Bisphosphonate treatment for skeletal complications in paediatric cancer—Experience from a single tertiary centre

The aim was to analyse the use and safety of bisphosphonate treatment for metabolic bone complications in paediatric cancer patients.


| INTRODUC TI ON
The prognosis of childhood cancer has markedly improved during the past several decades with approximately 80% of paediatric cancer patients currently surviving long term in Europe. 1 With the increasing survival rate, the adverse effects and long-term sequelae of childhood cancer treatment have become greatly important.
Skeletal complications are common in paediatric cancer patients.
Approximately, 2%-10% of children with acute lymphoblastic leukaemia (ALL), the most common paediatric malignancy, develop symptomatic osteonecrosis. 2 The risk of osteonecrosis in paediatric ALL is increased among teenagers as compared with younger children, [3][4][5][6] and in those with higher corticosteroid exposure. 4,6Some studies also show a higher risk for female, 3,5,6 and overweight patients. 6Low-bone mineral density (BMD) and an increased risk of osteoporotic fractures during cancer treatment have also been well described.Vertebral fractures in children are always a sign of osteoporosis.The cumulative incidence of vertebral fractures within 6 years after ALL diagnosis was 32.5% when systematically screened with magnetic resonance imagining, most diagnosed during the first 2 years after diagnosis. 7BMD deficits in paediatric cancer patients have multifactorial causes but specific risk factors are acknowledged, including corticosteroid treatment and haematopoietic stem cell transplantation (HSCT). 2,7Nevertheless, both osteoporotic vertebral fractures and osteonecrosis can also manifest already at ALL diagnosis. 2,8Although skeletal complications are well known, their management remains unestablished.Treatment optimisation is warranted as skeletal problems can significantly impair the quality of life of the survivors.
Bisphosphonates reduce bone resorption by inducing osteoclast apoptosis.They have an established role in the treatment of high-turnover osteoporosis in adults and children with osteogenesis imperfecta.Their use in other paediatric indications, including secondary osteoporosis, remains less well established.One study found that bisphosphonates relieved osteonecrosis-related pain in paediatric leukaemia patients. 9A systematic review that was published in 2021 and included five original articles also found that bisphosphonate treatment alleviated pain and improved mobility in ALL-related osteonecrosis. 10 There are only a few studies reporting the outcome of bisphosphonate treatment in paediatric cancer patients.
In this retrospective study, we describe our experience in the use of bisphosphonate in paediatric cancer patients in a single tertiary centre.

| Study cohort
The retrospective cohort included all paediatric (age <18 years at cancer diagnosis) patients treated with bisphosphonates between 1999 and 2020 at the Children's Hospital, Helsinki University Hospital, the largest, tertiary university hospital in Finland.Altogether 26 cancer patients had been treated with bisphosphonates during the study period in our clinic.One subject was excluded because of a discontinued follow-up after moving to another hospital district.The final cohort thus included 25 children with a primary cancer diagnosis between 1994 and 2017.
Patient data were retrieved manually from the hospital records.
Data were first retrieved by J.S. and then reviewed by P.U., a specialist in paediatric haematology.

| Bone mineral density measurement
The dual-energy x-ray absorptiometry (DXA) measurements performed on patients before and after bisphosphonate treatment were collected from the medical records.The bone mineral content (g) and areal BMD (g/cm 2 ) were assessed for lumbar spine, femoral neck, and whole body with DXA (DXA; paediatric software, Discovery A, versions 12.1 to 13.5.1;Hologic, Marlborough, MA).The DXA measurements were converted to standard Z-scores based on equipment-specific age-and sex-adjusted reference data for Caucasian American children.These reference values have been validated for Finnish children. 11The Z-scores were used in the statistical analyses.

| Diagnosis of osteonecrosis and vertebral fractures
Osteonecrosis is defined radiographically as a circumscribed lesion with a distinct rim.Because of the retrospective setting, the diagnosis of both osteonecrosis and vertebral fractures was based on the concurrent radiologists' view.Osteonecrosis was not routinely screened but radiological diagnostics were initiated because of symptoms like pain and stiffness.

| Data analyses
For evaluating the outcome of the bisphosphonate treatment, we developed a three-stage classification for the status of bone

Key Notes
• Skeletal complications are common in childhood cancer and can significantly impair the quality of life in survivors, but optimal treatment remains to be determined.
• This single-centre retrospective study evaluated the use and outcome of bisphosphonate treatment for metabolic bone complications in childhood cancer patients.
• Bisphosphonate treatment proved well tolerated in this patient group, also when given during cancer treatment.complication the end of follow up, comparing symptoms and findings pre-and post-bisphosphonate treatment: (1) the absence of symptoms and/or findings of the complication; (2) symptoms and/or findings improved but still present; (3) no improvement in the symptoms or findings, or a need for a surgical intervention because of the bone complication.The presence or absence of symptoms, like pain, was based on the hospital recordings by the treating physician.An increase in the BMD of 1 standard deviation or more was regarded as significant if the patients' symptoms had been relieved in tandem.Osteonecrosis was regarded to be improved if the pain was relieved and the findings in imaging studies were either stable or improved.

