Biomaterials 23 (2002) 2615–2619

Effects of preservation on the mechanical strength and chemicalcomposition of cortical bone: an experimental study in sheep femora

Jose Moreno, Francisco Forriol*

Orthopaedic Research Laboratory, School of Medicine, University of Navarra, Pamplona, Spain

Received 15 May 2001; accepted 22 November 2001

Abstract

Preservation methods have enabled bone banks to furnish cortical bone grafts to orthopaedic surgeons. However, cortical bone

preserved by freezing and autoclaving, may be weakened by these treatments. To test this hypothesis we compared the ultimate

tensile strength of freshly harvested sheep femora with that of femora which were frozen at �201C for 60 days, or autoclaved at

1341C for 8min.

We measured the collagen and mineral contents (calcium, phosphorus, magnesium) and hydroxyproline of the specimens and

tested for changes induced by preservation. Mechanical three point tests showed that frozen femora were significantly stronger than

either fresh or autoclaved femora ðpo0:05Þ: Frozen specimens also had the highest phosphorus level, indicating these measures are

related to strength.

Cortical bone is not significantly weakened by autoclaving or freezing. This result does not imply that preserved grafts are

clinically interchangeable with fresh grafts, rather, it suggests that future studies should focus on post surgical issues, such as the rate

of remodeling and integration, which may be sensitive to preservation technique. r 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Bone allograft; Frozen; Dry freezing; Autoclaving; Calcium; Magnesium; Hydroxyproline; Phosphorus

1. Introduction

Cortical bone allografts are coming into ever morefrequent use. Simple preservation and storage techni-ques make it possible for tissue banks to deliver corticalbone specimens whenever they are needed, however, thephysical changes wrought by preservation proceduresare a source of concern. Because cortical allografts areused to restore skeletal integrity, often under weight-bearing conditions, the mechanical strength of the graftis crucial. Grafts weakened by preservation proceduresmay be liable to fracture.

Changes in the mechanical properties of bone dependon underlying changes in its composition and structuralorganization [1]. Mineral content and amount of bonedetermines bone density, which is important for themechanical properties of bone [2–6].

Different preservation methods are currently used inbone banks, although their effects on allograft qualityare not fully understood. To be broadly applied suchmethods must be simple, inexpensive, not toxic, sterile,and should effectively preserve the biological andmechanical attributes of the bone.

Bone banks typically store bones in sterile conditionsat temperatures below �401C for periods longer than 6months or between �181C and �281C for periodsshorter than 6 months [4]. In bone conserved at �201Cthere is no change in physical or mechanical propertiesbut Laforest et al. [7] found deep freezing at –801Cdecreases the cortical bone stiffness.

Many authors believe that freezing and freeze-dryingare inadequate to prevent infection and insist that bonefor allografts should be sterilized by autoclaving,irradiation [3,8] or chemical treatments [9]. Unfortu-nately, while the processes of freezing and lyophilizationdo not alter the mechanical properties of bone, heat andirradiation weaken grafts. Voggenreiter et al. [10]indicated autoclave treatment, at 1341C for 5min,reduces bone stiffness (28% of the control bone),although treatment for only 3min maintains acceptable

*Corresponding author. Dpt COT, Clinica Universitaria, University

of Navarra, Apdo 192, 31080 Pamplona, Spain. Tel.: +1-349-48-274-

650; fax: +1-349-48-172-294.

E-mail address: fforriol@unav.es (F. Forriol).

0142-9612/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.

PII: S 0 1 4 2 - 9 6 1 2 ( 0 1 ) 0 0 4 0 2 - 1

properties. The amount of radiation needed to deacti-vate HIV in bone is unknown, although it would seemthat, at least in vitro, over 3Mrad are necessary [11,12].Some authors report that while such high radiationdoses destroy most bacteria and viruses, they maydiminish stiffness to compression by up to 50% [12].Lower radiation doses (0.01–0.05Mrad) did not appearto reduce the ultimate tensile strength of cortical bone[10].

Heat treatments may also slow the process ofintegration with host bone [13], although Taguchi et al.[14] found that autoclaved grafts serve as a scaffold forgraft incorporation. Boiling bone caused a decrease instiffness [3] and according to Borchers et al. [15] corticalbone autoclaving and boiling reduces bone strength butfreezing and freeze-drying increase bone stiffness [16].

