Effective Bone Hemostasis and Healing Using Radiofrequency and Conductive Fluid

Abstract

Hemostasis in bone is difficult to achieve because of the mineral content. Current techniques often are ineffective, can have systemic effects, or leave residual material in the wound. Our hypotheses were that a wand device coupling radiofrequency energy with a cooling conductive saline solution, applied topically to bone, could produce superior hemostasis compared with conventional electrocautery or no treatment, and not impede bone healing. Immediate hemostasis and subsequent bone healing for 6 and 12 weeks were evaluated in an iliac crest ostectomy (cancellous bone) and a drilled tibia defect (cortical bone) sheep model. Outcome variables were amount and intensity of bleeding, serial radiography, quantitative computed tomography, histology and mechanical testing. Control of bleeding was nearly complete (93%) and greater with the radiofrequency/saline treatment compared with electrocautery (56%) or no treatment (0%) in cancellous bone and cortical bone. Electrocautery induced surface char (black carbon debris) that could be seen at 6 and 12 weeks. There were no differences in bone healing between the radiofrequency and electrocautery device applications or untreated bone. At 12 weeks, all healing tibiae defects were as strong as undrilled tibiae. This may be an effective method to produce rapid hemostasis in bone without char or healing complications.

Blood loss in orthopaedic procedures can be significant and associated with postoperative morbidity.³ At most risk are patients with clotting disorders,^6,11,12,16 pediatric patients and patients with senility,^17,21 and patients with a high risk of bleeding, such as patients with cancer, sepsis, or patients having reoperations.^3,15,20 Blood transfusions provide significant risks and negative outcomes and are decreasing in supply and increasing in cost.¹⁵ Reduction in blood loss provides clinical benefits, in the visual field at surgery and postoperatively, particularly in patients at high risk.^18,19 Techniques for hemostasis in bone include pressure, antifibrinolytics,^7,15, electrocautery, or topical wax/foam/powders such as absorbable porcine gelatin and cellulose in fibrillar and powder forms.¹⁴ These methods may be ineffective, provide a risk of pharmacologic reactions, and leave char or residual foreign material in situ. A safer method that could control bleeding effectively in bone would offer clinical advantages. Any method used should not impede bone healing.

Application of radiofrequency (RF) energy offers potential advantages over other techniques to induce hemostasis by avoiding foreign materials and proteins and reducing tissue carbonization and necrosis. Use of RF has been described for tissue ablation and collagenous tissue shrinkage.^1,4,10 However, the use of saline-cooled RF thermal modification of connective tissues to induce collagen shrinkage of blood vessels and subsequent coagulation is novel and potentially would not induce tissue dessication, smoking, char, or necrosis. Radiofrequency energy coupled with saline delivery cools tissue in contrast to conventional electrocautery which reaches temperatures of 300°C and routinely induces char (carbonization of tissues) at the surface. The use of RF in a conductive saline to achieve hemostasis has not been evaluated for bone, and RF thermal energy could induce delayed bone healing. Our hypotheses were that superior hemostasis could be achieved in cancellous and cortical bone after application of a bipolar RF device as compared with an electrocautery device, and that healing of RF-treated bone and surrounding soft tissues would be similar to electrocautery treatment and untreated bone.

MATERIALS AND METHODS

Two bone injuries were created in each hind limb of 10 sheep (n = 20 limbs), an ostectomy of the iliac crest and a drilled defect of the medial tibia. Both defects on one hind limb received RF treatment, and both defects on the other hindlimb received electrocautery treatment. Bone healing was evaluated at 6 (n = 5 sheep) and 12 (n = 5 sheep) weeks. Two additional sheep did not receive any treatment at either site or limb (n = 4 specimens per site) and were evaluated at 6 weeks. A power calculation using the variations (standard deviation) and similar bone values (means) reported in another study of bone healing of inbred species using a similar paired-limb design and outcome measures² and accepting an α error of 0.05 (significance of detecting a difference set a p < 0.05) and a β error of 0.2, suggested 3.5 animals per group would be required. We selected a larger group (n = 5) for our study. Outcome measures for hemostasis were bleeding intensity score and percent surface without bleeding. Outcome measures for bone healing were serial radiographs, peripheral quantitative computed tomography (qCT), histomorphometry, and biomechanical testing. The sheep had a bilateral iliac crest ostectomy (3 × 1 cm) for a cancellous bone model and a bilateral medial distal tibia unicortical drilling (5.5 mm diameter) for the cortical bone model (Fig 1). All procedures were performed with the animals under general anesthesia. The animals were allowed unrestricted activity in a 12 × 12-foot box stall until euthanasia by overdose of barbiturate (Euthasol, Virbac Animal Health, Fort Worth, TX; 1 mL/4.5 kg) at 6 weeks or 12 weeks. The protocol was approved by The Ohio State University Institutional Laboratory Care and Use Committee.

