FOOD COMPOSITION AND ADDITIVES

Determination of Total, Saturated, and Monounsaturated Fats In

Foodstuffs by Hydrolytic Extraction and Gas Chromatographic
Quantitation: Collaborative Study
HOUSE: JOURNAL OF AOAC INTERNATIONAL VOL. 80, NO. 3, 1997
STEPHEN D. HOUSE 1
General Mills, Inc., 54 S. Michigan Ave, Buffalo, NY 14203
Collaborators: M. Lawson; T. Hammill; R. Mazal; K. Meyer; B. Balch; J. Ngeh-Ngwainbi; P. Oles; S. Bailey; R. Bakowski; T.
Phillipo; M. Phist; J. Polywacz; S. Hill; L. Menke; B. Wise; S. Powell; R. Johnson; D. Martin

Using gas chromatography (GC), 10 collaborating

laboratories measured total, saturated, and
monounsaturated fats in 8 blind duplicate pairs of
foodstuffs. The method involves a hydrolysis/ether
extraction of fat followed by quantitative GC analysis versus an internal standard. Calculations were
designed to comply with federal regulations as
specified in the Nutrition Labeling and Education
Act of 1990. The range of fat contents was about 1
50%
%. Collaborators received and analyzed (in triplicate) a pre-collaborative sample of known fat content as a practice sample. After satisfactory results
were obtained, participants received the 16-sample
set. The repeatability standard deviations (RSDr)
for total fat ranged from 2.04 to 10.6%
%; the reproducibility standard deviations (RSDR) for total fat
ranged from 3.74 to 15.8%
%. The hydrolytic extractionGC method for determination of fat (total, saturated, and monounsaturated) in foodstuffs has
been adopted first action by AOAC INTERNATIONAL.

he Nutrition and Labeling Education Act of 1990

(NLEA) stipulates requirements for labeling of food
products and includes strict definitions of descriptors
such as low fat and light, as well as chemical definitions
for some food components, including fat (1). The current definition of fat was adopted by the U.S. Food and Drug Administration (FDA) and the U.S. Department of Agriculture
(USDA), which have enforcement authority of the NLEA, after
a specific recommendation from the AOAC INTERNATIONAL Task Force on Methods for Nutrition Labeling Analy-

Submitted for publication March 8, 1996.

The recommendation was approved by the Methods Committee on
Food Nutrition and was adopted by the Official Methods Board of the
Association. See Official Methods Board Actions (1996) J. AOAC Int. 79,
76A, and Official Methods Board Actions (1996) The Referee, June issue.
1
Current address: General Mills, Inc., QRO Dept., 54 S. Michigan
Ave, Buffalo, NY 14203.

ses (2). One problem with fat analysis is the wide variety of
available methods (36), each with advantages and deficiencies
with regard to quantitation of fat. In general, all ether-extractable material that is unbound from the food matrix is
gravimetrically determined to be fat. Often, exhaustive extraction using hydrolysis followed by ether phase partition yields
high results because extraneous ether-soluble materials also are
measured. In some cases, simple refluxing in ether is insufficient
to extract all fat from the matrix, resulting in low recoveries.
The task force recommended that FDA adopts a chemical
definition of fat, to dissuade use of methods that might yield a
biased result. Therefore, the comments prefacing the regulations state: the agency has decided to define total fat as total
lipid fatty acids, that is, the sum of fatty acids from mono-, di-,
and triglycerides, free fatty acids, phospholipid fatty acids, and
sterol fatty acids, and the agency has decided to require
that the declaration of total fat be expressed as the amount of
triglyceride that would provide the analytically measured
amount of total lipid fatty acids in the food (1). These statements had several consequences on the state-of-the-art analysis
of fat.
The definition of total fat specifically requires measurement
of fatty acids that originate from all lipidic sources, but it also
limits what can be measured and reported as fat. Ether-soluble
materials that are not fatty acids are not to be included in the
total fat content. The further stipulation that total fat be expressed as the equivalent triglyceride amount of analytically
measured fatty acids means each fatty acid measured must be
converted to its triglyceride equivalent. Then these triglyceride
equivalents are summed to give total fat. The chemical and
mathematical nuances of this requirement are described in the
paper by House et al. (7). Furthermore, non-fatty-acid material
that might be extracted into the organic phase cannot be
counted as fat. This material includes glycerol, low molecular
weight carbohydrates and their polymerization products...
amino acids, and urea salts (2). It was clear that conventional
gravimetric analysis was not sufficient to meet the new NLEA
requirements.

