Journal of Orthopaedic Research 96o0608 Raven Press, Ltd., New York 0 1991 Orthopaedic Research Society
Modulation of Bone Loss During Disuse by Pulsed Electromagnetic Fields T. M. Skerry, *M. J. Pead, and *L. E. Lanyon Comparative Orthopaedic Research Unit, Department of Anatomy, University of Bristol, Bristol, U.K., and *Department of Veterinary Basic Science, Royal Veterinary College, London, U . K .
Summary: The effect of pulsed electromagnetic fields (PEMFs) on bone loss associated with disuse was investigated by applying 1.5 Hz repetitions of 30 ms bursts of asymmetric pulses, varying from +2.5 to - 135 mV, to bones deprived of their normal functional loading. The proximal portion of one fibula in each of a group of ovariectomised adult female beagle dogs was isolated from functional loading in vivo by proximal and distal osteotomies. Comparison of these prepared bones with their intact contralateral controls after 12 weeks, showed a 23% reduction in cross-sectional area. In similarly prepared bones exposed to PEMFs for 1 h per day, 5 days per week, this bone loss was substantially and significantly reduced to 9% Cp = 0.029). There was no evidence of any new bone formation on the periosteal surface of prepared fibulae in treated or untreated situations. PEMF treatment was not associated with any significant change in number of osteons per mm2 formed within the cortex of the bones, their radial closure rate, or their degree of closure. The modulation in loss of bone area associated with exposure to PEMFs can, therefore, be inferred to be due to a reduction in resorption on the bone surface. Key Words: Pulsed electromagnetic fields-Surface modeling/remodelingIntracortical remodeling-Disuse osteopenia.
The concept of using electrical stimulation to prevent disuse osteoporosis dates from the discovery of strain-generated potentials in bone (10) and the development of the hypothesis that functional strains in bone tissue influence modelingkemodeling through the agency of such potentials (1,2,14). Subsequent experiments have investigated the effects of direct (4,5,13), and indirect (6) electrical stimulation, as well as magnetic stimulation (2,3) on modeling and remodeling in normal and abnormal bones. Recent interest has centered on noninvasive means of electrical stimulation, which involve either capacitive coupling (6-8) or pulsed electromagnetic fields (3,16,19). Brighton and his colleagues (7,8) have shown, in a rat sciatic neurectomy model, that continuous exposure to capacitively-
The normal mass and architecture of load-bearing parts of the skeleton are only maintained as a consequence of continued functional loading. Removal or reduction of this stimulus is associated, in a number of situations, with bone loss (12,18,21-23). Since restoration of bone mass following restored function is slow, and may be incomplete ( 1 1), there are obvious advantages to maintaining bone mass during periods of bed rest, paralysis, or immobilization. The possibility of preserving bone mass during other disorders, such as postmenopausal osteoporosis, has additional significance. Received April 18, 1989; accepted November 14, 1990. Address correspondence and reprint requests to Dr. L. E. Lanyon at The Department of Veterinary Basic Sciences, The Royal Veterinary College, London NWl OTU, U.K.
600
MAGNETIC FIELD EFFECTS ON DISUSE OSTEOPENIA
coupled fields appears to modulate bone loss by an increase in bone formation, with no effect on resorption. McLeod et al. (16) showed that short periods of stimulation with PEMFs not only prevent development of disuse osteoporosis-in a functionally-isolated avian ulna model, but also engender new bone deposition. The experiments reported here were designed to investigate the effects of PEMFs in a mammalian preparation, similar to that used by McLeod et al. (16), in birds, for electromagnetic stimulation, and by Rubin and Lanyon (18) for mechanical stimulation. To do this, we used the fibula shaft of the adult ovariectomized beagle and deprived it of functional loading, while maintaining its neurovascular connections. This model allows direct comparison of modeling and remodeling between the intact fibula on one side and the corresponding region of the contralateral prepared bone, which is isolated from load-bearing. This preparation has the advantage over whole limb casting in that it is only the load-bearing of the fibular segment that is impaired. Function of the whole limb, its innervation, and its blood flow remain normal.
601
luck longus muscles. The bone is transected and a polytetrafluoroethylene (PTFE) cap fitted over the end to prevent union of the osteotomy. The bone is approached, similarly transected and capped 4 cm distal to the first osteotomy, so that the whole of the free portion of the fibula is isolated from functional loading. This free segment maintains all its normal soft tissue attachments but cannot be significantly loaded by the animal (Fig. 1). This surgery does not affect the animals’ gait, and 2 days postoperatively, all dogs in this study were weight-bearing normally. The dogs had a coil strapped to their left leg for a period of 1 h per day, 5 days per week. The coil was
I
MATERIALS AND METHODS
A group of 17 adult (18 months old) female beagle dogs were ovariohysterectomized and allowed to recover for 4 months, in order that their bone remodeling could equilibrate to the new hormonal environment. Ovariohysterectomy does not have a major effect on cortical bone remodeling parameters in the beagle, although there may be a slight increase in the resorption phase (20). However the principal benefit of ovariohysterectomy is the abolition of the estrous cycle. The female beagle has a 6 month estrous cycle and the natural estrogen levels fluctuate over this period (20). To our knowledge, there have been no studies of changes in remodeling parameters in relation to this fluctuation. However, it remains a confounding variable, which may be eliminated by ovariohysterectomy. Four months after ovariohysterectomy , dogs were anesthetized and the fibula on one side prepared, to protect it from mechanical loading. This preparation was made by isolation of the bone’s shaft by proximal and distal osteotomies. The surgical preparation, under general anaesthesia, involves exposure of the head of the left fibula by incising over the proximal one third of the bone and separating the lateral digital extensor and flexor hal-
f”
FIG. 1. Schematic diagram of the functionally isolated canine fibula preparation. Cranial view of the tibia and fibula. The free portion of the fibula, at A, is isolated from the effects of functional load by the proximal and distal osteotomies (X and Y). Healing of these is prevented by the teflon caps at B and B‘.
J Orthop Res, Vol. 9, No. 4, 1991
602
T. M . SKERR Y ET AL.
connected to a power pack, strapped to the contralateral leg with a light harness. This arrangement allowed consistent positioning of the coils on the lateral side of the limb, over the prepared sections of the left fibula. The coils are elliptical in shape, and contoured to fit the lateral surface of the limb so that the fibula segment lies at the center of the ellipse along the axis of the longest elliptical diameter (Fig. 2). In nine of the dogs (Group I), the coils were not energized. In the remaining eight (Group 2), power packs were activated so that the coil produced a pulsed field. The field was commercially available from Electro Biology Inc., and consisted of 1.5 Hz repetitions of 30 ms bursts of an asymmetric pulse, varying from + 2.5 to - 135 mVs (Fig. 3). The characteristics of the magnetic field induced by these coils are shown by the traces (Fig. 4) from a search coil, placed by the coil in the same orientation and position as the fibula during the experiment. The contralateral fibulae in these beagles were -18 cm
FIG. 2. Photograph to show the positioning of the field coil over the dog’s leg, using the foam harness.
