3

3. ESTIMATION OF RADIATION RISKS OF THYROID CANCER IN THE POPULATION
OF THE ORYOL OBLAST

There is evidence that exposure of thyroid to ionizing radiation results in an increase of thyroid cancer incidence. The magnitude of radiation risk, that is the probability of thyroid cancer occurrence, is high and comparable to the risk of leukemia.

The risk of thyroid cancer depends on age at exposure and increases with decreasing age. Given thyroid exposure to incorporated iodine isotopes, the dose is also a function of age at exposure and increases with decreasing age at exposure. The consequences of exposure show themselves as an increase in thyroid cancer incidence rate after a latent period of at least 4 years.

One of possible causes of the increase in thyroid cancer incidence in the Oryol oblast may be associated with thyroid exposure to incorporated iodine isotopes. The relationship of increase in thyroid cancer incidence and exposure to incorporated iodine isotopes after the Chernobyl accident has been demonstrated in various studies [1-6]. In Russia such studies were conducted only in the Bryansk oblast. This section examines a potential linkage between increase in thyroid cancer incidence and radiation dose among residents of the Oryol oblast.

A General description of the population being studied

The analysis of radiation risks and of the dose response relationship of thyroid cancer incidence focuses on the population of the Oryol oblast represented by residents of 3006 settlements. The demographic characteristics of the population are those obtained during the 1989 census and include the size of age groups by rayons and the number of people living in separate settlements. Adequate assessment of thyroid doses from incorporated ¹³¹I for cases and the population in general requires more details at the level of settlement. Since children and adolescents (at time of exposure) form a group who have an increased radiation risk, this group was considered separately. This group consists of members of the population born between 1969 and 1986.

The size of the whole population of the Oryol oblast in 1989 was 874046 persons, of them 207,624 were children and adolescents (0-17 years) and 511,716 were adults (18-60 years). Excluding the city of Oryol, the number of children and adolescents at exposure is 126,058 and 305,485 adults. The age interval up to 60 at exposure for adults was selected taking into account considerations of reliability of the state-level health statistics on incidence.

The follow-up period for the studied population covers 1991 to 2001. The time of the beginning of follow-up 1991 was selected with allowance for minimal latent period of radiogenic cancers of 5 years.

Cases

The available data on thyroid cancer cases in the Oryol oblast are official findings of the Oryol oncological clinic. The diagnosed cases between 1991 and 2001 for those exposed as children and adolescents and adults is shown in Figs. 3.1 and 3.2. Figures 3.3 and 3.4 present the cases for the population excluding the city of Oryol.

As can be seen from Figs. 3.1-3.4, the peak incidence occurred between 1996 and 1998. A total of 78 thyroid cancers in children and adolescents at exposure were diagnosed and 777 cases among adults between 1991 and 2001. With exclusion of Oryol, the number of cases in children and adolescents is 34 and 338 cases of thyroid cancers among adults.

Fig. 3.1. The thyroid cancer cases in those who were children and adolescents at time of exposure.

Fig. 3.2. The thyroid cancer cases in adults (18-60 years of age at exposure).

Fig. 3.3. The thyroid cancer cases in those who were children and adolescents at exposure
(excluding the city of Oryol).

Fig. 3.4. The thyroid cancer cases in those who were adults at exposure
(excluding the city of Oryol).

Table 3.1 contains the size of the populations and the number of cases for the whole population (0-60 year of age at exposure) in major towns (Oryol, Livny and Mtsensk) and rayons of the Oryol oblast. It follows from Table 3.1 that 486 cases, which make more than half of all cases, have been identified in the city of Oryol. Considering that the public health care level is better in Oryol than in other settlements and this may influence the registration level, the analysis of radiation risks was carried out both for the whole population of the oblast (separating out children and adolescents) and with the deduction of the population of Oryol. Moreover, excluding Oryol makes it possible to take into account the difference in the uncertainty associated with thyroid dose determination. The key contributor to dose for thyroid cancer is milk consumption and dose reconstruction for major towns may involve a significant uncertainty due to the fact that milk is supplied to major towns from a range of surrounding rayons which may have different levels of radioactive iodine contamination.

