FOOD BIOLOGICAL CONTAMINANTS

Iron Milk Medium Method for Recovering Clostridium

perfringens from Shellfish: Collaborative Study
ABEYTA & WETHERINGTON: JOURNAL OF AOAC INTERNATIONAL VOL. 77, NO. 2, 1994
CARLOS ABEYTA, JR, and JUNE H. WETHERINGTON
U.S. Food and Drug Administration, Seafood Product Research Center, 22201 23rd Dr, SE, Bothell, WA 98041
Collaborators: J.S. Cholensky; J.H. Cutting; P.A. Dabney; J.P. Dudley; R. Rivera; M.J. Roetting; S. Simpson; F.A. Stanley; R.O.
Torres; G.W. Varney; W.D. Watkins; J.H. Wetherington; M.R. Zipkes

Eleven laboratories participated in a collaborative

study analyzing shellfish (oysters, Crassostrea gigas) for the detection and enumeration of Clostridium perfringens by the iron milk medium (IMM)
method. The IMM method was compared to AOAC
Official Method 976.30. Shellfish were artificially inoculated with C. perfringens cells (vegetative and
spores) at low (1 103 colony forming units
[cfu]/g), medium (1 104 cfu/g), and high (1
106 cfu/g) levels. Negative controls (zero level)
were analyzed by each laboratory. C. perfringens
FD-1, the strain involved in a foodborne illness,
was used. Blind duplicates of each inoculum level
were analyzed, giving a total of 16 samples per
laboratory. The selectivity of IMM relies solely on
the rapid growth of C. perfringens at 45C indicated
by stormy fermentation reaction within 18 h. C. perfringens is detected and enumerated using the
most probable number technique. A statistical
evaluation of the data found no significant differences between the estimates from the 2 methods.
The IMM method for detection of C. perfringens
from shellfish has been adopted first action by
AOAC INTERNATIONAL.

he iron milk medium (IMM) method (1, 2) was developed to circumvent problems caused by using plate media that contain inhibitory agents and to avoid laborious,
costly, and time-consuming preparations of plates (36).
AOAC Official Method 976.30 for enumeration of C. perfringens is a plating method using tryptosesulfitecycloserine
(TSC) medium (6) and requires selection of a representative
number of colonies for isolation followed by biochemical test-

Submitted for publication December 17, 1992.

The recommendation was approved by the Committee on Microbiology
and Extraneous Materials. The method was adopted by the Official
Methods Board of the Association. See AOAC International Official
Methods Board News (1993) J. AOAC Int. 76, 33A, and Methods
Adopted First Action (1993) The Referee, 17, March issue.

ing. The IMM method is a convenient, rapid procedure for

identification of C. perfringens, and is especially useful for
field work (1).
The selectivity of the medium is based solely on the rapid
growth of C. perfringens at 45C resulting in the typical stormy
fermentation reaction within 18 h (1, 2). Stormy fermentation
in IMM at 45C, defined as the production of an acid curd with
subsequent disruption of the curd by a large volume of gas, has
long been used as one of the highly indicative confirmation
tests for C. perfringens. The acid clot is produced by the acidification of milk (lactic acid fermentation) to pH 4.5 (7).
The use of IMM for enumeration of C. perfringens has been
well documented (1, 2, 812). In recent studies of C. perfringens inoculated into shellfish and lean meats there were no significant differences in recovery abilities between IMM and
plating media such as TSC and Shahidi-Ferguson-perfringens
(SFP) (5). Similar results were found with naturally contaminated cuttlefish (9).
The IMM is more selective for differentiation of C. perfringens from closely related species (C. paraperfringens, C.
barati, C. abonsum, and C. perenne) than TSC. These similar
species would not give false positive results for C. perfringens
in suspect foods by using IMM incubated at 45C for 18 h.
They may, however, give false positives on other media such as
TSC and SFP because of their similarity in colony appearance
and biochemical reactions (11).
Collaborative Study

Reagents and Media

Reagents and solid media were prepared according to the
AOAC Official Method 976.30 (13). The solid media were prepared according to label instructions (Difco) and used only on
the day of preparation.
The IMM was prepared by placing 10 mL homogenized pasteurized milk and 0.2 g iron powder in 16 150 mm straight-wall
tubes, and then was sterilized for 10 min at 116C and 10 lb
pressure. The medium was used within 2 h of preparation.

