Evaluation of toxicity of dioxins and dioxin-like PCBs: Executive Summary

Executive summary

• ES1 New Zealand exposure data
• ES2 Toxic effects of dioxin-like compounds and health risk appraisal considerations
• ES3 Appraisal of the New Zealand population exposure data
• ES4 Recommendations

The Ministry for the Environment has collected information on the sources and environmental levels of dioxin-like compounds in New Zealand, and on population exposures to these chemicals, including data on dietary intakes and concentrations in serum. These, and other data collected by the Ministry of Health on PCDD/F body burdens of New Zealanders, have been used in this health risk appraisal. This report also presents a concise review of our current understanding of the health risks associated with exposure to dioxin-like compounds, and reviews various health guidelines that have been established by other jurisdictions.

From consideration of the New Zealand exposure data, and a review of the published scientific literature on dioxin-like compounds, the following observations and conclusions can be made.

ES1 New Zealand exposure data

For the general population, over 90% of exposure to dioxin-like compounds is through the diet, with foods of animal origin such as meats, dairy products and fish usually the main source. Unborn children are exposed to dioxin-like compounds in utero, and nursing infants are exposed to these contaminants present in breast milk.

Based on a dietary study for dioxin-like compounds (Buckland et al., 1998c), the level of dietary intake of these chemicals for the New Zealand population is lower than exposures reported for any other country where a comparable study has been undertaken. For adult males with a median energy diet, the intake is estimated as 0.37 pg TEQ/kg bw/day, and for adolescent males with a high energy (90^th centile) diet, the intake is estimated as 0.84 pg TEQ/kg bw/day, where the toxic equivalents (TEQ) are based on the toxic equivalent factors (TEFs) developed in 1997 for dioxin-like compounds by the World Health Organization (WHO) (Van den Berg et al., 1998).

Similarly, the results of the study of dioxin-like compounds present in serum (Buckland et al., 2001) show that levels for the general New Zealand population are at the low end of the scale of levels reported internationally. The mean serum concentration across the population aged 15 years and older was determined as 19.7 ng TEQ kg^-1 on a lipid-adjusted basis (range: 9.71-38.5 ng TEQ kg^-1). From these serum data, body burdens and average lifetime daily exposures (ALDE) were calculated. The mean ALDE for all data was estimated as 1.4 pg TEQ/kg bw/day (minimum of 0.35 pg TEQ/kg bw/day for the population aged 15-24 years; maximum of 3.4 pg TEQ/kg bw/day) for the population aged 65+ years). The higher ALDE estimate compared to the estimated dietary intake is because the ALDE includes historical exposures, which are likely to have been higher than current exposures, as well as intakes from non-dietary exposure pathways.

A study undertaken in the late 1980s measured concentrations of PCDD/Fs in New Zealand mother's milk in the range 6.2-40 ng TEQ kg^-1 of milk fat. The preliminary indications from a second study currently underway is that over a 10-year period from 1987/88 to 1997/98, concentrations of PCDD/Fs in breast milk have fallen by about two-thirds.

Whilst noting the comparatively low levels of PCDD/F emissions relative to most other industrialized countries, as documented in the New Zealand dioxin inventory (Buckland et al., 2000), there is no clear evidence to show that the emissions of PCDD/Fs from known sources correlate proportionally with general population exposures. Although the inventory estimates the relative contribution of the various sources to total emissions, the uncertainties in these estimates mean it cannot be assumed that these sources make the same relative contributions to human intakes as calculated. Although unlikely, it is even possible that the major sources of PCDD/Fs in foods are not the sources with the largest fractions of estimated total emissions in New Zealand. However, it is clear that to prevent or minimise general population intakes of dioxin-like compounds, protection of the quality of New Zealand's agricultural lands and aquatic environments used for food production is paramount.

Higher-level exposures, such as may occur in the work place, are normally restricted to smaller groups of people. In New Zealand, historical occupational exposures to PCDD/Fs would probably have been restricted to individuals involved in the handling and use of the pesticides pentachlorophenol (PCP) and 2,4,5-trichlorophenoxyacetic acid. These chemicals are no longer used in New Zealand.

More detailed information on the New Zealand exposure data is provided in Section 3 of this report.

ES2 Toxic effects of dioxin-like compounds and health risk appraisal considerations

The most widely studied of all the dioxin-like compounds is 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD). It has been shown to affect a wide range of organ systems in many animal species, it can induce a wide range of adverse biological responses, and it is extraordinarily potent. Recent animal studies suggest that the most sensitive endpoints of TCDD exposure relate to immunotoxicity, and reproductive and neurobehavioral effects.

Our current knowledge of the mechanisms of TCDD toxicity clearly suggests that binding of TCDD to the aryl hydrocarbon receptor (Ah receptor) is the first step in a series of events that manifest themselves in biological responses, including changes at the biochemical, cellular and tissue level. Binding to the Ah receptor occurs in humans and also in monkeys and rats - the two animals used most in experimental studies. This being the case, it is reasonable to use the results of the animal studies to predict health effects that may not yet have been demonstrated in human studies.