In addition, the agencies adopted definitions for saturated

and monounsaturated fat. The declaration of saturated fat
should include only those fatty acids that contain no double
bonds; while that of monounsaturated fat should include only
those fatty acids that contain one cis double bond. Omission of
trans fatty acids from the monounsaturated declaration was ostensibly due to the saturated character of trans fatty acids in
human metabolism. Finally, these fat subclass declarations
have to be made on a fatty acid weight basis, not an equivalent
triglyceride basis.
House et al. (7) described a method appropriate for determining NLEA-defined fat. It used a standard Mojonnier-type
hydrolysis, which provides a complete breakdown of the food
matrix, followed by ether extraction, transesterification of fatty
acids to their methyl esters (FAMEs), and gas chromatographic
(GC) quantitation. Trinonadecanoin was used as an internal
standard, mixed with the weighed sample prior to hydrolysis.
It is thus hydrolyzed and transesterified along with the fat from
the sample matrix and participates in all phases of the determination. A commercial reference mixture of FAMEs was used to
provide retention time checks and quantitative response factors. Individual fatty acids were measured, and each fatty acid
was converted to its triglyceride equivalent prior to summation
of triglycerides to give total fat. Saturated and monounsaturated
fatty acids were also measured as subclasses of fat from the GC
traces; however, calculation of these for labeling purposes was
performed on a fatty acid weight basis rather than as
triglycerides, as stipulated in the regulations (1). The work presented in this paper served as a framework from which the current collaborative study was formulated.
We have conducted this collaborative study to validate a
proposed method as an AOAC Official Method for the determination of total fat in foodstuffs, to comply with the specific
regulations of the NLEA.
Collaborative Study
This method used validated acid/alkaline hydrolysis and
ether extraction methods to extract fat from foodstuffs. These
residues were then reacted with BF3 to transesterify all fatty
acids to their respective FAMEs. An internal standard, which
was added when the samples were initially weighed, allowed
quantitative GC measurement of individual fatty acids. Triundecanoin was used as the internal standard.
The study consisted of 16 samples (Table 1), that is, 8 sets
of blind duplicates randomly numbered, with 10 laboratories
participating. Fat contents ranged from 150%. Samples represented a wide range of food types, including samples from various areas of the AOAC food triangle (8). Each laboratory
representative was asked to follow the method as closely as
possible, noting any unavoidable or accidental changes, and to
provide raw chromatograms as well as filled data tables, the
blanks for which were provided with the protocol. The method
used was based on the method described above, although several improvements and precautions were added to the method
protocol. The most significant difference between the literature
reference (7) and the collaborative protocol was the use of tri-

undecanoin as internal standard to minimize coelution of the

internal standard with other FAME peaks, particularly for samples containing hydrogenated fat, which has numerous trans
and positional isomers that elute over a rather broad time interval.
996.06 Fat (Total, Saturated, and Monounsaturated)
in Foods, Hydrolytic ExtractionGas
Chromatographic Method
First Action 1996
(Applicable to determination of fat in foods.)
Caution: See Appendix: Laboratory Safety for Safe Handling of Special Chemical Hazardsethanol, ethyl ether, and
petroleum ether; Safe Handling of Alkaliessodium hydroxide; and Safe Handling of Acidshydrochloric acid.
Boron trifluoride may be fatal if inhaled. See Material Safety
Data Sheets, or equivalent, for each reagent. Reagents may be
either flammable or caustic or may form explosive peroxides.
Work in hood with gloves and safety glasses. Dispose of waste
solvents in an appropriate manner compatible with applicable
environmental rules and regulations.
Method Performance:
See Tables 996.06AC for method performance data.

A. Principle
Fat and fatty acids are extracted from food samples by hydrolytic methods (acidic hydrolysis for most samples, alkaline
hydrolysis for dairy products, and combination for cheese samples). Pyrogallic acid is added to minimize oxidative degradation of fatty acids during analysis. A triglyceride, triundecanoin
(C11:0), is added as internal standard. Fat is extracted into ether
and then methylated to fatty acid methyl esters (FAMEs) using
BF3 in methanol. FAMEs are quantitatively measured by capillary gas chromatography (GC) against C11:0 internal standard.
Total fat is calculated as the sum of individual fatty acids expressed as triglyceride equivalents. Saturated and monounsaturated fats are calculated as the sum of respective fatty acids.
Monounsaturated fat includes only the cis form.

B. Apparatus
(a) Gas chromatograph.Equipped with hydrogen flame
ionization detector, capillary column, split mode injector, oven
temperature programming sufficient to implement a holdramp-hold sequence. Operating conditions: temperature (C):
injector, 225; detector, 285; initial tempperature, 100 (hold
4 min); ramp, 3C/min; final temperature, 240; hold 15 min;
carrier gas, helium; flow rate, 0.75 mL/min; linear velocity,
18 cm/s; split ratio, 200:1.
(b) Capillary column.60 m 0.25 mm id, 0.20 m film,
fused silica capillary, SP-2340 phase. (Note: FAMEs of C18:3
and C20:1 may coelute after several hundred injections.)
(c) Analytical balance.Capable of weighing to
0.0001 g.
(d) Mojonnier flasks.
(e) StoppersSynthetic rubber or cork.
(f) Mojonnier centrifuge basket.

(g) Hengar microboiling granules.

(h) Baskets.Aluminum and plastic.
(i) Shaker water bath.Capable of maintaining 7080C.
(j) Steam bath.Capable of supporting common glassware.
(k) Water bath.Equipped with nitrogen stream supply.
Capable of maintaining 40 5C.
(l) Wrist action shaker.Designed for Mojonnier centrifuge baskets.
(m) Mojonnier motor driven centrifuge.Optional; capable of maintaining 600 rpm.
(n) Gravity convection oven.Capable of maintaining
100 2C.
(o) Vortex mixer.
(p) Gas dispersion tubes.25 mm, porosity A, extra
coarse, 175 m.
(q) Three dram vials.About 11 mL.
(r) Phenolic closed-top caps.With polyvinyl liner,
15425 size.
(s) Teflon/silicone septa.15425 size.

