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
5
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.
6
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.
7
Many gaps still exist, however, and this study has tried to address one of the neglected
areas of occupational risks for breast cancer development.