Content uploaded by A.U. Daniels
Author content
All content in this area was uploaded by A.U. Daniels on Mar 19, 2018
Content may be subject to copyright.
Six Bioabsorbable Polymers:
In
Vitro
Acute Toxicity
of
Accumulated
Degradation Products
M.
S.
Taylor,*
A.
U.
Daniels,*
K.
P.
Andriano,*
and
J.
Hellert
*Division
of
Orthopedic Surgery, University
of
Utah Health Sciences Center, Salt Lake City, Utah and
+SRl
International, Menlo Park, California
Bioabsorbable polymer implants may provide
a
viable alternative to metal implants for internal
fracture fixation. One of the potential difficulties with absorbable implants is the possible tox-
icity of the polymeric degradation products especially if they accumulate and become concen-
trated. Accordingly, material evaluation must involve dose-response toxicity data as well as
mechanical properties and degradation rates.
In
this study the toxicity and rates of degradation
for six polymers were determined, along with the toxicity of their degradation product compo-
nents. The polymers studied were poly(glyco1ic acid) (PGA), two samples of poly(L-lactic acid)
(PLA) having different molecular weights, poly(ortho ester) (POE), poly(t-caprolactone)
(PCL), and poly(hydroxy butyrate valerate)
(5%
valerate) (PHBV). Polymeric specimens were
incubated at 37°C in 0.05
M
Tris buffer (pH 7.4 at 37°C) and sterile deionized water. The
solutions were not changed during the incubation intervals, providing a worst-case model of the
effects of accumulation
of
degradation products. The pH and acute toxicity of the incubation
solutions and the mass
loss
and logarithmic viscosity number of the polymer samples were
measured at
10
days, 4,
8,
12, and 16 weeks. Toxicity
was
measured using a bioluminescent
bacteria, acute toxicity assay system. The acute toxicity of pure PGA, PLA, POE, and PCL
degradation product components was also determined. Degradation products for PHBV were
not tested. PGA incubation solutions were toxic at
10
days and at all following intervals. The
lower molecular weight PLA incubation solutions were not toxic in buffer but were toxic by 4
weeks in water. The other materials did not produce toxic responses during the 16-week
exposures. The degradation products components in order from most toxic to least toxic are:
lactic acid (PLA), c-caproic acid (PCL), glycolic acid (PGA), cyclohexane dimethanol (POE),
propionic acid (POE), 1,6 hexane diol (POE), pentaerythritol dipropionate (POE), and penta-
erythritol (POE).
0
1994
John
Wiley
&
Sons,
Inc.
INTRODUCTION
Metal implants are traditionally used for internal fracture
fixation. Problems with this mode of treatment include in-
complete healing due to stress shielding and the need for
surgical removal of the implant.'
A
potential solution is
implants made from bioabsorbable materials that are less
stiff and should not require surgical removal. An addi-
tional advantage
is
the gradual transfer
of
load from the
implant to bone as the implant loses stiffness, allowing
gradual bone remodeling that decreases the likelihood of
re fracture.
However, there are a number of concerns related to the
use of absorbable implants. One of the major concerns is
the potential for adverse reactions due to accumulated
degradation products. Poly(a-hydroxy acids) are the most
Requests
for
reprints should
be
sent to M.
S.
Taylor, Division
of
Orthopedic
Surgery,
University
of
Utah Health Sciences Center,
50
N.
Medical
Dr.,
Salt Lake
City,
UT
841
32.
Journal
of
Applied Biomaterials,
Vol.
5,
15
1
-
I57
(
1994)
0
1994
John Wiley
&
Sons,
Inc.
