The authors have no conflict of interest.
Delayed Osteotomy but Not Fracture Healing in Pediatric Osteogenesis Imperfecta Patients Receiving Pamidronate†
Article first published online: 23 AUG 2004
Copyright © 2004 ASBMR
Journal of Bone and Mineral Research
Volume 19, Issue 11, pages 1779–1786, November 2004
How to Cite
Munns, C. F., Rauch, F., Zeitlin, L., Fassier, F. and Glorieux, F. H. (2004), Delayed Osteotomy but Not Fracture Healing in Pediatric Osteogenesis Imperfecta Patients Receiving Pamidronate. J Bone Miner Res, 19: 1779–1786. doi: 10.1359/JBMR.040814
- Issue published online: 2 DEC 2009
- Article first published online: 23 AUG 2004
- Manuscript Accepted: 24 JUN 2004
- Manuscript Revised: 27 MAY 2004
- Manuscript Received: 16 FEB 2004
- fracture healing;
- intramedullary rodding;
- osteogenesis imperfecta;
This study evaluated factors influencing fracture (n = 197) and osteotomy (n = 200) healing in children with moderate to severe OI. Pamidronate treatment was associated with delayed healing after osteotomy, but not after fracture. The data suggest that both pamidronate and mechanical factors influence bone healing in this cohort.
Introduction: Intravenous pamidronate is widely used to treat children with moderate to severe osteogenesis imperfecta (OI). However, the effect of this treatment on bone healing is not well characterized. We therefore retrospectively analyzed the healing of lower limb fractures and osteotomies in children with moderate to severe OI, both before and after the start of pamidronate treatment.
Materials and Methods: Bone healing was evaluated on standard radiographs after 197 lower limb fractures (132 femur and 65 tibia) in 82 patients (age at fracture, 0.0–19.9 years) and 200 intramedullary rodding procedures in 79 patients (age at surgery, 1.2–19.8 years). Delayed healing was diagnosed when a fracture or osteotomy line was at least partially visible 12 months after the event.
Results: Delayed fracture healing was observed more frequently during than before pamidronate treatment. However, the effect of pamidronate was no longer significant when age differences were taken into account (odds ratio [OR], 1.76; 95% CI, 0.61–5.10). Better mobility status was a strong independent predictor of delayed healing after fractures that occurred during pamidronate treatment. After osteotomies, delayed healing was more frequent when pamidronate had been started before surgery (OR, 7.29; 95% CI, 2.62–20.3), and this effect persisted after adjustment for multiple confounders. During pamidronate treatment, older age (OR per year of age, 1.25; 95% CI, 1.06–1.47) and osteotomy of the tibia (OR, 3.51; 95% CI, 1.57–7.82) were independent predictors of delayed healing.
Conclusions: This study suggests that pamidronate therapy is associated with delayed healing of osteotomy sites after intramedullary rodding procedures. Better mobility status, but not pamidronate treatment, seems to be predictive of delayed healing after fractures.
Osteogenesis imperfecta (OI) is a heritable bone disorder characterized by bone fragility, short stature, and bone deformities.(1) Intravenous infusions of pamidronate, a nitrogen-containing bisphosphonate, have been reported to be beneficial in children with moderate to severe OI, resulting in increased vertebral bone mass and size, muscle force, and growth rate, and reduced bone pain and fracture rates.(1)
The primary mechanism of action of bisphosphonates is to suppress osteoclastic bone resorption by reducing the number and activity of osteoclasts on the bone surface.(2) In children with OI, histomorphometric analysis of iliac bone biopsy specimens has shown that cyclical pamidronate markedly reduces trabecular remodeling activity, reducing both bone resorption and formation.(3)
Fracture healing requires coordinated bone resorption and formation.(4) Because bisphosphonates decrease both of these parameters and have a half-life of many years, concern has been raised that these medications may have a negative effect on fracture and osteotomy healing.(5), (6) A number of animal experiments have shown that bisphosphonates can affect bone healing, resulting in increased callus size and a retardation of callus remodeling.(6–10) In contrast, studies in adults have found no evidence that bisphosphonates interfere with fracture healing in a clinically relevant manner.(6), (11), (12)
The effect of bisphosphonates on fracture healing in children with moderate and severe OI has not been studied systematically, although various investigators have reported the impression that fracture healing proceeded normally during intravenous pamidronate therapy.(13–15) To address this issue in detail, we evaluated the incidence of delayed healing of femur and tibia fractures and osteotomy sites in children with moderate and severe OI, both before and after the start of cyclical intravenous pamidronate therapy.
