Home » Dioxins » Level 3 » Question 4


4. What are the effects of dioxins on human health?

  • 4.1 Have dioxins caused cancer to humans?
    • 4.1.1 Which studies have been used to evaluate dioxins carcinogenicity?
    • 4.1.2 Were cancers observed in highly contaminated populations?
  • 4.2 What non-cancer effects have been observed in children?
  • 4.3 What non-cancer effects have been observed in adults?

4.1 Have dioxins caused cancer to humans?

    • 4.1.1 Which studies have been used to evaluate dioxins carcinogenicity?
    • 4.1.2 Were cancers observed in highly contaminated populations?

4.1.1 Which studies have been used to evaluate dioxins carcinogenicity?

The source document for this Digest states:


In the evaluation of the evidence of effects of PCDDs, PCDFs and PCBs, only studies with serum or adipose tissue measurements were considered.

Human carcinogenicity data

The most informative studies for the evaluation of the carcinogenicity of 2,3,7,8-TCDD are four cohort studies of herbicide producers (one each in the United States and the Netherlands, two in Germany), and one cohort of residents in a contaminated area from Seveso, Italy. In addition, the multi-country cohort study from IARC includes three of the four high-exposure cohorts and other industrial cohorts, many of them not reported in separate publications, as well as some professional herbicide applicators.

In most epidemiological studies considered exposure was to mixtures of PCDDs including TCDD, as contaminants of phenoxy herbicides and chlorophenols. The cohorts examined in these epidemiological studies do not allow an evaluation of the risk associated with exposure to higher PCDDs separate from exposure to TCDD. These studies involve subjects with the highest recorded exposures to 2,3,7,8-TCDD. In these cohorts the blood lipid levels of 2,3,7,8-TCDD estimated to the last time of exposure were 2000 ng/kg (mean) (up to 32,000 ng/kg) in the US cohort, 1434 ng/kg geometric mean (range 301 -3683 ng/kg) among workers involved in the clean up of a TCP reactor accident in the Dutch cohort, 1008 ng/kg (geometric mean) in the group of workers with severe chloracne in the accident cohort in Germany, and up to 2252 ng/kg in the Boehringer cohort in Germany. These calculated blood 2,3,7,8-TCDD levels of workers at time of exposure were in the same range as the estimated blood levels in the Kociba two-year rat carcinogenicity study. Exposures in Seveso (median in zone A, 443 ng/kg; median in Zone B, 94 ng/kg) were, on average, lower than those of the industrial cohorts. The upper range of the high-exposed individuals was similar to that of the occupational cohorts (upper 75th percentile in Zone A, about 2000 ng/kg); there were 736 persons in Zone A.

Source & ©: WHO-IPCS  Assessment of the health risk of dioxins:
re-evaluation of the Tolerable Daily Intake (TDI)
page 8

4.1.2 Were cancers observed in highly contaminated populations?

The source document for this Digest states:

Increased risks for all cancers combined were seen in the occupational cohort studies. The magnitude of the increase was generally low; it was higher in sub-cohorts considered to have the heaviest 2,3,7,8-TCDD exposure. Positive dose - response trends for all cancers combined were present in the largest and most heavily exposed German cohort, and in the smaller German cohort where an accident occurred with release of large amounts of 2,3,7,8-TCDD. Increased risks for all cancers combined were also seen in the longer-duration longer-latency sub-cohort of the United States study, and among workers with the heaviest exposure in the Dutch study. These positive trends with increased exposure tend to reinforce the overall positive association between all cancers combined and exposure. The large German cohort evaluated dose-response both for estimated exposure to TCDD and for PCDDs/PCDFs using I-TEQ and identified a positive trend in both analyses.

In Seveso, all-cancer mortality did not differ significantly from that expected, in any of the contaminated zones, although excess risks were seen for specific cancers. Follow-up for the Seveso cohort was shorter than for the occupational cohorts. In most of these studies excess risks were observed for soft tissue sarcoma and also for lung cancer, non-Hodgkin lymphoma and digestive tract cancers. Statistically significant excess risks were observed in individual cohorts for a variety of other cancer sites including multiple myeloma, oral cavity, kidney cancer, leukaemia and breast cancer in women.

A single study in Seveso examined cancer in children 0-19 years of age. Excess risks were observed for ovarian and thyroid cancer and for some neoplasia of the haematopoietic tissue; these results were based on small numbers.