Fig 1: This illustration shows the surgical sheep model used to create bleeding in cancellous (ilium) and cortical (tibia) bone.

Based on pilot studies, a bipolar RF device (BPS5.0™, Bipolar Sealer, TissueLink Medical, Inc, Dover, NH) designed to accept a saline (0.9% NaCl) drip and a Force FX generator (Valley Labs, Boulder, CO) were used to seal the bone in treated sheep at 200 J fluence/cm (50 W, 4 cc/minute flow rate for 11 seconds /2.9 cm )(Fig 2). Standard monopolar electrocautery wands in the coagulation mode at 50 W (Force FX generator, Valley Labs) were used on the paired limbs for the same time. The exposed bone surface was measured at surgery (ostectomy length and width for the ilium and a 1-cm² cortical bone surface over the drill hole for the tibia) to estimate the application time for both devices. Areas of 1.0, 2.0, 2.5, 3.0, and 3.5 cm² were treated for 4, 8, 10, 12 or 14 seconds, respectively. The roller-ball wand tip with saline flow was applied directly to the surface of the bone area in a paint-brush pattern (Fig 3). Power was applied uniformly across the entire exposed surface for the assigned time. Surgeon discretion was used in the assigned time to return to persistently bleeding sites.

Fig 2: The Tissue Link Medical radiofrequency wand with a bipolar saline drip rollerball tip attached to a saline solution drip set and power source is shown.

Fig 3: The radiofrequency bipolar wand with saline drops at the tip is about to be applied to the surface of the bone.

Bleeding intensity and surface area were estimates of hemostasis. Bleeding intensity was scored by one investigator (AB), just before and 1 minute after completion of device application as 0 = none, 1 = minimal, 2 = mild, 3 = moderate, or 4 = marked. At this same time, the percent bone surface bleeding was estimated and converted to percent hemostasis (100-X).

Clinical parameters were assessed as estimates of healing. Surgical sites were scored by two investigators (AB, JK) for drainage (score 0-4; 0 = no drainage, 1 = 0-25% of the bandage surface stained with drainage, 2 = 26-50% of the bandage surface stained with drainage, 3 = 51-75% of the bandage surface stained with drainage; 4 = 76-100% of the bandage surface stained with drainage) and measured (mm) for extent of swelling (distance of swelling from the incision at the widest point). Lameness was scored by two investigators (AB, JK) as 0-5 for each hind limb at walking and trotting (0 = no lameness, 1 = minimal lameness, 2 = mild lameness, 3 = moderate lameness, 4 = marked lameness (only placing part of the foot), and 5 = nonweightbearing lameness).

Serial bone healing was assessed with radiography. A baseline radiograph of each iliac crest (ventrodorsal view) and of the tibiae (anteroposterior view) were obtained immediately after surgery, at Weeks 3 and 6 (Group 1), and Weeks 9 and 12 (Group 2). Tibial radiographs were scored by one investigator (AB), without knowledge of the group or week, for new bone formation in the defect and adjacent bone remodeling (0 = none, = minimal, 2 = mild, 3 = moderate, 4 = marked). Bone remodeling was noted as a local heterogeneous loss of radiodensity. Ilial radiographs were scored by one investigator (AB) for new bone formation at the surface and edges of the ostectomy site and adjacent bone remodeling (0 = none, 1 = minimal, 2 = mild, 3 = moderate, 4 = marked).