MATERIALS AND METHODS
This study was comprised of 131 patients (65 girls and 66 boys) with moderate to severe OI that were treated at the Shriners Hospital for Children in Montreal between 1984 and 2003. The distribution of OI types was as follows: type I, N = 15; type III, N = 39; type IV, N = 58; type V, N = 8; type VI, N = 7; type VII, N = 3; unclassified, N = 1. After 1992, an increasing number of OI patients received cyclical intravenous pamidronate therapy. Pamidronate was initiated in patients who had long bone deformities, had sustained more than two fractures per year during the previous 2 years, or had vertebral compression fractures.
Subsequent to the initial radiograph, which was taken at the time of fracture or osteotomy, follow-up radiographs were taken as clinically indicated to assess healing and/or annually as part of a skeletal survey to evaluate pamidronate therapy.
The dates of femoral and tibial fractures were obtained from patient records. Only fractures confirmed radiographically were included in this study. Most fracture management occurred at the child's primary health care facility.
Femur and tibia fractures were included in this analysis if they occurred either >12 months before (to eliminate those fractures where pamidronate was started during the 12-month evaluation period) or at any time after the start of pamidronate therapy. Radiographs from the first 10 subjects were evaluated by all authors. Subsequent radiographs were evaluated by a single observer (CM). Radiographs taken after 390 fractures in 112 patients were assessed (Fig. 1). In 193 of these fractures, radiographic follow-up was insufficient to judge whether or not the fracture site had completely healed within 12 months. Thus, the analysis was comprised of a total of 197 fractures (Fig. 1; Table 1).
The dates of femoral and tibial rodding procedures were obtained from the patients' medical records. Femoral and tibial osteotomies were performed by orthopedic surgeons of the Shriners Hospital for Children, Montreal, to facilitate the straightening of lower limb deformities for intramedullary rod placement. Femoral and tibial osteotomies were performed by standard techniques(16) using an oscillating power saw. Fibula osteoclasis was performed by applying direct force to the bone after completion of the tibial osteotomy.
Intramedullary rodding procedures of the lower limbs were included in this analysis if they occurred either >12 months before or at any time after the start of pamidronate therapy. Radiographs from the first 10 subjects were evaluated by all authors. Subsequent radiographs were evaluated by a single observer (CM). Radiographs after 317 interventions were assessed (Fig. 2). In 117 interventions, radiographic follow-up was insufficient to judge whether or not osteotomy sites had completely healed within 12 months. Thus, the final analysis was comprised of a total of 200 intramedullary rodding surgeries (Fig. 2; Table 1).
Delayed healing was diagnosed when a fracture or osteotomy line was at least partially visible 12 months after the event (Figs. 3 and 4). This definition corresponds to grade 3 of the Hammer classification of fracture healing.(17)
The long-term outcome of the fractures and osteotomies with delayed healing at 12 months was assessed by evaluating subsequent radiographs for healing, angulation, or fracture through the site or the requirement for surgical intervention (Fig. 5).
At each pamidronate treatment cycle, physiotherapists experienced in the care of children with OI evaluated patient mobility on a four-point scale as follows: 0 (does not walk), 1 (able to walk during therapy sessions only), 2 (walking only within the house), and 3 (able to walk within the community).(18) For this analysis, the maximal mobility score between 90 and 365 days after the fracture or osteotomy event was used.
The period of immobilization was minimized after fracture or osteotomy to reduce disuse osteoporosis. If there was radiological evidence of callus formation, active physiotherapy and weight bearing with the aid of orthoses were commenced 5–8 weeks after a fracture and 3 weeks after intramedullary rod placement.