The aim of our study was to determine howconservation methods (freezing and autoclaving) affectthe mechanical properties of the bone grafts. We deviseda systematic experimental approach using a reproduci-ble model (sheep femora) to test the hypothesis thatcortical bone is significantly weakened by freezing andautoclaving. Mechanical strength was measured andfollowed with extensive chemical analyses of the speci-men to determine if any changes in mechanical proper-ties depended on changes in the mineral composition.

2. Materials and methods

We used 54 left femora from 3–4 month-old sheep anda 5 cm long bone strip were obtained from the anterioraspect of each femur diaphysis. In each strip, thethickness of the wall, and the width and height of thespecimen, were measured in three different sections and21 were analyzed fresh, 21 frozen at –201C during60 days and 12 were autoclaved at 1341C during 8min.

The three-point loading test was used to determine theultimate tensile strength of bone strips. Tension atbreakage was measured using a test machine (Instron4502s) with a three points clamp (Instrons, UK, ref.no. 2810–300). The speed of displacement of thecrosshead was 5mm/min, and the environmental con-ditions were the same in all cases. Once the geometricdata for each strip of bone had been ascertained, wecalculated the moment of inertia of the central section,beginning by approximating the section of the femoralcylinder to a regular section. To calculate the moment ofinertia, we first determined the position of the centre ofmasses in the section. We then determined the momentof inertia with respect to the geometrical centre of thecomplete section of the bone without cutting it, andapplied Steiner’s theorem of parallel axes. Using theforce at breakage, the moment of flexion, and themoment of inertia, we determined the maximum tensilestress at failure. Since the supports were equidistant

from the point of loading, the reaction at each supportpoint is half of the applied load and its lever arm is halfthe distance between the support points.

Mineral contents (calcium, magnesium, hydroxypro-line and phosphorus), collagen content, and hydroxya-patite content were determined. Each specimen testedwas deffated and ashed. Bone ashes were assayed formineral contents by atomic absorption spectrophoto-metry [2,17,18]. The spectrophotometer (Perkin–Elmer460) was set to read at a wavelength of 422 nm tomeasure calcium and 285 nm to measure magnesium.Calcium (1000mg/l, Mercks) and magnesium (1000–0.002 g/l, Panreacs) solutions were used as standards.

Hydroxyproline assays were performed on dried bonespecimens by Woessner’s colorimetric method [19–21]using an Ultrospec Plus Spectrophotometer (PharmaciaLKBs) and trans-4-hydroxy-l-proline (1.0mg/ml in0.001n HCl, Sigma) as a standard. Phosphorus levelswere obtained by colorimetry [4] using K2HPO4 indeionized water (10 mg/ml) as a standard.

Statistical analysis was carried out on the valuesobtained, with reference to the preservation methodapplied to each specimen and the data from themechanical tests and those from the chemical analysis.

The statistical package used was SPSS 4.0s forMacintoshs. We checked whether the data obtainedfollowed a normal distribution using the Shapiro–Wilkstest and found the requirements of homoscedasticitywere fulfilled using Levene’s test for the homogeneity ofvariances. We then performed a descriptive analysis andthe values of each of the variables studied werecompared for the different forms of preservation andthe different aspects using two-way ANOVA. In thosevariables in which statistically significant differenceswere found, we performed multiple ‘post hoc’ compar-isons using Tukey’s test.

In addition, we studied the possible correlationsbetween each of the chemical variables (calcium,magnesium, hydroxyproline, phosphorus) and the me-chanical forces and each of the chemical variablesanalyzed using Pearson’s r correlation test, analyzing allthe possible combinations two by two. Probabilityvalues of o0.05 were taken as significant.

3. Results

Analyzing the bones strips according to the treatmentapplied, the group with the highest ultimate tensilestrength was the frozen group (317MPa), then the freshgroup (250MPa), and last the autoclave group(234MPa). The frozen bones were the most stiffnessðpo0:05Þ: There were no significant differences ðp >0:05Þ between the fresh bones and the autoclaved(Fig. 1).