Method
In AOAC Official Method 976.30 (13) a 50 g sample is
transferred to a sterile blender container. A total of 450 mL peptone water is added and homogenized 2 min at low speed (13
000 rpm). Serial dilutions are prepared and then plated on TSC
agar using either pouring or spreading techniques. Typical
colonies are picked and Gram stained. Sporeforming bacilli
that reduce nitrates to nitrites, produce acid and gas from lactose, and liquefy gelatin are provisionally identified as C. perfringens. Suspect colonies not meeting the above criteria must
be confirmed by determination of acid and gas from salicin and
raffinose. Salicin is usually not fermented (a few strains produce acid), and acid is usually produced from raffinose.
As an alternative to the AOAC procedure, a rapid screening
procedure based on the stormy fermentation of IMM was
tested. The fermentation is characterized by the rapid coagulation of milk followed by the fracturing of the curd into a spongy
mass which usually rises above the medium surface. The IMM
method is based on the most probable number (MPN) technique using iron milk as the enrichment medium. Tubes exhibiting stormy fermentation are streaked to a modified Clostridium perfringens medium (mCP) for confirmation.
Other related C. perfringens-like species (C. paraperfringens, C. barati, C. abonsum, and C. perenne) grow on mCP at
45C and have the ability to cleave phosphate from phenolphthalein diphosphate, but fail to produce rapid stormy fermentation in IMM within 1618 h. Cultures exhibiting a positive
stormy fermentation reaction and a positive phosphatase test on
mCP agar are considered C. perfringens (Figure 1).
Study Design

Statistical Analysis
The quality of the method depends on its repeatability and
reproducibility (14). The repeatability of the method refers to
how well any laboratory can get the same answer or nearly the
same answer upon replication. The reproducibility refers to
how well different laboratories can match their respective estimates to one another.
The relative standard deviations for both methods and difference between the concentration estimates were calculated.
Bacterial counts were converted to logarithms.
993.10 Clostridium perfringens from
ShellfishIron Milk Medium Method
First Action 1993
(Applicable to determination of 103106 Clostridium perfringens cells/g shellfish)
(Caution: Ammonium hydroxide may be fatal if inhaled or
ingested; vapors are extremely irritating. Avoid contact with
eyes, skin, or clothing. Use only with adequate ventilation,
preferably in fume hood.)
Method Performance
Oysters, 103 C. perfringens cells/g shellfish
Log average, 3.76
sr = 0.18; sR = 0.46; RSDr = 4.9%; RSDR = 12.2%
Oysters, 104 C. perfringens cells/g shellfish
Log average, 4.61
sr = 0.30; sR = 0.33; RSDr = 6.6%; RSDR = 7.2%
Oysters, 106 C. perfringens cells/g shellfish
Log average, 5.59
sr = 0.29; sR = 0.31; RSDr = 5.2%; RSDR = 5.5%

A. Principle
Eleven laboratories analyzed Pacific oysters (Crassostrea
gigas) for the presence and enumeration of C. perfringens using the IMM method and the AOAC Official Method 976.30
(13). Oysters were sterilized for 15 min at 121C and artificially inoculated with C. perfringens strain FD-1 (spores and
vegetative cells) at low (1 103 colony forming units [cfu]/g),
medium (1 104 cfu/g), and high (1 106 cfu/g) levels. Negative controls (zero level) were included in analyses. Blind duplicates of each inoculum level per product were analyzed giving a total of 16 samples per laboratory.

