Cancers of the cervix uteri and breast: changes in incidence rates in Bombay over the last two decades

Bulletin of the World Health Organization, 64 (3): 431-435 (1986) © World Health Organization 1986 Cancers of the cervix uteri and breast: changes in incidence rates in Bombay over the last two decades KASTURI JAYANT1 The age-adjusted incidence rate Qf cervical cancer in Bombay has decreased over the last 20 years, whereas over the sameperiod that ofbreast cancer increased slightly. Studies indicate that the incidence ofcervical cancer is lower in successively later-born cohorts, but no cohort effect has been observed for breast cancer. Age at marriage is an important factor in the etiology of both these cancers, and it is plausible that the decrease in cervical cancer has resultedfrom the increase in age at marriage of women over the last 20 years in the city. The changing pattern ofincidence ofcancerfor women in Bombay, among whom the incidence of breast cancer is the highest of all forms of cancer, indicates that, from a long-term perspective, cancer control programmes in India as a whole should aim at minimizing the risk of both cervical and breast cancer. The Bombay Cancer Registry, which was set up roughly 20 years ago, provides data on chronological trends in the incidence of various cancers in the city (1-3). The age-adjusted incidence rates for cervical and breast cancer in selected Indian cities together with the highest and lowest rates in the world are shown in Fig. 1. CANCER BREAST CERVIX UTERI BOMBAY 1964 - 20s4 24-7 1966-72 20 1 23 2 1973 -75 21 2 23 3 1938 23 1 167 POONA 1973-7 2330a 7 1e7-6o0 256 31 2 BANGALORE 1962 M., 39-7 MADRAS 1962 1ii 4 1 151 HIGHEST REPORTED E 6795 52*9 LOWEST REPORTED * J 39 Age-adjusted rate/ 100 000 WHO 551723 Fig. 1. Age-adjusted incidence rates (per 100 000) of breast cancer and cancer of the cervix uteri for different cities in India, as well as the highest and lowest rates reported globally. l Scientific Officer, Epidemiology Unit, Cancer Research Insti- tute, Tata Memorial Centre, Parel, Bombay 400 012, India. The incidence of cancer'of the cervix uteri in Bombay decreased from 24.7 per 100 000 in the 1960s to 23.2 per 100 000 in the 1970s, and was 18.7 per 100 000 in 1983. On the other hand, the incidence of breast cancer, which was about 20 per 100 000 in the 1960s and 1970s, has increased and in 1983 was 23.1 per 100 000. Although the incidence of breast cancer in Bombay is low compared to that in many other countries (2), it is the leading form of cancer among women in Bombay as in the developed countries. This is in contrast to the situation in the rest of India, where cancer of the cervix uteri predominates (Fig. 1 ) (2, 4, 5). We therefore studied the incidence of breast and cervical cancers in successive birth cohorts in Bombay and attempted to identify factors that might be responsible for the observed incidence patterns over the last 20 years. METHODS AND RESULTS Data on the incidence rates of cervical and breast cancer in Bombay are available for 1964-66, 1968-72, 1973-75, and 1983. If the mid-year for each of the first three of these periods is taken, i.e., 1965, 1970, and 1974, then, together with 1983, the inci- dence rates for successive 5th, 4th (which is approxi- mated as 5th for the analysis), and 9th (approximated as 10th) years are obtained. Table 1 shows the incidence rates for cervical cancer selected in this way plotted for 5-year birth cohorts from 1913 to 1948. It can be noted that the 482 -431- Table 1. Incidence rates (per 100 000) of cervical cancer among selected birth cohorts in Bombay Birth cohort Year reported 1948 1943 1938 1933 1928 1923 1918 1913 1965 0.2 (17)° 0.7 (22) 6.6 (27) 17.7 (32) 25.6 (37) 45.0 (42) 58.3 (47) 82.8 (52) 1970 1.2 (22) 4.2 (27) 13.7 (32) 29.7 (37) 50.8 (42) 66.4 (47) 72.7 (52) 53.9 (57) 1974 5.6 (27) 12.2 (32) 25.9 (37) 45.6 (42) 65.4 (47) 70.9 (52) 64.1 (57) 78.1 (62) 1983 18.2 (37) 40.1 (42) 45.6 (47) 66.1 (52) 70.0 (57) 53.4 (62) 42.5 (67) 41.1 (72) a Figures in parentheses indicate the mid-age in years of the birth cohort in the year reported. age-specific rates for successive cohorts generally decrease. In some cases, e.g., age group 30-34 years and cohorts born between 1933 and 1943, the de- crease is gradual, and in others, e.g., age group 35-39 years and