Risk Factors for Breast Cancer: A Case-Control Study of 315 Women Followed in the Gynecology and Oncology Departments of Two University Teaching Hospitals in Yaounde, Cameroon ()
1. Introduction
Breast cancer is a major health problem worldwide, accounting for 16% of all cancers in women [1]
In the United States, it is the second most common cancer in women after skin cancer [2] . The incidence in France is 88 per 100,000 women making it the most common cancer in women [3] .
The incidence of breast cancer is lowest in African countries, with an average of 19.3 per 100,000 women compared to 89.7 per 100,000 women in Europe [4] . However, this incidence has been increasing in developing countries. Although considered a disease of developed countries, in 2008 nearly 50% of all cases of breast cancer occurred in developing countries [5] .
In Cameroon, the incidence of breast cancer is difficult to assess accurately. According to the 2012 Yaoundé cancer registry, 32% of cancers in women involved the breast [6] , making it the most common cancer in women in Cameroon. In 2014, Sando et al. at the Yaounde Gyneco-Obstetric and Pediatric Hospital (YGOPH) reported a 34% prevalence of breast cancer, making it the second most common gynecological cancer after cervical cancer [7] .
Worldwide, breast cancer mostly occurs after the age of 50 years [8] [9] . The situation is different in our context where it occurs at a relatively younger age, 46 years [7] . The majority of these cases in our setting are diagnosed at advanced stages of the disease because of difficulties in access to health care among other factors [9] , underscoring the importance of early diagnosis through screening.
According to the WHO, breast cancer is responsible for 25% of cancer deaths in the world with the majority of these deaths occurring in the developing countries [8] . Although incidence of breast cancer is lowest in African countries, survival rates are also lowest [10] . The majority of breast cancer deaths, 69%, occur in developing countries [11] .
Risk factors for breast cancer are described in the literature. Increasing age, never married, widowed, nulliparity, smoking, alcohol use, low fruit and vegetable intake, a positive family history of breast cancer, post-menopausal use of hormone replacement therapy, early age at menarche or late menopause, great age at first pregnancy severe obesity were cited by authors as remarkable risk factors [12] - [15] . Danaei et al. in 2008 found that 7 million breast cancer related deaths worldwide were caused by nine potentially modifiable risk factors [16] . They designated tobacco smoking, alcohol consumption and low fruit and vegetables intake as risk factors for low-and-middle income countries.
In Cameroon, few data are available on breast cancer associated risk factors. Epidemiological data including some nutritional risk factors for breast cancer remain insufficient.
Some studies have incriminated feeding habits in the induction of cancers in general [17] [18] and breast cancer in particular [19] [20] . Westernization of lifestyles of people in the developing countries can partially explain this increase in the incidence of breast cancer [21] . Overweight and obesity are modifiable risk factors for cancer related deaths [16] .
The aim of this study was to look for risk factors for breast cancer in patients treated for breast cancer at two university teaching hospitals in Yaoundé. The findings of this study could help improve the preventive methods based on the modifiable risk factors identified.
2. Methods
For 5 months, from February to July 2015, a case-control study was conducted in the gynecology unit of the Yaoundé Gyneco-Obstetric and Pediatric Hospital (YGOPH) and the Medical Oncology unit of the Yaoundé General Hospital (YGH). We compared women treated for breast cancer (cases) with women who did not have breast cancer (controls).
We obtained the approval of the protocol for the study from the ethical committee of the YGOPH and YGH. Cases were recruited from the departments concerned. Their medical records were consulted. Controls were recruited randomly from the outpatient departments of the same hospitals. All women who gave their free and informed consent were interviewed by an investigator and examined. The data collection was done using a pretested questionnaire. Pregnant women were excluded from the study.
The variables of interest were: age, level of education, marital status, income level, family and or personal history of breast cancer, the existence of another cancer in the family, age at menarche, age at first pregnancy, parity, duration of breastfeeding, a past history of use of hormonal contraception and hormone replacement therapy, duration of hormonal contraception use, infertility, menopause, benign breast pathology, alcohol and tobacco consumption, physical activity, obesity, diet (consumption of vegetables, leguminous plants, red meat, fish, conserves, palm oil, refined oil).
The calculated minimal sample size was 23 subjects for each group based on 34% prevalence of breast cancer in Cameroon [22] and according to the Schlesselman formula [23] . The chosen precision for statistical calculations was 5%. To enhance the strength of association, we considered one case for two controls.
We analyzed the data using Epi Info 3.5.4 software and Microsoft Excel 2010. The differences between proportions were analyzed using contingency tables and by applying the Khi-square test. A P-value < 0.05 was considered to be statistically significant. The association between variables was assessed by calculating the Odds Ratio (OR) with a confidence interval (CI) of 95%. Logistic regression was used to eliminate confounding factors and all variables with p < 0.02 were included in the regression model.
