Trends in breast, ovarian and cervical cancer incidence in Mumbai, India over a 30-year period, 1976-2005: an age-period-cohort analysis.

BACKGROUND: Demographic, socioeconomic and cultural changes in India have increased longevity, delayed childbearing, decreased parity and resulted in a more westernised lifestyle, contributing to the increasing burden of cancer, especially among women. METHODS: We evaluated secular changes in the incidence of breast, cervical and ovarian cancer in Mumbai women aged 30-64 between 1976 and 2005. Age-standardised incidence rates were calculated and presented by site and calendar period. An age-period-cohort (APC) analysis quantified recent time trends and the significance of birth cohort and calendar period effects. The estimated annual percent change (EAPC) was obtained from the drift parameter, expressing the linear time trend common to both calendar period and birth cohort. RESULTS: Over the 30-year study period, the age-standardised rates significantly increased for breast cancer (EAPC: 1.1% (95% confidence interval (CI): 1.0, 1.3)), significantly decreased for cervical cancer (EAPC: -1.8% (95% CI: -2.0, -1.6)) and there was no statistically significant change for ovarian cancer (EAPC: 0.3% (95% CI: -0.1, 0.6)). For breast and cervical cancer, the best-fitting model was the APC model. CONCLUSIONS: The rates of breast, cervical and ovarian cancer remain low in comparison with western countries, and the divergent trends of breast (increasing) and cervical cancer (decreasing) in Mumbai were similar to those observed in several other Asian countries. The changing risk profile in successive generations - improved education, higher socioeconomic status, later age at marriage and at first child, and lower parity - may in combination partially explain the diverging generational changes in breast and cervical cancer in Mumbai in the last decades.

Cancer survival in Mumbai (Bombay), India, 1992-1999.

The Bombay cancer registry is the second oldest population-based cancer registry in Asia, and the first of its kind in India. It was established in 1963, and registration of cases is done by active methods. Data on survival from 28 cancer sites or types registered during 1992-1999 are reported. Follow-up has been carried out predominantly by active methods, with median follow-up ranging between 1-51 months for different cancers. The proportion of histologically verified diagnosis for various cancers ranged between 41-100%; death certificates only (DCOs) comprised 0-15%; 84-99% of total registered cases were included for survival analysis. Complete follow-up at five years ranged from 85-92% for different cancers. The 5-year age-standardized relative survival rates for common cancers were breast (48%), cervix (44%), lung (11%), oesophagus (14%), oral cavity (35%) and non-Hodgkin lymphoma (34%). The 5-year relative survival by age group portrayed either an inverse relationship or was fluctuating. Cases with a regional spread of disease were the highest for cancers of the tongue, oral cavity, larynx and cervix; survival decreased with the increasing extent of disease for all cancers studied.

Cumulative risk and trends in prostate cancer incidence in Mumbai, India.

BACKGROUND: Information relating to cancer incidence trends in a community forms the scientific basis for the planning and organization of prevention, diagnosis and treatment of cancer. We here estimated the cumulative risk and trends in incidence of prostate cancer in Mumbai, India, using data collected by the Bombay Population-based Cancer Registry from the year 1986 to 2000. METHODS: During the 15 year period, a total of 2864 prostate cancer cases (4.7% of all male cancers and 2.4% of all cancers) were registered by the Bombay Population-based Cancer Registry. For evaluation of the trend, we applied a linear regression model based on the logarithm of the observed incidence rates. The annual percentage changes were also computed for the evaluation. Cumulative incidence rates percentages were calculated by adding up the age specific incidence rates at single ages and then expressed as a percentage. RESULTS: Analysis of the trends in age-adjusted incidence rates of prostate cancer during the period 1986 to 2000 showed no statistically significant increase or decrease and the rates proved stable across the various age groups (00-49, 50-69 and 70+) also. The probability estimates indicated that one out of every 59 men will contract a prostate cancer at some time in his whole life and 99% of the chance is after he reaches the age of 50. CONCLUSION: The stability in age adjusted-incidence rates indicates that there are no changes in the etiological factors for prostate cancer in Mumbai, India. These findings may be of general interest because changes in diagnostic practices are confounded in the time trends of prostate cancer change in many western countries preventing inferences on the changes in risk.

An epidemiological assessment of increasing incidence and trends in breast cancer in Mumbai and other sites in India, during the last two decades.

A great deal is known on the epidemiology of breast cancer. In this paper an attempt has been made to discuss the epidemiology and trends in incidence of breast cancer in various populations of India with special reference to the data available at Mumbai Cancer Registry. For discussing descriptive epidemiology of breast cancer the data collected for most recent year, 1999, by Mumbai Cancer Registry has been utilized. For studying time trends in breast cancer the data collected for the Mumbai Cancer Registry for the years 1982-99 and for Bangalore and Chennai 1982-96 and for Barshi, Bhopal and Delhi for the years 1988-96 has been employed. A linear regression model based on the logarithms of the various incidence rates, a method frequently used for studying time trends, was applied to the entire dataset. Age specific incidence rates for breast cancer for most of the urban population in India were found to show steep increase till menopause years, after which the curves plateau. Most of the registries data indicate that Christians in India have the greatest risk of breast risk and Muslims have the lowest rate. In all the populations breast cancer was found to be less prevalent at the lower education level and the incidence increased with the education level. The trends for increase in breast cancer incidence over time for most of the populations in India were found to be statistically significant.

Expression of cell surface glycoprotein CD44 and integrins in breast cancers among Indian women.

Parsis, the sole surviving group of followers of Zoroaster who are settled in Bombay, have a fourfold higher incidence of breast cancer than the general population of Greater Bombay. CD44 expression was studied by immunohistochemistry in breast cancers of 50 non-Parsi and 35 Parsi women, 10 normal breast tissues, 10 proliferative lesions and 49 tissues adjoining a tumor mass. Alpha2 and beta1 integrins could be studied in only 42 malignant cases and five normal tissues. The immunohistochemistry results were correlated with other parameters including tumor grade and size, estrogen and progesterone receptor status, lymph node involvement and mitotic index. CD44 was not expressed in normal areas. Benign areas and tissues adjacent to tumor masses showed increased staining. Both Parsi and non-Parsi women showed significantly high CD44 expression. All Parsi ILCs were strongly positive for CD44. In both groups ER negativity was associated with strong CD44 positivity, while mitotic counts correlated with decreased CD44 expression in Parsis but not in non-Parsis. Alpha2 and beta1 integrins were strongly expressed on the basolateral surface of normal epithelium. However, they were downregulated in tumors. In general breast tumor tissues from Parsi and non-Parsi patients did not differ significantly with respect to most parameters. However, among Parsis lymph node involvement and CD44 correlated weakly whereas the mitotic index was inversely correlated with CD44. The reverse was true for non-Parsis. The deviation from the general pattern needs further study based on a large number of samples and appropriate use of splice variants.