J Orthop Res, Vol. 9, No. 4,1991
PEMF Waveform Characteristics
0.27 msecs
30 msecs
-
0.67~-
FIG. 3. Input waveform to coils, which produced electromagnetic field to which one group of dogs were exposed. [electrobiology source-(EBI)].
apart, when standing or sitting, and measurement of the magnetic field in animals undergoing active treatment at the lateral side of the limb, adjacent to the intact fibulae, showed a 90% attenuation of field strength. During application of the coils, animals were in individual pens, in the same kennel area. Pens were divided to a height of 1.4 m by a 6 cm thick reinforced masonry wall, which attenuated the field by 75% alone. Animals with nonenergized coils were in one kennel bank and those with energised coils in the opposite bank, separated by 3.6 m. There was no detectable pulsed field in the kennels of dogs with nonenergized coils during the treatment period. No difference in ambient field was detected between kennels. All animals received a series of fluorochrome markers over the 12 week period of the experiment. A single intravenous dose of calcein (British Drug House; Poole, Dorset) at 4 mg/kg was given at the time of surgery, a double intravenous label of tetracycline (Terramycin Q-50: Pfizer Ltd; Sandwich Kent) at 14 mg/kg in the seventh and eighth week, and a final intravenous label of xylenol orange (British Drug House; Poole, Dorset) at 30 mg/kg 1 week before the end of the experiment. All labels were made up as solutions of neutral pH and sterilized by passage through a micropore filter. Twelve weeks after the fibular preparation was made, the dogs were sacrificed by an intravenous overdose of pentobarbitone, and the fibulae re-
MAGNETIC FIELD EFFECTS ON DISUSE OSTEOPENIA
603
A
B
C
D
FIG. 4. Oscilloscope recordings from a search coil placed in a similar position in relation to coil as that of the isolated fibula in the experiment. Sensitivity in mv per division is given on the left of each photograph and sweep speed on the right. (SourceEBI). A, width of burst of pulses; B, rate of repetition of pulse burst; C, length and character of positive section of individual pulse; D, length and character of positive section of individual pulse.
moved from both hind legs of each animal. The two bones were marked and cut in equivalent positions corresponding to the midpoint of the isolated section of the left bone in each case. This point was chosen so that the analysis could be performed as far from the surgical sites as possible. Sections, 100 pm thick, were cut at this point, using a diamondedged annular saw (Microslice I1 Cambridge Instruments). Two further blocks, 6 mm long, were taken from other equivalent regions of each bone. One of these was decalcified in ethylene diamine tetraacetic acid (EDTA) (British Drug House: Poole, Dorset) and embedded in wax to provide sections for histology, while the other was embedded in poly-
methyl methacrylate (British Drug House: Poole, Dorset) to provide sections for fluorescent light microscopy. Thin sections were all cut on a Reichert Jung Polycut microtome, fitted with steel or tungsten carbide-edged blades. The 100 pm sections were radiographed using a Faxitron machine and Kodak high definition film. Microradiographs were enlarged and digitized so that the perimeter of the mineralized bone could be defined and the enclosed area measured. The 10 pm undecalcified sections were examined using incident ultraviolet light. The numbers of osteons containing each of the labels were ascertained. The radial closure rate of double-labeled osteons was es-
J Orthop Res, Vol. 9, No. 4 , 1991
T. M . SKERRY ET AL.
604
timated by making not less than four measurements of the distance between the two labels in each osteon. In addition, the overall diameter of the central canal of these infilling osteons was measured to provide an index of the degree to which they had completely infilled. This measurement was only performed on osteons, double-labeled at 7 weeks, and in which the diameter of labeled rings was between 20-30 pm. Such osteons would have been able to close completely by the end of the study period at the measured radial closure rate and, thus, measuring their central canal diameter at the end of the study gave an estimate of the amount of incomplete osteonal closure taking place. Undecalcified sections were examined, and the position and identity of the various cell types recorded. However, since these sections only represented activity at the end of the period of study, they were not used for quantitative estimations. The significance of the differences of bone area and fluorochrome-labeled parameters between left and right fibulae in each dog was assessed using a Student’s paired t-test. Data from groups of dogs were compared with an unpaired t-test. Analyses were performed on the measured values. Results from area measurements were converted to percentages of original area, for convenient comparison. RESULTS
Cross-sectional Area Isolated fibulae of animals in Group 1, which carried the power packs and nonenergized coils, had a significant loss in cross-sectional area when compared with the intact contralateral control. The mean loss in cross-sectional area in this group was 23.1%. In contrast, after functional isolation and treatment with the PEMFs for the same 12 week period, bone loss in the treated group (Group 2) was
9.5%. The difference in bone loss between isolated treated and isolated untreated fibulae was statistically significant (p = 0.029). These results are summarized in Table 1. Exposure to PEMFs was, therefore, associated with a substantial and statistically significant reduction in amount of bone lost in the disuse preparation (Fig. 5). There was no significant difference between cross-sectional areas of intact fibulae from Groups 1 and 2, indicating that the residual component of the magnetic field from the treatment side had no effect on the intact bones. Therefore, reduction in bone loss associated with PEMF stimulation was sufficient to eliminate any significance in the difference between intact and isolated bones, treated in this way. Frequency of Intracortical Remodeling Events Functional isolation without PEMF treatment (Group 1) caused a significant increase in the number of remodeling events within the cortex. Whereas there was no significant change between the number of active osteons per mm2 at the time of surgery and the number at 7 weeks in intact fibulae, there was a significant increase in number of active osteons in isolated fibulae over the same time period 0, < 0.001) (Table 2). Functional isolation with PEMF treatment (Group 2) was associated with a similar statistically significant increase in the number of active osteons per mm2 in the isolated fibulae (p < 0.001). Again, there was no significant difference between values for intact fibulae from treated and untreated groups. There was also no significant difference, at 7 weeks, between the number of active osteons for isolated treated and isolated untreated fibulae at (4.7 k 0.48 and 3.6 k 0.37). Neither was there any significant difference in the number of active osteons in intact fibulae in each group or between groups over the experimental period (Table 2). Bone Formation Rate in Secondary Osteons
TABLE 1. Percentage bone loss
Isolated, not treated (nonenergized coil) Percentage bone loss (crosssectional area)
23.1
.rt
3.82
J Orthop Res, Vol. 9, No. 4, 1991
0, = 0.029)
Isolated, treated (energized coil)
9.5
.rt
4.07
Comparison of intact and isolated fibulae showed that functional isolation was associated with a statistically significant increase in osteonal closure rate in untreated (p = 0.022) and treated (p = 0.023) groups (Table 2). However, there was no significant difference between closure rates in treated and untreated fibulae (1.8 2 0.04 pm per day and 1.7 2 0.7).