C. Reagents
(a) Pyrogallic acid.Analytical reagent grade.
(b) Hydrochloric acid (HCl).12N (concentrated) and
8.3N. To make 8.3N HCl, add 250 mL 12N HCl to 110 mL
H2O. Mix well. Store at room temperature.
(c) Ammonium hydroxide (NH4OH).58% (w/w), analytical reagent grade.
(d) Diethyl ether.Purity appropriate for fat extraction.
(e) Petroleum ether.Anhydrous; analytical reagent grade.
(f) Ethanol.95%; analytical reagent grade.
(g) Toluene.Nanograde.
(h) Chloroform (CHCl3).Analytical reagent grade.
(i) Sodium sulfate(Na2SO4).Anhydrous.
(j) Boron trifluoride (BF3) reagent.7% BF3 (w/w) in
methanol, made from commercially available 14% BF3 solution. Prepare in the hood.
(k) Diethyl etherpetroleum ether mixture.1 + 1, v/v.
(l) Triglyceride internal standard solution.C11:0 - triundecanoin; 5.00 mg/mL in CHCl3. Accurately weigh 2.50 g
C11:0 - triundecanoin into 500 mL volumetric flask. Add ca
400 mL CHCl3 and mix until C11:0 is dissolved. Dilute to volume with CHCl3. Invert flask at least 10 additional times.
Triglyceride internal standard solution is stable up to 1 month
when stored in refrigerator.
(m) Fatty acid methyl esters (FAMEs) standard solutions.(1) Mixed FAMEs standard solution.Reference
mixture containing series of FAMEs, including C18:1 cis and
trans (available as GLC-85 from Nu Chek Prep, Elysian, MN,
or equivalent). To prepare mixed FAMEs standard solution,
break top of glass vial, open, and carefully transfer contents to
3 dram glass vial. Wash original vial with hexane to ensure
complete transfer and add washes to 3 dram glass vial. Dilute
to ca 3 mL with hexane. (2) C11:0 FAME standard solution.
C11:0 - undecanoic methyl ester in hexane. Use only in preparation of individual FAME standard solutions, (3). To prepare
C11:0 FAME standard solution, break top of glass vial open and
carefully transfer contents to 50 mL volumetric flask. Wash

original vial with hexane to ensure complete transfer and add

washes to 50 mL volumetric flask. Dilute to volume with hexane. C11:0 FAME standard solution is stable up to 1 week when
stored at 0C. (3) Individual FAME standard solutions.
Standard solutions of each of following FAMEs: C4:0 - tetranoic
methyl ester, C6:0 - hexanoic methyl ester, C8:0 - octanoic
methyl ester, C10:0 - decanoic methyl ester, C12:0 - dodecanoic
methyl ester, C13:0 - tridecanoic methyl ester, C14:0 - tetradecanoic methyl ester, C14:1 - 9-tetradecenoic methyl ester, C15:0 pentadecanoic methyl ester, C15:1 - 10-pentadecenoic methyl
ester, C16:0 - hexadecanoic methyl ester, C16:1 - 9-hexadecenoic
methyl ester, C17:0 - heptadecanoic methyl ester, C17:1 - 10-heptadecenoic methyl ester, C18:0 - octadecanoic methyl ester, C18:1
- 9-octadecenoic methyl ester, C18:2 - 9,12-octadecadienoic
methyl ester, C18:3 - 9,12,15-octadecatrienoic methyl ester,
C20:0 - eicosanoic methyl ester, C20:1 - 8-eicosenoic methyl ester, C20:2 - 11,14-eicosadienoic methyl ester, C20:3 - 11,14,17-eicosatrienoic methyl ester, and C22:0 - docosanoic methyl ester.
Prepare individual FAME standard solutions as follows: Break
top of glass vial open and carefully transfer contents to 3 dram
glass vial. Wash original vial with hexane to ensure complete
transfer and add washes to 3 dram glass vial. Add 1.0 mL C11:0
FAME standard solution, (2). Dilute to total volume of ca
3.0 mL with hexane. Individual FAME standard solutions are
stable up to 1 week when stored in refrigerator.

D. Extraction of Fat
Finely grind and homogenize test samples prior to extraction of fat.
(Note: When running matrixes of unknown composition, it
may be necessary to analyze sample without addition of internal standard to ensure against interferences. Should interfering
peak be found, the area of C11:0 internal standard peak must be
corrected before performing calculations. Instead, use 2.0 mL
chloroform.)
(a) Foodstuffs excluding dairy products and cheese.Accurately weigh ground and homogenized sample (containing ca
100200 mg fat) into labeled Mojonnier flask. Force sample
into flask as far as possible. Add ca 100 mg pyrogallic acid,
C(a), and 2.00 mL triglyceride internal standard solution, C(l).
Add a few boiling granules to flask. Add 2.0 mL ethanol and
mix well until entire sample is in solution. Add 10.0 mL 8.3N
HCl and mix well. Place flask into basket in shaking water bath
at 7080C and set at moderate agitation speed. Maintain
40 min. Mix contents of flask on Vortex mixer every 10 min to
incorporate into solution particulates adhering to sides of flask.
After digestion, remove flask from bath and allow to cool to
room temperature. Add enough ethanol to fill bottom reservoir
of flask and mix gently.
(b) Dairy products.Accurately weigh ground and homogenized sample (containing ca 100200 mg fat) into labeled
Mojonnier flask. Force sample into flask as far as possible. Add
ca 100 mg pyrogallic acid, C(a), and 2.00 mL triglyceride internal standard solution, C(l). Add a few boiling granules to
flask. Add 2.0 mL ethanol and mix well until entire sample is
in solution. Add 4.0 mL H2O and mix well. Add 2.0 mL 58%
NH4OH, C(c), and mix well. Place flask into basket in shaking

