A blinded comparison of the safety and efficacy of hot biopsy forceps electrocauterization and conventional snare polypectomy for diminutive colonic polypectomy in a porcine model.

BACKGROUND: Although linked with perforation, serositis, delayed bleeding, and incomplete resection, hot biopsy forceps electrocauterization (HBF) is still widely used for diminutive colonic polypectomy. OBJECTIVE: To evaluate the safety and efficacy of HBF in comparison with conventional snare polypectomy (CSP). DESIGN: Randomized, blinded, controlled trial. SETTING: Academic endoscopy unit. SUBJECTS: Ten swine. INTERVENTION: Eighty-two paired polypectomies (41 HBF, 41 CSP) of small, minimally elevated, artificial lesions. Standardized technique using coagulating current at 25 W. HBF: the tissue was avulsed after 1 to 2 seconds of current caused blanching of the artificial pedicle. CSP: the polyp was removed by snare diathermy. MAIN OUTCOME MEASUREMENTS: Histopathology of resected specimens and polypectomy sites in colectomy specimens at necropsy (lateral mucosal and depth of ulceration, necrosis and inflammation). RESULTS: Some (21%) of the HBF specimens were ablated and uninterpretable. All CSP specimens yielded interpretable specimens. Mucosal necrosis adjacent to HBF resection sites varied widely, between 1.5 and 9 mm (mean 5.7 mm, standard deviation ± 2). There was visible mucosa under the HBF ulcer in 14% of cases. The depth of necrosis in the colon wall was significantly different between the two techniques, with partial muscularis propria (MP) necrosis in 14 of 41 lesions (34%) with HBF, compared with 1 of 41 (2%) of CSP (P < .001), and full-thickness MP necrosis in 9 of 41 lesions (22%) with HBF, compared with 1 of 41 (2%) of CSP (P = .014). There was full-thickness MP inflammation in 13 of 41 lesions (32%) with HBF compared with 5 of 41 (12%) of CSP (P = .06). Transmural subserosal inflammation was seen in 13 of 41 lesions (32%) with HBF compared with 4 of 41 (10%) of CSP (P = .027). There was no relationship between visible lateral mucosal injury and depth of injury (rs = -0.07). LIMITATIONS: Animal study. CONCLUSION: Despite use of the standardized HBF technique, there is a wide range of lateral mucosal and deep thermal injury as well as residual target mucosa. HBF also results in a significantly greater depth of tissue injury, with a high proportion of transmural necrosis. Ensuring minimal blanching of the mucosa during the procedure does not protect from deep injury. In comparison to conventional snare polypectomy, HBF is imprecise, potentially ineffective, and hazardous.

Beyond the snare: technically accessible large en bloc colonic resection in the West: an animal study.

BACKGROUND: Endoscopic submucosal dissection (ESD) and circumferential submucosal incision endoscopic mucosal resection (CSI-EMR) are techniques for en bloc excision of large sessile colonic lesions. Our aims were to compare the efficacy, safety and learning curve of colonic hybrid knife (HK) ESD versus CSI-EMR for en bloc excision of 50 mm diameter hemi-circumferential artificial lesions in a porcine model. PATIENTS AND METHODS: Two separate 50 mm diameter areas of normal recto-sigmoid mucosa were marked out in each of ten pigs. One was excised with HK-ESD using succinylated gelatin (SG) submucosal injection. The other was isolated with CSI with the Insulated Tip Knife 2 followed by SG submucosal injection then EMR with a large snare. Euthanasia and colectomy was performed at 72 h followed by blinded histopathology assessment. RESULTS: En bloc excision rates were: HK-ESD 100% versus CSI-EMR 20% (P = 0.008). The mean number of resections per lesion was HK-ESD 1 versus CSI-EMR 3 (P = 0.001). The mean dimensions of the largest specimen per technique were HK-ESD 63 × 54 mm versus CSI-EMR 49 × 41 mm (P = 0.005). Procedure duration mean was HK-ESD 54 min versus CSI-EMR 22 min (P < 0.001). When procedure duration was adjusted for the size of the resected en bloc specimen, a statistically significant and accelerated learning effect was noted for HK-ESD (r = -0.83, P = 0.003). There were no perforations and no significant bleeding. CONCLUSIONS: HK-ESD with SG submucosal injection is superior to CSI-EMR for en bloc excision of 50 mm diameter lesions in a porcine model. The technique is rapidly learnt. This novel approach may lower the barrier to colonic ESD for Western endoscopists.