605
MAGNETIC FIELD EFFECTS ON DISUSE OSTEOPENIA A
B
C
D
FIG. 5. Microradiographs from cross-sections of canine fibulae taken at midpoint of the isolated section and at an equivalent level on the contralateral side. A, intact contralateral control (untreated group); B, functionally isolated for 12 weeks (untreated group); C, intact contralateral control (treated group); D, functionally isolated and subjected to PEMF stimulation for 1 h per day, 5 days per week.
Degree of Osteonal Closure Measurement of the diameter of the osteonal canals remaining at 12 weeks and labeled at 7 weeks showed the mean diameter to be significantly greater in isolated fibulae than in intact fibulae, in treated (p = 0.003) and untreated (p = 0.044) groups (Table 2). There was no significant difference in the mean diameter between isolated treated and isolated untreated fibulae (21.7 pm k 1.29 and 19.7 2 1.22). Surface Modeling/Remodeling Microscopic examination of fluorochrome-labeled bones showed that the new periosteal surface in treated and untreated isolated fibulae was either resorptive or quiescent, with no evidence of
bone formation having occurred during the experimental period. Numerous surface lacunae, containing osteoclasts, were present on the periosteal surfaces of the isolated bones. DISCUSSION In the beagle, the fibula is a slim bone, without a medullary cavity, which articulates with the tibia proximally and is tightly attached to the tibia in its distal portion. The influence of load-bearing on fibular structure is evidenced by the dramatic decrease in the bone’s cross-sectional area when loadbearing is withdrawn. The proximal fibula has no specific nutrient artery, which could be damaged by surgery, but derives its blood supply from the periosteal vessels in the surrounding musculature (17).
J
Orthop Res, Vol. 9, No. 4, 1991
T. M . SKERRY ET AL. TABLE 2. Comparison of area and osteonal data
Intact Nontreated Isolated Intact Treated Isolated
Cross-sectional area (mm’) (at 12 weeks)
No. active osteones per mmz (at surgery)
No. active osteones per mm2 (at 7 weeks)
Residual diameter osteonal canal (at 12 weeks)
Osteonal radial closure rate (I*m per day)
7.37 i 0.49 p < 0.001 5.72 t 0.57
0.6 If- 0.21 n.s. 1.1 i 0.28“
0.7 t 0.15 p < 0.001 3.6 i 0.37”
13.9 t 2.75 p = 0.044 19.7 i 1.22
p = 0.022
7.13 t 0.58 n.s. 6.31 i 0.25
1.3 t 0.49 ns. 1.3 t 0.21“
1.2 i 0.23 p < 0.001 4.7 t 0.48d
15.3 t 1.58 p = 0.003 21.7 t 1.29
1.4 t 0.12 p = 0.023 1.8 rt 0.04
1.5 .t 0.11 1.7 f 0.07
Figures quoted are mean values t the Standard error of the mean, p values between the figures in the columns represent the statistical significance of the difference between those values as computed using a paired t-test and n s . indicates no significance. The level of statistical significance between the values marked a and is p < 0.001 as is that between those marked and ‘.
”
The dogs in this study were ovariectomized. This procedure abolished the six month ovarian cycle and, thus, the animals had a low but constant estrogen level over the study period. The possibility that ovariectomy may change the remodeling parameters in the beagle has been investigated, showing little effect in the parameters for cortical bone (20). The possibility of a transient effect on cortical bone over the 3 months post ovariohysterectomy (15) was allowed for by the 4 month wait between ovariohysterectomy and fibula preparation. It is also possible that the onset of disuse osteopenia may be affected by the lowered steady estrogen levels in these dogs, compared to the higher cyclic levels in entire animals. Snow (20) postulated that the resorption phase might be lengthened in ovariohysterectomised beagles either directly or by an uncoupling of resorption and formation. However, if such a process is hormonally regulated, it will be changed to the same degree on both prepared and intact fibulae. Such a phenomenon might affect rate of bone loss, but the basic cellular mechanism will be the same in the ovariohysterectomized or entire animal. Ovariohysterectomy should therefore, not affect the nature of the PEMF effect qualitatively but also eliminate any confounding effect of estrogen-related variation in estrogen level. Results of the histomorphometric study show that functional isolation produced a profound effect on internal remodeling and modeling on the surface. This was demonstrated by an increase in the number of active osteons, their radial closure rate, and degree of incomplete infilling. In no case was there a significant difference in any of these parameters between isolated treated and isolated untreated fibulae. This indicates that whereas functional isolation had a substantial effect on intracortical remod-
J Orthop Res, Vol. 9, No.4, 1991
eling parameters, this effect was not influenced by PEMFs. Since isolated treated fibulae lost significantly less bone than isolated untreated fibulae, it can be inferred that PEMFs had their effect on the bone surface, as this is the only other site at which cross-sectional area could be lost. This inference is consistent with the appearance of the bone surface. If modeling/remodeling activity at the bone surface had consisted of both formation and resorption, this effect of PEMFs could have been achieved by an alteration in the balance of these two processes. However, since there was no fluorochrome evidence of surface bone formation over the experimental period, it seems reasonable to infer that the PEMFs exerted a direct effect in modulating resorption alone (Fig. 5). In the system that we used to apply PEMFs to the limb, there was a fall off in field strength, detectable with a search coil of -90% between the limb to which the coil was applied and the contralateral limb. The lack of difference in any of the remodeling parameters or cross-sectional areas between intact fibulae opposite treated or untreated limbs suggests that there was no effect of the residual PEMF on this bone. This may be because the attenuation of the magnetic field is sufficient enough so that residual PEMFs have no effect on the contralateral limbs. Alternatively, it is possible that PEMFs only exert an influence on bones undergoing accelerated remodeling. If this were so, it would be of considerable relevance in the applicability of this technique. It is of considerable interest that the effect of PEMFs is not a generalized modulation of the remodeling response but appears, instead, to be a specific effect on resorption at the bone’s surface. The mechanism by which pulsed fields affect the devel-
MAGNETIC FIELD EFFECTS ON DISUSE OSTEOPENIA
