Chapter 3

 

THYROID CANCER INCIDENCE AMONG ADOLESCENTS AND ADULTS
IN THE BRYANSK REGION OF RUSSIA FOLLOWING THE CHERNOBYL ACCIDENT
(preliminary analysis of follow-up to 1998)

 

 

Introduction

 

As a result of the Chernobyl accident extensive areas of Russia (more than 60 thousand km2) were affected by radioactive contamination. The worst contaminated were the western areas of the Bryansk region where the maximum density of 137Cs contamination in the soil was as high as 4 MBq/m2. Somewhat lower contamination was reported in the Kaluga, Tula and Orel regions. The maximum level of 137Cs contamination in the soil in these areas was 0.6 MBq/m2, and the maximum contamination of 131I in the soil in the Bryansk region was 15 MBq/m2 on 10 May 1986.

After the Chernobyl accident a dramatic growth of thyroid cancer incidence in the contaminated areas of Russia, Ukraine and Belarus occurred. One of the possible causes of this increase was exposure of the thyroid to incorporated iodine-131 (131I). However, it cannot be ruled out that registration of these diseases was improved after the accident due to increased attention of health care authorities to the problem that led to wide coverage by specialized medical examination, higher attendance of patients and higher quality of check-ups. This is called the screening effect. Thyroid cancer is quite a rare disease with a low mortality rate and the screening effect can become an important factor leading to an increase of incidence rates. This is the reason why the issue is given so much attention in the paper.

Unlike the extensively studied carcinogenic effects of childhood exposure to radiation, there is little evidence of increase in thyroid cancer incidence for the age older than 20 years (Ron et al., 1995; Shore, 1992; Thompson et al., 1994). It should also be mentioned that the available studies mostly relate to external radiation sources. In this respect, the atomic bombing survivors in Japan show a negative radiation risk at the age of more than 40 years.

There are a large number of publications on radiation induced thyroid cancers after the Chernobyl accident. Most of the results of these studies have been described in the seminar proceedings Radiation and Thyroid Cancer (1999).  In one of the papers in that seminar the cancers are discussed only on a descriptive approach and consideration is limited to a simple analysis of the incidence rate and the standardized incidence ratio (SIR).   Other studies related to the analysis of thyroid cancer incidence after the Chernobyl accident that are worth mentioning are those of  (Ivanov et al., 1999a; Heidenreich et al., 1999; Jacob et al., 1999).

The above studies were concerned with the analysis of thyroid cancer incidence in children and adolescents at exposure for which the risk of induction of thyroid cancer is the highest. It is worth noting that studies of thyroid cancer incidence in the adult population exposed to radiation after the Chernobyl accident are few and, as a rule, are limited to analysis of incidence rates.

In the paper by Ivanov et al., (1999b) thyroid cancer incidence was studied in the population of the Bryansk region after the Chernobyl accident based on descriptive analysis without using radiation dose estimates. The analysis is indicative of the absence of exposure effects among the adult population.

The present work aims at analysis of the thyroid cancer incidence since 1986 and a study of the risk of radiogenic thyroid cancer among adolescents and adults (age at exposure 15-69) in the worst contaminated areas of Russia, namely the Bryansk region. The size of the population under consideration is sufficiently large to allow a statistical analysis.

 

Materials and methods

 

General description of medical and demographic data

 

Study area and population

 

The primary source of demographic information in the study were data of the federal statistics organization  and regional statistics committees. The Bryansk region is located in the south-west of Russia and its total population is 1470129 (according to the 1989 census). The region consists of 28 rayons (rayon is a territorial subdivision smaller than a region) and 3085 territorial units. The ratio of the urban and rural population is about 2. The largest cities are Bryansk (the population is 448026), Dyatkovo (34423), Novozybkov (44697) and Klintsy (70908 persons).

As the precise distribution of the age at exposure, broken down by rayons, in the period under study is unavailable, the data of the 1989 census are used. These are data about age structure of rayons and population size for practically each administrative unit of the Bryansk region, which makes possible more precise estimation of collective dose (for the population groups having no thyroid cancer) and individual doses (for cases). Migration in the Bryansk region in the studied period was not significant. For example, in 1991 the migration increment factor (the difference between those arriving and leaving the region was 10 thousand people) according to the state statistics was 25 for the urban population and -41 for the rural population. Major changes in the population structure were, most probably, due to relocation of people from heavily contaminated areas to the areas with lower contamination. The large-scale relocation of people from the contaminated areas was started in 1989 (three years after the accident) and by now more than 52 thousand people have been resettled, which is about 0.35% of the whole population of the Bryansk region.

The comparison of the age structure in the major cities and most contaminated areas from 1991 to 1998 census shows that the age structure remained practically unchanged since the time of the exposure to radiation.

We therefore assume that the age structure in the territorial units and rayons did not change significantly either. For this reason, the gender and age distribution in specific territorial units were calculated under the assumption that such distributions are identical to those in the rayon in which the territorial unit is located.

The size of the population under study was 1,019,047 people according to the census of 1989. The analysis of the dose relationship of incidence rates was based on using the cases diagnosed in the post latent period from 1991 to 1998.

Table 1 shows demographic characteristics of the population in the Russian region under consideration according to the census of 1989.

 

Table 1

Key demographic characteristics of the study population

 

Age at exposure

15-29

30-44

45+

15-69

Females

Cases of thyroid cancer

153

215

287

655

Size of population

153529

139915

260147

553592

Males

Cases of thyroid cancer

20

39

55

114

Size of population

154019

146142

185449

485610

Both sexes

Cases of thyroid cancer

173

254

342

769

Size of population

304811

283701

437163

1025685

 

Registration of thyroid cancer cases

 

In the Russian Federation (as in the former USSR) in accordance with the regulations of the Ministry of Health of Russia, cancer care relies on two principal functional units: the oncological dispensary at the regional level and the oncological consulting room (oncological cabinet) at the rayon level (Winkelmann et al., 1998).