3. Results
During the five months recruitment period, we studied 315 women from the two services. 105 cases were compared with 210 controls. The mean age of the cases (49 ± 13 years) was statistically different (P < 0.001) from that of the controls (40 ± 12 years). The women in both groups were more often married, with low income level and had secondary level education (Table 1). The socio-demographic risk factors found (Table 1) were: age groups [50 - 60] (P = 0.012; OR = 2.14; CI = 1.17 - 3.9), [60 - 70] (P = 0.002; OR = 4.93; CI = 2.15 - 11.34) and [70 - 88] (P = 0.02; OR = 12.6; CI = 1.5 - 106.7), widowhood (P = 0.006; OR = 3.45; CI = 1.7 - 6.9]), monthly income level < 86 US dollars (P = 0.002; OR = 2.19; CI = 1.31 - 3.65), primary education level (P = 0.005,
Table 1. Comparison of socio-demographic variables among women with breast cancer (cases, n = 105) and those without breast cancer (controls, n = 210).
OR: Odds Ratio. CI: Confidence Interval. %: Percentage.
OR = 2.11; CI = 1.25 - 3.56),
Most of our patients in the two groups were multiparous and had more often practiced breastfeeding, 84.4% and 81.4% respectively in cases and controls. Duration of breastfeeding > 96 months (P = 0.01; OR = 2.65; CI = 1.59 - 4.39) was founded to be a risk factor (Table 2). Concerning other personal history such as age at menarche, previous breast pathology, personal history and family history of breast cancer, none was contributive as well as alcohol and tobacco consumption, physical activity and obesity (Table 2).
The unhealthy dietary habits identified were (Table 3): red meat consumption > 3 times weekly (P = 0.002; OR = 2.14; CI = 1.33 - 3.45); consumption of palm oil > 2 times per week (P = 0.001; OR = 2.38; CI = 1.4 - 4.1).
Protective factors for breast cancer were (Table 1 and Table 2): menopause (P = 0.002; OR = 0.37; CI = 0.22 - 0.62]), nulliparity (P = 0.008; OR = 0.46; CI = 0.26 - 0.08) and the higher level of education (P = 0.005; OR = 0.36; CI = 0.18 - 0.73).
After multivariate analysis, the risk factors for breast cancer identified were (Table 4): age > 50 years (P = 0.010; aOR = 41.48; CI = 2.46 - 699.1) and heavy consumption of red meat (P = 0.014; aOR = 7.33; CI = 1.49 - 36).
4. Discussion
In Cameroon, breast cancer onset is usually between the ages of 35 to 50 years [6] implying that the mean age at diagnosis is significantly below 50 years. Also breast cancer is often diagnosed in advanced stages of the condition [7] , implying an insidious onset to the disease long before presentation at the hospital. Although breast cancer onset is before 50years, we found out that the risk was greatest in patients above 50 years. This result is similar to findings by Nkondjock et al. [24] who found an increased risk of 7.1% between 50 and 74 years. Nevertheless, previous studies in our setting show that the average age of patients with breast cancer at diagnosis was lower than 50 years [7] [25] . This suggests involvement of other factors independent of patients’ ages.
Iraj et al. in Iran reported that breast cancer was diagnosed with advanced lesions in patients presenting at a state-owned hospital, and in low-income earners [26] . We carried out our study in two state-owned hospitals that manage the majority of cancer cases due to their relatively low treatment costs. Low purchasing power is a factor limiting access to health care. Income level < 86 dollars was identified as a risk factor for breast cancer. Low income is associated with low levels of education which was significantly associated with breast cancer in our study. Landolsi et al. in Tunisia found that most patients with breast cancer had low levels of education [27] . Breast cancer is often diagnosed at an advanced stage [9] corresponding to women who do not have easy access to health care. Conversely, Nkondjock et al. found that the risk of breast cancer is slightly increased in high income earners because they tend to carry fewer pregnancies or have children at an older age [24] . Mandana et al. found that women who had never been married before and widows were at increased risk for breast cancer [13] . Widowhood was not found to be an independent risk factor in our study despite the fact
Table 2. Comparing the history of women with breast cancer (cases, n = 105) with those of women without breast cancer (controls, n = 210).
OR: Odds Ratio. CI: Confidence Interval. %: Percentage.
Table 3. Comparison of dietary habits of women with breast cancer (cases, n = 105) and those without breast cancer (controls, n = 210).
OR: Odds Ratio. CI: Confidence Interval. %: Percentage.
OR: Odds Ratio. CI: Confidence Interval. %: Percentage.
that widows were the most represented group of women who have had breast cancer. This status is probably a confounding factor because most widows are often older than 50 years.