The correlation between time to skin grafting and hypertrophic scarring following an acute contact burn in a porcine model.

One sequelae of burn injury remains the development of hypertrophic scarring. This appears more likely when the healing has been prolonged. Early excision of deep dermal burns and subsequent split skin grafting (SSG) may provide a more favorable result. The optimal timing of grafting for deeper dermal burns remains controversial. This study sought to establish evidence for the optimal grafting time using a porcine model. Five Large White female pigs were exposed to four contact burn injuries for duration of 20 seconds at 92°C. Each site was randomized to a treatment arm: dressing only as the control, SSG day 3, SSG day 14, and SSG day 21. Burn wound biopsies were obtained at days 0, 3, 14, 21, and 99 after the burn injury, together with microbiological swabs. Digital photographs were taken to assess scarring using the Vancouver scar scale. All biopsies were subject to histological and immunohistochemical analysis. Vancouver scar scale scores and histopathological analysis indicated that areas grafted on day 3 had the least fibrosis and scarring (P = 0.031). There was a strong correlation between the histological evaluation of the degree of fibrosis and α-smooth muscle actin levels (r = .60, P = .014). A greater degree of fibrosis was observed in the presence of infection (P = .028). Sites grafted on day 3 consistently exhibited the best clinical and histological scar outcome. The increased fibrosis observed in delayed grafting may have been be related to progression of burn depth and infection. These results suggest that early grafting of deep dermal burns may be preferential.

Succinylated gelatin substantially increases en bloc resection size in colonic EMR: a randomized, blinded trial in a porcine model.

BACKGROUND: Succinylated gelatin (SG) is an inexpensive colloid that may combine ease of use with the advantages of a colloid to potentially increase EMR specimen size, leading to a higher rate of en bloc resection. OBJECTIVE: To evaluate the safety, efficacy, and impact on EMR specimen size of SG as a submucosal (s.m.) injectant in comparison with normal saline solution (NS). DESIGN: Randomized, blinded, controlled trial conducted with Animal Ethics Committee approval. SETTING: Academic hospital. SUBJECTS: Ten swine. INTERVENTIONS: Sixty EMRs (30 using SG vs 30 using NS as 3 paired experiments per animal) of the largest possible en bloc snare resection of normal colonic mucosa after s.m. injection of a fixed volume of either SG or NS. MAIN OUTCOME MEASUREMENTS: EMR specimen size, duration of s.m. cushion, duration of procedure, ratio of vertical elevation to lateral spread of injectant, ease of resection, adverse effects, perforation, histopathology of EMR sites in colectomy specimens at necropsy (for inflammatory cell content, depth of ulceration, and vascular or ischemic changes). RESULTS: The mean subject weight was 53 kg. The mean EMR specimen dimensions and surface area were significantly larger with SG (length 37 vs 31 mm, P = .031; width 32 vs 26 mm, P = .022; surface area 9.5 cm(2) vs 6.7 cm(2), P = .044, respectively). The median s.m. cushion duration was 60 minutes with SG versus 15 minutes with NS (P = .005). The median procedure duration with SG was 2.6 minutes vs 2.5 minutes with NS (P = .515). The ratio of vertical elevation to lateral spread of injectant (mean score on a 3-point scale) was 3 with SG versus 2 with NS (P = .228). Ease of resection score (mean score on a 10-point scale) was 8 with SG versus 7 with NS (P = .216). There were no systemic adverse effects, hypersensitivity reactions, or bleeding episodes. There were 2 perforations (treated with clips) with SG and 1 with NS (P = 1.0). Blinded histopathologist assessment of necropsy colectomy specimens did not identify any significant differences between SG and NS EMR sites. LIMITATIONS: Animal study. CONCLUSIONS: SG is safe and results in a 42% increased surface area for en bloc EMR. Given its other favorable properties, it represents a significant step toward defining the ideal EMR solution.

Intermittent PTH(1-34) does not increase union rates in open rat femoral fractures and exhibits attenuated anabolic effects compared to closed fractures.

