KIYOSHI TSUJI, E. M. STAPERT, JOHN H. ROBERTSON, AND PETER M. WAIYAKI'
Control Analytical Research and Development, The Upjohn Company, Kalamazoo, Michigan 49001
Received for publication 9 July 1970
A sensitive sterility testing procedure for the detection of microbial contamination
in petrolatum-based ointments is described. The method involves dissolving
the ointment in filter-sterilized isopropyl myristate and filtering through a membrane
filter. Improved sensitivity is obtained by blending the membrane in Trypticase
Soy Broth before incubation. Filter-sterilized isopropyl myristate is shown to
be less toxic to microorganisms than heat-sterilized isopropyl myristate. The isopropyl
myristate method is more sensitive than the polyethylene glycol-ether method
for the detection of microbial contamination.
Microbiological contamination of pharmaceutical
preparations has been of increasing concern
(5). Techniques described for the testing of nonsterile
products without filtration (1, 6, 8) are not
ideal in many instances because of the presence of
antibiotics or other inhibitory substances. Petrolatum-
based ointments present a particular problem
since they are difficult to solubilize without
adversely affecting contaminants that may be
present. Sokolski and Chidester (9) reported a
method for detection of microorganisms in petrolatum-
based ointments by using isopropyl myristate
as the solvent. The method is described in the
Code of Federal Regulations as a sterility test for
ophthalmic ointments (2). However, the method
lacks sensitivity since less than 50% of added cells
MATERIALS AND METHODS
Test organisms. The following microorganisms were
used to determine recovery and toxicity: Staphylococcus
aureus ATCC 6538P; Bacillus subtilis ATCC
6633; Escherichia coil Upjohn Culture Collection
(UC) 3114; several clinical isolates of Pseudomonas
aeruginosa UC no. 95, 104, 350, 393, 394, 399, 744,
and 3363; and Salmonella newington UC 3476.
Vegetative cells were grown in Trypticase Soy Broth
(TSB; BBL) for 24 hr at 35 C and stored in liquid
nitrogen until used. Spores of B. subtilis were obtained
Solvent. Isopropyl myristate was obtained from
three suppliers: Givauden-Delawanna, Inc., New
York, N.Y. as Deltyl Extra; Goldschmidt Chemical,
Division of Wilson Pharmaceutical, New York, N.Y.;
and M. W. Parsons-Plymouth, Division of S. B.
Penick & Co., New York, N.Y. It was sterilized
I Present address: Department of Microbiology, University
of Hawaii, Honolulu, Hawaii.
either by dry heat or by filtration. The dry-heat
sterilization was accomplished by heating 50-ml
quantities in 125-ml Erlenmeyer flasks at 180 C for
4 hr. Sterilization by filtration was accomplished by
filtering through a 0.45-,um membrane filter (Millipore
Corp., Bedford, Mass.).
Rinse medium. A 0.5% concentration of Brain
Heart Infusion broth (Difco) was prepared and
filtered through a 0.45-pm membrane filter. A 1-ml
amount of Tween 80 (Atlas Chemical Industries,
Inc., Wilmington, Del.) was added to 1,000 ml of
broth and steamed for 10 min to dissolve the Tween
80. The medium was dispensed in 200-ml quantities
and autoclaved for 15 min at 121 C.
Filtration procedure. Approximately 0.5 g of ophthalmic
ointment was transferred aseptically onto
the blades of a sterile Waring Blendor. A 50-mi
amount of warm (42 C) isopropyl myristate was
added, and the mixture was blended for 20 sec. After
blending, the isopropyl myristate-ointment mixture
was filtered through a 0.45-,pm membrane filter. As
soon as the mixture passed through the filter, usually
within 30 sec, 200 ml of warm (42 C) rinse medium
was added to the filter. Immediately after filtration
and rinsing, the membrane filter was aseptically
removed and (i) placed on Trypticase Soy Agar
(TSA) plates or placed into TSB or (ii) placed in a
sterile Waring Blendor and blended for 30 sec with
100 ml of TSB. The blended suspension was transferred
into a sterile test tube (38 by 200 mm) and
incubated for 7 days at 35 C. Sample preparation
and filtration should be conducted in a laminar-flow
hood to minimize contamination.
