in 43 (22%) of the 196 patients at final follow-up visit, with a mean volumetric BMD Z score of −1.0 (-4.8 to 4.0) and a mean areal BMD of 0.74 g/cm2 (0.31-1.29 g/cm2). The growth of the patients after kidney, liver, or heart transplantation is shown in Fig. 2.
Predictors of fractures
Univariate analyses showed that male sex, older age at transplantation, higher body mass index (≥19 kg/m2) at the time of transplantation and at 1 year after transplantation, and fractures already before transplantation (recorded in 15 patients) increased the risk of fractures during the follow-up (Table 3). Fractures before transplantation increased the risk of all (3.50 [1.07-11.5]), vertebral (2.73 [0.43-17.3]), and nonvertebral fractures (4.61 [1.12-18.9]) after transplantation. In contrast, the type of graft (kidney, liver, or heart), retransplantation, number of acute rejections, or poor growth (defined as a decrease of >0.5 in height SD score units or more per year) were not associated with the fracture risk. Because the fractures occurred from a few weeks to 14 years after transplantation, the cumulative doses of immunosuppressive drugs before the development of fractures were very variable. However, the daily doses (mg/kg/day) for methylprednisolone, cyclosporine A, and azathioprine, and the cyclosporine blood levels were not higher in patients who developed fractures compared with those without fractures, as analyzed at several time-points after transplantation (Fig. 3). Cumulative methylprednisolone, cyclosporine A, or azathioprine was not associated with increased risk of fractures in a Cox model.
In a multivariate analysis, older age, male sex, liver transplantation, and fractures before transplantation were the most significant independent risk factors for all fractures after transplantation (Table 3).
We analyzed the incidence and type of bone fractures in a national cohort of pediatric transplant patients who were on a triple drug immunosuppressive medication. The results showed that the transplant children have a highly increased risk for fractures with an annual incidence for vertebral and nonvertebral fractures of 5.7% and 3.8%, respectively. Strategies for fracture prevention should be seriously considered for pediatric transplant patients.
The overall occurrence of fractures has ranged from 9.3% to 44% in adult transplant patients,(4,5,23,24) and during the first 3 years after transplantation, the incidence has been 18-35%.(6,25) In the pediatric age group, only two small retrospective studies on fractures have been reported. In a retrospective study of 117 liver transplant children, at least one fracture was observed in 19 (16%) children,(10) and in a cross-sectional study of BMD in 36 patients after cardiac, liver, or bone marrow transplantation, 3 children (8%) had radiographic evidence of spinal fracture.(13) In our cohort of kidney, liver, and heart transplant children, 38% of the 196 patients had fractures, and in 19% of the patients, fractures were diagnosed on several occasions. This clearly higher occurrence and incidence is most probably caused by the longer follow-up and the prospective nature of our study. The incidence of fractures was higher during the prospective part of our study. This was caused by, at least in part, a more careful evaluation during the prospective phase. The fractures occurred at any time after transplantation, and it seems that, in accordance with adults, the increased risk of fractures continues long term.(26) Whether this continuing risk is caused by transplantation or more vulnerable age to fractures, such as pubertal years, remains unclear. The incidence of all fractures in the control population coincided well with a previous report in our country.(22) The incidence of vertebral fractures in the control population was based on clinical fractures only, whereas in the patient population, asymptomatic fractures were also included. To diminish this bias, we separately evaluated the risk of both symptomatic and all vertebral fractures in the patient and control populations. The incidence of symptomatic vertebral fractures in the patient population remained 81-fold compared with that of control population. In addition, the incidence of vertebral fractures in our control population was five times higher than that reported previously from Sweden.(27)
Volumetric BMD values were calculated to minimize the impact of bone size on BMD values.(15) Kröger et al.(28) and Cochat et al.(29) have previously reported a constant volumetric BMD before puberty using this method. Klaus et al.(15) reported that volumetric BMD values using a similar method were not decreased in pediatric kidney transplant patients compared with reference values. The present data suggest that every fifth transplant patient shows decreased volumetric BMD values.
