Occupations and breast cancer in women treated at a tertiary hospital in Johannesburg
Abrahams, Odette Natasha
Introduction This is the first study in South Africa to look for an association between breast cancer and occupations in black women. Breast cancer is one of the commonest forms of malignancy experienced by women in South Africa and its incidence is increasing (1). Approximately six million women work in South Africa, some of these women are likely to be exposed to carcinogenic agents during their daily working lives. Many occupational carcinogens to the breast have been described and there is some evidence suggesting that many more synthetic chemicals used in different industries may also have carcinogenic properties that have not fully been explored as yet (2). This study plans to identify occupations that place black women at risk of breast cancer in the South African context. Thereafter, it will assess if there is an association between shift work (a known carcinogen to the breast) and breast cancer in black women in South Africa. The findings of the study could be of particular significance to the local context, given that women are entering the workforce in increasing numbers possibly putting more women at risk. Aims The aims of the study are to identify occupations that possibly increase the risk of breast cancer in South Africa, and to specifically assess if there is an association between shift work and breast cancer in black South African women. Objectives To determine whether there are associations between different occupations and breast cancer in black women by calculating the odds ratios (ORs) for breast cancer in different occupations in South Africa. To examine the association between shift work and breast cancer in black women adjusting for variables that may confound the association. Methods The study is an unmatched case-control study using secondary data from the existing Johannesburg Cancer Case Control Study (JCCS) database. The JCCS study is a large ongoing study that recruits black male and female cancer patients with all types of cancers receiving treatment at Charlotte Maxeke Johannesburg Academic Hospital (CMJAH), a tertiary hospital in Johannesburg. All women patients recruited from 1 January 2001 to 31 December 2009 were included in the analysis. This included 1 903 cases and 3 990 controls. An expert group was set up to estimate the likelihood of occupational engagement in shift work for the existing occupational groupings present in the JCCS database. These occupations were classified into: a) high probability of shift work; b) possibility of shift work and c) unlikely to do shift work. ORs were estimated using logistic regression. Those who had never worked were the reference category. Bivariate analysis was then carried out to estimate ORs between individual occupations and breast cancer and later to estimate ORs between the likelihood of involvement in shift work and breast cancer. Multivariate logistic regression followed a forward stepwise approach and all the possible confounders present in the JCCS database were considered. These included age, smoking, drinking alcohol, age at first menarche, parity, age at menopause, use of oral contraception and retroviral status. Results With regards to occupation, the bivariate analyses showed significant ORs for breast cancer in the occupational categories of health, education, social services, retail, hospitality, construction, food, textile and manufacturing, with the highest OR in education (OR 2.33, 95% CI: 1.44 – 3.78) and social service (OR 2.39, 95% CI: 1.24 – 4.58) followed by office workers (OR 2.17, 95% CI: 1.47-3.20) and health workers (OR 2.01, 95% CI: 1.33-3.04). Agriculture (OR 0.55, 95% CI: 0.32 – 0.94) and domestic workers (OR 0.96, 95% CI: 0.75 – 1.22) had ORs under one. Following the adjustments for possible confounders, no statistically significant ORs were found between specific occupations and breast cancer. However, manufacturing had the highest OR (OR 1.44, 95% CI: 0.42- 4.94), followed by office workers (OR 1.44 95% CI: 0.31- 5.94) and health workers (OR 1.31, 95% CI: 0.36-4.76) as compared to the never worked group. In the bivariate analyses there was an association between a possibility of doing shift workers and breast cancer (OR 1.66, 95% CI: 1.41-1.97). Similarly, women who had a high likelihood of participating in shift work had an elevated odds ratio for the disease (OR 1.76, 95% CI: 1.44-2.15). After adjusting for confounders in the multivariate model the ORs for both shift work groups were increased with women who possibly did shift work having an OR of 2.18 (95% CI: 1.34- 3.56) and those who had a high likelihood of carrying out shift work having an OR of 2.13 (95% CI: 1.26- 3.61). Conclusions The bivariate analyses identified elevated ORs for breast cancer in some occupations but in multivariate analyses no statistically significant OR were found. Nevertheless ORs>1 were found for the occupational categories of manufacturing, office and health workers. This study reports a strong association of breast cancer and shift work, which is supported by the literature. However shift work still remains necessary for the functioning of many industries. Many gaps still exist, however, and this study has tried to address one of the neglected areas of occupational risks for breast cancer development.
Submitted a requirements for the degree of Master of Medicine in the branch of Community Health (Occupational Medicine). It has not been used, either wholly or partially, for any other degree or examination at this or any other university.