563

REVIEW

Effects of mechanical stimulation on the differentiation of hard tissues

Stuart Goodman* and Per Aspenberg+ *Division of Orthopaedic Surgery, Stanford University Medical Center, Stanford, CA 94305-5326, USA; and ‘Department of Orthopaedic Surgery, University Hospital, Lund, Sweden

In 1892, J.L. Wolff believed that bone was a dynamic organ that responded to the biomechanical environment. Research has shown that mechanical stimulation can have a profound effect on the differentiation and development of mesenchymal tissues. It would appear that a ‘window’ of mechanical strain exists which may facilitate or discourage the accretion of bone. With respect to processes such as fracture healing and ingrowth of bone into porous coated prostheses, it may be possible to modulate the mechanical environment with the application of well-defined, exogenous loads in order to promote a more favourable outcome.

Keywords: Bone growth, mechanical stimulation, fracture healing, tissue differentiation

Received 16 October 1992; revised 25 November 1992; accepted 27 January 1993

Recent theoretical and experimental studies have provided convincing evidence that the mechanical environment plays an important role in the differentiation and development of mesenchymal tissue. While per- forming studies on the effects of mechanical strain on tissue ingrowth into the micromotion chamber in rabbits, the authors reviewed current studies on the effects of the mechanical environment on differentiation of mesenchymal tissue. In this review paper, the authors re-examine this literature emphasizing two areas of orthopaedic research, namely the healing of fractures and bone ingrowth into porous coated prostheses. By understanding the interaction between the mechanical environment and the differentiation of mesenchymal tissue, it may be possible to promote a more favourable outcome for various musculoskeletal conditions by the application of well-defined, exogenous loads.

GENERAL STUDIES ON LOAD AND BONE REMODELLING

Load in intact bones

The deformation of bone with the application of load has been quantitated in vivo by Lanyon et al. ‘* ‘. Preliminary studies were performed in sheep, attaching strain gauges to the medial surface of the tibia1 shaft’. Recordings of the bone strain were performed with the animal walking normally. The strain gauge tracings showed a cyclic pattern of strain that was generally similar between sheep, with some individual variation according to the particular style of gait. The authors then attached a strain gauge to the anteromedial aspect of the tibia1 shaft of one

Correspondence to Dr S. Goodman.

of the members of their research team’. It was shown in this subject that, with running, the principal tension reached 850 microstrain at 13 X lo3 microstrain/s along the axis of the tibia1 shaft. Prior to toe-off during normal gait, -400 microstrain of compression qt -4 X lo3

microstrain/s was documented at 37’ to the longitudinal axis of the tibia. Furthermore, Rubin and Lanyon3 demonstrated that, despite vast differences in the size and style of locomotion of a wide range of animal species, the peak strains in load-bearing bone are remarkably similar, possibly indicating similar strain thresholds for bone remodelling.

The mechanical environment and bone remodelling

Wolff postulated that the mechanical environment maintains bone mass and determines bone remodelling4. Rubin and Lanyon tested this hypothesis using a unique animal mode15-7. The diaphysis of the ulna of skeletally mature roosters was functionally isolated by surgical means, by excising the articular segments of the ulna and applying an external fixation device. When load bearing was removed for 6 wk, the isolated ulnar segment demonstrated cortical thinning secondary to endosteal resorption, intracortical porosis and reduction of bone mass. When the isolated ulnar segment was subjected to physiological loading of 4 cycles/d of 0.010-0.012 strain at 0.5 Hz for 6 wk, the disuse osteopenia was prevented. If the loading was increased to 36 cycles/d, endosteal and periosteal accretion of bone was seen. Increasing the frequency further had no incremental effect. Thus alterations in the mechanical environment were associated with dramatic changes in bone mass.

In similar studies using the ulna of mature male

0 1993 Buttetworth-Heinemann Ltd Biomaterials 1993, Vol. 14 No. 8 0142-9612/93/060563-07

564 Differentiation of hard tissues: S. Goodman and P. Aspenberg

turkeys, these authors developed a dose-response curve for peak strain magnitude and alteration in bone mass’. Using 100 daily loading cycles of 1 Hz for 8 wk and strains in the physiological range, it was demonstrated that peak longitudinal strains below 0.001 produced loss of bone mass, whereas those above 0.001 produced marked periosteal and endosteal bone formation. Skerry et al.‘, using the same model and a loading pattern of 360 sinusoidal cycles from 0 to 600 N [producing a longitudinal strain of 0.002) at a frequency of 1 Hz showed that a short period of dynamic loading could alter the orientation of the proteoglycan molecules in bone. This phenomenon was observed immediately after the application of the load and reverted to normal 46 h after cessation of the load. It was proposed that this phenomenon provided a mechanism whereby the bone’s recent dynamic strain history could be recorded, enabling the bone to adapt to changes in the functional biomechanical environment.

