Environ. Sci. Technol. 1999, 33, 1359-1365
Distribution of 2,4-Dichlorophenoxyacetic Acid in Floor Dust throughout Homes Following Homeowner and Commercial Lawn Applications: Quantitative Effects of Children, Pets, and Shoes MARCIA G. NISHIOKA,* HAZEL M. BURKHOLDER, AND MARIELLE C. BRINKMAN Battelle Memorial Institute, 505 King Avenue, Columbus, Ohio 43201 ROBERT G. LEWIS National Exposure Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711
Transport of lawn-applied 2,4-D into 13 actual homes was measured following both homeowner and commercial application of this herbicide to residential lawns. Collection of floor dust in five rooms of each house, corresponding to an entryway, living room, dining room, kitchen, and a child’s bedroom, both prior to and after application, indicated that turf residues are transported indoors and that the gradient in 2,4-D dust loading (µg/m2) through the house follows the traffic pattern from the entryway. Removal of shoes at the door and the activity level of the children and pets were the most significant factors affecting residue levels indoors after application. Spray drift and fine particle intrusion accounted for relatively little of the residues on floors. Prior to application, the median 2,4-D bulk floor dust loading was 0.5 µg/m2; one week after application, the median 2,4-D floor dust level in the living room was 6 µg/ m2, with a range of 1-228 µg/m2 on all carpeted floors in occupied homes, and 0.5-2 µg/m2 in unoccupied homes. The 2,4-D loadings on the carpet surface (dislodgeable residue/dust) were highly correlated with the 2,4-D bulk dust loadings. From these data we estimate that approximately 1% of the bulk dust is on the carpet surface, and it is this surface dust that may be readily available for dermal contact.
Introduction Approximately 80-90% of U.S. households report using pesticides (1, 2). With detection of pesticides in indoor air and house dust months to years postapplication, researchers have concluded that pesticides are highly persistent in the indoor residential environment (3-9). The ubiquitous presence of insecticides such as chlorpyrifos and permethrins in indoor air and dust suggests primary indoor use. However, * Corresponding author phone: (614)424-4964; fax: (614)424-3638; e-mail:
[email protected]. 10.1021/es980580o CCC: $18.00 Published on Web 03/31/1999
1999 American Chemical Society
migration of residues from the house foundation, crawlspace, or basement and track-in from lawn and garden may be contributory (3-5). For 2,4-dichlorophenoxyacetic acid (2,4-D), carbaryl, and chlorothalonil, which are applied exclusively outdoors, their presence in indoor air and floor dust implies that residues have been transported indoors via one or more transport mechanisms, including track-in (i.e., transport via foot traffic). In recently reported studies, residential track-in was simulated following lawn applications of 2,4-D, chlorpyrifos, and chlorothalonil (10, 11). These studies showed that walking over treated turf as much as one week after application resulted in transport of measurable residues and that the residues in carpet dust following track-in were proportional (3-4%) to the dislodgeable residues on the turf. The simulation studies also showed that pesticide residues after track-in were present in the bulk carpet dust and also on the carpet surface (as dust-bound or neat residue) and, thus, could contribute to the exposures of children in contact with that surface. The public health impact of residential pesticide use is largely unknown, especially that arising from chronic exposure of children in the indoor environment. In-vitro tests suggest that 2,4-D is not mutagenic; however, increases in malignant tumors have been found in rodent studies with high or prolonged exposure levels (12). Increases in the incidence of non-Hodgkin’s lymphoma were found among farm populations of Kansas and Nebraska using 2,4-D, but the cancers could not be ascribed solely to 2,4-D exposure. Inasmuch as these results were not found in other epidemiologic studies, albeit populations were small and followup times were short, the carcinogenic potential of 2,4-D is not clear (13). This manuscript details results from a pilot study of 13 homes, providing data on the extent to which lawn-applied 2,4-D was tracked into these homes and disbursed throughout the floors along the family traffic pattern, following lawn applications by both homeowners and commercial applicators. Sampling was conducted for both bulk floor dust (on carpeted and bare floors) and for residues on the surface of the carpet (dislodgeable dust/residues). This report also shows the impact of familial activity patterns on levels of 2,4-D brought indoors.