At the rayon level, residents attending as outpatients can be given oncological consultation in a rayon hospital. Rayon oncologists are to provide a clinical diagnosis of malignant and non-malignant neoplasms and refer patients to the regional oncological dispensary for more specific diagnosis and treatment.

The regional oncological dispensary is responsible for cancer diagnosis and treatment in the territory of the region. Specialized departments of dispensary include radiology, chemotherapy, surgery, X-ray, and others. For diagnosis and treatment of difficult tumours, rare tumours, such as thyroid cancer, or tumours of uncertain origin, patients are further referred to more specialized institutions, such as state hospitals, clinics associated with research institutes.

Cancer patients diagnosed and treated in a medical establishment are reported to the regional cancer oncological dispensary by means of the extract from the medical card. On discharge of the patient, the extract is mailed to the regional oncological dispensary of the patient's place of residence.

1051 cases of thyroid cancer were detected among the residents of the Bryansk region who were aged 15-69 at the time of accident between 1986 to 1998. These are considered in the analysis of the Standard Incidence Ratio (SIR). Of them, 769 cases detected from 1991 to 1998 were used for radiation risk analysis (Heidenreich et al., 1999; ICRP Report 60, 1990; BEIR V, 1990). These exceeded the minimal latent period for radiation induced thyroid cancer of five years.  655 cases were females and 114 were males. Information on thyroid cancer cases used in this analysis is stored in the cancer registry functioning in the Bryansk oncological dispensary. The persons with thyroid cancer were operated on in hospitals of Bryansk (the majority), or in the state research centers of Moscow and Obninsk.

After the Chernobyl accident thyroid cancer incidence in residents of the contaminated areas has received special attention of the health care authorities of Russia. Therefore, we assume that the incidence data are fairly complete and diagnoses are reliable.

Table 2 shows the distribution (%) of thyroid cancer cases by the method of confirmation. The majority (92% cases from 1986 to 1998) of thyroid cancer cases were confirmed histologically. The portion of diagnoses verified by examination of histological sample has grown from 86% in 1986 up to 99% in 1998.

 

Table 2

Distribution (%) of thyroid cancer cases by method of confirmation

 

Year

Histology

Only clinical

Other

1986

86

14

0

1987

86

14

0

1988

88

12

0

1989

90

8

2

1990

85

15

0

1991

84

9

6

1992

92

2

6

1993

95

3

2

1994

93

1

6

1995

96

4

0

1996

97

3

0

1997

96

1

3

1998

99

0

1

Average 1986-1998

92

5

3

Table 3 shows the distribution (%) of thyroid cancer cases by histological types. The dynamics of distribution of thyroid cancer cases by histological types is shown. As Table 3 indicates, the most common types of thyroid cancer among the population of the Bryansk region since 1986 were follicular and papillary cancers (29% and 51%, respectively).

Table 3

Distribution (%) of thyroid cancer cases by histological types

 

Year

Follicular carcinoma

Papillary carcinoma

Medullary carcinoma

Other

Non value*

1986

24

47

0

15

14

1987

25

51

1

10

13

1988

30

51

1

7

11

1989

33

49

1

7

10

1990

33

52

0

0

15

1991

21

32

1

32

14

1992

24

33

3

32

8

1993

30

52

2

12

4

1994

39

46

0

9

6

1995

48

39

1

8

4

1996

33

52

2

10

3

1997

20

62

2

14

2

1998

14

77

0

8

1

Average 1986-1998

29

51

1

12

7
 
* Histological material is not sufficient for definitive diagnosis.

 

 

Information used in the analysis of the dose-response relationship includes the date of birth, gender, address of residence at exposure, the date of diagnosis or surgery and the estimated thyroid dose from incorporated iodine radioisotopes.

 

Thyroid doses in the Bryansk region

 

The personal mean thyroid doses for the population of the Bryansk region were calculated based on the “Methodology for reconstruction of thyroid doses from iodine radioisotopes in residents of the Russian Federation exposed to radioactive contamination as a result of the Chernobyl accident in 1986” (2000) the latest revision of which was issued on 31.05.2000.

In our analysis “individual dose” is understood as a dose applied to an individual person. For cancer cases a dose value was assigned based on address of residence at time of exposure, gender and age at exposure. The collective dose was calculated for each populated point with allowance for age and sex structure.

The densities of deposisted 137Cs in the Bryansk region that are required for reconstruction of the population doses are taken from the cartographic database (Atlas of radioactive contamination of the European part of Russia, Belarus and Ukraine, 1998).

The methodological base of the study are the models accounting for thyroid doses in residents of the Russian Federation exposed to radiation as a result of the Chernobyl accident. Parameters of the dosimetric models for people of different age are derived using field radiation measurements and data about the grazing regime for milk cattle and intake of local milk in the contaminated regions of Russia at different time after the accident.

For determination of the internal radiation doses the most crucial data were more than 45 thousand measurements of 131I thyroid levels in the residents of the four worst contaminated regions of Russia (the Bryansk, Tula, Oryol and Kaluga regions) and over 5 thousand measurements of local milk samples collected in the same regions in May-early June 1986.

The methodology presented in the paper “Methodology for reconstruction of thyroid doses” (2000), was developed based on the following principles. Radiological data of different types and reliability are available for reconstruction of thyroid dose. The order in which these data are used depends on type of radiological data. The measured levels of 131I in the thyroid, which are more closely related to the internal dose, were the first to be used for the dose calculation. The measured 131I concentrations in milk consumed by local population ranked second in importance for calculation. If measurements of 131I in humans and/or milk samples were not available, the thyroid dose was estimated using statistical models relating the dose with the 137Cs contamination level and based on results of extensive radiation measurements. For reconstruction of doses in the areas for which the number of measurements are not sufficient, the ratios of 131I and 137Cs activities in depositions are used, as well as data about evolution of the accidental situation and protection measures taken.

