FOOD COMPOSITION AND ADDITIVES

Vitamin D in Infant Formula and Enteral Products by Liquid

Chromatography: Collaborative Study
SLIVA & SANDERS: JOURNAL OF AOAC INTERNATIONAL VOL. 79, NO. 1, 1996
MATTHEW G. SLIVA 1 and JAMES K. SANDERS
Bristol-Myers Squibb, 2400 W Lloyd Expressway, Evansville, IN 47721
Collaborators: M. Arendsen; A. Baars; S. Bhandari; B. Boerma; J. Broge; M. Bueno; G. Cherix; G. Gates; J. Hollembaek; M.
Kaufman; C. Kraft; L. Oerl; A. Santos; W. Schuep; J. Wehrmann

Results from a collaborative study of a new liquid

chromatographic (LC) method for determination of
vitamin D in infant formulas and enteral products
are presented. Each of 15 laboratories was provided with 11 blind duplicate samples covering a
range of approximately 200500 International
Units/quart (normal dilution), a system suitability
sample, and the U.S. Pharmacopeia ergo- and
cholecalciferol standards. Product types included
liquid and powder forms of milk (whey and casein),
soy, and hydrolyzed protein-based infant formulas
and enteral products. The method includes a single
liquidliquid extraction following saponification,
solid-phase extraction, and then concentration by
evaporation. An isocratic, nonaqueous, chromatographic system with reversed-phase, zero endcapped C18 column, and UV detector set at 265 nm
are used. Statistical evaluation of data from participating laboratories show the average reproducibility and repeatability of the method across all samples to be excellent, with RSDR and RSDr values of
13.48 and 9.44, respectively, after elimination of outliers. The LC method for determination of vitamin D
in infant formulas and enteral products has been
adopted by AOAC INTERNATIONAL.

easurement of vitamin D in infant formulas and enteral products has been an analytical issue for a
number of years. The current official method for
analysis of vitamin D in infant formula is the rat bioassay
(1). A new liquid chromatographic (LC) method was
adopted first action for milk-based infant formula only (2) at
the 1992 AOAC Annual Meeting. Analysts have expressed

Submitted for publication August 10, 1994.

The recommendation was approved by the Committee on Food
Nutrition, and was adopted by the Official Methods Board of the
Association. See Official Methods Board Actions (1995) J. AOAC Int. 78,
45A, and Official Methods Board Actions (195) The Referee 19, March
issue.
1
Current address: NAmSA, 2261 Tracy Rd, Northwood, OH 43619.

concerns with both these methods. Most analysts consider the

variability and cost of performing the rat bioassay unacceptable. In addition, a strong desire to reduce use of laboratory
animals exists. The LC method approved for milk-based products, although an improvement over the rat bioassay, is labor
intensive and has a higher than desired variability. The vitamin
D method developed by Bristol-Myers (3) was studied collaboratively because it was considered to adequately address these
concerns.
Collaborative Study
Fifteen collaborators were provided with 11 blind duplicate
samples. The sample assortment attempted to cover the range
of product types in the infant formula/enteral product industry.
Liquid and powder forms of infant formulas and enteral products made up of protein sources including milk (casein, whey,
and partially hydrolyzed whey), soy, and hydrolyzed milk protein were analyzed. The range of vitamin D in these products,
per label claims, was 200500 International Units (IU)/quart
normal dilution (as directed on the label), covering the majority
of the infant formula regulation specifications. The collaborative study protocol indicated expected variabilities of 6% between laboratories and 4% within laboratories. Although results did not meet these high expectations, significant
improvement in variability over previously studied methods
was seen, and therefore the study was considered successful.
The method was evaluated for within-laboratory variability
(repeatability) and between-laboratory variability (reproducibility). Grubbs and Cochran outliers were determined according to AOAC guidelines (4). The following statistics were reported on each sample: repeatability standard deviation (sr) and
relative standard deviation (RSDr); reproducibility standard deviation (sR) and relative standard deviation (RSDR); overall
mean; and outliers determined with the Grubbs and Cochran
tests. In addition, mean RSDr and RSDR values were calculated
for all samples, liquids only, powders only, infant formulas
only, and enteral products only.