water bath at 7080C set at moderate agitation speed. Maintain 10 min. Mix contents of flask on Vortex mixer every 5 min
to incorporate into solution particulates adhering to sides of
flask. After digestion, remove flask from water bath and add a
few drops of phenolphthalein. Keep solution basic (pink) with
addition of ammonium hydroxide. Add enough ethanol to fill
bottom reservoir of flask and mix gently.
(c) Cheese.Accurately weigh ground and homogenized
sample (containing ca 100200 mg fat) into labeled Mojonnier
flask. Force sample into flask as far as possible. Add ca 100 mg
pyrogallic acid, C(a), and 2.00 mL triglyceride internal standard solution, C(l). Add a few boiling granules to flask. Add
2.0 mL ethanol and mix well until entire sample is in solution.
Add 4.0 mL H2O and mix well. Add 2.0 mL 58% NH4OH,
C(c), and mix well. Place flask into basket in shaking water
bath at 7080C set at moderate agitation speed. Maintain
20 min. Mix contents of flask on Vortex mixer every 10 min to
incorporate into solution particulates adhering to sides of flask.
Add 10.0 mL 12N HCl and place flask into boiling steam bath
and maintain 20 min. Mix flask contents every 10 min on Vortex mixer. Remove flask from steam bath and allow to cool to
room temperature. Add enough ethanol to flask to fill bottom
reservoir and mix gently.
Add 25 mL diethyl ether to Mojonnier flask from (a), (b),
or (c). Stopper flask and place in centrifuge basket. Place basket
in wrist action shaker, securing flask in shaker with rubber tubing. Shake flask 5 min. Rinse stopper into flask with diethyl
etherpetroleum ether mixture, C(k). Add 25 mL petroleum
ether, stopper flask, and shake 5 min. Centrifuge flask (in basket) 5 min at 600 rpm. (Note: If centrifuge is not available, allow contents to set at least 1 h until upper layer is clear.) Rinse
stopper into flask with diethyl etherpetroleum ether mixture.
Decant ether (top) layer into 150 mL beaker and carefully rinse
lip of flask into beaker with diethyl etherpetroleum ether mixture. Slowly evaporate ether on steam bath, using nitrogen
stream to aid evaporation. Residue remaining in beaker contains extracted fat.

E. Methylation
Dissolve extracted fat residue in 23 mL chloroform and 2
3 mL diethyl ether. Transfer mixture to 3 dram glass vial and
then evaporate to dryness in 40C water bath under nitrogen
stream. Add 2.0 mL 7% BF3 reagent, C(j), and 1.0 mL toluene,
C(g). Seal vial with screw cap top containing Teflon/silicone
septum. Heat vial in oven 45 min at 100C. Gently shake vial
ca every 10 min. (Note: Evaporation of liquid from vials indicates inadequate seals; if this occurs, discard sample and repeat
the entire procedure.) Allow vial to cool to room temperature.
Add 5.0 mL H2O, 1.0 mL hexane, and ca 1.0 g Na2SO4, C(i).
Cap vial and shake 1 min. Allow layers to separate and then
carefully transfer top layer to another vial containing ca 1.0 g
Na2SO4. (Note: Top layer contains FAMEs including FAME of
triglyceride internal standard solution.)
Inject FAMEs onto GC column or transfer to autosampler
vial for GC analysis.

F. GC Determination
Relative retention times (vs FAME of triglyceride internal
standard solution) and response factors of individual FAMEs
can be obtained by GC analysis of individual FAME standard
solutions and mixed FAME standard solution.
Inject ca 2 L each of individual FAME standard solutions
and 2 L of mixed FAMEs standard solution. Use mixed
FAMEs standard solution to optimize chromatographic response before injecting any test samples.
After all chromatographic conditions have been optimized,
inject test samples from E.