In calculation of thyroid doses two key pathways of radioiodine intake were considered: ingestion and inhalation. Since the main source data for the calculation of thyroid dose (Methodology for reconstruction of thyroid doses, 2000) are measurements of thyroid 131I activity in people, both pathways are taken into account. According to the radiation monitoring data after the Chernobyl accident the radioiodine intake by residents of the contaminated areas was primarily due to consumption of milk and other foods (green vegetables etc.) exposed to surface contamination.

The thyroid dose from all iodine radioisotopes released into the environment from the Chernobyl accident, including daughter products resulting from decay of tellurium radioisotopes, was determined according to the report “Methodology for reconstruction of thyroid doses”, (2000).

The basic territorial subdivision used in thyroid dose reconstruction is a territorial unit with surrounding areas. Depending on where milk and diary products were supplied from in May 1986, the territorial units are categorized as:

·       settlements and villages;

·       cities or towns.

In the period that is being considered the rural population was mainly consuming milk from private farms, while cities and towns received milk and diary products from both collective and private farms of a given territorial subdivision.

The thyroid dose was estimated for the period from the beginning of the Chernobyl fall-out to 1 July 1986, which covers the time period when radiologically significant iodine radioisotopes occurred in the environment. In the methodology used the fall-out is assumed to occur at one time.

The source information for estimating the thyroid dose for residents of all territorial units of the Bryansk region was the following:

·       the date when the fall-out began and its duration in the region, area or territorial unit;

·       the mean soil 137Cs contamination density in a territorial unit (this value was used for estimating doses from all sources of 131I in rural territorial units and doses due to inhalation and consumption of green vegetables for cities and towns, as well as estimating doses due to inhalation for the city of Bryansk);

·       the mean 137Cs soil contamination density in the rayon (this value was used for estimating doses due to inhalation and consumption of green vegetables in cities and towns, and estimating doses due to consumption of green vegetables for the city of Bryansk);

·       the mean 137Cs soil contamination density in the areas surrounding the Bryansk areas (was used for estimating doses due to milk consumption in the city of Bryansk).

It was assumed that the rural population consumes milk from private cows only and residents of towns and cities uses milk from collective farms.

Using these source data, thyroid doses were estimated for each territorial unit of the Bryansk region.

The method for dose estimation is detailed in the Annex II.

The thyroid dose among persons who have developed cancer was deter­mined using the dependence of dose on age at exposure (for adolescents only) and residence address at exposure time (the name of the territorial unit). For the population without diagnosed cancers we calculated the distribution of the col­lective dose by age for each territorial unit. This estimation was made using the size of population in a par­ticular territorial unit (according to the census), sex and age distribution of the population in the rayon to which the territorial unit belongs and the age dependence of the thyroid internal dose (for adolescents only). The basic dosimetric data for the studied cohort is shown in Table 4. The excess relative risk (ERR) is defined as the increase in relative risk at the dose under consideration compared with zero dose, as described in detail below.

 

Table 4

The results of risk estimation for the population under consideration (age at exposure 15-69).

The number in parentheses in the table are the upper and lower 95th percentiles  of the uncertainty distribution.

 

Females

Mean dose (Gy)

0.026

0.021

0.025

0.022

External control

Excess relative risk

-1.02 (-3.91, 1.92)

-0.62 (-3.48, 2.21)

-1.90 (-4.07, 0.23)

-1.30 (-2.80, 0.15)

SIR (for unexposed)

3.08 (2.59, 3.57)

2.12 (1.84, 2.41)

1.59 (1.40, 1.77)

1.97 (1.82, 2.12)

Internal control

Excess relative risk

5.81 (0.63, 12.57)

3.06 (-0.70, 7.89)

-2.42 (-4.15, -0.23)

-0.92 (-2.35, 0.76)

SIR

3.07 (2.61, 3.60)

2.16 (1.88, 2.47)

1.56 (1.38, 1.75)

1.96 (1.81, 2.12)

Males

Mean dose (Gy)

0.023

0.023

0.020

0.022

External control

Excess relative risk

6.01 (-7.2, 19.4)

2.38 (-7.10, 9.57)

-1.96 (-4.97, 1.11)

-0.37 (-3.46, 2.74)

SIR (for unexposed)

2.21 (1.24, 3.18)

1.86 (1.27, 2.44)

1.14 (0.84, 1.44)

1.46 (1.19, 1.73)

Internal control

Excess relative risk

15.91 (-0.26, 45.4)

2.92 (-2.75, 12.79)

-1.96 (-3.89, 2.32)

0.74 (-2.32, 5.20)

SIR

2.52 (1.54, 3.89)

2.00 (1.42, 2.74)

1.08 (0.81, 1.40)

1.45 (1.20, 1.74)

Both sexes

Mean dose (Gy)

0.024

0.022

0.023

0.023

External control

Excess relative risk

0.74 (-2.69, 4.21)

0.03 (-2.68, 2.68)

-1.43 (-3.64, 0.51)

-0.63 (-2.06, 0.78)

SIR (for unexposed)

2.94 (2.51, 3.38)

2.00 (1.76, 2.25)

1.39 (1.25, 1.54)

1.78 (1.65, 1.91)

Internal control

ERR

8.65 (0.81, 11.47)

3.01 (-0.31, 7.20)

-1.63 (-3.24, 0.47)

0.01 (-1.44, 1.69)

SIR

3.03 (2.62, 3.55)

2.06 (1.82, 2.33)

1.37 (1.23, 1.53)

1.80 (1.67, 1.93)

 

 

The geographical pattern of the average thyroid dose (by rayons) for adolescents and adults at exposure is presented in Figure 1.  Figure 2 shows the cumulative normalized distribution of cases and population as a function of the thyroid dose. As is seen from Figure 2, the mean dose in those who developed cancer in the Bryansk region was nearly the same as that in the whole population. The above properties of this dependence may already be indicative of the absence of influence of the radiation factor.

