Chlorinated and brominated dioxins and dibenzofurans in human tissue following exposure.

With substantial improvements in analytic techniques over the past decade, it has become possible to measure polychlorinated dioxins (PCDDs) and dibenzofurans (PCDFs) in human tissue in a congener-specific fashion down to the low parts per trillion level. This paper reviews findings using these new techniques from a number of recent medical and environmental case studies. These studies include those of workers exposed to a polychlorinated biphenyl (PCB) transformer fire in the United States, German chemical workers exposed to 2,3,7,8-tetrachlorodibenzodioxin (2,3,7,8-TCDD) while cleaning up after an explosion, workers at a municipal incinerator in New York City, a chemist exposed to brominated and chlorinated dioxins, U.S. veterans and also Vietnamese civilians exposed to Agent Orange contaminated with TCDD in Vietnam, and victims of the polychlorinated dibenzofuran and PCB contaminated rice oil (Yusho) incident in Japan.


Introduction
During the 1980s, markedly improved extraction and separation methods, improved capillary columns, availability ofdioxin and dibenzofuran standards, and improvements in gas chromatography-mass spectrometry, greatly increased the ability of chemists to identify and quantitate chlorinated dioxins and dibenzofirans in human tissues down to the low parts per trillion level, and sometimes even lower. These technical advances permit dioxin measurements in human tissue to be used in clinical medicine and epidemiology. ' Despite these rapid analytical advances and demonstrations of the usefulness ofdioxin measurements in human tissue following potential exposure, the measurement of dioxins in human tissues to characterize exposure ofindividuals or groups is rarely used at the present time. This is partly due to the cost ofanalyses, between $1,500 and $2,500 each, the scarcity of qualified laboratories (there are between 10 and20 worldwide), the unwillingness of health insurance carriers to pay for these analyses, and, perhaps most important, the lack ofawareness among health care professionals concerning such analyses usefulness and availability.
In an attempt to show how such measurements can be of use in either clinical medicine or epidemiology, we review here selected congener-specific findings from our work this past decade, from various medical case histories. Measurement of dioxins in human tissue and food was first performed by Baughman between 1970 and1974 (1,2). He measured 2,3,7,8-TCDD in milk collected from Vietnamese women then living in villages that had been sprayed with Agent Orange and also in food samples. Measurement of2,3,7,8-TCDD, the dioxin characteristic ofAgent Orange, in adipose tissue of U.S. Vietnam veterans was first reported in 1982 and published in 1984 (3). Elevated dioxin and dibenzofuran levels in blood and fat of American workers following exposure and measured levels in controls from the general population were first reported in the SCHECTER ETAL. United States in 1983 (4) and subsequently published (5-7). At the same time, levels of dioxins and dibenzofurans in adipose tissue from the general population were reported for Canadians (8). Later work refined the techniques used for the measurement of dioxins in American workers and controls (9).

Methods
The analytical methods used for identifying and quantifying dioxins and dibenzofurans have been described elsewhere and are therefore only referenced here (10)(11)(12)(13)(14).
Human tissue was placed in chemically clean containers, usually glass bottles with Teflon-lined caps, and frozen. The tissue was stored frozen between -200C and -700C until analyzed.

Results and Discussion
PCB Transformer Fire in Binghamton, New York, 1981 Figure 1 presents total dioxin and dibenzofuran levels in adipose tissue from four potentially exposed workers and the mean values from matched controls obtained at a local hospital (7). All workers were potentially exposed to polychlorinated dibenzofurans and polychlorinated dibenzodioxins generated from a 1981 PCB transformer fire in Binghamton, New York (15)(16)(17)(18). Fat biopsies and dioxin analyses were firstperformed 2 years after the fire. These were the first dioxin and dibenzofuran measurements reported in the United States for potentially exposed workers and also for the general population. At the time, finding dioxins and dibenzofurans in the general population controls was unexpected. Worker #1 did not have elevated dioxin or dibenzofuran levels compared to these controls. Worker #2 had elevated dioxins but not dibenzofurans, however, more dibenzofurans were produced by PCB pyrolysis from the transformer fire than were dioxins. The findings ofboth elevated PCDDs and PCDFs in worker #3, is consistent with intake of these compounds from the incident, the only known special exposure in his case. Worker #4 shows elevated PCDF levels, but not elevated PCDD levels. The PCDD/F control values found in this human tissue study at about 1500 ppt were slightly higher than those in later studies, where we found mean PCDD/F levels of 1200 ppt in fat tissue.
