0031-8655190$03.00+0.00 Copyright @ 1990 Pergamon Press plc
Photochemistry and Photobiology Vol. 51, No. 5 , pp. 579-582, 1990 Printed in Great Britain. All rights reserved
ULTRAVIOLET-RADIATION AND SKIN CANCER. EFFECT OF AN OZONE LAYER DEPLETION THORMOD HENRIKSEN*, ARNEDAHLBACK, SBRENH.H. LARSEN and JOHAN MOAN Institute of Physics, University of Oslo. Blindern, 0316 Oslo 3, Norway. (Received 29 July 1989; accepted 28 November 1989) Abstract-The effect of changes in the ozone layer on the incidence of skin cancer was explored using data for Norway. Attempts were made to arrive at a relationship between the “environmental effective UV-dose” and the skin cancer incidence. Norway is well suited for this purpose because of the large variation in the annual UV-dose from north to south. Furthermore we have a well developed cancer registry and a homogeneous population with regard to skin type. Four different regions of the country, each with a broadness of 1” in latitude (approximately 111 km), were selected (located around 69.5, 63.5, 60 and 58.5” N). The annual effective UV-doses for these regions were calculated, assuming normal ozone conditions throughout the year and the action spectrum proposed by CIE, which extends u p to 400 nm. The incidence rate (in the period 1970-1980) of malignant melanoma and non-melanoma skin cancer (mainly basal cell carcinoma) increased with the annual environmental UV-doses. For both these types of cancer a quadratic dose-effect relationship seems to be valid to a first approximation. The oresent data indicate that the incidence of skin cancer would increase by approximately 2% for each percent ozone reduction.
information is used in an attempt to arrive at the relationship between the environmental effective UV-dose and the different types of human skin cancer. Norwegian data are particularly well suited for this purpose for three reasons; (1) Norway is quite long in the north-south direction (a latitude difference of 13”) which implies a variation in the annual UV-dose of about 50%. ( 2 ) The population is more or less of the same skin type (Caucasian race). (3) Norway has a good National Cancer Registry.
INTRODUCTION
A depletion of the atmospheric ozone layer would undoubtedly result in larger UV-dose rates as well as in larger annual “effective UV-doses”. This may in turn result in an increase in detrimental effects such as the incidence of skin cancer. In order to evaluate the extent of this increase, knowledge about the ozone layer thickness and the environmental “carcinogenic sunlight” (the part of the solar spectrum which overlaps with the action spectrum for skin cancer} is needed. Furthermore, information about the relationship between the environmental UV-dose and skin cancer is essential. In previous papers (Dahlback et al., 1989; Moan et al., 1988, 1989) we have calculated effective UVdoses both for a normal as well as a depleted ozone layer, including ozone holes such as that observed in Antarctica. It was found that a small depletion of the ozone layer would increase the effective environmental UV-dose with a “radiation amplification factor” close to 1.0 (this factor depends on the latitude as well as on the action spectrum used). This implies that 1% depletion of the ozone layer throughout the year would increase the annual dose of carcinogenic sunlight by 1% at our latitude (60” N). For normal ozone conditions the annual effective UV-dose was found to vary by approximately 4 5 % per degree of latitude in the region 6C-70” N. At mid-latitudes this factor increases to about 7 % per degree. Consequently, a depletion of the ozone layer of 4 5 % throughout the year in Scandinavia would yield a UV-radiation milieu corresponding to that found for normal ozone conditions at approximately 1” of latitude (111 km) to the south. This
MATERIALS AND METHODS
UV-doses. The effective annual UV-dose of carcinogenic sunlight can be defined as the integrated product of the biological action spectrum for carcinogenesis, A ( h ) , and the intensity of the UV-radiation, I(h,t):
D
=
1
(IWW
r A ( A ) x I(X,t)dhdt
(1)
291)
The integration in time includes a full year. I(h,t) includes the direct solar radiation as well as all orders of scattered light (Rayleigh scattering) from the atmosphere. We applied a numerical implementation of the discrete ordinate method for radiative transfer introduced by Stamnes eta/. (1988). I(A,t) varies with time of the day and of the year, geographical latitude and with the thickness of the ozone layer. The action spectrum, A ( h ) , is quite essential in these calculations. Because knowledge about the action spectrum for human skin cancer is lacking, we have used the action spectrum proposed by Commission Internationale d e I’Eclaire (CIE) (McKinlay and Diffey, 1987). This action spectrum is in good agreement with that published for skin cancer in mouse epidermis by Sterenborg and van der Leun (1987) as well as a spectrum for mutation of cells in the basal layer of the skin (Moan et al.. 1988). It must be pointed out that the CIE action spectrum extends up to 400 nm and thus includes a wavelength
*To whom correspondence should be addressed. 579
580
THORMOD HENRIKSEN et al. Table 1. Ultraviolet-radiation and skin cancer in Norway Region
Population
UV-dose
M.M.