The particularly large step in the cumulative distribution function in the dose range around 0.003 Gy is due to the contribution of the cases detected in the city of Bryansk (the population is some 450 thousand people). Doses in the regional centers are difficult to estimate because estimation is based on radioecological information from the areas surrounding a city and specific contribution of separate areas to a general regional dose is difficult to distinguish. Therefore, it was assumed in calculating the dose for a regional center that different areas gave equal contribution to the dose due to milk consumption.

 

Figure 1. Geographical pattern of the average thyroid dose (by rayons) for adolescents and adults
aged 15-69 years at exposure.

 

Figure 2. Distribution of cases and healthy persons as a function of the absorbed dose.

 

 

 

Risk assessment

 

Background incidence rate of thyroid cancer

 

In the calculations, the SIR and the spontaneously induced thyroid cancer rate in Russia from 1989 to 1998 was used. The confidence intervals for SIR were estimated in accordance with the description in the paper by Breslow and Day (1987). According to the federal state statistics, the spontaneous thyroid cancer incidence rate has changed over the period in question.

In the period that was considered, 1991-1998, the incidence rate in Russia as a whole for persons between the ages of 15 to 69 years varies by more than a factor of 1.7 (from 3´10-5 to 5´10-5). Most likely, the increase in the spontaneous rate is attributable to the fact that more attention was paid to diagnoses of thyroid cancer after the Chernobyl accident.

As the background rate varies with time, the non-stationary Poisson process of events was used in the risk analysis of the incidence. The risk estimates were made using both the external control group – the spontaneous thyroid cancer incidence in Russia as a whole, and the internal control group.

Excess relative risk per 1 Gy (b) was determined assuming a linear dependence of the thyroid cancer incidence rate with dose.

In calculations using the external control group the risk model takes the form:

,

where  is the spontaneous incidence rate of the thyroid cancer in Russia for attained age (e+t) at the time t, for person i;  is the factor taking into account the difference between the incidence rates in the considered region and Russia as a whole, f , depends upon sex,  is the spontaneous incidence rate for the i-th person under study.

This difference can be attributed to both the differences in screening effect levels for the population in general and to the difference in actual incidence levels in the study area. It is assumed that the shape of the incidence age distribution in Russia and in the region under study is identical.

It was shown in (Ivanov et al., 1999b) that the relative age distribution is a conservative quantity that doesn’t vary strongly in different countries. In this model f has the meaning of the SIR for spontaneous incidence.

di is the absorbed dose in the thyroid gland for the i-th person; b is the excess relative risk per unit dose. This value is a function of age at exposure.

When risk coefficients were estimated using the internal control, data were stratified by attained age and calendar time (index k), and the spontaneous incidence was determined from the balance of the observed and expected number of cases in a given stratum.

The second risk model is written as:

.

The 95% likelihood intervals were determined from the likelihood function profile.

The values of risk model parameters (b, f) from model 1 were used for prediction of the thyroid cancer incidence among the population under study. The number of anticipated spontaneous cases was calculated with allowance for the factor f, the dynamics of follow-up person-years and the trend of incidence rates in Russia in general. The number of radiogenic cancers was estimated as a product of the spontaneous incidence rate and the risk factor (b).

The method for risk estimation and prediction of thyroid cancer incidence is detailed in the Annex I.

 

Results

 

The dynamics of the thyroid cancer incidence in the study regions in general is presented in Figures 3-5.

Figure 3 shows the dynamics of the thyroid cancer incidence rate in the Bryansk region residents as a function of age at diagnosis for three age intervals 15-29, 30-44 and >45.

In Figure 4 the dynamics of the thyroid cancer incidence rate among the residents of the  Bryansk region is considered as a function of age at exposure in the same age intervals.

As follows from Figures 3 and 4, an increase in thyroid cancer incidence rate was observed in the post-accident period in each of the studied age groups. This increase is most probably due to a better detection level resulting from increased attention of the health care authorities to this problem.

Figure 5 presents a standardized incidence ratio with 95% confidence intervals (SIR = observed number of cases/expected number of cases) for males and females. The expected number of cases is calculated using the age-specific thyroid cancer incidence rates in the considered period in Russia in general.

 

Figure 3. Dynamics of thyroid cancer incidence rate in the Bryansk region residents
of the considered age groups as a function of age at diagnosis.

 

Figure 4. Dynamics of thyroid cancer incidence rate in the Bryansk region residents
of the considered age groups as a function of age at exposure.

 

Figure 5. Dynamics of the standardized thyroid cancer incidence ratio in the Bryansk region.

(SIR = observed number of cases/expected number of cases)

 

 

The mean value of SIR for the follow-up period considered and in the age interval at exposure between 15 and 69 years is 2.0 (1.8, 2.1 95% CI) for females and 1.5 (1.2, 1.7 95% CI) for males. The value of SIR does not vary significantly with time.

Table 4 illustrates the values of the risk coefficients (ERR1Gy) and SIRs for exposed and unexposed population (f coefficients taking into account the difference between the incidence in the considered region and Russia as a whole) with 95% confidence intervals. Estimates of radiation risk are based on using external and internal control groups. As can be seen from Table 4, the radiation risk of thyroid cancer for adolescents and adults in the Bryansk region in the given time period is not confirmed.

The fact that the values of SIR for unexposed members of the population (derived with model 1) and the exposed population suggests that the observed difference from unity is most probably due to the regional differences in the incidence rates and a possible effect of better registration of diseases as a consequence of paying more attention to the problem of cancer incidence in the contaminated territories after the Chernobyl accident.