Since age at exposure is a factor of risk, Figs. 3.5-3.8 show distribution of cancer cases as a function of age at exposure.

Table 3.1. Number of cases over the follow-up period 1991-2001 by rayons and major towns.

Rayon	Korsakovsky	Bolkhovsky	Verkhovsky	Glazunovsky	Dmitrovsky
Population	4196	17390	19278	13782	13072
Number of cases	3	42	15	10	6
Incidence rate per 100000 persons	53.8	182.1	60.9	58.8	32.8
Mean dose, mSv	17.0	29.5	12.7	22.3	33.1

Rayon	Dolzhansky	Zalegoshchensky	Znamensky	Kolpnyansky	Krasnoozerensky
Population	11748	14392	4808	17289	6956
Number of cases	5	14	3	10	3
Incidence rate per 100000 persons	32.4	74.4	45.6	44.4	31.7
Mean dose, mSv	7.8	14.0	12.8	11.4	16.3

Rayon	Kromsky	Livensky	Maloarkhangelsky	Mtsensky	Novoderevenkovsky
Population	18888	26861	12236	15889	11652
Number of cases	23	10	6	18	1
Incidence rate per 100000 persons	92.3	29.0	38.2	80.1	6.8
Mean dose, mSv	22.6	11.9	30.0	15.4	13.6

Continuation of Table 3.1.

Rayon	Novosilsky	Oplovsky	Pokrovsky	Sverdlovsky	Soskovsky
Population	8831	50340	15105	15109	6689
Number of cases	2	55	8	16	2
Incidence rate per 100000 persons	11.8	86.6	39.7	82.2	21.7
Mean dose, mSv	15.7	13.8	15.2	21.1	17.5

Rayon	Trosnyansky	Uritsky	Khotynetsky	Shablykinsky	Oryol
Population	10432	14400	9949	8265	287695
Number of cases	10	15	9	5	486
Incidence rate per 100000 persons	70.3	78.1	69.0	46.5	145.9
Mean dose, mSv	21.5	16.6	10.2	15.6	16.1

Rayon	Livny	Mtsensk
Population	45265	42919
Number of cases	43	38
Incidence rate per 100000 persons	83.1	78.0
Mean dose, mSv	7.7	12.6

Fig. 3.5. Distribution of cases in children and adolescents by age at exposure.

Fig. 3.6. Distributions of cases in adults by age at exposure and age distribution
of the population of the Oryol oblast (0-60 years of age).

Fig. 3.7. Distribution of cases in children and adolescents by age at exposure
(excluding the city of Oryol).

Fig. 3.8. Distribution of cases in adults by age at exposure
(excluding the city of Oryol).

For those exposed as children and adolescents the incidence increases with age, which is consistent with the age dependence of spontaneous incidence. However, it is worth pointing to the increase in the number of cases at 1 year of age (the age when the radiation risk is maximal). For adults the peak incidence occurs at the age of 35-37 and 45-50 years. In those more than 50 years of age the number of cases decreases. The gap in the age distribution for persons between the ages of 40-45 years is explained by the demographic gap in the age distribution of the population (the right part of Fig. 3.6).

Radiation doses

As was mentioned in Chapter 1, thyroid doses have been reconstructed using the official guidelines adopted by the Russian Scientific Committee on Radiation Protection. The mean doses by rayons of the Oryol oblast for the studied group (the ages of 0-60 years at exposure) are included in Table 3.1. As can be seen from the table, the largest dose values of about 30 mSv occurred in the Bolkhovsky, Dmitrovsky and Maloarkhangelsky rayons. Table 3.2 includes the mean thyroid dose values (mSv) for different age groups.

Table 3.2. Mean thyroid doses (mSv) for different age groups in the Oryol oblast (persons exposed at different age).

Age group	0-17	18-60
Oblast as a whole	36.2	7.6
Excluding the city of Oryol	31.4	7.8

It follows from Table 3.2 that the mean dose for those exposed as children and adolescents is about 5 times higher that in adults.

