Food
Elemental Analysis Manual: Section 4.3A Graphite Furnace Atomic Absorption Spectrometric Determination of Cadmium and Lead in Food Using Microwave Assisted Digestion
Appendix A - Supplemental Information on In-house Method Validation
Version 1 (June 2008)
Authors: William R. Mindak
William C. Cunningham
In-house validation of EAM Method 4.3 (draft E, January 2002) was performed using a CEM Corporation model MDS 2000 microwave digestion system with UDV vessels and a PerkinElmer 5100PC atomic absorption spectrometer equipped with a 5100 ZL furnace module (transverse heated graphite furnace), end-capped graphite tubes and AS71 autosampler.
4.3A.1 ANALYTICAL LIMITS
Analytical limits were determined by the analysis of fortified method blanks (FMBs). Method blanks were fortified at a level estimated to be approximately 3-5 times the detection limit. Results are summarized in 4.3A Table 1. The characteristic mass, m0, for cadmium was 1.1 pg and for lead was 25 pg.
4.3A Table 1. Estimate of Analytical Limits
| Cadmium | Lead | |
|---|---|---|
| Number of FMBs | 10 | 10 |
| Fortification level (µg/L) | 0.05 | 1 |
| Standard deviation (µg/L) | 0.00363 | 0.0911 |
| ASDL (µg/L) | 0.014 | 0.35 |
| ASQL (µg/L) | 0.11 | 2.7 |
| LOD (µg/kg) based on 1 g anal. portion | 0.35 | 8.8 |
| LOD (µg/kg) based on 4 g anal. portion | 0.087 | 2.2 |
| LOQ (µg/kg) based on 1 g anal. portion | 2.7 | 69 |
| LOQ (µg/kg) based on 4 g anal. portion | 0.68 | 17 |
4.3A.2 REFERENCE MATERIAL RESULTS
Three replicates of several reference materials (RMs) were prepared for the method validation of accuracy. Analyte concentrations were determined using both standard curve and standard additions procedures. When using the standard curve procedure, the required dilution of the analytical solution was based on recovery of the fortified analytical solution (FAS). Matrix interference was assumed if the FAS recovery was <90%, which necessitated increasing the dilution factor. Standard addition results compared favorably with standard curve results. A moisture correction factor was determined for each RM and applied to the results, which are summarized in 4.3A Table 2.
4.3A Table 2. Reference Material Results
| Cadmium | Lead | |||||
|---|---|---|---|---|---|---|
| Reference Materiala | Ref. Value (mg/kg) | Meanb Result (%) | RM Rec. (%) | Ref. Value (mg/kg) | Meanb Result (%) | RM Rec. (%) |
| Dogfish Muscle (NRCC DORM-1) | 0.086 | 0.093 | 108 | 0.4 | 0.43 | 107 |
| Spinach (NIST 1570) | - | - | - | 1.2 | 1.13 | 94 |
| Mussel (NIES-6) | - | - | - | 0.91 | 0.84 | 92 |
| Bovine Liver (NIST 1577) | 0.27 | 0.29 | 107 | 0.34 | 0.37 | 109 |
| Oyster Tissue (NIST 1566) | 3.5 | 3.09 | 88 | 0.48 | 0.5 | 104 |
| Cocoa Powder (FDA CP) | 0.39 | 0.371 | 95 | - | - | - |
| Wheat Flour (NIST 1567) | 0.032 | 0.032 | 100 | - | - | - |
| Rice Flour (NIST 1568a) | 0.022 | 0.021 | 95 | - | - | - |
a NRCC=National Research Council of Canada, NIST=National Institute of Standards and Technology, NIES=National Institute for Environmental Studies of Japan, FDA=Food and Drug Administration.
b n=3; corrected for moisture.
Cadmium
Six RMs were analyzed to evaluate method performance accuracy for cadmium (4.3A Table 2). These RMs provided a variety in matrix (vegetative, organ meat and animal protein) and cadmium concentration (0.022-3.5 mg/kg). The dilution factor required for the analytical solution varied from 1 (Dogfish Muscle, Wheat and Rice Flours) to 50 (Oyster). The dilution factors were based on cadmium concentration and not interference. FAS recoveries were acceptable for the dilutions used. Results were within the reference value uncertainty except for Oyster, which was slightly below the uncertainty window, but was recovered at 88%.