Pamidronate was administered intravenously on 3 consecutive days in all patients, as described.(1) The total annual pamidronate dose was 9 mg/kg in all children, although timing and dosage of the 3-day cycles varied with age. Children <2 years of age received 0.25 mg/kg on the first day of the first cycle, 0.5 mg/kg on days 2 and 3 of the first cycle, and 0.5 mg/kg daily on all 3 days of subsequent cycles. Cycles were repeated every 2 months. Children from 2 to 3 years of age received 0.38 mg/kg on the first day of the first cycle, 0.75 mg/kg on days 2 and 3 of the first cycle, and 0.75 mg/kg daily on all 3 days of subsequent cycles. Cycles were repeated every 3 months. Above 3 years of age, the first 3-day cycle consisted of a dose of 0.5 mg/kg on the first day and 1 mg/kg on days 2 and 3. In subsequent cycles, the dose was 1 mg/kg daily for 3 consecutive days. Cycles were repeated every 4 months. Each dose was diluted in 0.9% saline solution and administered slowly over 4 h.
All patients were advised to maintain adequate calcium intake according to the recommended daily allowance at all times before and during pamidronate treatment. All patients underwent physiotherapy and occupational therapy evaluation and support, including exercises and design of special devices for transportation and sitting.
Anthropometric and biochemical measurements
Weight and height measurements were converted to age- and sex-specific z scores based on reference data published by the National Center for Health Statistics.(19)
Urine creatinine concentration was determined colorimetrically. Urinary cross-linked N-telopeptides of type I collagen (NTX) were quantified by enzyme-linked immunoabsorbent assay (Osteomark; Ostex, Seattle, WA, USA) using the second void sample of the morning. Results for urinary NTX/creatinine ratios in OI patients were expressed as a percentage of age-specific mean values using published reference data.(20) Patients were fasting at the time of urine sampling.
Radiographs were performed using the Siemens Multix H radiological unit (Siemens AG, Erlangen, Germany) onto Kodak medical film (Eastman Kodak, Rochester, NY, USA). Where possible, both the anterior-posterior and lateral radiographs were taken with the patient in the standing position. Unless requested otherwise by the treating physician, femora and tibias were included on a single cassette (lower extremity radiograph).
Logistic regression analysis was used to evaluate the relationship between clinical characteristics and the presence of delayed fracture or osteotomy healing. Results were expressed as odds ratios (ORs) with 95% CI. The effect of potential predictor variables was initially assessed in univariate models and in multivariate models (all variables entered). To assess the effect of pamidronate therapy, pamidronate treatment status was included as a dichotomous variable (no/yes). OI types V, VI, and VII were coded as one group, because the number of subjects of individual groups was insufficient for the analysis. OI type III was selected as the reference category, because this group of patients constitutes one end of the spectrum of disease severity.
Healing of fractures and osteotomy sites might be influenced by the duration and timing of prior and subsequent pamidronate treatment, patient mobility, and the suppression of bone turnover during therapy (as expressed by urinary NTX levels). Therefore, a second logistic regression analysis including these variables was performed, including only fractures or osteotomies that had occurred while patients were receiving pamidronate treatment. Differences in the incidence of complications after delayed healing were tested for significance using the χ2 test only, because the number of events was too small for multivariate analysis.
All tests were two-tailed. In the logistic regression analysis, the association with delayed healing was considered to be significant, when the 95% CI of an independent variable did not include 1.00. Otherwise, p values <0.05 were considered significant. These calculations were performed using the SPSS software, version 12.0 for Windows (SPSS, Chicago, IL, USA).
Forty-two fractures with informative follow-up occurred >1 year before the start of pamidronate therapy, and 155 fractures were sustained after the start of pamidronate treatment. Fractures that occurred before pamidronate therapy on average happened at a younger age than fractures that were sustained during pamidronate treatment (Table 1).