What are the cellular mechanisms by which load affects tissue remodelling? Recent in vitro studies have demonstrated the ability of the osteoblast to respond to strain by alterations in the production of PGE,, CAMP, cGMP, DNA, and other substances’O~“, Cyclic but not static loading of osteoblasts in tissue culture stimulated cellular proliferation but suppressed the synthesis of proteoglycans, collagen and non-collagenous protein”. Other studies have suggested that the extracellular matrix may provide a 3D framework that conveys or transforms the mechanical stimulus for the osteoblast to interpretZ3. Recent studies have also underlined the importance of electrical fields in the maintenance, remodelling and repair of bones14, 15.

Finite element studies and bone remodelling

Theoretical investigations using finite element modelling have demonstrated that the mechanical loading history can direct the development and differentiation of the tissues of the musculoskeletal system’6-20. Studies by Carter and co-workers have shown that the process of endochondral ossification and the eventual architectural configuration of bones are determined by the strain history of the tissue. This process apparently begins at a very early stage in skeletal development. Using mathe- matical ‘rules of construction’, the cartilage anlage of a particular bone matures into a structure according to the pervading biomechanical environment. This hypothesis provides a mechanism for the development, differentiation and repair of tissues of the musculoskeletal system of different species and the adaptation of the organism to alterations in the biomechanical milieu.

THE EFFECTS OF LOAD ON FRACTURE HEALING

The healing of fractures of long bones follows an orderly sequence of events that has been documented exten- sivelyzl* “. Unstable fractures heal by callus formation through enchondral ossification [spontaneous, indirect or secondary union). If the fragments are anatomically reduced and rigidly stabilized, little external callus forms and the bones unite by a process of direct or primary union. In this process, healing of bone occurs directly by internal remodelling in areas of intimate contact: in gaps,

woven bone is laid down perpendicular to the long axis of the bone and later remodelled to more mature, longi- tudinally oriented lamellar bone,

Several research groups have tried to influence the progression of fracture healing by the application of mechanical load. Sarmiento et aLz3 have provided convincing experimental evidence that early weight bearing can accelerate the process of fracture healing. Rats subjected to a standardized femoral fracture and non-rigid intramedullary fixation were allowed to bear weight at an early stage, or were prevented from weight bearing by cast immobilization. Loading of the limb at an early stage was associated with earlier, more pronounced fracture healing radiographically, histologically and mechanically. The authors concluded that early functional weight bearing facilitates the maturation of the callus produced by enchondral ossification. These principles have been extended to human studies with encouraging resultsz4.

Different methods of fracture fixation provide different degrees of stability of the fracture fragments. Chao’s group studied the effect of fracture stability on fracture healing. They employed a standardized, bilateral tibia1 osteotomy in mature dogs. In one experiment, a more stable construct [compression plating of the tibia1 osteotomy) was compared with a less stable construct [external fixation using half pins) performed on the contralateral sidez5. After 120 d, the bone from the side with the external fixator was weaker to torsional loading and demonstrated less intracortical new bone formation, more porosity and resorption of bone compared with the tibia that had undergone more rigid fixation with a plate. It was concluded that the more rigid method of fixation using a plate and screws facilitated earlier healing and remodelling of the fracture, using this model. Early motion of the fracture fragments proved to prolong fracture healing.

The effect of rigidity of the external fixation system on the healing of the canine tibia1 osteotomy was also explored. The authors placed a more rigid external fixator employing six half pins on one side; the contra- lateral osteotomy was stabilized with a more flexible system using four half pins?‘. Although periosteal callus formation was greater on the side with the less rigid system at 90 and 120 d, earlier radiographic union was seen with the more rigid construct. However, torsional strength was similar using both methods at 120 d. A more rigid, two-plane external fixation system was then compared with a less rigid, one-plane construct”. In the early stages of fracture healing, the more flexible system was associated with more displacement of the osteotomy and yielded more callus formation and total porosity of bone. However, by 13 wk, the biomechanical properties of the osteotomy site were indistinguishable in the two groups. The repair and remodelling of fractures were found to be longer when an external fixation device with less rigidity was used. As in previous experiments, this study emphasized that stability of the fracture fragments per primum promotes earlier fracture healing.

Does compression aid in the healing of tibia1 osteotomies? Although the rigidity of the construct was increased by the application of 80 N of compression, there were no significant differences in the pattern or strength of healing of tibia1 osteotomies after 60 d of

Biomaterials 1993, Vol. 14 No. 8

Differentiation of hard tissues: S. Goodman and P. Aspenberg 565

external fixation”. Furthermore, when a telescoping external fixation system producing dynamic compression during gait was compared with a static system in the neutralization mode, there were no significant differences with respect to the technetium-MDP uptake, formation of new bone, bone porosity or maximum torque of canine tibia1 osteotomies”. These results in dogs suggest that compression, whether dynamic or static, does not alter the progression of healing of tibia1 osteotomies.