995.05 Vitamin D in Infant Formulas and Enteral

Products, Liquid Chromatographic Method
First Action 1995
(Applicable to determination of 82600 IU vitamin D/qt
[1 qt = 0.946 L] in infant formulas and enteral products.)
Caution: See Appendix: Laboratory Safety for Safe Handling of Organic Solvents and Safe Handling of Special
Chemical Hazardsacetonitrile, dichloromethane, ethanol,
ethyl acetate, and hexane. Dispose waste solvents in an appropriate manner compatible with applicable environmental rules
and regulations.
Method Performance:
See Table 995.05A for method performance data.

A. Principle
Test sample is saponified, extracted, and evaporated to concentrate nutrient. Vitamin D is determined by reversed-phase
LC with UV detection at 265 nm wavelength.

B. Apparatus
(a) Liquid chromatograph (LC).Equipped with UV detector and capable of meeting system suitability requirements.
Operate at 27 1C; use at higher temperature results in loss
of resolution.
(b) LC column.250 4.6 mm id, C18, 5 m particle size.
Operating conditions: injection volume, 250 L; column temperature, 27C; wavelength, 265 nm; flow rates, see Table 995.05B for flow rates; stop time, 35 min; retention times,
vitamin D2, 19.5 min; vitamin D3, 23 min. Note: The column
must not be end-capped.
(c) Solid-phase extraction (SPE) column.Silica;
500 mg/2.8 mL.
(d) Vacuum manifold.For SPE column, (c).
(e) Evaporator.With nitrogen flow.
(f) Rotary evaporator.
(g) Water bath shaker.Capable of maintaining 60C.
(h) Separatory funnel.250 mL capacity.
(i) Amber volumetric flask.250 mL.
(j) Reflux apparatus.

C. Reagents
(a) Solvents (LC grades).n-Hexane, dichloromethane,
acetonitrile, isopropanol, methanol, ethyl acetate.
(b) Ethanol.Absolute, pharmaceutical grade.
(c) Phenolphthalein solution.1%. Dissolve 1 g phenolphthalein in 100 mL absolute ethanol.
(d) Dichloromethaneisopropanol (IPA) solutions.(1)
99.8 + 0.2, v/v.Transfer 2 mL isopropanol into 1 L volumetric flask. Dilute to volume with dichloromethane and mix. (2)
80 + 20, v/v.Transfer 200 mL isopropanol into 1 L volumetric flask, dilute to volume with dichloromethane, and mix.
(e) Acetic acid solution.10%. Transfer 10 mL glacial
acetic acid (AR grade) into 100 mL volumetric flask, dilute to
volume with H2O, and mix well.

(f) Ethanolic potassium hydroxide (KOH) solution.Dissolve 140 g KOH pellets (AR grade) in 310 mL absolute ethanol and add 50 mL H2O. Prepare fresh on day of use.
(g) Mobile phase.Gradient mixture of acetonitrile,
methanol, and ethyl acetate. See Table 995.05B for concentrations of mobile-phase components.
(h) Vitamin D2 standard solutions.Certified reference
standard or traceable secondary standard. (1) Stock standard
solution.180 g/mL. Accurately weigh 45 mg vitamin D2
and quantitatively transfer to 250 mL amber volumetric flask.
Dilute to volume with absolute ethanol. (2) Working standard
solution.2.88 g/mL. Transfer 4.0 mL stock standard solution into 250 mL amber volumetric flask, and dilute to volume
with absolute ethanol. Store in refrigerator 7 days. (3) Internal standard solution.46 ng/mL. Use to quantitate vitamin
D3. Transfer 4.0 mL working standard solution into 250 mL
amber volumetric flask, and dilute to volume with absolute
ethanol. Store in refrigerator 7 days.
(i) Vitamin D3 standard solutions.Certified reference
standard or traceable secondary standard. Prepare and store in
refrigerator 7 days. (1) Stock standard solution.180 g/mL.
(2) Working standard solution.2.88 g/mL. (3) Internal
standard solution.46 ng/mL. Use to quantitate vitamin D2.
(j) System suitability standard solution.Dissolve 125 mg
certified vitamin D2 standard and 125 mg certified vitamin D3
standard in 10 mL mobile phase, (g). Heat solution 45 min at
90C under reflux, then cool. Transfer 1.0 mL solution from
reflux to 100 mL amber volumetric flask and dilute to volume
with mobile phase.
Note: Protect vitamins D2 and D3 standard solutions from
high temperature, oxygen, and light, to minimize degradation.