Non-M.
1. 69.5“ 2. 63.5” 3. 60.0“ 4. 58.5“
133000 307000 1487000 398000
0.66 0.86 1.00 1.07
5.1 6.6 10.6 11.2
22.8 34.8 47.4 52.5
The UV-dose is given in relative units and set equal to 1.OOfor 60” N (corresponds t o Oslo) with normal ozone conditions (the zonal average) throughout the year. M.M. is malignant melanoma and the numbers are the annual incidence rate per 100000. Non-M. is nonmelanoma skin cancer. The data are valid for the period 197Cb1979.
Figure 1. The localization of the different regions in Norway used for the evaluation of the dose effect curves for UV-induced skin cancer. The four regions cover 1” of latitude (111 km) in the north-south direction. Region 1 includes all people living between 69 and 70” N (133 000). The number of inhabitants in region 2 is 307 000, in region 3 (which includes both Oslo and Bergen) 1.487 million and in the fourth region 400 000. 0.2
0.4
0.6
0.8
1.0
1.2
1.4
Annual UV-dose (relative units)
region (35C-400 nm) where ozone has a negligible absorption (it is a region with a “window” in the ozone absorption spectrum). Since the solar UV-radiation is rather strong in the UV-A-region the application of the CIE action spectrum implies that variations in the ozone layer has a smaller biological effect than previously assumed. Annual effective UV-doses were calculated according to Eq. 1 and given in relative units. The dose for 60’” (region 3, Fig. l ) , with normal ozone conditions throughout the year, was set equal to 1.00. The details of the dose calculations as well as the variation of dose t ’ ~latitude are given by Dahlback et al. (1988). Incidence ofskin cancer in Norway. Norway extends in the north-south direction from 71 to 58” N and has a population of approximately 4 million. Four different areas were selected as shown in Fig. I , each covering a latitude region of 1” (111 km). Skin cancer data from the Norwegian Cancer Registry (Glattre et ul., 1985) for the period 1970-1979 were used. The skin cancer data were divided into two categories: “malignant melanoma“ and “non-melanoma skin cancer“. The malignant melanoma group includes a total of 3681 cases for the 10 yr period. The average age-adjusted incidence rate for the country was 8.1 (men) and 10.0 (women) new cases per 100 000 per year in the period. The incidence increased by approximately 7% per year for the period. The mean value of the incidence rate for men and women in the different regions was used in this work. The data include both cities and rural areas. “Non-melanoma skin cancer” constitutes by far the largest group and includes both “basal cell carcinoma“ and “squumous cell carcinoma”. In another paper (Moan er al.. 1989) the nonmelanoma skin cancer has been treated separately. The cancer registry did not include basal cell carcinoma until 1972 and it is reasonable to assume that this type was underreported during the first years. However, we have no reason to believe that there have been any regional differences in reporting to the cancer registry.