| Statistics
IBM SPSS Statistics for Windows (Version 26.0.Armonk, NY: IBM Corp) was used for statistical analyses.The mean or median with range are shown for numerical parameters.When comparing the prevalence of adverse effects of the bisphosphonate treatment between those treated during or after the chemotherapy, Fisher's exact test was used.For comparison of the BMD values before and after bisphosphonate treatment, the paired t-test was used.A pvalue <0.05 was regarded to be statistically significant.

| Ethics
The study was conducted ethically in accordance with the World Medical Association Declaration of Helsinki.No written informed consent was obtained from participants because the study was retrospective and registry-based and did not involve contacting the patients or any additional examinations as approved by the research committee of Helsinki University Hospital (approval number HUS185/2018).

| Patient characteristics
The mean age of the 25 patients at the diagnosis of the primary cancer was 8.4 years (range 0.8-16.4).Two patients, who had previously experienced primary cancer (Hodgkin disease and ALL), later developed secondary acute myeloid leukaemia (AML) where the age of diagnosis of the secondary cancer was 11 and 17 years, respectively.Eleven patients (44%), ALL in 10 and lymphoma in one, had been treated with HSCT, one of which had a history of both autologous and allogenic HSCT.Three had received highdose chemotherapy with autologous HSCT.Sixteen had received radiotherapy-11 as total-body irradiation (TBI) prior to allogeneic HSCT.Altogether 21 of the 25 patients were alive at the time of our retrospective analysis.The remaining four had died of the disease or treatment complications.Details of the patients' diagnoses and other characteristics are shown in supplements (Table S1).
The mean age at the diagnosis of the bone metabolic complication was 11.6 years (range 2.2-19.4),averaging 3.3 years after the primary cancer diagnosis (median 4.9; range 0-12 years).The time interval was not only longest among the transplanted subjects and those with secondary AML but also with one patient treated for hepatoblastoma with liver transplantation in early childhood (Table S1).

| Bisphosphonate treatment
The most common indication for the bisphosphonate treatment was osteoporosis with vertebral compression fractures (n = 13), osteonecrosis (n = 6) or osteoporosis together with peripheral fractures or osteonecrosis (n = 4).In two patients, the original indication was severe hypercalcaemia.In total, nine of the bisphosphonate-treated patients presented with osteonecrosis, and 14 with vertebral compression fractures.In 11 cases, the bone complication had developed after allogeneic HSCT, including TBI as conditioning.
Seven (28%) of the 25 patients were treated with bisphosphonates during their active cancer treatment.The other 18 received their bisphosphonate treatment after completing the cancer therapy.The delay between the diagnosis of the bone complication and bisphosphonate treatment was, on average, 13 months (range 0-76 months).The most employed bisphosphonate was intravenous pamidronate in 19 patients.Zoledronic acid infusions were given to five (in two as the primary bisphosphonate; in three as a secondary bisphosphonate after pamidronate), per oral alendronate to four (in one as maintenance after pamidronate), and per oral ibandronate and clodronate (as maintenance after pamidronate) both to two patients.The duration of the bisphosphonate treatment varied between 2 weeks for one patient with hypercalcaemia and 5 years for another with severe osteoporosis and hip osteonecrosis.In most cases (14 of the 21 with exact data available), the bisphosphonate treatment was given for up to 2 years.
The most used schedule for the bisphosphonate treatment was IV pamidronate up to 9 mg/m 2 per year, given for 1 to 3 consecutive days every 1 to 4 months, depending on the disease severity, treatment phase and other clinical characteristics.The dose and infusion intervals were individually tailored, and the total annual dose was often lower than the maximal dose.In two patients, bisphosphonate treatment was discontinued for 1-2 years and then restarted.In four patients, IV pamidronate was later changed to zoledronic acid or peroral bisphosphonate regimen.Zoledronic acid was mostly given 0.1 mg/kg/year, divided into two infusions per year.Patients received peroral calcium and vitamin D supplementation during the entire bisphosphonate treatment period; the dose of calcium supplementation was doubled during the week preceding infusion and during the infusion.Parameters of calciumphosphate metabolism and kidney function were followed by blood samples before and after each bisphosphonate administration in all patients.Additional medications, such as antiinflammatory medications, and therapies, like physiotherapy, were determined on an individual basis.Hypocalcaemia was seen in 3/7 patients in the former group and in 3/18 in the latter (p = 0.30).Bone pain was related to the bisphosphonate infusions in 5/18 treated after cancer treatment and one treated during the cancer treatment (p = 0.64).