Wolf et al. 3o hypothesized that the intermittent cyclic strain produced by an externally applied, well controlled cyclic load was a more important stimulus to fracture healing than simple static compression. Mature female rabbits were subjected to bilateral tibia1 osteotomies fixed with external fixation. On one side, 80 N of constant compression was applied across the osteotomy site; on the contralateral side the osteotomy was subjected to controlled cyclic loading (40 N at a frequency of 55 times/min, 3 h every morning and evening). At 3-4 wk, the side with constant compression but without cyclic loading exhibited slightly improved healing biomechanically; however, by 4-6 wk, the tibia under- going cyclic loading demonstrated higher torque and energy absorption to failure and lower stiffness compared with the contralateral tibia in which constant compression was applied. At 8 wk, no significant differences were noted. In the early stages of fracture healing, it would appear that stability of the fracture fragments hastens the repair process. Thereafter, controlled cyclic loading may provide a mechanical stimulus to the osteoblast and extracellular matrix which results in accelerated remodelling of the callus.

The results reported by Wolf et al. do not appear to be species specific. Goodship and Kenwright31’32 came to the same conclusion, that the healing of fractures was responsive to an imposed, controlled, mechanical environment using sheep. A unilateral tibia1 osteotomy was performed in mature sheep and stabilized with an external fixation device. In one group of animals, the osteotomy site was distracted 3 mm; in the second group, the tibia was subjected to a controlled mechanical stimulus (500 cycles of 360 N delivered axially at 0.5 Hz for 17 min daily) for 12 wk without initial distraction. The tibiae undergoing controlled axial loading sub- sequently demonstrated more external callus, increased fracture stiffness after 8 wk, higher torsional stiffness at post-mortem and more advanced fracture healing histo- logically. A larger initial gap between the bone ends (0.5 versus 2.0 mm) proved to be detrimental to the healing of the osteotomy, even with the application of micro- movement3’. Furthermore, the healing of the osteotomy was sensitive to the level of the applied force: healing was biomechanically and histologically more sound in osteotomies subjected to 200 N compared with 1000 N of intermittent micromotion delivered over 17 min at 0.5 Hz for 12 wk.

Controlled cyclic load and fracture healing in humans

Goodship and Kenwright have utilized the principle of an imposed mechanical load in a controlled randomized trial of the treatment of tibia1 fractures stabilized with external fixation3’, 33. A specially designed pneumatic

pump was attached to the unilateral frame of one group of patients and delivered a cyclical axial displacement of 1.0 mm at 0.6 Hz for 20-30 min/d. The group subjected to micromovement healed their fracture in a shorter time clinically, radiographically and biomechanically compared with the group treated with external fixation without micromotion.

Summary of the effects of load on fracture healing

It would appear that, after a period of protected loading, fractures stabilized with external fixation demonstrate an increased rate of healing with the application of controlled, cyclic loading. Variables such as the size of the initial gap, the stability of the bone-fixator construct, and the parameters of the applied motion (magnitude of the load, frequency, duration, etc.) appear to be important. Whether these same principles are applicable to techniques providing more rigid fixation (e.g. with internal fixation] has not been elucidated.

MICROMOTION AND BONE INGROWTH INTO POROUS COATED PROSTHESES

Bone cement has been used extensively to stabilize joint arthroplasties within bone. Recently, because of possible adverse effects of bone cement, press fit, porous coated implants have been used as a method of prosthetic stabilization349 35. These implants depend on bone ingrowth into small pores on the surface of the prosthesis. Two different systems of porous coating have evolved, the microporous system (pore size approximately lOO-500pm) and the macroporous or madreporic system (pore size of l-2 mm). In the microporous system, the pore size appears to be crucial to the degree of bone ingrowth36.

Several research groups have measured the amount of micromotion at the bone implant interface in vivo and in vitro. Yang et a1.37, using fresh frozen canine tibiae implanted with a tibia1 tray, showed that the amount of tangential displacement ranged from approximately 40 pm during application of 100% body weight to 100 ,am with 400% body weight. Volz et al. 36 tested four different uncemented tibia1 trays implanted in paired cadaveric tibiae. After 300 000 cycles of loading from 5 to 115 kg, motion ranging from less than 100,um to greater than 500,~m was found at the implant-bone interface, for different component designs. Vanderby et al. 3g compared the micromotion of a cemented hip replacement in canines with the motion after 0 and 4 months of bone ingrowth into a porous coated titanium alloy prosthesis. Torsional (craniocaudal) loads yielded the greatest amount of motion, which ranged from less than 50 ,um for cemented components, to 300 ,am (4 months of ingrowth) to 500 pm (zero ingrowth) for porous coated prostheses. Zalenski et aL4’, using a titanium fibre mesh porous coated hip prosthesis implanted into the canine femur, confirmed that torsional loads caused the greatest displacement. These authors measured 7-56pm of rotational micromotion immediately after prosthetic implantation but only 0.6-25pm of motion after 6 months. Anderson et aL41q 42, using a madreporic, cobalt chrome canine hip prosthesis and axial loads of up to 300 N, documented motion in the order of 65 pm or less at time zero, less than 20 ,am after 6 months, and less than

Biomaterials 1993. Vol. 14 No. 8

566 Differentiation of hard tissues: S. Goodman and P. Aspenberg

27 ,um after 2 yr. The micromotion using comparable Using human cadaveric specimens, Sugiyama et al.54 cemented prostheses ranged from approximately 7 to confirmed that under-reaming of the femur distally by 18 pm. 1/2-l mm significantly improved torsional stability.

