J. BIOMEI). MATEH. KES. VOL. 4, PP.

341-356 (1970)

Bio-Conipatibility in Selection of Materials for

Implantation*

CHARLES A. HOJISY, The Methodist Hospital, Baylor College of

Medicine, and Rice University, Houston, Texas 77025

SUMMARY
I n this paper the factors which contribute importantly to incompatibility and
functional failure of implant materials are examined. The interaction of these
criteria in the etiology of implant failure is discussed in terms of the host response
to an inert implant and the modulation of this response in the presence of cyto-
toxic sequellae. Cellular injury either of traumatic or biochemical origin is seen
to stimulate sequestration of the offending implant by fibrous tissue. Difficulties
associated with routine quantification of such response by animal implantation
studies have led to development of rapid in vitro techniques to predict the quan-
tity and cytotoxicity of moeities released by an implant. These techniques are
based on infrared analysis and primary mammalian tissue culture with a pseudo
body fluid to which materials are exposed a t elevated temperature. The protocols
allow formulation of rational criteria for rejection of candidate materials for
clinical implantation and acceptance for definitive preclinical animal implant
studies. The procedures are applicable for routine monitoring of materials
accepted for fabrication of prototype or production implants and of fabricated
prostheses.

INTRODUCTION
Bio-compatibility of implant materials has been evaluated most
often in the past by limited animal implantation followed by tentative
clinical application. The criterion for rejection was usually gross
tissue reaction adjacent to the implant. The absence of systematic
preclinical procedures for rational selection of prosthetic materials
must be counted incongruous and intolerable for this era.
Selection of a prosthetic implant must involve appraisal of in vivo
chemical reactivity and functional suitability. Although contiguous
* Presented a t the National Meeting, American Institute of Chemical Engi-
neers, Atlanta, Georgia, February, 1970.
341
@ 1970 by John Wiley & Sons, Inc.
342 HOMSY

tissue may respond to the simple physical presence of an inert implant

in ways which modulate function, these criteria are usually inter-
related. This interaction exhibits a spectrum over which either
criterion may dominate the etiology of implant failure.
At one end of this spectrum is the situation in which non-toxic
chemical interaction of an implant with the in vivo milieu leads to
changes in surface physical chemistry and related bulk mechanical
performance. This circumstance occurs in the absorption of lipid by
silicone rubber spheres in heart valve prostheses.' The absorption
aggravates surface stress crack tendency; cracking can lead to sphere
deformation and valve malfunction. The reported cracking of
prosthetic hand joints of silicone rubber is probably caused by this
mechanism. An analogous situation with metals is stress corrosion
cracking of stainless steel orthopedic appliances when attacked by
body fluid chloride ion.2-4
A more balanced combination of chemical reaction and functional
failure is illustrated by replacements for the femoral head made from
polyamide or polyacrylic. At resection there are substantially worn
areas in the articulating surface of the p r o ~ t h e s i s ; the
~ . ~ observed
proliferation of motion-restricting fibrotic tissue within the capsular
and extracapsular tissue is consistent with cytotoxic activity by the
wear debris. Frank mechanical failure was also frequently observed
with these prostheses and was probably the primary cause for their
abandonment.
Fortunately in recent times there are only a few clinical examples of
acute chemical toxicity with implant materials where failure did not
relate to function. This has been observed in the use of polyurethane
foam-in-place bone adhesive; these materials have a brief and dis-
astrous clinical history which well illustrates the chemical pitfalls of
premature clinical trial.'
I n this presentation we attempt to delineate the role of chemical
reactivity in the suitability of an implant; rapid in vitro screening
procedures for ordering of materials in terms of this reactivity will be
discussed.

