Author: Reames, Ginger; Charlton, Valerie
Date published: January 1, 2013
Journal code: PENV
Assessing lead in paint, dust, and soil is an important element of every childhood lead poisoning case investigation. The Centers for Disease Control and Prevention (CDC) have stated that "the most common sources of high-dose lead exposure for U.S. children are lead-based paint and lead-contaminated house dust and soil (CDC, 2005)." Handheld X-ray fluorescence (XRF) instruments have become one of the essential components of lead exposure assessment, due to the advantages of the technology: rapid, on-site screening capability, nondestructive testing, and immediate results.
The California Department of Public Health (CDPH) Childhood Lead Poisoning Prevention Branch (CLPPB) has previously validated the use of portable XRFs for lead in paint testing and lead in soil testing (Reames & Lance, 2002) and has developed a series of guidance documents with XRF testing procedures for local environmental health jurisdiction staff. CLPPB administers an XRF loan program that makes Niton model XLp303A instruments available to local county and city health jurisdictions throughout California. These instruments have been approved by the U.S. Environmental Protection Agency (U.S. EPA) and the U.S. Department of Housing and Urban Development (HUD) for leadbased paint testing (U.S. EPA, 2004).
Attempting to identify as many of the sources of lead exposure as possible is critical for a childhood lead poisoning case in order to reduce or eliminate all of the sources. Portable XRF instrument usage has allowed California lead programs to quickly inform the family of a child lead poisoning case of lead exposure hazards in paint, dust, and soil. Although established state (California Department of Public Health [CDPH], 2008) and federal guidance documents (Lead, 2001; U.S. Department of Housing and Urban Development [HUD] , 1995) are available on field sampling and identification of housing-based lead hazards, guidance is lacking on using XRFs to screen nonhousing items.
Discussions with Food and Drug Administration (FDA) officials and review of their previous work with XRF analysis led CLPPB to conduct a study to determine if the onsite screening capability of the CLPPB XRF instruments could be expanded to include food, medicinal, and similar relatively homogeneous samples. Staff affiliated with FDA have previously highlighted the suitability of XRF for rapid screening of toxic elements such as lead (Palmer, Jacobs, Baker, Ferguson, & Webber, 2009), and issued a laboratory information bulletin on the same topic for FDA staff (Palmer, Webber, Ferguson, & Jacobs, 2007). Some prior studies that concentrated on specific foods have also been performed, such as spices (Al-Bataina, Maslat, & Al-Kofahi, 2003) and Indian spices and cultural powders (Lin, Schaider, Brabander, & Woolf, 2010). Other than the FDA evaluations, however, previous XRF studies have involved the use of bench-top laboratorygrade XRF instruments.
CLPPB undertook an evaluation to focus on a handheld XRF that could easily be brought into homes and used both for testing lead-based paint and other environmental hazards and for screening lead in food and medicinal items. CLPPB's objective was to confirm that the XRF used for identifying environmental lead hazards could also be used as a screening tool for food, medicinal, and ceremonial items. If the XRF was found to be a valid screening methodology, the future addition of on-site food and medicinal item testing would be a major advance in rapid evaluation of suspect lead exposure sources for lead-poisoned children.
Sample Collection and Testing
Fifty-eight test samples were acquired from seven local California childhood lead programs and through childhood lead poisoning investigations by CLPPB staff over a 10-month period (June 2009 through March 2010). The majority of samples were collected because they were suspected lead exposure sources. Some additional samples were purchased at ethnic grocery stores. These samples were similar to those recalled by FDA due to food labeling concerns (FDA, 2008) and those tested in a prior lead in spices study (Lin et al., 2010). Samples consisted of seven general categories: imported candies, miscellaneous baby products (powder, lotion, and astringent), imported retail medicines, home remedies, tea, foods and spices, and ceremonial items (powders, herbs, incense, and camphor associated with Hindu worship). Foods and spices constituted the majority of samples tested (Figure 1).
The lead content of the samples included in our study was unknown prior to testing. Since the goal of testing was to develop a rapid screening method, the samples were not altered by grinding or other practices used by laboratories to further homogenize a sample. The majority of samples were tested in the original packaging, plastic bags, and XRF test cups. When initial testing suggested that lead-containing ink might be present in the packaging, some items were removed from the packaging and evaluated separately. Some bulk powders and liquids were transferred to an XRF test cup with a thin Mylar covering. An instrument quality control sample, consisting of a National Institute for Standards and Technology (NIST) bulk lead in soil standard, was tested prior to XRF testing of samples. XRF sample testing proceeded only if the quality control XRF result was within plus or minus 10% of the NIST lead value in parts per million (ppm).