Lead
Five RMs were analyzed to evaluate method performance accuracy for lead (4.3A Table 2) These RMs provided a variety in matrix (vegetative, organ meat and animal protein) and lead concentration (0.34-1.2 mg/kg). The dilution factor (based on FAS recovery) required for the analytical solution varied from 1 (Oyster and Bovine Liver) to 4 (Spinach). The Spinach exhibited unacceptably low FAS recoveries and low recovery based on reference values at dilution factors less than 4. The Spinach reference value recoveries for dilution factors of 2 and 3 were 42% and 75%, respectively. All results were within the reference value uncertainty except for Mussel, which was slightly below the uncertainty window, but was recovered at 93%.
4.3A.3 FOOD RESULTS
Fourteen foods were selected from the FDA Total Diet Study market basket 1999-1 and analyzed to evaluate method performance accuracy and precision (see 4.3A Table 3). Foods were chosen based on analytical challenge (either sample preparation or analysis) or analyte level. Two replicate portions of both unfortified and cadmium and lead fortified portions were prepared and analyzed. Analytical portions were limited by the energy content of the food but with an upper limit of 5 g to limit dilution of the nitric acid and ensure a complete digestion. Portions varied from approximately 0.5 g for high-fat foods like nuts to 5 g for high-water content foods like pickles. 4.3A Table 3 lists the average mass used for each food. TDS results were obtained using a dry ashing mineralization and measurement of lead and cadmium by graphite furnace atomic absorption spectrometry1.
4.3A Table 3. Foods and Analytical Portion Mass for In-House Validation
| Food | Portion (g) | Food | Portion (g) |
|---|---|---|---|
| American cheese | 0.87 | Iceberg lettuce | 3 |
| Beef liver | 1.4 | Canned spaghetti | 2.5 |
| Peanut butter | 0.47 | Dill pickles | 5 |
| White bread | 1.1 | Catsup | 2.8 |
| Raisin bran cereal | 0.85 | Mixed nuts | 0.51 |
| Canned fruit cocktail | 3.8 | Canned peaches | 4.5 |
| Spinach | 4.4 | Yellow mustard | 3.3 |
Cadmium
The results for cadmium are listed in 4.3A Table 4. Cadmium results ranged from 1 µg/kg in canned fruit cocktail to 170 µg/kg in spinach. All results were >LOQ. Agreement between duplicates was generally good with a few exceptions. American cheese, iceberg lettuce and mixed nuts had relatively high relative percent differences (RPD) but only American cheese was greater than the control limit of 20%. American cheese was challenging because the cadmium level was near the LOQ and the cheese has a very high mineral content. Even though there was an acceptable FAS recovery for American cheese, a matrix effect (peak broadening) was noted which probably contributed to the high RPD.
The required dilution factors (DFs) based on FAS recovery are listed in 4.3A Table 4. Foods highest in salt had the greatest matrix effect indicated by the low FAS recovery (<90%) when analyzed without further dilution. 4.3A Table 5 lists results for three high-salt foods at various dilutions and illustrates the utility of using the FAS recovery to detect interference and determine the necessary dilution to lower matrix interference to an acceptable level. Without the necessary dilution factor the results would be biased low. The high dilution factor for spinach was the result of the high cadmium concentration and not matrix interference.
Results were in good agreement with TDS results. The largest difference was for raisin bran cereal, which was probably the result of nonhomogeneity.