G. Calculations
Total fat is the sum of fatty acids from all sources, expressed
as triglycerides. Expressing measured fatty acids as
triglycerides requires the mathematical equivalent of condensing each fatty acid with glycerol. For every 3 fatty acid molecules, 1 glycerol (HOCH2CHOHCH2OH) is required. Essentially, 2 methylene groups and 1 methine group are added to
every 3 fatty acids.
Calculate retention times for each FAME in individual
FAMEs standard solutions, C(m)(3), by subtracting retention
time of C11:0 peak from retention time of fatty acid peak. Use
these retention times to identify FAMEs in mixed FAMEs
standard solution. Use additional FAME solutions (from the
same supplier) when necessary for complete verification of
FAME identity.
Calculate response factor (Ri) for each fatty acid i as follows:
Ri =

Psi
WC11:0

PsC11:0
Wi

where Psi = peak area of individual fatty acid in mixed FAMEs

standard solution; PsC11:0 = peak area of C11:0 fatty acid in
mixed FAMEs standard solution; WC11:0 = weight of internal
standard in mixed FAMEs standard solution; and Wi = weight
of individual FAME in mixed FAMEs standard solution.
Note: For any unknown or uncalibrated peaks, use the nearest calibrated fatty acid response factors and conversion factors
(see Table 996.06D).
Calculate amount of individual triglycerides (WTGi) in test
sample as follows:
Pti WtC11:0 1.0067
Ri
PtC11:0
WTGi = WFAMEi fTGi

WFAMEi =

where Pti = peak area of fatty acid i in test sample; WtC11:0 =

weight of C11:0 internal standard added to sample, g; 1.0067 =
conversion of internal standard from triglyceride to FAME;
PtC11:0 = peak area of C11:0 internal standard in test sample; and
fTGi = conversion factor for FAMEs to triglycerides for individual fatty acids (see Table 996.06E).
(Note: If procedure is followed exactly, WtC11:0 should be
0.010 g.)

Calculate amount of total fat in test sample (sum of all fatty

acids expressed as triglycerides [including cis and trans forms
of monounsaturated acids]) as follows:
Total fat, % =

WTGi
Wsample

100

where Wsample = weight of test sample, g.

Calculate weight of each fatty acid (Wi) as follows:
Wi = WFAMEi fFAi
where fFAi = conversion factors for conversion of FAMEs to
their corresponding fatty acids (see Table 996.06D).
Calculate percent saturated fat in test sample (w/w; expressed as saturated fatty acids; sum of C4:0, C6:0, C8:0, etc.) as
follows:
Saturated fat, % =

Saturated Wi
100
Wsample

Calculate percent monounsaturated fat in test sample (w/w;

expressed as sum of only cis form of monounsaturated fatty
acids [C16:1, C17:1, C18:1 cis, C20:1, etc.]) as follows:
Monounsaturated fat, % =

Monounsaturated Wi
100
Wsample

(Note: Samples containing hydrogenated fat will yield complicated chromatograms due to large numbers of isomers
formed during hydrogenation process. One general indication
of hydrogenation is presence of C18:1 trans peak(s). For chromatograms of hydrogenated fats, use the following guidelines
to calculate FAME peak areas: Trans peaks elute prior to cis,
therefore, include all peaks between C18:1 cis and C18:2 cis,cis
in calculation of C18:2 peak area. Often C18:1 trans peak consists of broad series of peaks [due to positional isomers from
hydrogenation]; include all of these in C18:1 trans peak area.)
Ref.: J. AOAC Int. 80, 555563(1997)
Results and Discussion
Each laboratory received 3 replicates of an oat-based cereal
(fat content about 7%) to serve as a precollaborative test sample
(Table 2). Results were analyzed for accuracy (versus the
authors laboratory) and precision. Once a laboratory submitted
satisfactory results for total fat, it was sent the collaborative
study samples. No laboratory was required to repeat analysis of
pretest samples.
Samples consisted of 8 pairs of blind duplicates, representing a wide variety of matrixes (Table 1). Also listed in Table 1 is the approximate sector of the food triangle, recommended for use by the AOAC Task Force on Methods for
Nutrition Labeling Analyses (8), into which each sample belongs.
Each laboratory was instructed to analyze each sample
(116) only one time. The exception would be obvious mistakes that would necessitate reruns. Sample weights for each

sample were suggested to ensure about 100200 mg fat in the

extracted residue. Results are presented on an as received basis, that is, not corrected for moisture.
Total fat.Results for total fat are shown in Table 3. Many
points from laboratory 10 were removed on the basis of Cochran and Grubb tests. All other data points excluded from statistical analysis by the Cochran and Grubb tests are noted in the
table. Other than the outlier results from laboratory 10, only
2 other pairs of data were considered outliers. Table 996.06A
contains the statistical summary of the total fat data. The average intralaboratory relative standard deviation (RSD) is about
5%; the average interlaboratory relative standard deviation is
about 10%, demonstrating the reproducibility and within-laboratory agreement of this method.
Saturated fat.Results for saturated fat are presented in Table 4. Outliers are marked as before. The statistical summary
for saturated fat is shown in Table 996.06B. The intralaboratory
and interlaboratory variation for saturated fat is similar to that
for total fat.
Monounsaturated fat.Results for monounsaturated fat are
presented in Table 5. Outliers are marked as before. The statistical summary for monounsaturated fat is shown in Table 996.06C. Additional variation is expected from the greater
level of interpretation of the chromatographic data. The intralaboratory and interlaboratory variations for saturated fat are
similar to that for total fat. Samples 6 and 9 are exceptions, with
inordinately high RSDs. Although low-fat, high-sugar products
should not be difficult to prepare and analyze, the small amount
of monounsaturated fat, coupled with the requirement of including only the cis form, reinforces the need to understand the
NLEA requirements and to maintain a high-resolution GC system.