Figure 2. Skin cancer in Norway vs the annual environmental UV-dose. Open triangles, with scale to the left, are non-melanoma skin cancer (basal cell carcinoma and squamous cell carcinoma). Solid circles, with scale to the right, are for malignant melanoma. The dotted straight line is fitted to the data, whereas the linear and quadratic curves are the theoretical ones through the point for region 3, with dose 1.0. RESULTS AND DISCUSSION
The incidence of skin cancer in four different regions of Norway is given in Table 1 and in Fig. 2 as a function of the annual environmental effective UV-dose. There is a significant increase in the incidence of skin cancer (both melanomas and non-melanomas) when the environmental effective annual UV dose increases. As shown in Fig. 2 , the results can be described either by a quadratic function:
R
=
CD2
where C is a constant and R is the annual incidence per lo5or by a linear relationship including a threshold dose:
R = C, + CzD
(3)
where C, and Cz are constants. The concept of “threshold dose” is discussed in the case of ionizing radiation as well as with regard
Ozone depletion and skin cancer to the effect of chemicals. The results in Fig. 2 may be interpreted in a similar way, implying that below a certain threshold dose no skin cancer can be observed. It can be pointed out that the experiments of Morrison et al. (1979) and Daniels (1977) seem to indicate that low UV-doses strengthen the immune system, whereas large UV-doses induce a transient weakening of the immune system. It would be extremely interesting to gather more data from low dose regions, like the Scandinavian countries, in order to see if there is further support for a threshold dose theory. It must be pointed out that the UV-dose in this work is the environmental effective annual dose to the area and not the UV-dose to the human skin. Even though the results in Fig. 2 are well suited for discussions including ozone layer depletions it remains to be seen whether the data can be used to arrive at the dose effect curve for UV-carcinogenesis. A dose-response relationship of the type: R
=
CDP
(4)
where C and p are constants has been proposed on the basis of epidemiological data (Green et al., 1976; Green and Hedinger, 1978; Van der Leun, 1984). One factor which may influence the present results is travel out of the regions--in particular summer vacations to the Mediterranean beaches. In the course of a couple of weeks the UV-dose to the skin may be significant even compared with the total annual dose in the Nordic countries. The cancer data used in this work are from the period 1970-1979 which implies that the carcinogenic exposures occurred mainly before that time (all radiation induced cancers appear after a latency period which is largely unknown) and consequently before charter travel became popular. It is of interest to note that malignant melanoma exhibits approximately the same relationship to the environmental UV-dose as that found for non-melanoma skin cancer (Fig. 2 ) . While nonmelanomas are predominantly located at sun-exposed areas of the body (face, hands, etc.), melanomas tend to be more evenly distributed although they occur about five times as densely on facial skin as on the body as a whole (Pearl and Scott. 1986). It can be noted that the data for melanomas include both small and large cities as well as rural areas. If, for example, the two cities Oslo and Bergen with more than 75 000 inhabitants are excluded, the incidence rate for region 3 (60" N) would decrease by approximately 4%. Another and larger effect is observed if the data for region 3 are divided into East- and West-Norway. Thus, the incidence rate for the western part of the country is about 17% lower than that found for East-Norway . This difference presumably reflects the weather and the possibilities for sun exposure.
581
The incidence of non-melanomas seem to be related mainly to the accumulated dose of carcinogenic sunlight, whereas melanomas seem to be related to episodes of sunburn which implies that the dose rate is important (Pearl and Scott, 1986; Bsterlind et al., 1987; Brodthagen et al., 1987).
O z o n e layer depletion and skin cancer In a previous paper (Dahlback et al., 1988), we calculated effective UV-doses for a normal ozone layer as well as for various depletions, transient, like the Antarctic ozone hole, as well as depletions which are assumed t o remain constant throughout the year. Any depletion of the ozone layer will increase the effective dose rate as well as the annual effective UV-dose. This may in turn increase the incidence rate of skin cancer. This increase, which sometimes is called the "amplification factor, can be defined as the percentage increase in skin cancer incidence due to a 1 % decrease of the ozone layer. The amplification factor can be split into a product of two other amplification factors: a radiation arnplification factor (A,) and a biological amplification factor ( A b ) . The radiation amplification factor is defined as the percent increase in the effective environmental UV-dose per percent decrease in ozone (a persistent depletion of 1% throughout the year). A s shown by Dahlback et al. (1989) the radiation amplification factor is not a constant. However for depletions u p to 10"/0 in the Nordic countries the radiation amplification factor is close to 1.0. This implies that a 4% depletion throughout the year would yield UV-doses and dose rates similar to those obtained 111 km (corresponding to 1" of latitude) to the south. D u e to the fact that the ozone absorption in the UV-A-region is negligible, the parameter A , depends strongly on the action spectrum used as well as on the extent of the wavelength integration in Eq. 1. The biological amplification factor, Ah, defined as the percent increase in skin cancer per percent increase in the effective annual UV-dose, can be estimated from the slopes of the curves in Fig. 2. For a quadratic dose response curve the amplification factor is 2.0 at all dose levels ( A h = DlR,dRldD). For a linear dose response curve with a threshold value like that shown in Fig. 2 the biological amplification factor will be between 1.7 and 2.3 in the actual dose range. Several other investigations have estimated Ah to be of this magnitude (van der Leun, 1984, and references therein). According to the present work, the overall amplification factor would be -2.0. This implies that an ozone depletion of 1% throughout the year will result in an increase in the incidence of skin cancer of about 2%. REFERENCES
Brodthagen, H., K. Thestrup-Pedersen and A. Bsterlind (1987) Sunshine and malignant melanoma. Nord. Cun-
582
THORMOD HENRIKSEN et al.