Preparation of experimentally contaminated oint.
ment. A sample of ointment was melted by heating to
approximately 80 C. The melted ointment was
poured into a water-jacketed semimicro Waring
Blendor cup maintained at 44 C with a Lauda K-2/R
circulating water bath. When the temperature of the
ointment reached 44 C, 0.1 ml of an aqueous suspension
of microorganisms containing 100 to 200
cells per ml was added. The inoculated ointment was
blended vigorously at 44 C for 30 sec to achieve a
uniform suspension and then cooled while gently
mixing until it congealed.
RESULTS AND DISCUSSION
Toxicity of the solvent. Since the possible presence
of microorganisms such as S. aureus, E. coli,
P. aeruginosa, and Salmonella is of concern in
ophthalmic ointments, the test method should
permit quantitative recovery of these microorganisms.
A 0.1-ml amount of a cell suspension,
containing approximately 100 organisms, was
added directly to isopropyl myristate. Test samples
were filtered at various time intervals. The
experiments were repeated at least three times.
Figure 1 demonstrates the survival of E. coli in
heat- and filter-sterilized isopropyl myristate.
Since rate of cell death was assumed to follow
first-order kinetics, the number of surviving organisms
was plotted against time on semi-log
graph paper, and a linear regression line was
0 0 o
0 ° o 0
\ 2 00o o
\e °o ° o
X, 0@ 9 FILTER STERILIZED
-A ---A HEAT STERILIZED
drawn by using the least squares method. D
values (time in minutes required to kill 90% of
the microorganisms) were calculated by using the
following formula (10): D = U/(log a - log b),
where U = time in minutes to decrease the number
of microorganisms from a to b (Ub - Ua),
a = number of microorganisms at time Ua, and
b = surviving number of microorganisms at time
Ub. Figure 2 shows the survival of P. aeruginosa
in heat- and filter-sterilized isopropyl myristate.
It is apparent from these figures that the filtersterilized
isopropyl myristate is significantly less
toxic than heat-sterilized isopropyl myristate
which was reported by Sokolski et al. (9). D
values also were obtained for other species of
organisms that were placed in heat- and filtersterilized
isopropyl myristate (Table 1).
Of the organisms tested, E. coli appeared to be
the more resistant to filter-sterilized isopropyl
myristate. However, when isopropyl myristate
was heat-sterilized, E. coli was highly sensitive to
Since P. aeruginosa appeared to be one of the
most sensitive organisms to the solvent system,
several isolates were tested to verify this observation.
Eight different clinical isolates were obtained
and exposed to filter-sterilized isopropyl
myristate for as long as 10 min. Table 2 shows the
20-,\ FILTER STERILIZED
^ ^-sHEAT- STERILIZED
10 ' ISOPROPYLMYRISTATE
¶ I Ip
FIG. 1. Survival ofEscherichia coli UC 3114 in heatand
in filter-sterilized isopropyl myristate.
FIG. 2. Survival ofPseudomonas aeruginosa UC 104
in heat- and filter-sterilized isopropyl myristate.
VOL. 20, 1970 799
TSUJI Er AL.
TABLE 1. Survival of microorganisms in heat- and
filter-sterilized isopropyl myristate
D value in isopropyl
Bacillus subtilis spores........ 45. la 100.3
Staphylococcus aureus ........ 11.0a 188.0
Escherichia coli..........4... 5b 329.0
Salmonella newington .1.....1.7b 18.0
Pseudomonas aeruginosa ...... 1.9b 19.0
a Indicates a significant difference from filtersterilized
isopropyl myristate at the 95% level.
b Indicates a significant difference from filtersterilized
isopropyl myristate at the 99% level.
TABLE 2. Survival of various isolates of Pseudomonas
aeruginosa in filter-sterilized
Isolate D value
UC 95 27.2
UC 104 19.2
UC 350 44.6
UC 393 64.2
UC 394 60.0
UC 399 24.6
UC 744 19.4
UC 3363 23.0
D values obtained for each of the isolates. Even
though there was considerable variation in the D
values for the eight isolates, these values did not
prove to be statistically different. P. aeruginosa
UC 104 was inoculated into isopropyl myristate
obtained from three suppliers. The D values in filter-
sterilized solvent were: Givaudan-Delawanna,
19.2; Goldschmidt, 10.8; M. W. Parsons-Plymouth,
13.3. These D-value differences were not significant
at the 95% level.