cohorts born between 1928 and 1948, it fluctuates, ultimately reaching a low level. Although data are limited, the incidence of cervical cancer among cohorts born in 1928, 1933, 1938, 1943, and 1948 has successively decreased, indicating progres- sively less exposure to etiological factors. Table 2 shows the incidence rates for breast cancer in Bombay for 1965, 1970, 1974, and 1983 plotted against the 5-year birth cohorts from 1913 to 1948. The overall trend is not so apparent as in the case of cervical cancer, but for some cohorts the incidence rate increased over the period 1965 to 1983. For breast cancer no cohort effect is apparent. Etiology of cervical cancer For the etiology of cervical cancer the following variables are important: age at first coitus, sexual promiscuity, circumcision status of the partner (or more generally his genital hygiene), and infection with viral agents. A review of data from several studies has indicated that onset of sexual activity CANCER BREAST 1970-75 HINDUS 18- 4 MUSLIMS 24-4 CHRISTIANS 31-0 PARSIS 46- 0l WHO 851724 CERVIX UTERI 1973 -75 27.7 15S7 177 6-6 I I TI I I I 40 20 0 20 Age-adjusted rate / 100 000 Fig. 2. Incidence rates (per 100 000) of breast cancer and cancer of the cervix uteri among religious groups in Bombay. before the age of 17 years was the most powerful discriminating variable for cervical cancer (6). Circumcision status of the partner was not found to be of importance. More recently infection with Table 2. Incidence rates (per 100 000) of breast cancer among selected birth cohorts in Bombay Birth cohort Year reported 1948 1943 1938 1933 1928 1923 1918 1913 1965 0.4 (17) 1.3 (22) 2.9 (27) 10.1 (32) 19.4 (37) 29.4 (42) 41.6 (47) 65.3 (52) 1970 0.6 (22) 2.1 (27) 8.4 (32) 22.4 (37) 35.8 (42) 45.7 (47) 49.8 (52) 58.3 (57) 1974 3.5 (27) 8.1 (32) 21.0 (37) 31.4 (42) 52.5 (47) 60.9 (52) 68.7 (57) 64.2 (62) 1983 21.9 (37) 32.7 (42) 44.9 (47) 71.7 (52) 71.3 (57) 80.2 (62) 85.0 (67) 41.1 (72) a Figures in parentheses indicate the mid-age in years of the birth cohort in the year reported. 432 K. JAYANT INCIDENCE OF CERVICAL AND BREAST CANCER IN BOMBAY 433 Table 3. Average age at marriage of women in various religious groups in Bombay, India, 1966° Average age at marriage Religious group (years) Hindus 16.77 Muslims 16.05 Christians 19.65 Parsis 24.03 a See ref. 10. human (alpha) herpesvirus 1 or 2 and papillomavirus has been implicated in the etiology of cervical cancer (7). Incidence of cervical and breast cancer among religious groups The variation in the incidence of cervical cancer among different religious groups in Bombay as well as their sociocultural patterns may have contributed to the reduction in the incidence of cervical cancer described here. The age-adjusted incidence rates for cervical and breast cancer (Fig. 2) for these groups in Bombay have been studied (8, 9). Hindus exhibit the highest incidence rate of cervical cancer and Parsis the lowest, whereas Muslims and Christians have similar intermediate rates. Parsis have the highest incidence rate for breast cancer followed by Chris- tians and Muslims, while Hindus have the lowest rate. Parsi women enjoy a higher socioeconomic status, lead a more westernized lifestyle, and marry at a later age than their counterparts in other religious groups in Bombay (Table 3). The proportion of unmarried Parsi women is also high. Among Parsis these factors lower the risk of cervical cancer to one of the lowest in the world, but increase the risk of breast cancer. The difference in the incidence rates of cervical cancer in Hindus and Muslims does not arise because of different average ages at first coitus among the two religious groups, since age at marriage, which closely correlates with the age at first coitus,a is similar for Hindus and Muslims (Table 3). Among Muslims circumcision of males at a young age is performed for religious reasons, and this may account for the lower rate of cervical cancer among Muslim women. Of interest in this respect is the twofold higher risk of cervical cancer among Hindus a Data in ref 10 indicate that age at consummation of marriage was the same as age at marriage for over 90% ofwomen who married when t 14 years, andc80% of marriages in which the women's age at marriage was 13 years were consummated before she reached her 14th birthday. than Muslims (11). The low rate of cervical cancer among Christians in Bombay reported almost a decade ago is interesting, but remains unexplained (12). Data from other parts of India show that the relative frequency of cervical cancer among Christians is similar to that among Muslims (5). It is unlikely that the different incidence rates of cervical cancer among Hindus and Christians in Bombay arise because of differences in the circumcision status or genital hygiene, since in Bombay the incidence of penile cancer, which is closely related to the level of personal hygiene, is almost the same in these two groups (crude incidence rate: Christians, 1.5 per 100 000; Hindus, 1.3 per 100 000) (12).b The average age at marriage of Christian women in Bombay is 2-3 years higher than that of Hindus or Muslims (Table 3), and it is therefore possible that the lower rate of cervical cancer among Christians than Hindus is caused by this age difference. Further- more, it seems that, compared to Hindus, the higher age of Christian women at marriage has reduced the incidence of cervical cancer among them almost as much as has the practice of male circumcision among Muslims. In contrast, both the overall incidence and age-specific incidence of breast cancer among Chris- tians are much higher than those among Hindus or Muslims (8). The higher incidence of breast cancer in pre-menopausal women may arise because, on average, Christians are older at the first full-term birth (13). The age-specific incidences of pre-meno- pausal breast cancer among Muslims and Hindus are similar (8), perhaps because of the similar ages at marriage of women in these two religious groups. DISCUSSION Although there are no data on the age at marriage for women in successive birth cohorts, estimates compiled from decennial census surveys indicate that for the whole of India this increased, on average, from 12.5 years (over the period 1921-31) to 17.2 years (over the period 1961-71) (14). In Greater Bombay the mean age at marriage for women was 18.5 years in 1960 and 20.1 years in 1965 (10). The decline in the incidence of cancer of the cervix uteri in Bombay over the last 20 years therefore probably arises because of the increase in the average age at marriage over this period. The population of Bombay has doubled since 1964 and it is unlikely that the amenities for cleanliness have improved. So, im- proved personal hygiene in the population could not have contributed to the reduction in the rate. b The crude incidence rates among different religious groups in Bombay, except Parsis, are comparable, since the age structures of these subpopulations are similar. 434 K. JAYANT The results described here could have an important bearing on cancer control programmes in India as a whole. First, the changing pattern in the incidence of cancers among women in Bombay indicates that, from a long-term perspective, cancer control pro- grammes should aim at minimizing the risk of cervi- cal and breast cancer. While the situation in Bombay may be different from that in the rest of India at present, it is likely that the trends reported here may be relevant in future for the country as a whole. Since age at marriage is an important factor in the etiology of cervical and breast cancer, there is a need for epi- demiological studies in the same group or community to determine the optimum age at marriage that will minimize the risk of both these cancers. A secondary prevention programme for cervical cancer aimed at identifying a high-risk group among women who married while young might also prove to be cost- effective. RtSUMt CANCER DU COL UTERIN ET CANCER DU SEIN: MODIFICATIONS DE L'INCIDENCE A BOMBAY AU COURS DES DEUX DERNIERES DECENNIES Le taux comparatif d'incidence pour le cancer du col uterin a Bombay (Inde) est passe de 24,7 pour 100 000 dans les ann&es 1960 a environ 23 pour 100 000 dans les annees 1970 et itait d'environ 19 pour 100 000 au debut des annees 1980. En revanche, l'incidence du cancer du sein qui 6tait d'environ 20 pour 100 000 dans les ann&es 1960 et 1970 a augmente jusqu'a atteindre environ 23 pour 100 000 au debut des ann&es 1980. L'incidence du cancer du col dans les cohortes de naissance ulterieures a diminue au cours des deux derni&es d&cennies, ce qui indique que ces cohortes ont ete progressivement moins expos6es aux