CCC
1045-486
1/94/020
I5
1-07
commonly used absorbable polymers for medical applica-
tions, and, much of the literature focuses
on
these materi-
als. Animal studies by Spenelhauer et al.' suggest a high
rate of noninfectious inflammatory response associated
with intermediate molecular weight degradation products
of poly(a-hydroxy acids). A high rate of sterile inflamma-
tory response occurred when poly(L-lactic acid) (PLA)
weight average molecular weight fell below
20,000
D. Use
of a higher initial molecular weight
PLA
delayed but did
not eliminate the response. Additional work by Schaken-
raad and Dijkstra3 demonstrated a sharp increase in solu-
bility for
PLA
and poly(glyco1ic acid) (PGA) as number
average molecular weight drops to about
5,000
to
10,000
Da. The molecular weight determinations were made us-
ing gel permeation chromatography (GPC) analysis, and
solubility by
in vivo
and
in vitro
weight loss. The apparent
difference in these results could be due to the polydisper-
sity of the material. Pistner et aL4 observed a difference in
degradation rate in a comparison of poly(L-lactides) of
three molecular weights. Two of the samples were amor-
phous and the third semicrystalline. The amorphous poly-
mers had a slower initial drop in inherent viscosity than
the semicrystalline polymer, equalizing at about week
20.
Additionally, the amorphous samples had a delayed
151
152
TAYLOR ET
AL.
weight loss. These results indicate the relative importance
of initial molecular weight and morphology. Another
study showed that PLAs exhibit accelerated degradation at
the center of bulk implants. The degraded polymer would
have limited diffusion out initially, possibly causing a large
delayed release of degradation products when the implant
exterior is suddenly breached. Also, both amorphous and
crystalline PLA develop increased crystallinity as degrada-
tion progre~ses.~ This indicates the potential for a change
in degradation rate due to morphology even as the implant
degrades. Other canine experiments showed bone resorp-
tion against PLA surfaces once weight loss from the im-
plant commences.6 When considered together, these re-
sults indicate the need for further evaluation prior to use
of poly(a-hydroxy acids) as bulk implants.
These concerns are amplified by the clinical results ob-
served in Europe with rods of PGA and PGA/PLA copol-
ymers or plates and screws of PLA for fracture fi~ation.~ Jn
this review by Bostman frequency of sterile inflammatory
response ranged from 5.7 to 47% of clinical cases. The
highest response rate was observed for a high molecular
weight (HMW) PLA and occurred 2 to
3
years postopera-
tively. The inflammation may be due in part to a high lo-
calized concentration of the degradation products. Inter-
estingly, the highest response rate for PLA occurred at the
ankle, the site ofthe lowest response rate for PGA. Clearly,
there are contributing factors that are not well understood
at this time, yet PGA and PLA are routinely used in
smaller volumes for suture materials and elicit much
smaller inflammatory responses in such cases. This sug-
gests a dose-response relationship for toxicity.
Little information is available comparing the toxicity
of degradation products as they accumulate or as a func-
tion of their concentration. One of the difficulties in com-
paring the toxicity of materials is the need to perform an
extensive testing battery to completely evaluate any mate-
rial.* The most common
in
vitro cytotoxicity tests rely on
cultures of animal or human cells with exposure intervals
ranging from a few minutes (Trypan Blue exclusion assay)
to 72 h. Discrepancies have been reported for different
time exposures, cell lines, and different methods
of
evalu-
ation. The difficulty in assessing toxicity is one of the indi-
cations for a battery of tests, as one may identify a poten-
tial problem with a material that might not be apparent
from the results of another test. This study relies upon a
single test, the bacterial bioluminescence test (BBT), for
comparison of toxicity. Accordingly, the present study
provides
a
basic comparison
of
the selected materials, but
does not provide definitive answers regarding the relative
toxicity.
The present study was undertaken to compare the deg-
radation and acute toxicity of accumulated degradation
products of six absorbable polymers under uniform
exposure conditions. A second phase of the study mea-
sured the toxicity of known quantities of pure degradation
products for five absorbable polymers.