4.3A Table 4. Cadmium in Food Results
| Founda | LOQ | RPD | Fortification | Recoverya | TDS Resultb | ||
|---|---|---|---|---|---|---|---|
| Food | (µg/kg) | (µg/kg) | DF | (%) | (µg/kg) | (%) | (µg/kg) |
| American cheese | 3 | 3.0 | 1 | 26.9 | 14 | 100.2 | <10 |
| Beef liver | 37 | 1.9 | 1 | 4.2 | 34 | 90.2 | 36 |
| Peanut butter | 87 | 5.8 | 1 | 0.5 | 106 | 99.9 | 79 |
| White bread | 20 | 2.5 | 1 | 7.8 | 23 | 95.6 | 14 |
| Raisin bran cereal | 43 | 3.2 | 1 | 1.4 | 59 | 96.1 | 33 |
| Canned fruit cocktail | 1 | 0.7 | 1 | 9.2 | 7 | 94.8 | <5 |
| Spinach | 170 | 6.2 | 10 | 9.4 | 234 | 97.9 | 176 |
| Iceberg lettuce | 18 | 0.91 | 1 | 12.2 | 17 | 94.1 | 15 |
| Canned spaghetti | 12 | 1.1 | 1 | 2.3 | 20 | 96.7 | 10 |
| Dill pickles | 5 | 2.2 | 4 | 6.8 | 5 | 91.1 | <7 (6) |
| Catsup | 24 | 2.9 | 3 | 7.5 | 18 | 98.6 | 15 |
| Mixed nuts | 20 | 5.3 | 1 | 11.6 | 48 | 84.2 | 20 |
| Canned peaches | 1 | 0.61 | 1 | 6.5 | 11 | 95.0 | <5 |
| Yellow mustard | 34 | 3.3 | 4 | 5.6 | 40 | 109.2 | 28 |
| Mean: | 8.0 | 96.0 | |||||
a Results are the mean of two replicate analytical portions. Results displayed in parentheses are trace levels <LOQ.
b Foods are from FDA Total Diet Study (TDS) market basket 1999-1. Results displayed with "<" are less than the LOQ; the LOQ is provided and a trace level is provided in parentheses if analyte was detected. LOQ = 10xσ.
4.3A Table 5. Cadmium Results for High-Salt Foods at Various Dilutions
| DF =1 | DF =2 | DF =4 | ||||
|---|---|---|---|---|---|---|
| Food | FAS Rec. (%) | Found (µg/kg) | FAS Rec. (%) | Found (µg/kg) | FAS Rec. (%) | Found (µg/kg) |
| Dill pickles | 13 | 1 | 58 | 3 | >90 | 5 |
| Catsup | 57 | 0 | 51 | 22 | >90 | 24a |
| Yellow mustard | 0 | 23 | 52 | 31 | >90 | 34 |
a DF=3
Lead
The results for lead are listed in 4.3A Table 6. Result for canned peaches was 15 µg/kg. All other results were trace levels <LOQ. Agreement between duplicates was generally good with a few exceptions. American cheese was challenging because of very low lead level and a very high mineral content, which resulted in some matrix interference. A dilution factor of 4 was required for lead in mustard and pickles, which resulted in a very low lead level in the analytical solution. The 41% difference between raisin bran cereal duplicates for lead was most likely due to nonhomogeneity (inconsistent ratio of raisins to cereal in the analytical portions even though the sample was processed through a food mill) and the low levels measured (<LOQ).
The required DFs based on FAS recovery are listed in 4.3A Table 6. Foods highest in salt had the greatest matrix effect and thus low FAS recovery (<90%) when analyzed without further dilution. 4.3A Table 7 lists results for three high-salt foods at various dilutions and illustrates the utility of using the FAS recovery to detect interference and determine the necessary dilution to lower matrix interference to an acceptable level. Without the necessary dilution factor, the results would be biased low.
The levels of lead were quantifiable in 10 of the foods but TDS results were all below LOQ. Results were generally higher than TDS results when considering TDS LOQ and trace results. For example, the level of lead found in dill pickles, catsup, mixed nuts, and yellow mustard are above the TDS LOQ but the TDS results were nonquantifiable. The disagreements with TDS results could be due to issues related to homogeneity or interferences compensation based on FAS recovery.
