Influence of Powder Particle Size Distribution on ComplexViscosity and Other Properties of Acrylic Bone Cement forVertebroplasty and Kyphoplasty

Lidia Hernandez, Marilo Gurruchaga, Isabel Goni

POLYMAT, Dep. de Ciencia y Tecnologıa de Polımeros, Facultad de Quımica UPV-EHU, Apdo. 1072,20018 San Sebastian, Spain

Received 11 January 2005; revised 20 April 2005; accepted 7 June 2005Published online 20 October 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.b.30409

Abstract: For use in vertebroplasty and kyphoplasty, an acrylic bone cement should possessmany characteristics, such as high radiopacity, low and constant viscosity during its applica-tion, low value of the maximum temperature reached during the polymerization process(Tmax), a setting time (tset) that is neither too low nor too high, and high compressive strength.The objective of this study was to investigate the influence of the powder particle distributionon various properties of one acrylic bone cement; namely, residual monomer content, Tmax,tset, complex viscosity, storage and loss moduli, injectability, and quasi-static compressivestrength and modulus. It was found that the formulations that possessed the most suitablecomplex viscosity-versus-mixing time characteristics are those in which the ratio of the largepoly(methyl methacrylate) beads (of mean diameter 118.4 �m) to the small ones (of meandiameter 69.7 �m) was at least 90% w/w. For these formulations, the values of the otherproperties determined were acceptable. © 2005 Wiley Periodicals, Inc. J Biomed Mater Res Part B: ApplBiomater 77B: 98–103, 2006

Keywords: biomaterial; acrylic bone cements; compression; viscosity; rheology

INTRODUCTION

To date, acrylic bone cements (ABCs) are the most widelyused injectable cements in vertebroplasty and kyphoplasty.1–3

In these applications, the ideal ABC must possess a widerange of properties, such as high radiopacity, a low value ofthe maximum temperature (Tmax) reached during the poly-merization process, a setting time (tset) that is neither shortnor too long, a viscosity that remains low and constant for theduration of the setting period, and high compressive strength.Of the commercially available ABCs in current clinical use,none possess all these properties. Thus, clinicians routinelymodify the cement composition so as to obtain a specificdesirable property.4 Two common approaches are usuallytaken. The first is to add large amounts of radiopacifier that isalready present in the cement,5 such as barium sulphate(BaSO4) or small amounts of another radiopacifier6,7 (such astungsten powder). The second is to increase the liquid mono-mer-to-powder ratio, in a bid to decrease the viscosity of the

polymerizing cement dough and increase the cement’s work-ing time. A key issue is the extent to which these changesaffect other important properties, such as the mechanicalproperties. In this regard, for Simplex P, Kurtz et al.8 haveshown that an increase in the amount of BaSO4 added directlyto the cement has a deleterious effect on its quasistatic tensileand fatigue properties. For Simplex P and Cranioplastic, ithas been shown that the monomer liquid-to-powder ratio hasa marked effect on compressive properties.9,10

The philosophy behind the present study is that it ispossible to obtain desirable cement properties through a ra-tional approach by determining its composition; one waybeing to adjust the powder particle size distribution, whichcan be achieved by combining poly(methyl methacrylate)(PMMA) beads of different sizes. This aspect is the subject ofthe present study, whose objective is to investigate the influ-ence of the powder particle distribution on residual monomercontent, Tmax, tset, complex viscosity, storage and loss mod-uli, injectability, and quasistatic compressive strength andmodulus of one acrylic bone cement.

MATERIALS AND METHODS

Materials

The cements were formulated by mixing two components,one liquid and the other solid, in a solid-to-liquid ratio of 2/1.

This article is dedicated to Cecilia SarasolaCorrespondence to: M. Gurruchaga (e-mail: popgutom@sc.ehu.es)Contract grant sponsor: MCYT; contract grant number: MAT2001–3874-C02–02Contract grant sponsor: Saiotek (Government of the Basque Country)Contract grant sponsor: University of the Basque Country (Predoctoral Fellowship);

contract grant number: 9/UPV 00203.215–13540/2001

© 2005 Wiley Periodicals, Inc.