2,3,7,8-TCDD is an extremely potent animal carcinogen. It has been shown to cause both benign and malignant tumors at multiple sites in several animal species. Increases in lung cancer and all-cancers combined have been observed in highly exposed cohorts of workers in industrial settings. From the occupational studies, and an understanding of biological plausibility as shown by animal studies, the International Agency for Research on Cancer has concluded that TCDD is carcinogenic to humans. Although the mechanisms of TCDD carcinogenicity are not fully known, the available evidence suggests that they do not involve direct damage to cellular DNA. Whatever the precise mechanism of TCDD carcinogenicity, no threshold for effects has been established. It is therefore plausible to include low-dose linearity in the range of possibilities, and for regulatory considerations we believe that the linear model should be the first choice unless it can be shown that it does not adequately fit the data.

The available epidemiological data indicate that dioxin-like compounds produce a variety of biochemical responses in humans, some of which occur at relatively low exposure levels. Enzyme induction, changes in hormonal levels and reduced glucose tolerance are examples of subtle changes that may occur at comparatively low exposures. However, these subtle effects are of unknown clinical significance, and may or may not indicate a toxic response or potential for a toxic response.

Despite the lack of data for toxic effects from dioxin-like compounds other than 2,3,7,8-TCDD, there is reason to infer that biochemical, cellular and tissue-level effects that are elicited by exposure to TCDD are also induced by other chemicals that have a similar structure and that bind to the Ah receptor. There is widespread agreement within the scientific community that the use of TEFs to estimate the relative toxicities of dioxin-like compounds has an empirical basis, is theoretically sound, and is a useful procedure. Furthermore, agencies such as the WHO and the United States Environmental Protection Agency (US EPA) have noted that the derivation and use of TEQ levels is a pragmatic and feasible approach for assessing human health risks from exposure to dioxin-like compounds. We concur with this approach, and endorse the use of the most recent (1997) WHO TEFs (Van den Berg et al., 1998), applying the concept of additivity for the PCDDs, PCDFs and dioxin-like PCBs to provide total TEQ estimates.

There are a number of key studies that document a variety of health effects of TCDD toxicity in laboratory animals, and these have been instrumental in establishing human health guidelines for exposure to dioxin-like compounds. Some of these studies are summarized in Table ES.1.

Table ES.1 Key studies used in the derivation of human health criteria

From these studies, the WHO, the United Kingdom, Germany, Japan and the Netherlands have all set a tolerable daily intake (TDI), and the United States Agency for Toxic Substances and Disease Registry (ATSDR) has set a minimal risk level (MRL) on the basis of the lowest observable adverse effect level (LOAEL) and the application of safety factors. In most cases, based on the non-cancer effects data, the TDI or MRL value set by these jurisdictions is, or falls within the range, 1-4 pg TEQ/kg bw/day. However, we note that the margins of safety used in deriving these values are very small, and we can have no assurance that some people with the highest exposures may not have deleterious effects even when the overall New Zealand population average intake, based on ALDE estimates, is approximately 1.4 pg TEQ/kg bw/day. It is also noted that, for non-cancer effects, the US EPA does not recommend the derivation of a reference dose (RfD) for dioxin-like compounds, because any RfD that the Agency would set is likely to be 2-3 orders of magnitude below current background intakes and body burdens.

The cancer study of Kociba et al. (1978) using Sprague-Dawley rats has been used to assess the carcinogenic risk from exposure to 2,3,7,8-TCDD. From the findings of this study, the US EPA has determined a risk-specific dose (RsD) of 0.006 pg TCDD/kg bw/day for a one in a million lifetime cancer risk. This was derived from an upper bound unit risk estimate of 1.6 x 10^-4 (pg/kg bw/day)^-1 (US EPA, 1985). Further analysis of the Kociba data has derived an oral intake RsD of 0.01 pg TEQ/kg bw/day, corresponding to a unit risk estimate of 1 x 10^-4 (pg/kg bw/day)^-1 (US EPA, 1994b), and most recently the US EPA have proposed an upper bound cancer risk estimate of 1.4 x 10^-3 (pg/kg bw/day)^-1 (US EPA, 2000) based on the Kociba rat data.

Although the evidence has limitations, more recent epidemiological studies in humans have led to TCDD being classified as a human carcinogen, and it is now possible to derive the cancer potency directly from human studies of exposed industrial workers. Thus, the US EPA has recently calculated a cancer potency factor from a meta-analysis of human data from three occupational cohorts of 1 x 10^-3 (pg/kg bw/day)^-1 (US EPA, 2000). This potency turns out to be much higher than initially estimated from the rat studies by the US EPA in their 1985 health assessment report (US EPA, 1985). However, recent reassessment of the Kociba rat study including incorporating the half-life differences between rats and humans has shown that the estimates for human cancer risks derived from this animal study come quite close to those that can be estimated directly from the pertinent studies of TCDD-exposed workers. This information adds to our confidence in recommending the use of the cancer potency estimates derived directly from the human studies to appraise the risks to the New Zealand population from exposure to dioxin-like compounds.