Twelve months after fracture, delayed healing was observed in 47 of these 197 fractures (Table 2). According to univariate analysis, delayed fracture healing was associated with increasing age and was significantly more frequent during pamidronate treatment. However, multivariate analysis indicated that pamidronate treatment was no longer significantly associated with delayed fracture healing when differences in age were accounted for (Table 2). Neither univariate nor multivariate analyses yielded significant associations of delayed healing with gender, fractured bone, OI type, height, and weight.
Next we attempted to elucidate factors that were associated with delayed fracture healing during pamidronate treatment. Compared with fractures that healed normally, fractures with delayed healing occurred at an older age, after a longer duration of pamidronate treatment, and after a longer interval following the most recent pamidronate treatment (Table 3). Although all of these factors were significantly associated with delayed healing in the univariate analysis, only age remained a significant predictor of delayed healing in the multivariate analysis (Table 3).
Mobility scores were available after 136 fracture events during pamidronate treatment. Delayed healing was seen after 7 of 62 (11%) fractures in patients with a mobility score of 0 or 1 and after 31 of 74 (42%) fractures in patients with a mobility score of 2 or 3 (OR, 5.66; 95% CI, 2.28, 14.1). When mobility score was added to the multivariate regression model shown in Table 3, it emerged as the strongest predictor of delayed fracture healing (Table 4). In addition, male gender and lower height z scores became significant predictors of delayed healing. However, time since the previous or until the subsequent pamidronate cycle was not significantly associated with delayed fracture healing, whether or not mobility score was taken into account.
To assess the clinical consequences of delayed fracture healing, information at the last available follow-up examination was considered. The median follow-up period was 1.9 years (range, 1.3–2.6 years) after fractures that occurred before pamidronate treatment and 1.7 years (range, 1.0–4.7 years) after fractures that occurred during pamidronate treatment. As shown in Fig. 1, refracture, angulation deformity, and eventually intramedullary rodding surgery occurred in both groups. The incidence of these complications was similar after fractures that occurred before pamidronate was started and after those that occurred during pamidronate treatment (p > 0.2 in each case; χ2 test).
Thirty-eight intramedullary rodding procedures with informative follow-up occurred >1 year before the start of pamidronate therapy, and 162 such interventions were done after the start of pamidronate treatment (Table 1). Rodding procedures performed before pamidronate therapy occurred at a slightly older age than interventions done during pamidronate treatment.
Twelve months after intramedullary rodding, delayed healing was observed after 103 of the 200 interventions where radiographic follow-up was sufficient for analysis (Table 5).
In univariate analysis, delayed osteotomy healing was significantly more frequent in patients who had received pamidronate, in patients with OI type IV (compared with OI type III), and in osteotomies involving the tibia. Delayed osteotomy healing was also associated with height and weight z scores (Table 5). In the multivariate analysis, only pamidronate treatment and osteotomy of the tibia remained significant predictors of delayed healing (Table 5).
Next we attempted to elucidate factors that were associated with delayed osteotomy healing during pamidronate treatment. Compared with osteotomies that healed normally, osteotomies with subsequent delayed healing were performed more frequently at the tibia, in OI type IV patients, at an older age, and at a higher height z score (Table 6). However, the multivariate analysis indicated that only tibia osteotomy and age remained significant predictors of delayed healing. Mobility score (available after 141 osteotomy procedures performed during pamidronate treatment) was not significantly associated with delayed osteotomy healing, neither in the univariate nor in the multivariate analysis (data not shown).
To assess the clinical consequences of delayed osteotomy healing, information at the last available follow-up examination was taken into account. The median follow-up period was 2.9 years (range, 1.3–7.4 years) in osteotomies that were performed before pamidronate treatment and 3.0 years (range, 1.0–5.9 years) in osteotomies that occurred during pamidronate treatment. In the former group, no complications were noted, whereas in osteotomies that were performed during pamidronate treatment, fractures through the osteotomy site and angulation of >10° occurred in 10 and 21 cases, respectively (Figs. 2 and 5). Further surgical intervention was performed in 20 of these unhealed osteotomy sites. This difference in complication rate between osteotomies before and during pamidronate treatment was statistically significant with regard to angulation of >10° and further surgical intervention (p = 0.02, χ2 test), but not for refracture (p = 0.12).