Collaborators Comments
On the basis of comments from collaborators, the following
additions, corrections, and other recommended changes to the
method are reflected in the method described:
(1) Additional instructions on how to interpret complicated
chromatograms, usually from samples containing hydrogenated fat, are now included.
(2) Conversion factors to transform from FAME weights to
fatty acid weights in the report sheet have been changed to exhibit 4 significant figures rather than 2.
(3) There were a few comments regarding the use of hazardous chemicals. However, we cannot at this time eliminate
liquid organic solvents.
Recommendation
On the basis of the results of this study it is recommended
that the hydrolytic extractionGC method for determining fat
(total, saturated, and monounsaturated) in foods be adopted
first action.
Acknowledgments
I thank each of the collaborators and their associates for their
contribution, patience, and persistence in completing this study:

Marijane Lawson and Tom Hammill, FDA Atlanta Center

for Nutrient Analysis, Atlanta, GA
Ross Mazal, Land O Lakes, Inc., Minneapolis, MN
Keith Meyer and Barbara Balch, Kraft General Foods,
Glenview, IL
Jerry Ngeh-Ngwainbi, Kelloggs, Battle Creek, MI
Philip Oles and Sandy Bailey, Lancaster Laboratories, Lancaster, PA
Ralph Bakowski and Tom Phillipo, USDA FSIS Eastern
Laboratory, Athens, GA
Martin Phist, Hormel Foods, Austin, MN
Joe Polywacz and Sharon Hill, Hazelton Laboratories,
Madison, WI
Larry Menke, Bill Wise, and Steven Powell, USDA FSIS
Western Laboratory, Alameda, CA
Rodney Johnson and Dianna Martin, General Mills, Inc.,
Minneapolis, MN
I thank my other colleagues at General Mills, Paul Larson
and Jon DeVries, for their valuable assistance and advice. Finally, I acknowledge the guidance and assistance of David Firestone, AOAC General Referee for Fats and Oils.

References
(1) Fed. Reg. (1990) 58, 6312964
(2) Carpenter, D.E., Ngeh-Ngwainbi, J., & Lee, S. (1993) in
Methods of Analysis for Nutrition Labeling, Chapter 5, D.M.
Sullivan & D.E. Carpenter (Eds), AOAC INTERNATIONAL, Arlington, VA
(3) Official Methods of Analysis (1990) 15th Ed., AOAC, Arlington, VA, sec. 920.39C
(4) Daugherty, C., & Lento, H. (1983) J. Assoc. Off. Anal. Chem.
66, 927932
(5) Sheppard, A.J., Hubbard, D.W., Prosser, A.R., Shen, J.C.-S.,
ODell, R.G., & Jones, S.T. (1989) Lipid Manual, U.S. Food
and Drug Administration, Washington, DC
(6) Ngeh-Ngwainbi, J. (1992) Analysis of Fat in Low Fat Cereal
Products, presented at the 106th Annual AOAC INTERNATIONAL Conference, Cincinnati, OH
(7) House, S.D., Larson, P.A., Johnson, R.R., DeVries, J.W., & Martin, D.L. (1994) J. AOAC Int. 77, 960965
(8) Wolf, W.R. (1993) in Methods of Analysis for Nutrition Labeling, Chapter 7, D.M. Sullivan & D.E. Carpenter (Eds),
AOAC INTERNATIONAL, Arlington, VA

Table 996.06A. Method performance for determination of total fat in foodstuffs by hydrolytic extractiongas
chromatography

Samplea

_
x, %

sr

sR

rb

Rc

RSDr, %

RSDR, %

No. of
laboratories
excludedd

Wheat-based cereal
Peanut butter
Fish sticks
Parmesan cheese
Chocolate cake (baked)
Fruit snack
Ground beef
Yogurt

1.96
46.3
11.2
26.5
13.3
3.92
21.91
1.46

0.208
0.86
0.354
0.540
0.929
0.087
1.11
0.131

0.260
2.37
0.541
4.17
1.95
0.146
1.82
0.222

0.582
2.41
0.991
1.51
2.60
0.244
3.11
0.367

0.728
6.64
1.51
11.7
5.46
0.409
5.10
0.622

10.6 1
1.85
3.14
2.04
7.00
2.22
5.06
8.98

13.3
15.12
14.80
15.8
14.7
13.74
18.32
15.2

2/10
2/10

1/10
1/10

a
b
c
d

Blind duplicates.
r = 2.8 sr.
R = 2.8 sR.
Outliers by either Cochran or Grubbs tests.

Table 996.06B. Method performance for determination of saturated fat in foodstuffs by hydrolytic extractiongas
chromatography

Samplea
Wheat-based cereal
Peanut butter
Fish sticks
Parmesan cheese
Chocolate cake (baked)
Fruit snack
Ground beef
Yogurt
a
b
c
d

_
x, %

sr

sR

rb

Rc

RSDr, %

RSDR, %

No. of
laboratories
excludedd

0.493
8.72
3.00
17.4 1
3.56
1.27
9.98
0.986

10.0391
0.257
0.223
0.311
0.171
0.0242
0.636
0.119

0.0522
1.81
0.572
2.46
0.304
0.0362
1.39
0.170

0.109
0.720
0.624
0.871
0.479
0.0678
1.78
0.333

0.146
5.07
1.60
6.89
0.851
0.101
3.89
0.476

7.92
2.95
7.44
1.79
4.81
1.90
6.38
12.1

10.6
20.7
19.1
14.1
8.55
2.83
13.9
17.2

1/10
1/10
1/10

2/10
1/10

Blind duplicates.
r = 2.8 sr.
R = 2.8 sR.
Outliers by either Cochran or Grubbs tests.