Experimental Methods Study Design. Homeowners in the Columbus, OH area who routinely use lawn chemicals were recruited for this study in which application to the lawn was made by the homeowner in year 1 and by a commercial applicator in year 2. Each family consisted of two adults, two to three school-age children, and one pet (one home had no pets). Homes were single story with a basement (except one split level), surrounded on all sides by turf, and carpeted in the main living room and a child’s bedroom. The sampling period at each home consisted of two one-week periods: a preapplication background week and a postapplication week. Preapplication sampling took place late March through April, and postapplication sampling took place mid-April through May. The sampling in pre- and postapplication weeks was nearly identical and consisted of indoor air sampling for 24 h on day 1 and day 3; wipe sampling of sills, tables, and bare floors in four rooms on day 8; collection of a carpet surface dust sample from the living room on day 8; and vacuum sampling of five floors, both carpeted and bare, on day 8. VOL. 33, NO. 9, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1359
Floor dust sample collection was carried out in the entry room (Entry), the central/main living area (Liv) of the home, a dining room (Din), kitchen (Kit), and a child’s bedroom (Bed). Although homes had at least two entrances, discussions with homeowners were used to identify locations in the five designated rooms that would fall along a frequently used traffic pattern or would constitute high to low use within the house. Locations of shoes, coats, school books, bicycles, and carpet wear were important secondary indicators. Sampling required approximately 15-20 h total at each home. An additional indoor air sample was collected during the actual 2,4-D lawn application, and this application event initiated the postapplication week. Deposition coupons on the lawn were used to estimate 2,4-D application rates. An integrated air exchange rate measurement was made during the postapplication week (14). The 2,4-D floor dust measurements are the subject of this report; analyses of air and surface wipe samples are the subject of a forthcoming article. During both the pre-and postapplication week, homeowners were asked to refrain from cleaning floors (sweeping, vacuuming, mopping) so as not to disturb the normal deposition and distribution of residues. Since approximately 47% of Americans vacuum floors only once or twice a week (15), standardization of this activity for this study is not inconsistent with typical activity patterns. However, this prohibition on cleaning presents a worst case scenario for residue levels on floors. Otherwise, families had no constraints on their normal activities. Due to mild weather during the monitoring period, heating and air-conditioning were not needed; windows were frequently open. The above sampling design was used in both the first and second years of the study with minor modifications. Seven families participated in the first year; four of these families, representing several important activity patterns, were included in the second year, together with two recently constructed unoccupied homes. Homeowners applied any one of about eight commercially available postemergence herbicide formulations consisting of dicamba, mecoprop, and 2,4-D (e.g., KMart K-Gro), with a desired lawn application rate of approximately 80 mg 2,4-D/m2. Homeowners used their own application equipment, either a hose-end sprayer or pressurized pump sprayer. The commercial applicator applied K-Gro formulation with a commercial sprayer. In accordance with Health and Human Services regulations, the study design, protocol, and informed consent were reviewed and approved by Battelle’s Human Subjects Review Panel. Sampling and Analysis Methods. A 0.2 m2 area of each bare (uncarpeted) floor was wiped with a cotton gauze wipe, one-half of a Johnson&Johnson SOF-WICK dressing sponge moistened with 2 mL of a solution (70:30 phosphate buffer: acetonitrile) that bears some similarity to sweat (16). The surface dust from the Liv carpet was collected with the EPA/ SwRI Polyurethane Foam (PUF) Roller (3, 10). The PUF Roller collection surface, a cylindrical sleeve of surface area 72 cm2, was moistened with the aforementioned “sweat simulant”. With this solvent mixture, the otherwise rigid PUF becomes soft, pliable, and slightly moist to the touch, so that the PUF surface is consistent with the intent of the roller to simulate a child’s hand contact with a surface. (Water-moistened PUF is somewhat rigid, with discontinuous beads of water, and may not be a good surrogate for skin.) A 0.48 m2 area was sampled, this being the perimeter of the area to be vacuumed. Bulk floor dust samples were collected using the HVS3 vacuum sampler, using four passes over the designated area, two passes each in orthogonal directions (10). The sampled area was as close to a 2 m2 area as possible, while remaining in an area of general foot traffic. HVS3 sampling of bare floors 1360