Technique of analysis of the dose response relationship
of thyroid cancer incidence

The analysis was performed using modern approaches. For descriptive analysis the standardized incidence ratio (SIR) is used equal to the ratio of the observed cases to the number of cases expected with no exposure. SIR is estimated using the indirect standardization method usually applied for rare diseases such as leukemias and thyroid cancers. Standardization of an indicator makes it possible to take into account differences in the age structure of compared groups. For rare diseases nationwide age-specific indicators are normally used as a control. In this study the controls were age-specific thyroid cancer incidence rates for Russia during the time period from 1991 to 2002.

For calculation of SIR and 95% confidence intervals the statistical package EPICURE [7] developed specifically for study of health effects of exposure in the cohort of atomic bomb survivors in Japan. The SIR was calculated using the formula:

where summation was made by time (index j) and age groups (index i); cases_i,j is the number of cases in the age group (i) at the time moment (j); PY_i,j is the number of person-years of follow-up in the age group (i) at the time moment (j); are age-specific thyroid cancer incidence rates in Russia in the age group (i) at the time moment (j).

The radiation risks and dose response parameters were calculated using the EPICURE package (the module AMFIT estimating regression coefficients by the maximum likelihood technique for grouped and stratified data).

For consideration of age and time differences in the compared groups the data were stratified by attained age and calendar time and divided into 4 dose groups: 0, 0.012, 0.02, 0.05, >0.05 for children and adolescents and 0, 0.006, 0.007, 0.0010, >0.01 Sv for adults.

The calculations were made for models with internal and external controls. The risk model with an internal control takes the form:

where l_i,j is the thyroid cancer incidence rate in the stratum (i,j); is the spontaneous incidence rate in the Oryol oblast in the stratum (i,j); ERR_1Sv is the excess relative risk at the dose of 1 Sv; d_i,j is the thyroid dose in the stratum (i,j).

The model with an external control is described by the equation:

where SIR^un is the standardized incidence ratio for unexposed population.

The confidence (95%) intervals were estimated by the likelihood function profile. In addition, for analysis of the dose response relationship, the distribution of standardized incidence ratio (SIR) was derived. This can be considered as an approximate distribution of relative risk by radiation doses.

Since the method of thyroid dose reconstruction and the dose values themselves are still debatable, the effect of radiation exposure on thyroid cancer incidence was studied using the same technique as in the analysis of the Bryansk oblast data [8]. This approach allows qualitative assessment of the role of the radiation exposure in thyroid cancer incidence without using dose values. The gist of the method is as follows. The radiation risk of thyroid cancer per unit dose and thyroid dose tends to increase with decreasing age at exposure. This means that exposure effects should be seen among children of smaller age. For ascertaining it is sufficient to draw a distribution of the population and cases by age at exposure as a factor of risk. The frequency of spontaneous cases, given no exposure, should be proportional to the size of a particular age group and the distributions should coincide. If the radiation response makes itself evident in some age groups, then the distribution of cases and that of the population in general will be different. The workability of the method was demonstrated in the study of thyroid cancer incidence in the Bryansk oblast [8] in which the distributions were different for children and adolescents at exposure and were identical for adults and these results were in agreement with estimates of radiation risk.

Main results

Children and adolescents at exposure (0-17 years old)

The dynamics of SIR for children and adolescents at exposure is shown in Figs. 3.9, 3.10. As can be seen from the figures, the SIR reaches its maximum in 1994-1996 exceeding the nationwide incidence level by a factor of 6-7 and then decreases to the control level (values 1-2). Such a time dependence can be attributed either to screening effect (detection of earlier diseases) or early manifestation of exposure effects in this category of the population.

Fig. 3.9. The standardized incidence ratio (SIR) for children and adolescents at exposure.

Fig. 3.10. The standardized incidence ratio (SIR) for children
and adolescents at exposure (excluding the city of Oryol).

Let us now consider the distribution of cases and the population as a whole by age at exposure (Figs. 3.11 and 3.12). As can be seen from Fig. 3.11 for the entire population, the distribution patterns are almost identical, which in a way suggests an absence of an exposure effect. Excluding Oryol from the study, the distribution of cases is shifting towards younger ages, which may be an indication of the exposure factor, however, to a lesser extent than in those exposed as children and adolescents living in the Bryansk oblast (Fig. 3.13).