98

In Tables I and II, compositions of the liquid and solid phaseof the cements and the mean particle diameter of the differentmixtures of beads used in this work are detailed. The mix-tures of PMMA powders are formulated as weight percent-ages and the mean particle diameter used is the populationmean.

Methods

The morphology of the PMMA powder was observed using aHitachi S-2700 scanning electron microscope (SEM), aftergold coating (Fine Coat Jeol Ion Sputter JFC-1100, SanSebastian, Spain). The accelerating voltage was 15 kV.

The PMMA bead particle size analysis was carried outwith a laser diffraction analyser (LS Beckman Coulter, Bar-celona, Spain) using the Fraunhofer optical model, with arunning length of 90 s. Previously, samples were dispersed inethanol and 3 min of ultrasound was applied so as to ensuregood dispersion of the particles.

The proton nuclear magnetic resonance (1H NMR) spec-troscopy was selected to accurately determine the percentageof residual monomer. Twenty milligrams of cement (curedfor 1 h at 37°C) was dissolved in deuterated chloroform,containing tetramethylsilane as internal standard, to obtainthe 1H NMR spectra from a FT-NMR Bruker spectrophotom-eter operating at 300 MHz and located at San Sebastian. Toassign the peaks and estimate the residual monomer content,we took into account the references.11 Measurements wereperformed in triplicate.

The preparation of the ABCs was carried out following themethod used for classical bone cement, as described in ISO5833 (1992) standard.12 The evolution of the polymerizationprocess was observed at a working temperature of 23°C usinga Teflon mould, described in a previous paper,11 connected toa high sensitivity thermotester. The maximum temperatureand the setting time were measured and calculated as de-scribed in ISO 5833 standards and the results are the averageof at least two measurements.

Evolution of complex viscosity (�*) as a function of timewas measured using a Rheometric Ares rheometer (San Se-bastian) in dynamic oscillation mode, at a frequency of 1 Hzusing plate–plate configuration. The radii of the plates were25 mm, and the gap between the plates was 2 mm. Therheometer was used in a constant strain mode, with a strainamplitude of 1%. Measurements were made at room temper-

ature at which cements may be mixed and injected. To carryout the experiment, the two components of the cement weremixed until good homogenization was achieved, and then, thesample was quickly placed on the lower plate. The measure-ments were started between 2 and 3 min after the mixing ofthe cement. Data were taken until the viscosity of the cementwas so high that the sample upset the even movement of theplates. Moreover, with this technique we can characterize thesample as a viscoelastic material by means of the storagemoduli (G�), loss moduli (G�), and loss tangent (tan �).

To measure the injectability of the various formulations,we followed the method proposed by other authors.13 Solidand liquid components of the cements were kept for 2 h at(23 � 1)°C before measurement was taken. A total amount of(3.0 � 0.1) cc of cement was prepared and charged in a 2-ccdisposable syringe. An 8-gauge needle (bone marrow biopsy/aspiration needle-Surecut BMB) of 150 mm length was fixedto the syringe and the cement injected to a Teflon recipient.Injectability was defined as the weight percent of cementinjected into the recipient, expressed as a percentage of thetotal amount of cement charged into the syringe and injectedafter the adequate mixing of its components. Mixing time wasmeasured as the time between the mixing of the two compo-nents of the cement and the loading of the syringe. Injectiontime was defined as the time between the beginning and theend of the injection, including the purge of the needle. Thesemeasurements were run in duplicate.

In the case that concerns us here, the main mechanical testincluded was that for compressive strength. Quasistatic com-pressive tests were carried out in this work using specimensprepared in a Teflon mould, with cylindrical holes of height12 mm and a diameter of 6 mm. The mould was filled withthe cement mixture and placed between two steel platescovered with Teflon sheets and kept in an oven at 37°C for 60min. A total of six cements’ samples were tested for eachformulation. Yield compressive strength and compressivemodulus were determined according to ISO 5833. The com-pressive testing was conducted on an Instron 4301 testingmachine at room temperature (San Sebastian), working at 22mm/min.

Statistical Analysis

One-way ANOVA statistical technique was used for the testof significance, with significance being taken as 5% level.