Two studies have evaluated cancer risk among subjects exposed to contaminated rice oil in Japan (Yusho) and Taiwan (Yucheng). The Japanese oil contained in the order of 1000 mg/kg PCBs and 5 mg/kg PCDFs. Estimates of intake are based on a study of 141 cases (Masuda, 1994). These patients consumed about 600 ml of oil over about one month, and ingested about 600 mg of PCBs and 3.5 mg of PCDFs total. Assuming a body weight of 60 kg, the daily dose was thus: 0.33 mg PCBs/kg/day and 0.002 mg PCDFs/kg/day. The Taiwanese oil contained about 100 mg/kg PCBs and 0.4 mg/kg PCDFs. Estimates are based a study of 99 cases. Patients consumed about 1 gram of PCBs and 3.8 mg of PCDFs over a period of about 10 months. Daily doses were approximately 0.06 mg PCBs/kg/day and 0.0002 mg PCDFs/kg/day. The contaminated rice oil contained a complex mixture of chlorinated ring compounds, including dioxin- and non-dioxin-like PCBs, PCQuaterphenyls, PCTerphenyls, as well as the PCDFs. There was an excess liver cancer risk in Japan (OR = 3.1) at 22 years of follow-up, and no excess risk in Taiwan (OR = 0.8) at 12 years.

In summary, the epidemiological evidence from the most highly 2,3,7,8-TCDD- exposed cohorts studied produces the strongest evidence of increased risks for all cancers combined, along with less strong evidence of increased risks for cancers of particular sites. The relative risk for all cancers combined in the most highly exposed and longer-latency sub-cohorts is 1.4. While this relative risk is not likely to be explained by confounding, this possibility cannot be excluded. It should be borne in mind that the general population is exposed to 2-3 orders of magnitude lower levels of TCDD, and 1-2 orders of magnitude lower levels of PCDDs/PCDFs than those experienced, as an equivalent lifetime dose in the industrial populations examined or the population at Seveso.

Source & ©: WHO-IPCS  Assessment of the health risk of dioxins:
re-evaluation of the Tolerable Daily Intake (TDI)
page 8-9

4.2 What non-cancer effects have been observed in children?

The source document for this Digest states:

Non-cancer effects in children

Two US birth cohorts with measured background exposure to PCBs have been followed since 1980, and 2 Dutch birth cohorts with measured background levels of PCBs, PCDDs and PCDFs have been followed since 1990. In Asia, some data are available on Japanese children exposed transplacentally to contaminated rice oil, and detailed follow-up is available on transplacentally exposed children in Taiwan. The estimated upper 10 percentile of total PCBs in breast milk lipid among mothers in the US cohorts was around 1.5 mg/kg; measurement of specific PCBs was limited by analytical methods available at the time of the studies. Mothers in these US cohorts were also exposed to other chlorinated pesticides and heavy metals. In the Dutch cohorts the mean TCDD I-TEQ concentration in human milk was 30.2 pg/g lipid (range 11.1 - 76.4 pg/g) and the estimated PCB concentration was 0.64 mg/g lipid.

Neurodevelopmental delays and neurobehavioral effects including neonatal hypotonia occurred in the three largest cohorts, two in the US and one in The Netherlands, although the age at which the effects occurred and the tests used to detect them were not the same. In the two US cohorts the observed neurobehavioral effects were limited to the infants with the highest decile of transplacental exposure, with some indication of a non-linear effect.

Thyroid hormone levels were evaluated in the two cohorts in The Netherlands with similar exposure to PCDDs/DFs and total PCBs. In utero exposure to total TEQs, as measured in mother's milk, may have influenced thyroid hormone status (TT4, TSH) in infants up to 3 months of age. In Japan and Taiwan effects on children exposed transplacentally to the contaminants in the rice oil included ectodermal defects, global persistent developmental delays, low birth-weight, mild persistent behaviour disorders, decrease in penile length at puberty, reduced height among girls at puberty and hearing loss. It should be noted however, that it is not clear to what extent dioxin-like and/or non-dioxin-like compounds are contributing to these effects when considering the complex mixtures that human individuals are exposed to. In all the studies of infants and children, effects were primarily associated with in utero, rather than lactational exposure. Breast fed infants in the Rotterdam/Groningen cohort were shown to have better neurobehavioural development compared to formula fed infants. Within the group of breast-fed infants, however, those with higher exposure within the cohort to total TEQs (> 50 pg/g milk fat) tended to have poorer neurobehavioral test results (Bayley PDI) compared to those with lower exposure (< 50 pg TEQs/g milk fat).