The SIR dose response relationship is presented in Figs. 3.14 and 3.15. It can be seen that the SIR dependence has a slight positive trend for thyroid dose, both for the population as a whole and with exclusion of Oryol.

Fig. 3.11. Distribution of the population and cases by age at exposure.

Fig. 3.12. Distribution function of the studied population and cases
(excluding Oryol) by age at exposure.

Fig. 3.13. Distribution of the population and the population as a whole
by age at exposure (girls of the Bryansk oblast).

Fig. 3.14. Standardized incidence ratio (SIR) as a function of dose.

Fig. 3.15. Standardized incidence ratio (SIR) as a function of dose
(excluding Oryol).

Adults of 18-60 years of age at exposure

Figs. 3.16 and 3.17 present the SIR for adults. As can be seen from the figures, SIR has similar features for adults to those in children and adolescents: an increase in the initial follow-up period, the peak in 1996-1998 and a decrease towards the control level at the end of the follow-up period.

The dependence of SIR on dose for adults is shown in Figs. 3.18 and 3.19. As follows from the figures, the dependence has a slight positive trend, i.e. the incidence increases with increasing radiation dose.

Results of calculating the standardized incidence ratio with 95% confidence intervals over the whole follow-up period from 1991 to 2001 for different age groups are shown in Table 3.3.

Fig. 3.16. The standardized incidence ratio (SIR) for adults at exposure.

Fig. 3.17. The standardized incidence ratio (SIR) for adults at exposure
(excluding Oryol).

Fig. 3.18. Standardized incidence ratio (SIR) as a function of dose.

Fig. 3.19. Standardized incidence ratio (SIR) as a function of dose
(excluding Oryol).

Table 3.3. Results of calculating the standardized incidence ratio over the follow-up period from 1991 to 2001 for different age groups.

Population group	Children and adolescents (0-17)	Adults (18-60)
Whole population	4.04 (3.20, 5.03 95% CI)	2.42 (2.25, 2.60 95% CI)
Excluding Oryol	2.88 (1.98, 4.04 95% CI)	1.72 (1.55, 1.92 95% CI)

It follows from the Table 3.3 that incidence in children and adolescents at exposure living in the Oryol oblast in the follow-up period is 3-4 times higher than the nationwide indicators. For adults this excess is a factor of 1.7-2.5.

A question about the influence of radiation exposure on incidence can be better answered based on the regression analysis of the dose relationship of incidence. For this purpose, as was mentioned above, the statistical package EPICURE was used.

In calculations with EPICURE the data were grouped into 4 dose groups. The key parameters of the dose groups for the whole population are illustrated in Table 3.4.

Estimates of radiation risk coefficients (excess relative risk per unit dose 1 Sv (ERR_1Sv)) calculated with EPICURE are presented in Table 3.5.

Table 3.4. Key parameters of dose groups for the whole population.

Dose groups	Mean dose (mSv)	PY (person years)	Cases	Spontaneous cases*	Mean dose (mSv)	PY (person years)	Cases	Spontaneous cases
Dose groups	Children and adolescents (0-17)				Adults (18-60)
1	0.9	487100	27	25.6	0.4	1370510	141	183
2	15.5	514512	25	20.0	6.5	580565	55	77
3	31.8	757000	16	15.6	7.4	2851080	488	373
4	88.1	524147	10	5.5	14.4	826425	93	108
Total	36.2	2282759	78	66.7	7.6	5628080	777	741

* estimate of spontaneous cases was derived from calculations with EPICURE with an internal control.

Table 3.5. Estimates excess relative risk per unit dose 1 Sv (ERR_1Sv).