In the adult transplant recipients, the fracture risk was highest during the first 2 years after transplantation.(6) In this study, the risk for fractures remained almost the same throughout the study period. The reason for this difference remains unknown. Growing bone, however, has increased capacity to resist a fracture, and it seems possible that a longer period is needed for the development of permanent damages in the growing skeleton.
Transplant children had a 160-fold higher incidence of vertebral fractures compared with controls. What is the significance of this observation? Vertebral fractures may cause severe pain and functional impairment as well as result in spinal deformity. They may also be the cause of recurrent back pain reported by one-fifth of the transplant children. On the other hand, some vertebral fractures are completely asymptomatic, but with the lifelong immunosuppressive medication, spontaneous recovery is unlikely. The possibility that some asymptomatic vertebral fractures could have occurred already before transplantation seems rather unlikely, because an analysis of thoracic lateral radiographs at the time of transplantation did not reveal vertebral compression fractures. It seems justified to screen all children of ≥5 years of age for vertebral fractures with regular intervals after organ transplantation.
Male sex, older age, liver transplantation, higher body mass index, and fractures before transplantation were risk factors for post-transplant fractures in our study. On the other hand, we did not find differences in the medication of patients with or without fractures. All patients received a similar combination of cyclosporin A, azathioprine, and methylprednisolone. The dosing of the first two drugs was “traditional,” with the average maintenance doses of 5.2 and 1.3 mg/kg/day, respectively. The exposure to methylprednisolone was minimized so that the dose was tapered to 0.25 m/kg/day at 2 weeks after transplantation, and every other day dosing was commenced 3-6 months after the operation. With the average maintenance dose of 0.1 mg/kg/day, the overall side effects remained modest, as evidenced by relatively stable body mass index postoperatively. However, it is clear that even small doses of glucocorticoids may affect the bone and possibly increase the risk for fractures.(9) Higher body mass index increased the risk of fractures during follow-up. Previous studies both in chronically ill children(21) and in obese children(30) suggest that increased body weight may predispose to compression fractures possibly through mechanical forces.
Determinants of post-transplantation bone disease have not been clearly defined in children.(9) Hill et al.(10) observed a metabolic bone disease (rickets, osteopenia, or osteosclerosis) in 17 of 19 patients with fractures. Guthery et al.(14) and Boot et al.(31) observed an association between the cumulative glucocorticoid dose and decreased lumbar BMD. Daniels et al.(13) did not observe any correlation between steroid dose and areal or estimated volumetric BMD. It is likely that fracture risk can not be predicted by bone mass or BMD alone and that additional methods to assess the quality of bone are needed.(32)
The findings of this study warrant further research. The nature of the bone pathology resulting in increased fragility should be analyzed. Pharmacological treatment options need to be addressed. Bisphosphonates have been successfully used in adults to prevent bone loss after transplantation.(33,34) However, in pediatric patients with secondary osteoporosis, bisphosphonate treatment has been regarded as experimental.(35-38) Based on our results, controlled clinical trials on pharmacological fracture prevention in children after solid organ transplantation are urgently needed to avoid the deleterious long-term skeletal complications.
This study was financially supported by the Päivikki and Sakari Sohlberg Foundation, Foundation for Paediatric Research, Paulo Foundation, and Helsinki University Central Hospital. The sponsors of the study had no role in study design, data collection, data analysis, or writing of the report.
- 2Research Group SPLIT 2004 Studies of pediatric liver transplantation 2002: Patient and graft survival and rejection in pediatric recipients of a first liver transplant in the United States and Canada. Pediatr Transplant 8:273–284., , , ,
- 222000 Incidence and etiologies for fractures in European children. Finn Med J 41:4135–4140., ,
- 362003 Bisphosphonates in children with bone diseases. N Engl J Med 349:2068.