Preparation of Inocula and Test Samples

C. perfringens strain FD-1 was grown in trypticase peptone
glucose yeast extract (Difco) broth for 24 h at 37C. Inocula
levels were confirmed by plating them on tryptic soy agar
(Difco). A known number of C. perfringens cells was added
aseptically to oysters (50 g samples) and homogenized for 1.5
min at low speed (13 000 rpm). Samples were delivered to the
collaborators by overnight mail with instructions to add
450 mL 0.1% peptone. On the day of the test the samples were
homogenized again for 2 min at 13 000 rpm.

Serial dilutions of homogenized shellfish samples are added

to tubes of iron milk medium (IMM), which are then incubated
1618 h at 45. Presence of C. perfringens causes stormy fermentation, formation of turbid, yellow fluid with production
of considerable gas resulting in curdled milk rising to top of
tube.
Acid phosphatase test uses solid, presumptive medium
which incorporates D-cycloserine and polymyxin B sulfate
with in situ tests for sucrose fermentation and acid phosphatase
production to confirm C. perfringens.
The MPN technique with IMM as enrichment medium is
used for enumeration of C. perfringens in samples.

B. Apparatus
(a) Blender.With low speed, ca 13 000 rpm.
(b) Water bath.Capable of maintaining 45 0.5.
(c) Filters.0.45 m pore size, sterile.
(d) Applicator stick.Sterile, disposable.
(e) Anaerobic jar.Equipped with N2 and CO2 generator
envelopes and catalyst.
(f) Incubator.Capable of maintaining 45 0.5.
(g) Petri plates.50 12 mm, disposable, sterile.

C. Media and Reagents

(a) Milk.Homogenized, pasteurized whole milk. Do not
use after pull date stamp on container.
(b) Iron powder.Reduced powder.
(c) Iron milk medium (IMM).Prepare fresh (on day of
use). Place ca 0.2 g iron powder, (b), in 16 150 mm straightwall tubes. Pipet 11 mL milk, (a), into tubes. Sterilize 10 min
at 10 lb pressure, 116, then place into 45 water bath. Use
within 2 h.
(d) Phenolphthalein diphosphate solution.0.5%. Dissolve 0.5 g phenolphthalein diphosphate in 100 mL H2O. Filter-sterilize through 0.45 m filter. Store up to 1 week at 4 in
dark brown bottle, protected from light.
(e) Iron chloride solution.4.5%. Dissolve 4.5 g
FeCl36H2O in 100 mL H2O. Filter-sterilize through 0.45 m
filter. Store at 4.
(f) Modified C. perfringens medium (mCP).100 mL contains 3.0 g tryptose, 2.0 g yeast extract, 0.5 g sucrose, 0.1 g Lcysteine, 0.01 g MgSO47H2O, 0.004 g bromcresol purple, and
1.5 g agar, in H2O. Sterilize 15 min at 15 lb pressure, 121.
Cool to 50 and add 40 mg D-cycloserine; 2.5 mg polymyxin B
sulfate; 2 mL 0.5% phenolphthalein diphosphate solution, (d);
and 0.2 mL 4.5% FeCl36H2O solution, (e). Pour 5 mL portions into 50 12 mm plates. Store at 4 protected from light;
mCP must be used within 1 week.
(g) Peptone diluent.0.1% peptone in H2O. Sterilize
15 min at 121.
(h) Ammonium hydroxide solution.30% w/v in H2O.

D. Sample Preparation
Weigh ca 50 g shellfish sample into sterile blender jar. Add
450 mL peptone diluent, C(g), to prepare 1:10 dilution sample.
Blend 90 s at low speed (ca 13 000 rpm) to minimize aeration
of homogenate.
Use 1:10 homogenate to make serial dilutions from 101 to
6
10 by transferring 10 mL each successive dilution to dilution
bottle containing 90 mL peptone diluent. Shake dilution bottle
vigorously 25 in ca 30 cm arc within 7 s before each transfer.

E. IMM Determination
Transfer 1 mL each dilution from D, in triplicate, to IMM
tubes, C(c). Incubate tubes 1618 h in 45 water bath.
Examine tubes. Tubes positive for stormy fermentation will
show medium digested to turbid, yellow fluid with great quantities of gas produced. Spongy mass of curdled milk usually
rises above medium surface.