MATERIALS
AND
METHODS
Bioabsorbable Polymers and
Controls
The polymers studied were PGA, two samples of PLA
having different molecular weights, poly(ortho ester)
(POE), poly( t-caprolactone) (PCL), and poly(hydroxy bu-
tyrate valerate)
(5%
valerate) (PHBV). All of these poly-
mers undergo hydrolysis. Table I lists the source and deg-
radation products for each material. Two sources of PLA
were selected to reflect its broad availability for medical
applications. The
POE
used in this study contained a mix-
ture of rigid
trans-cyclohexanedimethanol
and flexible
1,6
hexane diols present in a
60:40
mole ratio. The initial hy-
drolysis yields pentaerythritol mono- or dipropionate, the
two diols, and propionic acid. The pentaerythritol mono-
or dipropionate is further hydrolyzed to yield pentaeryth-
ritol and propionic acid.' Two potential advantages are
that the degradation is pH sensitive and can be controlled
and that the initial degradation products are pH neutral.
The PHBV used in this study contained 5 mol
%
hydroxy-
valerate to improve ductility. PHBV degrades primarily
by microbial attack, but was included in this study to eval-
uate its degradation in a sterile environment."
A low molecular weight polyethylene (PE-LMW) ob-
tained from Abiomed was used as negative control mate-
rial. The positive control material was the same PE with
40
wt%
trans-cinnamic acid added before hot molding the
specimens. Selection of the positive and negative control
materials was based on Burton et al.
I I
PHBV was provided in a powdered form.
All
other
polymer samples were milled to -40 mesh then freeze
dried for 48 h to minimize moisture and/or static charge
for ease of processing.
For the positive control, trans-cinnamic acid was finely
ground using a mortar and pestle and was slowly stirred
into milled PE. Mixing continued for
15
min at
less
than
100
rpm.
Milled polymers were hot compression molded. The
molding temperatures are given in Table 11. Amorphous
polymers (60:40 POE) were pressed at
50
"C above the
glass transition temperature, pressed at
5000
psi for
5
min,
and cooled to ambient temperature. Semicrystalline poly-
mers (PLA, PGA, PHBV, PCL, PE) were pressed
5"
below
their respective melting temperatures and pressed at 5000
psi for 5 min prior to cooling to ambient temperature. The
resulting samples were
1
-in. diameter disks weighing
1
g.
Thickness varied slightly with density of the polymers.
The disks were quartered using a jeweler's saw and an alu-
minum jig and stored in a vacuum desiccator jar with a
moisture absorber after molding until sterilization.
Each sample was weighed, numbered, and placed in a
vial. The samples in opened vials were sterilized by Nelson
Laboratories using ethylene oxide (ETO). The steriliza-
tion protocol was
1.5
h humidification at
38
"C
and 45%
relative humidity followed by
2.5
h exposure to
12%
ETO
carried in Freon@, and 48 h of degassing at
55
"C.
For PCL,
BIOABSORBABLE POLYMERS:
IN
VITRO
TOXICITY
153
TABLE
1.
Bioabsorbable Polvmers
Material Source, Trade
Name
Degradation Products
PGA DuPont, Medisorb lOOPGA Glycolic acid
PLA-LMW DuPont,
Medisorb
IOOL
Lactic acid
PLA-HM
W
Birmingham
Polymers, BP-0600 PLA Lactic acid
POE
SRI
International, no trade
name
Intermediate stage:
Pentaerythritol dipropionate, trans-cyclo hexane
dimethanol, 1,6 hexane diol
Pentaerythritol,
propionic
acid
Second
stage:
PCL Birmingham Polymers, BP-0800 PCL t-Hydroxy caproic acid
PHBV
ICI
America, no trade name Butyric acid, valeric
acid
the degassing temperature was lowered to
37
"C
and the
time extended to
72
h. This change was necessary due to
the low melting point of PCL.
Once sterilized, samples of each material were tested
for ETO residuals and sterility. All samples tested sterile
and had residual levels well below maximum acceptable
limits. The highest level of ETO was
<10.2
ppm, and the
most stringent FDA guideline for implants is
<25
ppm.
Incubation
Ten milliliters of sterile solution was added to each vial.