To correct deformities of the lower extremity a tibial osteotomy has to be combined with an osteoclasis (manual breaking) of the fibula. It was noted that in the 44 lower extremities with unhealed tibial osteotomies at last follow-up, only 5 of the adjacent fibulas remained unhealed (p < 0.001 for difference between the two bones; χ2 test).
This study suggests that the administration of cyclical intravenous pamidronate in children with OI is associated with delayed bone healing after osteotomy but not after fracture. Delayed osteotomy healing led to clinically relevant complications, such as angulation deformity and fracture, which necessitated further surgical intervention.
A number of previous studies have examined fracture and osteotomy healing in children with OI not receiving medical therapy.(21–23) However, the findings of these reports cannot be directly compared with the present observations, because they differ considerably with regard to case selection and surgical technique. In addition, comparisons between studies are complicated by the fact that there is no generally accepted definition of delayed fracture union within the orthopedic surgical community.(24) The criteria used in this study (total or partial persistence of a fracture or osteotomy line 12 months after the event) represent a conservative approach to the diagnosis of delayed bone healing.
It may come as a surprise that pamidronate should have a differential effect on osteotomy and fracture healing, as was observed in this study. However, similar observations have been made in both animal and clinical studies. In fracture healing studies, bisphosphonates did not have a major effect irrespective of the type of bisphosphonate and dosing regimen used.(11), (12), (25), (26) In contrast, bisphosphonate application after osteotomy has been reported to interfere with callus remodeling and to lead to the persistence of the osteotomy line on radiographs.(7), (9), (10) Our data showing a difference between the healing of fibulas and tibias after tibial rodding further support a discrepancy between fracture and osteotomy repair. Fibulas, which are broken manually by osteoclasis when lower limb deformities are corrected, had a much lower rate of delayed healing than adjacent tibias, which undergo osteotomy with an oscillating saw and receive the intramedullary rod.
The reasons for the discrepancy between fracture and osteotomy healing are not clear. Possibly, the presence of the rod may affect the healing process when at the same time the bone is exposed to pamidronate. It is also possible that the use of an oscillating saw for osteotomies interferes with bone healing by cauterizing bone ends, whereas osteoclasis neither disrupts the periosteum nor alters the bone ends. In addition, the more rapid healing of the fibula may have further delayed tibial osteotomy healing by acting as a splint, thus disturbing apposition of the tibial bone ends. Thus, the etiology of delayed osteotomy healing after the start of pamidronate therapy may be multifactorial, with interplay between drug, mechanical, and surgical factors. This topic clearly warrants further investigation.
In this study, neither the duration of pamidronate therapy nor the timing of drug administration in relation to the osteotomy influenced outcome. This observation is in contrast to animal studies,(8), (9) which suggested that there is a critical period after an osteotomy when bisphosphonates are more likely to delay bone healing. Our observations may indicate that the timing of pamidronate administration does not play a significant role in osteotomy healing in children with OI. However, it must be noted that, in this study, pamidronate was given every 4 months, regardless of surgical interventions. It is therefore also possible that any administration of pamidronate within the 4-month cycle interval is sufficient to interfere with healing.
With regard to fracture healing, we found that better mobility status increased the risk for delayed healing during pamidronate treatment, whereas pamidronate by itself had no independent effect. This suggests that weight bearing on abnormal OI bone adversely affected healing.
Limitations of this study include the fact that this is an uncontrolled observational trial, and therefore, comparisons are limited to historical controls. Details in clinical management other than medical treatment may therefore have differed between the pre-pamidronate and the pamidronate-treated groups. Also, follow-up radiographs were not obtained systematically, but mostly as required for clinical management. This may have introduced bias, because normal healing is less likely to lead to radiographic studies than delayed bone healing.
Despite these limitations, it seems prudent to devise strategies that minimize the chances of delayed healing in OI patients receiving pamidronate. A longer pamidronate-free period around the time of intramedullary rodding procedures may be required. Altering surgical and rehabilitative techniques by avoiding the use of oscillating saws for osteotomies and extending the non-weight-bearing period after a fracture/osteotomy might also prove beneficial in improving bone healing. However, it must be acknowledged that there are no data to show that any of these proposed measures are effective in reducing the risk of delayed healing.