F. Confirmation by Acid Phosphatase Test

With sterile forceps, aseptically place 0.45 m filter on
mCP. Inoculate mCP plate containing filter using sterile applicator stick to streak single line for each tube, from left to right,
no more than 3 streaks (1 dilution) per plate. (Note: If 5 tube
MPN technique is used, streak 2 plates per dilution.)
Incubate plates in anaerobic jar 24 h at 45 in air incubator.
Check plates for acid phosphatase reaction by removing
cover and exposing agar surface (with filter) 30 s to fumes from

30% ammonium hydroxide solution, C(h). Dark pink color on

colonies indicates positive result for acid phosphatase test, confirming presence of C. perfringens. (Note: C. perfringens
grows heavily on mCP at 45; other related species such as C.
paraperfringens, C. barati, C. abonsum, and C. perenne grow
only sparsely.)

G. MPN Determination
See 966.24 for 3 tube MPN table.
Use only tubes positive for both stormy fermentation, E, and
acid phosphatase test, F, for MPN determination.
Ref.: JAOAC 77, 351 (1994).
Results and Discussion
All collaborators submitted data for the statistical evaluation. Bacterial counts were converted to logarithms (Table 1).
Blind duplicates of each inoculum level per sample were tested.
Initially, before any statistical treatment of the data, values from
Collaborators 4 and 8 were rejected due to temperature abuse
and improper handling of samples during shipment. In these
laboratories C. perfringens used to inoculate samples grew beyond the experimental design levels. In most cases values from
9 laboratories were used for the statistical evaluation. Part of
the data from Laboratory 10 (Table 1) was excluded from the
statistical evaluation (M [IMM], Table 2) because of overgrowth of C. perfringens, caused either naturally or by
overseeding the samples. No end points (growth in all serial
dilutions in all tubes) were achieved using the IMM MPN procedure identified greater than log 6.0. Summary of results before and after outlier removal is listed in Table 2 (15).
Repeatability relative standard deviation, RSDr, and reproducibility relative standard deviation, RSDR, between the 2
methods indicated satisfactory precision. For the 3 inoculum
levels, both the within-laboratory relative standard deviation
and the among-laboratory relative standard deviation are consistently lower for the AOAC Official Method 976.30 compared with the IMM method. However, the tighter results for
the AOAC method may be due to familiarity with the method.
Relative standard deviations (the crucial parameter) for either
method are satisfactory. The IMM method showed more variability than the AOAC Official Method 976.30, but the differences between the methods were insignificant. Results in this
collaborative study showed no significant differences between
the methods for Students t-test at each inoculum level at the
5% confidence level. Inherently, the MPN techniques are less
precise than standard anaerobic total plate count methods
(ATPC), because MPN is an estimation of bacterial density, and
ATPC is an actual cfu count. Previous studies have shown comparable values of estimate of variances and percent coefficient
of variation for MPN methods (10, 12, 16). The advantages of
the IMM MPN method over the AOAC Official Method
976.30 is the rapid evaluation of C. perfringens in shellfish and
the cost-saving factor. Currently the AOAC procedure requires
the preparation of plates with an agar overlay, anaerobic incubation, isolation and biochemical identification of isolates. All