The two exposure media were sterile distilled water and
sterile Tris buffer. The Tris buffer solution was selected
because it has little or no effect on the response of the tox-
icity test method described subsequently under the condi-
tions used: pH
7.4
at
37
"C
and
0.05
A4.I'
The samples
were incubated at
37
"C. The removal dates were
10
days,
4,
8,
12,
and
16
weeks. At each removal,
pH
and toxicity
of the incubation solution and mass
loss
and logarithmic
viscosity number of the polymer specimens were mea-
sured. There were three samples per solution per time in-
terval for a total of
30
samples of each polymer for
exposure testing. In addition, there were three positive
control samples and three negative control samples for
each exposure medium as well as two samples for ETO
residuals testing and one sample for measurement of vis-
cosity after processing. The positive and negative controls
were tested at the first time interval only.
TABLE
II.
Hot
Pressing Temperature for Bioabsorbable Polymers
Material Press Temp ("C)
PGA
PLA-LM
W
PLA-HMW
POE
PCL
PHBV
215
165
170
110
50
160
Toxicity
Testing
Toxicity was measured using a commercially available
BBT assay system.
l3
The test procedure involved reconsti-
tuting lyophilized
Photobacterium phosphoreum
that emit
light indicative of their level of metabolic activity. The
bacteria were placed in
2%
saline in five test vials main-
tained at
15
"C.
Light output from the bacteria was mea-
sured in a photometer located adjacent to the refrigerated
cells. After serial dilution, test samples were added to four
of the vials, with the fifth serving as a negative control.
The light levels are then measured at
5
and
15
min after
addition of the test sample. The negative control sample
allowed corrections due to changes in luminescence with
time and volume. The toxicity data presented were given
by percentage light remaining,
(NOIN=,)
X
(
T5/TO),
where
the symbols Nand
T
refer to the light levels for the nega-
tive control and test sample vials, respectively; the sub-
script
0
refers to measurements taken immediately prior
to addition of the sample; and the subscript
5
refers to
measurements taken
5
min after addition of the sample.
A remaining light reading of
50%
or less is considered
toxic, and the concentration necessary to cause a 50%
reading is known as the ECso. For known dilution concen-
trations, the result is an equation relating concentration
and toxicity.
A
study by Burton et
a1.l'
compared this technique to
five standard acute toxicity tests for multiple samples of
eight materials. The tests were minimum Eagle's medium
elution tissue culture (MEM), mouse safety
(MS),
rabbit
intramuscular implantation (RII), rabbit intracutaneous
injection (RI), and mouse systemic injection (MSI). The
MEM was performed on confluent monolayers of mouse
fibroblast cells incubated for
24
to
48
h. The cells were
fixed, stained, and scored for cytotoxicity. As shown in
Table 111, of the
10
1
samples tested with MEM and BBT,
the two tests agreed for
95
samples
(87
nontoxic and
8
toxic) and disagreed
on
six samples
(4
BBT toxic,
2
MEM
toxic). This was the best correlation observed. The
MS
test
was performed on
93
samples of those, none were toxic
with MS and
20
were toxic with BBT. Twenty-six samples
were tested to compare BBT to
in vivo
toxicity screens
(RI,
154
TAYLOR
ET
AL.
TABLE
111.
Bacterial Bioluminescence Test (BBT) Compares Favorably with Other Acute Toxicity Testsll
RII,
MSI,
RI
MS
MEM
Comparison to BBT
(26
Samples Tested)
(93
Samples Tested)
(
10
I
Samples Tested)
Agree
Disagree
~
16
S
10
E
9P
10
S
(BBT toxic)
16
E
(BBT toxic)
17
P
(BBT toxic)
73
20
(BBT toxic)
87
8
(both
toxic)
4
(BBT toxic)
2
(MEM
toxic)
~
S,
E,
and Prefer to extracts
of
saline, ethanol/saline,
or
polyethylene glycol
400,
respectively. Numbers indicate test results that agree or disagreed with
BBT.
Except
as
noted,
test
results were nontoxic.
MSI, and/or
RII).
These showed the greatest dis-
agreement. Autian describes these tests as generally show-
ing a low frequency of response, but worth performing as
part of toxicity screen.l4 The numbers listed in Table
111
indicate the number of samples for which the response of
the tests agreed and disagreed. Tests with toxic results are
listed in parenthesis.