Data on the clinical outcome of the patients' bone complication(s)
were available for 23 patients, while two died soon after the onset of the bisphosphonate treatment, of causes unrelated to bisphosphonate, and the outcome could not be evaluated.In the two patients receiving bisphosphonate for hypercalcaemia, the increased calcium levels quickly normalised with the treatment.Both later developed osteoporosis during their ALL treatment.One also presented with vertebral compression fractures, and the bisphosphonate treatment was recommenced later.Among the six patients with osteonecrosis as the main indication for bisphosphonates, one with severe osteonecrosis required hip arthroplasty after bisphosphonate treatment, while the pain was relieved for the others during the bisphosphonate treatment and follow up.Compared to the time of diagnosis of the metabolic bone complication, the symptoms and/or findings of the complication were completely alleviated (category 1) in five (20%) and ameliorated (category 2) in 12 (48%) of our 25 patients during the bisphosphonate treatment and by the end of the follow-up.In the remaining six patients (24%), the symptoms or findings of the bone complication did not improve (category 3).None of those with vertebral compression fractures developed new compression fractures during or after the bisphosphonate treatment.
The BMD was measured by DXA both before and after the bisphosphonate treatment in lumbar spine in 20 and femoral neck in 19 patients (Table 1).A clinically significant improvement (at least one Z-score unit) was seen in 10/20 in lumbar spine BMD and in 8/19 in femoral neck BMD, while among the other patients the BMD scores remained at the same level or increased only slightly.
The mean lumbar spine BMD Z-score was −1.17 before and −0.07 after the bisphosphonate treatment (n = 20, p < 0.001), while those for femoral neck were −1.75 and −0.37, respectively (n = 19, p = 0.023).The bisphosphonate treatment was well tolerated and safe with no severe complications in any of our patients.Expectedly, mild hypocalcaemia and hypophosphatemia were the most common side effects, all clinically asymptomatic.Importantly, we observed no dose limiting renal toxicity even after allogeneic or autologous HSCT, or during active chemotherapy.

TA B L E 1
At our institution, ALL and Hodgkin disease were the most common primary cancer diagnoses in the bisphosphonate-treated paediatric patients.Almost half of the bone complications developed after allogeneic HSCT.most indication for bisphosphonate treatment was osteoporosis with more than half of the patients suffering vertebral compression fractures.This is not surprising considering the high incidence of vertebral fractures in paediatric ALL 7 and low BMD in patients with a history of allogeneic HSCT. 12As for the allogeneic HSCT, an increased risk for skeletal complications is mostly linked to glucocorticoids used to treat graft-versus-host disease.
In our cohort, the BMD at the lumbar spine was significantly improved with the bisphosphonate treatment.In most of the patients with vertebral fractures, the BMD was normalised during the follow-up, and no new vertebral compressions developed in any of the patients.Nevertheless, the effect of the bisphosphonate treatment is impossible to verify with the current study setting, as the skeletal complications follow their natural course with a high rate of spontaneous healing in children.In a large Canadian cohort of ALL patients, the risk factors for vertebral fractures during and after leukaemia treatment were vertebral fracture at diagnosis, lower BMD, and higher glucocorticoid exposure.Incomplete vertebral reshaping was more common in older children. 7The bone mineral deficit persists long term after childhood HSCT, as well summarised in a review by