In the absence of measurements taken before the incident or serial measurements taken over time, it was not initially obvious that elevated levels were from this incident, although the levels in some of the potentially exposed workers were clearly above control values. With the potentially exposed worker whose PCDD/F levels were below the controls, it was also not possible or reasonable to exclude the possibility of some intake of PCDD/Fs from the incident. However, it appeared reasonable to conclude that if total dioxins, total dibenzofurans, and also individual congeners were elevated and corresponded to the congeners found in environmental samples at the site, then the elevated levels probably resulted from this specific incident ofexposure. From a health perspective, conversion ofthe dioxins and dibenzofurans from measured values to their 2,3,7,8-TCDD toxic dioxin equivalents (TEq), comparing their toxicity with TCDD is sometimes useful (19)(20)(21). This conversion is used for the controls and the four workers in Figure 2. This shows that for the age, sex, and geographically similar hospital controls, the average total TEq is 50 ppt with 35 ppt from dioxins and 15 ppt from dibenzofurans. We later found 20-40 ppt TEq to be the usual levels for North American adults. For worker 1, the total is 32, for workers 2, 3, and 4, the values are 110, 92, and 109 ppt, respectively. Ifthese values should prove to be above a threshold for toxic response, then there would be not only documentation of exposure and intake, but also cause for concern about the health of these workers.
In Figures 3-6, individual lipid-adjusted congener levels of adipose tissue are compared for these workers and controls to provide more detailed comparisons. The pyrolysis of PCBs and polychlorinated benzenes in the transformer oil produced a mixture ofPCDDs and PCDFs, especially ofthe higher chlorinated congeners (15)(16)(17)(18).PCDFs were generally present in the soot in greater amounts than were PCDDs. In Figure 3, 2,3,7,8-TCDD, quantitatively a minor component in some of the soot samples tested after the fire, is found to be somewhat elevated in all four FiGuRE  workers when compared to the controls. It should be noted that half-life for elimination of 2,3,7,8-TCDD is believed to be longer than for the dibenzofurans. Figure 4 presents the workers' level of 2,3,4,7,8-penta(Pe)CDF. Worker 1 is below the control level, whereas the remaining three workers have elevated levels. Figure  5 shows elevation above background for workers 3 and 4 of 1,2,3,4,7,8 and 1,2,3,6,7,8-hexa(Hx)CDF. Figure 6, showing levels of l,2,3,4,6,7,8-hepta(Hp)CDF, documents only worker 4, who had elevated levels for this congener, which was also present in some soot samples. Exposure to various congeners, intake, and elimination may be expected to vary, but the congener patterns as well as total PCDD and also PCDF values seem to be consistent with some PCDD/F intake for most in this group of workers. These levels are not as high as those later reported from other dioxin tissue analyses in different incidents (22)(23)(24). A different approach was used for another worker who was involved in the cleanup and who exhibited the highest dioxin and dibenzofuran adipose tissue levels of those in the Binghamton ./ 46 23  incident (12,25). For a variety of reasons, including an attempt to determine whether these chemicals were from the work-site exposure, serial fat and later blood samples were obtained for PCDD/F analyses. Figure 7 presents 2,3,4,7,8-PeCDF results from 10 samples collected over 7-year period, which began 2 years after this worker's 6-month exposure in 1981. In Figure 8, results for 1,2,3,4,7,8 and 1,2,3,6,7,8-HxCDF are presented, and the same general pattern of a steady decline over time is noted. This decline over time is consistent with the exposure in 1981. Slight increases in the 1990 levels are believed to be due to analytic variability but other explanations may exist. It is, of course, quite rare for a patient to volunteer for repeated fat biopsies or donation of 100 to 450 mL of blood for dioxin analyses. As techniques improve, smaller blood specimens may be used, and the possibility of serial blood sampling for dioxin measurement should become less difficult for