Another study'* determined the response of the BBT to
pH
of
water, buffer pH, and buffer concentration at phys-
iological pH. The pH of sterile water was shifted to basic
or acidic by addition of NaOH or HC1, respectively. Over
the pH range between 4 and
10,
no suppression
of
bacte-
rial light output was observed. The test of buffer pH and
concentration showed that sensitivity might be reduced by
use of a citrate phosphate borate buffer that mildly sup-
pressed the bacterial light output at all pH and concentra-
tions, but that Tris buffer was
a
well-suited medium for an
incubation solution for the studies described here.
Polymer Mass
Loss
The polymer specimen mass
loss
was determined by
weighing the samples after drying in a vacuum oven at
ambient temperature (approximately
23
"C) for at least 48
h. The samples were considered dry when the weight taken
on successive days was unchanged.
Polymer Viscosity Measurement
Reduced viscosity
of
the polymer specimens in solvent
so-
lutions was measured using a Cannon-Ubbelhode vis-
cometer according to ASTM D2857.15 Logarithmic vis-
cosity number was measured for the polymers as received,
as pressed, and at each time exposure. Measurement of the
polymers as received and as pressed identified any change
in molecular weight due to processing. Hexafluoro-2-pro-
panol was the solvent used for PGA, and chloroform was
the solvent used for all other polymers. For POE, triethyl
amine was added to the solvent to stabilize the solution.
The concentration used for all samples was approximately
5
mg/mL. The formula used for calculation of viscosity
was
77
=
[ln(t,/t,)]/c,
where
ta
is the average of three flow
times for the polymer in solution,
t,
is the average of three
flow times for the pure solvent, and
c
is the concentration
in grams/milliliter. To hasten dissolution,
PHBV
samples
were milled to --SO mesh after incubation and prior to vis-
cosity testing.
Toxicity
of
Degradation Product Components
Pentaerythritol dipropionate was synthesized at
SRI
by es-
terifying pentaerythritol with propionic acid. The lactic
acid, glycolic acid, caprolactone, pentaerythritol, propi-
onic acid, cyclohexane dimethanol, and 1,6 hexane diol
were obtained from Aldrich Chemical. Using the BBT,
EC50 values for each of the degradation products were de-
termined. As described previously, POE degrades
in
stages. For acute toxicity evaluation, the intermediate
stage was considered to consist of the two diols and penta-
erythritol dipropionate, and the final stage to consist of the
two diols, pentaerythritol and propionic acid.
ECS0s
were
determined for both combinations. Under actual
in
vivo
conditions,
POE
degradation products are more likely to
consist of a combination of all possible degradation prod-
uct components.
RESULTS
Table
IV
presents the toxicity and pH of the incubation
solutions and weight loss and viscosity for the incubated
polymer specimens at the 16-week time interval and the
interval at which materials first elicited a toxic response.
Only PGA and the PLA-HMW in water incubation solu-
tions elicited toxic responses during the exposure period.
Each material
is
discussed separately below.
Controls
The incubation solutions for the negative and positive
controls were tested for pH and toxicity at
10
days. Weight
loss and viscosity were not determined. The negative con-
trols were all nontoxic, with an average light output for
the strongest concentration tested
of
99%.
The average pH
levels were 7.7 for Tns exposures and 6.8 for the water
exposures. The positive controls all caused a complete
155
BIOABSORBABLE POLYMERS:
IN
VITRO
TOXICITY
TABLE
IV.
Data
for
Final Exposures and First
Toxic
Exposures
Polymer Final Incubation Solution
16-Week Data Mass Loss Viscosity Remaining
Material and Exp. Media
(%I
(% of
original) PH Light
(%)
0.0
t
0.00
PGA
(water)
64.85
t
0.63 18.7
t
2.4 2.49
f
0.06
0.0
k
0.00
PLA-LMW (water)
4.69
-t
1.36 47.0
k
1.7 3.03
f
0.04
6.80
f
0.02
82.1
t
2.88
PLA-LMW
(Tris)
3.62
k
0.09 48.7
k
2.6
PLA-HW (water)
-0.05
k
0.01
19.5
-t
3.4 1.30
k
0.06 96.2
t
6.32
POE (water)
1.29
t
0.12 95.4
t
1.6 6.56
f
0.05 57.2
t
7.95
POE
(Tris)
1.58
f
1.61 89.3
f
8.3 1.75
k
0.05 68.4
f
6.69
PCL (water)
0.49
t
0.04 80.9
k
7.4 6.41
k
0.21 98.3
t
4.62
PCL
(Tris)
0.69
f
0.09 86.8
+.
