Table 3 provides a description of the 2002 field samples along with the total and speciated arsenic results. Total arsenic was measured above a detection limit of 0.04 μg/g in all 27 field samples. Overall, the concentrations of total arsenic found in the present study were similar to the baseline data appearing in Table 1. The nominal ordering of total arsenic concentrations among the species tested in the present study was as follows: Summer flounder (incoming), summer flounder (outgoing), hard clam, striped bass (bay), striped bass (flats), and Atlantic croaker. This ordering, along with the center and spread in the total arsenic data, is shown in Figure 2. To our knowledge, the arsenic speciation results shown in Table 3 are the 1st such data to appear in the literature for marine fish and shellfish collected from Mid-Atlantic coastal waters.
Table 4 presents the detection frequencies for the various forms of arsenic in the field samples, and Table 5 presents summary statistics for total arsenic and DMA in the field samples. Table 5 also presents summary statistics for inorganic arsenic in hard clam and MMA in striped bass. Summary statistics are not presented for inorganic arsenic in species other than hard clam or for MMA in species other than striped bass because of the low frequency of detection for such cases. For the purposes of Table 5, in the few cases when nondetections were involved, true values were assumed to be present at one-half the detection limit.
Total arsenic in summer flounder
Total arsenic in the incoming flounder spanned from a minimum of 1.8 μg/g ww to a maximum of 3.33 μg/g ww. In the outgoing flounder, total arsenic ranged from a minimum of 0.95 μg/g ww to a maximum of 2.13 μg/g ww. The median total arsenic concentration in the incoming flounder, 2.58 μg/g ww, is statistically greater than the median concentration in the outgoing flounder, 1.49 μg/g ww (Mann-Whitney, p = 0.0367). This difference was not related to differences in length between the incoming and outgoing fish. Although the incoming flounder as a group were slightly longer than the outgoing flounder (median total length: incoming flounder, 475 mm; outgoing flounder, 445 mm), this difference was not statistically significant (Mann-Whitney, p = 0.3457). Furthermore, no underlying statistically significant relationship was found between individual lengths of flounder and individual total arsenic concentrations (ANOVA, p = 0.6315). We therefore conclude that the arsenic concentration in the incoming flounder is, indeed, greater than that in the outgoing flounder and that this difference is not explained by differences in length.
To provide broader regional context for the summer flounder data collected in the present study, we compared our results to summer flounder data collected approximately 200 km to the north in the New York Bight Apex. Scientists from the National Oceanic and Atmospheric Administration reported an average total arsenic concentration of 1.72 μg/g ww in 14 summer flounder fillet samples collected in September 1993 (Deshpande et al. 2000). The raw data of those authors indicate a minimum, maximum, and median of 1.22, 2.34, and 1.61 μg/g ww, respectively, for total arsenic. These values are similar to the results of the summer flounder samples collected from the Delaware Inland Bays in the fall of 2002 (outbound fish). However, the concentrations in the 1993 New York Bight samples are nominally lower than those in the summer flounder samples collected from the Delaware Inland Bays in the spring of 2002 (inbound fish).
Total arsenic in striped bass, Atlantic croaker, and hard clam
For striped bass, the total arsenic concentration in the fish collected from the Delaware Bay ranged from 1.07 to 2.19 μg/g ww, with a median of 1.17 μg/g ww. In comparison, the total arsenic concentration in the striped bass collected upstream at the Cherry Island Flats ranged from 0.36 to 1.58 μg/g ww, with a median of 0.88 μg/g ww. The median concentration in the Delaware Bay fish was greater than the median concentration in the fish from the Cherry Island Flats when viewed at the 90% confidence level but not when viewed at the 95% confidence level (Mann-Whitney, p = 0.0947). Median lengths between the 2 striped bass groups were not different (Mann-Whitney, p = 0.2948), and no underlying statistically significant relationship was found between total arsenic and length among individual striped bass (ANOVA, p = 0.7623).