Intermittent Parathyroid Hormone (PTH)((1-34)) has an established place in osteoporosis treatment, but also shows promising results in models of bone repair. Previous studies have been dominated by closed fracture models, where union is certain. One of the major clinical needs for anabolic therapies is the treatment of open and high energy fractures at risk of non-union. In the present study we therefore compared PTH((1-34)) treatment in models of both open and closed fractures. 108 male Wistar rats were randomly assigned to undergo standardized closed fractures or open osteotomies with periosteal stripping. 27 rats in each group were treated s.c. with PTH((1-34)) at a dose of 50 mug/kg 5 days a week, the other 27 receiving saline. Specimens were harvested at 6 weeks for mechanical testing (n=17) or histological analysis (n=10). In closed fractures, union by any definition was 100% in both PTH((1-34)) and saline groups at 6 weeks. In open fractures, the union rate was significantly lower (p<0.05), regardless of treatment. In open fractures the mechanically defined union rate was 10/16 (63%) in saline and 11/17 (65%) in PTH((1-34)) treated fractures. By histology, the union rate was 3/9 (33%) with saline and 5/10 (50%) with PTH((1-34)). Radiological union was seen in 13/25 (52%) for saline and 15/26 (58%) with PTH((1-34)). Open fractures were associated with decreases in bone mineral content (BMC) and volumetric bone mineral density (vBMD) on quantitative computerized tomography (QCT) analysis compared to closed fractures. PTH((1-34)) treatment in both models led to significant increases in callus BMC and volume as well as trabecular bone volume/total volume (BV/TV), as assessed histologically (p<0.01). In closed fractures, PTH((1-34)) had a robust effect on callus size and strength, with a 60% increase in peak torque (p<0.05). In the open fractures that united and could be tested, PTH((1-34)) treatment also increased peak torque by 49% compared to saline (p<0.05). In conclusion, intermittent PTH((1-34)) produced significant increases in callus size and strength in closed fractures, but failed to increase the rate of union in an open fracture model. In the open fractures that did unite, a muted response to PTH was seen compared to closed fractures. Further research is required to determine if PTH((1-34)) is an appropriate anabolic treatment for open fractures.

Delayed cooling of an acute scald contact burn injury in a porcine model: is it worthwhile?

The current Australia and New Zealand Burn Association recommended burns first aid treatment is to place the burn under cool running water for 20 minutes. Immediate cooling of a burn wound has been shown to reduce the depth of the injury. Cooling has also been recommended as beneficial for up to 3 hours after the burn. No scientific data currently exist to support this recommendation. The aim of this study was to identify the effect of delayed cooling of an acute scald contact burn wound in a porcine model. Four partial-thickness contact scald burn injuries were induced in 12 piglets each. First aid treatment consisting of cool running water for 20 minutes was instituted randomly to each wound at different time points: immediately and at time delays of 5, 20, and 60 minutes. The group receiving immediate first aid with cool running water for 20 minutes served as the control group. At day 1 and day 9, biopsies were obtained and assessed in a blinded manner. Histologic analysis of burn depth on days 1 and 9 demonstrated no significant difference in the depth of the burn in the various treatment groups in comparison to the control group receiving immediate first aid. No significant differences in the surface areas of each burn were noted between the various treatment groups on day 9. Core body temperature did not fall below 35 degrees C throughout the cooling process. This study provides scientific evidence that in an animal model delayed cooling for up to 60 minutes postacute contact scald burn is still effective compared with immediate cooling at reducing burn depth.

An optimum animal model for neonatal thoracoscopy.

PURPOSE: This study attempted to define an optimum animal model for neonatal thoracoscopy. MATERIALS AND METHODS: Seven rabbits and three 5-7-kg piglets were subjected to thoracoscopy by three pediatric surgeons. The outcome for the successful completion of esophageal anastomosis and additional procedures, including diaphragmatic plication and lung biopsy, were noted. RESULTS: Whereas esophageal anastomosis could be accomplished in all piglets, it could be achieved in only 1 rabbit. Additional procedures were possible in 2 piglets and no rabbits. Anesthetic complications were seen in 4 rabbits and 1 piglet. CONCLUSION: Our findings suggest that the piglet may be the optimum animal model for replicating neonatal thoracoscopy. The fragility of the rabbit, coupled with a limited intrathoracic working space, limits its use for thoracoscopy.

Optimal duration of cooling for an acute scald contact burn injury in a porcine model.