2.5 7.68
?
0.00
100.5
f
5.58
PHBV
(water)
0.03
f
0.03 84.5
f
1.1
6.40
f
0.14 94.8
f
6.63
PHBV
(Tris)
0.02
k
0.02
80.4
t
0.8 7.75
t
0.03 101.5 k2.51
First Toxic Data
PGA
(Tris)
63.38
t
0.66 25.9
?
2.8 2.85
f
0.04
0.0
k
0.00
PLA-HW
(Tris)
-0.10
t
0.05 77.0
k
4.6 7.72
k
0.02 88.4
f
6.24
PGA (water),
10
day
PGA
(Tris),
10
day
PLA-LMW (water),
4
week
3.25
?
0.3
3.46
t
0.4
0.47
?
0.57
35.4
f
6.7 2.81
t
0.05
0.00
t
0.00
29.3
f
3.3 5.05
t
1.25
0.00
f
0.00
87.9
t
3.6 3.57
?
0.13 7.16t 10.13
Average
values
k
SD.
suppression of the light output, with average pH of
5.5
for
the Tris exposure media and 3.8 for the water.
PGA
For PGA, the pH dropped in both exposure media quite
rapidly to less than
5.
For the 10-day buffered exposure
media, the average pH was
5.05,
which would not cause
the toxic response observed. The incubation solution for
PGA was toxic at all time intervals for both water and
buffer exposures. All tested dilutions of the incubation
so-
lution were toxic. The concentration of the most dilute of
the test solutions was 5.6% of the concentration of the ini-
tial solution, indicating that the exposure solution was
well beyond the threshold of toxicity. The measured
ECsO
for glycolic acid was
100
X
M.
The average mass loss
for the first toxic exposures was
9
mg
(3.5%).
The corre-
sponding calculated concentration would be between
1200
X
Mif degraded to glycolic acid and 180
X
1
0-6
M
if degraded to a molecular weight of
5000.
Either con-
centration would produce
a
toxic response. However, the
most dilute sample (a concentration of 5.6% ofthe original
sample) had no light output suggesting that the latter con-
centration was
a
low estimate. The viscosity decreased
with time indicating a continued drop in average molecu-
lar weight of the polymer specimen.
PLA-LMW
Results for the PLA-LMW were substantially different for
the water exposures than for the buffer exposures.
Water Exposures.
The pH dropped rapidly, to an aver-
age of 3.6 by the 4-week exposure period, and remained
relatively constant afterward. The incubation solution was
toxic at 4 weeks and at all subsequent time intervals. The
average mass loss associated with the first toxic results was
1.2 mg
(0.5%).
The calculated concentration was between
1300
X
M
if degraded to a molecular weight of
5000.
The measured
ECsO
for lactic acid is
100
X
M.
The polymer viscosity
decreased nearly linearly with time to a value of about
50%
of the original value, indicating a large drop in average
molecular weight.
Mif degraded to lactic acid and
24
X
Tris
Exposures.
The pH dropped only slightly, to an
average of 6.8 by the 16-week exposures. The incubation
solution was not toxic at any time interval. The average
mass loss by the sixteenth week was 8.8 mg (3.6%). The
calculated concentration was
9700
X
M
if degraded
to lactic acid or
170
X
1
OP6
M
if degraded to a molecular
weight
of
5000.
The
ECso
for lactic acid was
100
X
1
0-6
M
indicating that the actual concentration in the incubation
solution lies closer to the latter value than to the former.
The polymer viscosity drop was similar in buffered
exposures to that seen for water exposures.