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 9, 1999
was suspended for the yr 2 study. Samples were transported on ice to the laboratory and then stored at -78 °C until analysis. The analytical method for bulk dust and PUF Roller collected samples has been described in detail elsewhere (10, 11); these reports detail the techniques for precleaning, moistening, and extracting the PUF sleeves as well as the precision of this sampling method. A summary of the analytical method applied to both media follows. The surrogate recovery standard (SRS), 3,4-D, was added prior to extraction at a level similar to that expected for 2,4-D: 100 ng for floor wipes and carpet surface dislodgeable residue samples and 500 ng for dust. After extracting the sample with 70:30 acetonitrile:0.1 M sodium acid phosphate at pH 3, distilled/deionized water was added, the pH was adjusted to 12, the extract was partitioned with hexane, and the aqueous extract was subjected to further separation on a C18 SPE cartridge. The SPE eluate was concentrated to near dryness; the internal standard for quantification, 2,6-D, was added, and the extract was adjusted to 1 mL with 5% methanol in methyl tert-butyl ether and methylated using ethereal diazomethane. Multilevel calibration standards were derivatized similarly. GC/ECD analyses were carried out as described before (10). Samples with analyte concentrations exceeding the calibration range were diluted, remethylated, and reanalyzed. Average recoveries for the SRS 3,4-D in field-collected samples were 92 ( 15% (n ) 101) for floor dust samples, 76 ( 9% (n ) 26) for floor dust wipe samples, and 86 ( 24% (n ) 24) for PUF Roller dislodgeable carpet surface residue samples.
Results and Discussion Establishing Track-In in the Residential Environment. The 2,4-D bulk floor dust loadings (µg 2,4-D/m2) in six occupied homes (A-F) of the yr 1 homeowner applicator study are shown in Figure 1. (Home G, which was not included in the yr 2 commercial applicator study, is not included in Figure 1 to facilitate visual comparisons between Figures 1 and 2.) Figure 1 includes both the preapplication 2,4-D levels and the levels of 2,4-D one week after the homeowner’s application. Three phenomena are readily identified. First, 2,4-D is detectable on all floors in all homes prior to, and one week after, the lawn application. Second, 2,4-D floor dust loadings one week postapplication are considerably higher than those levels at the end of the preapplication week. Third, there is a gradient in the 2,4-D levels throughout each home that corresponds to the traffic pattern through the house which family members follow when entering from outdoors. For completely, or nearly completely, carpeted homes, the 2,4-D levels in floor dust were highest in the entry area, and dropped to sequentially lower levels throughout the house along the traffic pattern. This gradient in 2,4-D floor dust from high to low was evident whether calculated on the basis of 2,4-D loading (µg 2,4-D/m2) or 2,4-D concentration (µg 2,4-D/g dust) and is consistent with our expectation of track-in from an external location. This same gradient in the 2,4-D floor levels was evident in both the pre- and postapplication bulk floor dust samples, although much more pronounced in the postapplication period. The average preapplication 2,4-D level in all homes, 0.5 µg/m2, was similar to the average level reported previously for the nine homes in which sampling was done approximately 5-6 months after the general 2,4-D application period in Columbus, OH (10). In this current study, the postapplication levels were approximately 2-200 µg/m2 or a 4-400-fold increase over preapplication levels. The track-in gradient is most readily discernible in this data set in homes A and B which had carpeted entryways and carpeting throughout most of the house. Any bias in