After XRF testing, samples were submitted to either a commercial laboratory or the CDPH Environmental Health Laboratory Branch (CDPH EHLB) following appropriate chain-of-custody procedures. Five samples that appeared to be low in lead when tested by XRF were sent to the CDPH EHLB for analysis, which employs a more rigorous modified sample analysis technique that results in lower detection limits. An imported candy sample was sent from CDPH EHLB to the CDPH Food and Drug Branch to determine if the sample exceeded the California Health and Safety Code Section 110552 level of 0.1 ppm, the California standard for lead in candies (CDPH, 2005). The majority of the samples were analyzed using U.S. EPA reference method 3050B/7421 (graphite furnace atomic absorption). The remaining samples that contained higher concentrations of lead were analyzed using U.S. EPA reference method 3050B/7420 (flame atomic absorption).
Data Analysis Plan
Review of previous findings by FDA indicated that quantitative XRF results derived from proprietary algorithms may not always be accurate due to a variety of factors including sample homogeneity, sample density, depth of the XRF readings, interferences from unexpected elements, the exact focal point of the X-ray beam, and limitations of preprogrammed algorithms used to calculate quantitative results. Therefore, two screening criteria were used to evaluate the Niton XRF lead detection results. Both criteria were based on a cut-point of 10 ppm, which is significantly lower than the lead-based paint standard of 5,000 ppm (Lead, 2001). This cut-point was thought to represent the lowest lead level for which results could reliably be obtained for the CLPPB model of XRF, based on prior FDA work and initial pilot testing by CLPPB.
The Niton XRF provides a test result for a given element in a bulk sample in two ways: a reading with units in ppm and a graph of the spectra of the elemental peaks that are present in the sample. For the first criterion of our evaluation, we hypothesized that if the laboratory sample result was >10 ppm lead then the unique spectral emissions produced by lead should be observed. Graphing software on the XRF and a companion proprietary PC software program (Thermo Scientific NDT® Software Suite) were used to observe whether both characteristic L-shell peaks of lead were present at specific energy levels: 10.5 kiloelectron volts (KeV) for the alpha peak and 12.6 KeV for the beta peak, respectively. For the second criterion, we hypothesized that if the laboratory sample result for lead was >10 ppm, then the XRF lead reading should also be >10 ppm.
Lead Content of Samples
Fifty-eight samples were included in the data analysis. These samples were first tested by XRF and then analyzed by the laboratory (Table 1). Both the Niton readings and the corresponding laboratory values were fairly low for the majority of samples, with 74% (43/58) of Niton readings reported as below method detection limits (<4.0 ppm-<17.3 ppm). Niton results that were preceded by a "less than" sign were classified as a nondetectable lead result. Seventy percent (41/58) of laboratory results were reported as below method detection limits (<0.1 ppm-<7.0 ppm). Detectable Niton lead readings ranged from 12.3 ppm to values that exceeded detector algorithm limits. Two samples were very high in lead, which caused the XRF to report values higher than one million ppm. This is due to an assumption made in the Niton analyzer's soil mode calibration that the sum total of metallic content is below 10% and that the metals present in the sample are not expected to affect one another. Since both samples contained over 10% lead, the numeric values provided by the analyzer exceeded the maximum value for ppm results, demonstrating that the XRF algorithm was oversaturated. Laboratory sample results above method detection limits ranged from 0.2 ppm to 340,000 ppm. One outlier was omitted from the data set because when tested by XRF, the sample appeared to have a very large amount of mercury (very large spectra peaks that overshadowed the area of the lead peaks) that caused the instrument to give a meaningless value (<91,018) for lead that could not be interpreted as being above or below the cut-point of 10 ppm.
Evaluation of Results Based on Screening Criteria
Laboratory sample results were grouped according to the cut-point of 10 ppm. Nine out of 58 samples (16%) were >10 ppm. Samples that exceeded the cut-point consisted of home remedies, imported retail medicine, and ceremonial items (Table 2). Based on the first criterion (confirmation of the presence of the alpha and beta L-shell elemental peaks of lead using spectra graphing software), lead peaks were observed for all samples that contained >10 ppm (9/9). Conversely, lead peaks did not appear to be present in all of the samples with laboratory results that were <10 ppm (49/49). This demonstrates that the Niton XRF could be used to reliably classify samples as having lead above the cut-point of 10 ppm by observation of the presence of lead peaks.
