Component Description
Measurement of concentrations (ng/mL) of aflatoxin B1-lysine, derived from aflatoxin B1-albumin adducts, in a 2,104 serum sample subset of NHANES 1999-2000.
Eligible Sample
A one-third subset of examined participants ≥ 12 years of age from NHANES 1999-2000.
Description of Laboratory Methodology
Measurement of serum aflatoxin-albumin adducts
Modification of the method of McCoy et al. (2005) was used to assess aflatoxin-albumin adducts in serum. NHANES serum specimens were obtained from McKesson Plasma and Biologics (San Francisco, CA) where they were stored at –70°C; specimens were residual sera that were previously freeze/thawed one or more times. In brief, the laboratory used a high throughput method in which 250µL serum was robotically aliquoted into 96-well plates and spiked with 100µL of tetradeuterated internal standard (aflatoxin B1-lysine-d4) to account for recovery. Aflatoxin B1 adducts were hydrolyzed from serum albumin using enzymatic digestion with pronase which was incubated in a 40°C water bath for 4 hours or overnight at 37°C. Once digestion was complete, samples were transferred to an automated solid phase extraction (SPE) system.
The method employed a mixed-mode anion exchange sorbent in an automated SPE of AFB1-lysine (AFB-lys) resulting from the enzymatic hydrolysis of serum albumin. SPE columns were eluted with 2% formic acid in methanol. An isotope dilution mass spectrometry based method (LC-MS/MS triple quadrupole MS using electrospray ionization) was used for routine quantitation of AFB-lys as a measure of individual exposure to aflatoxin B1.
Concentrated samples were injected onto a C18 HPLC chromatographic column and eluted with a gradient of acetonitrile in water with 0.06% formic acid to resolve the AFB-lys from other matrix constituents. Analytes were detected using positive ionization selective reaction monitoring mode and quantified using 4-point linear calibration curves that used 1/x2 weighting; quantitation was based on peak area ratios using the internal standard to correct for recovery. Peak area ratios (aflatoxin B1-lysine m/z: aflatoxin B1-lysine-d4 m/z) were linearly related to concentration. Instrument voltages and temperatures were optimized to yield optimum intensity of AFB-lys, as previously published. AFB-lys concentrations were quantitated in a subset of NHANES 1999-2000 serum specimens in 62 separate assays.
Transitions for aflatoxin B1-lysine (AFB-lys) and deuterated internal standard
| Analyte |
Q1 (Da) |
Q3 (Da) |
| Aflatoxin B1-lysine |
457.3 |
394.2 |
| Aflatoxin B1-lysine-d4 |
461.3 |
398.3 |
The average (±SD) calibration slope for assays containing NHANES samples was 1.1±0.2 and average intercept was 0.0047±0.016. Calibration curves had an average (±SD) r2 of 0.999±0.008. Calibrators were required to be within 15% of their nominal values for all assays in which results above the lower limit of detection were reported. Two different calibration ranges were used; the lower range was used in the majority of the NHANES analyses.
Four calibration standards (ng/mL) used in 62 assays to measure serum concentrations of aflatoxin B1-lysine (AFB-lys) in NHANES 1999-2000
| No. Assays |
Calibrator 1 |
Calibrator 2 |
Calibrator 3 |
Calibrator 4 |
| 6 |
0.07 |
0.34 |
0.68 |
2.73 |
| 56 |
0.01 |
0.05 |
0.5 |
1.5 |
Blanks were analyzed concurrently with unknown samples and were used to determine if any analyte contamination had occurred. Blanks contained 250µL 25% methanol in place of serum; one blank was included in each assay. In none of the 62 assays containing NHANES samples, was AFB-lys detectable in the blank.
Relative retention times were used to select the correct peaks for integration and quantitation. AFB-lys eluted 0.02±0.01 min after the internal standard in the QC pools which were known to contain aflatoxin B1-albumin adducts and similarly, 0.02±0.02 min after the internal standard in the unknown NHANES samples. Peaks greater than 0.06 minutes from the internal standard were not considered to be representing AFB-lys.
Limit of Detection (LOD)
Using a combination of approaches, the lower limit of detection (LOD) was determined to be 0.02 ng/mL. Results near the LOD are subject to greater uncertainty. Several schemes were used to estimate the LOD including 1) Taylor series to estimate the SD of a zero sample via extrapolation of a linear dilution series to zero concentration, 2) visual evaluation of chromatography around the LOD, and 3) estimation of the LOD by calculating 3 times the s0 where s0 is the standard deviation of the concentration of a near-background sample that was measured multiple times.