At tissue harvest, bone surface discoloration was assessed during dissection to the bone. Treated surfaces were inspected grossly and scored by one investigator (AB) for the presence of discoloration (0 = none [white tissue], 1 = minimal [tan tissue], 2 = mild [brown tissue], 3 = moderate [brownish-black tissue], and 4 = black tissue).

To quantify amount and density of bone healing in tibiae, computed tomography (CT) scans (Picker PQS helical CT, Philips Medical Systems for North America, Bothell, WA) were performed through the drilled defect of the tibiae at 1-mm slices. Corresponding software (Philips 4.6B1, Philips Medical Systems for North America) was used on the central slice to trace the amount of bone in the original drilled area (region of interest [ROI]). The central slice was standardized for xray attenuation differences for density measurements using a potassium phosphate standard. After standardization, the software calculated the conversion of the potassium phosphate ROI to ash density.⁸ Tracings of the ROIs were made for the original defect and bone in the defect. Mean bone density was calculated similarly in adjacent cortical bone in a 3-mm circular ROI of plantar cortical bone at least 3 mm from the defect. Calculations were made by one investigator (JK) for percent new bone in the defect by dividing the area of bone in the defect by the original area of the defect. Density of the new bone in the defect was expressed as a percent of the normal adjacent bone density. All area and density values were obtained directly by the CT-unit ROI tracings and calculated by the CT software.

Return of structural strength was assessed as an estimate of healing in cleaned distal tibiae that were halved and loaded at 0.5 N/second in 3-point bending (Servohydraulic Materials Testing Machine, MTS Bionics 858, Company, Eden Prairie, MN) in a customized jig until failure. The medial halves were used for testing healed defects. The lateral halves were used for testing empty drilled defects (n = 5 per group) or undrilled tibiae (n = 5 per group). A preliminary study on three pilot sheep showed a low coefficient of variance of 2.1% in peak load to failure between medial and lateral tibial halves. The thickness of the bone specimen at the drill-hole site was measured with calipers before mechanical testing. Peak load to failure (N) and bone strength (N/mm bone thickness) were calculated and recorded. The mode of bone failure was recorded as (1) through the defect, (2) through intact bone but directly adjacent or around the edge of the defect, or (3) through intact tibia not involving the defect.

Bone necrosis, bone quality, and bone density of healing bone were assessed with histologic evaluation. After 3-point bending tests (tibiae) and fixation (10% formalin) of bones (tibiae and iliac crest), specimens were cut on a band saw to create bone blocks for calcified processing and sectioning (15 (xm; Exakt System, Zimmer Inc, Freiburg, Germany). Careful alignment produced sections perpendicular to the treated surface for the iliac crest and in central cross section for the tibial defect. Sections were stained with Masson's trichrome stain and evaluated by two investigators (AB, SW) without knowledge of the treatment group. Sections were scored for amount of inflammatory cells and edema (0 = none, 1 = minimal, 2 = mild, 3 = moderate, 4 = marked), bone morphology (normal = 0, 1 = minimal, 2 = mild, 3 = moderate, 4 = severe distortion of normal architecture), bone necrosis (0 = none, 1 = minimal, 2 = mild, 3 = moderate, 4 = marked), and measured for necrosis depth (mm). Scores of the investigators were summed to produce an inflammation and bone necrosis index for the tibiae and iliac crest.

Fluorochrome bone labeling was performed to assess rate of bone formation and bone activity using calcein green (Calcein (20mg/kg); C0875, Sigma Chemical C, St Louis, MO) dissolved in a 2% solution of sodium bicarbonate and administered intravenously at the start of Week 4 and again at the start of Week 5 in the 12-week group. Unstained histologic sections were evaluated for bone porosity under UV light at ×40 magnification. Five 1-mm² areas were counted by one investigator (JK) in the bone filling the defect in the tibiae, under the treated surface of the ilia, and in control adjacent areas of bone deep in the ilium and on the lateral side of the tibial cortex. A 1-mm² ocular grid containing 10 boxes measuring 1 um each was scrolled over the region of interest and viewed through the microscope. Point counting at the intersected lines of bone or no bone (ie, marrow space or intertrabecular space) was done, and percent bone was expressed as number of points on bone/number of points total (bone and no bone) × 100. Bone formation rate was estimated by quantifying bone-forming surface area using similar grids and point counting the number of intersections on fluorescent bone (active bone formation) expressed as a percent of the sum of fluorescent and nonfluorescent bone. Mineral apposition rate (MAR) of osteons, and trabecular and woven bone were evaluated under UV light at ×600 magnification. In labeled bone, the distance between the outer borders of the two adjacent labels were determined in five different images in the same region. Mineral apposition rate was calculated by determining the mean value of this distance and dividing by the number of days between administration of the calcein marker. Mineral apposition rate is expressed at μ per day.