cer Union Symp. Radiat. Cancer Risk 35 (Abstr.). Dahlback, A., T. Henriksen, S. H. H. Larsen and K. Stamnes (1989) Biological UV-doses and effect of an ozone layer depletion. Photochem. Photobiol. 49, 621-625. Daniels, F. (1977) Physiological and pathological extracutaneous effects of light on man and mammals, not mediated by pineal or other neuroendocrine mechanisms. In Sunlight and Man (Edited by T .B. Fitzpatrick, M. Pathak, L. C. Haber, M. Seiji and A. Kukubita), pp.217-258. University of Tokyo Press, Tokyo. Glattre, E., T. E. Finne, 0. Olesen and F. Landmark (1985) Atlas of Cancer Incidence in Norway 1970-1979. The Norwegian Cancer Society, Oslo. Green, E. E. S., G. B. Findley, K. F. Klenk, W. M. Wilson and T. Mo (1976) The ultraviolet dose dependence of non melanoma skin cancer incidence. Photochem. Photobiol. 24, 353-362. Green, E. E. S. and R. A. Hedinger (1978) Models relating ultraviolet light and non-melanoma skin cancer incidence. Photochem. Photobiol. 28, 283-291. Leenhouts, H. P. and K. H. Chadwick (1987) Fundamental aspects of the dose effect relationship for ultraviolet radiation. In Human Exposure to Ultraviolet Radiation (Edited by W. F. Passchier and B. F. M. Bosnjakovic), pp. 21-25. Excerpta Medica, Amsterdam. van der Leun, J. (1984) UV-carcinogenesis. Photochem. Photobiol. 39. 851-859.
McKinlay, A. F. and B. L. Diffey (1987) A reference spectrum for ultraviolet induced erythema in human skin. CIE-J. 6 , 17-22. Moan, J., S. Larsen, A. Dahlback and T. Henriksen (1988) UV-radiation and skin cancer (In Norwegian). J . Norwegian Med. Assoc. 31, 2838-2840. Moan, J . , A. Dahlback, T. Henriksen and K. Magnus (1989) Biological amplification factor for sunlightinduced nonmelanoma skin cancer at high latitudes. Cancer Res. 49, 5207-5212. Morrison, W. L., J . A. Parrish, and J. H. Epstein (1979) Photoimmunology. Arch. Dermatol. 115, 350-355. Bsterlind, A , , M. A. Tucker and 0. M. Jensen (1987) The Danish case-control study of skin melanomaimportance of UV-light exposure. Nord. Cancer Union Symp. Radiat. Cancer Risk 36 (Abstr.). Pearl, D. K. and E. L. Scott (1986) The anatomical distribution of skin cancers. Int. J . Epidemiol. 15, 502-506. Stamnes, K., S. Tsay, W. Wiscombe and K. Jayaweera (1988) Numerically stable algorithm for discrete-ordinate method radiative transfer in multiple scattering and emitting media. Appl. Opt. 27, 2502-2509. Sterenborg, H. J. C. M. and J. van der Leun (1987) Action spectra for tumorigenesis by ultraviolet radiation. In Human Exposure to Ultraviolet Radiation (Edited by W. F. Passchier and B. F. M. Bosnjakovic), pp. 173-191. Excerpta Medica, Amsterdam.