FIGURE 1. Preapplication and postapplication 2,4-D house dust loadings (µg/m2) with homeowner application of 2,4-D to lawn; year 1 study. accumulation mode or sampling between bare and carpeted floors may have been largely eliminated by virtue of having carpeting throughout the house. A slightly different track-in gradient is observed for homes such as homes C and E (Figure 1) which had a substantial number of uncarpeted floors in the early sections of the house traffic pattern. Two distinct gradients in 2,4-D floor dust loadings appeared within these homes: one gradient established for the uncarpeted floor areas and a second established for the carpeted areas. Note that the bare floor areas (sheet vinyl, wood, etc.) are designated in Figure 1 with an asterisk (*). The postapplication 2,4-D loadings on these bare floors are 5-20-fold lower than the loading on the nearest sequential carpeted area. The difference in 2,4-D loadings in bulk dust between bare and carpeted areas is due to factors other than the dust loading, as illustrated by data from two representative homes,
shown in Table 1. For home B (carpeted throughout), the dust loading was remarkably similar in all rooms (0.5-1.1 g/m2), and the single 2,4-D gradient throughout the house was observed in both the 2,4-D loading and the 2,4-D dust concentration. This suggests that if track-in is the primary transport/intrusion mechanism, the 2,4-D initially tracked into the home at the entry is diluted as it is dispersed along the traffic pattern. For home C (and also E, both of which had bare and carpeted floors), the concentration of 2,4-D in the dust is nearly equivalent in all bare floor samples, so that the 2,4-D loading (µg/m2) is governed by the amount of dust. The concentration of 2,4-D in the bare floor dust is quite low, 1-2 µg/g, relative to its concentration in the carpeted Liv floor dust, 14 µg/g. For homes with bare floor entries, the data suggest that 2,4-D tracked in at the beginning of the week, presumably at higher concentrations, is transported VOL. 33, NO. 9, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1361
FIGURE 2. Preapplication and postapplication 2,4-D house dust loadings (µg/m2) with commercial application of 2,4-D to lawn; year 2 study. to, and accumulates in, the carpeted areas of the house by in-house activity. The 2,4-D tracked-in later in the week at lower levels may be that which is found on the bare entry floors at sampling time. The 2,4-D floor dust loadings in homes one week after commercial lawn application are shown in Figure 2. (Note that floor dust samples were not collected in the Din area this time, and only wipe sampling was used for bare floors.) Trends identified above are again evident. First, 2,4-D is detected, with one exception, in all postapplication floor dust samples, including the floor dust from the unoccupied homes (X and Y). Second, with exception of a few wipe samples from bare floors, 2,4-D is present in preapplication dust samples, however, at levels that are more variable than in the previous year, ranging here from 0.2 to 5 µg/m2. Again, the postapplication levels are significantly higher than 1362
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 9, 1999
preapplication levels. Third, the track-in traffic gradient is again evident in these homes. This latter observation indicates that track-in cannot be attributed solely to track-in upon reentry by the applicator, since in no case did the commercial applicator enter the house. Several additional trends were also identified. First, in the main living areas of homes A and B there was approximately a 3-fold reduction in 2,4-D levels relative to the homeowner application study, contrasted with a 3-fold increase in 2,4-D levels in homes E and F. In the case of home B, there was a 2-fold increase in the 2,4-D level on the lawn, and in home F there was a 5-fold decrease in the level on the front lawn. Since the homeowners in homes B, E, and F removed or thoroughly rinsed shoes after self-application and before reentry, the differences between homeowner and commercial application tend to indicate that activity patterns
TABLE 1. Comparison of 2,4-D Floor Dust Loadings and Concentrations from HVS3 Vacuum Collection in Two Representative Homes (One with and One without a Carpeted Entry) Following Homeowner Application to Turf home B: carpeted throughout room
floora
2,4-D loading, µg/m2
entry kit liv din bed
C C C C C
74 35 13 12 5.3
a
home C: many bare floors throughout
2,4-D concn, µg/g
dust, g/m2
room
floora
2,4-D loading, µg/m2
2,4-D concn, µg/g
dust, g/m2
67 57 28 20 7.8
1.1 0.62 0.48 0.58 0.67
ent/kit hall din liv bed
V W W C C
0.71 3.1 1.7 70 27
1.6 1.2 1.7 14 11
0.45 2.6 0.98 5.0 2.5
Flooring types: C, carpet; V, sheet vinyl (bare floor); W, wood (bare floor).
