The Spine Journal - (2014) -

Basic Science

Evaluation of resorption and biocompatibility of collagen hemostats inthe spinal epidural space

Kentaro Mizuno, MD, Takumi Ikeda, MD, PhD*, Kazuya Ikoma, MD, PhD,Hidenobu Ishibashi, MD, Hitoshi Tonomura, MD, PhD, Masateru Nagae, MD, PhD,Yuji Arai, MD, PhD, Hiroyoshi Fujiwara, MD, PhD, Yasuo Mikami, MD, PhD,

Toshikazu Kubo, MD, PhDDepartment of Orthopaedics, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, 465 Kawaramachi-Hirokoji, Kamigyo-ku,

Kyoto 602-8566, Japan

Received 1 October 2013; revised 5 December 2013; accepted 22 January 2014

Abstract BACKGROUND CONTEXT: Collagen hem

FDA device/drug

Author disclosures

KI: Nothing to disclos

MN: Nothing to disc

close. YM: Nothing t

All the authors hav

tent of this article and

1529-9430/$ - see fro

http://dx.doi.org/10.10

ostats have different characteristics depending ontheir properties and configuration. In vivo serial evaluation of local reactions because of placementof hemostats in the epidural space has not been reported.PURPOSE: This study compared the resorption and biocompatibility of two types of collagen he-mostats placed in the epidural space.STUDY DESIGN: This in vivo study used experimental animals to evaluate collagen hemostatsthat were placed in the epidural space.METHODS: A ligamentum flavum resection model was created in Japanese white rabbits (n565).A microfibrillar collagen hemostat (MCH group, n55), cotton-type collagen hemostat (CCH group,n55) that was chemically cross-linked, or no hemostat (control group, n54) was placed in the spi-nal epidural space. For histologic evaluation, each group was euthanized 1, 2, 4, and 8 weeks post-operatively (PO), and hematoxylin-eosin and immunohistochemical (IHC) staining forinflammatory cytokines (tumor necrosis factor [TNF]-a, interleukin [IL]-6), cyclooxygenase(COX)-2, and macrophages (CD68) was performed. To evaluate exudate accumulation and the de-gree of inflammation in the epidural space, magnetic resonance imaging at 7.04 T was serially per-formed in each group (n53) under anesthesia and sedation.RESULTS: The collagen hemostats in both groups were reabsorbed at 4 weeks PO. In the MCHgroup, there was inflammatory cell infiltration and granuloma formation around the hemostat, TNF-a–positive cells were seen up to 1 week, and IL-6–, COX-2–, and CD68-positive cells were seen atall evaluation times. In the CCH group, no inflammatory cell infiltration around the hemostat wasobserved, and IHC staining showed no positive cells at 4 weeks PO and later. T2*-weighted MRimages showed significantly higher mean signal intensity of the epidural space in the MCH groupthan in the CCH group but only at 1 week PO (p!.05).CONCLUSIONS: Resorption of both hemostats was similar. In the MCH group, there was intensetissue inflammation around the hemostatic material, and MR images showed high signal intensitybecause of exudate accumulation in the epidural space. This indicated a strong foreign-body

status: Not applicable.

: KM: Nothing to disclose. TI: Nothing to disclose.

e. HI: Nothing to disclose. HT: Nothing to disclose.

lose. YA: Nothing to disclose. HF: Nothing to dis-

o disclose. TK: Nothing to disclose.

e no conflicts of interest directly relevant to the con-

received no financial support.

* Corresponding author. Department of Orthopaedics, Graduate School

of Medical Science, Kyoto Prefectural University of Medicine, 465

Kawaramachi-Hirokoji, Kamigyo-ku, Kyoto 602-8566, Japan. Tel.: (81)

75-251-5549; fax: (81) 75-251-5841.

E-mail address: tikeda@koto.kpu-m.ac.jp (T. Ikeda)

nt matter � 2014 Elsevier Inc. All rights reserved.