Host Response to Inert Implant

Implantation of a prosthetic material necessarily involves surgical
trauma a t the implant site. If the surgeon were to decide against
actual implantation of a foreign material and proceed to close the
 BIO-COMPATIBILITY FOR IMPLANTATION 343

incision using best technique, there will proceed a complex series of

biochemical processes directed to heal the incisional site. According
t o a simplified scheme and in the absence of infection, there is acute
cellular activity with an influx of polymorphonuclear cells and edema
which is soon followed by the appearance of monocytes and fibro-
blasts.* The latter cells undertake reintegradation of the operatively
disturbed area by generation of collagenous tissue (scar tissue).
This fibrous tissue provides a framework on which reconstitution of
pre-operative tissue structure may preceed. I n time a major part of
scar tissue may be replaced by normal tissue s t r u ~ t u r e . ~The
exuberance of scar tissue development and its persistence is quite
variable among individuals and is greatest in epithelial or skin tissue.L0
The presence of an implant is an impediment to the normal healing
sequence just described. If the implant surface is sufficiently ex-
tensive (the order of centimeters in length and breadth) to present an
impervious barrier to body fluid, the healing process described above
leads to a fibrocartilagenous membrane of low cellularity which
isolates the implant from normal tissue. This membrane is the
body’s response to the stimulus of an inert foreign material which is
impervious to body fluid.
Figures 1, 2, and 3 show typical histological appearance of the
fibrous connective tissue membrane encapsulating t in. diameter
discs of perfluorocarbon polymer (TFE and FEI’ resins) which had
been implanted for 1 and 2 yr in canine ear. A fibrocartilaginous
tissue network is apparent with fiber orientation parallel to the
implant surface. There is little vascularity within the connective
tissue and no inflammatory cells.
On the other hand, i t has been shown by Oppenheimer, Hulbert,
Calnan,11-13and others that perforated structures and porous struc-
tures (pore size the order a millimeter or less) which allow sufficient
body fluid movement through and around the implant leads to a more
or less thorough infiltration of the implant with tissue rather than
sequestration. The woven or knitted fiber vascular prosthesis is an
example of the latter phenomenon; it is impregnated by a fibrin clot
at implantation. Seamless, knitted tubular grafts of polyethylene
terepthalate retrieved up to 7 yr postimplantation in most cases
exhibited luminal surfaces of fibrin and outer surfaces of thicker
fibrous connective tissue; these structures extended into and blended
within the interstices of the graft. Endothelial cells were not usually
demonstrable beyond approximately 1 cm from thc suture lines.14
344 HOMSY

Fig. 1. Fibrous tissue membrane adjacent, to TFE implant, X475, eosin arid
hematoxylin, 1 yr.

Fig. 2. Fibrous Lissiie membrane adjacent t,o FEP implant, X475, eosiri arid
hematoxylin, 1 yr.
 BIO-COMPATIBILITY FOR IMPLANTATION 345

Fig. 3. Fibrous membrane encapsdating perflnorocarbori polymer ( T F E Com-

posite), X 125, eosiit and hernatoxylin, 2 yr.
346 HOMRY

Orthopedic appliances of cobalt alloy (3oyOchromium, 5% mo-

lybdenum) which have been cast fabricated and therefore exhibit a
relatively rough and porous surface tend to be found a t re-section
intimately adherent to adjacent tissue and bone. The excellent
corrosion resistance of this alloy and the opportunity for metabolic
activity within the interstices of the implant surface allows intimacy
of tissue or bone. Ironically, this turns out to be a disadvantage to
the clinician since it greatly complicates the removal of symptomatic
cobalt alloy appliances. The important element for invasive rather
than sequestration action appears to be permeability sufficient for
effective metabolic activity.
Figure 4 presents a histological section of material directly apposed
to a cobalt alloy plate which had been used to fix canine longbone
fracture. Mature bone is prominent with only small areas of thin
fibrous membrane. When the surface porosity is reduced by filling
the surface with polytetrafluoroethylene polymer, development of
sequestration membrane is favored over intimate adherence of bone
or tissue.29

Fig. 4. Bone arid tissue adjacent to cobalt alloy implant, X400, eosin and hema-
toxylin, 10 months.
 BIO-COMPATIBILITY F011 I M P L A N T A T I O N 347

