Correlation of Environmental Carbaryl Measurements with Serum and Urinary 1-Naphthol Measurements in a Farmer Applicator and His Family Dana B. Shealy,I John R. Barr,1 David L. Ashley,1 Donald G. Patterson, Jr,1 David E. Camann,2 and Andrew E. Bond 3 'Division of Environmental Health Laboratory Sciences, National Center of Environmental Health, Centers for Disease Control and Prevention, Atlanta, GA 30341-3724 USA; 2Department of Environmental Chemistry, Chemistry and Chemical Engineering Division, Southwest Research Institute, San Antonio, TX 78228-0510 USA; 3National Exposure Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711 USA. In exposure or risk assessments, both environmental and biological measurements are often used. Environmental measurements are an excellent means for evaluating regulatory compliance, but the models used to esimae body b n from e e arec Unless .ll pos- sible routes of exposure (i.e., inhalation, derma abso n, inetion) are evaluated, eosure to a toxicant can be underestimated. To crcumvent this problem, measurements of the internal dose of a toxicant in blood, serum, urine, or tissues can be used sinlaly or in combination with environmentl data for exposure assessment. In three separate laboratories, carbaryl or its primary metabolite, l-naphthol, was measured i personal air, dermal samples, blood serum, and urine from farmer applicators and their fimilies. The u ess of both environmenl ad biological data has been demonstrated. For the farmer applicator, he environmental levels of carbaryl would have been sufficient to determine that. an exposure had occurred. However, bio- logical measurements were necessary to determine the absorbed dose of eacmember of the applicator's family. In addition, a correlation between serum and urinary 1--naphthol measure- ments has been shown; therefore, either matrix can be used to accurately evaluate occupational carbaryl exposure. Key words: air, carbaryl, farmer, 1-naphthol, serum, urine. Environ Healh Perspect 105:510513 (1997) Carbaryl is one of the most widely used industrial insecticides (1). In 1988, about 25 million pounds of carbaryl was applied to crops on farms in the United States to eliminate chewing and sucking insects (1). Carbaryl does not generally linger in the environment. It is readily absorbed into the soil, where it quickly breaks down, and does not leach into groundwater. For these rea- sons, residual levels of carbaryl in the envi- ronment are not believed to be hazardous to the general population. Agricultural workers, however, are often exposed to much greater levels of carbaryl than the general population. Because expo- sure symptoms associated with cholinesterase inhibition usually appear long before toxic quantities of carbaryl are absorbed, acute poisonings involving carbaryl are rare. Although toxic quantities are rarely absorbed in occupational settings, improper use of safety equipment (i.e., gloves, respirators, protective clothing) may lead to high-level exposures to carbaryl. In humans, carbaryl does not accumu- late in tissues or persist in blood. It is quick- ly metabolized into a nontoxic compound, 1-naphthol, which is excreted in urine as the glucuronide or sulfate ester (2). The body burden of 1-naphthol as measured in urine or serum is the most common indica- tor of exposure to carbaryl. Carbaryl exposure may also be evaluated indirectly by measuring carbaryl in personal respiratory air, dermal patches, and dermal wipes. Although these environmental mea- surements are not necessarily related to the amount of carbaryl absorbed by the body, they provide useful information about exposure routes and absorption potential. However, if data are obtained that show a clear correlation between environmental and biological measurements, that correla- tion could be used to evaluate the mecha- nisms of carbaryl entry into and distribu- tion within the body. Environmental measurements of car- baryl and biological measurements of 1- naphthol in selected agricultural farmer applicators and their families are reported. These data, when viewed collectively, pro- vide useful information for occupational carbaryl exposure assessment. Materials and Methods Sample collection. AU samples were collected from six farmer applicators and their fami- lies, who voluntarily, with informed consent, participated in the pilot Agricultural Health Study (3) conducted by the National Cancer Institute (NCI), the U.S. Environmental Protection Agency (EPA), and the National Institute of Environmental Health Science (NIEHS). In this study, biological (i.e., blood and urine) and environmental (i.e., air, house dust, food) residues of a variety of pesticides and/or their metabolites were measured and detailed questionnaire data obtained. The collective data obtained