The following experiment was performed to
investigate the cause of the increased toxicity observed
with heat-sterilized isopropyl myristate. A
5-ml amount of water was added to 100 ml of
heat-sterilized (4 and 18 hr at 180 C) and filtersterilized
isopropyl myristate. The mixtures were
shaken vigorously for 1 hr. The pH of the water
extract was then determined, and the titrable
acidity of the water extract was determined by
neutralizing 5 ml of water extract with 0.01 N
NaOH. The results obtained were as follows.
With filter-sterilized isopropyl myristate, the pH
of the water extract was 5.0, and 0.058 ml of
0.01 N NaOH was needed for neutralization.
With heat-sterilized (4 hr) isopropyl myristate,
the pH of the water extract was 3.3, and 0.600 ml
of 0.01 N NaOH was needed for neutralization.
TABLE 3. Comparative study of microbial recovery
from an inoculated petrolatum-based ophthalmic
ointment by the filter-sterilized isopropyl
myristate method and by the polyethylene
glycol-ether method of the
Determination Vcioaubnlte mIysroipsrtopaytle egtlhyycloeln-e
Day 1 26 +fi 5 12 a 6 6 2
Day 2 25 -4i 2 13 -fi 7 6 2
Avg per cent 49 23
Day 1 53 i 6 44 2 3 zt 4
Day 2 35 i 2 26 9 3 L 1
Avg per cent 80 8
TABLE 4. Comparison of recoveries of microorganisms
by plate count, with blended and nonblended
membrane filter methods
No. of positive
of tests with
.* No. of membrane filter
Microorganism A colonies in TSB"
t; Blended blended
Escherichia colib.. 4 2.5 4/4 3/4
.... .... 6 0.5 6/6 0/6
a Abbreviations: TSA, Trypticase Soy Agar;
TSB, Trypticase Soy Broth.
b Number of cells added was 2.1/0.5 g of ointment.
¢ Number of cells added was 1.8/0.5 g of ointment.
With heat-sterilized (18 hr) isopropyl myristate,
the pH of the water extract was 2.7, and 1.220 ml
of 0.01 N NaOH was needed for neutralization.
The increase in toxicity of the heat-sterilized isopropyl
myristate may be due to the increase in
acidity. Gas chromatography and mass spectrometry
were used to attempt to identify the nature of
the acid produced. Results were inconclusive.
Microbial recovery. The efficiency of recovery
of microorganisms by the isopropyl myristate
method and the polyethylene glycol-ether method
of the Millipore Corp. (7) was compared by studying
a purposely contaminated ophthalmic ointment
(Table 3). The Millipore method recovered
23% of the inoculated E. coli and only 8% P.
aeruginosa. The isopropyl myristate method made
possible recovery of twice as many E. coli cells and
800 APPL. MICROBIOL.
10 times more P. aeruginosa than the Millipore
method. The inefficiency of the Millipore method
may be due to its inherent difficulty of transfer of
microorganisms from an oil layer to an aqueous
layer. The apparently low recovery of the isopropyl
myristate method, although better than
the Millipore method, may be that microorganisms
on a membrane filter are still encased in an
oil film and thus are denied access to nutrients
required for multiplication. If this hypothesis is
correct, then the efficiency of the isopropyl myristate
method might be improved by removing the
oil film from the microorganisms, e.g., by blending
the membrane filter in a nutrient medium.
The following experiment was designed to evaluate
such a blending procedure. Ophthalmic ointment
was purposely contaminated with approximately
four cells per g of either E. coli or P.
aeruginosa. Samples (0.5 g) of the contaminated
ointments were dissolved in isopropyl myristate
and filtered by the procedure described above. The
filter pads were (i) placed directly on TSA plates,
(ii) placed into tubes containing 100 ml of sterile
TSB, or (iii) blended with 100 ml of TSB in a
sterile Waring Blendor jar. The plates, tubes, and
blendor jars were incubated for 48 hr at 37 C, and
any resulting growth was confirmed as E. coli or
P. aeruginosa by streaking on EMB Agar (BBL)
or Pseudosel Agar (BBL), respectively (Table 4).