Photochemistry and Photobiology Vol. 51, No. 5 , pp. 579-582, 1990 Printed in Great Britain. All rights reserved
ULTRAVIOLET-RADIATION AND SKIN CANCER. EFFECT OF AN OZONE LAYER DEPLETION THORMOD HENRIKSEN*, ARNEDAHLBACK, SBRENH.H. LARSEN and JOHAN MOAN Institute of Physics, University of Oslo. Blindern, 0316 Oslo 3, Norway. (Received 29 July 1989; accepted 28 November 1989) Abstract-The effect of changes in the ozone layer on the incidence of skin cancer was explored using data for Norway. Attempts were made to arrive at a relationship between the “environmental effective UV-dose” and the skin cancer incidence. Norway is well suited for this purpose because of the large variation in the annual UV-dose from north to south. Furthermore we have a well developed cancer registry and a homogeneous population with regard to skin type. Four different regions of the country, each with a broadness of 1” in latitude (approximately 111 km), were selected (located around 69.5, 63.5, 60 and 58.5” N). The annual effective UV-doses for these regions were calculated, assuming normal ozone conditions throughout the year and the action spectrum proposed by CIE, which extends u p to 400 nm. The incidence rate (in the period 1970-1980) of malignant melanoma and non-melanoma skin cancer (mainly basal cell carcinoma) increased with the annual environmental UV-doses. For both these types of cancer a quadratic dose-effect relationship seems to be valid to a first approximation. The oresent data indicate that the incidence of skin cancer would increase by approximately 2% for each percent ozone reduction.
information is used in an attempt to arrive at the relationship between the environmental effective UV-dose and the different types of human skin cancer. Norwegian data are particularly well suited for this purpose for three reasons; (1) Norway is quite long in the north-south direction (a latitude difference of 13”) which implies a variation in the annual UV-dose of about 50%. ( 2 ) The population is more or less of the same skin type (Caucasian race). (3) Norway has a good National Cancer Registry.
INTRODUCTION
A depletion of the atmospheric ozone layer would undoubtedly result in larger UV-dose rates as well as in larger annual “effective UV-doses”. This may in turn result in an increase in detrimental effects such as the incidence of skin cancer. In order to evaluate the extent of this increase, knowledge about the ozone layer thickness and the environmental “carcinogenic sunlight” (the part of the solar spectrum which overlaps with the action spectrum for skin cancer} is needed. Furthermore, information about the relationship between the environmental UV-dose and skin cancer is essential. In previous papers (Dahlback et al., 1989; Moan et al., 1988, 1989) we have calculated effective UVdoses both for a normal as well as a depleted ozone layer, including ozone holes such as that observed in Antarctica. It was found that a small depletion of the ozone layer would increase the effective environmental UV-dose with a “radiation amplification factor” close to 1.0 (this factor depends on the latitude as well as on the action spectrum used). This implies that 1% depletion of the ozone layer throughout the year would increase the annual dose of carcinogenic sunlight by 1% at our latitude (60” N). For normal ozone conditions the annual effective UV-dose was found to vary by approximately 4 5 % per degree of latitude in the region 6C-70” N. At mid-latitudes this factor increases to about 7 % per degree. Consequently, a depletion of the ozone layer of 4 5 % throughout the year in Scandinavia would yield a UV-radiation milieu corresponding to that found for normal ozone conditions at approximately 1” of latitude (111 km) to the south. This
MATERIALS AND METHODS
UV-doses. The effective annual UV-dose of carcinogenic sunlight can be defined as the integrated product of the biological action spectrum for carcinogenesis, A ( h ) , and the intensity of the UV-radiation, I(h,t):
D
=
1
(IWW
r A ( A ) x I(X,t)dhdt
(1)
291)
The integration in time includes a full year. I(h,t) includes the direct solar radiation as well as all orders of scattered light (Rayleigh scattering) from the atmosphere. We applied a numerical implementation of the discrete ordinate method for radiative transfer introduced by Stamnes eta/. (1988). I(A,t) varies with time of the day and of the year, geographical latitude and with the thickness of the ozone layer. The action spectrum, A ( h ) , is quite essential in these calculations. Because knowledge about the action spectrum for human skin cancer is lacking, we have used the action spectrum proposed by Commission Internationale d e I’Eclaire (CIE) (McKinlay and Diffey, 1987). This action spectrum is in good agreement with that published for skin cancer in mouse epidermis by Sterenborg and van der Leun (1987) as well as a spectrum for mutation of cells in the basal layer of the skin (Moan et al.. 1988). It must be pointed out that the CIE action spectrum extends up to 400 nm and thus includes a wavelength
*To whom correspondence should be addressed. 579
580
THORMOD HENRIKSEN et al. Table 1. Ultraviolet-radiation and skin cancer in Norway Region
Population
UV-dose
M.M.