in this study will assist the cosponsoring agencies evaluate the role of agricultural exposures, as well as dietary and lifestyle factors, in the development of cancers and of neurologic and other chronic diseases (3). The final interpretation of the collective study data will be reported elsewhere. Only the data from biological and personal environmental sam- ples involving carbaryl exposures will be dis- cussed here. During the application season, up to four biological samples were collected from each participant and two personal air, der- mal, and handwipe samples were collected from the farmer applicator. One sample was collected on the day prior to pesticide appli- cation and another on the day of pesticide application. Biological samples were also collected on the two days following pesti- cide application. Additional environmental samples included one handwipe sample from each family member and one house- hold indoor air sample taken on the day of application. Personal air samples were not obtained from the family members because previous studies (4,5) suggest that concen- trations of toxicants in their personal air are similar to indoor air concentrations. Personal air, dermal patch, and hand- wipe analysis. Personal air and dermal sam- ples were collected only from the farmer applicator. In accordance with established sampling procedures, the applicator wore a portable air sampler and an Q-cellulose der- mal patch during the entire sampling period (4). Handwipe samples were collected from the applicator and family members with cot- ton swabs and isopropanol (4). Following a previously published solvent extraction Address correspondence to D.B. Shealy, CDC, 4770 Buford Highway, Mailstop F-17, Atlanta, GA 30341-3724 USA. We thank Joe Wooten, Cheryl McClure, Vince Maggio, and Carolyn Newman for analytical assis- tance. The use of trade names is for identification purposes only and does not constitute endorsement by the Public Health Service, the Department of Health and Human Services, the Centers for Disease Control and Prevention, or the U.S. Environmental Protection Agency. Received 4 November 1996; accepted 6 January 1997. Volume 105, Number 5, May 1997 * Environmental Health Perspectives 510 Articles * Correlation of carbaryl measurements procedure (5), carbaryl was isolated from sample matrices and analyzed by gas chro- matography with mass spectrometry. The limits of detection (LODs) for the personal air, dermal patch, and handwipe analyses were 0.7 ng/m3, 14.3 ng/sample, and 1.74 ng/sample, respectively. The coefficients of variation (CVs) ranged from 5 to 15%. Urine analysis. Urine samples (10 ml) were prepared according to a previously published method that involved solvent extraction and chemical derivatization (6). The derived extracts were analyzed by gas chromatography coupled with isotope-dilu- tion, low-resolution tandem mass spectrom- etry (7). Two ions were monitored for both the native 1-naphthol and the 13C-labeled analogue. One ion was used for 1-naphthol quantification and the other for analyte con- firmation. The calculated concentrations of unknowns were normalized on the creati- nine content in each urine specimen. The LOD [calculated as three times the standard deviation at zero concentration(3so)] of the method was 1.2 pig/l (ppb) with an average CV of 7.5% on repeat measurements of unknown samples at concentrations span- ning the entire linear range. Serum analysis. Serum samples (8 g) were prepared according to an established method (J.R. Barr, unpublished data) involving protein denaturation and solid phase extraction (SPE). 1-Naphthol con- centrations in the serum extracts were ana- lyzed by gas chromatography and isotope- dilution, high-resolution mass spectrometry U.R. Barr, unpublished data). The LOD (3so) of the method was 19 ng/l (ppt) with an average CV of 19% at 100 ng/l. Data analysis. All data were evaluated statistically with SAS statistical software (SAS Institute, Cary, NC). Most of the data points available for correlation analyses were very near the LOD. To avoid using a very skewed distribution of data points in statis- tical analyses, only quantitative data above the limit of quantitation (LOQ; 10so) (8) of the biological methods were used in correla- tion evaluations. The elimination of the data between the LOD and LOQ did not critically affect the correlation analysis or its significance. Pearson correlations were con- sidered significant if p<0.05. Results and Discussion Of the six farmer applicators studied, only one was actively applying carbaryl on his crops at the time of monitoring. The car- baryl and 1-naphthol concentrations mea- sured in environmental and biological sam- ples associated with this applicator are shown in Table 1. Only biological data (not shown) were