Population category	Children and adolescents (0-17)	Adults (18-60)
Internal control
Whole population	7.8 (n.d., 124.7 95% CI)*	6.3 (-16.7, 43.7 95% CI)
Excluding Oryol	13.5 (n.d., 296.3 95% CI)	12.9 (-12.8, 55.2 95% CI)
External control
Whole population	0.8 (-6.4, 16.9 95% CI)	1.1 (-19.9, 34.5 95% CI)
Excluding Oryol	9.9 (-5.1, 65.0 95% CI)	9.4 (-14.8, 49.0 95% CI)

*n.d. - the lower confidence limit has not been defined.

It can be seen from Table 3.5, that the values of the radiation risk coefficients are positive both for children and adolescents and adults, but are not statistically significant (the lower confidence limit is negative). In two calculations, the confidence limit was not determined for children and adolescents because of the shapes of the likelihood function (no convergence in solution of the system of likelihood equations). It was an unexpected result that the trends for adults were positive. In the studies for the Bryansk oblast the trend for adults was negative.

Descriptive statistical analysis of the relationship of the thyroid cancer incidence
in the Oryol oblast and the radioactive ¹³¹I contamination of soil and
the stable iodine content in soil

This section provides results of a descriptive analysis of the dependence of thyroid cancer incidence in the Oryol oblast on two environmental factors: the radioactive ¹³¹I contamination of soil and the stable iodine content in soil. Each factor was represented by 6 levels of concern, as can be seen from Table 3.6.

Two follow-up periods were studied: from 1982 to 1991 assuming that no radiation-induced cancers occurred during this period (the latent period was from 1986 to 1991) and the time between 1992 and 2001.

Table 3.6. Levels of concern of environmental factors.

Levels of factors	Level of iodine in soil	Radioactive ¹³¹I in soil, kBq/m²
1	Extremely low	10-20
2	Very low	20-40
3	Low	40-100
4	Moderately low	100-300
5	Lower limit of norm (forested area)	300-600
6	Lower limit of norm (forest-steppe)	600-1200

The descriptive analysis is represented by the values of thyroid cancer incidence rate (95% confidence intervals) as a function of the mentioned environmental factor levels. The source data were the rayon incidence rates and each rayon was related to levels of stable and radioactive iodine in soil in accordance with Table 3.6. Results of calculation of the incidence rate for both genders in different follow-up periods are shown in Table 3.7. As can be seen from Table 3.7, the population of the Oryol oblast is mainly concentrated in the territories with levels 4-6 with respect to stable iodine and levels 2-3 with respect to contamination with radioactive iodine.

The influence of iodine levels on incidence was studied for persons exposed between the ages of 0-60 years. Children and adolescents were not considered separately, as the number of cases for them is limited.

The data of Tables 3.7 and 3.8 are presented in graphs of Figs. 3.20 and 3.21. The indicators for the follow-up period between 1992 and 2001 are noticeably (3-4 times) above those for the period from 1982 to 1991.

Table 3.7. Calculated incidence rates for both genders (follow-up period 1982-1991).

Level	Number of cases	Follow-up person years	Rate per 100000 persons
Stable iodine in soil
1	-	-	-
2	-	-	-
3	-	-	-
4	9	297126	3.0
5	51	1039504	4.9
6	176	4968586	3.5
Radioactive ¹³¹I
1	4	117486	3.4
2	32	884882	3.6
3	195	5128947	3.8
4	5	173901	2.9
5	-	-	-
6	-	-	-

Table 3.8. Calculated incidence rates for both genders (follow-up period 1992-2001).

Level	Number of cases	Follow-up person years	Rate per 100000 persons
Stable iodine in soil
1	-	-	-
2	-	-	-
3	-	-	-
4	19	297126	6.4
5	160	1039504	15.4
6	648	4968586	13.0
Radioactive ¹³¹I
1	5	117486	4.2
2	81	884802	9.2
3	701	5128947	13.7
4	5	173901	23.0
5	-	-	-
6	-	-	-

Fig. 3.20. Dependence of the thyroid cancer incidence rate on the level
of stable iodine in soil in different follow-up periods.

Fig. 3.21. Dependence of the thyroid cancer incidence rate on the level
of radioactive ¹³¹I iodine in soil in different follow-up periods.

The incidence rate in the period 1982 to 1991 remained almost unchanged with variations in the levels of radioactive iodine and stable iodine (Figs. 3.20 and 3.21).