D. Preparation of Standards Mixture and Test

Samples
Note: Perform assay in subdued lighting to minimize vitamin degradation.
(a) Standards mixture.Transfer 4.0 mL vitamin D2 internal standard solution, C(h), and 4.0 mL vitamin D3 internal
standard solution, C(i), into 125 mL Erlenmeyer flask, and add
15.0 mL H2O.
(b) Test sample.Prepare formula or enteral product according to label directions and use as test sample. Vitamin D
concentration will be 0.5 IU/mL. Transfer 15.0 mL test sample
into 125 mL Erlenmeyer flask, using pipet or syringe (depending on viscosity). Add 4.0 mL internal standard solution to
flask. Use vitamin D2 standard solution, C(h)(3), to quantitate
vitamin D3. Use vitamin D3 standard solution, C(i)(3), to quantitate vitamin D2.

E. Saponification and Extraction

Add 15.0 mL aliquots of ethanolic KOH solution, C(f), to
standards mixture, F(1), and to test sample from D(b). Close
flasks with stoppers and place them in water bath shaker for
30 min at 60C. Remove flasks from water bath shaker and
allow to cool to room temperature.
Treat each solution similarly, as follows: Transfer contents
of flask to 250 mL separatory funnels. To empty flask add

15.0 mL H2O, close with stopper, and shake vigorously. Transfer contents of each flask to corresponding funnel, rinse empty
flask again with 60 mL hexane, and transfer rinse to funnel.
Close funnel with stopper and shake vigorously 90 s. Allow
layers to separate ca 10 min. Drain and discard aqueous layer.
Add 15.0 mL H2O to hexane layer remaining in funnel, close
funnel with stopper, and shake vigorously. Allow layers to
separate; drain and discard aqueous layer.
Add 1 drop phenolphthalein solution, C(c), and 15.0 mL
H2O to funnel. Add 10% acetic acid solution, C(e), to funnel,
dropwise with shaking, until wash is neutral to phenolphthalein
(colorless). Drain and discard aqueous layer. Drain hexane
layer through sodium sulfate, supported by small cotton plug,
into 100 mL round-bottom flask. Rinse funnel and sodium sulfate with few milliliters of hexane, collecting hexane in same
round-bottom flask.

F. Evaporation and Solid-Phase Extraction

Treat extract solutions from E as follows: Place round-bottom flask on rotary evaporator and evaporate hexane to dryness
at 40C. Immediately after evaporation, add 2.0 mL dichloromethaneIPA solution, C(d)(1), to flask.
Wash SPE column with 4.0 mL dichloromethaneIPA solution, C(d)(2), and then with 5.0 mL dichloromethaneIPA solution, C(d)(1). Transfer solution from round-bottom flask to
SPE column. Rinse empty flask with 1.0 mL dichloromethane
IPA solution, C(d)(1), and transfer rinse to SPE column. Wash
SPE column with 2.0 mL dichloromethaneIPA solution,
C(d)(1), and discard this fraction. Elute vitamins D2 and D3
with 7.0 mL dichloromethaneIPA solution C(d)(1) into 16
100 mm disposable culture tube.
Place culture tube in water bath and evaporate dichloromethaneIPA solution at 40C using nitrogen. When tube is
dry, immediately add 1.0 mL acetonitrile and swirl to rinse
down sides of tube. Transfer solution to LC vial with disposable
dropper.

G. Chromatographic Determination
Inject standards mixture, D(a), onto LC column at beginning, middle, and end of each run of test solutions.
To calculate content of vitamins D2 and D3, use internal
standard procedure and peak heights.