Non-M.
1. 69.5“ 2. 63.5” 3. 60.0“ 4. 58.5“
133000 307000 1487000 398000
0.66 0.86 1.00 1.07
5.1 6.6 10.6 11.2
22.8 34.8 47.4 52.5
The UV-dose is given in relative units and set equal to 1.OOfor 60” N (corresponds t o Oslo) with normal ozone conditions (the zonal average) throughout the year. M.M. is malignant melanoma and the numbers are the annual incidence rate per 100000. Non-M. is nonmelanoma skin cancer. The data are valid for the period 197Cb1979.
Figure 1. The localization of the different regions in Norway used for the evaluation of the dose effect curves for UV-induced skin cancer. The four regions cover 1” of latitude (111 km) in the north-south direction. Region 1 includes all people living between 69 and 70” N (133 000). The number of inhabitants in region 2 is 307 000, in region 3 (which includes both Oslo and Bergen) 1.487 million and in the fourth region 400 000. 0.2
0.4
0.6
0.8
1.0
1.2
1.4
Annual UV-dose (relative units)
region (35C-400 nm) where ozone has a negligible absorption (it is a region with a “window” in the ozone absorption spectrum). Since the solar UV-radiation is rather strong in the UV-A-region the application of the CIE action spectrum implies that variations in the ozone layer has a smaller biological effect than previously assumed. Annual effective UV-doses were calculated according to Eq. 1 and given in relative units. The dose for 60’” (region 3, Fig. l ) , with normal ozone conditions throughout the year, was set equal to 1.00. The details of the dose calculations as well as the variation of dose t ’ ~latitude are given by Dahlback et al. (1988). Incidence ofskin cancer in Norway. Norway extends in the north-south direction from 71 to 58” N and has a population of approximately 4 million. Four different areas were selected as shown in Fig. I , each covering a latitude region of 1” (111 km). Skin cancer data from the Norwegian Cancer Registry (Glattre et ul., 1985) for the period 1970-1979 were used. The skin cancer data were divided into two categories: “malignant melanoma“ and “non-melanoma skin cancer“. The malignant melanoma group includes a total of 3681 cases for the 10 yr period. The average age-adjusted incidence rate for the country was 8.1 (men) and 10.0 (women) new cases per 100 000 per year in the period. The incidence increased by approximately 7% per year for the period. The mean value of the incidence rate for men and women in the different regions was used in this work. The data include both cities and rural areas. “Non-melanoma skin cancer” constitutes by far the largest group and includes both “basal cell carcinoma“ and “squumous cell carcinoma”. In another paper (Moan er al.. 1989) the nonmelanoma skin cancer has been treated separately. The cancer registry did not include basal cell carcinoma until 1972 and it is reasonable to assume that this type was underreported during the first years. However, we have no reason to believe that there have been any regional differences in reporting to the cancer registry.