Commercial software programs were used for data analysis (SAS, version 8.2, SAS Institute Inc, Cary, NC). Quantitative data were analyzed by multiple analysis of variance (MANOVA) for between treatments and among times with animal as a dependent variable. (p < 0.05) If significance was identified between groups or across time, a least squares difference) post test was done to identify the subgroups that were significantly different (p < 0.05). Scored data were analyzed by a Mann-Whitney rank test (p < 0.05).

RESULTS

Bleeding surface area and intensity in ilial cancellous bone was reduced by RF and electrocautery (p < 0.02) and reduced greater by RF than electrocautery (p < 0.015), achieving nearly complete (mean, 93%) hemostasis with RF. (Table 1; Fig 4) Bleeding intensity in tibial cortical bone was reduced significantly after RF application (p < 0.05). In the tibiae defects treated with RF that continued to bleed, the bleeding started at a mean of 42 seconds and appeared to well up from the medullary cavity. Subjectively, surface bleeding from the cortex and periosteum was eliminated.

TABLE 1: Hemostatis of Cancellous Sheep Bone with Radiofrequency or Electrocautery Devices

Fig 4A: B. The ilial bone surface is shown after ostectomy. (A) The radiofrequency-treated surface was blanched with 100% hemostasis. (B) The electrocautery-treated surface appears black with char with several areas of continued bleeding.

Sheep were clinically normal in health and gait throughout the study. There were no differences across time or between treatments in tibiae or ilial soft tissue healing, including incisional swelling, incisional drainage, and histologic fibrosis. At 6 weeks, the bone that had electrocautery treatment had an abundance of black material visible on the bone surface, greater than with the RF treatment (p < 0.05). This was observed sporadically at 12 weeks. Bone healing progressed in a normal manner and similarly in RF-treated, electrocautery-treated and untreated bone. Bone formation was activated (more bone label) at treated sites. There were no differences between the RF and electrocautery treatment groups in any bone activity parameter in either the tibiae or ilia. (Table 2; Figs 6) New bone radiographically increased from Weeks 3, 9, and 12 in the ostectomy and the tibia sites (p < 0.01). Bone remodeling radiographically increased at Weeks 3, 6, 9, and 12 at the ostectomy site (p < 0.01) and Weeks 3, 6, and 9 at the tibia site (p < 0.01). Pattern of bone healing progressed centripetally in the tibia defects, and the amounts and densities of bone filling the defects were similar in RF-treated, electrocautery-treated, and untreated bone and greater at 12 weeks than 6 weeks. (Fig 7; Table 2) A circular ring pattern of remodeling was identified on radiographs of two tibial defects, one in the RF treatment group and one in the electrocautery treatment group. This pattern was confirmed with CT as surface bone resorption to 2 mm depth (Fig 8). Bone healing of tibiae was mechancially completed by 12 weeks when RF-treated, electrocautery-treated, and untreated tibiae failed similarly to undrilled tibiae. (Table 2) There was no difference in mode of failure in the RF treatment group (median mode of failure score = 2.0) as compared with the EC treatment group (median mode of failure score = 1.0) (p = 0.1). Histologic assessment showed that, regardless of treatment, the predominant tissue type at the ilial surgery site was mature fibrous tissue and new bone (Fig 5). The predominant tissue filling the tibial defect was normal woven bone. Bone necrosis was minimal, present at 6 weeks, more frequent in tibial than ilial specimens, and not different between treatment groups. The depth of necrosis ranged from 100 to 300μm. Inflammation in bone or soft tissue was not observed. Bone and soft tissue healing in RF and electrocautery treatment groups were similar to those of untreated bone, but char was evident in electrocautery-treated specimens.

