The Use of Dried Blood Spots for Characterizing Children’s Exposure to Organic Environmental Chemicals

Abstract

Biomonitoring is a commonly used tool for exposure assessment of organic environmental chemicals with urine and blood samples being the most commonly used matrices. However, for children’s studies, blood samples are often difficult to obtain. Dried blood spots (DBS) represent a potential matrix for blood collection in children that may be used for biomonitoring. DBS are typically collected at birth to screen for several congenital disorders and diseases; many of the states that are required to collect DBS archive these spots for years. If the archived DBS can be accessed by environmental health researchers, they potentially could be analyzed to retrospectively assess exposure in these children. Furthermore, DBS can be collected prospectively in the field from children ranging in age from newborn to school-aged with little concern from parents and minimal risk to the child. Here, we review studies that have evaluated the measurement of organic environmental toxicants in both archived and prospectively collected DBS, and where available, the validation procedures that have been performed to ensure these measurements are comparable to traditional biomonitoring measurements. Among studies thus far, the amount of validation has varied considerably with no studies systematically evaluating all parameters from field collection, shipping and storage contamination and stability to laboratory analysis feasibility. These validation studies are requisite to ensure reliability of the measurement and comparability to more traditional matrices. Thus, we offer some recommendations for validation studies and other considerations before DBS should be adopted as a routine matrix for biomonitoring.

Introduction

Exposure assessment is inherently the most difficult part of epidemiology, sometimes being called its Achilles’ heel (Kogevinas, 2011). While personal measurements and ecologic assessments are often used for assessing exposure, the measurement of chemicals and their metabolites or reaction products in biological matrices has several distinct advantages over these techniques. For example, biological measurements quantify chemicals that enter the body and integrate all routes of exposure into one measurement (Barr et al., 2005b; Calafat, 2012; Needham et al., 2005). With advances in analytic instrumentation and the recognized advantages of direct measurements in human matrices, biomonitoring of exposure has burgeoned over the past 20 years (Barr et al., 2005b; Needham et al., 2005). It remains one of the primary and most widely accepted tools for exposure assessment used today (Barr et al., 2005b; Calafat, 2012).

Biomonitoring has been used for decades to measure exposure to bioaccumulative organic chemicals, also called persistent organic pollutants (POPs) (Adgate et al., 2001; Barr et al., 2007; Barr et al., 2005b; Needham et al., 2005). These chemicals, such as polychlorinated biphenyls (PCBs), organochlorine (OC) pesticides and polybrominated flame retardants, are usually measured in serum or plasma (Barr et al., 2005b). These measurements provide excellent information on body burden and are considered the gold standard for exposure assessment for these chemicals. More recently, biomonitoring in serum and plasma is also being extended to a growing number of biologically non-persistent organic chemicals (NPOPs) such as bisphenol A, phthalates and current-use pesticides (Adgate et al., 2001; Aylward et al., 2010; Barr et al., 2005a; Barr and Angerer, 2006; Barr et al., 2007; Barr et al., 2004; Barr et al., 2006; Biggs et al., 2008; Bouchard et al., 2011; Hanari et al., 2006; Horii et al., 2010; Kannan et al., 2004; Kannan et al., 2007; Tao et al., 2008; Wang et al., 2019). However, the use of serum and plasma matrices for biomonitoring has several distinct limitations, especially for children’s studies: (1) they provide only current body burden information, thus the timing and magnitude of exposure is unknown; retrospective biomonitoring is only possible if serum or plasma samples have been previously and properly collected and archived; (2) the invasive collection of blood (usually ∼10 mL) typically precludes the enrollment of infants and children in such studies; (3) for many NPOPs, serum is subject to contamination from pre-analytic processes (e.g., phthalate contamination from venipuncture) or the parent chemical can be quickly metabolized by blood enzymes; (4) sampling requires a trained phlebotomist; and (5) transfer of samples from a field location to the laboratory is cumbersome and costly (Barr et al., 2005b; Needham et al., 2005).

Dried blood spots (DBS) — drops of capillary whole blood obtained through heel prick in infants ≤ 6 months old and finger stick in those > 6 months old -- collected on standardized filter paper represent a minimally invasive alternative to venipuncture. Since the 1960s, DBS have been collected routinely as a part of each US state’s newborn screening (NBS) program to facilitate screening of newborns for congenital metabolic disorders (Guthrie and Susi, 1963; Mei et al., 2001) such as phenylketonuria (Carreiro-Lewandowski, 2002). The number of diagnostic tests performed and the number of blood spots (up to 15 spots) collected vary by state (Carreiro-Lewandowski, 2002), with some programs screening over 60 disorders [https://data.newsteps.org/newsteps-web/].

As a result of the successful NBS program, the pre-analytic and analytic procedures for the use of DBS in NBS have been fully vetted and validated (Adam et al., 2011; Chace et al., 1999; Denniff et al., 2013; Li et al., 2012; McHugh et al., 2011; Mei et al., 2001; Mei et al., 2010). For example, the filter papers or Guthrie cards are certified to meet performance standards for absorption and lot-to-lot consistency of the materials (Mei et al., 2010). Storage and shipping procedures have been evaluated to test for analyte stability under varied conditions (Adam et al., 2011; Bowen et al., 2011; Flores et al., 2017; Mei et al., 2011). Furthermore, testing procedures are guided and evaluated using a standardized quality assurance program developed and administered by the Centers for Disease Control and Prevention (CDC) for almost 40 years (De Jesus et al., 2010) that includes proficiency testing and quality control materials.

Overall, newborn screening procedures for DBS blood collection in NBS are minimally-invasive, low cost, and can be carried out by people with minimal training (De Jesus et al., 2010). Unlike serum or plasma, DBS samples do not need to be centrifuged, separated, or immediately frozen following collection for use in NBS. A cold chain (i.e., a documented process ensuring appropriate refrigeration or freezing at each step in the pre- and post-analytic processes) from the point of sample collection to receipt in the laboratory is not required. Capillary blood is simply dripped, not pressed, onto filter paper, allowed to dry thoroughly, and then wrapped, stacked, and stored. Most analytes measured in NBS remain stable at room temperature for a week or more, providing considerable flexibility in procedures for sample collection and transport. In addition, DBS samples are stable in laboratory freezers for longer periods (>1 week up to years) of time and can be stacked in relatively small amounts of space. For example, a standard 27 cubic foot lab freezer can hold 8,000 to 10,000 DBS sampling cards. After testing has occurred on the DBS samples, the state health departments retain the DBS for a specified period of time, either a short duration (< 3 years) or a long duration (indefinitely), for further diagnostic testing or for other novel uses (Therrell et al., 2011)[ https://data.newsteps.org/newsteps-web/]. In 2011, the Secretary of Health and Human Services convened an Advisory Committee on Heritable Disorders in Newborns and Children to develop recommendations on retention and future uses of DBS. These recommendations largely focused on genetic testing and confidentiality issues with no discussion of potential biomonitoring applications (Therrell et al., 2011; Vladutiu, 2010). However, the promise of increased repositories of residual DBS poses an opportunity for population-based research.