opment of disuse osteopenia is not clear. It is possible that the fields could affect bone cells directly in a unique manner, or that they cause changes that mimic those induced by loading. Since the means by which mechanical strain affects bone cells is also unknown, this distinction is, at present, unhelpful. The finding that the PEMF waveform used in this study acts primarily on surface resorption contrasts with Brighton’s findings (8), where capacitatively coupled field reduced bone loss during disuse, by an effect on bone formation. This difference between the effects of two different electrical regimes should not be taken as lessening the potential usefulness of PEMFs to cortical bone remodeling, but quite the reverse. It is inconceivable that all electrical treatments should have a similar, and beneficial, effect on modeling/remodeling activity. Other workers have also concluded that the response of bone to electrical stimulus may be signal specific (3,9,16). If electromagnetic fields can influence the processes of remodeling, it is encouraging that different effects can be produced by different electrical regimes. This provides the opportunity for selective control of different aspects of the modeling/ remodeling process. Potential implications of the results we present here are considerable. PEMF treatment of human patients for 1 h a day for 5 days a week is not only manageable but convenient. One hour is a considerably shorter time than some PEMF treatments for non-unions (3) but it is possible that even shorter treatment times may be effective. Brighton (9) has recently shown that cell metabolism may be affected by low energy electric fields, applied for periods as short as 2 min per day. Our results in dogs show a potential benefit from such PEMF treatment, which would be useful even if applicable only to immobilization osteoporosis. However, if bone loss in postmenopausal women could be similarly modulated, then the potential benefit of pulsed electromagnetic fields would be enormous. This possibility depends on whether PEMFs have a similar effect in regions of the human skeleton affected by postmenopausal bone loss. It may be significant that the effect of PEMFs in our study has been to modulate surface resorption, as this is the means by which trabeculae loose bone and cortices become thinned. CONCLUSIONS
Application of pulsed electromagnetic fields to a functionally-isolated bone for 1 h each day, in ex-
607
perimental dogs, reduces the loss of bone associated with disuse. This reduction in bone loss appears to be due to inhibition of resorption on the bone surface, with no effect on remodeling within the cortex. Acknowledgment:This work was funded by Electro Biology Inc, and performed while TMS was a Wellcome Veterinary Research Fellow. MJP is a Horserace Betting Levy Board Fellow.
REFERENCES 1. Bassett CAL, Pawluk RJ, Becker RO: Effects of electric
currents on bone in vivo. Nature 204:652-654, 1964 2. Bassett CAL, Chokshi HR, Hernandez E, Pawluk RJ, Strop M: The effect of pulsing electromagnetic fields on cellular calcium and calcification of nonunions. In: Electrical Properties of Bone and Cartilage, ed by CT Brighton, Black J, Pollack SR. New York, Grune and Stratton, 1979, pp 42744 1 3. Bassett CAL: The development and application of pulsed electromagnetic fields (PEMFs) for ununited fractures and arthrodesis. Orthop Clin North Am 15:61-87, 1984 4. Brighton CT, Friedenberg ZB, Zemsky LM, Pollis PR: Direct current stimulation of non-union and congenital pseudarthrosis. J Bone Joint Surg 57A:368-377, 1975 5. Brighton CT: The semi-invasive method of treating nonunion with direct current. Orthop Clin North Am 15:33-45, 1984 6. Brighton CT, Katz MJ, Go11 SR, Nichols CE, 111, Pollack SR: Prevention and treatment of sciatic denervation disuse osteoporosis in the rat and capacitively coupled electrical stimulation. Bone 6:87-97, 1985 7. Brighton CT, Taddunni GT, Pollack SR: Treatment of sciatic denervation disuse osteoporosis in the rat tibia with capacitive coupled electrical stimulation. J Bone Joint Surg 64A:1022-1028, 1985 8. Brighton CT, Tadduni G, Go11 SR, Pollack SR: Treatment of denervation/disuse osteoporosis in the rat with a capacitively coupled electrical signal: effects on bone formation and bone resorption. J Orthop Res 6:676-684, 1988 9. Brighton CT, Jensen L, Pollack SR, Tolin BS, Clark CS: Proliferative and synthetic response of bovine growth plate chondrocytes to various capacitively coupled electrical fields. J Orthop Res 7:759-765, 1989 10. Fukada E, Yasuda I: On the piezoelectric effect of bone. J Physiol Soc Japan 12:115&1162, 1957 1 1 . Jaworski ZFG, Uhthoff HK: Reversibility of non-traumatic disuse osteoporosis during the active phase. Bone 7:431439, 1986 12. Krolner B, Toft B: Vertebral bone loss: an unheeded side effect of therapeutic bed rest. Clin Sci 64537-540, 1963 13. Lavine LS, Lustrin I, Shamos MH: An experimental model for studying the effect of electric current on bone in vivo. Nature 224: 11 12-1 113, 1969 14. McElhaney JH, Stalnaker R, Bullard R: Electrical fields and bone loss of disuse. J Biomech 1:47-52, 1968 15. Martin RB, Butcher RL, Sherwood LL, Patterson-Allen P, Boyd R, Dannucci G, Fanis D, Heitter D, Sharkey NA: Skeletal effects of ovariectomy in the dog. Orthop Trans 10:332, 1986
J Orthop Res, Vol. 9, No. 4, 1991
608
T . M . SKERRY ET AL.
16. McLeod KJ, Rubin CT, Lanyon LE: Optimisation of induced electric field frequency in the prevention of osteoporosis. Orthop Trans 11:334, 1987 17. Miller ME, Christiansen GC, Evans HE: The Anatomy of the Dog. Philadelphia PA, WB Saunders and Co, 88, 1964 18. Rubin CT, Lanyon LE: Regulation of bone mass by mechanical strain magnitude. Culc Tim Znt 37:411-417, 1985 19. Shimizu T, Zerweth JE, Videman T, Gill K, Mooney V, Holmes RE, Hagler HK: Bone ingrowth into porous calcium phosphate ceramics: Influence of pulsing electromagnetic field. J Orthop Res 6:248-258, 1988 20. Snow GR, Cook MA, Anderson C: Oophorectomy and cor-
J Orthop Res, Vol. 9, No. 4, 1991
tical bone remodeling in the beagle. CuZc Tiss Znt 3 6 5 8 6 590, 1984 21. Tilton FE, Degioanni TTC, Schneider VS: Long-term followup on Skylab bone demineralisation. Aviut Space Environ Med 51:1209-1293, 1980 22. Uhthoff HK, Jaworski ZFG: Bone loss in response to long term immobitisation. J Bone Joint Surg 60B:420-429, 1978 23. Woo SL-Y, Akeson WH, Coutitis RD, Rutherford L, Doty D, Jemmott GF, Amiel D: A comparison of cortical bone atrophy secondary to fixation with plates with large differences in bending stiffness. J Bone Joint Surg 58A:190-195, 1976
Modulation of Bone Loss During Disuse by Pulsed Electromagnetic Fields T. M. Skerry, *M. J. Pead, and *L. E. Lanyon Comparative Orthopaedic Research Unit, Department of Anatomy, University of Bristol, Bristol, U.K., and *Department of Veterinary Basic Science, Royal Veterinary College, London, U . K .