The amount of bone ingrowth into porous coated prostheses has been disappointing4” 44. One possible reason for the paucity of bone ingrowth is the initial biomechanical environment at the bone-implant interface. The instability produced at the interface secondary to muscle forces and intermittent loading results in motion between the implant and bone45. Several researchers have shown that some degree of motion is compatible with bone ingrowth. Pilliar et a1.46 have shown in dogs that up to 28 ,um of micromovement can still permit bone ingrowth to occur. Movement at the interface measuring l50pm or greater promoted the formation of fibrous tissue. Similar results were reported by Burke et al.47. Hollis et a1.4* implanted multiple porous coated titanium plugs transcortically into mature dogs. Using a special device, adjacent implants were rotated 25, 50, 100 or 200 pm twice per day for 10 min. With 25 ,um of motion, bone grew consistently into the pores: little or no bone ingrowth was seen with 200 pm of rotational motion. With motion in the range 50-100 pm, the amount of bone ingrowth was related to the magnitude of motion and the size of the pore. Thus, with 50 ,um of movement, some of the smaller pores and most of the larger pores had ingrowth of bone. With 100 pm of micromotion, few of the smaller pores but some of the larger pores demon- strated bone formation. This study also showed that only a short period of daily motion is necessary to inhibit bone ingrowth.

In order to improve the fixation of cementless components in bone, hydroxyapatite (HA) has been used to coat the surface of metallic implants. Soballe et al.55 have shown that coating a porous titanium implant placed in the femoral condyles of mature dogs improves bone ingrowth from 8% (uncoated) to 47% [HA coated) for stable implants after 4 wk. When the implants were unstable (a dynamically loaded device produced 500 ,um of axial translation during each gait cycle), coating the implant with HA improved pushout strength but not the amount of bone ingrowth. When the degree of micromotion was reduced to 150 pm, the shear strength of HA coated implants was still improved compared with uncoated implants56. A thick fibrous membrane was noted around unstable titanium implants but a thinner fibrous/fibro- cartilaginous membrane with a radiating rather than a random collagen pattern surrounded HA coated implants. When micromotion was discontinued between 4 and 16 wk, the fibrous and fibrocartilaginous membranes around both types of implants were replaced by bone57, indicating the capability of tissue differentiation when presented with the appropriate stimulus.

The placement of a screw within bone is somewhat analogous to the process of bone ingrowth into a porous coated prosthesis in that the bone in the threads of the screw must undergo a process of repair and remodelling in order to maintain adequate stability of the construct. Schatzker et a1.4g placed tightly fitting and overdrilled screws into loaded and minimally loaded long bones in dogs. Screws located in a stable, minimally loaded environment, whether overdrilled or not, were surrounded by new bone at 6 wk. However, overdrilled screws subjected to intermittent loading during gait were surrounded by a radiological lucent line containing a synovial-like lining, fibrous tissue and osteoclastic resorption of the surrounding bone. This study points to the importance of the biomechanical environment in stabilizing an implant with a poor initial fit within bone.

Controlled cyclic load and bone ingrowth

Can mechanical strain enhance bone ingrowth into porous coated prostheses? Rubin and McLeod” implanted 5 mm diameter titanium alloy rods with porous coating into the functionally isolated ulna of skeletally mature male turkeys for 8 wk. A servo-hydraulic actuator provided controlled dynamic loading of the ulna at 1 or 20 Hz for 1000 cycles/d. Ingrowth of bone was enhanced by both levels of applied load, yielding means of 21 and 74% bone ingrowth with the lower and higher frequencies, respectively, compared with 13% ingrowth without any load. Thus, controlled, frequency specific, low amplitude mechanical load may stimulate the bone surrounding a porous coated implant, thereby increasing bone ingrowth.

The micromotion chamber

Prosthetic design and surgical technique have been shown to affect the degree of prosthetic micromotion and bone ingrowth. Evans et aL5’ demonstrated a linear, inverse relationship between canal fill and femoral component motion using a straight stemmed, collarless titanium alloy prosthesis in dogs. The distal, non-porous coated part of the femoral stem was shown by Krushell et al. 51 and Zalenski et al. 52 to enhance initial stability to axial and rotational loads by 100 and 170%, respectively. However, the smooth portion of the stem was found to be superfluous at 6 and 24 months. after implantation. Increasing the femoral component offset (which was shown to be an advantage in cemented femoral com- ponents) was shown to be a possible disadvantage in cementless femoral hip components because of the resulting increase in micromotion during stairclimbing53.