DBS offer unique opportunities for children’s health studies. Because of their ease of collection, shipment and storage, great interest has been expressed in using DBS in broad-ranging applications including infectious (Iyer et al., 2018; Lange et al., 2017a; Lange et al., 2017b; van Loo et al., 2017; Won et al., 2018) and chronic disease (Bjornstad et al., 2018; Brindle et al., 2010; Henderson et al., 2017; Hu et al., 2015; Lacher et al., 2013; Maleska et al., 2017; McDade et al., 2004; McDonald et al., 2017; Miller and McDade, 2012; Nguyen et al., 2014; Samuelsson et al., 2015) screening, genetic profiling (Bassaganyas et al., 2018; McDade et al., 2016; Segundo et al., 2018), pharmaco-management (Gallay et al., 2018; Page-Sharp et al., 2017; Schauer et al., 2018; Spooner et al., 2009), forensic testing (Hamelin et al., 2016; Perez et al., 2015; Sosnoff et al., 1996; Stove et al., 2012) and biomonitoring of chemical exposures (Archer et al., 2012; Basu et al., 2017; Batterman and Chernyak, 2014; Batterman et al., 2016; Chaudhuri et al., 2009; Funk et al., 2013; Funk et al., 2008; Kato et al., 2009; Kim and Kannan, 2018; Krishnan et al., 2013; Ladror et al., 2017; Ma et al., 2014b; Ma et al., 2013; Murphy et al., 2013; Pedersen et al., 2017; Spector al., 2014; Spliethoff et al., 2008a). In fact, procedures for measuring over 100 analytes in DBS have been published including indicators of endocrine, immune, reproductive, and metabolic function, nutritional status and infectious disease status (McDade et al., 2007). On average, about 50–100 μL of whole capillary blood are collected onto each spot and measurement assays may use anywhere from a 3-mm punch to one or more entire spots. The use of a controlled 3-mm punch enables further use of the spot since about three to four 3-mm punches can be obtained from one DBS. When an entire spot is used, multiple analyses may be performed from the spot eluate if the general elution is compatible with multiple assays. Recently, more than 35,000 DBS samples have been collected by NIH-funded studies in the United States, including large surveys like the National Longitudinal Study of Adolescent Health (McDade et al., 2007; Nguyen et al., 2014) and the Health and Retirement Study (McDade, 2011; McDade et al., 2007; Sonnega et al., 2014). The National Health and Nutrition Examination Survey (NHANES), widely regarded as the gold-standard for assessing the health of children and adults in the United States, has initiated an effort to implement DBS sampling to complement the mobile examination centers currently used to collect biological specimens (Miller et al., 2015). These applications indicate that DBS sampling is feasible for large-scale research applications, and that it is generally acceptable to research participants. For example, in 2007–2008, the National Longitudinal Study of Adolescent Health implemented DBS sampling in a large, nationally representative sample of 24–32 year old adults across the United States and found that 94% of participants were willing to provide a finger stick DBS sample which was higher than the percentage that willingly reported their income (93%) (McDade et al., 2007). Similarly, in the recent Household Air Pollution Intervention Network Trial (2017–2021)(Barr et al., 2020; Clasen et al., 2020), a multi-national birth cohort study with 3600 pregnant women, their children and other women in their households, longitudinal DBS were successfully collected in 91–99% participants (unpublished data), even in children from birth to 1 year of age.

Given the ease of implementation into population-based studies and the biorepository of archived blood spots available in each state, the implications of the use of DBS in children’s health studies are profound. Blood samples collection in infants or small children is difficult, at best, but more often infeasible, essentially precluding them from participation in many important studies. DBS represent a potential opportunity to include this vulnerable segment in more population-based studies as parents are more likely to allow a less painful, minimally-invasive heel or finger prick sample be taken than the more invasive venous blood draw (Bell et al., 2018; Ghassabian et al., 2018; Yeung et al., 2016), albeit the DBS measurements would be subject to the same limitations as blood in interpretation. Furthermore, the existence of perinatally collected spots may enable retrospective analysis of early indicators or markers of exposure or disease that may occur later in life. In fact, over the last two decades, archived DBS have been tested for feasibility for the analysis of environmental chemical exposures (Burse et al., 1997; Kato et al., 2009; Ma et al., 2013; Spliethoff et al., 2008a). While these studies have shown that the physical measurement of a chemical or metabolite in the DBS is feasible, despite the low volume of blood provided by a single spot, few studies have taken the process past the laboratory measurement stage to also test and validate the impact of field sampling, shipping and storage techniques on the viability of DBS as a matrix for biomonitoring. A recent review by Parsons et al. provided a historical view of DBS analysis for inorganic environmental chemicals in biomonitoring studies and summarized the current challenges and limitations in measuring inorganic chemicals in archived DBS with recommendations and technical justifications for improvement (Parsons et al., 2020). In this paper, we provide an overview of the published work on the use of DBS in exposure biomonitoring applications for organic pollutants and suggest some critical gaps that need to be addressed before DBS should be routinely used in biomonitoring studies, especially those that include children. Although other reviews have reported on feasibility of analysis of environmental and clinical biomarkers (Freeman et al., 2018; McClendon-Weary et al., 2020; McDade, 2011; McDade et al., 2007; Sharma et al., 2014), we specifically provide a laboratory-based viewpoint that addresses issues involved in proper measurement and interpretation of DBS-based biomonitoring data.