Summary of activities to estimate the lower limit of detection (LOD) for AFB-lys in serum using a LC-MS/MS assay
| Date of Analysis |
LOD Scheme |
Design |
Estimate of LOD (ng/mL) |
| Aug-06 |
Taylor series (3*SD of a zero sample) |
5 dilutions, singlicates per dilution over 3 days (15 points) |
0.018 |
| Feb-07 |
Taylor series (3*SD of a zero sample) |
3 dilutions, 5 replicates per dilution over 3 days (45 points); also evaluated chromatography for appropriateness |
0.018 |
| Jan-08 |
Chromatographic evaluation |
Evaluated 7 chromatograms at about 0.02, 0.48, 1.5 ng/mL for appropriateness of 0.02 ng/mL |
0.02 |
| Aug-10 |
3*SD of a sample close to zero |
Nearly zero sample, singlicate over 3 days (3 points) |
0.04 |
| Feb-12 |
Chromatographic evaluation |
Evaluated 27 chromatograms =LOD from NHANES and dozens |
0.02 |
Standard stability
A concentrated stock standard solution stored at -70°C was diluted to an intermediate concentration that was further diluted and used to prepare individual calibrators. The highly concentrated stock material, synthetic AFB-lys, was purified using semipreparative HPLC-UV. Stability of this stock solution over the course of 7 years was periodically determined. Degradation was no more than 2%. Calibrators were adjusted accordingly after each determination.
Stability of the concentrated stock of AFB-lys stored at -70°C
| Date of analysis |
[AFB-lys] µg/mL |
| Jul-04 |
4.23 |
| Oct-07 |
4.19 |
| Jun-11 |
4.17 |
To test for stability of the intermediate stock standard solutions that were diluted from the concentrated stock, we stored materials at -20°C or -70°C for up to 2 years. We demonstrated complete stability at -70°C.
Stability of the intermediate stocks of AFB-lys stored at -20°C or -70°C
| Initial[AFB-lys] ng/mL |
Storage temperature |
Storage time |
Stability |
| 50 |
-20°C |
3 mo |
90% |
| 4 |
-70°C |
2 yr |
100% |
Serum stability
To test for stability of aflatoxin-albumin adducts in serum, high (HS04501) and low (LS04500) bench QC pools were stored at variable temperatures and times; AFB-lys concentrations were compared to concentrations in pools that were continuously stored at -70°C for 3.3 years. Pools stored at -20°C or 4°C for 1.6 years, or 37°C for 0.25 years showed no evidence of degradation. QC serum maintained at room temperature for 16 h prior to sample processing for quantitative analysis showed no evidence of degradation.
QC pools were also submitted to three freeze/thaw cycles and subsequently analyzed within a single analytical run. On average, there appeared to be a small decrease in the AFB-lys concentrations (average 97±2% of expected result) over the course of three cycles but all concentrations from freeze-thaw samples were within QC limits thus indicating no significant decrease could be established.
From the literature, Scholl and Groopman reported that using two closely comparable chromatographic methods (LC-MS/MS and HPLC-fluorescence), a set of 19 serum specimens with AFB-lys concentrations ranging from 0.3 to 6 ng/mL stored for 15 years at -80°C, gave AFB-lys results that were essentially unchanged; currently, this appears to be the longest term demonstration of aflatoxin-albumin adduct stability in serum.
Reproducibility
Approximately 75 percent of the NHANES samples (1,572/2,104) were tested once for AFB-lys. Considering all NHANES samples, of those with AFB-lys results greater than or equal to the LOD (27/2,104), 93 percent (25/27 cases) were measured more than once to attempt to confirm the finding; only those with insufficient serum were not retested (2/27 cases). Because most positive results were around the LOD and because of the inherent variability of results around the LOD (>30% CVanalytical), we expected inconsistency in repeating and confirming results. Confirmation was considered achieved if a subsequent result was ≥LOD. We defined AFB-lys-positive samples having at least one result that conformed to the relative retention time requirement for AFB-lys peak selection (see earlier). Thus, not all AFB-lys-positive results were confirmed, however, all of the highest results (>0.05 ng/mL) were repeated and confirmed (see following table).