all involved. The serial determinations done here are consistent with and similar to studies ofYusho patients and others exposed to dibenzofurans where all findings suggest shorter half-lives ofelimination for certain ofthe dibenzofur-ans than the 5-10 year value, currently believed to represent an approximate half-life for TCDD in humans (26)(27)(28) TCDD Exposure of Workers at the BASF Factory in Ludwigshaven, Germany, 1953 Figure 9 presents adipose tissue levels ofTCDD in six German workers 32 years after they worked as cleanup workers in the TCDD-contminated BASF factory in Ludwigshaven, Germany (31). When exposure is high, dioxins can be found in human tissues at elevated levels for decades after exposure. These workers developed chloracne and other more serious illnesses during and after the cleanup. The workers wanted to have evidence for a type ofworkers' compensation insurance hearing to document intake of dioxins and its possible relationship to subsequent illnesses, which led to their request for fat tissue dioxin analysis. The adipose tissue TCDD levels in 1985 varied from 11 to 141 ppt lipid. We estimate levels in 1953, the time of exposure, to have been between 917 and 11,750 ppt. The mean level for German controls is about 4 ppt. The estimated body burden for TCDD for these workers, shown in Figure 10, varied from 9.7 to 124 jg. Their estimated dose, in nanograms per kilogram of body weight, varied from 140 to 1800. For these calculations we assumed first-order kinetics, a 5 year half-life, a singlecompartment model, and also assumed that most TCDD is found in fat tissue. These levels were sufficient to cause chloracne and apparently other illnesses in each of these workers.
Municipal Incinerator Workers, New York City, USA, 1990 Municipal incinerators characteristically produce a mixture of congeners, including tetra through octachlorinated dioxins and dibenzofurans, by burning organic compounds in the presence of chlorine. Because of concerns about bioavailability of dioxins and dibenzofurans from incinerator ash where elevated PCDD/Fs were found in meat and milk from cattle grazing near incinerators in Holland (32), incinerator workers' blood was examined (33). Pooled blood of workers from a relatively old incinerator in New York City was measured for PCDD/Fs, and the percentage increase, by congeners, was compared to a matched New York City control group ofpooled blood (Fig. 11). A consistent pattern of PCDD/F congener elevation was found, with the exception of TCDD. This is similar to the findings in both meat and milk from Dutch cattle grazing near a municipal incinerator. The levels found document bioavailability and suggests that there is a hazard to workers from this currently popular method ofhousehold garbage disposal. To a lesser extent, it suggests a hazard to those living near incinerators or the dump sites for their ash. Better and more modem incinerators, and those which do not burn chlorinated products, may sometimes produce less PCDD/Fs. Personal protective measures were used for these municipal incinerator workers after these findings.   Chemist Accidentally Exposed to TCDD and TBrDD in 1956 Figure 12 illustrates an occupational hazard for dioxin chemists. A chemist was exposed to dioxins 34 and 35 years before blood samples were taken for analysis (13). He had first synthesized and then purified 2,3,7,8-tetrabromo-p-dioxin aches, and severe pain in his legs when he walked. He was hospitalized primarily for medical evaluation and rapidly recovered, with only symptomatic treatment for his skin lesions. This incident was first described by Baughman in his 1974 Ph. D. thesis (1). In 1990 and again in 1991, approximately 35 years after exposure, blood dioxin analysis was performed to determine if a residual increase in dioxin levels existed. We found 20 ppt of TXDD and 625 ppt ofTBrDD on a blood-lipid basis. This compares to the 3-5 ppt ofTCDD usually measured in adults in the United States; 2,3,7,8-TBrDD has not been found in the general population despite measurements with a detection level in the low part per trillion range. exposure. Corrections were made for assumed half-life and also for assumed increased fat content with age in these workers, which is believed to dilute present levels.