16/j.spinee.2014.01.050

2 K. Mizuno et al. / The Spine Journal - (2014) -

reaction to the MCH, thus demonstrating a difference in biocompatibility with the CCH. � 2014Elsevier Inc. All rights reserved.

Keywords: Collagen hemostat; Spinal epidural space; Biocompatibility; Resorption; Antigenicity; Inflammation

Introduction

Collagen hemostats in spinal surgery are mainly used tocontrol bleeding from the epidural venous plexus. After he-mostasis, removal of as much of the collagen hemostat aspossible is recommended [1,2], but we have encounteredcases in which rebleeding after removal occurs, thus requir-ing retention of a topical hemostat. However, neurologicdeficits because of swelling and granulation of topical he-mostats have also been reported as a postoperative (PO)complication [3–8]. In addition, differences in resorptionand biocompatibility of hemostatic materials dependingon their properties and configuration have been reported[1,2,9–16]. When a hemostatic material is placed in the epi-dural space, inflammation and cytotoxicity because of thehemostatic material may adversely affect surrounding tis-sue such as the spinal nerves. Nevertheless, to our knowl-edge, local reactions that may occur because of theplacement of hemostatic materials in the spinal epiduralspace have not been investigated in vivo.

The purpose of this study is to compare the resorptionand biocompatibility of two types of collagen hemostatswhen placed in the epidural space. We created a ligamen-tum flavum resection model in Japanese white rabbits,and we placed two different types of collagen hemostatsused in spinal surgery in the epidural space. Histologicanalysis was performed to evaluate resorption and the bio-logical response to each hemostatic material. In addition,high magnetic field magnetic resonance imaging (MRI) inthe experimental animals under anesthesia and sedationwas performed for serial imaging to evaluate exudate accu-mulation and the degree of inflammation in the epiduralspace.

Materials and methods

Hemostatic materials placed in the spinal epiduralspace

We used two types of hemostatic materials purifiedfrom collagen that was obtained from bovine dermis.One type was the microfibrillar collagen hemostat(MCH) (Avitene; Zeria Pharmaceutical Co. Ltd, Tokyo,Japan). Microfibrillar collagen hemostat, which was de-veloped in the United States in the 1970s, is a hemostaticmaterial developed by purifying collagen so that the natu-ral fibril cross-links are maintained. The other type wasthe cotton-type collagen hemostat (CCH) (Integran; Nip-pon Zoki Pharmaceutical Co. Ltd and Koken Co. Ltd,

Tokyo, Japan). The raw material was atelocollagen, inwhich telopeptide, the main antigenic determinant site incollagen, had been removed to reduce antigenicity [12].CCH was produced by spinning atelocollagen intocotton-like fibers and chemical cross-linking the fibers us-ing a polyepoxy compound. CCH is a type of collagen he-mostat developed in Japan in the 1990s to reduceantigenicity and to improve handling and resorption. Inthis study, to insert the hemostatic material under the ver-tebral arches, the press sheet type with cotton fibers spreadalong a plane was used.

Resected ligament flavum model for spinal dura materexposure

This animal study was approved by the ExperimentalAnimal Committee at our institution. We created a liga-mentum flavum resection model using 65 Japanese whiterabbits (males, age 13 weeks, 2.5–2.8 kg). The animalswere purchased from Oriental Bio Service (Kyoto,Japan).

Surgery was performed under inhalation anesthesia. A 3-cm posterior longitudinal incision was made in the dorsumat the level of L7/S1. The fascia was opened, the spinousprocesses were spread, and the ligamentum flavum at L7/S1 was exposed. Then, the ligamentum flavum was re-sected, and the epidural fat was removed to expose the duramater.

For histologic and immunohistochemical (IHC) evalua-tion, 0.5 mg of each material was placed on the dura underthe first sacral vertebral arch in 56 rabbits. For MRI, 5 mgof each material was inserted on the dura between the L7and S1 vertebral arches in nine rabbits. The fascia and skinwere closed with 3-0 nylon suture.