Table 996.06C. Method performance for determination of monounsaturated fat in foodstuffs by hydrolytic
extractiongas chromatography

Samplea
Wheat-based cereal
Peanut butter
Fish sticks
Parmesan cheese
Chocolate cake (baked)
Fruit snack
Ground beef
Yogurt
a
b
c
d

_
x, %

sr

sR

rb

Rc

RSDr, %

RSDR, %

No. of
laboratories
excludedd

0.280
22.3
1.83
6.43
3.79
1.08
8.88
0.345

0.0320
0.411
0.165
0.271
0.413
0.0453
0.930
0.0222

0.0560
1.11
0.313
1.09
1.27
0.734
1.96
0.0542

0.0896
1.15
0.462
0.759
1.16
0.127
2.60
0.0621

0.157
3.11
0.876
3.05
3.56
2.06
5.49
0.152

11.4
1.84
9.02
4.20
10.9
4.17
10.5
6.42

20.0
4.96
17.1
17.0
33.5
67.7
22.0
15.1

2/10

1/10

2/10

1/10

Blind duplicates.
r = 2.8 sr.
R = 2.8 sR.
Outliers by either Cochran or Grubbs tests.

Table 996.06D. Factors for conversion of FAMEs (fi) to

their corresponding fatty acids
Fatty acid
4:0
6:0
8:0
10:0
11:1
12:0
13:0
14:0
14:1
15:0
15:1
16:0
16:1
17:0
17:1
18:0
18:1 (cis or trans)
18:2
18:3
18:4
20:0
20:1
20:5
20:6
22:0
22:1
22:5
22:6
24:0

fi
0.8627
0.8923
0.9114
0.9247
0.9300
0.9346
0.9386
0.9421
0.9417
0.9453
0.9449
0.9481
0.9477
0.9507
0.9503
0.9530
0.9527
0.9524
0.9520
0.9517
0.9570
0.9568
0.9557
0.9554
0.9604
0.9602
0.9593
0.9590
0.9963

Table 996.06E. Factors (fTGi) for conversion of FAMEs

to triglyceride equivalents
Fatty acid
4:0
6:0
8:0
10:0
11:0
12:0
14:0
14:1
15:0
15:1
16:0
16:1
17:0
17:1
18:0
18:1 (cis or trans)
18:2
18:3
18:4
20:0
20:1
20:5
20:6
22:0
22:1
22:5
22:6
24:0

fTGi
0.9868
0.9897
0.9915
0.9928
0.9933
0.9937
0.9945
0.9944
0.9948
0.9947
0.9950
0.9950
0.9953
0.9952
0.9955
0.9955
0.9954
0.9954
0.9954
0.9959
0.9959
0.9958
0.9957
0.9962
0.9962
0.9961
0.9961
0.9965

Table 1. Samples used in collaborative study

Sample
Wheat-based cereal
Fruit snack
Peanut butter
Ground beef
Fish sticks
Yogurt
Parmesan cheese
Chocolate cake (baked)

ID No.

Food triangle location

1, 15
6, 91
2, 16
7, 10
3, 14
8, 11
4, 13
5, 12

5
5
3
4
8
6
4
6

Table 2. Preliminary test results (%

% total fat w/w) for
oat-based cereal
Sample ID

Statistics

Test 1

Test 2

Test 3

Mean

CV, %

1
2
3
4
5
6
7
8
b
9
10

6.32
7.32
NAa
7.45
7.42
7.43
7.06
7.50
7.15
7.29

6.71
7.28
NA
7.52
7.41
7.59
7.05
7.40
7.19
6.59

6.70
7.23
NA
7.52
7.34
7.54
7.07
7.50
7.16
6.97

6.58
7.28

7.50
7.49
7.52
7.06
7.47
7.17
6.95

3.34
0.62

0.53
0.77
1.09
0.14
0.77
0.29
5.04

Average
SD
CV, %

7.21
0.39
5.41

7.27
0.29
3.93

7.26
0.29
3.93

Laboratory ID

a
b

NA, not completed.

Control laboratory: General Mills.

Table 3. Collaborative study results for total fat (%

% w/w), by laboratory ID and sample number
Sample
Laboratory ID
11
12
13
14
15
16
17
18
19
10
a
b
c

15

16

1.83
2.15
2.21
2.01
1.58
2.49
2.10
2.10
1.96
2.02

1.66
2.14
1.9
1.93
1.66
2.07
1.91
2.06
2.06
1.31

44.10
48.60
47.92
45.62
44.15
47.74
a
51.41a
46.40
49.54
b
30.97c

42.20
48.96
46.23
43.82
43.59
47.27
a
45.96a
45.22
49.40
b
30.18c

14

11.10 10.80
11.34
11.15
12.02 11.45
10.59 11.15
10.58 10.76
11.09
11.40
a
12.84a a10.54a
11.22
11.25
11.63 12.68
1b7.22c 1b7.37c

Outliers removed by Cochran test.