PLA-HMW
The pH for the PLA-HMW remained relatively constant
throughout the test intervals for both exposure media. The
bacterial light output was not suppressed for any of the
tests. PLA-HMW showed a slight weight loss at the initial
156
TAYLOR
ET
AL.
time interval and no weight loss in the subsequent in-
tervals. The actual data for the final exposure indicated a
slight gain in mass
(<O.
1%).
The polymer viscosity drop
was about 25% in both media, indicating that degradation
occurred, although solvation did not.
POE
The pH of the POE incubation solutions remained rela-
tively constant throughout the test in both exposure me-
dia. None of the incubation solutions were toxic, although
there was some light suppression at the later time in-
tervals. Average weight losses for the 16-week samples of
POE were 3.2 mg (1.3%) in water and
4.0
mg (1.6%) in
buffer. The resulting concentration was between 1300
x
1
0-6
Mand
2600
X
M.
The former value represents
degradation to the two diols and pentaerythritol dipropio-
nate, and the latter value represents degradation to the two
diols, pentaerythritol and propionic acid.
PCL
The pH of the PCL incubation solutions remained rela-
tively constant throughout the test in both exposure me-
dia. The average mass losses for the 16-week PCL samples
were 1.2 mg
(0.5%)
in water and 1.7 mg
(0.7%)
in buffer.
The corresponding concentrations of degradation prod-
ucts were
530
X
M
in
buffer. The measured ECSO for caprolactone was
450
X
M.
The calculated concentration was close to the
measured ECSo, yet the bacterial light output was not sup-
pressed. This suggests that the actual concentration was
lower than calculated. One possible explanation is that in-
termediate molecular weight polymer chains become
sol-
uble in a manner similar to that seen with PLA and PGA.
The polymer viscosity for PCL decreased by
25%
during
the exposure period, indicating a drop in average molecu-
lar weight.
M
in water and
750
X
PHBV
The pH of the PHBV incubation solutions remained rela-
tively constant throughout the test in both exposure me-
dia. None of the incubation solutions were toxic. PHBV
showed virtually
no
mass
loss
(<0.05%),
but polymer vis-
cosity did drop, indicating degradation occurred in a ster-
ile environment. One can conclude that microbial attack
is not the only method of degradation for PHBV.
Degradation Products
The EC5o values of the degradation products are shown in
Table
V.
Two values are shown for POE. The two values
represent the difference in toxicity between the first and
final stage products. PHBV is not included because its deg-
radation products were not tested. The tendency for poly-
mers to become soluble at intermediate molecular weights
TABLE
V.
Relative Toxicities
of
Degradation Products
Toxic
Concentration
PGA 150 87
PLA
100
72
POE (intermediate stage)
1000
336
POE (final) 270
91
PCL 120 137
makes it difficult to predict the mass loss that correlates
with toxicity. The value presented assumes degradation to
single repeat units.
DISCUSSION
This study provides insight into the relative acute toxici-
ties of bioabsorbable polymer degradation products. In or-
der from most toxic to least toxic they are: lactic acid
(PLA), c-caproic acid (PCL), glycolic acid (PGA), cyclo-
hexane dimethanol (POE), propionic acid (POE), 1,6 hex-
ane diol (POE), pentaerythritol dipropionate (POE), and
pentaerythritol
(POE).
The difference in concentration
that may be tolerated differs by a factor of 10. The toxicity
is a factor that should be considered in material selection,
but must be combined with degradation rate and local tis-
sue clearance to predict the concentration present in the
tissue, and the resulting response.
in vivo
degradation rates
are affected by not only the polymer crystallinity and mo-
lecular weight but also by the implant size, shape, and im-
plantation site. The order of degradation could be consid-
ered by either mass loss
or
viscosity decrease. Because vis-
cosity decrease represents a loss of mechanical properties,
that is, modulus, it is presented here. The order of viscosity
decrease, from most affected to least affected, is: PGA,
PLA-LMW, PLA-HMW, PHBV, PCL, and POE. Note
that the data presented in Table
4
indicates that PLA-
LMW and PLA-HMW exhibited markedly different be-
haviors, thus demonstrating the role of molecular weight
in bioabsorbables. Similarly, two otherwise identical ma-
terials with different morphologies may exhibit different
degradation rates, but this aspect was not considered in
this study.