The experimental model was divided into MCH (place-ment of MCH), CCH (placement of CCH), and control (re-section of ligamentum flavum only, but no placement of ahemostatic material) groups.

The experimental animals were returned each to their in-dividual cages. They were free to move about and eat anddrink ad libitum. The surgical wounds were examined, in-cluding a check for any lower limb paralysis.

Histologic evaluation of inflammatory cell infiltration tothe collagen hemostats and resorption

Histologic analysis was performed to evaluate inflamma-tory cell responses to the collagen hemostats and their resorp-tion when placed in the epidural space. At 1, 2, 4, and 8 weeksPO, MCH (n55), CCH (n55), and control (n54) group

3K. Mizuno et al. / The Spine Journal - (2014) -

animals were euthanized with an overdose of anesthesia. Thevertebral bodies fromL7 to S1were excised and fixed in Zam-boni solution at 4�C for 1 week. These specimens were decal-cified in an aqueous solution of 22.5 vol% formic acid and 10wt/vol% sodium citrate. The first sacral vertebra (S1) was cutin cross section and embedded in paraffin.

Paraffin sections (3 mm thick) were prepared and stainedwith hematoxylin-eosin (HE). These were examined withlight microscopy (BX53; Olympus, Tokyo, Japan) to eval-uate resorption of the hemostatic material, inflammatorycell distribution (neutrophils and macrophages) in the epi-dural space, and the condition of the nerves and dura mater.

Evaluation of the immunologic reaction to collagenhemostats placed in the epidural space

To evaluate the immunologic and inflammatory re-sponses for the placement of the topical hemostatic materi-als placed in the epidural space, IHC staining forinflammatory cytokines (tumor necrosis factor [TNF]-a, in-terleukin [IL]-6), cyclooxygenase (COX)-2, and macro-phages (CD68) was performed.

The 3-mm paraffin sections were deparaffinized and im-mersed in phosphate-buffered saline (PBS). After antigenactivation, endogenous peroxidase was deactivated withPBS containing 0.3 vol% hydrogen peroxide. Next, toblock nonspecific binding, 10 vol% normal goat serum(Jackson Immunoresearch Laboratory, West Grove, PA,USA) in PBS for anti-TNF-a and anti-COX-2 antibodiesand 10 vol% normal rabbit serum (Vector Laboratories,Burlingame, CA, USA) in PBS for anti-IL-6 and anti-CD68 antibodies were added to the primary antibodiesand incubated overnight. The primary antibodies used inthis study were Armenian hamster anti-TNF-a (1:500, sc-12744; Santa Cruz Biotechnology), goat anti-IL-6(1:1600, sc-1265; Santa Cruz Biotechnology), goat anti-COX-2 (1:200, sc-1747; Santa Cruz Biotechnology), andmouse anti-CD68 (1:50, EBM11; Dako Cytomation,Glostrup, Denmark).

After washing with PBS, the sections for TNF-a wereincubated with peroxidase-conjugated anti-Armenian ham-ster IgG antibody as the secondary antibody and avidin-biotinylated enzyme complex (Vectastain ABC System;Vector Laboratories) for 30 minutes. For IL-6 and COX-2, Histofine Simple Stain MAX-PO (G) (anti-goat; Nichir-ei, Tokyo, Japan) and for CD68, Histofine Simple StainMAX-PO (M) (anti-mouse; Nichirei) were used for the re-actions. The sections were developed with nickel 3,30-dia-minobenzidine and counterstained with nuclear fast red.

MRI evaluation of fluid accumulation or inflammation inthe epidural space

To evaluate exudate accumulation and the degree of lo-cal inflammation because of placement of the topical hemo-stats in the same animals, imaging was performed with highmagnetic field MRI.