Mostoufi-Moab and Ward 2
A recent systematic review investigating bisphosphonate therapy in ALL-related vertebral fractures found only two previous studies eligible for evaluation, neither of them controlled.The authors concluded that there was insufficient evidence to support bisphosphonate treatment in this patient group, but well-designed, controlled studies were needed. 13A previous study compared bisphosphonate treatment together with calcium and vitamin D supplementation to supplementation only for osteoporosis in the setting of HSCT with chronic graft-versus-host disease in children.They showed that the annual BMD improvement was better in the bisphosphonate treatment arm. 14e idea of using bisphosphonate for osteonecrosis is to prevent the progression of the osteonecrosis lesions and help in managing the pain.In six of our patients, the main indication for bisphosphonate was severe, painful osteonecrosis.One patient with joint involvement later went through a hip arthroplasty.In all the other patients, pain was relieved with the bisphosphonate treatment.In previous studies on ALL children presenting with osteonecrosis, bisphosphonate treatment was able to alleviate the osteonecrosisrelated pain. 9,10There is no evidence, however, for an effect on the healing of the osteonecrosis per se in this patient group.Our previous study in children with ALL-related osteonecrosis also indicates that bisphosphonate treatment does not help if it has already led to joint damage. 15The effect of bisphosphonates on the healing process of osteonecrosis lesions cannot be confirmed with the current study setting.Whether the bisphosphonates can prevent the osteonecrosis lesions from progressing joint collapse remains unanswered.Interestingly, recent studies have linked osteonecrosis with low BMD in ALL children. 16,17This connection would further justify bisphosphonate treatment in osteonecrosis.There are only a few studies on the natural evolution of ALL-related osteonecrosis, and even less data on osteonecrosis in other paediatric cancers.In a prospective study in ALL children treated with a dexamethasonebased protocol, osteonecrosis was diagnosed in 6.1% of patients and symptoms completely resolved in 40% in the follow-up, while 40% had mild symptoms, and 20% severe symptoms interfering with daily activities. 3In a retrospective analysis of the Children's Cancer Group, the CCG-1882 protocol for ALL, osteonecrosis-related symptoms (pain and/or immobility) were defined as chronic in 84%, and 24% had undergone an orthopaedic procedure. 4 most patients, there was a marked delay between the diagnosis of the skeletal complication and application of the bisphosphonate treatment, as was also seen in our previous study on 10 children with ALL-related osteonecrosis treated with bisphosphonates. 15The potential beneficial effects of bisphosphonate treatment are probably compromised if administered too late when the active metabolic process in the bone has already ceased.
Recent preclinical studies have raised a concern about decreased antileukemic effect when bisphosphonates are given simultaneously with chemotherapy. 18,19First, Janke et al. showed in 2019 that zoledronic acid given continuously from the beginning of leukaemia treatment in a murine model led to prevention of osteonecrosis but also increased the risk of relapse. 18Another recent in vitro study showed that when using fivefold peak plasma concentrations of zoledronic acid or pamidronate, a slight or moderate antagonistic effect, respectively, on dexamethasone-induced cell death was observed. 19No antagonistic effect was shown for other drugs including daunorubicin, 6-mercaptopurine, pegylated asparaginase, and vincristine.
While awaiting confirmation studies, it is prudent to postpone the bisphosphonate therapy in ALL to after the patient has received the chemotherapy, including glucocorticoids or at least until negative measurable residual disease is shown in the bone marrow.Further studies are needed to elucidate the potential effects of bisphosphonate treatment on the antileukaemic effects of chemotherapy.

| Strengths and limitations
Our study involved the careful analysis of a non-selected group including all paediatric cancer patients treated with bisphosphonates for bone complications in a tertiary university clinic.Although all patients with bisphosphonate treatment were included, our cohort may be biased towards more severe cases of skeletal complications.
In the retrospective analysis, we unfortunately could not find all the relevant details from the patient records.For example, the BMD measurements could not be corrected for height, and we could not specify if BMD was biased by vertebral compression fractures in some of the patients.Because of the limited number of patients, the outcome factors could not be analysed separately for each bisphosphonate regimen, for different bisphosphonate treatment schedules, or for each bone complication.A major limitation is the lack of a control group because of the retrospective setting.

| CON CLUS IONS
Large, controlled studies are warranted before bisphosphonates can be routinely employed for bone complications in childhood cancer.
Still, bisphosphonates provide a safe therapeutic option for severe bone metabolic complications, also after HSCT and during active chemotherapy when administered after glucocorticoids in ALL.
Whether bisphosphonate treatment appears beneficial in the healing process of childhood cancer-related bone complications is important to investigate in controlled studies.

AUTH O R CO
Altogether eight patients presented with mild hypophosphatemia and six with mild hypocalcaemia within 1 week after the bisphosphonate administration.No patients showed any symptoms associated with hypophosphatemia or hypocalcaemia.Other adverse effects included bone pain (n = 6), fever (n = 5) and abdominal pain (n = 4), although no relation could be demonstrated between bisphosphonate administration and fever.Rare findings included increased aminotransferase levels in one and hypotension in another patient.Six patients had mildly compromised renal function parameters, probably related to other medications like cyclosporine.No renal toxicity related to bisphosphonate treatment was seen.The adverse symptoms and findings were mostly related to the first intravenous bisphosphonate infusion.To determine how the cancer treatment affected the bisphosphonate-related adverse effects, we analysed the results in two subgroups: patients treated during or after chemotherapy.Despite the lack of statistical significance, patients with concurrent chemotherapy tended to be more commonly present with hypophosphatemia and hypocalcaemia.Hypophosphatemia was registered in 4/7 of those treated with bisphosphonate during chemotherapy as compared to 4/18 of those treated after chemotherapy (p = 0.16).
Bone mineral density (BMD) measurements by dualenergy x-ray absorptiometry in the 25 child cancer patients with skeletal complications treated with bisphosphonates.
Note: Age-and sex matched Z-scores are shown for BMD measurements at the WB, FN and LS measured at the onset and at the end of BP treatment.Abbreviations: FN, femoral neck; LS, lumbar spine; WB, whole body.