of the family overshadow effects such as the applicator’s reentry into the home or the level on the turf. Second, in the four occupied homes participating in both studies (A, B, E, and F), the 2,4-D loadings in the child’s bedroom were nearly identical for homeowner and commercial application. This may suggest that a child will establish an individualized trackin pattern that is most evident in his/her bedroom. Third, the increase in 2,4-D loadings in the floor dust of the unoccupied homes after lawn treatment, though minimal, point to transport mechanisms other than track-in, such as resuspension from foliar surfaces followed by fine particle intrusion. Due to the ethical obligation of informing homeowners of results of the first study before inviting them to participate in the second study, the design of the commercial-applicator study was somewhat compromised. Changes in family behavior were evident, and the results tended to confirm the overarching importance of family activity patterns to trackin. In homes A and B, greater vigilance was exercised with respect to track-in by pets and children in the first few days after commercial application. In homes E and F, parental reminders to children to remove shoes at the door was not enforced as stringently as in the first year of the study, with the higher 2,4-D floor loadings suggestive of increased trackin of residues. This change in lifestyle resulted from the E and F homeowners’ conclusion, reached upon reviewing the data from the first study, that they had been overly cautious relative to other participants. Although homes X and Y are designated as unoccupied, some foot-traffic did occur in these homes during the study. In home X, the builder’s agent spent 4 h/day there answering phone calls; this agent entered through the garage and spent her time indoors. Access to home Y was restricted, although one client inadvertently visited the home near the end of the week. In both homes, sampling teams made multiple visits to the homes but limited their potential foot track-in by removing shoes at the door. These scenarios, in comparison with fully occupied homes, suggest that the postapplication 2,4-D floor dust levels of homes X and Y were caused primarily by intrusion mechanisms other than track-in and are presumed to involve resuspension of 2,4-D from turf followed by fine particle intrusion of the closed house (home Y) and an incremental addition of fine particle penetration as doors and windows were opened (home X). Comparison of Wipe and Vacuum Sampling for Bare Floors. While not the focus of this sampling effort, minor conclusions may be drawn about wipe vs vacuum sampling from data obtained. The surface loadings of 2,4-D are listed in Table 2 for bare floors where both wipe and vacuum samples were collected. Data are categorized by the sampling time (pre- or postapplication), by the floor type, and whether outdoor shoes were worn indoors. In the preapplication period, wipe and vacuum sampling appear to give comparable results for relatively smooth wood floors, as indicated by the ratio approximately equal to 1. In contrast, the collection efficiencies of these techniques differ
TABLE 2. Comparison of 2,4-D Loading on Bare Floors with Wipe and Vacuum Sampling 2,4-D loading, µg/m2 preapplication postapplication flooring type and shoes worn indoors? wipe vacuum W/Va wipe vacuum W/V vinyl yes no wood-smoothc yes no wood-groovedd yes no
0.32
0.08
4
22.7 9.17 0.67 0.56
1.6 0.71 0.59 0.26
14 13 1.1 2.2
NTb
NTb
NTb
0.34 0.55 NTb
0.30 0.61 NTb
1.1 0.9 NTb
5.09
1.65
3.1
NTb
NTb
NTb
0.32 NTb
5.74 NTb
0.06 NTb
2.50 1.58
3.13 1.47
0.8 1.1
a W/V: ratio of 2,4-D loading, wipe to vacuum. b NT: not tested (sample not collected or no home available with those characteristics). c Tongue and groove wood floor with few gaps or breaks in surface. d Parquet flooring and/or worn tongue and groove flooring with uneven surface.