TABLE II. Mean Diameter of the Particle Size Distribution of theBeads Employed in This Work and Their Mixtures

Colacryl/Plexigum (w/w) Dmean (�m)

Colacryl 118.490/10 111.980/20 111.160/40 101.240/60 87.120/80 78.0Plexigum 69.7

TABLE I. Composition of the Liquid and Solid Phases of theCements Studied in This Work

Liquid Solid

N, N�, Dimethyl-4-toluidine(DMT) (1% v/v)

Benzoyl peroxide (BPO)(1.25% w/w)

Methyl methacrylate(MMA) (99% v/v)

PMMA beads mixture (98.75%w/w) of Colacryl (Bonar) andPlexigum (Rohm and Haas)a

a We prepared five mixtures of these beads, as well as pure Colacryl and Plexigumfrom 90–20% (w/w) Colacryl.

99INFLUENCE OF PMMA PARTICLE SIZE ON BONE CEMENTS

RESULTS AND DISCUSSION

The morphologies of the Colacryl and Plexigum beads areshown in Figure 1; the powder distribution diagrams arepresented in Figure 2; the variation of the mean particlediameter with % Colacryl beads in the powder is given inFigure 3; typical temperature-versus-time plots are given inFigure 4; the variation of Tmax and tset with % Colacryl beadsin the powder is given in Figure 5; and typical complexviscosity-versus-time plots are given in Figures 6 and 7.

The percentage of residual monomer detected in all thecement samples was very similar without significant statisti-cal differences between them, showing a value of approxi-mately 2.25% (% w/w), which is lower than that of most ofthe commercial bone cements.

We observe that, when the content of big PMMA beadsincreases, the tset becomes longer and the Tmax decreasesnoticeably. We can say that only the largest PMMA beads11

(Colacryl) offer an effective way to dissipate the heat pro-duced during the setting process. The cement with 100%Colacryl beads shows the slowest polymerization rate, andhence, the slowest increase in viscosity. Conversely, with anincrease in the fraction of small beads, the onset of increasein viscosity appears sooner, because of the dissolution of thesmallest PMMA beads, and subsequently, the onset of theincreasing rate of polymerization appears earlier, with Tmax

being more elevated. In the magnification (Figure 7), we canalso see that during the first stages of the process there areclear differences in the viscosity of all the cements, showing,in the case of 100% Colacryl formulation, a marked differ-ence in the viscosity increase from the very beginning of thepolymerization process. Finally, we tried to relate the com-plex viscosity with handling or setting parameters. As in thecase of other researchers,14 we were not able to relate thesetting time with a concrete value of the viscosity or vis-coelastic parameters, and so we considered the time of theonset of the increasing viscosity (tons) in a similar way toLewis and Carroll,15 but calculated as indicated in Figure 8.This means we take tons as the time at which the intersectionbetween the linear fits of the initial and final stages of thecomplex viscosity is produced. This time gives us the startingmoment of the rapid increase of the viscosity. Figure 9 showsthat a linear relationship exists between tons and the corre-sponding tset. Thus, this result suggests that tons can be con-sidered as the moment at which the dissolution of the PMMAbeads in the monomer gives up being in the main process andthe polymerization becomes the process that controls theincrease of viscosity.

Typical plots of storage (G�) and loss moduli (G�), andloss tangent (tan �) versus time from mixing are given inFigures 10 and 11. The magnitude and physical significanceof G� are related to the recoverable portion of the energyimparted by the applied strain. G� is a measure of the viscous

Figure 1. The morphologies of the PMMA beads examined by SEM.

Figure 2. Particle size distribution of the big and small PMMA beadsand their mixtures.

Figure 3. Mean diameter of the mixtures of PMMA beads as afunction of the Colacryl content.

100 HERNANDEZ, GURRUCHAGA, AND GONI

response, and reflects that the viscous flow leads to thedissipation of part of the energy imparted by the appliedstrain. Moreover, the ratio G�/G� is called tan �. This lastparameter will move from � to 0 if we pass from a Newtonianliquid to an ideal elastic material. The very beginning of themoduli plots in Figure 10 indicates the moment when asolid-like behavior becomes more noticeable than that of theliquid-like. However, if we compare carefully Figure 10 withFigure 8, we can observe that the moment signaled as tons

appears later than the moment when G� becomes higher thanG�. Thus, Figure 10 shows the behavior of the cementsprevious to their complete curing. All the dough cementsshow an increase of both moduli as the setting goes on. As wecan see, such increase of G� and G� becomes faster as thePlexigum content increases, as in the case of the complexviscosity. In addition, we observe that, except for the Co-lacryl cement, in all cases, elastic and storage modulus followvery similar plots: practically the same at high small particlecontent and parallel but close as the large particle fractionincreases. In the case of Colacryl, we must point out that theperformance of both moduli is different since G� increasesfaster than G�. Obviously, the plot of tan � shows us thesedifferences. In this case, a rapid decrease of the slope indi-cates a predominant elastic behavior.