J. Moreno, F. Forriol / Biomaterials 23 (2002) 2615–26192616

The data obtained in the chemical analysishad a normal distribution pattern ðp > 0:05Þ for allvariables, which meets the requirements for homosce-dasticity ðp > 0:05Þ: No significant differences werefound when we compared the amount of calciumðp ¼ 0:29Þ; magnesium ðp ¼ 0:31Þ; and hydroxyprolineðp ¼ 0:12Þ contained in the fresh, frozen andautoclaved groups, which indicates that the amount ofthese substances is not affected by the preservationtreatment.

On the other hand, the preservation process affectedthe phosphorus content. There were statistically highlysignificant differences in the amounts of phosphorus(p=1.0E�4) found in the differently treated specimensof bone (Table 1). The highest levels were found in thefrozen group (13.42%), which was seen to havesignificant differences with respect to the autoclavegroup (12.83%) and the fresh group (13.05%). Nosignificant differences were found between these last twogroups ðp > 0:05Þ: The chemical data show a moderate,highly significant, positive correlation (r ¼ 0:4326;po0:01) between the values for calcium and themagnesium values in the specimens analyzed. Magne-sium was found to have a low significant and negativecorrelation with the hydroxyproline values(r ¼ �0:1894; po0:05).

In the study of the relationship between the values forultimate tensile strength and the levels of the chemicalcomponents analyzed in these specimens, we detected alow significant, negative correlation (r ¼ 0:1599;po0:05) between the ultimate tensile strength and theamount of hydroxyproline. We also found a highlysignificant correlation (r ¼ �0:3704; po0:01) betweenthe ultimate tensile strength and the amount ofphosphorus level also had a highly significant correla-tion (r ¼ �0:3324; po0:01) with the ultimate tensilestrength.

4. Discussion

There is no reason why the method used forpreserving bone tissue should change the mineralcontent of the bone, although it is possible to findchanges in the chemical components and alterations inthe hydroxyproline and collagen levels caused byheating or freezing of the samples.

The mechanical properties of a material depend on itscomposition and structural organization. The composi-tion includes porosity and mineralization, whereasorganization covers the type of bone, whether it iscortical or cancellous [1]. Since the bones in our studywere all of the same type, we shall focus here on themineral contents, which are related to the mechanicalproperties of the bone. Schaffler and Burr [6] foundexponential increases in bone stiffness with the increasein minerals, while Currey et al. [2,3] discovered acorrelation between the modulus of elasticity and thecalcium content. McCalden et al. [5] found changes ofthe mechanical properties of the cortical bone and age,and significant relation between porosity and bonestrength.

Amtmann [22] describe variations in the distributionof density in the human femur, caused by irregularcalcium distribution. The mineral content has a relation-ship to density, which is an important variable indetermining the mechanical properties of bone, such asthe modulus of elasticity, the creep and the breakageload. Authors [2,4,6,13,23] have found a relationshipbetween density and the modulus of elasticity, detectinga correlation between strength and stiffness, porosityand mineralization. For Cubo and Casinos [24] magne-sium and phosphorus appear to be the most importantinorganic elements in relation to the mechanical proper-ties of the avian long bones. McCalden et al. [5] foundthat mineral content is not related to age and does nothave effect on the mechanical properties of ageing bone.

Markel et al. [25,26] found a strong correlationbetween the calcium content in the repair of a fracture,and the mechanical properties. The increase in stiffnessin the fracture callus is due to the ‘‘maturation’’ of thehydroxyapatite crystals. Their studies show that calcifi-cation is a process which develops over a period of time,so that the bone gradually develops adequate strength.

We now found differences between the types ofpreservation method, as far as calcium, magnesiumand hydroxyproline were concerned. However, the levelsof phosphorus were found to differ significantly betweenthe preservation methods.

We observed that frozen femora were significantlystronger than either fresh or autoclaving femoraðpo0:05Þ: Frozen specimens also had the highestmineral content. Mineral composition and mechanicalstrength of cortical bone are both significantly altered byfreezing. The data also indicate that autoclaving, which

Fig. 1. Cortical bone bending strength.

J. Moreno, F. Forriol / Biomaterials 23 (2002) 2615–2619 2617

had no significant effect on either composition orstrength. In summary, our study indicates that simplepreservation techniques do not adversely affect initialgraft strength.