Our findings reveal that lifestyle modifications could be a risk factor for breast cancer. Consuming palm oil more than 2 times per week was significantly associated with breast cancer occurrence. This finding is specific to our setting because of the frequency of consumption of the product. Palm oil is a food substance which is financially very accessible and rich in unsaturated fatty acids. It is rich in beta carotene, a precursor for vitamin A and E, and known to have anti-carcinogenic properties. However, during cooking, the oil is usually heated and bleached thereby converting the unsaturated fatty acids are converted to saturated fatty acids. Beta carotene is thus destroyed eliminating its anti-carcinogenic properties hence our findings. Bougnoux et al. in 2010 reported the involvement of saturated fatty acids in promoting breast cancer carcinogenesis [28] . Contrary to our findings however, Arab et al. [29] in 2016 disputed the involvement of conjugated fatty acids in breast cancer carcinogenesis based on their systematic review of studies concerning the issue.
Red meat consumption > 3 times weekly is an independent risk factor for breast cancer. This result is consistent with that of Farvid et al. in 2014 in the United States who found that eating red meat more than 5 times a week significantly increased breast cancer risk in contrast to those who consumed meat less than five times weekly [30] . Beef consumption in our setting is usually done after cooking at high temperatures and often following coal grilling. Steck et al. had described that beef cooked this way increased breast cancer risk [31] . The substances deemed carcinogenic are usually expressed when meat is heated to high temperatures [32] .
Although to varying degrees, some reproduction-related factors have been identified as risk factors for breast cancer. Some authors had reported a decreased risk of breast cancer with increasing parity [33] - [35] . Nulliparity in our study was found to be a risk factor. However, w with co-existing factors, parity appears to be a factor of confusion for breast cancer in our setting. This is probably because breast cancer is increasingly occuring in younger women [7] who are often nulliparous. Multiparity was not identified as a risk factor. Palmer et al. found out that in breastfeeding women, parity-related risks were eliminated hence implying that breastfeeding conferred protection against the disease [34] . This is especially because grand multiparae were women who breastfed their babies longest.
We could not find any correlations between breastfeeding and breast cancer. This finding is coherent with findings by Magnusson et al. [33] . However in breastfeeding women, duration of breastfeeding > 96 months was found to be a risk factor contrary to descriptions in literature where breastfeeding is a protective factor [16] [35] . Ortiz Mendoza et al. in Mexico found that women younger than 50 who have breast cancer had short lactation times at less than 12 months [36] .
Early onset of puberty was not identified as a risk factor in our setting. Literature describes that prolonged exposure to estrogens including precocious puberty, increases the risk of breast cancer [20] especially because early maternity constitutes a significant protective factor [35] .
Menopause was not found to be a risk factor in this study because breast cancer occurs in our environment at a young age, before menopause [7] . This correlates with the findings of Magnusson et al. who found no association between menopausal symptoms and breast-cancer risk [33] .
Danaei G. et al. in 2008 found that alcohol consumption and tobacco smoking play a significant role in carcinogenesis [16] . We did not find a similar result because tobacco consumption is not common amongst women in our context. This is also true for alcohol consumption which appears to be marginal amongst women in our setting. Furthermore, the relationship between alcohol and breast cancer appears to be related to the amount of alcohol consumed [37] [38] . In a 2014 study in 3 African countries, Qian et al. reported a significantly higher risk of breast cancer among women consuming alcohol at least once a day for at least six months [39] .
A family history of breast cancer or other type of cancer was not found to be a risk factor in our study. According to the American Cancer Society [40] , 50% of breast cancers occur in women with no known risk factors. However, when there is a positive family history, the closer the relationship the greater the risk [41] .
Our results do have their limitations though. Due to our limited financial resources, we were unable to perform a mammogram to rule out the presence of cancerous lesions in the control group. Furthermore, an information bias is possible in both groups, as far as recall on food habits and women’s lifestyles may be concerned.
5. Conclusion
We found a wide variety of risk factors. Age > 50 years and beef consumption were independent risk factors for breast cancer identified in this study. These factors should be considered when carrying out primary prevention of breast cancer in our context.
Acknowledgements
We acknowledge the administrative staff of the Yaoundé General Hospital and the Yaoundé Gyneco-Obstetric and Pediatric Hospital for facilitating the implementation of this study as well as the personnel of these hospitals for their support during the collection of data. Apart from the personal contribution of each author, the study was not funded.
Conflict of Interest
The authors declare no conflict of interests
Authors’ Contributions
Essiben F., Foumane P., Signing Soh P. conceived the study, participated in the study design, data collection and drafting the manuscript. Ngo Um Meka E. has been involved in analysis and interpretation of data. Eyongoben O. participated in data collection and review of the article. Dohbit proof-read the manuscript. Mboudou T. E. supervised the study. All authors have read and approved the final manuscript.