Summary: The effect of pulsed electromagnetic fields (PEMFs) on bone loss associated with disuse was investigated by applying 1.5 Hz repetitions of 30 ms bursts of asymmetric pulses, varying from +2.5 to - 135 mV, to bones deprived of their normal functional loading. The proximal portion of one fibula in each of a group of ovariectomised adult female beagle dogs was isolated from functional loading in vivo by proximal and distal osteotomies. Comparison of these prepared bones with their intact contralateral controls after 12 weeks, showed a 23% reduction in cross-sectional area. In similarly prepared bones exposed to PEMFs for 1 h per day, 5 days per week, this bone loss was substantially and significantly reduced to 9% Cp = 0.029). There was no evidence of any new bone formation on the periosteal surface of prepared fibulae in treated or untreated situations. PEMF treatment was not associated with any significant change in number of osteons per mm2 formed within the cortex of the bones, their radial closure rate, or their degree of closure. The modulation in loss of bone area associated with exposure to PEMFs can, therefore, be inferred to be due to a reduction in resorption on the bone surface. Key Words: Pulsed electromagnetic fields-Surface modeling/remodelingIntracortical remodeling-Disuse osteopenia.
The concept of using electrical stimulation to prevent disuse osteoporosis dates from the discovery of strain-generated potentials in bone (10) and the development of the hypothesis that functional strains in bone tissue influence modelingkemodeling through the agency of such potentials (1,2,14). Subsequent experiments have investigated the effects of direct (4,5,13), and indirect (6) electrical stimulation, as well as magnetic stimulation (2,3) on modeling and remodeling in normal and abnormal bones. Recent interest has centered on noninvasive means of electrical stimulation, which involve either capacitive coupling (6-8) or pulsed electromagnetic fields (3,16,19). Brighton and his colleagues (7,8) have shown, in a rat sciatic neurectomy model, that continuous exposure to capacitively-
The normal mass and architecture of load-bearing parts of the skeleton are only maintained as a consequence of continued functional loading. Removal or reduction of this stimulus is associated, in a number of situations, with bone loss (12,18,21-23). Since restoration of bone mass following restored function is slow, and may be incomplete ( 1 1), there are obvious advantages to maintaining bone mass during periods of bed rest, paralysis, or immobilization. The possibility of preserving bone mass during other disorders, such as postmenopausal osteoporosis, has additional significance. Received April 18, 1989; accepted November 14, 1990. Address correspondence and reprint requests to Dr. L. E. Lanyon at The Department of Veterinary Basic Sciences, The Royal Veterinary College, London NWl OTU, U.K.
600
MAGNETIC FIELD EFFECTS ON DISUSE OSTEOPENIA
coupled fields appears to modulate bone loss by an increase in bone formation, with no effect on resorption. McLeod et al. (16) showed that short periods of stimulation with PEMFs not only prevent development of disuse osteoporosis-in a functionally-isolated avian ulna model, but also engender new bone deposition. The experiments reported here were designed to investigate the effects of PEMFs in a mammalian preparation, similar to that used by McLeod et al. (16), in birds, for electromagnetic stimulation, and by Rubin and Lanyon (18) for mechanical stimulation. To do this, we used the fibula shaft of the adult ovariectomized beagle and deprived it of functional loading, while maintaining its neurovascular connections. This model allows direct comparison of modeling and remodeling between the intact fibula on one side and the corresponding region of the contralateral prepared bone, which is isolated from load-bearing. This preparation has the advantage over whole limb casting in that it is only the load-bearing of the fibular segment that is impaired. Function of the whole limb, its innervation, and its blood flow remain normal.
601
luck longus muscles. The bone is transected and a polytetrafluoroethylene (PTFE) cap fitted over the end to prevent union of the osteotomy. The bone is approached, similarly transected and capped 4 cm distal to the first osteotomy, so that the whole of the free portion of the fibula is isolated from functional loading. This free segment maintains all its normal soft tissue attachments but cannot be significantly loaded by the animal (Fig. 1). This surgery does not affect the animals’ gait, and 2 days postoperatively, all dogs in this study were weight-bearing normally. The dogs had a coil strapped to their left leg for a period of 1 h per day, 5 days per week. The coil was
I
MATERIALS AND METHODS
A group of 17 adult (18 months old) female beagle dogs were ovariohysterectomized and allowed to recover for 4 months, in order that their bone remodeling could equilibrate to the new hormonal environment. Ovariohysterectomy does not have a major effect on cortical bone remodeling parameters in the beagle, although there may be a slight increase in the resorption phase (20). However the principal benefit of ovariohysterectomy is the abolition of the estrous cycle. The female beagle has a 6 month estrous cycle and the natural estrogen levels fluctuate over this period (20). To our knowledge, there have been no studies of changes in remodeling parameters in relation to this fluctuation. However, it remains a confounding variable, which may be eliminated by ovariohysterectomy. Four months after ovariohysterectomy , dogs were anesthetized and the fibula on one side prepared, to protect it from mechanical loading. This preparation was made by isolation of the bone’s shaft by proximal and distal osteotomies. The surgical preparation, under general anaesthesia, involves exposure of the head of the left fibula by incising over the proximal one third of the bone and separating the lateral digital extensor and flexor hal-
f”
FIG. 1. Schematic diagram of the functionally isolated canine fibula preparation. Cranial view of the tibia and fibula. The free portion of the fibula, at A, is isolated from the effects of functional load by the proximal and distal osteotomies (X and Y). Healing of these is prevented by the teflon caps at B and B‘.
J Orthop Res, Vol. 9, No. 4, 1991
602
T. M . SKERR Y ET AL.
connected to a power pack, strapped to the contralateral leg with a light harness. This arrangement allowed consistent positioning of the coils on the lateral side of the limb, over the prepared sections of the left fibula. The coils are elliptical in shape, and contoured to fit the lateral surface of the limb so that the fibula segment lies at the center of the ellipse along the axis of the longest elliptical diameter (Fig. 2). In nine of the dogs (Group I), the coils were not energized. In the remaining eight (Group 2), power packs were activated so that the coil produced a pulsed field. The field was commercially available from Electro Biology Inc., and consisted of 1.5 Hz repetitions of 30 ms bursts of an asymmetric pulse, varying from + 2.5 to - 135 mVs (Fig. 3). The characteristics of the magnetic field induced by these coils are shown by the traces (Fig. 4) from a search coil, placed by the coil in the same orientation and position as the fibula during the experiment. The contralateral fibulae in these beagles were -18 cm
FIG. 2. Photograph to show the positioning of the field coil over the dog’s leg, using the foam harness.