The higher level of phosphorus in the frozen group isperhaps due because at temperatures of �201C enzymeactivity does not stop completely, and may act on thephosphorus, changing its state in the bone by breakingits union to the collagen matrix. Freezing may alter thehydroxyapatite crystals, causing them to become un-structured and to return to a state similar to that ofamorphous calcium phosphate, which is the unstableprecursor in solution of hydroxyapatite [27]. Whereascalcium is strongly bonded to the matrix, phosphorusdissolves more easily [28]. In the autoclaved treatment,heat seems not affect the state of hydroxyproline andcollagen, but it is not possible to say the same for itsstructure, which may be altered.

On the other hand, magnesium has a highly sig-nificant positive correlation with calcium as described byLappalainen and Knuuttila [29]. Posner [27] indicatethat a molar relationship Mg/Ca >0.2 prevents theformation of apatite from highly saturated solutions ofcalcium phosphate. We also found correlation betweenmagnesium and hydroxyproline, which were negative.

Three-point load test is a straightforward method,which produces a compression stress in the upper partand a tension stress in the lower part of the specimen.This is widely used to study biological materials, as theresults are reproducible (coefficient of variation between10% and 20%) [2,3,10] in comparison with torsion testswhich present coefficients of variation of up to 50%.

Bone boiling is a simple procedure, which trans-formed collagen into gelatine, and there is major loss ofwater. However, according to Taguchi et al. [14]autoclaved grafts serve as a structural element fordeveloping new cells.

Autoclaving reduces the compression module by 59%and compression strength by 58%, while in boiledtrabecular bone compression strength is reduced by26%, but the modulus of elasticity is not affected.Bending stiffness in boiled cortical human bone decreasethe stiffness but the preservation of bone grafts attemperatures of �201C to �1471C did not affect theirmechanical properties [3].

Jerosh et al. [23] studied the mechanical differencesbetween cortical bone preserved in different ways, andobserved that its ultimate tensile strength fell 8.7% whensterilized with autoclave, increased 18.9% with lyophi-lization, and 3.4% with lyophilization and argongamma radiation. It fell by 1.7% when this gas wasnot used. To draw these comparisons, they used controlgroups of bones frozen at �251C, and concluded thatbending stiffness did not vary with the temperature atwhich the bone was preserved.

Itoman and Nakamura [16] found, in rats, the bonestiffness increased after freezing to –801C and freeze-drying. On the other hand, decalcified bone was verysoft.

Our own results confirm a strength increase in frozenbone with a correlation with the chemical composition.As magnesium seems to play a part in the calcification ofthe matrix and therefore in the strength of the bone,when calcium increases, magnesium increases too, whichmakes for greater stiffness in the bone, while thehydroxyproline and collagen decrease, thereby loweringthe bone’s elasticity. The reverse would also be true.

References

[1] Martin RB. Determinants of the mechanical properties of bone.

J Biomech 1991;24:79–88.

[2] Currey JD, Brear K, Zioupos P. The effects of ageing and changes

in mineral content in degrading the toughness of human femora.

J Biomech 1995;29:257–60.

[3] Currey JD, Foreman J, Laketic I, Mitchell J, Pegg DE, Reilly GC.

Effects of ionizing radiation on the mechanical properties of

human bone. J Orthop Res 1997;15:111–7.

[4] Harrington KD, Johnston JO, Kaufer HN, Luck JV, Moore TM.

Limb salvage and prosthetic joint reconstruction for low-grade

and selected high-grade sarcomas of bone after wide resection and

replacement by autoclaved autogeneic grafts. Clin Orthop

1986;211:180–214.

[5] McCalden RW, McGeough JA, Barker MB, Court-Brown CM.

Age related changes in the tensile properties of cortical bone.

J Bone Jt Surg (Am) 1993;75-A:1193–205.

[6] Schaffler MB, Burr DB. Stiffness of compact bone: effects of

porosity and density. J Biomech 1988;21:13–6.

[7] Laforest P, Kempf JF, Follea G, Karger C, Kempf I.

Comparaison des qualit!es m!ecaniques de l’os cortical de banque

suivant la m!ethode de cong!elation. Rev Chir Orthop 1991;77:

389–95.