The micromotion chamber [MC), a modification of the bone harvest chamber developed by Albrektsson and co- workers5” “, is a commercially pure titanium device implanted in the rabbit tibia to measure the effects of manually imposed motion on tissue ingrowth into a single 1 mm pore. Using the MC, Aspenberg et al. showed that short daily periods of motion (20 cycles/d, 0.5 mm in amplitude, delivered over a 30 s period for 3 wk) inhibited the ingrowth of bone into the pore6*. A greater amount of fibrous tissue rather than bone was formed in the chamber when micromotion was instituted. Thus, as in studies by others5* ‘* 30-33, 48V 58, short daily periods of mechanical strain can have a dramatic effect on tissue differentiation. The effects of pore cross- sectional shape, amplitude and frequency of micromotion on bone ingrowth were also examined. In one set of animals, the outer cylinder was pierced by a square 1 mm hole; a second set of animals had a chamber with a round- holed cylinder ” In all cases the channel in the inner core . was a 1 X 1 X 5 mm quadrate. All chambers underwent 20 cycles/d of micromotion for 3 wk at 0.5 mm amplitude.

Biomaterials 1993, Vol. 14 NO. 8

Differentiation of hard tissues: S. Goodman and P. Aspenberg 567

The more congruent [square-square) interface with a greater cross-sectional area was found to have more tissue differentiation into bone compared with the more incongruent (round-square) interface. Indeed, the more congruent interface produced as much bone ingrowth as non-moved chambers! Increasing the amount of micro- motion from 0.5 to 0.75 mm at 20 cycles/d produced more tissue differentiation to fibrous tissue rather than bone, despite a congruous (square-square) interface63. Further- more, if the amplitude was kept constant at 0.5 mm and the frequency was increased from 20 cycles once per day to 20 cycles delivered twice per day, more tissue differentiation to fibrous tissue resulted (unpublished data). These studies have demonstrated that tissue differentiation into the micromotion chamber is very sensitive to the parameters of motion chosen for the experiment.

Finite element analysis and bone ingrowth and remodelling

Computer simulation of the prosthesis-bone construct has been carried out by Huiskes’ group using finite element analysis 64 Using the theories of adaptive bone . remodelling, the strain energy density has been noted to be a feedback mechanism determining the configuration and architectural appearance of bone with an implant. The bone surrounding a prosthesis remodels according to general principles similar to those outlined by other

authors5ss. The rigidity of the implant and the bonding characteristics of the implant to bone are important determinants of the remodelling process.

Summary of the effects of load on bone ingrowth

As in fracture healing, porous coated prostheses must be implanted so as to obtain excellent initial stability and close apposition between the porous coating and the adjacent bone. Thus, the anatomy, the design of the prosthesis and surgical technique are crucial variables. Motion above approximately 25-50 pm will inhibit bone ingrowth. However, the application of controlled, exogenous, specific loads may stimulate the bone surrounding a porous coated implant.

DISCUSSION AND CONCLUSIONS

In 1892, J.L. Wolff published a monograph entitled Das Gesetz der Transformation der Knochen [The Law of Bone Remodelling)4. Wolff believed that bone was a dynamic organ responding to changes in the biomechanical environment. Current theoretical and experimental studies have shown that the differentiation and develop- ment of mesenchymal tissue is determined in part, either directly or indirectly, by the loading parameters to which it is subjected. The genetic control of these processes seems in some respects to be more indirect than traditionally thought. Rather than determining an exact musculoskeletal phenotype, it would appear that the genome provides more general rules and standards by which the developing mesenchymal tissue should react to various types of mechanical stimuli.

The possibility of modulating processes such as

fracture healing and the degree of bone ingrowth into prosthetic implants by altering the biomechanical environment is intriguing. It is clear from this review, however, that exogenous loads cannot be applied indiscriminately. With respect to bone, it would appear that a ‘window’ of mechanical strain exists which may facilitate or discourage the accretion of bone’, 31V 32V 63.

With further research, it is likely that orthopaedic surgeons and rehabilitation specialists will modify their treatment protocols in order to optimize the biomechanical variables affecting fracture healing and other musculo- skeletal conditions. Manufacturers will undoubtedly redesign orthopaedic implants and appliances to take advantage of the mechanical milieu in which they are functioning. In this regard, theoretical studies and computer modelling are valuable methods in the pre- clinical assessment of different loading parameters.

ACKNOWLEDGEMENTS

This work was funded in part by Swedish Medical Council Project numbers 09509 and 2031, and the Medical Faculty of the University of Lund. The authors are grateful to Lela Blankenberg for typing the manuscript.