H. Calculations
Perform following calculations:
(1) Calculate concentration of vitamin D2 in standards mixture (CSD2) as follows:
(W 4 4 4 40000)
CSD2, IU/mL =
1.05
(250 250 250)
where W = weight of vitamin D2 certified reference standard
(mg); 4 = dilution factor; 40000 = IU vitamin D/mg; 250 =
volumes of subsequent dilutions of vitamin D2 standard solutions; 1.05 = correction factor for previtamin D.
(2) Calculate concentration of vitamin D3 in standards mixture (CSD3) as above, using weight of vitamin D3 certified reference standard.

(3) Calculate response ratio of vitamin D2 in standards mixture (RSD2) as follows:

RSD2 =

PS D2
PS D3

where PSD2 = peak height of vitamin D2 in standards mixture;

and PSD3 = peak height of vitamin D3 in standards mixture.
(4) Calculate response ratio of vitamin D3 in standards mixture (RSD3) as above, using peak heights of vitamins D3 and D2,
respectively.
(5) Calculate response ratio of vitamin D2 in test sample
(RTD2) as follows:

RTD2 =

PT D2
PT D3

where PTD2 = peak height of vitamin D2 in test sample; and

PTD3 = peak height of vitamin D3 in test sample.
(6) Calculate response ratio of vitamin D3 in test sample
(RTD3) as above, using peak heights of vitamins D3 and D2,
respectively.
(7) Calculate concentration of vitamin D2 (CTD2) or D3
(CTD3) in test sample as follows:

CTD2,IU/mL =

RTD2

CTD3,IU/mL =

RTD3

RSD2
RSD3

CSD2
UD
Vt

CSD3
UD
Vt

where Vt = volume of test sample, mL (usually 15); U = conversion factor to appropriate units (if necessary); and D = dilution factor for diluted powders or liquids.

I. System Suitability
For system suitability use standards mixture, D(a), during
routine operation. Resolution factor (defined by U.S. Pharmacopeia) between vitamin D2 and vitamin D3 should be 2.0.
Separation between these peaks should be sufficient to allow
additional peak to be resolved between vitamin D2 and D3,
namely, previtamin D3.
To verify that previtamin D3 is resolved, inject system suitability standard solution, C(j). The system suitability standard
solution contains, in order of elution, previtamin D2, vitamin
D2, previtamin D3, and vitamin D3.
Optimize chromatography by adjusting amount of methanol
in mobile phase; decreasing methanol content increases retention times. Maintain column temperature at 27 1C; increased temperature decreases retention, and vice versa. Sixreplicate injections of standard should have <2% relative
standard deviation (RSD). RSD values for standards throughout run should be <4%.
Ref.: J. AOAC Int. 75, 566(1992); J. AOAC Int. 79, 73(1996)
Collaborators Comments
Comments received from collaborators were generally favorable with regard to the ease of use of the method as well as
the clean chromatography. Two suggestions regarding spe-

cific procedures are as follow: (1) increase in equilibration

time after each post-injection column cleanup would assure stable baseline; and (2) addition of KOH as a pellet rather than
preparing ethanolic KOH would improve saponification. These
suggestions have merit and individual laboratories can consider
adopting these revised procedures, if internal validation is successful. Changes such as these do not change the performance
of the method, therefore, no changes to the method as written
were made.
A general comment from one collaborator confirmed the
temperature sensitivity of the chromatographic system, as addressed in the method.

In summary, this collaborative study shows the method to

have very good repeatability and reproducibility. Comparisons
to previously studied LC methods for vitamin D indicate that
the method provides results with significantly improved analytical variability.
The method validation described in 1992 (3) defined the
specificity (separation of vitamins D2 and D3 as well as their
precursors), linearity (coefficient of determination of 1.000), a
working range equal to 82600 IU/quart and the broad range of
applicability, and proved its ruggedness.

Results and Discussion

This collaborative study, designed to test the LC method for

vitamin D in infant formulas and enteral products developed by
Bristol-Myers Squibb, indicates that the method is a significant
improvement over other methods studied for this determination. Overall average repeatability of 9.44% RSD and reproducibility of 13.48% RSD show that the method performed
well in all laboratories while testing a variety of products, regardless of protein source.
On the basis of the data obtained in this collaborative study
and the validation work presented in 1992, it is recommended
that this LC method for the analysis of vitamin D in infant formulas and enteral products be adopted first action.