Cellular Injury in Host-Implant Interaction

The above discussion related to changes which occur in the absence
of toxic interaction between host arid implant. Laing and associates
have observed that solid metallic implants are incapsulated by
fibrous membrane the thickness of which appears to be proportional
t o the degree of metallic dissolution, i.e., biochemical irritation
adjacent t o the implant causes more vigorous sequestration response.15
In addition, the presence of toxic moeities is usually signaled by a
large population of inflammatory cells.
Thicker membrane is also observed whtn fracture fixation appli-
ances become loose allowing relative motion to occur between the
appliance and adjacent tissue or bone.
Figure 5 shows a section of tissue which was adjacent to a cobalt
alloy fracture fixation plate implanted in canine for 10 months. The
plate became loose and relative motion occurred between the metal
surface and adjacent tissue. The section shows well organized
fibrous tissue analogous to that sho~vnin Figures 1, 2, and 3; this may

Fig. 5. Fibrous tissue membrane adjacent to unstable cobalt, alloy implant,

x475, eosin and hematosylin, 10 months.
348 HOMSY

be compared with the biopsy adjacent to a stable plate shown in

Figure 4.
These observations suggest t h a t cellular injury either of traumatic
or biochemical origin stimulates fibroblastic response to encapsulate
the offending locus within the body. Of course, gross cytotoxicity by
moeities released by the implant would develop generalized necrosis
around the implant; however, such an aseptic necrotic volume usually
n-ould be encapsulated by a fibrous membrane (encystment).
Migration of cytotoxic moeities to body parts distant from implant
site may also occur. However, such effects should be limited t o
rapidly reacting implants where diffusional time scale is very short
in comparison to times for fibrous membrane generation (2-3 weeks).
Both local and distant cytotoxicity should be understood t o include
sublethal cellular effects which may induce bizarre changes in cellular
metabolism, e.g., cancer.

Quantification of Host/Implant Chemical Interaction

1;ollowing the discussion above, implant material compatibility
may be viewed in terms of the level and kind of migratible chemical
moeities released by the implant. It is clear t h a t degradation of
polymer t o relatively high molecular weight uriits which remain
mechanically associated within the polymer matrix would riot be
expected t o produce toxic effects. On the other hand, mechanical
properties of the implant could be significantly altered.
Low molecular weight degradation products should diffuse readily
into implant surroundings. This circumstance is analogous to release
of metallic ions consequent to corrosion. I n both instances, the
thickness of the sequestration membrane would be a gross measure of
degree of implant compatibility.
The low molecular weight migratible moeities may also have been
in a polymeric implant prior to implantation. These may be stabi-
lizing additives, plasticizers, or polymer degraded during fabrication
of a functional part.
The large number of potential materials for implantation precludes
routine use of complex and lengthy in vivo animal implantation
cvaluation. Reproducible histological assay of tissue adjacent to an
implant requires closest attention to implantation arid biopsy tech-
niques. I n addition, the movement of an implant relative t o con-
tiguous tissue is not strictly controllable; as \\e have shown, this
 BIO-COMPATIBILITY FOR IMPLANTATION 349

factor may complicate interpretation based on sequestration mem-

brane thickness.
I n order to sharply reduce the number of in vivo implantation
studies needed in general screening of implant candidates, in vitro
protocols have been developed to grade candidates by degree of
generation of low molecular weight, migratible compounds over
exposure to simulated body fluid (PECF, Table I) and by apparent
cytotoxicity of these compounds in primary mammalian tissue
~ulture.1~J’By these screening schemes only candidates of great
compatibility potential would need to be evaluated in definitive
animal implantation studies.

TABLE I
Pseudo-Extmcellular Fluid (PECF)
(NaHCO3, K,HPOr, NaCI, KCl)a

Concent,ration (meq/liter)
Ion Physiological PECF

NA+ 145 145

K+ 5 5
c1- 113 118
HCO - 30 30
HPO 9 2 2

8 Reagent grade chemica1.s.

Candidate materials in standardized form (for example, polymer

molding powders) and of the same surface area are exposed to the
115°C PECE’ for 62 hr. Test temperature of 115OC is used in order
t o simulate the effects of long-term in vivo implantation. It is
believed that elevated temperature efficiently models the influence of
long-term implantation in terms of degradation chemistry and
migration of additive chemicals from the implant material. The
protein constituents of body fluid are not simulated since these
would be expected to contribute but little to the chemical stress upon
an implant. The absence of protein buffers causes a slight elevation
from physiological pH of 7.2-7.3 to 8 . 2 ; the anionic strength of the
proteins is compensated by a slightly greater chloride ion concentra-
tion over physiologic. While the possible effects of enzyme chem-
istry are not included, the omission is not considered significant since
350 HOMSY