Total arsenic in the 5 Atlantic croaker samples ranged from 0.48 to 0.8 μg/g ww, with a median of 0.79 μg/g ww. No statistically significant relationship was found between length and total arsenic for croaker (ANOVA, p = 0.3554). The range of total arsenic in the 5 croaker samples from the Inland Bays was within the range of 0 to 2.1 μg/g ww reported by the USEPA for Atlantic croaker samples collected from the Louisianian Province (Summers et al. 1992). Finally, total arsenic in the 2 hard clam samples collected from the Inland Bays was 0.93 and 1.53 μg/g ww, with a median of 1.23 μg/g ww.
MMA and DMA
Monomethylarsonic acid was detected in 17 of the 27 samples, for an overall detection frequency of 63%. All 17 detections were “J” qualified. The minimum detected MMA concentration was 0.00115 μg/g ww, and the maximum was 0.00432 μg/g ww. The trimmed mean MMA concentration in the incoming flounder was 0.00229 μg/g ww, whereas that for the outgoing flounder was nearly equal at 0.00233 μg/g ww. The trimmed means for the striped bass samples were 0.00323 μg/g ww for the samples collected from the Delaware Bay and 0.0029 μg/g ww for the striped bass collected farther up estuary near the Cherry Island Flats. No significant relationship was found between MMA and fish length.
Dimethylarsinic acid was detected in 26 of the 27 samples, for an overall detection frequency of 96%. All 10 detections in summer flounder and all 9 detections in striped bass were “J” qualified. In contrast, only 1 of the 5 croaker detections was “J” qualified, and neither of the detections in clam were “J” qualified. The peak DMA concentration, 0.528 μg/g ww, was found in 1 of the hard clam samples. The median for the 2 clam samples, 0.398 μg/g ww, was much greater than any of the other species tested. Croaker had the 2nd highest median DMA concentration at 0.057 μg/g ww. The median DMA concentration in the incoming flounder, 0.0079 μg/g ww, was not statistically different than the median in the outgoing flounder at 0.0066 μg/g ww (Mann-Whitney, p — 0.83). Similarly, the median DMA concentration in the striped bass collected from Delaware Bay, 0.018 μg/g ww, was not significantly different from the median in the striped bass collected near the Cherry Island Flats at 20 μg/kg (Mann-Whitney, p = 0.60). Dimethylarsinic acid was not related to length for flounder or striped bass; however, DMA did show a strong inverse relationship to length for croaker (ANOVA, p = 0.0294, R2 = 83.7%).
Figure Figure 2.. Variation in total arsenic concentration in selected marine fish and shellfish species from the Mid-Atlantic region, USA.
Download figure to PowerPoint
With regard to the DMA levels in the hard clam and croaker samples, it is important to note that samples for both species were collected within 2 weeks of each other in the middle of the summer (15 July 2002 to 30 July 2002). It is postulated that the appearance of DMA in these species is related to arsenic transformations that occur in phytoplankton during this time period, coupled with attendant food-chain exposure. It is recognized that phytoplankton take up arsenate from aqueous solution, reduce the arsenate to arsenite, and methylate the arsenite to produce MMA and DMA, which are then excreted (Phillips 1990; Neff 1997). Recently, Hellweger and Lall (2004) successfully modeled the time course of this transformation by showing the gradual appearance of DMA during the summer period, when algal growth rates slow under phosphorus-limiting conditions. Considering that hard clams feed directly on phytoplankton by filtering large quantities of water, the clams likely were exposed to DMA primarily through their algal diet and, to a lesser extent, through the water that they pump. Croaker forage for a variety of organisms on and in the surface sediments, including small mollusks (e.g., clams and snails) (Diaz and Onuf 1985). Being bottom-feeders, they also inadvertently ingest detritus, which may include algae in various stages of decay. Croaker therefore are exposed to DMA through their diet. We speculate that DMA concentrations are lower in the croaker than in the clam because the croaker is migratory whereas the clam is not, resulting in shorter durations of exposure for the croaker. Finally, the inverse relationship between DMA and croaker length may reflect a higher proportion of phytoplankton in the diet of smaller croaker. This also may explain why DMA was detected only in the smallest flounder.