In the period from 1992 to 1999 an increase in the levels of stable and radioactive iodine results in increased incidence rates (Figs. 3.20, 3.21). The positive correlation of thyroid cancer incidence and radioactive iodine level may suggest the induction of radiogenic cancers. On the other hand, the increase in incidence rate with increasing level of stable iodine in soil does not fit the hypothesis that the endemia of stable iodine during fallout of radioactive iodine may be contributing to increase in radiogenic cancers. The result we have arrived at may be explained by insufficient accuracy of stable iodine data and by the fact that the iodine ratio in soil of different territories and in people living there may be different.

For better visualization of analysis results the incidence rates (per 100,000 persons) were presented as a matrix (I´J), where I (rows) is the number of levels of stable iodine and J (columns) is the number of levels of radioactive ¹³¹I. Results of analysis have been differentiated by follow-up periods, oblasts and sex.

An example of a matrix of thyroid cancer incidence rate (both genders, 0-60 years of age at exposure and follow-up period from 1992 to 1999) is provided by Table 3.9.

Table 3.9. Example of a matrix of thyroid cancer incidence rates.

	0	1	2	3	4	5
0	0	0	0	0	0	0
1	0	0	0	0	0	0
2	0	0	0	0	0	0
3	0	7.7	4.3	0	0	0
4	0	0	13.9	23.0	0	0
5	4.3	9.5	13.9	0	0	0

The cell with maximum incidence rate is shaded.

The matrix was presented as a map with outlines (the upper left-hand corner of the matrix is in the left-hand corner of the picture). The red color in the map corresponds to the local maxima of incidence rates. The abscissa axis is stable iodine level and the axis of ordinates is radioactive iodine level. Higher number of level means higher iodine content.

Maps with outlines are shown in Fig. 3.22. As can be seen from the figure, these maps reflect the dependence of incidence rate on level of stable and radioactive iodine. They also show the values of incidence rates per 100,000 persons as a function of radioactive and stable iodine level. If stable and radioactive iodine levels do influence thyroid cancer incidence, the peak incidence should shift towards the upper left-hand corner, the region of decreased content of stable iodine and increased content of radioactive iodine. Figure 3.22 shows a shift in the peak towards higher content of radioactive iodine.

For quantifying the influence of the considered factors on thyroid cancer incidence we used the chi-square statistic based on analysis of differences between expected and observed number of cases [10]. Six levels of each factor were convoluted into 2 and the data took the form of table 2´2. The convolution of levels was done by the scheme shown in Table 3.10. Table 3.10 is a matrix of levels of stable (rows) and radioactive iodine. For example, cell (5, 4) represents the 5^th level of stable iodine and 4^th level of radioactive iodine. Table 2´2 and levels included in the cells of table 2´2 are bold-faced and delineated. Other cells (not delineated) were not considered, because such levels of stable and radioactive iodine do not occurein the Oryol oblast.

Whole population (0-60 years) 1982-1991

Whole population (0-60 years) 1992-1999

Fig. 3.22. Maps with outlines of thyroid cancer incidence rate as a function of level
and a combination of factors.

Table 3.10. Matrix of the levels of stable (rows) and radioactive iodine.

1.1	1.2	1.3	1.4	1.5	1.6
2.1	2.2	2.3	2.4	2.5	2.6
3.1	3.2	3.3	3.4	3.5	3.6
4.1	4.2	4.3	4.4	4.5	4.6
5.1	5.2	5.4	5.4	5.5	5.6
6.1	6.2	6.4	6.4	6.5	6.6

Comments to the delineated cells of table 2´2 differentiated by the level of influence of factors of stable and radioactive iodine are illustrated by Table 3.11. The cell “control” means that the stable iodine content is close to norm and the radioactive iodine contamination is minimum.

Table 3.11.

Level	0	1
0	Stable iodine deficiency	Radiation exposure + iodine deficiency
1	Control	Radiation exposure

Results of testing of the null hypothesis (the relative risk is 1) are shown in Table 3.12. The expected number of cases for the criterion hi-square is calculated by the incidence rate in the cell - “control’ of table 2´2 and the number of person years in other cells of table 2´2.