TABLE 2: Mechanical and Histologic Variables

Fig 5A: F. Representative transverse histologic sections of the ilial ostectomy site at 12 weeks are shown. (A) Normal trabecular bone of approximately 55% porosity was seen with Masson's trichrome stain at a magnification of ×40. (B) Unstained sections of normal ilial bone under ultraviolet light showed the calcein fluorescent label at a magnification of ×40, and approximately 25% of the bone surface was actively forming bone. (C) Double rings of calcein label administered 1 week apart in normal ilial bone (Magnification, ×100) showed discreet labels of the surface and a mean mineral apposition rate of 2.34 mm/day. (D) Radiofrequency-treated (shown) and electrocautery-treated ilial bone had more bone and smaller pores than adjacent normal bone and (E) greater active bone forming surface than normal bone. (F) Double rings of calcein label were wider in RF-treated and electrocautery-treated bone because the mineral apposition rate was greater.

Fig 6A: F. Representative transverse central histologic sections of sites of drilled tibial defects are shown. (A) Normal tibial cortical bone shows surface lamellar bone, endosteal osteonal cortical bone, and a few active osteons (Stain, Masson's trichrome; magnification, ×40). (B) Unstained sections of normal tibial cortical bone under ultraviolet light showed the calcein fluorescent label at a magnification of ×40. Relatively inactive bone formation was seen in only one osteon. (C) Double rings of calcein label administered 1 week apart in normal tibial bone (Magnification, ×100) showed discreet labels in active osteon with a mean mineral apposition rate of 2.1 mm/day. (D) Bone in RF-treated (shown) and electrocautery-treated drilled tibial defects at 12 weeks contained as much bone as normal cortical bone, but in a woven rather than osteonal pattern. (E) Bone in RF-treated (shown) and electrocautery-treated drilled tibial defects at 12 weeks had a significantly greater amount of label representing greater bone formation than normal cortical bone. (F) Bone in RF-treated (shown) and electrocautery-treated drilled tibial defects at 12 weeks had a twofold greater mineral apposition rate than normal cortical bone.

Fig 7A: B. Quantitative CT of a transverse central slice of healing drilled tibial bone defects treated with radiofrequency showed (A) a normal centripetal formation of bone in the defect at 6 weeks. (B) Centripetal formation of bone had increased to a mean of 84% bone hin the defect by 12 weeks.

Fig 8: The qCT scan shows a transverse central slice of a 6-week defect with a line of separation in original bone (arrow) that may have represented a ring of necrosis seen in two sheep in the 12-week group - one treated with RF and the other treated with electrocautery. This no longer was visible at 12 weeks.