U.S. Military Personnel Exposed to Agent Orange in Vietnam
Vietnamese Exposed to TCDD from Agent Orange Similar findings, but with sometimes higher levels of 2,3,7,8-TCDD, have been reported from the Vietnamese population who lived in areas sprayed with Agent Orange in the south of Vietnam (35)(36)(37)(38). Breast milk 2,3,7,8-TCDD levels from residents of Tan Uyen Province are shown in Figure 14 (1,35). One sample measured in 1970 (the last year of spraying), had 1450 ppt lipid ofTCDD, the highest dioxin level found in human breast milk to date. In 1973, levels from 77 to 400 ppt were found in human milk lipid from Vietnamese women nursing in these same villages. Between 1985 and 1990, we (3S-38) collected and measured dioxins in milk from nursing mothers residing in sprayed areas in Vietnam and found much lower levels ofTCDD. In the north ofVietnam, where no spraying occurred and where little industrial pollution existed, even lower levels ofTCDD (as well as other dioxins from other sources) were found; here there was a TCDD level of 2.1 ppt lipid) from a pool of 30 nursing mothers living in Hanoi in 1984. The elevated TCDD findings document exposure to and actual intake of varying amounts of this dioxin from the same incident in both the workers who performed the spraying and the residents who lived in the sprayed area. In this instance, the same incident produced both an occupational and an environmental hazard. Lesser concentrations of dioxin-contaminated phenoxyherbicides, such as 2,4,5trichlorophenoxyacetic acid, were widely used in the United States and elsewhere. The more intense application ofphenoxyherbicides in Vietnam may provide estimates ofan upper exposure to and intake ofTCDD from these herbicides in agricultural sprayers, and other military personnel, and also in the sprayed Vietnamese, who are the largest group potentially contaminated. Agent Orange, characterized by 2,3,7,8-TCDD contamination, generated a great deal of interest in this dioxin congener. Figure  15 compares 2,3,7,8-TCDD levels to total dioxin and dibenzofuran TEq levels in adipose tissue obtained in 1984 from Vietnamese hospitalized in the south ofVietnam in Ho Chi Minh City and from residents of the north of Vietnam in Hanoi (38). The Vietnamese population's exposure to Agent Orange was of lengthy duration due to their continued residence in the contaminated area, whereas the U.S. veterans served in the affected area for 1 year, or rarely for 2 years. Total PCDD/F TEq levels in adipose tissue from Vietnamese living in Vietnam in the 1980s varied between 12 and 143 ppt in Ho Chi Minh City (formerly Saigon and surrounding areas) with TXDD ranging from not detected with a detection limit of2 ppt to 103 ppt. In light of recent epidemiological findings linking phenoxyherbicides (in some cases not contaminated with TCDD) and 2,3,7,8-TCDD to cancer in humans (39)(40)(41)(42)(43)(44)(45)(46), these findings may be of clinical concern.
In Figure 16, which depicts 2,3,7,8-TCDD and total PCDD/F TEqs in blood from Vietnam, geographical variation in dioxin levels can be observed. The northern samples have characteristically low levels, and in the south, long after Agent Orange contamination, there are still populations with higher levels, with regional variations noted in central and southern Vietnam. These large populations with higher levels (in the south) and lower levels (in the north) present a unique opportunity for studying dioxin body burdens associated with disease states.
Yusho Rice Oil Poisoning with PCDFs and PCBs, Japan, 1968  (48,49). In this incident, rice oil used for cooking was contminted with PCBs, PCDFs, and PCDDs. About 2000 persons became sick and were treated by a special medical team. The illnesses were similar to those documented in animal studies with dioxins and with PCBs and from occupational exposures to dioxins and dioxin-like chemicals. It is believed that the toxic effects, based on amounts present and relative congener toxicity were primarily from the dibenzofurans, especially 2,3,7,8-substituted penta-, hexa-and hepta-chlorinated congeners. PCBs are believed to have contributed to a lesser extent, and a small contribution to total toxicity is thought to have come from the dioxins and other chemicals. An adipose sample was taken from a victim 17 years after the 1968 incident to see if elevated levels could still be detected. Due to the very high dose, three ofthe dibenzofurans typical ofthe Yusho poisoning were still found to be elevated in this patients' adipose tissue. These are 2,3,4,7,8-PeCDF, total HxCDF, and 1,2,3,4,6,7,8- HpCDF. An almost identical incident occurred in Taiwan in 1979 with involvement of similar congeners (50,5). The Binghamton incident, as noted previously, showed elevation ofsome ofthese congeners in some exposed workers. Figure 18 presents the Yusho data above for total PCDD, PCDF, and PCDD/F levels after conversion to TEqs. The PCDFs are seen to be markedly elevated (5386 versus 113 ppt for controls). Because ofthis, total PCDD/Fs are higher than in controls, 5764 as compared to 1648 ppt. PCDD toxic equivalents are identical in both, but PCDF TEqs are elevated, at 1363 ppt in this Yusho patient compared to 18 ppt in the controls. Thus, the total TEq is still markedly higher (1387 ppt compared to 42 ppt in controls), even in the specimen obtained 17 years after exposure. A4Kc$K" FIGURE 18. Yusho poisoning in Japan. PCDD/F and total (TEq) levels in a Yusho patient and in controls (ppt, lipid).