Table III shows the injectability data, the mixing timebefore syringe charge, and the time lapse, until the formula-

tion is liquid enough to be injected. Typically the surgeonwill have between 6 and 8 min to mix and inject the cement.1

Differences in mixing time are associated with differences inthe tset of each formulation. With respect to injection inter-vals, formulations with high content of Colacryl beads (90/10and 80/20) showed the longest injection times, allowing be-tween 5 and 6 min of injection without rheological problems.In practically all cases, the injection percentage is near 80%,except for the pure Colacryl and Plexigum samples. In thecase of Plexigum, this is due to the rapid increase in viscosityand setting. However, in the case of Colacryl, this fact mustbe attributed to a more solid-like behavior of this cement, aswe have seen in the moduli plot. Under the strong shearstrength produced inside the needle, we can deduce that thisrheological behavior is intensified. The performance of tan �of all cements (Figure 11) highlights this theory, since we cansee that, in the case of 100% Colacryl formulations, tan �slope decreases markedly over time, while all the others showa more constant tan � value.

The values of the yield compressive strength (�c) andcompressive modulus (Ec) are given in Figure 12. Theseproperties were determined for the cured cement and they aredirectly related to the performance of the repaired vertebralbody. Changes in particle size do not cause a statistically

Figure 4. Exotherms of polymerization of all formulations studied.

Figure 5. Maximum temperature and setting time versus percentageof Colacryl beads content.

Figure 6. Evolution of complex viscosity after the mixing for all for-mulations.

Figure 7. Magnification of the viscosity in the initial stages, after themixing.

101INFLUENCE OF PMMA PARTICLE SIZE ON BONE CEMENTS

significant effect on either of these parameters. For all for-mulations, �c exceeds the minimum level required, as statedin ISO 5833; which is 70 MPa, while for Ec the values arewithin the range given in the literature.16

To study the influence of the size of PMMA beads withoutthe interference of other components, the addition of a ra-diopaque agent was avoided. Although it is unquestionablewhether the radiopaque agent has an effect on the cement’sbehavior, we avoided its use in order to enhance the influenceof the particle size on the cement’s properties. A subsequent

Figure 8. Graphic calculation of the time of the onset of increasingviscosity for Colacryl and Plexigum cements.

Figure 9. Onset time versus setting times of all ABC formulations.

Figure 10. Rheological values of loss and storage modulus versustime from mixing.

Figure 11. Rheological evolution of tan � as a function of time.

Figure 12. Compressive properties of the cements studied: (a) yieldcompression strength, (b) compressive modulus.

TABLE III. The Mixing Time, Injection Time, and Injectability ofthe Cement Formulations

Colacryl/Plexigum(w/w)

MixingTime (min)

InjectionTime (min)

Injectability(%)

Colacryl 4.0 (0.3)a 3.28 (0.38) 40.99 (2.35)90/10 3.5 (0.2) 6.13 (0.56) 82.55 (0.15)80/20 3.5 (0.3) 5.43 (0.12) 83.17 (0.11)60/40 2.0 (0.1) 3.84 (0.22) 78.24 (0.38)40/60 1.0 (0.1) 1.71 (0.09) 77.94 (0.25)20/80 1.0 (0.2) 1.29 (0.34) 82.08 (2.49)Plexigum 1.0 (0.0) 1.13 (0.08) 32.24 (8.40)

a Values in parentheses indicate SD values.

102 HERNANDEZ, GURRUCHAGA, AND GONI

study will determine which is the most suitable radiopacifieror even the selection of radiopaque monomers.17,18

In conclusion, we have shown that

● it is feasible to obtain injectable bone cements with optimalrheological properties by the simple means of using beadsof selected size distribution.