This collaborative study was designed to evaluate BristolMyers Squibbs method for the analysis of vitamin D in infant
formulas and enteral products. Products from both classes, liquid and dry (powder) formulations, were included in the sample
pool. The study covered as broad range of vitamin D levels as
possible in an attempt to encompass the entire Infant Formula
Act range (256640 IU/quart for 20 Kcal/oz product, or 40
100 IU/100 Kcal).
Tables 13 contain a summary of the data generated by
15 collaborators. A total of 330 analyses were performed on
blind duplicates of samples AK. Results are reported in
IU/quart normal dilution for liquids and IU/g for powders
(samples E, H, and K) (1 mcg vitamin D = 40 IU).
The data were statistically analyzed using AOAC software.
Results of that analysis are presented in Tables 995.05A, 4, and
5. The data were checked for outliers using Grubbs and Cochran tests. Thirty of 330 values were eliminated: 14 as Grubbs
outliers, 4 as Cochran outliers, and 12 due to lack of one of a
duplicate samples (6 analyses reported no value due to lost
sample). The calculations for repeatability and reproducibility
do not include these outliers. Mean values compared favorably
to label claim values when compensated for overages. The
RSDR values ranged from 7.28 to 23.11% with an overall average of 13.41% (7.2819.44% with an average of 12.30% for
infant formulas only). At the concentration levels analyzed in
this procedure (0.184 mg/mL), Horwitz suggests an RSDR
value of 20.7% (5).
Because of stability concerns with vitamin D in retorted
products, a true value from which to calculate recovery was
not available. (The NIST vitamin D standard is in oil and does
not reflect this product matrix. Attempts to run this standard in
the authors laboratory were unsuccessful due to the vitamin
D:fat ratio.) Thus, each sample was tested by the rat bioassay
to provide a comparative value. Figure 1 (data presented in Table 6) illustrates the comparison of results. A correlation coefficient of >0.95 indicates good correlation with the bioassay.
Table 7 provides a comparison of results from this collaborative study with studies performed on other vitamin D methods (5). As these data show, the method developed by BristolMyers Squibb compares very favorably with the methods
studied previously.

Recommendation

Acknowledgments
We thank John Phillips for his assistance with the statistical
evaluation of the data, Janice Saucerman for her technical assistance, and the following collaborators and associates for
their contributions to this study:
John Hollembaek, Wyeth-Ayerst Laboratories, Inc., Mason, MI
Genevieve Cherix, NQAL, Dublin, OH
W. Schuep, F. Hoffman-LaRoche Ltd, Basel, Switzerland
Bertus Boerma, Nutricia BV, Zoetermeer, The Netherlands
Jim Broge, Sandoz Nutrition Corp., St. Louis Park, MN
Gloria Gates, Lancaster Laboratories, Lancaster, PA
Adre Baars, Mead Johnson BV, Nijmegan, The Netherlands
James Wehrmann, Hazleton Laboratories, Madison, WI
Aida Santos, Nutricia-Loma Linda, Mt. Vernon, OH
Martin Bueno, U.S. Food and Drug Administration, Washington, DC
Sneh Bhandari, Silliker Laboratories, Chicago Heights, IL
Margaret Arendsen, Mead Johnson & Co., Zeeland, MI
Lisa Oerl, Southern Testing, Wilson, NC
Charlene Kraft, Mead Johnson & Co., Evansville, IN
Masako Kaufman, Takeda USA, Orangeburg, NY
References
(1) Official Methods of Analysis (1990) 15th Ed., AOAC, Arlington, VA, sec. 936.14
(2) Sertl, D.C., & Molitor, B.E. (1985) J. Assoc. Off. Anal.
Chem. 68, 177182
(3) Sliva, M.G., Green, A.E., Sanders, J.K., Euber, J.R., &
Saucerman, J.R. (1992) J. AOAC Int. 75, 566571

(4) Youden, W.J., & Steiner, E.H. (1975) Statistical Manual of

the AOAC, AOAC, Arlington, VA
(5) Horwitz, W. (1982) J. Assoc. Off. Anal. Chem. 65, 525529
(6) Tanner, J.T., Barnett, S.A., & Mountford, M.K. (1993) J.
AOAC Int. 76, 10421056