these time-consuming steps can limit the number of samples

processed.
Tests performed in our laboratory as well as in others (1, 2,
8, 9, 12) showed that stormy fermentation using IMM is highly
indicative of the presence of C. perfringens. All of our biochemically confirmed C. perfringens strains isolated from seafoods, water, and sediment have exhibited stormy fermentation. Biochemical confirmation for C. perfringens in the
AOAC Official Method 976.30 requires isolates to be tested for
motility, nitrate reduction, gelatin liquefaction, and lactose fermentation. In the IMM method C. perfringens is confirmed using acid phosphatase test, first developed for rapid confirmation (17). Later, Bisson and Cabelli (18) modified the acid
phosphatase test medium and developed a membrane filter test
for detection of C. perfringens from aquatic environmental
samples. They developed a solid, presumptive Clostridium perfringens medium which incorporates D-cycloserine and polymyxin B sulfate together with in situ tests for sucrose fermentation, acid phosphatase production, and -D-glucosidase
activity. Appropriate volumes of water samples are passed
through membrane filters, which are then placed on agar plates
and incubated at 45C for 24 h. After incubation plates are
scored for the presence of C. perfringens and then exposed to
ammonia vapors for acid phosphatase detection. This medium
was further modified for the collaborative study by not including the indoxyl--D-glucoside (mCP medium). Other related C.
perfringens-like species (C. paraperfringens, C. barati, C.
abonsum, and C. perenne) grow on mCP, but the significant
difference is that C. perfringens grows heavily on mCP and the
related species grow sparsely.
Materials for the IMM method are inexpensive and readily
available, and decrease the need for inhibitors and biochemical
confirmation testing; therefore the IMM method is useful for
rapid evaluation of shellfish samples in field laboratories.
Recommendation
On the basis of this study, it is recommended that the iron
milk medium method for detection of C. perfringens in shellfish be adopted first action.
Acknowledgments
The authors would like to thank the following:
Robert O. Torres, FDA, Los Angeles, CA
Joseph P. Dudley and Paul A. Dabney, FDA, Atlanta, GA
Jeffrey H. Cutting, FDA, Denver, CO
William D. Watkins, FDA, North Kingstown, RI
George W. Varney, FDA, Winchester, MA

James S. Cholensky, FDA, Minneapolis, MN

Steven Simpson, FDA, Brooklyn, NY
Rafael Rivera, FDA, San Juan, Puerto Rico
Mary J. Roetting and Marilyn R. Zipkes, FDA, Cincinnati, OH
Fred A. Stanley and June H. Wetherington, FDA, Bothell, WA
The authors would also like to thank Wallace Andrews, the
General Referee for Non-Dairy Food Microbiology, U.S. Food
and Drug Administration, Washington, DC, for helpful advice
throughout this study, and Mary H. Krane for technical assistance.
References
(1) St. John, W.D., Matches, J.R., & Wekell, M.M. (1982) J. Assoc. Off. Anal. Chem. 65, 11291133
(2) St. John, W.D., & Matches, J.R. (1982) J. Assoc. Off. Anal.
Chem. 65, 15141516
(3) Angelotti, R., Hall, H.E., Foter, M.J., & Lewis, K.H. (1962)
Appl. Microbiol. 10, 193199
(4) Marshall, R.S., Steenbergen, J.F., & McClung, L.S. (1965)
Appl. Microbiol. 13, 559563
(5) Shahidi, S.A., & Ferguson, A.R. (1971) Appl. Microbiol. 22,
688692
(6) Harmon, S.M., Kautter, D.A., & Peeler, J.T. (1971) Appl. Microbiol. 22, 688692
(7) MacFaddin, J.F. (1980) Biochemical Tests for Identification
of Medical Bacteria, 2nd Ed., Waverly Press, Inc., Baltimore,
MD, pp. 194200
(8) Wekell, M.M., Hartman, W.J., & Dong, F.M. (1980) J. Nutr.
110, 21032108
(9) Abeyta, C. (1983) J. Assoc. Off. Anal. Chem. 66, 11751177
(10) Abeyta, C., Wekell, M.M., & Peeler, J.T. (1985) J. Food Sci.
50, 17321735
(11) Abeyta, C., Michalovskis, A., & Wekell, M.M. (1985) J.
Food Prot. 48, 130134
(12) Abeyta, C., Wekell, M.M., Kaysner, C.A., Stott, R.F.,
Raghubeer, E.V., Matches, J.R., & Peeler, J.T. (1988) Water
Sci. Technol. 20, 6367
(13) Official Methods of Analysis (1990) 15th Ed., AOAC, Arlington, VA, sec. 974.30
(14) Youden, W.J., & Steiner, E.J. (1975) Statistical Manual of the
Association of Official Analytical Chemists, 5th Ed., AOAC,
Arlington, VA
(15) Horwitz, W. (1988) J. Assoc. Off. Anal. Chem. 71, 168171
(16) Entis, P. (1984) J. Assoc. Off. Anal. Chem. 67, 812823
(17) Ueno, K., Fujii, H., Marui, T., Takahashi, J., Sugitano, T.,
Ushijima, T., & Suzuki, S. (1970) Jpn. J. Microbiol. 14, 171173
(18) Bisson, J.W., & Cabelli, V.J. (1979) Appl. Environ. Microbiol. 37, 5566