Inorganic arsenic was detected in only 5 of the 27 samples, for an overall detection frequency of 19%. A single detection occurred in each of the finfish species and in both clam samples. All detections were “J” qualified. The single detection in summer flounder, 0.00048 μg/g ww, occurred in the smallest flounder caught in the present study. It also was the 1st flounder caught in the fall. The detected concentration in croaker was 0.00057 μg/g ww, whereas the detected concentration in striped bass was 0.00168 μg/g ww. The peak inorganic arsenic concentration was found in 1 of the hard clam samples at 0.02 μg/g ww; this was the same sample that had the peak DMA detection. The median inorganic arsenic concentration for the 2 clam samples was 0.0145 μg/g ww. These inorganic arsenic concentrations are exceedingly low but, nevertheless, are consistent with levels reported by the EPA and by Schoof for several marine fish species (Schoof et al. 1999; USEPA 2003).
Figure 3 shows the percentage inorganic arsenic in the samples assuming that the nondetected inorganic arsenic results actually were present at one-half the detection limit of 0.03 μg/g ww. Using this assumption, mean percentage inorganic arsenic values ranged from a high of 1.7% for croaker to a low of 0.7% for summer flounder. For the 2 hard clam samples, in which no assumptions were necessary regarding detection levels, the percentage of arsenic in the inorganic form was 0.96% and 1.3%, yielding an average of 1.1%. The overall mean for all 27 samples was 1.2%, again assuming that nondetected values were present at one-half the detection limit. The percentages found in the present study are within the range reported in the literature for a variety of marine fish and shellfish species (Nriagu and Simmons 1990; Chew 1996; Donahue and Abernathy 1999; Johnson and Roose 2002; Kirby and Maher 2002; De Gieter et al. 2002; USEPA 2003).
Table Table 4.. Frequency of detection of arsenic in Mid-Atlantic (USA) marine fish and shellfisha
|Speciesb||Sample size (n)||Total arsenic (μg/g wet wt)||Inorganic arsenic (μg/g wet wt)||MMA (μg/g wet wt)||DMA (μg/g wet wt)|
|Summer flounder (inbound)||5||5||0||3||5|
|Summer flounder (outbound)||5||5||1||3||5|
|Striped bass (bay)||5||5||0||4||5|
|Striped bass (flats)||5||5||1||4||4|
Table Table 5.. Summary statistics for arsenic in Mid-Atlantic (USA) fish and shellfisha
| ||Total arsenic (μg/g wet wt fillet)|
| ||Summer flounder (inbound)||Summer flounder (outbound)||Atlantic croaker||Hard clam||Striped bass (bay)||Striped bass (flats)|
| ||DMA (μg/g wet wt fillet)|
| ||Summer flounder (inbound)||Summer flounder (outbound)||Atlantic croaker||Hard clam||Striped bass (bay)||Striped bass (flats)|
| ||Other (μg/g wet wt fillet)|
| ||Hard clam inorganic arsenic||Striped bass (bay) MMA||Striped bass (flats) MMA|| || || |
|Count||2||5||5|| || || |
|Average||0.01454||0.00359||0.00332|| || || |
|Median||0.01454||0.00342||0.00255|| || || |
|Standard error||0.00556||0.00038||0.00056|| || || |
|Minimum||0.00898||0.00276||0.00228|| || || |
|Maximum||0.02010||0.00500||0.00500|| || || |
Risk assessment and fish advisory
Human health risk associated with the consumption of chemical contaminants in seafood depends on the concentration of the contaminant in the seafood (Cf), the consumption rate of seafood (CR), the exposure frequency (EF) and duration (ED), the body weight of the consumer (BW), the averaging time (AT), and the dose-response metric for the chemical (e.g., q1* for the cancer endpoint and RfD [reference dose] for noncancer effects). For the cancer endpoint, lifetime risk can be calculated as (Cf)(CR)(EF)(ED)(q*1)/((BW)(AT)). For noncancer effects, risk is calculated as (Cf)(CR)(EF)(ED)/((RfD)(BW)(AT)).