Biomonitoring of organic environmental chemicals in DBS

To interrogate the literature on the use of DBS to measure organic environmental chemicals, we used the Preferred Reporting Items for Systematic Reviews and Meta-Analyses approach (PRISM) (Moher et al., 2009). We searched PubMed, Web of Science, and Google Scholar databases; the search terms we used were “dried blood spots,” “environ*,” combined with one of the following terms: “chemical,” “PCB,” “BPA,” “PFAS,” “PFC,” “PBB,” “PBDE,” “pesticide,” “phthalate,” or “persistent organic pollutant.” The search returned 133 articles with an additional 3 included from a priori knowledge of their existence. After removing replicates, 87 papers remained which were screened for content. We removed all papers referring to inorganic chemicals (e.g., metals, metalloids), because those were recently reviewed separately (Parsons et al., 2020). We further refined the search to eliminate articles that used non-human samples, forensic analyses, or measured drugs of abuse, clinical biomarkers, or therapeutic pharmaceuticals. We excluded these articles as their levels would be expected to be much higher than most environmental chemicals, the resulting data may be used exclusively qualitatively, or the collection, storage and analysis issues addressed are likely not analogous to those encountered in human environmental chemical biomonitoring. We fully evaluated 23 articles and excluded 6 because they were review articles on biomonitoring specific chemicals or focused exclusively on the collection process. Ultimately, 17 articles were included in our review (Table 1). Most of these papers focused on chemicals that are traditionally measured in serum or plasma such as polychlorinated biphenyls (PCBs), polybrominated diphenyl ethers (PBDEs), legacy organochlorine insecticides (OCs), poly- and per-fluorinated alkyl substances (PFAS), and cotinine (Barr et al., 2005b). Although the potential for use of DBS in environmental health studies has been demonstrated (Malsagova et al., 2020), many outstanding issues must be resolved. For example, in the studies listed in Table 1, researchers have assumed that the DBS tested had a standardized volume ranging from 50 μL to 200 μL. Even though filter cards have been more standardized for absorptivity and diffusion of blood, the volume in the spots can vary considerably based upon how completely the spot is filled, if the spots are “double covered” or allowed to have multiple drops on top of each other, the blood hematocrit, or if the finger or heel is allowed to touch the filter card during collection (Peck et al., 2009). Additionally, in most instances, pre-analytic procedures have not been fully evaluated or validated. While the promise of DBS in exposure assessment, and in particular, in children’s exposure assessment exists because of its utility in NBS, we have an opportunity to fully invest in its validation and to set its performance standards before we haphazardly adopt it as a valid technique for environmental biomonitoring.

Table 1.

Summary of literature on the analysis of organic contaminants in dried blood spots grouped by target chemical class.