REFERENCES

1

2

3

4

5

6

7

6

9

10

11

Lanyon, L.E. and Smith, R.N., Bone strain in the tibia during normal quadrupedal locomotion, Acta Qrthop. Stand. 1970, 41, 238-248 Lanyon, L.E., Hampson, W.G.J., Goodship, A.E. and Shah, JS., Bone deformation recordedin vivo from strain gauges attached to the human tibia1 shaft, Acta Ortbop. Stand. 1975, 46, 256-268 Rubin, C.T. and Lanyon, L.E., Dynamic strain similarity in vertebrates: an alternative to allometric limb bone scaling, I. Theor. Biol. 1984, 107, 321-327 Wolff, J.L., TheLaw ofBone Remodelling (Das Gesetz der Transformation der Knochen, 1982), translated by P. Maquet and R. Furlong, Springer, Berlin, Germany, 1966 Rubin, CT. and Lanyon, L.E., Regulation of bone formation by applied dynamic loads, 1. Bone ]oint Surg. 1984, 66-A, 397-402 Rubin, CT., McLeod, K.J. and Bain, S.D., Functional strains and cortical bone adaptation: epigenetic assurance of skeletal integrity, I. Biomech. 1990, 23, Suppl. 1, 43-54 Lanyon, L.E., Functional strain in bone tissue as an objective, and controlling stimulus for adaptive bone remodelling, I. Biomech. 1987, 20, 1083-1093 Rubin, C.T. and Lanyon, L.E., Regulation of bone mass by mechanical stimulation, Calcif. Tiss. Int. 1985, 37, 411-417 Skerry, T.M., Bitensky, L., Chayen, J. and Lanyon, L.E., Loading-related reorientation of bone proteoglycan in vivo. Strain memory in bone tissue?]. Orthop. Res. 1988, 6, 547-551 Somjen, D., Binderman, I., Berger, E. and Hare& A., Bone remodelling induced by physical stress is prostaglandin E2 mediated, Bioclin. Biophys. Acta 1980, 627, 91-100 Burger, E., Klein-Nuland, J., Jong, M. de and Zoelen, J, van, Enhancement of DNA synthesis in calvaria by mechanical stimulation may be mediated by TGFB, J. Bone Miner. Res. 1988, S176, 1988

Biomaterials 1993, Vol. 14 No. 6

566 Differentiation of hard tissues: S. Goodman and f. Aspenberg

12

13

14

15

16

17

16

19

20

21

22

23

24

25

26

27

28

29

30

31

Strafford, B.B., Brighton, C.T., Williams, J.L. and Pollack, S.R., The in vitro response of bone cells to cyclic biaxial mechanical strain, ‘Zkansactions of the 35th Annual Meeting of the Orthopaedic Research Society Las Vegas, NV, USA, 1969, p 35 Rodan, G.A., Baurret, L.A., Harvey, A. and Mensi, T., Cyclic AMP and cyclic GMP: mediators of the mechanical effects on bone remodeling, Science 1975, 189, 467-469 McLeod, K.J. and Rubin, C.T., The effect of low frequency electric fields on osteogenesis, I. Bone Joint slug. 1992,74-A, 920-929 Lavine, L.S. and Grodzinsky, A.J., Current concepts review. Electrical stimulation of the repair of bone, 1. Bone Joint Surg. 1967, 69-A, 626-630 Carter, D.R., Mechanical loading history and skeletal biology, J. Biomech. 1987, 20,‘ 1095-1109 Carter, D.R., Orr, T.E. and Fyhrie, D.P., Relationships between loading history and femoral cancellous bone architecture, J. Biomech. 1969, 22, 231-244 Carter, D.R., Blenman, P.R. and Beaupti, G.S., Correlation between mechanical stress history and tissue differen- tiation in initial fracture healing, J. O&hop. Res. 1988, 6, 738-748 Wong, M. and Carter, -D.R., A theoretical model of endochondral ossification and bone architectural reconstruction in long bone otogeny, Anat. Embryol. 1990,161,523-532 Carter, D.R., Wong, M. and Orr, T.E., Musculoskeletal ontogeny, phylogeny, and functional adaptation, J. Biomech. 1991, 24, Suppl. 1, 3-16 Perren, S.M., Bone healing, in Manual of Internal Fixation (Eds M.E. Muller, M. Allgower, R. Schneider and H. Willenegger), Springer, Berlin, Germany, 1990, pp 86-69 Cruess, R.L., Healing of bone, tendon and ligament, in Fractures in Adults (Eds C.A. Rockwood and D.P. Green), J.B. Lippincott, Philadelphia, PA, USA, 1984, pp 147-168 Sarmiento, A., Schaeffer, J.F., Beckerman, L., Latta, L. and Enis, J.E., Fracture healing in rat femora as affected by functional weight-bearing, J. Bone Joint Surg. 1977, 59.A,369-375 Sarmiento, A., Functional bracing of tibia1 and femoral fractures, Clin. Orthop. 1972, 62, 2-?3 Lewallan, D.G., Chao, E.Y.S., Kasman, R.A. and Kelly, P.J., Comparison of the effects of compression plated and external fixators on early bone-healing, J. Bone Joint Surg. 1964, 66-A, 1084-1091 Wu, Jiunn-Jer, Shyr, H.S., Chao, E.Y.S. and Kelly, P.J., Comparison of osteotomy healing under external fixation devices with different stiffness characteristics, J. Bone Joint Surg. 1984, 66-A, 1258-1264 Williams, E.A., Rand, J.A., Chao, E.Y.S. and Kelly, P-J., The early healing of tibia1 osteotomies stabilized by one- plane or two-plane external fixation, J. Bone Joint Surg. 1967, 69-A, 355-365 Hart, M.B., Wu, Jiunn-Jer, Shyr, H.S., Chao, E.Y.S. and Kelly, P.J., External skeletal fixation of canine tibia1 osteotomies - compression compared with no com- pression, J. Bone Joint Surg. 1985, 67-A, 598-605 Aro, H.T., Kelly, P.J., Lewallan, D.G. and Chao, E.Y.S., The effects of physiologic dynamic compression on bone healing under external fixation, Clin. Orthop. 2990, 256, 260-273 Wolf, J.W., White, A.A., Panjabi, M.M. and Southwick, W.O., Comparison of cyclic loading versus constant compression in the treatment of long-bone fractures in rabbits, J. Bone Joint Surg. 1981, 63-A, 805-610 Goodship, A.E; and Kenwright, J., The influence of induced micmmovement upon the healing of experimental