significantly for smooth vinyl and grooved wood (e.g., parquet or worn) floors, with wipe collection being more efficient on the vinyl floor, and vacuum collection showing higher loadings than the wipe on the grooved wood floor. The latter result is most likely due to collection of dust from within grooves not reached by a wipe. For estimation of dermal contact, wipe data may be preferable in rooms that have wood floors, since only the surface is sampled with a wipe. (Note: wipe sampling was not used uniformly in the yr 1 study; all available data are presented here.) If trends from these limited number of samples are meaningful, it appears that equivalent efficiency in sampling bare floors shifts in the postapplication period. Approximately equal loadings are now measured in the samples from vinyl floors in cases where outdoor shoes were not worn indoors and from grooved wood floors (with or without shoes worn). Data for smooth wood floors is equivocal. The major difference in collection is observed in 2,4-D loadings from smooth vinyl floors where outdoor shoes are worn, with wipe sampling providing a more efficient collection of residues. Further directed studies would be required to confirm these results and elucidate mechanisms. These data are presented here so that other researchers can use current data and extrapolate to other sample collection methods. Comparison of 2,4-D Loadings in Dust and Carpet Surface Dislodgeable Residues. There is a very high degree of correlation (Pearson correlation of r ) 0.95) between the 2,4-D loading in bulk dust and dust on the carpet surface. This correlation appears to hold well despite the diversity of carpet types involved. For the linear relationship between 2,4-D in bulk dust/m2 vs 2,4-D in surface dust/m2, the slope of 0.0085 corresponds to an approximate 100:1 ratio between VOL. 33, NO. 9, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1363
TABLE 3. Descriptors for Families and Homes home descriptor no. of adults no. of children ages: M F child activity levelc pets activity areae descriptor pet activity levelf applicator’s shoes worn indoors family’s shoes worn indoors air infiltration, m3/h yr 1: yr 2: air exchange, L/h yr 1: yr 2: 2,4-D on turf, mg/m2 homeowner: yr 1 commercial: yr 2
A
B
C (X)a
D (Y)b
E
F
G
2 3 8, 10 6 hi dog in/out runs with kids hi yes
2 3
2 3 5, 7, 9
2 2 5, 10
low no
2 3 7, 10 12 hi dog in kenneled in house lowg no
2 3
hi 2 cats in/out sedate low yes
2 3 5 11, 13 low dog in/out old; sedate low no
5, 10, 12 low cat out kenneled in garage low no
yes
yes
yes
yes
no (yr2 ∼ yes)
no (yr2 ∼ yes)
yes
247 117
407 831
289 (203)
249 (70)
254 78
127 177
300
0.5 0.25
0.7 1.43
0.6 (0.17)
0.6 (0.10)
0.6 0.19
0.3 0.43
0.6
42.5 45.8
19.2 55.6
50.5 (48.2)
55.5 (48.3)
30.6 44.7
217, 40i 43.5
9.2
6, 11, 14 modd none
low dog out lowh no
a Data for unoccupied home X in yr 2 study (see Figure 2). b Data for unoccupied home Y in yr 2 study (see Figure 2). c Hi (based on observations): two boys close in age, share a bedroom, have friends in neighborhood, run, push, jostle, leave bicycles against house for ready use; low (based on observations): separate bedrooms, not observed playing together. d Three additional children, ages 8 (M), 8 (M), and 11 (F), at this house after school for 2 h. e Indoor and/or outdoor area where pet played or spent time. f Hi (as based on observations): jumps on furniture, runs with children; low (as based on observations): spends a lot of time sleeping or resting. g Dog was fairly active when allowed to run around inside, but overall activity contributing track-in was judged to be low. h Dog was observed running around outside, but overall activity contributing track-in was judged to be low because dog was kept outside or in the garage. i Deliberate difference by homeowner; high level to front yard, lower level to back yard where children played.