In Section 5, a critique is provided of the following risk assessment methodologies that are applicable to the current study:

• safety factor approach
• low-dose extrapolation
• benchmark dose or point of departure
• public health risk assessment.

The public health risk assessment approach, which takes into account current background exposures as well as allowing for consideration of single point sources of exposure adding to background, is considered to be the most appropriate means by which ongoing sources of dioxin-like compounds to the environment can be managed and exposures to the population reduced in the medium to long-term. Furthermore, we recommend that these chemicals be classified in the 'Class 2' category of the public health risk assessment framework.

Further discussion of the toxic effects of dioxin-like compounds is provided in Section 4, and on health risk appraisal considerations in Section 5 of this report.

ES3 Appraisal of the New Zealand population exposure data

The New Zealand serum data can be used to estimate exposure to dioxin-like compounds by relating body burden to an equivalent human daily intake, or ALDE estimate. Such intake estimates will include dietary intakes and other exposure pathways, such as inhalation. Because of the representative nature of the New Zealand serum study, both in terms of the large number of samples analyzed, and the incorporation of demographic, geographic, age, gender and ethnicity variables into the study design, it is a reasonable assumption that the ALDE estimates derived from the serum data are an accurate reflection of intakes across the population. Furthermore, these estimates of exposure are likely to be more reliable than the intake estimates derived from the comparatively smaller dietary study that were based on model diets.

The New Zealand intakes can be compared against the TDI target value established by the WHO and the MRL set by the ATSDR of 1 pg TEQ/kg bw/day. The average current dietary intake may be somewhat lower than this value (perhaps two times lower). However, the more reliable estimate of intake based on serum concentrations suggests that during approximately the last 25 years the average intake was probably close to 1.4 pg TEQ/kg bw/day. This being the case, about half the population would have exceeded this intake. It should also be noted again that these health criteria set by WHO and ATSDR involve very small margins of safety. Therefore, there would appear to be only a small margin of safety, if any, between New Zealand intakes and some non-cancer effects in animal studies; in particular, effects on the offspring of exposed mothers.

Cancer risk has been assessed by low-dose extrapolation, using both animal and human data. The cancer estimates from the human studies are similar to those from animal studies when the long human half-life is taken into account, and emphasis should therefore be placed on cancer risks derived from the human data. Overall, based on the human potency factors, the current appraisal has estimated that the upper bound lifetime risk for background intakes of dioxin-like compounds for the New Zealand population may exceed one additional cancer per 1000 individuals. This cancer risk estimate is 100 times higher than the value of 1 in 100,000 often used in New Zealand to regulate carcinogenic exposure from environmental sources. Of course, if there were a threshold above current exposures the actual risks would be zero. Alternatively, they could lie in a range from zero to the estimate of 1 in 1000 or more.

When assessing cancer risk using a benchmark dose derived from human data, the margin of safety between the average intake of dioxin-like compounds for the New Zealand population and the concentration estimated to result in a 1 in 100 cancer risk is found to be very small (less than an order of magnitude).

The exposure of breast-fed infants also warrants mention since after about six months of breast feeding, the body concentrations of dioxin-like compounds in infants exceed those in the mother as a consequence of the presence of these contaminants in breast milk. Whether adverse effects will result from this exposure is not known, although it is clearly known that breast feeding is in general beneficial. The concern about dioxin-like compounds in breast milk adds to the reasons for taking a prudent and precautionary approach concerning population exposures to these chemicals.

On the basis of the findings of this risk appraisal, the current background exposures to dioxin-like compounds for the New Zealand population has, in our opinion, an insufficient margin of safety, and steps should be taken to further reduce human exposure.This report also considers 'special' populations who may be exposed to dioxin-like compounds above background levels, including workers who were occupationally exposed to PCDD/F from the use of PCP in the timber industry. Based on a small number of cases, it is clear that workers who handled PCP have higher PCDD/F body burdens above those of the general New Zealand population. While it is not possible to draw conclusions about health effects from the small number of workers whose blood concentrations have been measured, it is clear that studies are needed of workers with these exposures. If a study is feasible, and would be of sufficient statistical power to assess health effects, we recommend that such a study of former New Zealand timber workers be undertaken.

More detailed information on the health risks to New Zealanders from exposure to dioxin-like compounds is provided in Section 7 of this report.

ES4 Recommendations

In the light of ever-increasing scientific information concerning the toxicity of dioxin-like compounds, and data on body burdens present in the New Zealand population, we make the following recommendations:

1. A precautionary approach should be adopted concerning dioxin-like compounds in New Zealand.
2. A goal of ongoing reduction in population body burdens of dioxin-like compounds should be stated.
3. Identifying a tolerable daily intake is not recommended.
4. A health exposure criterion (HEC) should be established to regulate point sources of exposure.
5. Application of the HEC should involve consideration of the plausible maximally exposed person from the point source activity.
6. The New Zealand population burden of dioxin-like compounds should be monitored periodically, perhaps every 5-10 years.
7. Policies and the HEC should be reviewed after consideration of trends revealed by future population monitoring.

Further information on each of these recommendations is provided in Section 8 of this report.