The use of bone morphogenetic protein 7 in fracture non‐unions

Abstract

Objective:  To assess the rate of clinical and radiological union with the use of bone morphogenetic protein 7 (BMP‐7) in a range of fractures.

Methods:  This case series retrospectively reviews a series of 16 fracture non‐unions in 13 patients. These patients were treated with the commercially available BMP‐7. Time to radiological and clinical union was assessed by serial out‐patient follow‐up.

Results:  At nine months post‐surgery in which BMP‐7 was added, 12 of 16 non‐unions had achieved clinical and radiographic union. Three patients required repeat grafting. The mean time to union was 5.1 ± 1.6 months after the application of BMP‐7.

Conclusion:  The use of the osteo‐inductive agent, BMP‐7 results in good clinical and radiological outcomes which are not restricted to tibial non‐unions.

Introduction

Approximately 5.6 million fractures are recorded in the United States each year. While most unite satisfactorily, up to 10% do not and these require some additional management ¹ . Bone formation and fracture healing is a complex process relying on both mechanical and biological factors. It is regulated by effects on mesenchymal cell differentiation, which themselves can be influenced by local and systemic factors ² . When this process falters, delayed‐ and non‐unions occur. Autogenous bone graft with stable fixation remains the mainstay of treatment in these patients ³ . However, significant donor site morbidity has been reported to be associated with the harvesting of autogenous graft from the iliac crest ⁴ . More recently there has been much interest in the stimulation of bone growth, in particular the role of the bone morphogenetic proteins (BMPs).

The BMPs belong to a family of proteins that also includes the transforming growth factor‐beta (TGF‐β) and inhibins. In 1984, Urist et al. purified this group from demineralized bone ⁵ , having previously observed their, as then uncharacterized, mechanism of action ⁶ . Since their first description in this regard, BMPs have been seen to induce new bone formation in ectopic sites and therefore to have enormous potential in bone repair ⁵ . It is now known that their mechanism of action is by binding to the cells exterior and inducing transcription of a number of osteogenic genes ⁷ .

Individual BMPs are prominent at many sites during embryogenesis and are thought likely to be key regulators of early development and organogenesis ⁸ . In vertebrates, one of the functions of the BMPs is to induce formation of bone, cartilage, and connective tissues associated with the skeleton. This osteoinductive ability has led to the use of BMPs as therapeutic agents for creation of new bone, which is useful in treatment of skeletal injuries and diseases.

Since their first description, much evidence has become available as to the efficacy of the BMPs, in particular of BMP‐7 (also called osteogenic protein‐1), in promoting bone formation ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ .

In this case series, we reviewed a total of 13 consecutive patients with 16 fracture non‐unions at various sites who had all been treated with BMP‐7, and measured a number of clinical and radiographic outcomes. Our aim was to assess the use of BMP‐7 in a heterogeneous cohort of fractures and to compare our results to the existing literature.

Materials and methods

This is a retrospective, uncontrolled case series of sixteen established fracture non‐unions in thirteen patients in a large university teaching hospital. Patients were identified from a specific log of BMP‐7 use. All spinal fusions were excluded. All patients were treated with fixation (internal or external) as well as BMP‐7. The patients were treated between May 2001 and March 2007, during which time 6601 fractures were managed operatively in this unit.

The graft substitute used was BMP‐7, available commercially as Osigraft (Howmedica International, Limerick, Ireland). Each vial of Osigraft contained 3.5 mg of eptotermin alfa (a recombinant human BMP‐7) in bovine collagen (a bioresorbable scaffold).

Union was assessed clinically and radiographically. Clinical assessment was on the basis of pain or mobility at the fracture site and the ability to fully weight bear (in the case of lower limb fractures). As this was a retrospective study, clinical assessment of union could not be fully standardized and was therefore based on clinical reviews documented in the patient notes. Union was defined (retrospectively) as ability to weight bear with no, mild or moderate pain at the fracture site, no fracture site mobility and no need for re‐operation.