The materials tested as degradation product compo-
nents were well controlled uniform solutions. The ex-
pected degradation product
in vivo
probably contain not
only a mixture of molecular weights, but also might exist
in different states of ionization. Both factors could alter
the biologic response.
The results observed are also affected by the particular
processing parameters. For the hot compression molding,
difficulty was encountered with PGA. In the samples used,
granularity was evident. Use
of
a higher processing tem-
perature was attempted without success to reduce the
BIOABSORBABLE
POLYMERS:
IN
VITRO
TOXICITY
157
granularity.
No
doubt
the
granularity
affected
the
rate
at
which
fluid
penetrated the
sample,
thus
affecting
the
deg-
radation
rate.
We
would
like to thank Osteotech
for
support
for
this project and SRI Interna-
tional
for
donation
of
POE.
REFERENCES
1.
Daniels, A.
U.;
Chang, M.
K.
0.;
Andriano,
K.
P.; Heller, J.
Mechanical properties of biodegradable polymers and com-
posites proposed
for
internal fixation of bone. J. Appl. Bio-
mater. 157-78; 1990.
2. Speneihauer, G.; Vert, M.; Benoit,
J.
P.; Boddaert, A.
In
vi-
tro
and
in
vivo
degradation of POIY(D,L lactide/glycolide)
type microspheres made by solvent evaporation method.
Biomaterials
10:
5
57-563; 19 89.
3. Schakenraad, J. M.; Dijkstra, P. J. Biocompatibility of
pOly(D,L-laCtiC acidiglycolide) copolymers. Clin. Mater. 7:
4. Pistner, H.; Bendix, D.; Muhling,
J.;
Reuther, J. Poly(L-lac-
tide): a long-term degradation study
in vivo.
Part
111.
Analyt-
ical characterization. Biomaterials 14:29 1-298; 1993.
5.
Li,
S.;
Garreay, H.; Vert, M. Structure-property relation-
ships in the case of the degradation of massive poly(cY-hy-
droxy acids) in aqueous media. J. Mater. Sci. Mater. Med.
1
:
6. Suganuma, J.; Alexander,
H.
Biological response of intra-
253-269; 1991.
198-206; 1990.
medullary bone to poly-L-lactic acid. J. Biomed. Mater. Res.
(in press).
7. Bostman,
0.
M. Current concepts review: absorbable im-
plants for the fixation of fractures.
J.
Bone Joint Surg.
8. Balls, M.; Clothier, R.
H.
Cytotoxicity assays for intrinsic
toxicity and imtancy. In: Watson, R. R., Ed.
In
vitro
meth-
ods
of
toxicology. Boca Raton,
FL
CRC Press; 1992:37-53.
9. Heller, J. Controlled drug release from poly(ortho esters)-
A
surface eroding polymer. J. Controlled Release
2:
167-
177; 1985.
10.
PHBV biodegradable polyesters. Wilmington, DE: ICI
Americas; 1988.
1
1.
Burton,
S.
A.;
Peterson,
R.
V.;
Dickman,
S.
N.;
Nelson, J. R.
Comparison of
in vitro
bacterial bioluminescence and tissue
culture bioassays and
in
vivo
tests for evaluating acute toxic-
ity. J. Biomed. Mater. Res. 20:827-838; 1986.
12. Daniels, A.
U.;
Andriano,
K.
P.; Felix, B.; Hamber, E. The
effect
of
pH and buffer solutions on the bacterial biolumi-
nescence test for acute toxicity. Trans. SFB
1
10; 199
1.
1
3. Microtox manual. Carlsbad, CA: Microbics Corporation;
1991.
14. Autian, J. Toxicological evaluation of biomaterials: primary
acute toxicity screening program. Artif. Organs 153-60,
1977.
15.
Annual book of standards. Philadelphia, PA: American
So-
ciety for Testing and Materials; 1987.
73A( 1): 148-1 53; 199
I.
Received March
15,
1993
Accepted February
10,
1994