Citation	Target chemicals	Methods/Design	Validation components	Findings
Burse 1997 (Burse et al., 1997)	PCBs, p,p’-DDE	N=10 neonatal DBS from African American Texan, non-breastfeeding infants; GC-ECD	None. No LODs provided.	Only p,p’-DDE was detectable above blank concentrations ranged from 130 pg/mL to 1870 pg/mL. Blanks were not defined in the paper so it is unknown whether they were filter papers. Assumed DBS was 75–100 μL blood.
Lu 2012 (Lu et al., 2012)	PCBs and PBDEs	Archived DBS and dried plasma spots (1987–2009); no N provided; GC-HRMS	Recoveries and precision; stability over 30 days at room temperature; card blanks	Card-matched blank levels increased over time for most common PCBs and PBDEs. 30-day stability was within 14% of original measurement; spikes at 50 pg/mL had recoveries (relative standard deviations of 71–11% (9–20%) for dried plasma spots and 67–108% (5–47%) for DBS. PCB 180 > PCB 138> PCB 153 which is not a normal pattern.
Ma 2013 (Ma et al., 2013)	PBDEs	N=51 composite DBS samples ( comprised of 24 6-mm punches from DBS) collected from 1224 infants collected from 1997–2011 in NY; GC-HRMS	Blanks included reagent blanks and punches of blank space from each of the cards that made up each pool. Corrected recoveries ranged from 77–107%. RSDs ranged from 1–14%.	BDE-47 (86%) and BDE-99 (45%) were detected most frequently. Mean concentrations were 128, 40, and 12 pg/mL for BDE-47, -99, and -100, respectively. By pooling samples, the detection frequency can be greatly improved.
Ma 2014 (Ma et al., 2014b)	PCBs and OCs	N=51 composite DBS samples (comprised of 24 6-mm punches from DBS) collected from 1224 infants collected from 1997–2011 in NY; GC-HRMS	Blanks included reagent blanks and punches of blank space from each of the cards that made up each pool. Corrected recoveries ranged from 52–102% for PCBs and 26–88% for OCs. RSDs were <15% for all analytes.	BDE-47 (86%) and BDE-99 (45%) were detected most frequently. Mean concentrations were 128, 40, and 12 pg/mL for Significant time trend decrease in sum PCBs and p,p’-DDE over the span of 14 years. HCB did not decline over time. Calculated half-lives ranged from 3–10 years. Retrospective samples were used to test temporal trends in chemical exposures
Batterman 2014 (Batterman and Chernyak, 2014)	PCB; PBB 153; PBDEs; TBBPA; trans-nonachlor; p,p’-DDE; β-HCH; HCB	N=6 adults with venous blood draw, venous blood spotted on filter cards (not capillary blood); GC-MS	LODs ranged from 5–55 pg/mL. Temperature stability to mimic shipping and storage conditions at room temperature, 4°C, −20°C, −80°C; background contamination on cards evaluated.	Background concentrations were similar to those chemicals detected at low levels and less persistent chemicals (e.g., 28 congener). Good overall agreement between serum and DBS. PBDE loss from 8–23% over 1 year; PCB loss from <1–12% over 1 yr; OCs loss 2–9% over 1 year
Batterman 2016(Batterman et al., 2016)	PCBs, PBDEs, p,p’-DDE; β-HCH, HCB	N=21 adults with venous blood draw to create whole blood, plasma and DBS	LODs ranged from 10–200 pg/mL in DBS	Concentrations exceeded LODs in 25% of the samples for 14 chemicals in DBS. Concentrations in plasma > whole blood ∼ DBS. Concentrations similar to US population estimates except HCB and β-HCH were 10 and 15 times higher.
Spliethoff 2008(Spliethoff et al., 2008a)	PFAS	N=110 composite DBS samples ( comprised of 24 6-mm punches from DBS) collected from 2460 infants collected from 1997–2007 in NY; LC-MS/MS	Blanks included reagent blanks and punches of blank space from each of the cards that made up each pool. Corrected recoveries ranged from 85–103%. RSDs ranged from <10–20%. LODs were 0.2–0.4 ng/mL.	Frequencies of detection of all analytes was >90%. Calculated half-lives of PFOS and PFOA ranged from 2–8 years.
Kato 2009(Kato et al., 2009)	PFAS	N-97 DBS collected in TX in 2007; LC-MS/MS	Recoveries ranged from 53–146% with RSDs ranging from 13–22%. LODs ranged from 01–0.4 ng/mL. Blanks were reagent blanks only.	PFAS were detected in 70–100% of the samples using 1 spot with assumed 75 μL blood. Concentrations ranged from <LOD to 11 ng/mL.
Ma 2013 (Ma and Kannan, 2013)	PFAS and BPA	N=192 DBS collected in N from 20082011; LC-MS/MS	Blanks included reagent blanks and punches of blank space from each of the cards that made up each pool. Recoveries at 1 ng/mL ranged from 44–95% with RSDs ranging from 6–21%. LODs were 0.03–0.3 ng/mL.	Filter cards not appreciable source of PFAS contamination but BPA levels on the cards ranged from 0.5–0.8 ng/mL. PFAS detected in 100% of samples (GM=1.62 for PFOS and 1.16 for PFOA; range 0.21–6.46 ng/mL) while BPA was detected in 86% of samples (GM=3.84 ng/mL; range <1–36 ng/mL). Assumed 1 spot was equivalent to 50 μL. It is not known if the measured concentrations of BPA in DBS were due to exposure in utero or exposures that could have occurred in hospitals immediately after the delivery.
Poothong 2019 (Poothong et al., 2019)	PFAS	N=708 samples from 59 adults in 2013–2014; UPLC-MS/MS	Good correlation between DBS and venous blood (n=57; Spearman rhos ∼ 0.72–0.97, p < 0.0001)	Extensive whole blood/DBS cross matrix validation; 6 levels of quality control materials, blanks included reagent and calves blood; negligible levels of PFAS in blanks; RSDs varied from 9–40%; volume studies conducted with visual identification of volume; estimated 3.3 μL used in analysis; Bland Altman plots revealed small biases from −0.08 ng/mL to 0.5 ng/mL
Spector 2007 (Spector et al., 2007)	Cotinine	N=20 (10 DBS from infants of smoking mothers and 10 from nonsmoking mothers); ¼ spot used; GC-MS	None provided	Infants of smokers had significantly higher cotinine concentrations than infants of nonsmokers (28 vs. 8 ng/mL). Assumed 200 μL blood per spot.
Joseph 2013 (Joseph et al., 2013)	Cotinine	N=1541 DBS collected in 2010–2011	LOD was 0.3 ng/g.	Cotinine was detected in 61% of DBS with 17% over 3 ng/g. Median cotinine levels were significantly higher among Black than White children (0.66 ng/g vs 0.30 ng/g) and among Medicaid recipients (0.94 ng/g vs < 0.3 ng/g).
Murphy 2013 (Murphy et al., 2013)	Cotinine and trans-3-hydroxycotinine	N=83 DBS from smoker and N=99 from non-smoker adults in MN; N=283 archived DBS from 2010–2011; LC-MS/MS	Correlation between plasma and DBS cotinine; LOD was 0.3 ng/g	Correlation between plasma and DBS cotinine was 0.99. Correlation between 3-hydroxycotinine to cotinine ratios in plasma and DBS was 0.94. Intraclass correlation among 139 duplicate pairs was 0.995. Cotinine was detected in 59% of archived DBS ranging from <0.3 to 55 ng/g.
Spector 2014 (Spector et al., 2014)	Cotinine	N= 1414 DBS from CA, MI, NY, and WA; LC-MS/MS	LOD=0.3 ng/g; maternal smoking status vs. DBS cotinine	Cotinine was detected in 35% of DBS, including DBS of 29% of newborns whose mothers reportedly did not smoke during pregnancy; 12% DBS had cotinine >9 ng/g which indicates active maternal smoking although 41% of the mothers self- identified as a nonsmoker.
Ladror 2017 (Ladror et al., 2017)	Cotinine	N=158 DBS from smokers and nonsmokers; N=50 DBS from 1–21 year olds	LOD was <0.25 ng/mL using a 3.2 mm punch of a DBS; comparison of blood to DBS values	Correlation of matched plasma and DBS samples was 0.94 with 100% sensitivity and 94% specificity to differentiate smokers and nonsmokers. Levels in DBS collected form 50 1–21 year olds ranged from <0.25–6 ng/mL.
Funk 2008 (Funk et al., 2008)	Benzene (hemoglobin adducts of it oxide)	N=9 adult DBS prepared from venous blood and N=9 infant DBS from NC; GC-MS	Method variation ranged from 13–28%	Geometric mean levels ranged from 27.7 to 33.1 pmol/g globin. Adult blood, DBS and infant DBS samples provided similar results. Correlation of plasma values and DBS values was 0.73.
Xue 2016 (Xue et al., 2016)	Aflatoxin B1 (lysine adduct)	N=36 paired serum and DBS from adults in Kenya; whole spot; LC-fluorescence detection	Accuracy 98–106%; RSD 1.6–6.8%. 2% degradation over 24 hrs at room temperature, <5% over 12 months at 4°C. LOD = 0.4 pg/mg albumin. Extraction recoveries 68–76%.	Geometric mean levels of 11.88 pg/g albumin. Pearson correlation of 0.784 between serum and DBS.

POPs=persistent organic pollutants; PCB = polychlorinated biphenyl congeners; PBB=polybrominated biphenyl congener 153; TBBPA=tetrabromobisphenol A; OC=Organochlorine insecticide; β-HCH=β-hexachlorocyclohexane; HCB=hexachlorbenzenehexachlorobenzene; PFAS=poly- and per-fluorinated alkyl substances; BPA=bisphenol A; LOD=limit of detection; RSD=relative standard deviation; GC-ECD=gas chromatography with electron capture detection; GC-MS=gas chromatograph-mass spectrometry; GC-HRMS=gas chromatography-high resolution mass spectrometry; LC-MS/MS=liquid chromatography-tandem mass spectrometry; UPLC-MS/MS=ultra high performance liquid chromatography-tandem mass spectrometry

Gaps Existing in our Current Knowledge of DBS Use in Biomonitoring Studies of Organic Chemicals

Although DBS sampling is increasingly used as a minimally invasive technique to acquire representative blood for purposes other than NBS, the analysis of DBS is associated with several analytical challenges such as high contamination risk, low blood volume spotted, blood spot heterogeneity and variation based upon hematocrit levels and chromatographic effects (Peck et al., 2009; Ren et al., 2010). Below are some of the issues encountered in using DBS for environmental chemical analysis and strategies that can be employed to evaluate and/or mitigate those issues.