Standard AP and lateral radiographs were used to assess radiological union. These were reviewed independently by both consultant orthopaedic surgeons and consultant radiologists. The criteria for radiological union were bridging callus at the fracture site on plain radiographs or the complete absence of visible fracture lines. In addition, all adverse events were recorded.

Results

Thirteen patients were chosen for this study based on a review on the use of BMP‐7 in a single orthopaedic unit. The mean age (mean ± standard deviation) of the study cohort was 62.1 ± 20.6 years (range, 29–90 years). There were seven female and six male patients. Four patients were current smokers and three were ex‐smokers. The mean inpatient length of stay was 10.8 ± 9.3 days. These details are summarized in Table 1.

Table 1.

Summary of patients' characteristics

Item	Amount
Sex
Female	7
Male	6
Age (years)	62.1 ± 20.6
Initial fracture site
Femur	6
Humerus	5
Tibia	2
Time from index procedure (month)	8.9 ± 6.1
Smoking status
Non	5
Current	4
Previous	2
Unknown	2

There were six humeral, two tibial and five femoral non‐unions initially. Although BMP‐7 is in fact only licensed for the treatment of tibial non‐unions, it was used off label for the other fractures. Three of the femoral non‐unions required repeat grafting with BMP‐7 and were re‐entered in the study. These three patients therefore underwent three procedures in total; a primary procedure and two subsequent procedures with BMP‐7 augmentation. These patients are discussed in detail below. There were three hypertrophic non‐unions and one loss of reduction. Although these may be described as mechanical failures rather than biological failures, they were included due to the retrospective nature of the study and the fact that the operating surgeon at the time judged a need for some form of biological augmentation. The remainder of the cases had atrophic non‐unions. One patient was treated with an Ilizarov frame, all others underwent internal fixation (plate fixation or intramedullary nail). Patients were followed up for between 4 and 24 months (mean, 8.3 ± 5.5 months). The mean time from initial treatment to surgery with BMP‐7 was 8.9 ± 6.1 months. The patients had a mean number of post‐operative outpatient clinic visits of 5.9 ± 4.0.

At 9 months post‐operation, 75% (12 of 16 non‐unions) had achieved a satisfactory radiological and clinical outcome (i.e. significant callus formation and full weight bearing with minimal/no pain). The mean time to observed union was an average of 5.1 ± 1.6 months.

Four of the original 13 patients treated with BMP‐7 did not go on to union. Of these, three underwent repeated grafting with BMP‐7. All of these three patients had femoral fractures, two subtrochanteric and one midshaft. One was a 76 year‐old non‐smoking woman with a sub‐trochanteric femoral fracture, which had initially been treated with a dynamic condylar screw (DCS—Synthes, Solothurn, Switzerland). After a period of eight months she had revision surgery with a second DCS, which was augmented with the addition of BMP‐7. After a further seven months and loss of position she was treated with a proximal femoral nail (PFN—Synthes) and further BMP‐7. She went on to achieve a good clinical result and radiological evidence of union.

The second patient in this group was a 90 year‐old non‐smoking woman, also with a subtrochanteric fracture. Her original dynamic hip screw (DHS—Synthes) was revised to an angled blade plate (Synthes) combined with BMP‐7 following a period of six months. After a further five months she had a repeated blade plate procedure and further BMP‐7 grafting. She went on to clinical and radiological union.

The third patient in the re‐operation group was a 76 year‐old female ex‐smoker with a comminuted sub‐trochanteric and mid‐shaft femoral fracture. She was initially treated with a DCS. After 10 months, at which stage union had clearly not manifested and the DCS plate had broken, her fracture was re‐plated and BMP‐7 was added. After an interval of a further 5 months without union, the new DCS plate having also broken, she had a further revision, again augmented with BMP‐7. Radiological and clinical union was established 8 months after the final procedure (1, 2, 3, 4, 5).

Figure 1.

Plain film of a broken DCS plate which had been applied 10 months previously in a 76 year‐old woman with comminuted sub‐trochanteric and diaphyseal femoral fractures.

Figure 2.

Intra‐operative check film after re‐fixation with DCS and use of BMP‐7.

Figure 3.

Plain film showing failure of second fixation at 5 months post‐operation.

Figure 4.