Repeat analysis of 27 AFB-lys-positive samples grouped by concentration at or grouped above the LOD (0.02 ng/mL)
| [AFB-lys] ng/mL |
N |
≥LOD Confirmed |
| 0.02 |
7 |
1/7 (14%) |
| 0.03 - 0.05 |
15 |
8/15 (53%) |
| 0.06 - 0.20 |
5 |
5/5 (100%) |
Laboratory Quality Assurance and Monitoring
The laboratory and method were certified according to the Clinical Laboratory Improvement Amendments (1988) guidelines.
Units of serum were obtained from Tennessee Blood Services (Memphis, TN) and were screened using the methodology described above; all units of serum that were screened exhibited undetectable AFB-lys. All quality control pools were prepared by spiking human serum with serum from rats that were dosed with aflatoxin B1 to create aflatoxin B1-albumin adducts. Targeted concentrations of AFB-lys were estimated and confirmed before dispensing the blended serum pools into 2-mL cryovials for storage at -70°C.
Before each analytical run, individual calibration standards, bench QC materials and serum samples were thawed from -70°C to room temperature. Calibrators and QC materials were prepared and analyzed in parallel with the NHANES samples.
Typical assays consisted of 4 calibration samples, 1 blank, 3 bench quality control samples, 53 subject person’s serum followed by 3 bench quality control samples. One in every 20 subject person’s serum in the assay was a random blind QC pool.
Blind Quality Control
Three blind QC pools were used during the analysis of the survey specimens. Their results were required to be within 3 standard deviations of the characterization means. At the end of the study, summary statistics were tabulated for the blind QC pools that were included in the 62 assays from this study.
The following two tables show the mean, standard deviation of individual results (SDi) and analytical coefficient of variation (CVa) for aflatoxin B1-lysine concentrations (ng/mL) from individual blind quality control pools during the characterization period and during the period of analysis of NHANES samples
Blind quality control pools during the characterization period
| Blind QC Pools |
Mean |
SDi |
CVa |
| High |
4.22 |
0.34 |
8% |
| Medium |
1.55 |
0.12 |
8% |
| Low |
0.49 |
0.06 |
11% |
Blind quality control pools during NHANES Analysis
| Blind QC Pools |
Mean |
SDi |
CVa |
| High |
4.22 |
0.35 |
8% |
| Medium |
1.53 |
0.01 |
9% |
| Low |
0.48 |
0.04 |
9% |
Bench Quality Control
Five bench QC pools, prepared as described above, were characterized to determine the quality control limits for duplicate samples. After establishment of the control limits of the pools, QC samples contained within each analytical run were evaluated for validity by use of a multi-rule quality control SAS program. Each analytical run contained 2-3 pools in duplicate.
The following two tables show the mean, standard deviation of pair mean (SDm) and analytical coefficient of variation (CVa) for aflatoxin B1-lysine concentrations (ng/mL) in bench quality control pools during characterization period and during the period of analysis of NHANES samples
Bench quality control pools during the characterization period
| Bench QC Pools |
Mean |
SDm |
CVa |
| HS04501_QC03S |
1.63 |
0.11 |
7% |
| HS07502_QC01S |
1.03 |
0.05 |
5% |
| MS07501_QC01S |
0.65 |
0.03 |
5% |
| LS04500_QC03S |
0.26 |
0.02 |
8% |
| LS07500_QC01S |
0.2 |
0.02 |
8% |
Bench quality control pools during NHANES Analysis
| Bench QC Pools |
Mean |
SDm |
CVa |
| HS04501_QC03S |
1.65 |
0.13 |
8% |
| HS07502_QC01S |
1.03 |
0.1 |
10% |
| MS07501_QC01S |
0.64 |
0.04 |
6% |
| LS04500_QC03S |
0.26 |
0.03 |
10% |
| LS07500_QC01S |
0.2 |
0.02 |
11% |
External Quality Assurance
As an external quality assurance activity, a sample exchange was conducted to assess the comparability of three different methods for measuring serum AFB-lys concentrations. Published results showed that isotope dilution LC-MS/MS was the most sensitive technique and HPLC-fluorescence was the least sensitive method. Overall, this exchange showed excellent correlations among the three independent methodologies, including ELISA, which were conducted in different laboratories. The two chromatographic methods showed satisfactory agreement (LC-MS/MS = HPLC-fluorescence ÷ 0.71). In summary, the results supported the validation of these technologies for assessment of human exposure to this environmental toxin and carcinogen.
Data Processing and Editing
Data were received after all the laboratory testing was complete. The data were not edited for extreme values.
Data Access: All final data are publicly available.
Analytic Notes
None.