Sample collection and storage conditions:

Although DBS collection methods have been standardized for NBS (CLSI, 2013), for alternative uses, a variety of DBS sampling protocols have been proposed with varied levels of reliability testing and validation depending on the goal of the research. For DBS to be a valuable resource for trace analysis of organic environmental chemicals, standardization of DBS sampling procedures is necessary and may differ based upon the target chemicals for analysis. Considerable variability exists among states on how DBS samples are stored, the length of storage, and in policies for accessing specimens for research (Olshan, 2007). Approximately 57% of state laboratories that responded to a 2003 survey by CDC stored DBS for one year or less while about 15% of the laboratories stored DBS indefinitely (Olney et al., 2006). Of the 47 states that reported about storage conditions, 22 states followed at least one of the following criteria suggested for storage: (1) frozen (preferably at −20°C); (2) stored in sealed bags of low gas permeability; (3) inclusion of a desiccant in the bag; and (4) inclusion of a humidity indicator in a bag (Therrell et al., 1996). Only 12 labs stored the samples frozen (Therrell et al., 1996). According to this survey, 15 states (31%) had a written policy for how residual DBS could be used outside of NBS. The lack of written policies addressing laboratory, program, privacy and consent issues may limit the availability of DBS specimens for research (Olney et al., 2006). Duration and physical conditions of DBS storage and parental consent for the use of residual DBS for research can impose challenges in using DBS for research. In addition, the utility of DBS for environmental research may be hampered by substantial costs associated with retrieval of the specimens.

Quantitative analysis of trace levels of environmental chemicals in DBS in its current sampling format presents many technical challenges. The challenges in DBS analysis are largely ascribed to the way samples are collected and analyzed. First, in addition to standard alcohol wiping of the sampling area, the entire hand or heal area should be properly cleaned to remove any residual dust or grime that could contaminate the blood drops. Secondly, proper placement of whole blood on Whatman #903 filter paper is critical. If the blood is blotted or smeared onto the filter paper or if the blood is placed on top of a previously collected drop, then the quality of the DBS for quantitative biomarker analysis may be diminished. After wiping the initial blood drop from the finger or heel with sterile gauze, a large drop of blood should be allowed to form on the finger or heel and then fall onto the filter paper without any direct contact between the finger/heel and paper. Thirdly, variation in blood spot size can be minimized by collecting samples on filter papers with preprinted circles as guides to standardize the volume of whole blood collected from each individual. Most preprinted filter paper have ½ inch spots printed on them as guidelines. Techniques have been developed to spot accurate volume of blood onto the filter paper (Li et al., 2012). Alternately, the blood samples can be applied with calibrated micropipettes onto the filter paper to avoid potential variability in blood samples due to hematocrit effect, blood volume influence, and uneven distribution (chromatographic effect) of blood on the card although sometimes clotting of the blood is a problem, even in pre-heparinized capillaries. Finally, the filter paper matrix stabilizes many analytes in DBS, but the rate of sample degradation can vary by analyte and may be especially important for reactive or unstable analytes.

Drying and storage conditions can affect the target analyte and contamination following collection. Yet, no consensus has been reached with regard to the duration for air-drying of DBS after spotting the whole blood. An interval of at least 4 h or preferably overnight at room temperature (15–22° C) is proposed in some publications (Basavaraju and Pitman, 2014). The necessary drying time can vary considerably by climate as well; a particularly wet or humid climate may require more drying time than a warmer or more arid environment. Therefore, stability of target analytes should be evaluated prior to sample collection because this has direct implications for handling and storage of DBS. Refrigerating or freezing samples promptly after drying is advisable to minimize the chances of degradation but great care should be taken to avoid introducing humidity into the cards, so they should be frozen in sealed zipper top bags with desiccant packets and humidity card indicators to ensure proper storage over time. Because the majority of liquid blood or urine samples are stored frozen at −20°C or less for environmental chemical analysis, DBS should be stored at similar low temperatures if they are to be used for environmental chemical analysis. In order to avoid cross-contamination, samples should be separated from each other by inserting a clean blank paper between samples during transport and storage or storing each card in its individual zipper top bag. As with any biological samples, repeated cycles of freezing and thawing should be avoided although the filter paper matrix appears to provide some degree of protection against sample degradation that is not present with liquid blood samples (McDade et al., 2007). Requirements for shipping DBS samples are minimal unless the samples are known to contain an infectious or etiologic agent. Samples from normal, healthy individuals are considered diagnostic specimens and must be labeled as such for shipment. Overnight shipping of DBS in cold dry condition is preferred for environmental chemical analysis, however, for some applications where the target analytes are particularly stable in the filter paper, mailing through the regular post is acceptable.

Whatman 903 filter papers manufactured from 100% pure cotton linters with no wet strength additives are extensively used in NBS and are certified to meet performance standards for sample absorption and lot-to-lot consistency set by the Clinical and Laboratory Standards Institute (Hannon et al., 1997). The CDC, which maintains the Newborn Screening Quality Assurance Program (NSQAP) noted that collection on these filter papers achieves a similar level of precision and accuracy that analytical scientists and clinicians would expect from standard methods of blood collection using vacuum tubes or capillary pipettes (Mei et al., 2001).

Background contamination:

One of the major issues associated with the use of DBS for trace level analysis of organic chemicals is the background contamination of target chemicals present in DBS, which could arise during sampling, handling, transport, and storage. Considering the trace levels (on the order of pg/μL) of target chemicals present in DBS, contamination can significantly affect the accuracy of the measurements. Therefore, DBS cards must be handled carefully to avoid contamination (Antunes et al., 2016). Inclusion of procedural blanks (Milli-Q water in place of DBS) and method blanks (newly purchased Whatman grade 903 card) in the analysis would help evaluate the magnitude of contamination in DBS. One study showed that PFOS, PFOA, and BPA were found in both procedural blanks and method blanks (Ma et al., 2013). The mean concentrations of PFOS and PFOA in blanks (procedural and method blanks) were 0.01 and 0.1 ng/mL, respectively, while BPA was present in filter cards at concentrations of 0.5–0.8 ng/mL. These results suggest that the DBS filter cards themselves are not a significant source of PFOS and PFOA found in blanks; however, the levels of BPA in blanks likely preclude their use for general population exposures. Field blanks (unspotted portion of DBS filter paper) have been analyzed to determine background levels of contamination of PCBs and PBDEs in each of the filter papers (Ma et al., 2014a). Hexachlorobenzene (HCB), PCB-18, PCB-28, PCB-52, PCB-66, PCB-138 and PCB-153 were found in field blanks at concentrations in the range of 20–50 ng/L (Ma et al., 2014a).