such effects should become apparent during animal implantation

studies.
Following the exposure period the PECF is studied in two ways:
(1) concentration of low molecular organic moeities are measured by
infrared (IR) spectrophotometric detection of primary, secondary,
and tertiary carbon-hydrogen bonds; ( 2 ) aliquot portions of PECF
from a given experiment is used to prepare the nutrient medium for
new-born mouse heart tissue culture. In these ways a direct com-
parison is obtained between level of migratible moiety and tissue
culture cytotoxicity . Evaluation of metal compatibility could
follow the same general scheme excepting that mass spectrographic
analysis would be substituted for infrared.
The ef’ficacy of these procedures in evaluation of candidate poly-
meric materials is illustrated by the data presented in Table 11.
Cytotoxic effects in primary tissue culture for different polymers
ranged from a very mild reaction to total inhibition of the culture’s
growth. I n addition to a varying degree of reduction of cell migra-
tion (as represented by the zone of outgrowth), morphological
abnormalities could be found in the cells. Arbitrarily the cells’
response was classified into five categories :
0-Excellent outgrowth, normal cell morphology ; (typical for
control cultures).
+l-Outgrowth comparable to that of the control, but some
vacuolization of the cells.
+ 2-Distinct reduction in growth usually accompanied by ex-
tensive vacuolization.
+3--Growth almost totally inhibited, and in most cases only
fibroblastic ; severe vacuolization or granular degeneration of
cells.
+4--Growth totally inhibited, although the culture is not neces-
sarily dead (as indicated in rare cases by rhythmic contrac-
tions of the explant).
The moles per million moles of methyl (-CH,), methylene
(-CH,), and methyne (-CH) structural groupings detected within
the PECF is seen to generally correlate with the degree of tissue
culture pathology.
Several materials yielded very low total organic radical response
(<SO RlPRII) and relatively mild tissue culture response (up to $ 2 ) .
 HIO-COMPATIBILITY FOR IMPLANTATION 3.5I

TABLE I1
Bio Compatibility Screening Itesriltss
Resin Exposed to PECF, 63 hr, 115"C, 30 psia

Total CH3, and

CH in P E C F
Tissue via I R analysis
culture MPMb equiva-
Polymer response lent n-hexanol
I . Chopped graphite fiber +1 n.d.':
2. Silastic (Dow Corning 372 Non-reinforced) +I 5
372 Reinforced-fabric +1 5
3. Polyethylene (U. of Texas)d (SG: 0.96) +1 17
4. Vitreous carbon frit +1 to +2 n.d.
5. Polytetrafluoroethylene (TFE) bleached fiber + 1 to +2 n.d.
6. Fluorinated ethylene propylene (FEP T-160) +1 to +2 n.d.
7. Polypheriylene oxide (Grade 731) +1 to +2 27
8. Polyethylene (SG: 0.96) +2 n.d.
9. Acrylic molding powder (V-415) +2 n.d.
10. Polyphenylene oxide (Grade 534-801) +2 17
11. Polyethylene (SG: 0.925) +2 17
12. Fluorinated ethylene propylene (FEP T-100) +2 23
13. Ionomer (1550) +2 142
14. Polypropylene (Grade 114, Food Grade) +2 198
15. Vinylidene fluoride (Grade 200) +2 to +3 3
16. Nylon (Grade 101) +2 to +3 14
17. Ionomer (AD 8043) +2 to +3 30
18. Cellulose propionate $3 81.7
19. Polystyrene (HH401) (Grade 300) +3 168
20. Nylon (Grade 38) $4 12
21. Polyvinylchloride (Grade 5430) +:3 to +4 277
22. Polyurethane (Grade 58093) +4 89
23. Polyurethane (Grade 16139) +4 328
24. Polyvinylchloride (U. of Texas) f 4 514
25. ABS (Grade X7-1000) +4 516
-
B
Scale: +I-Some vacuolization and growth inhibition but nominally as
control cultures.
+2--Moderate vacuolization, morphological changes and growth
inhibition.
+3--Severe growth inhibition and vacuolization.
3.4-Total growth inhibition.
b MPM-Moles per million.
c n.d.-Not detectable.
d University of Texas, Austin, Drug/Plastic Research Laboratory, negative
standard.
e University of Texas, Austin, Drug/Plastic Research Laboratory, intensely