At 1, 2, 4, and 8 weeks PO, high magnetic field MRI(Varian MRI system 7.04 T; Agilent Technologies, Palo Al-to, CA, USA) was performed in the MCH (n53), CCH(n53), and control (n53) group animals under anesthesiaand sedation. The animals were sedated by inhalation anes-thesia and placed in a prone position. A 3�4 cm diametertransmit/received surface coil (Takashima Seisakusho, To-kyo, Japan) was placed on the dorsal midline at the levelof L7/S1, and T1-weighted (T1WI) and T2*-weighted(T2*WI) axial imaging was performed. Axial T1WI wasperformed using the following parameters: repetitiontime/echo time 102.4/5.14 ms, matrix 512�512 (zero fill),slice thickness 2 mm, slice gap 0 mm, field of view60�60 mm, and signal average 4. Axial T2*WI was per-formed using the following parameters: repetition time/echo time 350/15 ms, matrix 512�512 (zero fill), slicethickness 2 mm, slice gap 0 mm, field of view 60�60mm, and signal average 4. After MR imaging, the experi-mental animals were each returned to their individual ca-ges. They were free to move about and eat and drink adlibitum.

To evaluate exudate accumulation and the degree of in-flammation in the epidural space, imaging software (Agi-lent software, VnmrJ, version 3.1, revision A; AgilentTechnologies) was used. Mean signal intensity on axialT2*WI was measured. For axial T1WI used to evaluate be-tween the L7/S1 vertebral arches, a rectangular region ofinterest (ROI) 1 mm dorsally from the dura between the in-tervertebral joints bilaterally was established. For thesuperior articular process of the first sacral vertebra (S1),a 1-mm [2] round ROI was established. Refering toT1WI location, mean signal intensity in each ROI was cal-culated from axial T2*WI (Fig. 1). Mean signal intensityfor the ROI in the epidural space was standardized by themean signal intensity in the superior articular process.The images were examined separately by two spinal sur-geons. When there was a disagreement in an assessment,a consensus was reached after deliberation.

Statistical analysis

Mean signal intensity of the epidural space on MRI wascompared using one-way analysis of variance. Fisher pro-tected least significant difference was used for multiplecomparison testing. The level of statistical significancewas p value less than .05. Statistical analysis was per-formed using SPSS statistics 20 (IBM Ltd, Tokyo, Japan).All results are expressed as the mean6standard deviation.

Results

The surgical wounds healed in all animals, and therewere no deep-seated infections. After surgery, there wasno lower limb paralysis, and body weight progressivelyincreased.

Fig. 1. Region of interest (ROI) established in the epidural spinal space on axial magnetic resonance (MR) images. On axial T2*-weighted MR images, a

rectangular ROI 1 mm dorsally from the dura between the L7/S1 vertebral arches and a 1-mm round ROI in the superior articular process of the first sacral

vertebra (S1) were established refering to T1WI location. The mean signal intensity of the ROI in the epidural space was standardized by comparison with the

signal intensity of the superior articular process.

4 K. Mizuno et al. / The Spine Journal - (2014) -

Histologic evaluation of HE staining

To evaluate inflammatory cell infiltration and resorptionof the collagen hemostats placed in the epidural space,cross-sectioned tissue stained with HE was examined withlight microscopy.

MCH groupAt 1 week PO, swelling of the hemostatic material in the

epidural space under the vertebral arches was observed. Inall cases, there was inflammatory cell infiltration, predom-inantly neutrophils, near the hemostatic material. At 2weeks PO, there were still inflammatory cells near theshrunk hemostatic material and granuloma production withmany multinucleated giant cells. At 4 and 8 weeks PO, thehemostatic material had disappeared in all cases, the num-ber of inflammatory cells decreased, but granuloma wasstill seen (Fig. 2, Left).