Table 1. Enumeration of Clostridium perfringens, from oysters using the tryptose sulfite cycloserine (TSC) method
(AOAC Official Method 976.30) and the iron milk method (IMM)a
Inoculation Levels
Low

Medium

Ac

Bc

High
B

Zero

Temp.,
Lab.
Cb

TSC IMM

TSC IMM

TSC IMM

TSC IMM

TSC IMM

TSC IMM

1
2
3
4
5
6
7
8
9
10
11

3.8 3.2
4.0 4.0
3.9 3.6
6.0 4.9
3.9 4.0
3.6 3.6
4.1 3.6
8.8 >6.0g
3.9 4.0
4.7 5.0
3.9 3.6

3.7
3.9
3.9
6.5
3.8
2.9
3.9
8.9
3.9
4.4
3.7

4.8
5.0
5.0
6.4
4.8
4.6
4.9
8.7
4.9
7.3
4.6

4.8
4.9
4.9
7.1
4.8
4.5
5.0
8.9
4.9
6.4
4.7

5.7
5.9
5.9
8.2
5.9
5.5
5.9
8.8
5.9
6.6
5.7

5.7
6.0
5.9
f
5.9
5.4
6.1
8.9
5.8
8.3
5.6

a
b
c
d
e
f
g

5.4
5.6
16.0
28.0
15.0
17.0
10.0
28.5
20.0
21.0
11.0

3.1
3.6
3.4
6.0
4.0
3.6
3.4
>6.0
4.0
4.4
3.6

4.0
4.4
4.6
5.4
5.0
5.0
4.2
>6.0
4.8
>6.0
4.3

TSC expressed in CFU Log10/g; IMM expressed in MPN Log10/g.

Temperature of sample received in the laboratory.
Blind duplicates.
<2.0, negative controls.
<0.5, negative controls.
>2.0, <3.0.
No end point.

4.6
4.9
5.0
6.0
4.6
4.6
4.4
>6.0
5.2
>6.0
4.4

5.4
5.4
5.3
6.0
6.0
5.4
5.7
>6.0
5.7
6.0
5.4

6.0
6.0
5.4
2.4
5.7
5.4
5.7
>6.0
5.7
>6.0
5.2

TSC

IMM

3.0

2.4

TSC IMMe

5.4

5.0

Table 2. Method performance of the iron milk method (IMM) and the AOAC Official Method 976.30 (TSC) for detection
and enumeration of Clostridium perfringens in oysters
Before outlier removal
Codea
L (IMM)
L (TSC)
M (IMM)
M (TSC)
H (IMM)
H (TSC)
a
b
c

After outlier removal

Av.b

RSDR, %

RSDr, %

Labs

Av.

RSDR, %

RSDr, %

3.7611
3.8833
4.6125
4.8188
5.5944
5.8889

4.9
5.0
6.6
1.2
5.2
1.2

12.2
9.3
7.2
3.2
5.5
4.6

9
9
8
9
9
9

3.7611
3.8786
4.6125
4.8188
5.5944
5.8889

4.9
2.3
6.6
1.2
5.2
1.2

12.2
2.7
7.2
3.2
5.5
4.6

L = low inoculum level (1 103 cfu/g); M = medium inoculum level (1 104 cfu/g); H = high inoculum level (1 106 cfu/g).
Bacterial counts converted to logarithms.
Indicates removal of outlying lab(s).

Labs
9
7
8
8
9
9

a c

Figure 1. Analytical scheme for the determination of Clostridium perfringens using the Iron Milk Method (IMM) and
the AOAC Official Method 976.30 (TSC).