TABLE 4. Contribution of Activities to 2,4-D Loading in Living Room Floor Dust incremental addition of 2,4-D, µg/m2
parameter application spray drift ventilation of home particle intrusion of closed house particle intrusion with opening/closing of doors/windows track-in applicator’s shoes worn indoors high activity children moderate activity children low activity children high activity dog low activity dog a
∼0 0.3 1.4 51.2 (p ) 0.01)a 9.0 w/o shoes, 16.7 with shoes -0.2 w/o shoes, 7.5 with shoes -6.0 w/o shoes, 1.7 with shoes 117.5 (p ) 0.001) 14.6
Parameters with p level listed were the only parameters statistically significantly different from zero.
bulk dust and dust on the carpet surface, i.e., 1% of carpet dust is present on the surface. Role of Activity Patterns. The study presented here is limited by the small number of homes studied and the fact that formal survey instruments were not used; observations and recall with homeowners were used to interpret the data. To the extent that these homes represent important trends and factors in the general population, it appears that familial factors may have a greater effect on residential exposure than application factors such as spray drift. A comparison of the homes is shown in Table 3 with respect to occupants, activity patterns, average 2,4-D turf application rates, and air exchange/air infiltration rates. The levels of child and pet activities and whether family members wore their outdoor shoes indoors were examined as differences among these homes. A multivariate analysis of variance (ANOVA), using both yr 1 and 2 Liv 2,4-D bulk dust loadings, was used to estimate the contributions of different transport mechanisms to the postapplication 2,4-D loading on the Liv floor. The model fits the untransformed postapplication 2,4-D 1364
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 9, 1999
loading on the Liv floor as a simple linear function of indicator variables for low, moderate, and high child activity, family shoes worn indoors, applicator shoes worn indoors, opened or closed status of the home, and high and low dog activity levels. For simplicity, the air exchange/infiltration rates and 2,4-D turf loadings were not included. The analysis was run in SAS. Although the data are sparse, and assumptions of the model could not be tested, the analysis allowed estimation of the incremental contribution of each transport factor. Results, listed in Table 4, indicate that a high activity dog and the homeowner applicator’s shoes worn indoors were the most significant factors, followed by high activity children and their shoes. Only the high pet activity factor and applicator’s shoes were statistically significant at p < 0.05 confidence level. The high activity homes were those with at least two boys in the 7-10 age bracket who shared a bedroom and had friends in the immediate neighborhood. In the case of the home with the high activity dog, the dog was in contact with the treated turf within an hour of application, whereas
children were not; the dog was observed running and playing with the children and was often on the Liv sofa with family members. Implications for Indoor Exposure. To limit intrusion of this pesticide into the home, it may be advisable to limit the contact of indoor-outdoor pets with the treated turf and/or to wash the animals frequently in the first week after lawn treatment. It also appears that homeowners can apply lawn care products with no more indoor intrusion occurring than with commercial applications if the applicator’s shoes are removed before he/she enters the home. Although the role of his/her clothing (such as contamination on pant legs) could not be deduced from the small data set here, it appears reasonable to suggest that use of coveralls that are removed before reentering the house can also limit track-in intrusion of pesticides. Finally, consistent removal of shoes at the door not only by the applicator but also by all family members appears to result in substantially lower track-in of lawnapplied chemicals. Entry mats and uncarpeted floors have been suggested (17) as control devices for limiting track-in and accumulation of pollutants in the home and, thereby, reducing the potential for childrens’ indoor exposures to pesticides in dust via dermal contact and nondietary ingestion (hand-to-mouth). The data here provide an interesting corollary: a carpeted entry, where children are less likely to play, may serve as a retainer for tracked-in pollutants and prevent migration of much of it into carpeted living areas where children do play. Uncarpeted entries, or bare floors with a smooth-surface, may only exacerbate the migration of pollutants into carpeted living areas. The data presented here are being combined with the aforementioned data for 2,4-D in air and on solid surfaces of the homes for assessment of potential residential exposures. These data will be reported shortly. We emphasize that these results are based on a small sample set, with a single measurement for some activity factors, and thus would warrant further validation prior to extrapolation to all families and all homes.
Acknowledgments The authors acknowledge the participation of families in the Columbus area, the lawn application provided by TruGreenChemLawn, use of Key and Nicholas Homes model homes, and assistance by John Menkedick of Battelle for ANOVA analyses. The information presented in this paper has been wholly funded by the U.S. Environmental Protection Agency under Cooperative Agreement CR-822082 to Battelle. It has been subjected to Agency review and approved for publication. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.