J Orthop Res, Vol. 9, No. 4,1991
PEMF Waveform Characteristics
0.27 msecs
30 msecs
-
0.67~-
FIG. 3. Input waveform to coils, which produced electromagnetic field to which one group of dogs were exposed. [electrobiology source-(EBI)].
apart, when standing or sitting, and measurement of the magnetic field in animals undergoing active treatment at the lateral side of the limb, adjacent to the intact fibulae, showed a 90% attenuation of field strength. During application of the coils, animals were in individual pens, in the same kennel area. Pens were divided to a height of 1.4 m by a 6 cm thick reinforced masonry wall, which attenuated the field by 75% alone. Animals with nonenergized coils were in one kennel bank and those with energised coils in the opposite bank, separated by 3.6 m. There was no detectable pulsed field in the kennels of dogs with nonenergized coils during the treatment period. No difference in ambient field was detected between kennels. All animals received a series of fluorochrome markers over the 12 week period of the experiment. A single intravenous dose of calcein (British Drug House; Poole, Dorset) at 4 mg/kg was given at the time of surgery, a double intravenous label of tetracycline (Terramycin Q-50: Pfizer Ltd; Sandwich Kent) at 14 mg/kg in the seventh and eighth week, and a final intravenous label of xylenol orange (British Drug House; Poole, Dorset) at 30 mg/kg 1 week before the end of the experiment. All labels were made up as solutions of neutral pH and sterilized by passage through a micropore filter. Twelve weeks after the fibular preparation was made, the dogs were sacrificed by an intravenous overdose of pentobarbitone, and the fibulae re-
MAGNETIC FIELD EFFECTS ON DISUSE OSTEOPENIA
603
A
B
C
D
FIG. 4. Oscilloscope recordings from a search coil placed in a similar position in relation to coil as that of the isolated fibula in the experiment. Sensitivity in mv per division is given on the left of each photograph and sweep speed on the right. (SourceEBI). A, width of burst of pulses; B, rate of repetition of pulse burst; C, length and character of positive section of individual pulse; D, length and character of positive section of individual pulse.
moved from both hind legs of each animal. The two bones were marked and cut in equivalent positions corresponding to the midpoint of the isolated section of the left bone in each case. This point was chosen so that the analysis could be performed as far from the surgical sites as possible. Sections, 100 pm thick, were cut at this point, using a diamondedged annular saw (Microslice I1 Cambridge Instruments). Two further blocks, 6 mm long, were taken from other equivalent regions of each bone. One of these was decalcified in ethylene diamine tetraacetic acid (EDTA) (British Drug House: Poole, Dorset) and embedded in wax to provide sections for histology, while the other was embedded in poly-
methyl methacrylate (British Drug House: Poole, Dorset) to provide sections for fluorescent light microscopy. Thin sections were all cut on a Reichert Jung Polycut microtome, fitted with steel or tungsten carbide-edged blades. The 100 pm sections were radiographed using a Faxitron machine and Kodak high definition film. Microradiographs were enlarged and digitized so that the perimeter of the mineralized bone could be defined and the enclosed area measured. The 10 pm undecalcified sections were examined using incident ultraviolet light. The numbers of osteons containing each of the labels were ascertained. The radial closure rate of double-labeled osteons was es-
J Orthop Res, Vol. 9, No. 4 , 1991
T. M . SKERRY ET AL.
604
timated by making not less than four measurements of the distance between the two labels in each osteon. In addition, the overall diameter of the central canal of these infilling osteons was measured to provide an index of the degree to which they had completely infilled. This measurement was only performed on osteons, double-labeled at 7 weeks, and in which the diameter of labeled rings was between 20-30 pm. Such osteons would have been able to close completely by the end of the study period at the measured radial closure rate and, thus, measuring their central canal diameter at the end of the study gave an estimate of the amount of incomplete osteonal closure taking place. Undecalcified sections were examined, and the position and identity of the various cell types recorded. However, since these sections only represented activity at the end of the period of study, they were not used for quantitative estimations. The significance of the differences of bone area and fluorochrome-labeled parameters between left and right fibulae in each dog was assessed using a Student’s paired t-test. Data from groups of dogs were compared with an unpaired t-test. Analyses were performed on the measured values. Results from area measurements were converted to percentages of original area, for convenient comparison. RESULTS
Cross-sectional Area Isolated fibulae of animals in Group 1, which carried the power packs and nonenergized coils, had a significant loss in cross-sectional area when compared with the intact contralateral control. The mean loss in cross-sectional area in this group was 23.1%. In contrast, after functional isolation and treatment with the PEMFs for the same 12 week period, bone loss in the treated group (Group 2) was
9.5%. The difference in bone loss between isolated treated and isolated untreated fibulae was statistically significant (p = 0.029). These results are summarized in Table 1. Exposure to PEMFs was, therefore, associated with a substantial and statistically significant reduction in amount of bone lost in the disuse preparation (Fig. 5). There was no significant difference between cross-sectional areas of intact fibulae from Groups 1 and 2, indicating that the residual component of the magnetic field from the treatment side had no effect on the intact bones. Therefore, reduction in bone loss associated with PEMF stimulation was sufficient to eliminate any significance in the difference between intact and isolated bones, treated in this way. Frequency of Intracortical Remodeling Events Functional isolation without PEMF treatment (Group 1) caused a significant increase in the number of remodeling events within the cortex. Whereas there was no significant change between the number of active osteons per mm2 at the time of surgery and the number at 7 weeks in intact fibulae, there was a significant increase in number of active osteons in isolated fibulae over the same time period 0, < 0.001) (Table 2). Functional isolation with PEMF treatment (Group 2) was associated with a similar statistically significant increase in the number of active osteons per mm2 in the isolated fibulae (p < 0.001). Again, there was no significant difference between values for intact fibulae from treated and untreated groups. There was also no significant difference, at 7 weeks, between the number of active osteons for isolated treated and isolated untreated fibulae at (4.7 k 0.48 and 3.6 k 0.37). Neither was there any significant difference in the number of active osteons in intact fibulae in each group or between groups over the experimental period (Table 2). Bone Formation Rate in Secondary Osteons
TABLE 1. Percentage bone loss
Isolated, not treated (nonenergized coil) Percentage bone loss (crosssectional area)
23.1
.rt
3.82
J Orthop Res, Vol. 9, No. 4, 1991
0, = 0.029)
Isolated, treated (energized coil)
9.5
.rt
4.07
Comparison of intact and isolated fibulae showed that functional isolation was associated with a statistically significant increase in osteonal closure rate in untreated (p = 0.022) and treated (p = 0.023) groups (Table 2). However, there was no significant difference between closure rates in treated and untreated fibulae (1.8 2 0.04 pm per day and 1.7 2 0.7).