Post‐operative plain film showing new fixation. Osteogenic protein‐1 was again applied to the fracture site.

Figure 5.

Plain film showing union at the fracture site 8 months after the final procedure.

The fourth patient who did not progress to union had a diaphyseal tibial non‐union treated with BMP‐7. This patient died of unrelated causes seven months post‐operatively. Evidence of union had not been documented at her last review, 4 months post‐operatively.

There were no adverse reactions that could be attributed to the use of BMP. No cases of superficial or deep infection were documented. There was no incidence of wound dehiscence.

Discussion

This series supports the role of BMP‐7 in non‐unions at various sites. Notably, all humeral fractures progressed to satisfactory union. Although three femoral fractures required re‐operation, all cases eventually united. This study does not show superiority over traditional autologous grafting but rather presents BMP‐7 stimulation as a viable alternative.

The use of BMPs, in particular BMP‐7, either alone or with autogenous bone graft, produces good clinical and radiographic results in the majority of non‐unions ¹⁰ , ¹¹ , ¹² . What remains unclear is whether or not these agents are more efficient than the traditionally utilized autograft, or indeed, in combination with an autograft.

The largest trial available to date remains the prospective randomized control trial published by Friedlaender et al. in 2001 ⁴ . This study of 122 patients with tibial fractures randomized to either autograft or BMP‐7 showed encouraging results for BMP‐7, with no statistically significant difference in rates of union between the two groups. It is worth noting that 20% of patients treated with autogenous graft had chronic donor site (iliac crest) pain. The authors does accept that rates of morbidity may be lower for alternative donor sites, such as the intramedullary canal ¹⁶ . The rate of clinical union was slightly higher for both groups (81% for BMP‐7, 85% for autograft) at the primary end‐point (nine months) than in our sample, although not all patients were followed for nine months. In the BMP‐7 group, 75% of patients achieved radiological union in this time, a figure identical to our results.

In 2005, Bong et al. published a prospective study of 23 consecutive humeral non‐unions ¹⁷ . All patients in this study received an autograft or allograft as well as BMP‐7. All patients in this trial went on to union with a mean time of 144.3 days, slightly shorter than the mean 5.1 months recorded in our group.

In 2002, a randomized trial of 14 patients with fresh tibial fractures treated with either an external fixator alone, or in conjunction with BMP‐7, was published ¹⁸ . However, the study group was very small and failed to show any conclusive advantage in either group.

More recently, Garrison et al. produced a systematic review of the clinical effectiveness and cost‐effectiveness of the use of BMP in both non‐unions and spinal fusions ¹⁹ . While the review found fault with much of the published data advocating the use of BMPs, it did confirm that they result in at least equal rates of union to autograft in cases of non‐union, and that they shorten operating times. It did however, come to the conclusion that the likelihood of its use being cost‐effective was small (6.4%).

The major advantage of BMPs would appear to be the reduction in donor site morbidity, an advantage lost when the technique is combined with autogenous graft, as is the case in many published series. This advantage needs to be offset against the significant cost of their use. However, it has been shown that the addition of osteogenic proteins to autografts (and allografts) induces more rapid and more abundant new‐bone formation ¹⁵ .

One of the perceived limitations of BMP‐7 usage is the seemingly prohibitive cost. A recent cost‐analysis by Dahabreh et al. showed that, when all factors were included, the increased cost compared to autologous iliac crest bone grafting was only 6.7% ²⁰ . This figure reflects the fact that the bone graft group in that study had a higher rate of re‐operation and an increased length of inpatient admission.

While this case series does not contain a control group, is relatively small, and is heterogeneous with respect to the fracture site and technique of fracture fixation, it nevertheless offers a strong indication of the efficacy of BMP‐7 in difficult clinical situations. However, there remains a lack of long‐term follow up for BMP‐7 in humans and the area clearly requires further research. From this series we conclude that BMP‐7 is an effective and safe, in the short‐term, adjunct to the treatment of fracture non‐union at a range of sites and its use should not be limited to tibial fractures.

Disclosure

No funding or donations were received by any of the authors or their families in relation to this work.