obtained from those farm applicators and families from farms on Table 1. Concentrations of carbaryl in the environmental samples and 1-naphthol and carbaryl in biological samples from a carbaryl applicator Sample Preapplication Application Post application Post application day day day 1 day 2 Personal air (pg/m3 carbaryl) 0.0080a 640 NS NS 0.01 6b Dermal patch (pg/cm2 carbaryl) 0.oloa 11 NS NS 0.001 4b Handwipe (pg carbaryl) 20.0a 20,100 NS NS g.0b Urinec(pg/g creatinine 1-naphthol) 270 (860) 140 (500)a 7,100 (12,000) 1,500 (2,600) 9,300 (22,000)b Serum (pg/l 1-naphthol) 0.260 510 1.9 0.56 Serum (pg/l carbaryl) ND 0.12 ND ND Abbreviations: NS, no sample was collected; ND, below the limit of detection of the method. aMorning sampling on a day where two samples were obtained. bEvening sampling on a day where two samples were obtained. cUrinary measurements in parentheses are expressed in micrograms per liter. 8 7 0 I= co ._ *. 6 0c U -a 5 s e 4 -5 CD .0 -i Day-1 Day 3 Day-1 Day Day Day+1 Day+2 (A.M.) (P.M.) Sampling day Figure 2. Carbaryl uptake and elimination in serum (measured as 1-naphthol) and urinary 1-naphthol excretion in a carbaryl applicator. All sampling days are shown relative to the application day (Day). Both morning and evening urine samples were obtained on the application day, whereas only an evening serum sample was obtained. The dashed line indicates that a data point (application day morning sample) is missing in the serum profile. 1-naphthol concentrations of biological samples obtained on the same days rise at similar rates as the environmental samples. The biological measurements of sam- ples from the carbaryl applicator showed a distinct elimination pattern (Fig. 2). The serum uptake/elimination profile was con- sistent with the expected profile (9). Before pesticide application, the serum 1-naphthol level of this applicator was higher than that of applicators from farms that did not apply carbaryl. Following carbaryl applica- tion, the serum 1-naphthol concentration of the applicator had increased by about three orders of magnitude, peaking at about 0.5 mg/l (ppm). In addition, approx- imately 100 ng/l nonmetabolized carbaryl = I= co U- =0 Q .0 CD -i Sampling day Figure 1. Graph of the log of environmental car- baryl measurements of an applicator before (Day -1) and the on the day (Day) of carbaryl applica- tion (R2 = 0.85). The units of measurement are nanograms per cubic meter for personal air, nanograms per sample for dermal patches, and micrograms per 2 hands for handwipes. which carbaryl was not applied during the study. The concentrations of 1-naphthol in the urine samples from these individuals exhibited no elimination pattern and were below the 95th percentile of the reference range, suggesting only background exposure. The results of the environmental mea- surements for the carbaryl applicator sug- gest overt occupational exposure (Fig. 1). All environmental measurements taken the day carbaryl was applied were significantly higher than those taken before application, and all environmental measurements appeared to increase at similar rates. A regression analysis of the sampling day ver- sus the log of the carbaryl concentration of all environmental samples produced a coef- ficient of determination (X2) of 0.85, indi- cating an agreement between the measure- ments. In addition, the urinary and serum Environmental Health Perspectives * Volume 105, Number 5, May 1997 511 Articles - Shealy et al. was detected in the application day serum sample from the applicator. His serum 1- naphthol level decreased in the 2 days fol- lowing exposure until it closely approached the preapplication level. The carbaryl applicator's urinary 1- naphthol level was consistently higher than his serum 1-naphthol level (Fig. 2). The creatinine-adjusted urinary 1-naphthol con- centration in the applicator (preapplication sample) was about 250 pg/g creatinine, which is significantly greater than the 95% percentile reference range value of 36 pg/g creatinine (10,11). On the morning before application, the urinary 1-naphthol concen- tration dropped almost 50% from the pre- vious day, although the level was still over three times the reference range. This decrease could possibly be a result of con- tinued elimination from a previous expo- sure. After carbaryl application, the applica- tor's urinary 1-naphthol concentration rose sharply and peaked at 22,000 p'g/l (9,300 pg/g creatinine), a level comparable to other reported occupational carbaryl exposures (12,13). As the applicator's serum 1-naph- thol levels dramatically decreased in the days following application, the urinary 1- naphthol level remained much higher than that of preexposure samples as 1-naphthol was being eliminated from the body. Urinary 1 -naphthol measurements were