DISCUSSION

We used cancellous bone ostectomy and cortical bone drilled defects to investigate the capability of a new RF modality coupled with conductive fluid to control hemorrhage in bone. Our study was an in vivo study in normal bone that assessed healing for 12 weeks and until return of mechanical strength. We compared RF with a similar modality that is well accepted for control of hemorrhage during surgery. No comparison to hemostatic materials was done. This study is limited to assessment of RF use on normal, adult bone in an animal model. In clinical cases, diseased or inflamed bone may be more or less sensitive to thermal necrosis or have impaired healing capabilities. The RF application we used was satisfactory in nearly eliminating bleeding in normal bone, but in hypervascularized bone, hemostasis may be less satisfactory. We assessed bone healing and necrosis in a tightly controlled range-of-energy application based on a bone surface area formula. Any increase in wattage or time of application of the RF may not yield the same results. Familiarity with estimating the surface area and calculating the time of appropriate application is likely important in minimizing necrosis and optimizing bone healing. Bone adjacent to implants may be more susceptible to necrosis or negative effects of even minimal necrosis. One limitation of our study was that we used medial tibial halves to assess the healing defect and lateral tibial halves to assess undrilled and unhealed drilled tibiae. Assessment of our pilot work and the variances obtained in our data revealed this is unlikely to have affected the outcome of our comparisons. In a pilot study of three sheep used to develop the model (six tibiae), there was no significant difference in the peak load to failure of medial and lateral distal tibial halves, and the between-halves variability, evaluated by calculation of coefficient of variation (CV) and expressed as a percent, was low at 2.1%. To calculate within-halve variability (CV_halve), the standard deviation (SD) of medial and lateral halves was meaned for the three animals. The mean SD_halves was divided by the overall mean peak load to failure to determine the CV_halves as 2.1%. The low CV_halves supported a structural symmetry of the distal tibia at the site we selected, and we elected to collect data of the undrilled and drilled but unhealed lateral halves as relevant comparative data. Similarly, the CV_animal was calculated as 6.8% and represents the variability among sheep in peak load to failure. This relatively low within-animal CV⁵ attested to the homogeneity of our sheep in age, breed, body weight, gender, and use. As expected, the CV_animal was greater in bone of the 10 healing sheep (7.4%), with this additional source of potential variation (strength of healing tissue) contributing to among-animal variability. However, CVs less than 10% for animal data are relatively low.⁵ In consideration of our relatively large differences in peak load to failure between undrilled and healed tibiae at 6 weeks, the contribution of halve variation to the overall variability between the groups is proportionally low. Variability among observers for the histologic scoring was not calculated and circumvented by use of one index score for histologic parameters. In the instances in which two observers scored incisional drainage and lameness, scores were identical as there was no drainage or lameness.

In comparing the RF energy device to the electrocautery energy device, the advantages of RF were greater and more consistent hemostasis and lack of char or tissue carbonization. The char was visible grossly and histologically at 6 and 12 weeks in the electrocautery group. The consequences of the presence of carbon in the tissues is unknown. There did not seem to be a measurable negative effect on bone healing rate, quality, or mechanical properties. Inflammation was not a feature of the tissue surrounding the carbon. Because our application of electrocautery was short, the build-up of char was minor. However, hemostasis also was insufficient in the time of application constraints of this study. Overall, electrocautery was less effective in inducing hemostasis in bone. Direct comparison to other methods of hemostasis such as chemostatic and pharmacologic methods was not done. Pharmacologic methods may reduce bleeding by 50%,^11,12 but bleeding at the surgery site still is not specifically controlled as with a direct application method. Topical chemostatics offer the advantage of no risk of thermal injury, however, foreign material must be placed on the surface and may be removed inadvertently during manipulation. Incomplete hemostasis often occurs if bleeding is intense and if the material is dislodged. A relative comparison of available techniques is provided in Table 3.

TABLE 3: Advantages and Disadvantages of Methods of Hemostasis

We confirmed that RF energy coupled through a conductive fluid provided nearly complete hemostasis in cancellous and cortical bone, and it did not interfere with normal bone healing. When the RF energy was applied in the parameters of this study (200 Joules fluence/cm²), bone necrosis was minimal, cortical tibial defects maintained structural strength and healed to near original strength within 12 weeks. It is well established that RF energy can be applied to tissues and induce undesirable consequences, such as excessive necrosis.^9,13 Adherence to guidelines on application is important to prevent such consequences. However, use of this RF device can induce a bloodless field on bone that may be valuable for certain surgeries requiring high visibility or lack of fluid loss. Expectations for normal progression of bone healing can exist. The time commitment to apply (at moderate power) the device is minimal - less than 14 seconds to cover 3.5 cm² of bleeding bone.

We confirmed that using a RF device with a conductive fluid, as used in this study, was safe for cortical and cancellous bone and comparable in bone healing to electrocautery or no treatment. The advantage of the RF device for hemostasis is complete hemostasis in cancellous and cortical bone. Hemostasis was better than that achieved with standard electrocautery techniques, without materials or proteins added to the wound. No untoward clinical effects on the incision, limb gait, or animal discomfort occurred. This technique may result in greater patient comfort postoperatively and fewer blood transfusions.

Acknowledgments

We thank Viktor Krebs, MD, for assistance with model development, Amanda Johnson and Kate Hissam for technical assistance, and Dr. Paivi Rajala-Schultz for statistical consultation.