4.3A Table 6. Lead in Food Results
| Founda | LOQ | RPD | Fortification | Recoverya | TDS Resultb | ||
|---|---|---|---|---|---|---|---|
| Food | (µg/kg) | (µg/kg) | DF | (%) | (µg/kg) | (%) | (µg/kg) |
| American cheese | (76) | 237 | 3 | 16 | 283 | 95.5 | <50 |
| Beef liver | (44) | 49 | 1 | 6.5 | 171 | 92.8 | <50 (18) |
| Peanut butter | (55) | 145 | 1 | 8.1 | 531 | 89.4 | <50 |
| White bread | (38) | 124 | 2 | 10 | 233 | 88.7 | <30 |
| Raisin bran cereal | (53) | 160 | 2 | 41 | 295 | 96.6 | <30 (15) |
| Canned fruit cocktail | (13) | 18 | 1 | 5.8 | 66 | 94.8 | <20 (13) |
| Spinach | (23) | 31 | 2 | 0.6 | 58 | 96.7 | <20 (13) |
| Iceberg lettuce | (9) | 24 | 1 | 2.5 | 84 | 97.3 | <20 |
| Canned spaghetti | (17) | 55 | 2 | 4.9 | 100 | 96.6 | <20 (9) |
| Dill pickles | (54) | 55 | 4 | 27 | 49 | 95.2 | <30 (29) |
| Catsup | (43) | 98 | 4 | 4.2 | 92 | 93.3 | <30 |
| Mixed nuts | (64) | 130 | 1 | 11 | 478 | 90.9 | <50 (16) |
| Canned peaches | 15 | 15 | 1 | 2.5 | 55 | 91.9 | <20 (13) |
| Yellow mustard | (42) | 83 | 4 | 28 | 80 | 84.6 | <20 |
| Mean: | 12.1 | 93.2 | |||||
a Results are the mean of two replicate analytical portions. Results displayed in parentheses are trace levels <LOQ.
b Foods are from FDA Total Diet Study (TDS) market basket 1999-1. Results displayed with "lt;" are less than the LOQ; the LOQ is provided and a trace level is provided in parentheses if analyte was detected. LOQ = 10xσ.
4.3A Table 7. Lead Results for High-Salt Foods at Various Dilutions
| DF =1 | DF =2 | DF =4 | ||||
|---|---|---|---|---|---|---|
| Food | FAS Rec. (%) | Found (µg/kg) | FAS Rec. (%) | Found (µg/kg) | FAS Rec. (%) | Found (µg/kg) |
| Dill pickles | 43 | 9 | 50 | 18 | >90 | 54 |
| Catsup | 33 | 7 | 60 | 22 | >90 | 43 |
| Yellow mustard | 28 | 4 | 38 | 15 | >90 | 42 |
4.3A.4 CONCLUSION
The method is applicable to the analysis of food for cadmium and lead. A single microwave digestion program can be utilized by varying the analytical portion based on energy content. Foods of different nutritional composition (fat, protein, carbohydrate) can be digested in a single microwave carousel. Analyte recovery in the fortified analytical solution aids in determining the dilution necessary to minimize matrix interference. This interference might otherwise go unnoticed. Method performance was validated by the analysis of food type RMs and recovery of analyte from consumer foods. Recovery was acceptable in both the reference materials (mean 99% for cadmium and 101% for lead) and foods (mean 96% for cadmium and 93% for lead). By carefully adhering to the QC requirements and analyzing analytical solutions at dilutions dictated by the FAS recovery, accurate results can be obtained. Quality control analyses are necessary to ensure data quality.
REFERENCES
- U.S. Food and Drug Administration (2005) (INTERNET) FDA Total Diet Study Home Page [cited 26 February 2010].
Elemental Analysis Manual
Elemental Analysis Manual: Section 4.3 Graphite Furnace Atomic Absorption Spectrometric Determination of Cadmium and Lead in Food Using Microwave Assisted Digestion Elemental Analysis Manual: Section 4.3B Graphite Furnace Atomic Absorption Spectrometric Determination of Cadmium and Lead in Food Using Microwave Assisted Digestion Elemental Analysis Manual: Section 4 Analytical Methods Elemental Analysis Manual (EAM)