also obtained from the applicator's spouse and two children. The 1-naphthol levels in the spouse and first and second child approximately doubled following carbaryl application (from 13, 7.4, and 8.1 pg/g to 27, 12, and 19 1.g/g creatinine, respective- ly), indicating exposure to carbaryl during its application on their farm; however, the values were low and within the reference range of the U.S. population. Neither the indoor air concentration of carbaryl, which is representative of the personal air concen- trations of the family members, nor the house dust levels appeared to increase upon application of carbaryl on the farm, so it is unlikely that this would be the source of the low-level exposure in family members. Even at the low levels observed, the expo- sure of family members to carbaryl applied on the farm warrants further study. A Pearson correlation analysis of the log serum and log urinary 1-naphthol concentra- tions in the same individuals showed good correlation between the two methods (1?2 = 0.945, p = 0.0003). A plot of the method correlation is shown in Figure 3. The num- ber of data points for this analysis is small; therefore, corresponding data may not be accurately interpolated from the resultant regression line. However, the good correla- tion between serum and urinary concentra- tions of 1 -naphthol suggests that accurate .r-7 03 6 cn o SD 03 1.5 2.0 2.5 3.0 Log serum concentration (ng/l) 3.5 Figure 3. Correlation plot of serum and urinary 1- naphthol measurements. The correlation coeffi- cient is 0.945 (n = 7, p = 0.0003). assessments of carbaryl exposure can be obtained from either medium. When there is no significant measurement bias between biological matrices, the noninvasive sampling techniques used in urine collection are often preferred because samples are easier to obtain, especially from children. Additionally, a Pearson correlation analysis of handwipe carbaryl and urinary 1 -naphthol in the same individuals corre- lated well (R2 = 0.997, p = 0.0003). A plot of this correlation is shown in Figure 4. With three of the four data points close together at lower concentrations, the corre- lation was largely determined by a single data point. However, as shown in the inset (Fig. 4), the data correlate well (R2 = 0.996, p = 0.003) even when the high value is omitted. This correlation suggests that handwipe measurements may accurately reflect the change in internal dose of car- baryl following an exposure. Although primarily associated with car- baryl elimination, urinary 1-naphthol is also attributed to naphthalene exposure, especially from cigarette smoke (14,15). When 1-naphthol is mainly derived from naphthalene, a correlation usually exists between it and 2-naphthol, a secondary metabolite of naphthalene (11). A poor correlation between the two geometric iso- mers of naphthol is indicative of another 1- naphthol source, usually carbaryl exposure. For statistical purposes, we sorted sam- ples into two groups: an application and nonapplication group. Samples were in the application group if they were obtained from any person on the farm where carbaryl had been applied during the sampling week. Conversely, samples were in the nonappli- cation group if they were obtained from farms where carbaryl was not applied. Plots of log 1-naphthol concentration versus log co = 2.Oe+7 CN4 CD a .2 1.5e+7 0 .5+ o1.Oe+7 co 2,000 4,000 6,000 8,000 10,00( Urinary 1-naphthol concentration (ng/ml) Figure 4. Correlation plot of handwipe carbaryl and urinary 1-naphthol measurements. The corre- lation coefficient is 0.997 (n= 4, p = 0.0003). The lower concentration end of the plot is enlarged in the inset. 2-naphthol concentration in application and nonapplication samples are shown in Figure 5. A Pearson correlation analysis of each subgroup was performed (Fig. 5). The concentrations of urinary naphthol isomers among samples of the nonapplication group were significantly correlated (p = 0.0001), whereas the concentrations among samples of the application group were not (p>0.05). This supports the assumption that the increase in excretion of 1-naphthol in the urine of residents of a farm where carbaryl was applied is due mainly to the bodily absorption of carbaryl during the applica- tion period. Conclusion In most cases, the most useful exposure information is the biologically effective dose of a toxicant; however, this information is usually difficult, if not impossible, to obtain. In contrast, environmental measurements are usually easy to obtain and can be good indicators of the actual intake of a pesticide, especially if multiple intake routes are evalu- ated. However, the most direct indicator of the intake of toxicants during an exposure is the measurement of the body burden or internal dose of toxicants or their