Table 3.12.

Period	Both genders, age 0-60 at exposure
1982-1991	0.98 (a)	0.33 (c)
1982-1991	0.90 (b)	0.93 (d)
1992-1999	0.75 (e)	<0.001 (g)
1992-1999	0.93 (f)	<0.001 (h)

The cells in which the zero hypothesis was rejected are shaded. As can be seen from Table 3.12, the incidence in the period 1982-1991 in the territories with stable iodine (cell d) level close to norm, is not significantly different from the control (cell b). There is no statistically significant difference for incidence in the territories with increased density of radioactive iodine contamination either (cell c). However, in the period from 1992 to 2001 in the territories in which the influence of stable and radioactive iodine should be pronounced most of all, the incidence rate (cell g) is statistically significantly different from the control. A statistically significant difference from control was also observed for territories with the content of stable iodine close to the norm and with increased content of radioactive iodine (cell h) The verification of the hypothesis about the difference in incidence in cells (h) and (g) has revealed no statistical significance (p=0.8), i.e. no endemia effect on thyroid cancer incidence. The fact that no effect of iodine deficiency was observed is probably explained by a relatively small difference in the levels of stable iodine in control groups and compared territories (only three categories for stable iodine 4-6) and in addition, as was mentioned above, the iodine content in soil may differ from that in human body. Therefore the presented results should be considered as preliminary.

Discussion of the results

The analysis of thyroid cancer incidence showed a noticeable increase in the incidence in the Oryol oblast as compared with that for Russia in general. On average, the thyroid cancer incidence in the population of the Oryol oblast in the time period 1991 to 2001 was 3-4 times higher than the nationwide rates, while for adults this excess was a factor of 1.7-2.5. Possible reasons for this increase can be both regional differences in incidence rates and manifestation of exposure effects. The SIR since 1991 increases to the maximum values in 1996-1998, the values being 6-7 for children and adults at exposure and 3-3.5 for adults and then decreases to the levels close to nationwide rates. Such dynamics of the SIR can be explained by the screening effect during the studied period (diagnosis of latent diseases as a result of better registration), manifestation of the radiation response factor in this time period or a combination of both.

Estimates of the radiation risk reveal a positive trend of incidence rate as a function of thyroid dose, which, however, is not statistically significant. A positive trend has been observed for adults as well. An indirect confirmation to this trend is provided by the distribution of population and cases by age at exposure (Fig. 3.23) suggesting that the distributions differ and age of cases shifts to the region of smaller values. As was said in the section “Methods of analysis”, this may be a qualitative indication of the influence of radiation factor. For the adults of the Bryansk oblast these distributions were identical and the incidence trend with allowance for dose was even negative among adults.

Fig. 3.23. Distribution of adults by age at exposure (excluding Oryol).

In general, the radiation risk factors in the Oryol oblast are lower than those in the Bryansk oblast. For example, in the Bryansk oblast the excess relative risk for children and adolescents at exposure with 1 Sv dose was 40 and statistically significant. A similar value was derived for children and adolescents of Belarus and the Ukraine. This values has been taken by the UNSCEAR as a preliminary estimate of radiation risk of thyroid cancer after the Chernobyl accident. Naturally, the question arises: what is the proportion of radiogenic cancers among all thyroid cancer cases in the Oryol oblast?

To answer this question let us estimate the attributive risk which by definition is a proportion of radiogenic risks among all cases. For this purpose we use data of Table 3.4 showing spontaneous and observed cases. For children and adolescents the attributive risk is (78-66.7)´100/78=14.5% and for adults (777-741)´100/777=4.6%. This means that if estimates of radiation risk derived for the residents of the Oryol oblast are objective, 15% of thyroid cancers in children and adolescents at exposure are radiation-induced and 5% for adults.

Assuming that the estimates adopted by UNSCEAR are objective, the attributive risk for children and adolescents will be equal to (40´0.036)´100/1+40´0.036=59%.

We shall regard this estimate as conservative, with a margin of safety.