Literature Cited (1) Savage, E. P.; Keefe, T. J.; Wheeler, H. W.; Mounce, L. M.; Helwic, L.; Applehaus, F.; Goes, E.; Goes, T.; Mihlan, G.; Rench, J.; Taylor, D. K. Arch. Environ. Health 1981, 36, 304-309. (2) Whitmore, R. W.; Kelly, J. E.; Reading, P. L. National home and Garden Pesticide Use Survey, Final Report, Vol. 1; Report No. RTI/5100/17-01F, prepared for the U.S. Environmental Protection Agency by Research Triangle Institute, Research Triangle Park, NC, March 1992. (3) Lewis, R. G.; Fortmann, R. C.; Camann, D. E. Arch. Environ. Contamin. Toxicol. 1994, 26, 37-46. (4) Whitmore, R. W.; Immerman, F. W.; Camann, D. E.; Bond, A. E.; Lewis, R. G.; Schaum, J. L. Arch. Environ. Contamin. Toxicol. 1994, 26, 47-59. (5) Richter, E. D.; Kowalski, M.; Leventhal, A.; Grauer, F.; Marzouk, J.; Brenner, S.; Shkolnik, I.; Lerman, S.; Zahavi, H.; Bashari, A.; Peretz, A.; Kaplanski, H.; Gruener, N.; Ishai, B. P. Arch. Environ. Health 1992, 47, 135-138. (6) Mukerjee, S.; Ellenson, W. T.; Lewis, R. G.; Stevens, R. K.; Somerville, M. C.; Shadwick, D. S.; Willis, R. D. Environ. Int. 1997, 23, 657-673. (7) Anderson, D. J.; Hites, R. A. Atmos. Environ. 1989, 23, 20632066. (8) Roinestad, K. S.; Louis, J. B.; Rosen, J. D. J. AOAC Int. 1993, 76, 1121-1126. (9) Simcox, N. J.; Fenske, R. A.; Wolz, S. A.; Lee, I.-C.; Kalman, D. A. Environ. Health Perspect. 1995, 103, 1126-1134. (10) Nishioka, M. G.; Burkholder, H. M.; Brinkman, M. C.; Gordon, S. M.; Lewis, R. G. Environ. Sci. Technol. 1996, 30, 3313-3320. (11) Nishioka, M. G.; Burkholder, H. M.; Brinkman, M. C.; Gordon, S. M. Simulation of Track-In of Lawn-Applied Pesticides from Turf to home: Comparison of Dislodgeable Turf Residues with Carpet Dust and Carpet Surface Residues; U.S. Environmental Protection Agency Report No. EPA/600/R-97/108; Research Triangle Park, NC; 1997 (NTIS Accession No. PB98-103120). (12) Pesticide Profiles; Kamrin, M. A., Eds.; Lewis Publishers: Boca Raton, FL, 1997. (13) An SAB Report: Assessment of Potential 2,4-D Carcinogenicity@; U.S. EPA; EPA-SAB-EHC-94-005; March 1994. (14) Dietz, R. N.; Cote, E. A. Environ. International 1982, 8, 419433. (15) Exposure Factors Handbook; U.S. Environmental Protection Agency: 1996; pp 14-60. (16) Rousseau, R. W.; Ferreli, J. K.; Macnie, R. N. Perspiration Poisoning of Protective Clothing Materials-Part IV; Technical Report: Natick/TR/84/009L, AD-B088-681; U.S. Army Natick Research, Development and Engineering Center, Natick, MA, June 1983. (17) Roberts, J. W.; Budd, W. T.; Ruby, M. G.; Bond, A. E.; Lewis, R. G.; Wiener, R. W.; Camann, D. E. J. Exp. Anal. Environ. Epi. 1991, 1, 143-155.
Received for review June 8, 1998. Revised manuscript received February 1, 1999. Accepted February 15, 1999. ES980580O
VOL. 33, NO. 9, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1365