metabo- lites in blood, urine, or tissues. Because metabolic rates may vary from person to person, blood or serum measurements of the nonmetabolized toxicants are often consid- ered more accurate than urinary metabolite measurements. However, urine samples are much easier to obtain and are often the only samples available from small children. In this study, the usefulness of both environmental and biological data in evalu- ating environmental exposures has been demonstrated. The environmental data of the carbaryl applicator would have been Volume 105, Number 5, May 1997 * Environmental Health Perspectives O 512 Articles - Correlation of carbaryl measurements A =L C4 o.. . .... Cj CL~~~~~~~~~~~~~~~~~~~~iL Lo g I-a hlS cocetato (gg/1 eo 0 1 2 3 4 5 2 S. ;S ..... >tz $ #w w #? ? ?? if %#t ?s# :: S2 ?S -'''?< '$ ?WSt? - B Asm$?.@? ?.........Z ,. -?....... .$ H .... _ z. $w~fif1?# , $ ' ?'? ?2tt-?--?, w . ,,, ., ,t .,,?we,, 0 1 2 3.xx..<>>?rl ............Ei Lo 1 Iahto cocnrto (1jig/I) Figure 5. Correlation plots of naphthol isomers. The naphthol isomers are poorly correlated in the application group (A) (* = 0.22, p>0.05), indicating a non-naphthalene source of urinary 1-naphthol. The naphthol isomers correlate well in the nonap- plication group (B) (R = 0.62, p = 0.0001), indicat- ing primarily naphthalene-derived 1-naphthol. sufficient to decide whether an exposure had occurred. However, biological mea- surements were necessary to determine the extent of the toxicant absorbed in each member of the farmer applicator's family. A strong correlation between serum and urinary 1-naphthol measurements has been shown. This correlation indicates that either matrix can be used to accurately evaluate occupational exposure. In theory, since there was good agreement among the environmental and biological measure- ments, additional data points could be used to construct a curve from which environ- mental measurements could be used to esti- mate body burden. REFERENCES 1. U.S. EPA. EPA's pesticide programs. Washington, DC:Environmental Protection Agency, 1991. 2. Knaak JB, Tallant MJ, Kozbelt SJ, Sullivan LJ. The metabolism of carbaryl in man, monkey, pig, and sheep. J Agric Food Chem 16(3):465-470 (1968). 3. Alavanja MCR, Akland G, Baird D, Blair A, Bond A, Dosemeci M, Kamel F, Lewis R, Lubin J, Lynch C. Cancer and noncancer risk to women in agriculture and pest control: the Agricultural Health Study. J Occup Med 36(11):1247-1250 (1994). 4. Harding HJ, Merritt PM, Clothier JM, Camann DE, Bond AE, Lewis RG. Sample collection methods to assess environmental exposure to agricultural pesticides. In: Measurement of toxic and related air pollutants; Pittsburgh, PAAir & Waste Management Association, 1993;842-847. 5. Geno PW, Majumdar TK, Camann DE, Bond AE. A multiresidue GC/MS method for the determination of pesticides in environmental media. In: Measurement of toxic and related air pollutants, VIP 50. Pittsburgh, PA:Air & Waste Management Association, 1995;531-541. 6. Hill RH Jr, Shealy DB, Head SL, Alley CC, Bailey SL, Gregg M, Baker SE, Needham LL. Determination of pesticide metabolites in human urine using isotope dilution technique and tandem mass spectrometry. J Anal Toxicol 19(5):323-329 (1995). 7. Shealy DB, Bonin MA, Wooten JV, Ashley DL, Needham LL. Application of an improved method for the analysis of pesticides and their metabolites in the urine of farmer applicators and their families. Environ Int 22(6):661-675 (1996). 8. Taylor JK. Quality assurance of chemical mea- surements. Chelsea, MI:Lewis Publishers, 1987. 9. Bernard A, Lauwerys R. General principles for biological monitoring of exposure to chemicals. In: Biological monitoring of exposure to chemi- cals: organic compounds (Ho MH, Dillon HK eds). New York.John Wiley & Sons, 1987;l-16. 10. Needham LL, Hill RH Jr, Ashley DL, Pirkle JL, Sampson EJ. The priority toxicant reference range study: interim report. Environ Health Perspect 103(suppl 3):89-94 (1995). 11. Hill RH Jr, Head SL, Baker SE, Gregg M, Shealy DB, Bailey SL, Williams CC, Needham LL. Pesticide residues in urine of adults in the United States: reference range concentrations. Environ Res 71:99-108 (1995). 12. Shafik MT, Sullivan HC, Enos HF. A method for the determination of 1 -naphthol in urine. Bull Environ Contain Toxicol 6:34-39 (1971). 13. Comer SW, Staiff DC, Armstrong JF, Wolfe HR. Exposure of workers to carbaryl. Bull Environ Contam Toxicol 13(4):385-391 (1975). 14. Corner EDS, Young L. Biochemical studies of toxic agents. 7: The metabolism of naphthalene in animals of different species. Biochem J 58:647-655 (1954). 15. Bieniek G. The presence of 1-naphthol in the urine of industrial workers exposed to naphtha- lene. Occup Environ Med 51:357-359 (1994). Environmental Health Perspectives * Volume 105, Number 5, May 1997 513