The Australian and New Zealand Burn Association recommend 20 minutes of cold running tap water as burn first aid. Scientific evidence for the optimal duration of treatment is limited. Our aim was to establish the optimal duration of cooling using cold running tap water to treat the acute burn. Partial thickness contact scald burns were induced at five sites in each of 17 pigs. Treatments with cold running tap water for 5, 10, 20, and 30 minutes were randomly allocated to different sites together with an untreated control site. In the running water 5 and 10 minute treatments intradermal temperatures rose by 1 degrees C per minute when cooling was stopped, compared with 0.5 degrees C per minute for 20 and 30 minutes duration. No differences in the surface area of each burn were noted between the five treatments on day 9. Histological analysis of burn depth on days 1 and 9 revealed that a higher proportion of burns treated for 20 and 30 minutes showed improvement compared with those treated for 5 and 10 minutes only. This difference reached statistical significance (P < .05) only in the cold running water for 20 minutes treatment arm. There was a statistically significant (P < .05) improvement in burn depth in a porcine acute scald burn injury model when the burn was treated with cold running tap water for 20 minutes as opposed to the other treatment durations. This study supports the current burn first aid treatment recommendations for the optimal duration of cooling an acute scald burn.

Bolus or weekly zoledronic acid administration does not delay endochondral fracture repair but weekly dosing enhances delays in hard callus remodeling.

INTRODUCTION: It has been widely assumed that osteoclasts play a pivotal role during the entire process of fracture healing. Bisphosphonates (BPs) are anti-catabolic agents commonly used to treat metabolic bone diseases including osteoporosis, minimizing fracture incidence. Yet, fractures do occur in these patients and the potential for negative effects of BPs on healing has been suggested. We aimed to examine the effect of different dosing regimes of the potent BP zoledronic acid (ZA) on early endochondral fracture repair and later callus remodeling in a normal bone healing environment. METHODS: Saline, a Bolus dose of 0.1 degrees mg/kg ZA or 5 weekly divided doses of 0.02 degrees mg/kg of ZA commenced 1 week post operatively in a rat closed fracture model. Samples at 1, 2, 4 and 6 weeks post fracture were used to analyze initial fracture union, and 12 and 26 weeks post fracture to investigate the progress of remodeling. RESULTS: ZA did not alter the rate of endochondral fracture union. All fractures united by 6 weeks, with no difference in the progressive reduction of cartilaginous soft callus between control and treatment groups over time. ZA treatment increased hard callus bone mineral content (BMC), volume and increased callus strength at 6 and 26 weeks post fracture. Hard callus remodeling commenced at 4 weeks post fracture with Bolus ZA treatment but was delayed until after 6 weeks in the Weekly ZA group. By 12 and 26 weeks, Bolus ZA had equivalent callus content of remodeled neo-cortical bone to the Saline controls, whereas Weekly ZA remained reduced compared to Saline controls at these times (P<0.01). Callus material properties such as peak stress were significantly reduced in both ZA groups at 6 weeks. At 26 weeks, Bolus ZA-treated calluses generated peak stress equivalent to control values, whereas Weekly ZA callus peak stress remained significantly reduced, indicating remodeling delay. CONCLUSIONS: Osteoclast inhibition with ZA does not delay endochondral fracture repair in healthy rats. Bolus ZA treatment increased net callus size and strength at 6 weeks while allowing hard callus remodeling to proceed in the long term, albeit more slowly than control. Prolonged bisphosphonate dosing during repair does not delay endochondral ossification but can significantly affect remodeling long after the drug is ceased.

Bisphosphonate-laden acrylic bone cement: mechanical properties, elution performance, and in vivo activity.