toxic standard.
352 HOMSY

Chopped graphite fibers (No. 1, Table 11) and vitreous carbon frit
(No. 4) are, of course, not polymers; these were examined because
of increased interest in implant applications for these special sub-
stances. The medical grade silicone rubbers (No. 2) and perfluoro-
carbon polymers (TPE and FEP) (Nos. 5 and 6) represent polymer
families which have seen extensive and successful clinical implanta-
tion over the past 10-15 yr. I n addition, this ((benign” group
includes three polyethylene polymers (Nos. 4,8, and 11) an acrylic
molding powder (No. 9) and two polyphenylene oxide polymers
(Nos. 7 and 10).
It is important to emphasize that these results are specific to the
generic type, manufacturer, and preparation history (indicated by
manufacturer’s lot number) of a given material. The speed and low
cost of the evaluation procedures here discussed means that each
batch of material used in a commercial preparation of a n implant
may be certified. Moreover, not only raw material but also fabri-
cated forms of these materials should be certified for implant use
since fabrication conditions (especially thermal history with polymers)
can reduce the surface resistance of the implant to biochemical attack.
Polymers No. 3 and 24 were polyethylene (PE) and polyvinyl-
chloride (PVC) obtained from the University of Texas’ Drug-Plastic
Research Laboratory. These were used as external control materials
since they had been established as “negative” and “cytotoxic”
standards respectively in that laboratory’s extensive polymer toxicity
screening methodology.1s The data are in agreement with that
assessment. The negative control was identified as “high density”
polyethylene, i.e., the Ziegler catalysis product of high molecular
weight with little or no chain branching, narrow molecular weight
distribution, and pronounced hydrophobicity, all of which factors
should mediate against chemical reactivity.
The nylon resins (Nos. 16 and 20) were aberrant from the correla-
tion and were seen to cause pronounced inhibition of cell migration
in the culture dishes. This exceptional behavior and that of poly-
vinylidene fluoride (No. 15) and of Ionomer grade AD 8043 (No. 17)
are believed to reflect toxic degradation or migration moieties from
these polymers which do not contain appreciable C-H bonds. The
reIatively high infrared response for the samples of food grade
polypropylene (No. 12) and Ionomer grade 1550 (No. 13) is puzzling
and may indicate the presence of a relatively innocuous anti-oxidant
 BIO-COMPATIBILITY FOR IMPLANTATION 353

which migrated from the polymer. These aberrations show that

simple (C-H) bond infrared assay should not be exclusively relied
upon for implantation acceptance screening or monitoring. On the
other hand, the assay is believed to be efficient for rejection screening.
We re-emphasize that both the infrared and tissue culture screening
protocols should be viewed as preliminary to in vivo implantation
studies. The infrared spectrographic response is most rapidly
obtained a t modest cost. Results may be obtained for a large num-
ber of candidate polymers or series monitoring of the same polymer
(limited primarily by autoclave capacity) within 4 days. However,
the I.R. results should be used only as a basis for rejection when
response exceeds about 50 moles per million total C-H moiety.
Those polymers which exhibit a response of <50 RIP14 total
+
(C-H) groups and a 1 tissue culture reaction would be expected to
yield innocuous migration products a t very low rates and could be
safely considered for long-term implantation in humans. Only
limited animal implantations should be needed to verify this assess-
ment.
Those materials which exhibit a + 2 reaction would require more
caution since migration products from these polymers appear dis-
tinctly harmful to living cells. Although, in vivo concentrations of
these products would not accumulate to the same extent as in the
culture dish, and although several materials in the category have
been found to be well tolerated in animal and human use, any + 2
plastic should be thoroughly screened in vivo before it is used in
humans.
Those materials which score +3 or +4 should not be contemplated
for medical use, and would be inappropriate for most kinds of animal
work with implants.
Appropriately shaped specimens of fabricated polymer or metal
(e.g., micro-tensile, flexural strength, and compressive strength
specimens) could also be tested in these ways; changes in significant
mechanical properties (implant stability parameters) could thereby
be directly correlated with the bio-compatibility parameters.
Simultaneous Tissue and Blood Compatibility
The foregoing discussion has been limited to the question of im-
plant toxicity. I n many current and evolving applications, such as
cardiac bypass pumps, blood dialysis machines and artificial hearts,
354 HOMSY