Table 1

Percentage mass of dried specimens of calcium, magnesium, hydroxyproline and phosphorus

Calcium Magnesium Hydroxyproline Phosphorus

n %x SD %x SD %x SD %x SD

Fresh 21 24.19 10.30 0.43 0.18 2.75 1.08 13.05 6.07

Frozen 21 24.94 8.45 0.44 0.17 2.70 1.10 13.42 5.06

Autoclave 12 25.12 10.66 0.42 0.19 2.80 1.29 12.84 5.07

J. Moreno, F. Forriol / Biomaterials 23 (2002) 2615–26192618

[8] Zasacki W. The efficacy of application of lyophilized, radiation-

sterilized bone graft in orthopaedic surgery. Clin Orthop

1991;272:82–7.

[9] Zhang Q, Cornu O, Delloye Ch. Ethylene oxide does not

extinguish the osteoinductive capacity of demineralized bone.

Acta Orthop Scand 1997;68:104–8.

[10] Voggenreiter G, Ascherl R, Fr .uh HJ, Bl .umel G, Schmit-

Neuerburg KP. Konservierung und Sterilisation von Kortikalis-

biomechanische Untersuchungen an der Ratte. Unfallchir

1995;98:53–8.

[11] Tomford WW. Transmission of disease through transplantation

of musculoskeletal allografts. J Bone Jt Surg (Am) 1995;77-

A:1742–54.

[12] Triantafyllou N, Sotiorpoulus E, Triantafyllou J. The mechanical

properties of lyophylized and irradiated bone grafts. Acta Orthop

Bel 1975;41:35–44.

[13] K .ohler P, Kreicbergs A, Str .omberg L. Physical properties of

autoclaved bone. Acta Orthop Scand 1986;57:141–5.

[14] Taguchi Y, Pereira BP, Kour AK, Pho RWH, Lee YS.

Autoclaved autograft bone combined with vascularized bone

and bone marrow. Clin Orthop 1995;320:220–30.

[15] Borchers RE, Gibson LJ, Burchardt H, Hayes WC. Effects of

selected thermal variables on the mechanical properties of

trabecular bone. Biomaterials 1995;16:545–51.

[16] Itoman M, Nakamura S. Experimental study on allogenic bone

grafts. Int Orthop 1991;15:161–5.

[17] Beyer RE, Huang JC, Wilshire GB. The effect of endurance

exercise on bone dimensions, collagen, and calcium in the aged

male rat. Exp Gerontol 1985;20:315–23.

[18] Campo RD, Tourtellotte CD. The composition of bovine

cartilage and bone. Biochim Biophys Acta 1967;141:614–24.

[19] Wardale RJ, Duance VC. Quantification and inmunolocalisation

of porcine articular and growth plate cartilage collagens. J Cell Sci

1993;105:975–84.

[20] Woessner JF. The determination of hydroxyproline in tissue and

protein samples containing small proportions of this amino acid.

Arch Biochem Biophys 1961;93:440–7.

[21] Jonsson U, Ranta H, Stromberg L. Growth changes of collagen

cross-linking, calcium, and water content in bone. Arch Orthop

Trauma Surg 1985;104:89–93.

[22] Amtmann E. The distribution of breaking strength in the human

femur shaft. J Biomech 1968;1:271–7.

[23] Jerosch J, Muchow H, Clahsen H. Uber die Stabilit.at von

humaner Kortikalis nach verschiedenen Konservierungs- und

Sterilisationsverfahren. Z Orthop 1991;129:295–301.

[24] Cubo J, Casinos A. Mechanical properties and chemical

composition of avian long bones. Eur J Morphol 2000;38:112–21.

[25] Markel MD, Wikenheiser MA, Chao EYS. A study of fracture

callus material properties: relationship to the torsional strength of

bone. J Orthop Res 1990;8:843–50.

[26] Markel MD, Wikenheiser MA, Morin RL, Lewallen DG, Chao

EYS. Quantification of bone healing. Comparison of QCT, SPA,

MRI, and DEXA in dog osteotomies. Acta Orthop Scand

1990;61:487–98.

[27] Posner AS. The mineral of bone. Clin Orthop 1985;200:87–99.

[28] Obrant KJ, Odselius R. The concentration of Calcium and

Phosphorus in trabecular bone from the iliac crest. Calcif Tissue

Int 1986;39:8–10.

[29] Lappalainen R, Knuuttila M. Magnesium content of healthy

and chronically diseased human cancellous bone in relation to

age and some physical and chemical factors. Med Biol 1985;63:

144–8.

J. Moreno, F. Forriol / Biomaterials 23 (2002) 2615–2619 2619