In children in Seveso who were highly exposed to TCDD, small, transient increases in hepatic enzymes, total lymphocyte counts and subsets, complement activity, and non-permanent chloracne were observed. Also an alteration of the sex ratio (excess female to male) was observed in children born to parents highly exposed to TCDD.

Source & ©: WHO-IPCS  Assessment of the health risk of dioxins:
re-evaluation of the Tolerable Daily Intake (TDI)
page 9-10

4.3 What non-cancer effects have been observed in adults?

The source document for this Digest states:

Non-cancer effects in adults

Several persistent, exposure-related effects occurred in two or more adult populations exposed to PCDDs, PCDFs and PCBs. These populations include the industrial cohorts previously described: US Air Force Ranch Hands (exposed to TCDD during spraying of Agent Orange, median serum TCDD levels back-extrapolated at time of exposure around 50 pg/g lipid), Centers for Disease Control Vietnam Experience Study (exposed to TCDD during a one year tour in Vietnam, mean serum TCDD levels at time of study in 1987, 4 pg/g lipid), and the Seveso, Yusho and Yucheng cohorts.

The effects are elevated GGT in the NIOSH, Ranch Hands and Vietnam Experience cohorts (NIOSH: out-of-range GGT levels, OR=2.27, 95%CI, 1.17-4.39: Vietnam Experience Study: OR 1.3, 95%CI, 1.0-1.8: Ranch Hand mean GGT concentration in highest exposed group 33.3 pg/g lipid TCDD compared to referent group, p <0.001); statistically nonsignificant dose-related increases in Triglyceride levels in the NIOSH cohort (and significant increases in Ranch Hands with serum TCDD concentrations above 15 pg/g lipid; significantly increased mean fasting plasma glucose levels among Ranch Hands with 2,3,7,8-TCDD concentrations > 94 pg/g lipid (OR=1.5, 95%CI 1.2, 2.0), an increased prevalence of diabetes among workers in the NIOSH cohort with serum concentrations above 1500 pg/g lipid and mortality from diabetes among females in all zones of Seveso, particularly in zone B (Zone A, Obs=2, RR=1.8, 95%CI 0.4-7.3; Zone B, Obs=13, RR=1.9, 95%CI 1.1-3.2; Zone R, Obs=74, R.R. 1.2, 95%CI 1.0-1.6). Increased mortality from cardiovascular diseases occurred in multiple industrial cohorts and in males of Zones A and R of the Seveso cohort. Positive dose-response trends were also observed for ischaemic heart disease in the heavily exposed German occupational cohort study, the Dutch occupational cohort and the IARC multicenter study. Among Yusho and Yucheng adults, chronic exposure-related effects included chloracne, conjunctivitis, and sebaceous cysts and inflammation, decreased nerve conduction velocity, fatigue and malaise, hyperpigmentation and hyperkeratosis, and increased mortality from non-malignant liver disease.

In summary, noncancer endpoints were evaluated among groups exposed to dioxins, dioxin-like and non-dioxin-like polychlorinated aromatic compounds in a variety of exposure scenarios, from background to extremely high exposures. Among children exposed in utero to background levels, effects include subtle developmental delays (U.S.and Dutch children) and subtle thyroid hormone alterations (Dutch infants to age 3 month). Multiple, persistent effects occurred among highly exposed children in Yusho and Yucheng who had transplacental exposure. Of the many effects evaluated in exposed adult study populations, many were transient effects disappearing after the end of exposure. A few conditions appear to be in excess among the exposed cohorts when compared to unexposed referent groups including alterations in lipid, fasting plasma glucose and GGT concentrations as well as mortality from cardiovascular disease. Both of the Asian cohorts showed excess death from non-malignant liver disease.

See Table 2: Estimated tissue concentrations in human populations exposed to dioxin and dioxin-like compounds

Source & ©: WHO-IPCS  Assessment of the health risk of dioxins:
re-evaluation of the Tolerable Daily Intake (TDI)
page 10-13

Themes covered
Publications A-Z

Get involved!

This summary is free and ad-free, as is all of our content. You can help us remain free and independant as well as to develop new ways to communicate science by becoming a Patron!