The presented estimates of radiation risk given in Table 4 are confirmed by results of the linear regression of the SIR dependence on mean dose for separate rayons (Figure 6).

Figure 6. SIR for selected rayons of the Bryansk region as a function of the mean rayon dose.

(SIR = observed number of cases/expected number of cases)

 

 

Table 5 presents risk estimates for follicular and papillary forms of thyroid cancer. As incidence data for these cancer forms are not available from Russian statistics, the risk was estimated using internal control. As follows from Table 5, the radiation risk for these cancer forms is not confirmed. A statistically significant risk was obtained only for the follicular cancer form in males, but this estimate can be biased since the number of cases is rather limited (35 cases).

Table 5

The results of risk estimation for the population under study (age at exposure 15-69)
for different forms of thyroid cancer

 

Cancer form

Follicular

Papillary

Sex

males

females

both

males

females

both

Number of cases

35

197

232

48

334

382

ERR Gy-1

7.75

(0.52, 18.8)

-1.11

(-3.52, 2.11)

1.31

(-1.31, 4.65)

-1.31

(-5.43, 5.98)

-0.07

(-2.20, 2.49)

0.44

(-1.70, 3.00)

 

 

The results of thyroid cancer prediction are presented in Figure 7. It can be seen that the result of prediction using the parameters of model is in good agreement with the observed incidence.

According to the projection, by 2005 2000 thyroid cancer cases are expected to occur in the residents of the Bryansk region aged 15-69 in 1986.

Figure 7. Prediction of thyroid cancer cases among adolescents and adults at the

Chernobyl accident in the Bryansk region compared to the observed number.

 

 

Discussion

 

The presented work is the first analysis of dose-response of the thyroid cancer incidence in adolescents and adults in Russia exposed to incorporated 131I isotopes after the Chernobyl accident. The analysis was carried out for the Bryansk region, which is the worst contaminated territory in Russia.

The volume of the data about the disease cases used in the analysis (769 thyroid cancer cases in 1,019,047 people) is comparable to the body of information reported in well known studies such as Ron et al. (1995) the cohort of 120 thousand people, 700 cases, and Thompson et al. (1994) 80 thousand people and 225 cases. Results of the performed analysis are essentially in agreement with the conclusions made in both studies regarding the absence of noticeable radiation risks of thyroid cancer for the considered category of the population. However, it should be stressed that both studies were dealing with radiation risks arising from thyroid exposure to external source of ionizing radiation as distinct from the internal sources in this study.

The results presented in the dose response analysis should be considered as preliminary due to serious constraints of the study. First of all, the analysis uses personified, rather than individual dosimetry. A dose for a specific individual (for a case) and a group of individuals without thyroid cancer was determined based on place of residence at the time of exposure and age at exposure. Such an approach naturally involves a lot of assumptions. However, the authors used as much available information as possible. All available measurements of individual doses were used for dose reconstruction. In fact, these are data of the Russian Scientific Commission on Radiation Protection. Reconstruction of individual doses requires a survey on an individual basis which is difficult to realize in the present study because of the size of the population and the number of cases.

In our study the follow-up period is rather short (8 years), but the number of cases is quite significant (769) and, as was mentioned above, is comparable to that used in similar studies (Ron et al., 1995; Thompson et al., 1994).

According to (Ron et al., 1995; Thompson et al., 1994; Shore, 1992), the peak in occurrence of radiogenic cancers is to be expected 10-15 years after exposure. It may be noted, however, that these results were obtained for an external radiation source, and influence of incorporated iodine isotopes on development of thyroid cancer is still not well understood. Therefore, it remains an open question when the peak in induction of radiogenic cancers should be expected.

Since individual follow-up of such a large group is impossible, the analysis for persons who were not diagnosed with the disease was based on general demographic data of 1989, the year when the census was carried out. In other words, the analysis does not allow for migration and mortality of the population, while mortality may become a significant factor for older age groups.

There were two reference groups in the above study (external and internal control). As can be seen from the presented results, calculations using internal control, by and large, give higher estimates of excess relative risk. Yet, it is difficult to give preference to any of the approaches. On the one hand, internal control, given a large number of cases, takes better account of the study population. But in case of a rare disease such as thyroid cancer the use of an internal control can lead to a bias due to the small number of cases in strata and the decrease in accuracy of estimation of the spontaneous component of risk. The other approach uses an external control, the thyroid cancer incidence rate in Russia in general. But even when using external control data reliability remains an issue.

There is no operating national cancer registry in Russia and data on thyroid cancer incidence are obtained from official bodies of medical statistics. It is not unlikely that these data have a bias, but the risk analysis is based on the relative distribution of incidence rate by age, which is shown in Ivanov et al. (1999b), to be practically identical even for different countries and is weakly dependent on the registration level. The difference in the registration levels in the studied regions and in Russia in general will influence the SIR value only.

Moreover, for the population under study the percentage of cases of thyroid cancer that are histologically confirmed is rather high (95% from 1991 to 1998). Moreover, the quality of the diagnoses is currently improving due to increased interest to the problem from health care establishments. Nevertheless, the risk model (with external control) used in this study is not much dependent on this factor because the number of cases is quite large and the dose was unknown at diagnosis, so the distribution of wrong diagnoses is most likely to be proportional to the studied population size. As a result, the dose response curve is expected to shift, having the same slope, and the radiation risk will not change.

It is still disputable whether other iodine isotopes contribute to the internal irradiation. The methods for reconstructing doses from such sources of radiation are just being developed. Account should also be taken of external exposure of thyroid to long-lived isotopes.

There remains an open question how thyroid cancer is influenced by a combined effect of exposure to iodine isotopes and an external source of radiation due to long-lived isotopes.

The above restrictions of the analysis make us treat the presented results as preliminary.