Contamination levels of field blanks varied with time and the studies pointed out that the contamination occurred during sampling and storage (Krishnan et al., 2013; Ma et al., 2014a). Similarly, background contamination by PCB-105, PCB-170/190, PCB-194 (at 5–20 ng/L) and PBDE-47 (35 ng/L) has been reported on Whatman 903 filter cards (Batterman and Chernyak, 2014). However, the background levels were found to be consistent across the card samples and no additional contamination was found after storing DBS for 1 year (Batterman and Chernyak, 2014), suggesting that the contamination source originated from handling of the original card and storage did not add further contamination. Contamination from baby diaper rash cream, baby wet wipes, disinfectant, liquid infant formula, liquid infant formula, ultrasonic gel, breast milk, feces, and urine on DBS filter cards have been reported (Winter et al., 2018). Since lot-to-lot variation is likely, sampling and analysis of a punch taken close to the archived DBS is recommended to account for background levels of contamination (Lu et al., 2012). If background contamination is consistent across the card, and small relative to levels expected in the population, then background corrections can be used to avoid overestimating concentrations. Thus, monitoring of background levels of target analytes present in and across DBS filter cards is crucial for accurate determination of these chemicals in DBS samples. It is important to ascertain that the background levels of target analytes in DBS are several fold less than the concentrations found in the blood of newborns. The subtraction of background values from the concentrations determined in DBS sample is vital.

Hematocrit effect:

The hematocrit (HT) or packed cell volume fraction is the proportion of blood volume that is occupied by red blood cells. HT in newborns can range from 0.28 to 0.67 and from 0.35 to 0.42 for other children. Since HT is directly proportional to the viscosity of blood, it affects flux and diffusion of the blood spotted on the filter paper. At a high HT value, the distribution of a blood sample through the paper can be poor, resulting in a smaller blood spot when compared with aa blood sample with a lower HT. If a uniform punch of these spots was used for analysis, the high HT sample would have more blood per punch, leading to a higher measured analyte concentration than that from the low HT DBS samples (Adam et al., 2000; Mei et al., 2001). These differences are thus a potentially important source of exposure measurement error and subsequent bias in epidemiology studies.

One strategy to eliminate the HT effect is to utilize the entire DBS disk instead of taking punches from the disk. However, when only a punch is available, which is often the case for valuable samples, other strategies may accommodate the HT effect. Total protein, measured by a commercial protein kit, successfully reduces intraspot variability in measurements of butyrylcholinesterase (BChE) (Perez et al., 2015) and nerve agents (Shaner et al., 2018) in DBS punches. Potassium (K⁺), an intracellular analyte, can correct for differences in measured concentrations of caffeine and its major metabolite, paraxanthine, in experimentally generated DBS (Capiau et al., 2013; De Kesel et al., 2014), and used for removing nuisance variation in a metabolomics analysis of archived DBS (Petrick et al., 2017). While total protein and K+ have shown promise as HT correlates, the marker most studied is hemoglobin (Hb). Upon exposure to oxygen, Hb is converted to oxyhemoglobin (oxyHb), which is further oxidized to methemoglobin (metHb) and denatured to hemichrome (HC)(Bremmer et al., 2011b). The sum of oxyHb, metHb, and HC (i.e., ‘total Hb’) or the quantification of all Hb derivatives in DBS could be calculated as a proxy for hematocrit. For example, an experimentally derived UV-Vis absorbance maximum in the ‘Soret’ region was used by Yano et al. (Yano et al., 2019) and Petrick et al. (Petrick et al., 2019) to measure Hb in archived DBS extracts as a correlate of HT. These DBS punches were archived at −20°C and the oxidation process was slowed by freezing (Bremmer et al., 2011a), thus minimizing the presence of additional Hb species. An alternate approach is to use sodium lauryl sulfate to bind with all Hb complexes, which forms a single complex that can be measured at 550 nm using simple UV-Vis instrumentation (Oshiro et al., 1982; Richardson et al., 2018). Recently, advances in nondestructive measurements of Hb or total Hb ‘on-card’ were suggested. Miller et al. (Miller IV, 2013) proposed using an experimentally determined wavelength of 980 nm using near infrared reflectance (NIR) as a correlate of hematocrit, while Capiau et al. (Capiau et al., 2016; Capiau et al., 2018) showed that using diffuse reflectance spectroscopy for oxyHb, metHb, and HC or ‘total Hb’ measurement could adjust for caffeine in DBS punches. In the absence of a whole spot, accounting for the measurement variation induced by the HT effect in DBS punches can increase accuracy of measurements and therefore the ability to detect small differences in populations and reduce measurement error in epidemiology studies.

Chromatographic effect:

In addition to the effect of hematocrit and blood volume, the chromatographic effect (Mei et al., 2001) -- also known as distribution effect or coffee stain effect -- occurs during spotting blood because of the interaction of blood and/or analyte with the DBS filter paper. The chromatographic effect changes the way an analyte moves in the filter paper in a manner similar to thin layer chromatography, with the filter paper acting as the stationary phase and blood acting as the mobile phase. Depending on the structure of a given analyte, the chromatographic effect can drive the analyte to the edge of the spot or concentrate it in the center of the spot. O’Mara et al. (O’Mara et al., 2011) showed variations in the concentrations of five different compounds from punches taken at the center and the perimeter. Thus, during analytical method development, it should be assessed whether the same analyte concentration could be found from a punch from different locations of the same DBS in quality control (QC) samples at one or more concentrations. These QC samples are analyzed along with a set of calibration standards. Good reproducibility (<15% variability) suggests that the chromatographic effect is negligible (Liang et al., 2009).

The best strategy to resolve the hematocrit and chromatographic effects is through whole spot analysis. Partial sampling methods introduce unwanted bias owing to inconsistent spot sizes and possibly heterogeneous spots due to chromatographic effects. When possible, whole spot analyses can eliminate the variation from spreading and non-homogeneity and allow for more consistent DBS concentrations, even at different hematocrit levels (Li et al., 2011; Youhnovski et al., 2011).