facteurs 6tio- logiques. Cet effet de cohorte n'a pas ete observ6 pour le cancer du sein. On a examine la variation d'incidence du cancer du col et du cancer du sein chez les femmes appartenant a diff6rents groupes religieux de Bombay, ainsi que les pratiques socio- culturelles de ces groupes, afin d'identifier l'effet que ces facteurs pourraient avoir eu sur la baisse d'incidence du cancer du col. Le taux comparatif d'incidence pour le cancer du col est plus elev6 chez les femmes appartenant a la reli- gion hindouiste (27,7 pour 100 000) et plus faible chez les femmes de religion musulmane (15,7 pour 100 000) et chre- tienne (17,7 pour 100 000). Le faible taux chez les musul- manes est connu depuis un certain temps et est attribue a la pratique de la circoncision masculine. Toutefois, le faible taux observ6 chez les chretiennes, bien que connu depuis une dizaine d'ann6es, demeure inexplique. II se pourrait toute- fois que l'age moyen au mariage, qui est superieur de 2 a 3 ans chez les chretiennes par rapport a l'Age moyen au mariage chez les musulmanes ou les hindoues ait une in- fluence sur l'incidence du cancer du col. De plus, I'age moyen au mariage, plus 6leve chez les chr6tiennes, a reduit l'incidence du cancer du col chez ces femmes presque dans les memes proportions que la pratique de la circoncision masculine l'a reduite chez les musulmanes. En Inde, I'age moyen des femmes au mariage a augment6 au cours des dernieres d6cennies, passant de 12,5 ans en 1921-31 A 17,2 ans en 1961-1971. A Bombay, I'age moyen des femmes qui se sont mariecs en 1960 6tait de 18,5 ans, et de 20,1 ans en 1965. La baisse d'incidence du cancer du col uterin observ6e au cours des 20 derni&es annees A Bombay et donc probablement due a l'augmentation de l'Age moyen au mariage chez les habitantes de cette ville. La modification des taux d'incidence du cancer du col uterin et du cancer du sein a Bombay semble indiquer que, dans une perspective A long terme, les programmes de lutte contre le cancer chez les femmes indiennes doivent viser A r6duire les risques associes A ces deux types de cancer. Dc plus, des etudes cpidemiologiques doivent etre entreprises pour 6valuer de faqon critique le role de l'Age au mariage chez les femmes dans 1'etiologie du cancer du col et du cancer du sein. REFERENCES 1. DOLL, R. ET AL. Cancer incidence in 5 continents, vol. 2. New York, International Union against Cancer, 1970. 2. WATERHOUSE, J. ET AL. Cancer incidence in 5 con- tinents, vol. 3 and 4. Lyon, International Agency for Research on Cancer (Scientific Publication No. 15 (1976) and No. 42 (1982), respectively). 3. JUSSAWALLA, D. J. ET AL. Cancer morbidity and mortality in Greater Bombay 1983. Bombay, Bombay Cancer Registry, Indian Cancer Society, 1984. 4. JUSSAWALLA, D. J. Cancer morbidity and mortality in Poona city agglomeration 1978-80. Bombay, Indian Cancer Society, 1983. 5. SANGHVI, L. D. & JAIN, D. K. Annual Report of the National Cancer Research Project. New Delhi, Indian Council of Medical Research, 1982. INCIDENCE OF CERVICAL AND BREAST CANCER IN BOMBAY 435 6. ROTKIN, I. D. A comparison review of key epidemio- logical studies in cervical cancer related to current searches for transmissible agents. Cancer research, 33: 1353-1367 (1973). 7. ZUR HAUSEN, J. Human genital cancer: synergism between two virus infections or synergism between a virus infection and initiating events. Lancet, 2: 1370-1372 (1982). 8. JUSSAWALLA, D. J. ET AL, Histological and epidemio- logical features of breast cancer in different religious groups in Greater Bombay. Journal of surgical oncology, 18: 269-279 (1981). 9. JUSSAWALLA, D. J. & YEOLE, B. B. Epidemiology of cancer of the cervix in Greater Bombay. Journal of surgical oncology, 26: 53-62 (1984). 10. RELE, J. R. & KANITKAR, T. Fertility andfamily plan- ning in Greater Bombay. Bombay, Popular Prakashan, 1980, pp. 38-87. 11. WAHI, P. N. ET AL. Prevalence and distribution of cancer of the uterine cervix in Agra. Cancer, 30: 720-725 (1972). 12. Cancer incidence in Greater Bombay 1970-72. Bombay, Bombay Cancer Registry, Indian Cancer Society, 1976. 13. MACMAHON, B. ET AL. Etiology of human breast cancer: a review. Journal of the National Cancer Institute, 50: 21-42 (1973). 14. AGARWALA, S. N. India's population problems. New Delhi, Tata McGraw Hill Publishing Co. Ltd, 1977, p. 92.