Cemented total hip replacements generally fail after 10-20 years, often due to implant loosening from bone resorption. Bisphosphonates such as zoledronic acid (ZA) and pamidronate (PAM) are potent inhibitors of bone resorption. The local delivery of bisphosphonates via acrylic bone cement could decrease osteolysis and prolong implant lifespan. Conflicting studies suggest that bisphosphonate loading may or may not reduce the mechanical properties of acrylic bone cement. We assayed acrylic bone cement laden with ZA or PAM at different concentrations and diluent volumes. Four-point bend testing and compressive testing indicated that high volumes of diluent (with or without bisphosphonate) significantly reduced bending modulus and compressive strength. Radiography and electron microscopy indicated that high diluent volumes generated abnormal acrylic bone cement structure. After 6 weeks of incubation in saline, only 0.9% w/w of the total bisphosphonate incorporated in acrylic bone cement eluted in vitro, indicating a slow elution rate. In vivo testing was performed using a rat model. Cement cylinders were inserted into incisions in rat distal femora and ZA delivered locally (via elution from acrylic bone cement) or systemically (via injection). At 4 weeks postoperatively, dual energy X-ray absorptiometry demonstrated no significant increase in local bone mineral density (BMD) adjacent to ZA-laden implants. In contrast, systemic ZA delivery (0.1 mg/kg) led to a large (48.6%) and significant increase in BMD. Thus, systemic delivery appears more effective than local delivery.

Models of tibial fracture healing in normal and Nf1-deficient mice.

Delayed union and nonunion are common complications associated with tibial fractures, particularly in the distal tibia. Existing mouse tibial fracture models are typically closed and middiaphyseal, and thus poorly recapitulate the prevailing conditions following surgery on a human open distal tibial fracture. This report describes our development of two open tibial fracture models in the mouse, where the bone is broken either in the tibial midshaft (mid-diaphysis) or in the distal tibia. Fractures in the distal tibial model showed delayed repair compared to fractures in the tibial midshaft. These tibial fracture models were applied to both wild-type and Nf1-deficient (Nf1+/-) mice. Bone repair has been reported to be exceptionally problematic in human NF1 patients, and these patients can also spontaneously develop tibial nonunions (known as congenital pseudarthrosis of the tibia), which are recalcitrant to even vigorous intervention. pQCT analysis confirmed no fundamental differences in cortical or cancellous bone in Nf1-deficient mouse tibiae compared to wild-type mice. Although no difference in bone healing was seen in the tibial midshaft fracture model, the healing of distal tibial fractures was found to be impaired in Nf1+/- mice. The histological features associated with nonunited Nf1+/- fractures were variable, but included delayed cartilage removal, disproportionate fibrous invasion, insufficient new bone anabolism, and excessive catabolism. These findings imply that the pathology of tibial pseudarthrosis in human NF1 is complex and likely to be multifactorial.

Modeling bone morphogenetic protein and bisphosphonate combination therapy in wild-type and Nf1 haploinsufficient mice.

Recombinant bone morphogenetic proteins (BMPs) show promise in treating the orthopedic complications associated with neurofibromatosis type 1 (NF1), such as congenital pseudarthrosis of the tibia. Minimal scientific information regarding the effects of BMP in the context of NF1 is available. As abnormalities in both bone formation and resorption have been documented in Nf1-deficient mice, we hypothesized that inadequate BMP-induced bone formation could be augmented by cotreatment with the bisphosphonate zoledronic acid (ZA). First, primary osteoblasts isolated from wild type (Nf1(+/+)) and Nf1-deficient (Nf1(+/-)) mice were cultured in the presence and absence of BMP-2. While Nf1(+/-) cells exhibited less osteogenic potential than Nf1(+/+) cells, alkaline phosphatase expression and matrix mineralization for both genotypes were enhanced by BMP-2 treatment. To model this response in vivo, 20 microg BMP-2 was implanted intramuscularly into the quadriceps of mice to induce heterotopic bone. Radiographs revealed significantly less net bone formation in Nf1(+/-) mice compared to Nf1(+/+) controls. To test the effect of an antiresorptive agent, mice were cotreated twice weekly from postoperative day 3 with 0.02 mg/kg ZA or with saline. ZA treatment led to a synergistic increase in the amount of heterotopic bone in both Nf1(+/+) and Nf1(+/-) mice compared with saline controls, as measured by DEXA and histomorphometry. Thus, the anabolic deficiency noted in Nf1(+/-) mice is amenable to stimulation by BMP-2, but mineralized tissue formation remains below that of Nf1(+/+) controls. Bisphosphonate combination therapy is superior to BMP therapy alone in terms of net bone production in vivo in both wild-type and Nf1-deficient mice.

Optimal timing of a single dose of zoledronic acid to increase strength in rat fracture repair.