 

Conclusion

 

The suggestion often made that there is a radiation risk of thyroid cancer for the adolescents and adults (age at exposure 15-69) of the Bryansk region in the observation period 1991-1998 has not been confirmed. The excess relative risk ERR1Gy per unit dose 1 Gy among adolescents and adults at the time of the Chernobyl accident (age 15-69 years) in the Bryansk region (with using external control) was found to be -0.4 with 95% CI (-3.5, 2.7) for males, -1.3 with 95% CI (-2.8, 0.1) for females and -0.6 with 95% CI (-2.1, 0.8) for males and females together. With using internal control the excess relative risk ERR1Gy per unit dose 1 Gy was found to be 0.7 with 95% CI (-2.3, 5.2) for males, -0.9 with 95% CI (-2.4, 0.8) for females and 0.0 with 95% CI (-1.4, 1.7) for males and females together.

The spontaneous incidence rate in the region under consideration among adolescents and adults at exposure is about twice that of Russia as a whole. This excess is attributed to the differences in registration of diseases and regional differences in the spontaneous level of incidence.

The presented estimates of radiation risk should be treated as tentative because of the many assumptions and restrictions used in the analysis.

 

Acknowledgement

 

The presented study was carried out in the framework of the French-German Chernobyl Initiative (Project No 3 - Health Effects of the Chernobyl Accident, Specific Agreement No 3.1.3S - Thyroid Cancer in Adolescents and Adults in the Most Affected Territories of Russia after the Chernobyl Accident).

 

References

 

Atlas of radioactive contamination of the European part of Russia, Belarus and Ukraine. Developed in the Institute of Global Climate and Ecology of Roshydromet and Russian Academy of Sciences under direction of academician Izrael Y.A. - Moscow: Federal Service of Geodesy and Cartography of Russia, 1998 (in Russian).

Breslow N.E., Day N.E. Statistical methods in cancer research. Vol. II. The design and analysis of cohort studies. - Lyon: IARC, IARC Scientific Publication 82, 1987.

Heidenreich W.F., Kenigsberg Y., Jacob P., Buglova E., Gulko G., Paretzke H.G., Demidchik E.P., Golovneva A. Time trends of thyroid cancer incidence in Belarus after Chernobyl accident//Radiat. Res. - 1999. - V. 151. - P. 617-625.

International Commission on Radiological Protection. Report 60. Recommendations of the International Commission on Radiological Protection. - Oxford: Pergamon Press, 1990.

Ivanov V.K.(a), Gorski A.I., Pitkevitch V.A., Tsyb A.F. Risk of radiogenic thyroid cancer in Russia following the Chernobyl accident. In: Thomas G., Karaoglou A., Willliams E.D. eds. Radiation and thyroid cancer. Proceeding of an International Seminar on Radiation and Thyroid Cancer. - Brussels-Luxembourg: World Scientific Publishing, 1999: 89-96.

Ivanov V.K.(b), Gorski A.I., Tsyb A.F., Maksioutov M.A., Rastopchin E.M. Dynamics of thyroid cancer incidence in Russia following the Chernobyl accident//J. Radiol. Prot. - 1999. - V. 19, N 4. - P. 305-318.

Jacob P., Kenigsberg Y., Zvonova I., Gulko G., Buglova E., Heidenreich W.F., Golovneva A., Bratilova A.A., Drozdovitch V., Kruk J., Pochtennaja G.T., Balonov M., Demidchik E.P., Paretzke H.G. Childhood exposure due to the Chernobyl accident and thyroid cancer risk in contaminated areas of Belarus and Russia//British J. of Cancer. - 1999. - V. 80, N 9. - P. 1461-1469.

Methodology for reconstruction of thyroid doses from iodine radioisotopes in residents of the Russian Federation exposed to radioactive contamination as a result of the Chernobyl accident in 1986. Guidelines MU-2.6.1-00b, 2000 (in Russian).

Radiation and Thyroid Cancer. Proceeding of an International Seminar on Radiation and Thyroid Cancer. Thomas G., Karaoglou A., Willliams E.D., eds. - Brussels-Luxembourg: World Scientific Publishing, 1999.

Ron E., Lubin J.Y., Shore R.E., Mabuchi K., Modan B., Pottern L.M., Shneider A., Tucker M., Boice J.D. Thyroid cancer after exposure to external radiation: a pooled analysis of seven studies//Radiation Res. - 1995. - V. 141. - P. 259-277.

Shore R.E. Issues and epidemiological evidence regarding radiation-induced thyroid cancer//Radiation Res. - 1992. - V. 131. - P. 98-117..

Thompson D.E., Mabuchi K., Ron E., Soda M., Tokunaga M., Oshikubo S., Sugomoto S., Ikeda T., Terasaki M., Izumi S., Preston D.L. Cancer incidence in atomic bomb survivors. Part II: Solid tumors, 1958-1987//Radiation Res. - 1994. - V. 137. - P. S17-S67.

National Academy of Sciences Committee on the Biological Effects of Ionizing Radiation. Health effects on population of exposure to low levels of ionizing radiation. BEIR V Report. - Washington DC: US National Academy of Sciences, 1990.

Winkelmann R.A., Okeanov A., Gulak L., Remennik L., Rahu M., Storm H.H. Cancer registration techniques in the New Independent States of the former Soviet Union. - Lyon: IARC, IARC Technical Report No. 35, 1998. - P. 22-43.

Annex I

 

Risk assessment

 

The likelihood function for model under consideration is:

,

n is the number of cases; N is the number of healthy persons; parameter li for a person i is a function of age at exposure (åi), time since exposure (ti) and absorbed dose (di); ti is the time interval from the accident time to detection of the case and for healthy persons this is time interval from the accident time to the end of 1998.

As mentioned above, two linear models, one with external control (model 1) and the other with internal control (model 2), were used for assessing risk coefficients.

Within the first model:

.

Let us divide the ti interval in to mi intervals of the year length, then:

.