In seafood, arsenic can exist in many forms, including the inorganic species arsenite (III) and arsenate (V), as well as the organic species arsenobetaine, arsenocholine, MMA, DMA, arsenosugars, and arsenolipids (Cullen and Reimer 1989; Chew 1996). The question then arises as to which concentration should be specified for Cf in the above risk equations. Current guidance indicates that the sum of the inorganic species arsenite and arsenate should be used (USEPA 2000a). The various organic arsenic compounds in fish and shellfish are considered to be relatively nontoxic (ATSDR 2000). For purposes of the current risk assessment, Cf in the above equations was set equal to the single inorganic arsenic concentration detected in each of the fish species. For hard clam, Cf was taken as the average inorganic arsenic concentration detected in the 2 clam samples.
The EF value was set to 350 d/y, whereas the ED value was set to 30 y. This latter assumption is consistent with the standard assumption used in Delaware's fish contaminant monitoring program (DNREC and DHSS 2005). For the cancer endpoint, AT was set to 365 d/y multiplied by 70 y, or 25,550 d. For the noncancer endpoint, AT was set to 365 d/y multiplied by 30 y, or 10,950 d. Assumptions for EF, ED, and AT are consistent with Federal Superfund guidance (USEPA 1989; USEPA 1991). Body weight was taken as 70 kg, also to maintain consistency with standard risk assessment practice (USEPA 1991). A cancer slope factor (q*1) of 1.5 mg/kg/d was used for the cancer assessment, whereas an RfD value of 0.0003 mg/kg/d was used for the noncancer assessment (http://www.epa.gov/iris). Both the slope factor and the reference dose apply to inorganic arsenic.
Figure Figure 3.. Inorganic arsenic as a percentage of total arsenic in selected marine fish and shellfish from the Mid-Atlantic region, USA. Bars are means, and error bars are standard errors. Nondetected values are assumed to be present at half the detection limit.
Download figure to PowerPoint
With regard to seafood consumption rate, site-specific quantitative information for the Delaware Inland Bays is not available. Detailed information is available, however, for the nearby Delaware Estuary, where a study-wide average consumption rate of 17.5 g fish/d was determined for recreational anglers and their households (KCA 1994). Coincidentally, this is the same consumption rate used by the USEPA in deriving fish tissue chemical-contaminant screening values for recreational anglers (USEPA 2000b). This consumption rate was used for purposes of the current risk assessment.
Using the values and equations above, the calculated cancer and noncancer risks are shown in Table 6. From these results, consuming these species increases the lifetime cancer risk from as little as 7.4 × 10−8 (i.e., 1 additional possible cancer in a population of 13.5 million people) for flounder up to 2.2 × 10−6 (1 additional possible cancer in a population of ∼450,000 people) for hard clam. Given the uncertainties inherent in current cancer risk assessment, the actual risks may be lower and could actually be zero. The true cancer risk, which is unknown, is unlikely to be greater than that shown in the table. With regard to noncancer risks, we see that noncancer health risks are all well below a hazard index of 1 and, therefore, are of little concern. In sum, the cancer and noncancer health risks shown in the table are considered de minimus by the health and natural resources agencies that are responsible for issuing fish consumption advisories in Delaware. As such, no advisory was believed to be warranted based on the measured levels of inorganic arsenic in these species.
Table Table 6.. Summary of human health cancer and noncancer risks associated with ingestion of inorganic arsenic in Mid-Atlantic (USA) marine fish and shellfisha
|Species||Lifetime cancer risk||Noncancer health risk|
It is interesting to consider the risks and potential risk management actions that may have ensued if the concentration of inorganic arsenic had been estimated by multiplying total arsenic by 10%, which is a default assumption recommended by the US Food and Drug Administration when only total arsenic data are available (USFDA 1993). For example, the median concentration of total arsenic in the incoming flounder was 2.58 μg/g ww. Ten percent of that value yields an estimated inorganic arsenic concentration of 0.258 μg/g ww. Substituting this value into the cancer risk equation, we see that lifetime cancer risk becomes 4 × 10−5 (1 additional possible cancer in a population of 25,000). Under current risk management policies, an advisory may have been issued in this case, especially if considered in conjunction with other contaminants present in the fish.