bio-compatibility must include consideration of the thrombogenic

behavior of the polymer.
Studies have explored the relationship between electric charge
potential, distribution, and flow, a t natural and prosthetic surfaces in
contact with blood and the incidence of t h r o m b o s i ~ . ~A~ general
,~~
conclusion from these studies was that electronegativity of the
surface of prosthetic metals reduced the tendency for thrombosis.
The observation was consistent with the negative surface potential
of blood vessel wall with respect to blood flowing through the vessel.
Electronegativity developed on the surface of polymers through
attachment of the natural anionic anti-coagulant heparine sulfate
was seen to produce degrees of anti-thrombogeni~ity.2~-~~ Indeed,
Cross and co-workers,26were able to show that polyelectrolyte poly-
mers of a given level of polyanion excess (0.5 milli-equivalents per
dry gram) were substantially Iess thrombogenic than polymers of
greater or less electronegativity.
Earlier, Lyman et U Z . , ~ ’ showed an inverse correlation between
surface free energy and coagulation times of in vitro and in vivo dog
blood. l’olytetrafluoroethylene was somewhat aberrant from this
correlation; Lyman suggested that local high energy sites could
explain the incidence of coagulation even though the average free
energy of the material is relatively low. More recently, Lyman2*
has found that polymers synthesized to exhibit extraordinarily low
surface free energy, such as polyperfluoroacrylates, were substantially
anti-thrombogenic, and, this behavior reflected the reduced tendency
for blood platelet precipitation on these surfaces. This evidence of
the induction of thrombosis on foreign surfaces by platelet precipita-
tion scheme is offered over against that of activation of Hagemann
blood The former scheme corresponds to that believed to
obtain for thrombosis induction on autogeneous surfaces.
Understanding of thrombosis on foreign materials is incomplete.
However, tissue bio-compatibility must perforce be the first requisite
for successful blood contacting materials. Other factors pertaining
to surface physical-chemistry as those discussed above must also be
controlled to mimic the natural internal surfaces of the circulatory
system. Very active research on this problem area deserves con-
tinuing support if the hope for workable circulatory assist devices is
to become reality.
 BIO-COMPATIBILITY FOR IMPLANTATION 355

CONCLUSIONS
(1) Careful animal implantation studies can inform the severity of
host/implant interaction; however, duration, complexity and po-
tential ambiguity of such studies preclude routine screening and
monitoring of candidate implant materials by such techniques.
(2) Implant bio-compatibility is determined by the extent of
chemical interaction between host and implant as a result of which
cytotoxic moieties migrate to contiguous tissue. Implant stability
(mechanical integrity and properties) is not informed by this measure
and requires appropriate mechanical test criteria.
(3) Rapid in vitro techniques can be used to predict the quantity
and cytotoxicity of moeities released by a n implant. These tech-
niques can also be used to predict changes in mechanical properties
of a material during implantation.
(4) These techniques allow formulation of rational criteria for
rejection of candidate materials for clinical implantation and for
acceptance for definitive preclinical animal implant studies. These
procedures are applicable to routine monitoring of materials accepted
for fabrication of prototype or production implants and of completely
fabricated prostheses.

The author expresses appreciation for the assistance of M. S. Anderson, M.D.,

Department of Pathology, The Methodist Hospital, in preparation and interpre-
tation of histological sections. In addition, H. S. Tullos, M.D., and R. F. Stanley,
M.D., provided important aid in animal surgery. The i n vitro implant screening
protocols were developed in the Bio-Medical Engineering Laboratory and De-
partment of Biology of Rice University where work in this area continues to go
forward. The talented collaboration of Professor K. Ansevin of the Department
of Biology is warmly credited. I n addition, the technical assistance of Mr.
James Wykhart is appreciatively acknowledged.