For the second criterion (quantitative agreement relative to the 10 ppm cut-point), all of the samples containing >10 ppm lead by laboratory analysis corresponded with Niton readings of >10 ppm (9/9). Six Niton XRF results, however, were >10 ppm for samples that were <10 ppm by laboratory analysis (false positive results). This finding demonstrates that the observation of the spectra is the most accurate means to determine whether lead is present in samples above 10 ppm.
Since the objective of the Niton evaluation study was to determine whether the instrument could be used as a field screening tool using the two screening criteria, no additional statistical analyses were performed. The Niton XRF detection limits were as much as two orders of magnitude greater than those obtained by the laboratory, particularly for samples with lead content below 10 ppm, which comprised the majority of samples in our study. The Niton detection limits were considerably higher than the laboratory detection limits in part because the XRF is a screening methodology and lacks the rigor of sample preparation, further homogenization, and acid digestion such as that employed in laboratory analysis methods for lead.
Our study, although limited in scope, demonstrates the potential food and medicinal item screening capability of XRFs such as the Niton XLp303A. The instrument consistently identified the presence of the characteristic lead spectra for samples with >10 ppm of lead (10 ppm cut-point). Lead spectra were absent for samples with <10 ppm of lead. Although the XRF is a screening methodology, instrument readings were >10 ppm for all laboratory sample results of >10 ppm. For some samples with laboratory results <10 ppm, however, the XRF readings were >10 ppm. These results were considered to be false positive results relative to the 10 ppm cut-point used for evaluation purposes. This supports the findings cited by FDA (Palmer et al., 2007; Palmer et al., 2009) that spectra results should be given precedence when determining if a metal such as lead is present in the sample. Given this limitation, it should be noted that the instrument appears to be able to correctly classify samples that may be a lead exposure concern (>10 ppm), which supports the public health goal of rapid screening to identify lead-contaminated food and medicinal items.
Use of this methodology will allow childhood lead poisoning investigators to quickly identify food and medicinal items with >10 ppm lead that may be contributing to a child's lead exposure. The main limitation of the XRF is that the limits of detection are not low enough to determine if a given food or medicine is not a significant contributor to a child's elevated blood lead level. A food item that is tested by XRF with a result that is reported as below the detection limit of <10 ppm could still contain enough lead to be a concern. If the food item actually contains 7 ppm (mcg/g) lead and the child eats a gram a day, this would exceed the FDA 6 meg per day provisional tolerable total daily intake for children (Provisional Tolerable Total Daily Intake for Children, 1993). Laboratory analysis is therefore still required for samples with low XRF lead screening results. The two strengths of XRF technology, however, are the ability to view and categorize the unique spectra of various elements and to provide immediate feedback to the lead-poisoned child's family regarding food and medicinal items with high lead-exposure potential.
As testing procedures were developed for our evaluation study, it became clear that this screening method requires operators to become experienced in spectra identification. It is essential for XRF users to learn to recognize the presence of elemental lead peaks. This can be learned by testing samples with known lead content and "blank" samples. It is also important for operators to recognize elements that can overlap with lead spectra, resulting in inconclusive or inaccurate results. The outlier sample that had a high level of mercury required the operator to readily identify that the sample could not be adequately characterized in the field.
Although a relatively small number of lead samples >10 ppm were tested (nine samples, 16% of the samples in the study) this reflects CLPPB experience of testing these types of items during childhood lead poisoning investigations. The majority of suspect items appear to have relatively low lead content, while occasionally a very high lead item is identified. The distribution of lead results in our study provided an opportunity to determine if the XRF was able to distinguish high lead-level items from among the suspect items tested. The capability demonstrated in our study shows that XRF results can inform the family of a lead-poisoned child so that they can immediately remove high lead-level items from their child's environment.