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

tibia1 fractures, J. Bone Joint Surg. 1985, 65-B, 650-655 Kenwright, J. and Goodship, A.E., Controlled mechanical stimulation in the treatment of tibia1 fractures, Ch’n. Orthop. 1990, 241, 36-47 Kenwright, J., Richardson, J.B., Cunningham, J.L., White, S.H., Goodship, A.E., Adams, M.A., Magnussen, P.A. and Newman, J.H., Axial movement and tibia1 fractures. A controlled randomized trial of treatment, J. Bone Joint Surg. 1991, 73-B, 654-659 Galante, J., Restocker, W., Lueck, R. and Ray, R.D., Sintered fiber metal composites as a basis for attachment of implants to bone, J. Bone Joint Surg. 1971, 53-A, 101-114 Spector, M., Historical review of porous-coated implants, J. ArthropZasty New Orleans, LA, USA, 1987,2,163-177 Bobyn, J.D., Pilliar, R.M., Cameron, H.V. and Weatherly, G.C., The optimum pore size for the fixation of porous- surfaced metal implants by the ingrowth of bone, Clin. Orthop. 1980, 150, 263-270 Yang, A., Sumner, D.R., Choi, S., Natarajan, R. and Andriacchi, T.P., Direct measurement of micmmotion at the bone-implant interface: the tibia1 component in a canine model, Transactions of the 36th Annual Meeting of the Orthopaedic Research Society New Orleans, LA, USA, 1990, p 233 Volz, R.G., Nisbet, J.K., Lee, R.W. and McMurtry, M.G., The mechanical stability of various noncemented tibia1 components, Clin. Orthop. 1988, 226, 38 Vanderby, R., Manley, P.A., Kohles, S.S., Belloli, D.M. and McBeath, A.A., A micmmotion comparison of cemented and porous ingrowth total hip replacements in a canine model, Transactions of the 35th Annual Meeting of the Orthopaedic Research Society Las Vegas, NV, USA, 1989, p 577 Zalenski, E., Jasty, M., O’Connor, D.O., Page, A., Krushell, R., Bragdon, C.R., Russotti, G. and Harris, W.H., Micromotion of porous-surfaced, cementless prostheses following 8 months of in vitro bone ingrowth in a canine model, Transactions of the 35th Annual Meeting of the Orthopaedic Research Society Las Vegas, NV, USA, 1969. p 377 Anderson, G.I., Hearn, T.C., Cuncins, A., Waddell, J.P., Richards, R.R. and Ling, H., Micromotion between three femoral stem designs and canine femora immediately postoperatively and with six and twentyfour months ingrowth, Transactions of the 36th Annual Meeting of the Orthopaedic Research Society New Orleans, LA, USA, 1990, p 463 Anderson, G.I., Finkelstein, J.A., Waddell, J.P., Richards, R.R. and Schemitsch, E., Madreporic-surface femoral arthmplasties in the dog: histomorphometric analysis in relation to micromotion, Transactions of the 37th Annual Meeting of the Orthopaedic Research Society Anaheim, CA, USA, 1991. p 536 Cook, SD., Clinical, radiographic and histologic evaluation of retrieved human noncemented porous coated implants, 1. Long-term Effects Med. Implants 1991, 1, 11-51 Cook, S.D., Thomas, K.A. and Haddad, R.J., Histologic analysis of retrieved human porous-coated total joint components, Clin. Orthop. 1988, 234, 90 Cameron, H.U., Pilliar, R.M. and Macnab, I., The effect of movement on the bonding of porous metal to bone, J, Biomed. Mater. Res. 1973, 7, 301-310 Pilliar, R.M., Lee, J.M. and Maniatopoulos, C., Observation on the effect of movement on bone ingrowth into porous surfaced implants, CZin. Orthop. 1986, 206, 108-113 Burke, D.W., Bragdon, C.R., O’Connor, D.O., Jasty, M., Haire, T. and Harris, W.H., Dynamic measurement of the interface mechanicsin vivo and the effect of micmmotion on bone ingrowth into a porous surface device under controlled loadsin vivo, Transactions of the 37th Annual