The performed analysis of the influence of the levels of stable iodine in soil and radioactive iodine on the incidence shows that there is a correlation between the incidence and the radioactive iodine level and no correlation with the stable iodine level. Generally, this relationship is in agreement with results of radiation risk analysis (positive trend with dose).

The absence of an effect of iodine deficiency is probably explained by a relatively small difference in the levels of stable iodine in control and compared territories (only three categories for stable iodine 4-6) and in addition, iodine levels in soil can differ from those in human body. Therefore, the presented results regarding the influence of endemia of stable iodine should be treated as preliminary.

Conclusions

1. The analysis has revealed a marked increase in the thyroid cancer incidence in the Oryol oblast as compared to that for Russia as a whole. On the average, the thyroid cancer incidence in the Oryol oblast in the period 1991-2001 was 3-4 times higher than the nationwide rate, for adults the excess is a factor of 1.7-2.5. This increase may be explained by regional differences in incidence and a manifestation of the exposure effect. The values of the standardized incidence ratio SIR was increasing from 1991 and reached its maximum in 1996-1998: the value was 6-7 for children and adolescents at exposure and 3-3.5 for adults. Then it was decreasing to the levels close to nationwide indictors. Such dynamics of the SIR can probably be attributed to the screening effect in the considered period (diagnosis of latent diseases due to better registration), the effect of radiation exposure or a combination of these factors.

2. Estimates of radiation risk show a positive trend of incidence as a function of thyroid dose, which, however, is not statistically significant. A positive trend was also inferred for the adult population.

3. The attributive risk (a fraction of radiogenic cancers among all cancer cases) derived from direct estimates of radiation risk for residents of the Oryol oblast, is equal to 15% for children and adolescents at exposure and 4% for adults. A conservative estimate of attributive risk of 60% has been adopted by UNSCEAR and is used as a tentative estimate for children and adolescents of the Oryol oblast, given the risk coefficient of 40 for the dose 1 Sv is used.

4. The performed analysis provides sufficient evidence to assume that the radiation exposure factor did influence the thyroid cancer incidence in the Oryol oblast, although to a lesser extent than in the Bryansk oblast.

References

1. Ivanov V.K., Gorski A.I., Tsyb A.F. et al. Dynamic of thyroid cancer incidence in Russia following the Chernobyl accident//J. Radiol. Prot. - 1999. - V. 19. - P. 305-318.

2. Ivanov V.K., Gorski A.I., Pitkevitch V.A. et al. Risk of radiogenic thyroid cancer in Russia following the Chernobyl accident. In: Radiation and Thyroid Cancer. Proceeding of an International Seminar held in St. John’s College, Cambridge, UK, 20-23 July 1998. - Singapure: World Scientific, 1999. - P. 89-96.

3. Jacob J., Goulko G., Heidenreich W.F. et al. Thyroid cancer risk to children calculated//Nature. - 1998. - V. 392. - P. 31-31.

4. Jacob J., Kenigsberg, Zvonova I. et al. Childhood exposure due to the Chernobyl accident and thyroid cancer risk in contaminated areas of Belarus and Russia//Br. J. Cancer. - 1999. - V. 80. - P. 1461-1469.

5. Likhtarev I.A, Sobolev B.G., Kairo I.A. et al. Thyroid cancer in the Ukraine//Nature. - 1995. - V. 365. - P. 365.

6. Likhtarev I.A, Kairo I.A., Shpak V.M. et al. Radiation-induced and background thyroid cancer of Ukranian children (dosimetric approach)//Int. J. Radiat. Med. - 1999. - V. 3-4. - P. 51-66.

7. Preston D.L., Lubin J.H., Pierce DA., McConney M.E. EPICURE. - Sietle, USA: Hirosoft International Corporation, 1993.

8. Ivanov V.K., 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.

9. Ivanov V.K., Gorski A.I., Maksioutov M.A. et al. Thyroid cancer incidence among adolescents and adults in the Bryansk region of Russia following the Chernobyl accident (preliminary analysis)//Health Physics. - 2002. - V. 83(6).

10. Breslow N.E., Day N.E. Statistical methods in cancer research, 1987.