CCH groupAt 1 week PO, swelling of the hemostatic material under

the vertebral arches, similar to the MCH group, was ob-served. Mild inflammatory cell infiltration was present inall animals. At 2 weeks PO, the hemostatic material hadshrunk, and no migration of inflammatory cells towardthe hemostatic material was seen. At 4 weeks PO and later,the hemostatic material had disappeared in all cases, and fi-brous tissue was present without inflammation in the epi-dural space (Fig. 2, Right).

Control groupThe epidural space was covered by fatty tissue in all an-

imals. The spinal cord and dura were normal. Mild inflam-mation was seen in the epidural space at 1 and 2 weeks POand then decreased.

Immunohistochemical evaluation of TNF-a, IL-6, COX-2, and CD68

To evaluate foreign-body and inflammatory responses tothe topical hemostatic materials in the epidural space, IHC

staining for TNF-a, IL-6, COX-2, and CD68 wasperformed.

MCH groupCells positive for TNF-a surrounding the hemostatic

material were seen up to 1 week PO. Inflammatory cellsand multinucleated giant cells surrounding the hemostaticmaterial were positive for IL-6, COX-2, and CD68 at allobservation times (Fig. 3, A and B).

CCH groupNo TNF-a–positive cells in the epidural space near the

hemostatic material were seen at any observation time. In-terleukin-6–positive cells in the epidural space near the he-mostatic material were seen until 2 weeks PO. At 4 weeksPO, fibroblasts positive for IL-6 were seen, but at 8 weeksPO, these were almost disappeared (Fig. 3, C). Cyclooxyge-nase-2–positive inflammatory cells in the epidural spacewere present at 1 week PO, but these positive cells subse-quently disappeared. CD68 was negative at 2 weeks POand later (Fig. 3, D).

Control groupNo TNF-a–positive cells were present at any observa-

tion time. Interleukin-6–positive cells were observed inthe interstitial cells of the epidural fat at 1 week PO, butthese were negative at 4 weeks PO and later (Fig. 3, E).Cyclooxygenase-2 was positive until 4 weeks PO, andCD68 was positive until 1 week PO (Fig. 3, F).

MRI analysis of the mean signal intensity within the ROIestablished in the spinal epidural space

To evaluate exudate accumulation and the degree of in-flammation in the epidural space, serial imaging with highmagnetic field MRI was performed in the experimental an-imals under anesthesia and sedation (Fig. 4).

In both the MCH and CCH groups, the mean signal in-tensity in the epidural space was the highest at 1 week POand then decreased over time. At 1 week PO, the mean

Fig. 2. Histologic examination of the epidural space after placing the collagen hemostat. (Left) Microfibrillar collagen hemostat (MCH) group—1 week post-

operative (PO): there is swollen hemostatic material in the epidural space and extensive inflammatory cell infiltration; 2 weeks PO: there are many inflammatory

and multinucleated giant cells near the shrunk hemostatic material; 4 weeks PO: the fibrous component of the hemostatic material has disappeared, but there is

granuloma (black arrow); and 8 weeks PO: granulomatous tissue is still seen in the epidural space (black arrow). (Right) Cotton-type collagen hemostat (CCH)

group—1 week PO: there is no inflammatory cell infiltration near the hemostatic material or surrounding tissue; 2 weeks PO: there are no inflammatory cells, but

fibrous tissue is present in the epidural space; and 4 and 8 weeks PO: the hemostatic material has disappeared, and there is fibrous tissue in the epidural space.

5K. Mizuno et al. / The Spine Journal - (2014) -

signal intensities in the epidural space were 5.4260.43,4.3760.63, and 2.3960.33 in the MCH, CCH, and controlgroups, respectively. The mean signal intensities in theMCH and CCH groups were significantly higher than inthe control group (p!.01). In the MCH group, the meansignal intensity was significantly higher than in the CCHgroup (p!.05) (Fig. 5).

At 2 weeks PO and later, the mean signal intensity in theepidural space did not significantly differ among the threegroups.