Figure 2. Skin cancer in Norway vs the annual environmental UV-dose. Open triangles, with scale to the left, are non-melanoma skin cancer (basal cell carcinoma and squamous cell carcinoma). Solid circles, with scale to the right, are for malignant melanoma. The dotted straight line is fitted to the data, whereas the linear and quadratic curves are the theoretical ones through the point for region 3, with dose 1.0. RESULTS AND DISCUSSION
The incidence of skin cancer in four different regions of Norway is given in Table 1 and in Fig. 2 as a function of the annual environmental effective UV-dose. There is a significant increase in the incidence of skin cancer (both melanomas and non-melanomas) when the environmental effective annual UV dose increases. As shown in Fig. 2 , the results can be described either by a quadratic function:
R
=
CD2
where C is a constant and R is the annual incidence per lo5or by a linear relationship including a threshold dose:
R = C, + CzD
(3)
where C, and Cz are constants. The concept of “threshold dose” is discussed in the case of ionizing radiation as well as with regard
Ozone depletion and skin cancer to the effect of chemicals. The results in Fig. 2 may be interpreted in a similar way, implying that below a certain threshold dose no skin cancer can be observed. It can be pointed out that the experiments of Morrison et al. (1979) and Daniels (1977) seem to indicate that low UV-doses strengthen the immune system, whereas large UV-doses induce a transient weakening of the immune system. It would be extremely interesting to gather more data from low dose regions, like the Scandinavian countries, in order to see if there is further support for a threshold dose theory. It must be pointed out that the UV-dose in this work is the environmental effective annual dose to the area and not the UV-dose to the human skin. Even though the results in Fig. 2 are well suited for discussions including ozone layer depletions it remains to be seen whether the data can be used to arrive at the dose effect curve for UV-carcinogenesis. A dose-response relationship of the type: R
=
CDP
(4)
where C and p are constants has been proposed on the basis of epidemiological data (Green et al., 1976; Green and Hedinger, 1978; Van der Leun, 1984). One factor which may influence the present results is travel out of the regions--in particular summer vacations to the Mediterranean beaches. In the course of a couple of weeks the UV-dose to the skin may be significant even compared with the total annual dose in the Nordic countries. The cancer data used in this work are from the period 1970-1979 which implies that the carcinogenic exposures occurred mainly before that time (all radiation induced cancers appear after a latency period which is largely unknown) and consequently before charter travel became popular. It is of interest to note that malignant melanoma exhibits approximately the same relationship to the environmental UV-dose as that found for non-melanoma skin cancer (Fig. 2 ) . While nonmelanomas are predominantly located at sun-exposed areas of the body (face, hands, etc.), melanomas tend to be more evenly distributed although they occur about five times as densely on facial skin as on the body as a whole (Pearl and Scott. 1986). It can be noted that the data for melanomas include both small and large cities as well as rural areas. If, for example, the two cities Oslo and Bergen with more than 75 000 inhabitants are excluded, the incidence rate for region 3 (60" N) would decrease by approximately 4%. Another and larger effect is observed if the data for region 3 are divided into East- and West-Norway. Thus, the incidence rate for the western part of the country is about 17% lower than that found for East-Norway . This difference presumably reflects the weather and the possibilities for sun exposure.
581
The incidence of non-melanomas seem to be related mainly to the accumulated dose of carcinogenic sunlight, whereas melanomas seem to be related to episodes of sunburn which implies that the dose rate is important (Pearl and Scott, 1986; Bsterlind et al., 1987; Brodthagen et al., 1987).
O z o n e layer depletion and skin cancer In a previous paper (Dahlback et al., 1988), we calculated effective UV-doses for a normal ozone layer as well as for various depletions, transient, like the Antarctic ozone hole, as well as depletions which are assumed t o remain constant throughout the year. Any depletion of the ozone layer will increase the effective dose rate as well as the annual effective UV-dose. This may in turn increase the incidence rate of skin cancer. This increase, which sometimes is called the "amplification factor, can be defined as the percentage increase in skin cancer incidence due to a 1 % decrease of the ozone layer. The amplification factor can be split into a product of two other amplification factors: a radiation arnplification factor (A,) and a biological amplification factor ( A b ) . The radiation amplification factor is defined as the percent increase in the effective environmental UV-dose per percent decrease in ozone (a persistent depletion of 1% throughout the year). A s shown by Dahlback et al. (1989) the radiation amplification factor is not a constant. However for depletions u p to 10"/0 in the Nordic countries the radiation amplification factor is close to 1.0. This implies that a 4% depletion throughout the year would yield UV-doses and dose rates similar to those obtained 111 km (corresponding to 1" of latitude) to the south. D u e to the fact that the ozone absorption in the UV-A-region is negligible, the parameter A , depends strongly on the action spectrum used as well as on the extent of the wavelength integration in Eq. 1. The biological amplification factor, Ah, defined as the percent increase in skin cancer per percent increase in the effective annual UV-dose, can be estimated from the slopes of the curves in Fig. 2. For a quadratic dose response curve the amplification factor is 2.0 at all dose levels ( A h = DlR,dRldD). For a linear dose response curve with a threshold value like that shown in Fig. 2 the biological amplification factor will be between 1.7 and 2.3 in the actual dose range. Several other investigations have estimated Ah to be of this magnitude (van der Leun, 1984, and references therein). According to the present work, the overall amplification factor would be -2.0. This implies that an ozone depletion of 1% throughout the year will result in an increase in the incidence of skin cancer of about 2%. REFERENCES
Brodthagen, H., K. Thestrup-Pedersen and A. Bsterlind (1987) Sunshine and malignant melanoma. Nord. Cun-
582
THORMOD HENRIKSEN et al.