Biomaterials 1993, Vol. 14 No. 8

Differentiation of hard tissues: S. Goodman and P. Aspenberg 569

48

49

50

51

52

53

54

55

Meeting of the Orthopaedic Research Society Anaheim, CA, USA, 1991, p 103 Hollis, J.M., Hofmann, C., Stewart, C.L., Flahiff, CM. and Nelson, C., Effect of micromotion on ingrowth into porous coated implants using a transcortical model, Transactions of the 38th Annual Meeting of the Ortho- paedic Research Society Washington, DC, USA, 1992, p 584 Schatzker, J., Horne, J.G. and Sumner-Smith, G., The effect of movement on the holding power of screws in bone, Clin. Ortbop. 1975, 111, 258-282 Evans, B., Cohen, C., Mitchell, J., Heppenstall, R., Ducheyne, P. and Cuckler, J., Relationships between canal fill and interfacial displacement for an in viva loaded porous ingrowth femoral prosthesis, Transactions of the 36th Annual Meeting of the Orthopaedic Research Society New Orleans, LA, USA, 1990, p 201 Krushell, R., Zalenski, E.B., Page, A., Bragdon, C.R., O’Connor, D.O., Jasty, M. and Harris, W.H., The role of the distal non-porous portion of the stem in the stability of proximally porous coated canine femoral implants after bone ingrowth, Transactions of the 35th Annual Meeting of the Orthopaedic Research Society Las Vegas, NV, USA, 1989, p 404 Zalenski, E.B., Jasty, M., Bragdon, CR., Krushell, R., O’Connor, D.O., Page, A. and Harris, W.H., The effect of proximally coated prostheses on stress transfer and micmmotion with and without the distal stem, Transactions of the 36th Annual Meeting of the Ortbopaedic Research Society New Orleans, LA, USA, 1990, p 203 O’Connor, D.O., Davey, J.R., Zalenski, E., Burke, D.W. and Harris, W.H., Femoral component offset: its effect on micmmotion ln stance and stair climbing, Transactions of the 35th Annual Meeting of the Orthopaedic Research Society Las Vegas, NV, USA, 1989, p 409 Sugiyama, H., Whiteside, L.A. and Engh, A., Torsional fixation of the femoral component in total hip replace- ment: the effect of surgical technique, Transactions of the 36th Annual Meeting of the Orthopaedic Research Society New Orleans, LA, USA, 1990, p 258 Soballe, K., Hansen, ES., Rasmussen, H.B., Jorgensen, P.H. and Bunger, C., Hydroxyapatite coating provides a stronger fibrous anchorage of implants with controlled

58

57

58

59

80

61

62

63

84

micmmotion, Transactions of the 36th Annual Meeting of the Orthopaedic Research Society New Orleans, LA, USA, 1990, p 206 Soballe, K., Hansen, ES., Rasmussen, H.B. and Bunger, C., Hydroxyapatite implant coating modifies membrane formation during unstable mechanical conditions, Trans- actions of the 37th Annual Meeting of the Orthopaedic Research Society Anaheim, CA, USA, 1991, p 35 Soballe, K., Hansen, E.S., Rasmussen, H.B. and Bunger, C., Hydroxyapatite coating converts fibrous anchorage to bony fixation during continuous implant loading, Trans- actions of the 38th Annual Meeting of the Orthopaedic Research Society Washington, DC, USA, 1992, p 292 Rubin, C.T. and McLeod, J-J., Promotion of bony ingmwth by frequency specific, low amplitude mechanical strain, Transactions of the 38th Annual Meeting of the Orthopaedic Research Society Washington, DC, USA, 1992, p 89 Albrektsson, T. and Albrektsson, B., Osseointegration of bone implants. A review of an alternative mode of fixation, Acta Orthop. Stand. 1987, 58, 587-577 Albrektsson, T., Jacobsson, M. and Kllebo, P., The harvest chamber - a newly developed implant for analysis of bone remodelling in situ, in Biomaterials and Biomechanics 1983 (Eds P. Ducheyne, G. Van der Perre and A.El Ubert), Elsevier, Amsterdam, The Netherlands, 1984, pp 283-288 Aspenberg, P., Goodman, S., Toksvig-Larsen, S., Ryd, L. and Albrektsson, T., Intermittent micromotion inhibits bone ingrowth: experiment using titanium implants in rabbits, Acta Orthop. Stand. 1992, 63, 141-145 Goodman, S., Toksvig-Larson, S. and Aspenberg, P., Ingrowth of bone into pores in titanium chambers implanted in rabbits: effect of pore cross-sectional shape in the presence of dynamic shear, 1. Biomed. Mater. Res. 1993, 27, 247-253 Goodman, S. and Aspenberg, P., The effect of amplitude of micromotion on bone ingrowth into titanium chambers implanted in the rabbit tibia, Biomaterials [in press) Huiskes, R., Weinans, H., Grootenboer, H.J., Dalstra, M., Fudala, B. and Sloof, T.J., Adaptive bone-remodeling theory applied to prosthetic-design analysis, I. Biomecb. 1987, 20, 1135-1150

Biomaterials 1993. Vol. 14 No. 8