Table 995.05A. Method performance for determination of vitamin D in infant formulas and enteral products by liquid
chromatography
Samplea
IF liquid
IF liquid
IF liquid
IF liquid
IF liquid
IF liquid
E liquid
E liquid
IF powder
IF powder
E powder
a
b

c
d

__
Xb

sr

sR

RSDr, %

RSDR, %

rc

Rd

1086.7
983.5
585.0
494.7
488.1
472.5
357.6
346.1
5.69
3.96
1.09

112.4
76.89
35.59
46.27
36.08
20.99
31.03
27.49
0.22
0.54
0.25

142.13
135.98
42.61
62.69
57.81
46.46
58.19
33.55
0.60
0.77
0.25

10.34
7.82
6.08
9.35
7.39
4.44
8.68
7.94
3.87
13.71
23.11

13.08
13.83
7.28
12.67
11.84
9.83
16.27
9.69
10.46
19.44
23.11

314.72
214.73
99.65
129.56
101.02
58.77
86.88
76.97
0.62
1.51
0.70

397.96
380.74
119.31
175.53
161.87
130.09
162.93
93.94
1.68
2.16
0.70

IF = infant formula; E = enteral product.

Mean concentration of vitamin D in IU/QND for liquid formulas (IU = International Unit [1 mcg vitamin D = 40 IU]; QND = quart normal dilution
[1 qt = 0.946 L]), and in IU/g for powders.
r = 2.8 sr.
R = 2.8 sR.

Table 995.05B. Flow rates and concentrations of

mobile phase components for determination of vitamin
D in infant formulas and enteral products by liquid
chromatography
Time, min
0.0
28.0
28.5
31.0
31.5
33.0
34.0

Flow rate,
mL/min
0.7
0.7
2.5
2.5
2.5
2.5
0.7

Acetonitrile,
Ethyl
%
Methanol, % acetate, %
91.0
91.0
0.0
0.0
91.0
91.0
91.0

9.0
9.0
0.0
0.0
9.0
9.0
9.0

0.0
0.0
100.0
100.0
0.0
0.0
0.0

Table 1. Collaborative study results (including outliers) for determination of vitamin D in infant formulas and enteral
products by liquid chromatographic method developed by Bristol-Myers Squibba
Sample
C
Collaborator
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
c

N
Mean
SD
RSD, %
Label claim
a
b
c

1149
1189
1181
805
988
1030
1311
1078
1010
1060
772
918
1100
1300
920

1059
1145
NRb
1008
1006
1100
1299
1084
1280
1080
1096
1030
1090
1360
1230

1000
1081
1123
679
897
917
1208
1045
1010
966
NR
863
944
1160
610

960
1041
NR
698
910
997
924
1045
1180
942
964
878
1090
1170
1240

578
628
606
586
580
547
552
563
554
568
404
333
556
500
630

606
691
625
571
566
552
688
600
545
573
428
365
566
580
600

510
521
620
550
458
411
612
502
450
624
500
429
548
530
420

514
558
NR
514
472
437
476
497
513
497
472
487
509
530
310

29
1089.9
137.3
12.6
756

28
983.6
150.9
15.3
756

Results presented in IU/quart normal dilution, unless otherwise noted.

NR, not reported.
N, number of samples.

30
558.0
82.6
14.8
400

29
499.0
64.2
12.9
405

Table 2. Collaborative study results (including outliers) for determination of vitamin D in infant formulas and enteral
products by liquid chromatographic method developed by Bristol-Myers Squibba
Sample
D
Collaborator
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
c

N
Mean
SD
RSD, %
Label claim
a
b
c

489
482
538
502
536
452
546
556
530
430
324
361
466
510
520

508
523
NRb
526
497
430
519
523
587
436
448
466
464
500
535

234
495
536
538
455
456
459
526
507
403
436
372
431
490
490

442
524
523
471
464
439
439
492
508
422
488
370
442
500
490

350
358
337
324
357
373
486
322
321
322
244
380
307
420
410

345
404
367
342
347
375
360
358
321
321
232
420
305
430
490

313
363
385
440
307
314
342
326
352
315
332
349
310
350
370

316
375
NR
346
299
322
364
348
341
376
352
423
317
380
350

29
489.8
55.9
11.4
403

30
463.6
62.7
13.5
403

Results presented in IU/quart normal dilution, unless otherwise noted.