605
MAGNETIC FIELD EFFECTS ON DISUSE OSTEOPENIA A
B
C
D
FIG. 5. Microradiographs from cross-sections of canine fibulae taken at midpoint of the isolated section and at an equivalent level on the contralateral side. A, intact contralateral control (untreated group); B, functionally isolated for 12 weeks (untreated group); C, intact contralateral control (treated group); D, functionally isolated and subjected to PEMF stimulation for 1 h per day, 5 days per week.
Degree of Osteonal Closure Measurement of the diameter of the osteonal canals remaining at 12 weeks and labeled at 7 weeks showed the mean diameter to be significantly greater in isolated fibulae than in intact fibulae, in treated (p = 0.003) and untreated (p = 0.044) groups (Table 2). There was no significant difference in the mean diameter between isolated treated and isolated untreated fibulae (21.7 pm k 1.29 and 19.7 2 1.22). Surface Modeling/Remodeling Microscopic examination of fluorochrome-labeled bones showed that the new periosteal surface in treated and untreated isolated fibulae was either resorptive or quiescent, with no evidence of
bone formation having occurred during the experimental period. Numerous surface lacunae, containing osteoclasts, were present on the periosteal surfaces of the isolated bones. DISCUSSION In the beagle, the fibula is a slim bone, without a medullary cavity, which articulates with the tibia proximally and is tightly attached to the tibia in its distal portion. The influence of load-bearing on fibular structure is evidenced by the dramatic decrease in the bone’s cross-sectional area when loadbearing is withdrawn. The proximal fibula has no specific nutrient artery, which could be damaged by surgery, but derives its blood supply from the periosteal vessels in the surrounding musculature (17).
J
Orthop Res, Vol. 9, No. 4, 1991
T. M . SKERRY ET AL. TABLE 2. Comparison of area and osteonal data
Intact Nontreated Isolated Intact Treated Isolated
Cross-sectional area (mm’) (at 12 weeks)
No. active osteones per mmz (at surgery)
No. active osteones per mm2 (at 7 weeks)
Residual diameter osteonal canal (at 12 weeks)
Osteonal radial closure rate (I*m per day)
7.37 i 0.49 p < 0.001 5.72 t 0.57
0.6 If- 0.21 n.s. 1.1 i 0.28“
0.7 t 0.15 p < 0.001 3.6 i 0.37”
13.9 t 2.75 p = 0.044 19.7 i 1.22
p = 0.022
7.13 t 0.58 n.s. 6.31 i 0.25
1.3 t 0.49 ns. 1.3 t 0.21“
1.2 i 0.23 p < 0.001 4.7 t 0.48d
15.3 t 1.58 p = 0.003 21.7 t 1.29
1.4 t 0.12 p = 0.023 1.8 rt 0.04
1.5 .t 0.11 1.7 f 0.07
Figures quoted are mean values t the Standard error of the mean, p values between the figures in the columns represent the statistical significance of the difference between those values as computed using a paired t-test and n s . indicates no significance. The level of statistical significance between the values marked a and is p < 0.001 as is that between those marked and ‘.
”
The dogs in this study were ovariectomized. This procedure abolished the six month ovarian cycle and, thus, the animals had a low but constant estrogen level over the study period. The possibility that ovariectomy may change the remodeling parameters in the beagle has been investigated, showing little effect in the parameters for cortical bone (20). The possibility of a transient effect on cortical bone over the 3 months post ovariohysterectomy (15) was allowed for by the 4 month wait between ovariohysterectomy and fibula preparation. It is also possible that the onset of disuse osteopenia may be affected by the lowered steady estrogen levels in these dogs, compared to the higher cyclic levels in entire animals. Snow (20) postulated that the resorption phase might be lengthened in ovariohysterectomised beagles either directly or by an uncoupling of resorption and formation. However, if such a process is hormonally regulated, it will be changed to the same degree on both prepared and intact fibulae. Such a phenomenon might affect rate of bone loss, but the basic cellular mechanism will be the same in the ovariohysterectomized or entire animal. Ovariohysterectomy should therefore, not affect the nature of the PEMF effect qualitatively but also eliminate any confounding effect of estrogen-related variation in estrogen level. Results of the histomorphometric study show that functional isolation produced a profound effect on internal remodeling and modeling on the surface. This was demonstrated by an increase in the number of active osteons, their radial closure rate, and degree of incomplete infilling. In no case was there a significant difference in any of these parameters between isolated treated and isolated untreated fibulae. This indicates that whereas functional isolation had a substantial effect on intracortical remod-
J Orthop Res, Vol. 9, No.4, 1991
eling parameters, this effect was not influenced by PEMFs. Since isolated treated fibulae lost significantly less bone than isolated untreated fibulae, it can be inferred that PEMFs had their effect on the bone surface, as this is the only other site at which cross-sectional area could be lost. This inference is consistent with the appearance of the bone surface. If modeling/remodeling activity at the bone surface had consisted of both formation and resorption, this effect of PEMFs could have been achieved by an alteration in the balance of these two processes. However, since there was no fluorochrome evidence of surface bone formation over the experimental period, it seems reasonable to infer that the PEMFs exerted a direct effect in modulating resorption alone (Fig. 5). In the system that we used to apply PEMFs to the limb, there was a fall off in field strength, detectable with a search coil of -90% between the limb to which the coil was applied and the contralateral limb. The lack of difference in any of the remodeling parameters or cross-sectional areas between intact fibulae opposite treated or untreated limbs suggests that there was no effect of the residual PEMF on this bone. This may be because the attenuation of the magnetic field is sufficient enough so that residual PEMFs have no effect on the contralateral limbs. Alternatively, it is possible that PEMFs only exert an influence on bones undergoing accelerated remodeling. If this were so, it would be of considerable relevance in the applicability of this technique. It is of considerable interest that the effect of PEMFs is not a generalized modulation of the remodeling response but appears, instead, to be a specific effect on resorption at the bone’s surface. The mechanism by which pulsed fields affect the devel-
MAGNETIC FIELD EFFECTS ON DISUSE OSTEOPENIA