The usefulness of this methodology is best illustrated with an example from our study. Two different types of ceremonial cosmetic chalks brought from India were evaluated to determine if either one was a lead exposure source in a childhood lead poisoning case. One chalk was white, the other yellow. Neither chalk had any packaging or other information regarding the ingredients. The XRF quickly identified the yellow chalk as a very high lead exposure source (large elemental lead peaks and 414,000 ppm quantitative result). No lead peaks were confirmed for the white chalk, although a quantitative XRF result exceeded 10 ppm (34.5 ppm). The child's family was told to discontinue use of the chalks, pending the laboratory results. The laboratory results for these chalks were 340,000 ppm and 8.6 ppm, respectively.
Our study illustrates that it is possible to expand the capability of an XRF used to identify lead exposure hazards in paint, dust, and soil to screen food and medicinal items. Although the XRF has higher limits of detection than a laboratory, the instrument used in our study was able to consistently classify samples using a cut-point of 10 ppm. The effectiveness of this screening methodology is dependent on the operator's ability to discern the presence of the characteristic elemental lead peaks. Rapid identification of suspect lead exposure items enhances the ability of a childhood lead poisoning investigator to inform the family of immediate steps they can take to decrease their child's lead exposure.
Acknowledgements: CLPPB staff recognizes the assistance of Richard Jacobs, PhD, of the Food and Drug Administration, in providing technical guidance on the use of XRFs to screen food and medicinal items for lead. CLPPB also acknowledges the California Department of Public Health Environmental Health Laboratory Branch and Food and Drug Branch for assistance with sample characterization and analysis.
Lead Detection in Food, Medicinal, and Ceremonial Items Using a Portable X-Ray Fluorescence (XRF) Instrument
Al-Bataina, BA., Maslat, A.O., & Al-Kofahi, M.M. (2003). Element analysis and biological studies on ten oriental spices using XRF and Ames test. Journal of Trace Elemental Biology, 1 7(2), 85-90.
California Department of Public Health. (2005). Sherman food, drug, and cosmetic law, California Health and Safety Code, §110552.
California Department of Public Health. (2008). Accreditation, certification, and work practices for lead-based paint and lead hazards (California Code of Regulations, Title 17, Division 1, Chapter 8). Sacramento: Author.
Centers for Disease Control and Prevention. (2005). Preventing lead poisoning in young children. Atlanta: Author.
Food and Drug Administration. (2008). Raja Foods issues nationwide recall on Swad brand abil, gulal, kanku, kum kum, lagan samagri and pooja samagri containing high levels of lead. Retrieved from http://www.fda.gov/Safety/Recalls/ArchiveRecalls/2008/ucml 1 2341 .htm
Lead; Identification of Dangerous Levels; Final Rule (U.S. Environmental Protection Agency), 40 CER. Part 745, 66 (2001).
Lin, CG., Schaider, L.A., Brabander, DJ., & Woolf, A.D. (2010). Pediatric lead exposure from imported Indian spices and cultural powders. Pediatrics, 125(4) 828-835.
Palmer, P.T., Jacobs, R., Baker, P.E., Ferguson, K., & Webber, S. (2009). Use of field-portable XRF analyzers for rapid screening of toxic elements in FDA-regulated products. Journal of Agricultural andFood Chemistry, 57(7), 2605-2613.
Palmer, R, Webber, S., Ferguson, K., & Jacobs, R. (2007). On the suitability of portable X-ray fluorescence analyzers for rapid screening of toxic elements. Laboratory Information Bulletin, 4376, 1-15.
Provisional Tolerable Total Daily Intake for Children (U.S. Food and Drug Administration). Federal Register, 58, 33860 (1993).
Reames, G., & Lance, L.L. (2002). Childhood lead poisoning investigations: Evaluating a portable instrument for testing soil lead. Journal of Environmental Health, 64(8), 9-13, 25.
U.S. Department of Housing and Urban Development, Office of Lead Hazard Control. (1995). Guidelines for the evaluation and control of lead-based paint hazards in housing. Washington, DC: Government Printing Office.
U.S. Environmental Protection Agency. (2004). Performance characteristic sheet for Niton LLC model series XLp300 (1st ed.). Kansas City, MO: Midwest Research Institute.
Valerie Chariton, MPH, MD
Childhood Lead Poisoning
California Department of Public Health
Corresponding Author: Ginger Reames, Chief, Environmental Investigation Unit, Childhood Lead Poisoning Prevention Branch, California Department of Public Health, MS 7506, 850 Marina Bay Parkway, Building P, Third Floor, Richmond, CA 94804-6403. E-mail: Ginger.Reames@cdph.ca.gov.