REFERENCES
1. It. Carmen and P. Kahn, .I. Bzoitied. Muter. ties., 2 , 457-464 (1968)
2 J Cohen, J. Malerzals, 1 , 354 (1066).
3. J . R. Cahoori aiid H. W. Paxtori, .I. Hiomed. Mater. Res., 2, 1-22 (1968).
4. V. J . Colangelo arid N. I). (ireeiie, J . Hzomed. Mater. Res., 3,247-265 (1969).
5. D. King, L. It. Straub, and C. N. Lanibert, JBJS, 41A, 705 (1959).
6. E. El. J . Smyth, JBJS, 40B, 4 (1958).
7. I. Redler, JBJS, 44A, 8 (1962).
356 HOMSY

8. A. F. Szezpinski, The American Surgeon, 32, 372-374 (1966).

9. E . J. Kaminski, R. Oglesby, N. K. Wood, and J. Sandrik, J . Biomed. dlater.
Res., 2, 81-88 (1968).
10. L. V. Ordman, and T . Gillman, Archives Surg., 93, 857-911 (1966).
11. E. T. Oppenheimer, M. Willhite, I. Danishefsky, and A. P. Stout, Cancer
Research, 21, 132-134 (1961).
12. S. F. Hulbert, J. J. Klawitter, C. I). Talbert, arid C. T. Fitts, “Materials of
Construction for Artificial Bone Segments,” Clemson University, 1968.
13. J. Calnan, British J. Plastic Surg., 16, 1-22 (1963).
14. M. E. DeBakey, G. Jordan, B. Abbott., arid It. M. O’Neal, Archives Surg.,
89,755 (1964).
15. P. Laing, A. Fergusori, and E. S. Hodge, J . Biomed. Mater. Res., 1, 135-149
(1967).
16. C. A. Homsy, It. Hodge, N. Gordon, E. E. Braggs, and M. Estrella, J .
Biomed. Mater. Res., 3, 235-245 (1969).
17. C. A. Homsy, K. D. Ansevin, W. O’Bannon, S. A. Thompson, R. Hodge, and
M. Estrella, J . Macromol. Sri. (Chem.), A4, 615-634, (1970).
18. S.A. Rosenbluth, G. It. Weddington, W. L. Guess, and J. Aritian: J . Pharm.
S C ~ 54,
. , 156-159 (1965).
19. P. N. Sawyer, Ann. N . Y . Acad. Sci., 1 4 6 , 4 9 4 5 (1968).
20. J. It. Sadd, D. E. Koepke, L. R. Ihggett, W. C. Farnstarff, W. P. Young,
and V. L. Gott, Surg. Forum Am. Col. of Surg., 12, 252 (1961).
21. V. L. Gott, J. I). Whiffen, and It. C. Diltton, Science, 142,1297-1298 (1963).
22. V. L. Gott, J. D. Whiffen, D. E. Koepke, L. It. Ilaggett, W. C. Booke, and
W. P. Young, Trans. Amer. SOC.Artificial Int. Organs, 10, 213-217 (1964).
23. R. J. Leininger, C. I>. Falb, and G. A. Grode, Ann. N . Y . Acad. Sci., 146,
11-20 (1968).
24. E. W. Merrill, E. W. Salzman, B. J. Lipps, Jr., E. R. Gilliland, W. G. Aust,en,
and J . Joison, Trans. Amer. Soc. ArtiJicial Znt. Organs, 12, 139 (1966).
25. C. A. Homsy, M. Estrella, It. Hodge, and M. A. Serrato, Material Engineer-
ing and Science Conference, American Institute of Chemical Engineers,
Philadelphia, 1968.
26. R. A. Cross, L. N. Nelson, 13. J. Bixler, and V. L. Gott, Materials Engineer-
ing and Science Conference, American Institute of Chemical Engineers,
Philadelphia, 1968.
27. D. J. Lyman, W. M. Muir, and I. J. Lee, Am. SOC.Artificial Znt. Organs, 11.
301-306 (1965).
28. D. J. Lyman, personal commiinications, 1969.
29. C. A. Homsy, It. F. Stanley, M. S. Anderson, and J. W. King, Proceedings
Western Orthopedic Association, Colorado Springe, Colorado, September,
1969.