opment of disuse osteopenia is not clear. It is possible that the fields could affect bone cells directly in a unique manner, or that they cause changes that mimic those induced by loading. Since the means by which mechanical strain affects bone cells is also unknown, this distinction is, at present, unhelpful. The finding that the PEMF waveform used in this study acts primarily on surface resorption contrasts with Brighton’s findings (8), where capacitatively coupled field reduced bone loss during disuse, by an effect on bone formation. This difference between the effects of two different electrical regimes should not be taken as lessening the potential usefulness of PEMFs to cortical bone remodeling, but quite the reverse. It is inconceivable that all electrical treatments should have a similar, and beneficial, effect on modeling/remodeling activity. Other workers have also concluded that the response of bone to electrical stimulus may be signal specific (3,9,16). If electromagnetic fields can influence the processes of remodeling, it is encouraging that different effects can be produced by different electrical regimes. This provides the opportunity for selective control of different aspects of the modeling/ remodeling process. Potential implications of the results we present here are considerable. PEMF treatment of human patients for 1 h a day for 5 days a week is not only manageable but convenient. One hour is a considerably shorter time than some PEMF treatments for non-unions (3) but it is possible that even shorter treatment times may be effective. Brighton (9) has recently shown that cell metabolism may be affected by low energy electric fields, applied for periods as short as 2 min per day. Our results in dogs show a potential benefit from such PEMF treatment, which would be useful even if applicable only to immobilization osteoporosis. However, if bone loss in postmenopausal women could be similarly modulated, then the potential benefit of pulsed electromagnetic fields would be enormous. This possibility depends on whether PEMFs have a similar effect in regions of the human skeleton affected by postmenopausal bone loss. It may be significant that the effect of PEMFs in our study has been to modulate surface resorption, as this is the means by which trabeculae loose bone and cortices become thinned. CONCLUSIONS
Application of pulsed electromagnetic fields to a functionally-isolated bone for 1 h each day, in ex-
607
perimental dogs, reduces the loss of bone associated with disuse. This reduction in bone loss appears to be due to inhibition of resorption on the bone surface, with no effect on remodeling within the cortex. Acknowledgment:This work was funded by Electro Biology Inc, and performed while TMS was a Wellcome Veterinary Research Fellow. MJP is a Horserace Betting Levy Board Fellow.
REFERENCES 1. Bassett CAL, Pawluk RJ, Becker RO: Effects of electric
currents on bone in vivo. Nature 204:652-654, 1964 2. Bassett CAL, Chokshi HR, Hernandez E, Pawluk RJ, Strop M: The effect of pulsing electromagnetic fields on cellular calcium and calcification of nonunions. In: Electrical Properties of Bone and Cartilage, ed by CT Brighton, Black J, Pollack SR. New York, Grune and Stratton, 1979, pp 42744 1 3. Bassett CAL: The development and application of pulsed electromagnetic fields (PEMFs) for ununited fractures and arthrodesis. Orthop Clin North Am 15:61-87, 1984 4. Brighton CT, Friedenberg ZB, Zemsky LM, Pollis PR: Direct current stimulation of non-union and congenital pseudarthrosis. J Bone Joint Surg 57A:368-377, 1975 5. Brighton CT: The semi-invasive method of treating nonunion with direct current. Orthop Clin North Am 15:33-45, 1984 6. Brighton CT, Katz MJ, Go11 SR, Nichols CE, 111, Pollack SR: Prevention and treatment of sciatic denervation disuse osteoporosis in the rat and capacitively coupled electrical stimulation. Bone 6:87-97, 1985 7. Brighton CT, Taddunni GT, Pollack SR: Treatment of sciatic denervation disuse osteoporosis in the rat tibia with capacitive coupled electrical stimulation. J Bone Joint Surg 64A:1022-1028, 1985 8. Brighton CT, Tadduni G, Go11 SR, Pollack SR: Treatment of denervation/disuse osteoporosis in the rat with a capacitively coupled electrical signal: effects on bone formation and bone resorption. J Orthop Res 6:676-684, 1988 9. Brighton CT, Jensen L, Pollack SR, Tolin BS, Clark CS: Proliferative and synthetic response of bovine growth plate chondrocytes to various capacitively coupled electrical fields. J Orthop Res 7:759-765, 1989 10. Fukada E, Yasuda I: On the piezoelectric effect of bone. J Physiol Soc Japan 12:115&1162, 1957 1 1 . Jaworski ZFG, Uhthoff HK: Reversibility of non-traumatic disuse osteoporosis during the active phase. Bone 7:431439, 1986 12. Krolner B, Toft B: Vertebral bone loss: an unheeded side effect of therapeutic bed rest. Clin Sci 64537-540, 1963 13. Lavine LS, Lustrin I, Shamos MH: An experimental model for studying the effect of electric current on bone in vivo. Nature 224: 11 12-1 113, 1969 14. McElhaney JH, Stalnaker R, Bullard R: Electrical fields and bone loss of disuse. J Biomech 1:47-52, 1968 15. Martin RB, Butcher RL, Sherwood LL, Patterson-Allen P, Boyd R, Dannucci G, Fanis D, Heitter D, Sharkey NA: Skeletal effects of ovariectomy in the dog. Orthop Trans 10:332, 1986
J Orthop Res, Vol. 9, No. 4, 1991
608
T . M . SKERRY ET AL.
16. McLeod KJ, Rubin CT, Lanyon LE: Optimisation of induced electric field frequency in the prevention of osteoporosis. Orthop Trans 11:334, 1987 17. Miller ME, Christiansen GC, Evans HE: The Anatomy of the Dog. Philadelphia PA, WB Saunders and Co, 88, 1964 18. Rubin CT, Lanyon LE: Regulation of bone mass by mechanical strain magnitude. Culc Tim Znt 37:411-417, 1985 19. Shimizu T, Zerweth JE, Videman T, Gill K, Mooney V, Holmes RE, Hagler HK: Bone ingrowth into porous calcium phosphate ceramics: Influence of pulsing electromagnetic field. J Orthop Res 6:248-258, 1988 20. Snow GR, Cook MA, Anderson C: Oophorectomy and cor-
J Orthop Res, Vol. 9, No. 4, 1991
tical bone remodeling in the beagle. CuZc Tiss Znt 3 6 5 8 6 590, 1984 21. Tilton FE, Degioanni TTC, Schneider VS: Long-term followup on Skylab bone demineralisation. Aviut Space Environ Med 51:1209-1293, 1980 22. Uhthoff HK, Jaworski ZFG: Bone loss in response to long term immobitisation. J Bone Joint Surg 60B:420-429, 1978 23. Woo SL-Y, Akeson WH, Coutitis RD, Rutherford L, Doty D, Jemmott GF, Amiel D: A comparison of cortical bone atrophy secondary to fixation with plates with large differences in bending stiffness. J Bone Joint Surg 58A:190-195, 1976