● the storage and loss moduli could be a good way to predictthe injection suitability of this kind of ABC.

● cements in which the proportion of large PMMA beads(mean diameter of 118.4 �m) to small beads (mean diam-eter of 69.7 �m) is at least 90% w/w present high values ofthe compressive parameters and possess injectability andviscosity characteristics that make them suitable candidatesfor use in vertebroplasty and kyphoplasty.

We thank “Unitat de Citometria dels Serveis cientificotecnics”from the University of Barcelona for the particle size analysis.

REFERENCES

1. Hide IG, Gangi A. Percutaneous vertebroplasty: History, tech-nique and current perspectives. Clin Radiol 2004;59:461–467.

2. Heini PF, Berlemann U. Bone substitutes in vertebroplasty. EurSpine J 2001;10:205–213.

3. Truumees E, Hilibrand A, Vaccaro AR. Percutaneous vertebralaugmentation. Spine J 2004;4:218–229.

4. Jasper LE, Deramond H, Mathis JM, Belkoff SM. Materialproperties of various cements for use with vertebroplasty. J MatSci Mat Med 2002;13:1–5.

5. Moreland DB, Landi MK, Grand W. Vertebroplasty: Tech-niques to avoid complications. Spine J 2001;1:66–71.

6. Martin JB, Jean B, Sugiu K, San Millan Ruiz D, Piotin M,Murphy K, Rufenacht B, Muster M, Rufenacht A. Vertebro-plasty. Clinical experience and follow-up results. Bone 1999;25(Suppl. 1):11S–15S.

7. Theodorou DJ, Theodorou SJ, Duncan TD, Garfin SR, WongWH. Percutaneous balloon kyphoplasty for the correction of

spinal deformity in painful vertebral body compression frac-tures. Clin Imaging 2002;26:1–5.

8. Kurtz SM, Villarraga ML, Zhao K, Edidin AA. Static andfatigue mechanical behaviour of bone cement with elevatedbarium sulphate content for treatment of vertebral compressionfractures. Biomaterials 2005;26:3699–3712.

9. Belkoff SM, Sanders JC, Jasper LE. The effect of the monomer-to-powder ratio on the material properties of acrylic bone ce-ment. J Biomed Mater Res Part B: Appl Biomater 2002;63:396–399.

10. Jasper LE, Deramond H, Mathis JM, Belkoff SM. The effect ofmonomer-to-powder ratio on the material properties of cranio-plastic. Bone 1999;25 (Suppl. 1):27S–29S.

11. Pascual B, Vazquez B, Gurruchaga M, Goni I, Ginebra P, Gil J,Planell JA, Levenfeld B, San Roman J. New aspects of theeffect of size and size distribution on the setting parameters andmechanical properties of acrylic bone cements. Biomaterials1996;17:509–516.

12. ISO 5833/2. Implants for surgery–Acrylic resin cements. Or-thopaedic application. 1992.

13. Mendez JA, Fernandez M, Gonzalez-Corchon A, Salvado M,Collıa F, de Pedro JA, Levenfeld BL, Lopez-Bravo A, VazquezB, San Roman J. Injectable self-curing bioactive acrylic-glasscomposites charged with specific anti-inflammatory/analgesicagent. Biomaterials 2004;25:2381–2392.

14. Farrar DF, Rose J. Rheological properties of bone cementsduring curing. Biomaterials 2001;22:3005–3013.

15. Lewis G, Carroll M. Rheological properties of acrylic bonecement during curing and the role of the size of the powderparticles. J Biomed Mater Res Part B: Appl Biomater 2002;63:191–199.

16. Lewis G. Properties of acrylic bone cement: State of the artreview. J Biomed Mater Res Part B: Appl Biomater 1997;38:155–182.

17. Artola A, Gurruchaga M, Vazquez B, San Roman J, Goni I.Elimination of barium sulphate from acrylic bone cements. Useof radiopaque monomers. Biomaterials 2003;24:4071–4080.

18. Artola A, Goni I, Gil J, Ginebra P, Manero JM, Gurruchaga M.A radiopaque polymeric matrix for acrylic bone cements.J Biomed Mat Res Part B: Appl Biomater 2003;64:44–55.

103INFLUENCE OF PMMA PARTICLE SIZE ON BONE CEMENTS