When the filter pads were blended, growth of E.
coli was obtained in each of the four replicates
tested, and P. aeruginosa growth was obtained
from each of the six replicates tested. In contrast
to the efficiency of the blended method, only
three positives of four replicates for E. coli and
zero positives of six replicates for P. aeruginosa
were obtained when the filter pads were not
blended but placed directly into the TSB medium.
TABLE 5. Sterility test ofpetrolatum-based
No. of Mostpositive
in probable no.
Product 6 repiicate of micoointmnent
ttbbss per g of ointment
Baciguent ................. 5 3.6
Myciguent................. 4 2.2
Neo-Cortef 0.5%........... 1 0.2
Neo-Cortef 1.5%........... 2 0.7
Neo-Cortef w/Tetracaine. . 1 0.2
Neo-Delta-Cortef.......... 1 0.2
Neo-Medrol ............... 2 0.7
Neo-Predef........... ....1 ..0.2
Neo-Predef w/Tetracaine. . 2 0.7
Neosone .0 0
a Registered trademarks of
The Upjohn Co.,
The benefit obtained from blending the filter pad
supports the hypothesis that, in the nonblended
test, microbial cells remain coated with an oil film
or cells are trapped within the oily filter membrane
structure (3), or both occur, and thus are
denied access to nutrients required for multiplication.
Thus, the filter-sterilized isopropyl myristate
method when coupled with the blending of the
membrane filter would be a suitable sterility test
for petrolatum-based ophthalmic ointments.
Sterility testing of petrolatum-based ophthalmic
ointments. Ten different ophthalmic ointments,
manufactured by The Upjohn Co., were tested for
sterility. Filter-sterilized isopropyl myristate was
used, and the membrane filters were blended in
TSB medium. Table 5 shows the results of this
study. Although several positive tests were obtained,
none of the organisms isolated was a
pathogen. Assuming even distribution of microorganisms
in ointments, the most-probable number
of contaminating microorganisms per gram of
ointment was calculated by using the following
formula (4): X = 2,3026 log n/q, where: X = the
most-probable number of microorganisms per
sample, n = total number of replicate tubes, q =
number of negative tubes. The level of contamination
in these products is very low.
A. R. Lewis and R. J. Cole are acknowledged for their statistical
analysis. Technical assistance of S. C. Edwards is also
1. Buhlmann, X. 1968. Method for microbiological testing of
nonsterile pharmacuticals. Appl. Microbiol. 16:1919-1923.
2. Code of Federal Regulations. 1968. Sterile bacitracin-neomycin
sulfate-polymyxin B sulfate ophthalmic ointment;
sterile bacitracin-neomycin sulfate-polymyxin B sulfatehydrocortisone
acetate ophthalmic ointment. Title 21,
Part 41e.433, p. 239. U.S. Government Printing Office,
3. Friedman, B., P. Blais, and P. Shaffer. 1968. Fine structure of
Millipore filters. Cell Biol. 39:208-211.
4. Halvorson, H. O., and N. R. Ziegler. 1933. Application of
statistics to problems in bacteriology. I. A means of determining
bacterial population by the dilution method. J.
5. Kallings, L. 0. 1965. Microbiological contamination of
medical preparations. Report 1965 to the National Board
of Health (Sweden).
6. Kallings, L. O., 0. Ringertz, and L. Silverslope. 1966. Microbiological
contamination of medical preparations. Acta.
Pharm. Sniecica 3:219-228.
7. Millipore Corporation. 1969. Microbiological analysis of
toiletries and cosmetic products. Application report AR-16,
p. 6. Millipore Corp., Bedford, Mass.
8. Pederson, E. A., and L. Szabo. 1968. Microbial content in
non-sterile pharmaceuticals. II. Methods. Dan. Tidsskr.
9. Sokolski, W. T., and C. G. Chidester. 1964. Improved viable
counting method for petrolatum-based ointments. J. Pharn.
10. Stumbo, C. R. 1948. Bacteriological considerations relating
to process evaluation. Food Technol. 2:115-132