Discussion

Resorption of the two types of collagen hemostats wassimilar when placed in the epidural space, but tissue bio-compatibility was clearly different. Absorbable topical he-mostats are often used for bleeding that is difficult to

control with surgical ligation or coagulation hemostasis us-ing electrocauterization. Topical hemostats are used in sur-gery to control bleeding from cut surfaces of solid organssuch as the liver and spleen. Many types and configurationsof topical hemostats are now commercially available. Inspinal surgery, local hemostatic materials are often used es-pecially to control bleeding from the epidural venousplexus during laminectomy or intervertebral fusion, andcollagen and gelatin are commonly used. Gelatin has beenreported to have merits because of its hydrophilicity, lowantigenicity, resorbability, biocompatibility, activation ofblood coagulation ability, and other factors [1,2,13–16].Although similar merits have also been reported for colla-gen, antigenicity of MCH has previously been reported[3–8,17,18]. Therefore, in the present study, we comparedMCH with CCH, a collagen hemostat that also uses colla-gen obtained from bovine dermis as the raw material buthas a different shape and purification process.

Fig. 3. Immunohistochemical evaluation of tumor necrosis factor (TNF)-a, interleukin (IL)-6, cyclooxygenase (COX)-2, and CD68. Microfibrillar collagen

hemostat (MCH) group—(A) cells positive for TNF-a (black arrow) were present near the hemostatic material and in granulomatous tissue until 1 week

postoperative (PO); IL-6–positive cells (black arrow) were seen at all observation times; and (B) inflammatory and multinucleated giant cells positive for

COX-2 and CD68 (black arrow) were seen at all observation times. CCH group—(C) no TNF-a–positive cells near the hemostatic material were seen at

any observation time; there were IL-6–positive cells (black arrow) near the hemostatic material until 2 weeks PO; at 4 weeks PO, there were positive fibro-

blasts; staining for IL-6 was negative at 8 weeks PO; (D) COX-2–positive inflammatory cells (black arrow) in the epidural space were present at 1 week PO;

at later times, there were almost no positive cells; CD68-positive cells (black arrow) in the epidural space were present at 1 week PO; and at later times,

staining for CD68 was negative. Control group—(E) no TNF-a–positive cells were present at any observation time; there were IL-6–positive cells (black

arrow) in the epidural space until 2 weeks PO; and (F) there were COX-2–positive cells (black arrow) until 4 weeks PO and CD68-positive cells (black arrow)

until 1 week PO.

6 K. Mizuno et al. / The Spine Journal - (2014) -

Microfibrillar collagen hemostat is a hemostatic materialdeveloped by purifying collagen so that the natural fibrilcross-links are maintained. On the other hand, CCH wasmade from atelocollagen, in which telopeptides, the mainantigenic determinant site in collagen, had been removed.Their characteristics are different. In an animal study usinga wound-healing model in which hemostatic materials wereplaced in the auricle of white rabbits, Tomizawa et al.[1,10] reported that biocompatibility, resorption, vascularinvasion, and tissue healing differed based on propertiesand configuration of the hemostatic material.

Regarding resorption of the collagen hemostats, ourstudy showed disappearance of the topical hemostats in

both groups by 4 weeks PO, without any obvious differ-ence. However, the histologic and MRI findings duringthe resorption process were different. In the MCH group,there was inflammatory cell infiltration near the hemostaticmaterial at all observation times, and granuloma formationwas observed. IHC staining showed IL-6–, COX-2–, andCD68-positive cells near the hemostatic material at all ob-servation times. The hemostatic material was phagocytizedby macrophages because of an antigen-antibody reaction,and granulomatous inflammation occurred as the responseto the remaining hemostatic material. In the CCH group,no inflammatory cell infiltration near the hemostatic mate-rial was seen, and IHC staining showed no positive cells at

Fig. 3. (continued).

7K. Mizuno et al. / The Spine Journal - (2014) -

4 weeks PO and later, when the hemostatic material wasresorbed.