cer Union Symp. Radiat. Cancer Risk 35 (Abstr.). Dahlback, A., T. Henriksen, S. H. H. Larsen and K. Stamnes (1989) Biological UV-doses and effect of an ozone layer depletion. Photochem. Photobiol. 49, 621-625. Daniels, F. (1977) Physiological and pathological extracutaneous effects of light on man and mammals, not mediated by pineal or other neuroendocrine mechanisms. In Sunlight and Man (Edited by T .B. Fitzpatrick, M. Pathak, L. C. Haber, M. Seiji and A. Kukubita), pp.217-258. University of Tokyo Press, Tokyo. Glattre, E., T. E. Finne, 0. Olesen and F. Landmark (1985) Atlas of Cancer Incidence in Norway 1970-1979. The Norwegian Cancer Society, Oslo. Green, E. E. S., G. B. Findley, K. F. Klenk, W. M. Wilson and T. Mo (1976) The ultraviolet dose dependence of non melanoma skin cancer incidence. Photochem. Photobiol. 24, 353-362. Green, E. E. S. and R. A. Hedinger (1978) Models relating ultraviolet light and non-melanoma skin cancer incidence. Photochem. Photobiol. 28, 283-291. Leenhouts, H. P. and K. H. Chadwick (1987) Fundamental aspects of the dose effect relationship for ultraviolet radiation. In Human Exposure to Ultraviolet Radiation (Edited by W. F. Passchier and B. F. M. Bosnjakovic), pp. 21-25. Excerpta Medica, Amsterdam. van der Leun, J. (1984) UV-carcinogenesis. Photochem. Photobiol. 39. 851-859.
McKinlay, A. F. and B. L. Diffey (1987) A reference spectrum for ultraviolet induced erythema in human skin. CIE-J. 6 , 17-22. Moan, J., S. Larsen, A. Dahlback and T. Henriksen (1988) UV-radiation and skin cancer (In Norwegian). J . Norwegian Med. Assoc. 31, 2838-2840. Moan, J . , A. Dahlback, T. Henriksen and K. Magnus (1989) Biological amplification factor for sunlightinduced nonmelanoma skin cancer at high latitudes. Cancer Res. 49, 5207-5212. Morrison, W. L., J . A. Parrish, and J. H. Epstein (1979) Photoimmunology. Arch. Dermatol. 115, 350-355. Bsterlind, A , , M. A. Tucker and 0. M. Jensen (1987) The Danish case-control study of skin melanomaimportance of UV-light exposure. Nord. Cancer Union Symp. Radiat. Cancer Risk 36 (Abstr.). Pearl, D. K. and E. L. Scott (1986) The anatomical distribution of skin cancers. Int. J . Epidemiol. 15, 502-506. Stamnes, K., S. Tsay, W. Wiscombe and K. Jayaweera (1988) Numerically stable algorithm for discrete-ordinate method radiative transfer in multiple scattering and emitting media. Appl. Opt. 27, 2502-2509. Sterenborg, H. J. C. M. and J. van der Leun (1987) Action spectra for tumorigenesis by ultraviolet radiation. In Human Exposure to Ultraviolet Radiation (Edited by W. F. Passchier and B. F. M. Bosnjakovic), pp. 173-191. Excerpta Medica, Amsterdam.