Let us determine the mean incidence rate on the k-th time interval:

.

Assuming that the rate changes linearly within the interval, then:

.

The logarithm of the likelihood function is:

where  is the spontaneous incidence rate at attained age (e+k) for the i-th person at the k-th time interval.

As the personal data for healthy persons are not available and we have only the demographic data for M, the territorial units can be presented as a logarithm of the likelihood function using the formula:

,

where emin and emax is the minimal and maximal age at exposure, respectively; nj,l is the number of persons at the j-th age at exposure in the l-th territorial unit.

It may be assumed that the size of age groups is a constant in time since exposure.

When risk coefficients were estimated using the internal control (model 2), data were stratified by attained age and calendar time, and the spontaneous incidence was determined from the balance of the observed and expected number of cases in a given stratum.

The spontaneous mortality in the stratum by attained age j, at time moment k was taken to be as follows:

.

For prognosis of thyroid cancer incidence the dynamics of expected number of cases C(k) in the time interval k is accounted for by the formula:

,

where S(k) and R(k) are dynamics of spontaneous and radiogenic cancers, respectively.

:

.

The function F(k) accounts for the difference in the regional incidence rate and the incidence rate in Russia in general. The function F(k) provides for the 5 year latent period in induction of radiogenic cancers.

Function F(k)=1, if 1986+k<1991 (since no estimate of F(k) was made for this period) and F(k)=f at the same time; F(k)=0 if k<5 (latent period) and F(k)=1 if k³5; Dj,l is the collective dose in the territorial unit l in persons with age at exposure j.

It was assumed in the prediction that the size of the studied population remains unaltered due to spontaneous mortality.

Annex II

 

Description of the methodology for thyroid dose reconstruction

 

a) Basic formulae for estimation of mean thyroid dose (Zvonova et al., 2000).

The mean thyroid dose Dthj of residents of age u is calculated with the formula:

, mGy,                                                                                        (1)

where Dthst is the standard thyroid dose in 3-year old children in rural and urban population points (calculated by the regression equation (2)), mGy; p(u) is the mean relation of dose in 3-year old children to dose in persons of age u, relative units (Table 2); D1(u) is the dose of person of age u (calculated by relations (3-10) for intake i0=1 kBq/day and the actual times of the beginning of grazing period, consumption of milk and green vegetables at ratio (10), mGy; D1st(u) is the same as in D1(u) but for the times of beginning of grazing period, consumption of milk and green vegetables from the beginning of radioactive fallout and in the absence of countermeasures (t1=t2=t0; fms=0; f3=f4=1; t3=¥, f5=1) at the ratio (10), mGy.

The standard thyroid dose in rural and urban residents of the Bryansk region is related to the mean 137Cs soil contamination density in a population point and in its vicinity S137 in 1986 through the linear regression:

, mGy.                                                                                                (2)

The values of the regression equation parameters s and w are given in Table 1.

 

Table 1

Parameters of regression equation (2) for the population points of the Bryansk region

 

137Cs fallout density

s, mGy

w, mGy m2/kBq

Villages

Cities and towns

Villages

Cities and towns

  >500 kBq/m2

220±120

79±12

0.75±0.08

0.71±0.08

  (37-500) kBq/m2

105±14

79±12

0.95±0.10

0.71±0.08

  >37 kBq/m2

0

0

0.38±0.04

0.28±0.03

 

 

Table 2

Mean ratio of p(u) of dose of 3-year old children and dose of people of age u, relative unitsa

 

Age u,
full years

Cities and towns

Villages

Age u,
full years

Cities and towns

Villages

0

0.50±0.09

0.61±0.09

10

3.6±1.0

2.4±0.5

1

0.63±0.1

0.7±0.1

11

4.1±1.1

2.6±0.5

2

0.80±0.2

0.9±0.2

12

4.5±1.2

2.8±0.5

3

1.0±0.2

1.0±0.2

13

5.0±1.3

3.0±0.6

4

1.2±0.3

1.2±0.2

14

5.4±1.3

3.1±0.6

5

1.5±0.4

1.3±0.3

15

5.7±1.3

3.3±0.6

6

1.9±0.5

1.5±0.3

16

6.0±1.3

3.4±0.6

7

2.3±0.7

1.7±0.4

17

6.3±1.3

3.5±0.6

8

2.7±0.8

1.9±0.4

>17

7.3±0.9

4.1±0.4

9

3.1±0.9

2.2±0.5

 

 

 

 

a Mean arithmetic values with mean error are given.

 

 

 

 

,                                                                           (3)

where D(u) is thyroid dose for person of age u, mGy; Ih, Ig is the total intake of 131I with inhaled air (index “h”) and food (index “g”) respectively, kBq; dh(u), dg(u) are dose coefficients for 131I intake for persons of age u by the inhalation and ingestion pathways, mGy/kBq according to data of (ICRP Publication 67, 1993; ICRP Publication 71, 1995).

b) Basic formulae for estimating dynamics and integral intake of 131I for rural residents.

The total intake of 131I by the inhalation and ingestion pathways is determined as integral over time of a corresponding function of intake i(t), kBq/day:

, kBq,                                                                                                (4)

, kBq,                                                                                                (5)

where t0 is the time of beginning of radioactive fallout in a given region of Russia. The zero time is the moment of the accident, 26 April 1986, 01 a.m.

Considering shortage of source data on environmental (meteorological), economic (agricultural technology) and social (diet and behavioral) data, the function of 131I intake for residents of age u with inhaled air ih and food ig is used in the methodology in a simplified form as follows:

ih(u, S137) = i0 · φh (u, S137),                                                                                                         (6)

ig(t, u, S137) = i0 · φg (t, u),                                                                                                           (7)

where i0, kBq/day, is a constant value taken to be equal to daily intake of 131I with milk for a person of age u approximated to the time moment of beginning of radioactive fallout in a given region t0 and functions jh(u,S137) and jg(t,u), relative units, are determined using formula (5) and (6).