Magnetic resonance images showed that the mean signalintensity of the epidural space in the MCH group was signifi-cantly higher than that in the CCH group at 1 week PO. Re-garding hemostat resorption, MCH absorbs fluid(physiologic saline) 12 times its own weight and CCH 32times its own weight. Resorption of CCH was reported to bebetter [1]. Considering this fluid absorption, the mean signalintensity in the epidural space would be expected to be higherin the CCH group than in the MCH group. However, our re-sults were in conflict with this. We believe that exudate accu-mulation because of local inflammation as a result of MCHplacement was reflected by the results of MRI experiments.

In rat studies, granuloma formation with placement ofMCH under the intracranial dura and in dental sockets hasbeen reported [11,12]. Formation of granuloma with place-ment of MCH in the spinal canal has also been clinically re-ported [3–8]. On the other hand, in CCH, albumin impuritiesare removed by enzyme treatment during solubilization,

which further reduces antigenicity. To our knowledge, gran-uloma formation because of the placement of CCH in a sur-gical field has not been reported. Besides the use as ahemostat, Fujiu andMiyamoto [19] applied CCH to air leaksafter pulmonary resection and reported only mild tissue in-flammation and adhesions. Regarding hemostat configura-tion, better biocompatibility and tissue healing with CCHscompared with MCHs have also been reported [1].

In addition to HE staining, IHC staining and MR imagesin this study showed clear differences in biocompatibilitybetween the two types of hemostatic materials. In theMCH group, IHC staining showed a strong inflammationreaction to the hemostatic material, and MR images re-vealed high signal intensity that was attributed to exudateaccumulation in the epidural space. Thus, a foreign-bodyreaction was elicited to the hemostatic material. However,in the CCH group, the collagen fibers were resorbed duringthe healing process without any foreign-body reaction.

As also described in the previous case reports, a majorconcern with the placement of hemostatic materials during

Fig. 3. (continued).

8 K. Mizuno et al. / The Spine Journal - (2014) -

spinal surgery is nerve compression because of swelling ofthe hemostatic material. However, inflammatory cytokines,which were confirmed in our study, are also pain-inducingsubstances [20]. Local inflammatory reactions triggered byhemostatic materials, depending on the site of hemostatplacement, may be a cause of PO neuropathic pain. When

Fig. 4. Magnetic resonance images at 1 week postoperative (PO) high signal in

hemostat and cotton-type collagen hemostat groups until 2 weeks PO (white arr

a hemostatic material is placed in a surgical field, PO inflam-mation in the epidural space because of retained hemostaticmaterial should be considered based on the magnitude of sig-nal intensity in the epidural space on PO MR images.

Surgeons can select from a variety of topical hemostaticmaterials depending on their hemostatic efficacy and ease of

tensity in the epidural space were seen in both the microfibrillar collagen

owhead).

Fig. 5. Evaluation of the mean signal intensity within the region of inter-

est established in the spinal epidural space. Serial mean signal intensity of

the epidural space was compared among the three groups. At 1 week post-

operative, the mean signal intensities in the microfibrillar collagen hemo-

stat (MCH) and cotton-type collagen hemostat (CCH) groups were

significantly higher than that in the control group (p!.01). The mean signal

intensity in the MCH group was significantly higher than that in the CCH

group (p!.05). Bar shows mean6standard deviation of the mean. NS, not

significant; *p!.05; **p!.01.

9K. Mizuno et al. / The Spine Journal - (2014) -

handling. However, when placing collagen hemostats in theepidural space for PO hemostasis, the characteristics of a he-mostatic material should be carefully considered before use.

Acknowledgment

The authors thank Mr Kouji Koike (Department of Path-ophysiology, Institute of Bio-active Science, Nippon ZokiPharmaceutical Co. Ltd) for his technical assistance ofthe histologic stains.

References

[1] Tomizawa Y. Clinical benefits and risk analysis of topical hemostats:

a review. J Artif Organs 2005;8:137–42.