Blood spot volume:

A major challenge currently associated with the use of DBS for environmental chemical analysis is the inability to accurately measure the volume of blood present in a punch or entire disk. Although several approaches have been used to predict the approximate volume of blood present in a punch (for example, 6 mm ∼ 10 μL; 12 mm ∼ 50 μL), these estimates are associated with various levels of uncertainty which can affect accuracy of trace analysis of environmental chemicals present in DBS, again potentially introducing additional measurement error across individuals/samples in epidemiology studies. Several strategies have been proposed to alleviate the issues related to blood volume present in DBS. One of the approaches is volumetric application of whole blood on the filter paper, which can be performed by pipetting or using precision capillaries or other micro-sampling devices that deliver fixed amount of blood onto filter paper. Whereas such techniques can be used in prospective sample collection protocols, this requires additional steps and special training that are not followed by the NBS. Therefore, quantitative application is not a viable option when using archived DBS. Different blood volumes spotted onto DBS filter paper may result in different measured analyte concentrations from a fixed punch size and HT value. During the development of a DBS extraction method, the relationship between DBS area/weight and the amount of blood spotted on the paper/card should be examined by spotting increasing volumes of blood onto the filter paper (at least in duplicate) and measuring the areas of the obtained spots (ter Heine et al., 2008) or weighing the obtained spots (Hoogtanders et al., 2007; van der Heijden et al., 2009). A linear relationship represents an even distribution of blood on the filter paper but does not really address person-to-person differences in hematocrit. Another more practical approach is to spot QC samples at least at two concentrations (e.g. low and high) onto the filter paper in six replicates with three or more increasing volumes (e.g.10, 30 and 60 μL). After drying, a single punch (e.g. 3 mm) is taken from the center of each DBS sample and analyzed along with a set of calibration standards, for which the spotted blood volume could be the same as one of the above QCs, or at a different volume (Liang et al., 2009; ter Heine et al., 2008). As long as the QC sample results are within ±15% of their nominal spike values, precise sample pipetting may not be necessary. Currently no suitable method exists to accurately determine the volume of blood present in DBS. This is the major challenge encountered in the analysis of DBS for environmental chemicals.

Because HT, blood volume and blood distribution may have a direct impact on the use of DBS sampling for quantitative analysis, these factors must be carefully evaluated. Otherwise, accurate pipetting using a calibrated pipette, followed by cutting the whole blood spot from the card/paper in combination with a normalization technique (e.g., Hb or K+ standardization) appears to be the only option for accurate and precise quantification of the analyte of interest.

Method validation and assay sensitivity:

Despite the popularity of DBS as a resource for assessing environmental chemical exposures in newborns, the small amount of blood limits assay sensitivity and poses a significant challenge to laboratory analysts. Determination of trace levels of environmental chemicals in DBS requires assays with great sensitivity, as the typical volume of blood available on a DBS with a ½ inch diameter is <100 μL. Several studies have reported quantification of environmental chemicals, including hexachlorocyclohexanes (HCHs) and dichlorodiphenyltrichlorethane (DDT) metabolites, p,p′-DDE (Burse et al., 1997; Dua et al., 1996), perfluorinated compounds (Kato et al., 2009; Poothong et al., 2019; Spliethoff et al., 2008b), metronidazole (Suyagh et al., 2010), polybrominated diphenyl ethers (PBDEs) and polychlorinated biphenyls (PCBs) (Lu et al., 2012). The majority of the earlier studies have used pooled specimens of DBS collected from several individuals to achieve a sample volume adequate for analysis and detection of environmental chemicals in newborns. For environmental epidemiological studies, it is appropriate to analyze DBS from individual babies, so that the exposure levels of environmental chemicals in each infant can be associated with health outcomes. A few recent studies have used a single disk of 16-mm diameter (∼50–100 μL whole blood) for the analysis of PFOS, PFOA and BPA. However, such analytical methods required highly sensitive state-of-the art instrumentation such as high performance liquid chromatography (HPLC) with high-resolution mass spectrometry (HRMS) or HPLC-MS/MS. In other words, the analytical instrumentation should be capable of measuring sub picogram amounts of target chemicals.

Unlike whole blood, serum or plasma, for which sample homogeneity could be ensured by vortex-mixing, laboratory methods for the analysis of environmental chemicals in DBS samples require some modifications from those reported for plasma/serum. Because the whole blood has been dried on filter paper, target analytes must be brought into solution. Buffers have been used in the extraction for biomarkers of various endogenous biomolecules. However, for the extraction of environmental chemicals from the filter paper matrix, various organic solvents are used. Disintegration of the filter paper matrix with harsh acidic or alkaline conditions can ensure complete dissolution and release of chemicals into solution, but it is important to ensure that such harsh solvents do not affect the integrity of target chemicals. The extraction efficiency and matrix effect of the target analytes must be tested with various solvents to assure that all target chemicals are recovered from the filter paper matrix. This can be tested in the laboratory by fortifying fresh whole blood samples with environmentally relevant concentrations of target chemicals, spotting them on the Whatman 903 filter paper, allowing the spot to dry for 4 h at room temperature, followed by extraction and quantification. Calibration standards should be prepared by fortifying the fresh whole blood with various levels of target analytes for the assessment of recovery and matrix effects (as matrix-matched calibration curve). Furthermore, because a very small volume of blood is used for extraction (e.g., ∼50 μL if entire spot is used), the final extract should be concentrated to few microliter volumes in order to be able to detect the signal of the target analytes. Several earlier studies have concentrated the final sample extract to 20–50 μL prior to instrumental analysis for PCBs, PBDEs and OCPs (Krishnan et al., 2013; Ma et al., 2014a).

Detection of Chemicals versus Metabolites and Adducts

As indicted above, the use of plasma, serum and DBS analysis methods are best suited to chemicals that persist in the body, with long half-lives. Short-lived chemicals will indicate recent exposures. Adducts that are formed by reactive metabolites on blood proteins such as hemoglobin (Funk et al., 2008) and albumin (Xue et al., 2016) will persist following exposure, and accumulate following repeated exposure. These biomarkers can provide an integrated dosimeter, based on the lifespan of the erythrocyte for hemoglobin adducts (approximately 120 days, with a zero-order loss) (Fennell et al., 2005) or the half-life of albumin (approximately 19 days with a first order loss)(Sleep et al., 2013). The use of hemoglobin adduct monitoring methods in DBS from newborns may be limited by the differences in sequence between adult and fetal hemoglobin. Many methods rely on the modified Edman procedure for detection of valine adducts (N-terminus in both alpha and beta hemoglobin) and are standardized using these adducts (Fennell et al., 2005; Yano et al., 2019). However, fetal hemoglobin consists of alpha and gamma subunits with glycine is the N -terminus in the gamma subunit, so the standardized methods for measuring these adducts would not be acceptable (Fennell et al., 2005; Yano et al., 2019).