UNLABELLED: We hypothesized that ZA treatment would bolster fracture repair. In a rat model for closed fracture healing, a single dose of ZA at 0, 1, or 2 wk after fracture significantly increased BMC and strength of the healed fracture. Delaying the dose (1 or 2 wk after fracture) displayed superior results compared with dosing at the time of fracture. INTRODUCTION: Bisphosphonates are known to increase bone strength and thus the resistance to fracture by decreasing osteoclastic bone resorption. These properties may enable bisphosphonates to also increase the strength of fracture repair. Zoledronic acid (ZA) is a potent bisphosphonate with a high affinity for bone mineral, allowing bolus intravenous dosing in a range of indications. In this study, we examined the application of bolus dose ZA in endochondral fracture repair. MATERIALS AND METHODS: Carbon-14 labeled ZA was used in a closed rat fracture model. Rats were divided into five treatment groups (n = 25 per group): saline control, local ZA (0.01 mg/kg), and three systemic bolus ZA groups (0.1 mg/kg) with different administration times: at fracture, 1 wk after fracture, and 2 wk after fracture. Rats were killed 6 wk postoperatively. Postmortem analyses included radiography, QCT, microCT, biomechanical testing, scintillation counting, autoradiography, and histology. RESULTS: Single-dose systemic ZA administration significantly increased callus volume, callus BMC, and mechanical strength. Perioperative treatment increased mechanical strength by 30% compared with controls (p < 0.05). Administering the systemic dose at 1 or 2 wk after fracture further increased mechanical strength compared with controls by 44% and 50%, respectively (p < 0.05). No significant differences in mechanical parameters were seen with local injection at the dose studied. Autoradiographic analysis indicated that ZA binds significantly to bone that is present at the time of administration. ZA quantification indicated that delayed administration significantly increased the uptake efficiency in the callus. Histological and microCT analysis showed that ZA treated calluses had a distinctive internal structure consisting of an intricate network of retained trabecular bone. CONCLUSIONS: The timing of a single systemic dose of ZA plays an important role in the modulation of callus properties in this rat fracture model; delaying the single dose produces a larger and stronger callus.

Manipulation of the anabolic and catabolic responses with OP-1 and zoledronic acid in a rat critical defect model.

UNLABELLED: Bone repair involves both anabolic and catabolic responses. We hypothesized that anabolic treatment with OP-1 (BMP-7) and anti-catabolic treatment with zoledronic acid could be synergistic. In a rat critical defect, this combination therapy produced significant increases in new bone volume and strength. INTRODUCTION: When used to augment bone healing, osteogenic protein 1 (OP-1/BMP-7) and other BMPs stimulate the anabolic response, inducing osteoblast recruitment, differentiation, and bone production. However, BMPs can also upregulate catabolism by direct stimulation of osteoclasts and indirectly by osteoblasts through RANKL/RANK. We hypothesized that if such osteoclastic upregulation were modulated by zoledronic acid (ZA), the combination of OP-1 and ZA should produce increased new bone over OP-1 alone. MATERIALS AND METHODS: Rats with a surgically induced 6-mm femoral critical size defect were separated into five dosing groups: Carrier, Carrier + ZA, OP-1, OP-1 + ZA, and OP-1 + ZA administered 2 weeks after surgery (2W). Carrier +/- 50 microg OP-1 was placed in the defect, and 0.1 mg/kg ZA or saline was administered subcutaneously. Bone repair was analyzed by radiographs, QCT, mechanical testing, histology, and histomorphometry. RESULTS: Carrier alone and Carrier ZA groups did not unite by 8 weeks. Radiological union occurred in all OP-1 groups but was tenuous in some animals treated with OP-1 alone. BMC was increased by 45% in the OP-1 ZA group and 96% in the OP-1 ZA 2W group over OP-1 alone (p < 0.01). Callus volume increased over OP-1 alone by 45% and 86% in the OP-1 ZA and OP-1 ZA 2W groups, respectively (p < 0.01). The increased callus volume in the OP-1 ZA 2W group translated to increases in strength of 107% and stiffness of 148% (p < 0.05). BFR was not significantly different between OP-1 groups regardless of ZA treatment. CONCLUSIONS: ZA treatment significantly increased the BMC, volume, and strength of OP-1-mediated callus in a critical size defect in rats at 8 weeks. Thus, modulation of both anabolic and catabolic responses may optimize the amount and mineral content of callus produced, which could be of clinical benefit in obtaining bone union.