Conclusions
We have reviewed some relatively recent incidents of dioxin and dibenzofuran contamination where human tissue PCDD/F levels were determined to illustrate the use ofsuch measurements in evaluation of exposed populations. Human tissue measurement for dioxins and dibenzofurans is relatively new and is just now beginning to be exploited in medicine and epidemiology. This is in part because the recent origin ofmost chlorinated dioxins and dibenzofurans. Also, the measurement ofthem in the environment and in humans is historically quite recent (52)(53)(54)(55)(56)(57).
The serial determinations in a worker after exposure following a PCB transformer fire add to the evidence that some of the dibenzofuran congeners are eliminated in humans more rapidly than is the case for 2,3,7,8-TCDD. Thus, time after exposure is an important consideration in use and interpretation of dioxin tissue measurement. Also, relatively low levels, as seen in one of our Binghamton workers, may or may not have been higher than what existed before exposure, ifthe initial levels were below average general population values. Similar findings have been reported in published studies of "Ranch Hand" Agent Orange sprayers, where no detectable levels of TCDD were found in some subjects despite probable heavy exposure (58). Variation in absorption or excretion, laboratory error in analysis, or even mislabeling or mixing ofspecimens might be possible explanations for unexpectedly low values.
For the German workers and the dioxin chemist, detection of elevated chlorinated and brominated dioxins up to 35 years after exposure documents remarkable persistence for 2,3,7,8-substituted TCDD and TBrDD and points out potential long-term health hazards. The finding ofelevated levels in these cases also may be useful in workers' compensation hearings and as a way to estimate potential risk above background in populations or in individuals. The levels related to human health effects are not yet well characterized. There is recent evidence suggesting that humans may be at least as sensitive to certain dioxin-induced biochemical end points as are laboratory animals (59,60). Another case history illustrated dioxin contamination of American veterans and Vietnamese residents from herbicides contaminated with dioxins more than 20 years after the last application of Agent Orange in Vietnam.
There is also evidence that there may be neurodevelopmental and other adverse health outcomes in children exposed in utero and/or by nursing to the closely related dioxinlike compounds PCBs and PCDFs (61)(62)(63)(64)(65). It would seem prudent to consider adverse health outcomes such as neurological, immune deficiency, or reproductive outcomes, rather than focus exclusively on the better-characterized cancer-inducing capability ofthe dioxins and related chemicals.
Little workhas been done so far to correlate fat and blood dioxin levels to target organ levels (66). Much work also remains to be done to characterize effects on blood dioxin levels during fasting and weight loss, during starvation, or during the wasting syndromes noted in diseases such as cancer, AIDS, and some other infectious diseases.
To date, these early studies on human tissue levels of dioxins and dibenzofurans appear to be promising with respect to their usefulness in medicine and medical research, but a great deal of work remains to integrate this valuable new technique into clinical medicine. Not the least of the problems is determining a routine method for funding the analyses when they are medically or scientifically indicated. Some of the studies cited here have been made possible by the generous assistance of the Christopher Reynolds Foundation, the Samuel Rubin Foundation, the CS Fund, Church World Service, CIDSE, the Commonwealth of Massachusetts, and the American Association forthe Advancement of Sciences and the National Academy ofSciences for the work done originally in Vietnam and on Vietnamese milk, food, and wildlife specimen from 1969 to 1974. The expert technical assistance in preparation ofthis manuscript is due to Ruth Stento and Karan Charles. Most of all, we extend our thanks to the patients involved.