NA, not completed.
Outliers removed by Grubbs test; blind duplicates grouped together.

13

12

10

11

28.60
27.89
32.66
18.94
25.16
28.05
25.64
26.84
30.81
20.07

28.90
27.22
32.3
20.67
25.14
26.98
26.28
26.78
30.92
19.30

12.80
14.22
14.97
12.06
12.91
13.63
16.20
14.63
14.66
17.87

11.00
14.28
13.96
12.93
12.82
13.88
13.45
14.65
14.76
10.01

4.20
4.00
3.92
3.73
3.85
3.66
3.99
NAb
4.04
a
2.44a

4.00
3.87
3.91
3.97
3.92
3.65
3.98
NA
4.01
a
8.04a

20.70
24.79
21.46
20.87
19.86
23.80
22.51
20.74
23.97
a
5.33a

20.30
22.14
23.63
19.02
18.82
22.69
22.29
22.86
24.02
a
13.93a

1.63
1.46
1.50
1.32
1.31
1.38
1.25
1.19
1.67
1.33

1.69
1.46
1.95
1.38
1.33
1.67
1.25
1.26
1.88
1.30

Table 4. Collaborative study results for saturated fat (%

% w/w), by laboratory ID and sample number
Sample
Laboratory ID
11
12
13
14
15
16
17
18
19
10
a
b

15

0.49
0.54
0.57
b
0.79b
0.42
0.51
0.54
0.53
0.50
0.48

0.43
0.54
0.47
b
0.76b
0.44
0.51
0.51
0.52
0.52
0.37

16

8.31
7.97
9.82
9.82
8.88
8.63
b
11.40b1 b11.30b1
8.23
8.12
a
6.64a a8.91a
9.70
8.92
8.63
8.39
9.75
9.71
5.03
4.46

3
a

11.10a1
2.97
3.61
4.06
2.74
2.70
3.40
2.91
3.02
1.89

Outliers removed by Cochran test.

Outliers removed by Grubbs test; blind duplicates grouped together.

14

13

12

10

11

2.87a
2.92
2.97
3.99
2.82
2.99
2.84
2.91
3.29
1.95

19.20
17.70
20.50
13.50
16.30
18.20
17.30
17.90
20.20
13.10

19.40
17.30
20.3
14.60
16.30
17.80
17.70
17.90
20.30
12.70

3.39
3.56
3.64
3.98
3.10
3.54
4.04
3.53
3.58
2.29

3.17
3.56
3.43
4.22
3.08
3.58
3.43
3.56
3.60
1.27

1.26
1.31
1.25
b
1.68b
1.23
1.34
1.31
1.27
1.31
a
0.92a

1.20
1.28
1.27
b
1.78b
1.25
1.28
1.31
1.29
1.30
a
3.68a

9.88
7.87
9.91
11.001
9.05
11.301
11.101
9.79
10.901
a
1.87a

9.68
6.14
10.8 1
9.95
8.58
10.601
10.801
11.001
11.201
a
6.45a

1.11
0.94
0.95
1.04
0.85
0.90
0.85
0.82
1.09
0.80

1.12
0.94
1.37
1.27
0.86
1.09
0.85
0.87
1.21
0.80

Table 5. Collaborative study results for monounsaturated fat, by laboratory ID and sample number
Sample
Laboratory ID
11
12
13
14
15
16
17
18
19
10
a
b
c

15

16

14

0.23
0.32
0.39
0.36
0.21
0.28
0.29
0.27
0.27
a
0.29a

0.21
0.32
0.32
0.35
0.22
0.27
0.28
0.26
0.29
a
0.17a

20.70
23.10
22.20
23.80
21.30
22.90
a
24.10a
22.30
23.70
b
12.80c

20.40
23.30
21.8
22.40
21.10
22.70
a
21.20a
21.70
23.60
b
12.90c

1.78
1.92
2.43
2.16
1.66
1.79
2.13
1.79
1.94
1.12

1.74
1.89
1.97
1.94
1.69
1.91
1.65
1.79
2.13
1.14

4
a

2.11a
7.50
8.39
5.40
5.68
6.88
6.18
6.58
7.10
4.65

Outliers removed from statistical analysis by Cochran test.

NA, not completed.
Outliers removed from statistical analysis by Grubbs test; blind duplicates grouped together.

13

12

10

11

6.99a
6.78
8.21
6.00
5.65
6.32
6.36
6.55
7.14
4.41

1.02
2.76
4.65
5.14
3.47
4.39
5.79
2.53
4.63
2.93

2.32
2.80
4.34
5.43
3.44
4.50
4.80
2.47
4.65
3.67

0.57
1.09
0.50
2.25
0.24
0.87
2.04
NAb
1.04
1.27

0.53
1.03
0.55
2.38
0.23
0.96
2.03
NA
1.04
3.51

9.31
7.90
9.25
9.96
8.72
10.20
9.87
9.62
10.40
a
2.74a

9.13
6.11
10.4
9.29
8.32
9.96
9.90
10.401
10.101
a
6.11a

0.37
0.39
0.37
a
0.46a
0.29
0.34
0.30
0.26
0.40
0.31

0.39
0.38
0.38
a
0.25a
0.30
0.41
0.30
0.29
0.44
0.29