The dynamics of inhalation intake of 131I from a passing cloud are modeled as a homogeneous process during one day starting from the beginning of fallout t0:

jh(u, S137) = f1(u) × f2(S137),            if  t0<t<(t0+1);

jh(u, S137)T 0                                         at other t,                                                                    (8)

where f1(u), relative units, is the coefficient accounting for the relation between the 131I intake by the inhalation and ingestion pathways for children and adolescents of different age groups as compared with this parameter in adult rural residents (ICRP Publication 71, 1995) - (Table 2) due to differences in diet, the values f1(u) are given for rural and urban residents separately.

Table 3

Values of coefficient f1(u), relative units, in formula (6) for rural and urban residents
of different age groups

 

Age, years

< 1

1 - 2

3 - 7

8 - 12

13 - 17

> 17

f1(u), relative units,
(village)

0.1

0.2

0.4

0.6

0.9

1.0

f1(u), relative units,
(city and town)

0.1

0.2

0.4

0.8

1.5

1.8

 

 

f2(S137), relative units, is the coefficient accounting for the relation of inhalation and ingestion pathways of 131I transfer to human body for rural residents as a function of 137Cs soil contamination density:

f2(S137) = 0.15                    at  S137£100 kBq/m2,                                                                      (9)

f2(S137) = 2.0 × S137–0.56       at  S137>100 kBq/m2.

For a mixture of equal air-borne concentrations of 131I in the form of elemental iodine, methyl iodide and aerosol fraction with mathematical mean equal to 1 mm and fast absorption in the respiratory tract (ICRP Publication 71, 1995). Individual variations in dose coefficients dh(u) and dg(u) for persons of the same age group should be described by standard geometric deviation equal to 1.6.

Dynamics of 131I intake with local foods

Dynamics of 131I intake with local foods is modeled by a multi-component function accounting for milk contamination during in-house and grazing periods and contamination of green vegetables:

                    (10)

where t, days, is time since the Chernobyl accident on 26 April 1986, 01 a.m.;

t0, days, is time of the start of radioactive fallout in a given area;

t1, days, is time of the beginning of consumption of green vegetables (Table 4);

t2, days, is time of the start of grazing of milk cattle (Table 4);

t3, days, is time when contaminated milk and other local products in a given populated point or area were stopped to be consumed (Table 4);

fms, relative units, is the ratio of 131I intake with milk during indoors period and that during grazing period. According to the monitoring data in the first days after radioactive fallout the parameter fms is taken to be 0.1, relative units. After the start of the grazing period fms is taken to be zero;

fv, relative units, is the coefficient accounting for 131I intake with green vegetables. With allowance for age differences in diet fv is taken to be 0.05 for adults, adolescents and children older than 7 years, 0.03 for children of 3-7 years and 0 for children younger than 3 years;

f5, relative units, is the coefficient accounting for reduction in 131I transfer to human body due to stoppage of consumption of milk and other local food in May-June 1986. For persons who stopped consuming milk and local foods the value f5 is taken to be 0.1 starting from a specific date and for the rest - 1.0;

Ts, days, is the period of reduction of 131I concentration in the milk from cows kept indoors equal to 6.0±1.5 days;

Tec, days, is the period of reduction of 131I concentration in milk from cows grazing contaminated areas, taken to be 4.2 days based on monitoring data after the Chernobyl accident;

Tm, days, is the period of reduction of 131I concentration in milk of cows after a single intake to cow body taken to be 1.5 days (Korneev and Sirotkin, 1987).

 

 

 

 

Table 4

Times of beginning of grazing milk cattle and consumption of green vegetables by residents
of the Bryansk region in spring of 1986, days after the accident

 

Collective milk cattle

(0-4) ±5

Private milk cattle

(3-8) ±5

Annual green vegetables of rural residents

(3-8) ±5

Perennial green vegetables of urban residents

(24-28) ±7

 

 

The function of 131I intake for residents of cities and towns is similar to functions (8) and (10) for the rural population of the region where a given city or town is located.

 

Table 5

Mean time when milk was stopped to be consumed by residents of western areas
of the Bryansk region (days after the accident)

 

District

Time when milk was stopped to be consumed, days

Gordeevsky

13-20

Zlynkovsky

11-20

Klimovsky

15-18

Klintsovsky

14-18

Krasnogorsky

10-20

Novozybkovsky

10-18

 

 

 

The formula  for calculating collective doses

Given the assumptions made, it can be easily shown that the contribution of PDk of the j-th settlement to the collective dose of the thyroid dose is:

, person×mGy,                                    (11)

where k=1,…,K; K - number of settlements in the Bryansk region;  - population in the j-th settlement, persons; vc - settlement type: village, town, city;  - mean thyroid dose form 131I in rural and urban persons of the Bryansk region;  - age structure of the population in the k-th settlement.

 

References

 

International Commission on Radiological Protection. Age-dependent doses to members of the public from intake of radionuclides. Part 2. Ingestion dose coefficients. - Oxford: Pergamon Press; ICRP Publication 67, Part 2; Ann ICRP 23(3/4); 1993.

International Commission on Radiological Protection. Age-dependent doses to members of the public from intake of radionuclides. Part 4. Inhalation dose coefficients. - Oxford: Pergamon Press; ICRP Publication 71, Part 4; Ann ICRP 25(3/4); 1995.

Korneev N.A., Sirotkin A.N. Radiology of agricultural animals. - Moscow: Energoatomizdat, 1987. - 208 p. (in Russian).

Zvonova I., Balonov M., Bratilova A., Vlasov O., Shishkanov N. Update on thyroid dose reconstruction of population of Russia in 14 years after the Chernobyl accident: Method and dose estimation. - Hiroshima: IRPA-10, 2000. - P. 11-265.