[2] Gabay M, Boucher BA. An essential primer for understanding the

role of topical hemostats, surgical sealants, and adhesives for main-

taining hemostasis. Pharmacotherapy 2013;33:935–55.

[3] Awwad EE, Smith KR Jr. MRI of marked dural sac compression by

Surgicel in the immediately postoperative period after uncomplicated

lumbar laminectomy. J Comput Assist Tomogr 1999;23:969–75.

[4] Alander DH, Stauffer ES. Gelform-induced acute quadriparesis after

cervical decompression and fusion. Spine 1995;20:970–1.

[5] Friedman J, Whitecloud TS III. Lumbar cauda equine syndrome as-

sociated with the use of Gelfoam: case report. Spine 2001;26:

E485–7.

[6] Doita M, Nishida K, Kurosaka M. Radiculopathy due to microfibril-

lar collagen hemostat mimicking recurrence of disc herniation. Skel-

etal Radiol 2006;35:953–5.

[7] Sani S, Boco T, Lewis SL, et al. Postoperative acute disseminated en-

cephalomyelitis after exposure to microfibrillar collagen hemostat. J

Neurosurg 2008;109:149–52.

[8] Menovsky T, Plazier M, Rasschaert R, et al. Massive swelling of Sur-

gicel� Fibrillar� hemostat after spinal surgery. Case report and a re-

view of the literature. Minim Invasive Neurosurg 2011;54:257–9.

[9] Tomizawa Y, Komori M, Takada K, et al. In vivo evaluation of col-

lagen hemostats: biocompatibility and resorption. [in Japanese]. Jpn

J Cardiovasc Surg 2003;32:17–22.

[10] Tomizawa Y, Komori M, Takada K, et al. In vivo and in vitro evalua-

tion of the biocompatibility and cytotoxicity of local hemostatic

agents. [in Japanese]. Jpn J Cardiovasc Surg 2004;33:382–6.

[11] Magro-Ernica N, Magro-Filho O, Rangel-Garcia I. Histologic study

of use of microfibrillar collagen hemostat in rat dental sockets. Braz

Dent J 2003;14:12–5.

[12] Ereth MH, Schaff M, Ericson EF, et al. Comparative safety and effi-

cacy of topical hemostatic agents in a rat neurosurgical model. Neu-

rosurgery 2008;63:369–72.

[13] Ibarrola JL, Bjorenson JE, Austin BP, Gerstein H. Osseous reactions

to three hemostatic agents. J Endod 1985;11:75–83.

[14] Alpaslan C, Alpaslan GH, Oygur T. Tissue reaction to three subcuta-

neously implanted local hemostatic agents. Br J Oral Maxillofac Surg

1997;35:129–32.

[15] Schonauer C, Tessitore E, Barbagallo G, et al. The use of local

agents: bone wax, gelatin, collagen, oxidized cellulose. Eur Spine J

2004;13(1 Suppl):S89–96.

[16] Miyata T, Taira T, Noishiki Y. Collagen engineering for biomaterial

use. Clin Mater 1992;9:139–48.

[17] Nakajima M, Kamei T, Tomimatu K, Manabe T. An intraperitoneal

tumorous mass caused by granulomas of microfibrillar collagen he-

mostat (Avitene). Arch Pathol Lab Med 1995;119:1161–3.

[18] DeLustro F, Condell RA, Nguyen MA, McPherson JM. A compara-

tive study of the biologic and immunologic response to medical de-

vices derived from dermal collagen. J Biomed Mater Res 1986;20:

109–20.

[19] Fujiu K, Miyamoto H. Microscopic findings of sheet-type collagen

applied to air leaks after pulmonary resection. Ann Thorac Cardio-

vasc Surg 2012;18:212–5.

[20] Guillot X, Semerano L, Decker P, et al. Pain and immunity. Joint

Bone Spine 2012;79:228–36.