NR, not reported.
N, number of samples.

30
357.6
38.2
16.3
200

29
347.5
33.0
9.48
200

Table 3. Collaborative study results (including outliers) for determination of vitamin D in infant formulas and enteral
products by liquid chromatographic method developed by Bristol-Myers Squibba
Sample
K
Collaborator
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
b

N
Mean
SD
RSD, %
Label claim
a
b
c

5.55
6.36
6.25
6.20
5.56
5.10
5.30
6.70
5.44
5.78
4.40
3.01
5.99
8.69
6.33

5.66
5.99
6.12
6.00
5.43
4.94
5.80
6.50
4.89
5.62
4.56
2.83
5.48
8.85
6.00

4.55
3.65
4.12
3.50
3.67
3.75
4.30
4.02
4.67
3.71
2.97
3.66
3.98
6.59
4.10

4.11
3.69
4.64
11.6
3.71
3.62
4.30
4.30
3.47
3.95
2.16
2.66
3.95
4.46
4.16

0.87
0.97
1.32
1.03
1.03
0.94
1.20
1.30
1.16
0.97
0.89
3.94
1.07
2.55
1.10

0.90
1.01
1.23
1.10
1.11
1.03
1.40
0.40
1.05
1.80
1.12
3.10
1.17
1.23
1.20

30
5.71
1.21
21.66
6.25

Results presented in IU/g, unless otherwise noted.

NR, not reported.
N, number of samples.

30
4.20
1.58
37.7
3.33

30
1.31
0.71
54.51
0.83

Table 4. Laboratories eliminated from statistical

evaluation of LC method for determination of vitamin D
in infant formulas and enteral products
Sample
C
F
B
G
D
I
A
J
K
E
H
a
b
c

Outlier laboratory
3a
3 , 11a, 15b
11c, 12c
3a
3a
1b

3a
c
12 , 14c
4c
c
12 , 14c
a

Laboratory eliminated due to lack of duplicates.

Cochran outlier.
Grubbs outlier.

Table 5. Statistics of LC method for determination of

vitamin D in infant formulas and enteral products
Sample type

Na

Mean RSDr, %

Mean RSDR, %

All
Powders
Liquids
Infant formulas
Enteral products

11
3
8
8
3

9.34
13.56
7.76
7.88
13.24

13.41
17.67
11.81
12.30
16.36

N, number of samples.

Table 6. Comparison of the results of vitamin D

analysis in infant formulas by LC method and rat
bioassay
Concentration determined by
Sample
A
B
C
D
E
F
G
H
I
J
K
a
b

LC methoda

Rat bioassay

Unitsb

360
585
1090
488
396
984
495
1090
473
346
569

328
434
>745c
433
322
>550c
467
938
433
212
508

IU/QND
IU/QND
IU/QND
IU/QND
IU/g 100
IU/QND
IU/QND
IU/g 1000
IU/QND
IU/QND
IU/g 100

Results presented after elimination of outliers.

Concentration of vitamin D in IU/QND for liquid formulas [IU =
International Unit (1 mcg vitamin D = 40 IU); QND = quart normal
dilution (1 qt = 0.946 L)], and in IU/g for powders.
Values not included in Figure 1.

Table 7. Comparison of the results from previous

studies of LC methods (6) for determination of vitamin D
in infant formulas versus the results from the
collaborative study of AOAC Official Method 995.05
Method
a

1
a
2
995.05
a

RSDr, %

RSDR, %

No. of
samples

No. of
laboratories

17.7
30.2
7.88

26.6
30.2
12.30

3
3
8

12
4
15

Numbers 1 and 2 represent the methods reported in J. AOAC Int.

76, 1042 (1993).

Figure 1. Comparison of the results of vitamin D

analysis in infant formulas by LC method and rat
bioassay.