In summary, this report suggests that pamidronate is associated with delayed osteotomy healing in children and adolescents with moderate to severe OI and that better mobility is associated with a higher risk of having a delay in fracture healing.
- 12004 Osteogenesis imperfecta. Lancet 363: 1377–1385.,
- 21999 Bisphosphonates: Pharmacology, mechanisms of action and clinical uses. Osteoporos Int 9 (Suppl 2): S66–S80., ,
- 32002 The effects of intravenous pamidronate on the bone tissue of children and adolescents with osteogenesis imperfecta. J Clin Invest 110: 1293–1299., , ,
- 41999 Growth factor regulation of fracture repair. J Bone Miner Res 14: 1805–1815., , ,
- 52001 Bone loss accompanying medical therapies. N Engl J Med 345: 989–991.
- 62001 Can bisphosphonates be given to patients with fractures? J Bone Miner Res 16: 437–440.
- 71994 Use of a bisphosphonate (pamidronate) to modulate fracture repair in ovine bone. Ann Oncol 5 (Suppl 7): S53–S55., , , ,
- 81999 Effect of bisphosphonate (incadronate) on fracture healing of long bones in rats. J Bone Miner Res 14: 969–979., , , , ,
- 92001 Long-term effect of incadronate disodium (YM-175) on fracture healing of femoral shaft in growing rats. J Bone Miner Res 16: 429–436., , , , , ,
- 102002 Raloxifene, estrogen, and alendronate affect the processes of fracture repair differently in ovariectomized rats. J Bone Miner Res 17: 2237–22346., , , , , , , , , ,
- 112000 The effect of alendronate on bone mass after distal forearm fracture. J Bone Miner Res 15: 586–593., , , ,
- 122002 Alendronate in the prevention of bone loss after a fracture of the lower leg. J Bone Miner Res 17: 2247–2255., , , , ,
- 131998 Cyclic administration of pamidronate in children with severe osteogenesis imperfecta. N Engl J Med 339: 947–952., , , , ,
- 142002 Pamidronate treatment of osteogenesis imperfecta—lack of correlation between clinical severity, age at onset of treatment, predicted collagen mutation and treatment response. J Pediatr Endocrinol Metab 15: 163–174.,
- 152002 Beneficial effect of long term intravenous bisphosphonate treatment of osteogenesis imperfecta. Arch Dis Child 86: 356–364.,
- 161991 Bone cutting. Clin Phys Physiol Meas 12: 1–19., , , , ,
- 171985 Accuracy of radiologic assessment of tibial shaft fracture union in humans. Clin Orthop 199: 233–238., ,
- 181981 Nonoperative treatment of osteogenesis imperfecta: Orthotic and mobility management. Clin Orthop 159: 111–122.
- 191979 Physical growth: National Center for Health Statistics percentiles. Am J Clin Nutr 32: 607–629., , , , ,
- 201994 Bone resorption rates in children monitored by the urinary assay of collagen type I cross-linked peptides. Bone 15: 31–34.,
- 211988 Non-union of fractures in children who have osteogenesis imperfecta. J Bone Joint Surg Am 70: 439–443., , ,
- 221988 Complications of intramedullary rods in osteogenesis imperfecta: Bailey-Dubow rods versus nonelongating rods. J Pediatr Orthop 8: 645–649., , ,
- 231992 Osteogenesis imperfecta: A clinical study of the first ten years of life. Calcif Tissue Int 50: 36–41., , , ,
- 242002 A lack of consensus in the assessment of fracture healing among orthopaedic surgeons. J Orthop Trauma 16: 562–566., , , , ,
- 251998 No adverse effects of clodronate on fracture healing in rats. Acta Orthop Scand 69: 532–536., , , ,
- 262003 Long-term administration of clodronate does not prevent fracture healing in rats. Clin Orthop 408: 268–278., , , , , , ,