Validation of Entire Collection, Shipping and Analysis Process:

While we have mentioned validation needed in each step of the pre-analytic and analytic process, validation of the entire process from start to finish including field testing methods should be performed before adopting a DBS method as an acceptable means of exposure assessment. While each step may be independently validated, small changes that each step in the process may affect all subsequent steps. Undertaking a series of validation schemes similar to NBS is necessary to ensure the validity and reliability of measurements from DBS.

Proficiency testing:

The CDC has been operating the NSQAP for over 40 years. The program helps participating NBS program laboratories to evaluate and improve the quality of their testing efforts by providing quality control DBS materials and proficiency testing materials for the external evaluation of screening programs. These activities have created a mechanism for the validation of methods for the analysis of DBS. There is no regulatory guidance for bioanalytical method validation and subsequent sample analysis for DBS samples for use in routine biomonitoring applications. However, both the US Food and Drug Administration and the European Union have developed practical guidelines for suitability of bioanalytical methods that serve as a good model for suitable validation parameters to dictate the appropriateness of a given method for DBS analysis (EMA, 2012; FDA, 2018).

Ethical issues:

Most state NBS programs use informed refusal or dissent, meaning that parents may refuse the DBS collection and subsequent screening tests which are otherwise required by law. These consent forms do not cover other growing number of uses for DBS. The collection form and educational material for parents could indicate that the sample becomes the property of the state and that, unless the parents object in writing, the samples may be used without personal identifiers in studies related to preventing birth defects and disorders of the newborn or for protecting the public’s health (Therrell et al., 1996). It is also worth noting that although significant benefits may be gained from the use of DBS for epidemiologic research, the general public has great concern over potential genetic testing and its implications. Potential misuse of DBS can lead to discrimination, psychological harm, identification of incorrect assignment of paternity, and potential social injustices. NBS programs should begin to promulgate policies and rules for retention and the use of residual DBS specimens for public health research programs.

Overall Recommendations

Before routine use of DBS in epidemiologic studies, full validation procedures should be conducted. These validations should include:

1. Laboratory measurement feasibility and validation studies:

   1. Measurement feasibility on instrumentation should be evaluated. This can first be accomplished by injecting amounts of the target analytes anticipated in 50–100 μL blood on instrumentation. Given the ultra-sensitivity and selectivity required at such low concentrations, confirmatory analytical instrumentation such as tandem or high-resolution mass spectrometry should be used.
   2. A controlled laboratory evaluation of target analyte-spiked DBS of known and varied blood volumes and HTs to ascertain the homogeneity of spotting on the card, chromatographic and HT effects, volume dispersion on the cards, and recovery of the analytes from the spots, whether they be whole spots or punches. This can be done by sampling different spot volumes (e.g., 50, 75, and 100 μL of blood) from each of ∼5–6 different blood samples (where HT has been determined) and assessing variability of measured concentrations using punches (e.g., 3-mm punch) and whole spots. If available, a NIST Standard Reference Material (SRM) can also be used to evaluate overall accuracy. If NIST SRMs are unavailable, spiked recoveries at 2–3 different concentrations should be conducted.
   3. Laboratory validation of the methods should be performed. Validation experiments should include those commonly done for most bioanalytical methods (i.e., accuracy, precision, limits of detection and quantification, robustness, stability studies, and an assessment of 10–20 unknown DBS samples). Special consideration should be given to robustness studies with a focus on potential contaminants in the samples or on the cards. Unspotted portions of the DBS cards themselves can be used to evaluate contamination as well. In addition, the evaluation of analyte and DBS stability under varying conditions (e.g., heated temperature which may simulate postal shipping procedures, refrigerated and freezing temperatures). Stability tests should represent the amount of time anticipated from sample collection to analysis being sure to evaluate all potential pre- and post-analytic conditions. Furthermore, effects of humidity on the DBS/analyte stability should be considered, especially if desiccants or humidity indicators will not be used.

2. Matrix comparison studies. Without being able to appropriately compare DBS analyte levels to levels in more traditional matrices, interpretation may be difficult. Most organic chemicals are present in the blood serum or plasma which represents about 60% of whole blood. Thus, we would expect to find a ratio of 0.6 between DBS and serum measurement levels. For POPs, lipid adjustment is often employed; such adjustment should allow a 1:1 ratio. Evaluation of the necessity of such an adjustment is warranted as well. Similarly, knowing how DBS levels of NPOPs compare to the traditional urine matrix, will help us interpret resulting data and may alert us if a problem exists. Typically blood levels of NPOPs are about 2–3 orders of magnitude lower than urinary metabolite levels (Barr et al., 1999; Barr et al., 2005b), so ultra-sensitive methods must be employed for meaningful results.

3. Field collection and shipping studies:

   1. Field collection should be tested and refined. DBS should be collected by drip only without touching fingers or heels to cards. A variety of shipping procedures should be evaluated concurrently for acceptability.
   2. Preferably, spots should be collected from the same participants and shipped via postal service or overnight express at ambient temperature, cold (blue ice) or frozen (dry ice). Field technicians should be properly trained in collecting and shipping procedures.

4. A standard operating procedure (SOP) should be developed and adopted for DBS collection, shipping, storage, and analysis. Establishing an SOP is tedious and time-consuming but is necessary for proper operation and use of DBS in epidemiologic studies.

Conclusions

Although DBS have been demonstrated to be a feasible matrix for environmental exposure assessment, many areas related to their use remain non-standardized or unvalidated. In general, DBS analyses have been shown to be most promising for measurement of POPs. Given that POPs are typically measured in blood or blood serum/plasma and are inherently stable, this may be an acceptable way of capturing exposure measures in children or evaluating retrospective exposures. The use of DBS for other environmental chemicals still remains less understood. Regardless, even for POPs measurements, before adopting DBS measurements as credible measures for environmental health-based studies, special precautions should be taken to ensure the integrity of the sample from collection to analysis. Use of retrospectively collected DBS from newborn screening program may yield an estimate of chemical exposure levels, the margin of error associated with such measurements need to be carefully taken into consideration while interpreting data from those studies. In addition, standardized methods for estimating sample volume should be implemented.