Risk Factors of Breast Cancer in the Eastern Mediterranean Region : A Systematic Review and Meta-Analysis

BACKGROUND
Identifying risk factors of breast cancer is a key point for preventive strategies to reduce the incidence. The aim of current study was to determine most important risk factors for breast cancer in the Eastern Mediterranean Region (EMR) using a systematic review.


MATERIALS AND METHODS
PubMed, Scopus, Web of Science till August 24, 2012 and the reference lists of all included studies were searched. Analytic studies which had reported odds ratios (OR), relative risk (RR) or required data to calculate them were included. A total of 343 studies were critically appraised and finally 30 studies were meta-analyzed. Heterogeneity between the studies was assessed by I2 and Cochran's Q. Egger's test was used to assess publication bias.


RESULTS
Twenty five case- control studies, one nested case-control and four cohort studies were included. The largest ORs were obtained for history of no live birth (2.25; 95%CI: 1.58-3.18), body mass index (BMI) more than 30 (2.21; 95%CI: 1.71-2.36), age at first pregnancy more than 30 years old (1.52; 95%CI: 1.30-1.77) and meat consumption more than three times per week (1.39; 95%CI: 1.03-1.87). The other important predictors were higher education and smoking as risk factors, physical activity and ovulatory stimulating medication as protective factors.


CONCLUSIONS
The most important predictors of breast cancer in EMR were history of no live birth, BMI more than 30, age at first pregnancy more than 30 years old, physical inactivity and smoking. Almost all these risk factors are consistent with known risk factors for this cancer in other parts of the world.


Introduction
Breast cancer is the most common malignancy of women in the world (Minatoya et al., 2013).This cancer accounts about one fifth of all female malignancy.Breast cancer is the leading cause of death in high income countries and second leading cause in low and middle income countries (Jemal et al., 2011).This cancer is increasing in many developing countries (Yang et al., 2005) The known most important determinant factors for breast cancer included: age, family history, genetics, personal history of breast cancer, radiation to chest/ face before age 30, race/ethnicity, being overweight, pregnancy/breast feeding and menstrual history, using hormone replacement therapy (HRT), drinking alcohol, having dense breast, lack of exercise, and smoking (Das et al., 2012;Wafa et al., 2014).
The World Health Organization's (WHO) Eastern Mediterranean Region (EMR) covers a population of over 500 million, spread over a wide area of relative cultural and geographical similarity extending from Morocco to
The population, cultures, religious, habits, and knowledge of women in EMR is different from the women lived in developed countries (Habibzadeh, 2012).Majority of women in this region knows little about screening methods of breast cancer.Differences in cultures, habits, and ethnicity affect on differences in marital age, pregnancy features, age of menarche, and other important risk factors of breast cancer in this region (Habibzadeh, 2012).In many countries (such as Iran, Tunisia, Pakistan) of this region the breast cancer diagnosis age is younger than other parts of the world (Shaukat, 2013;Bidgoli and Azarshab, 2014).Also, the stage of diagnosed breast cancer in some countries in this area is highest than the other regions which can be due to lack of knowledge about breast cancer screening (Harirchi, 2012).
Knowing about risk factors of breast cancer in EMR can help policy makers to plan preventive strategies to reduce the incidence of it.Findings of systematic reviews on risk factors of breast cancer in developed countries mostly cannot be useful and generalized to women in EMR.The aim of current study is to determine most important breast cancer risk factors in EMR by reviewing analytic studies by a systematic review and meta-analysis.

Data sources
We conducted a systematic search of all studies which evaluated breast cancer risk factors (except genetics) in EMR using PubMed database, Scopus, EMbase, and Web of Science up to August 24, 2012.The last updated search was done on July 21, 2014.The reference lists of all included studies were searched for further evidences too.

Search strategy
The search strategy was according problem, intervention, comparison, and outcome (PICO) characteristics: The women of EMR countries, risk factor and breast cancer.Table one shows the key words of our search based on medical subjects heading (Mesh).

Eligible studies
We included analytic studies if they had reported Odds Ratio (OR), Relative Risk (RR) or required data to calculate them with 95% confidence interval (CI).The search was restricted to English papers that published as full text.There was no age limitation.We excluded studies if they were reviews, letters, case reports, case series, cross sectionals and the studies that did not report obviously the population sources or the available data to calculate the OR and RR.Also any study on men was excluded too. Figure one depicts the selection process of studies included in the systematic review and meta-analysis.

Quality evaluation of articles and data extraction
Two independent investigators (NN and SKR) conducted data extraction using pre-specified inclusion and exclusion criteria.Discrepancies were resolved by discussion between reviewers and reaching to consensus and if required, senior investigator (MN) qualified the studies.The Newcastle-Ottawa scale as a tool for assessing case-control and cohort studies (Wells, 2000) was used to evaluate the quality of studies.Each study is judged on eight items, categorized into three groups: the selection of the study groups; the comparability of the groups; and the ascertainment of either the exposure or outcome of interest for case-control or cohort studies respectively.
All of included studies (100%) fulfilled the key questions.Out of 30, just four studies fulfilled highest score (8 score out of 8).Finally, necessary information was extracted from each eligible study.
The extracted information included: first author name, the origin country, date of study, date of publication, sampled population, sample size, number of cases and controls, study design, reproductive, lifestyle, and sociodemographic risk factors of breast cancer.Selected life style factors contained body mass index, physical activity, smoking, alcohol consumption and dietary habits (especially meat consumption).The reproductive risk factors included age at first pregnancy, age at first menstruation, history of no live birth, induced abortion and nulliparity.Years of education was considered as a socio demographic risk factor.

Statistical analysis and synthesis of results
The OR with 95%CI was principle summary measure and if it was not presented in the individual study, OR with 95%CI was calculated by LaMorte-stat tool (Wayne and LaMorte, 2006) The reference cutoff for included risk factors were as follow: Body mass index between 25-30 (over weight) and more than 30 (obese), physical activity more than 5.5 hour/week/year, positive history of smoking (not considering pack/year), meat consumption more than three times per week, alcohol consumption (not considering standard unit of alcohol consumption per day), age at first menstruation less than 12 years old and age at first pregnancy more than 30 years old.
Heterogeneity between the studies was defined as p<0.1 or I² Index >50% using Cochran's Q and I 2 index respectively.I² Index was calculated as 100 × (Cochrane Q -df)/Cochrane Q.In case of heterogeneity between the studies random-effect model was used in meta-analysis.
If not, the fixed-effect model was done.Egger's linear regression test was used to indicate the publication bias (P<0.05set as significant level).Meta-analysis was performed by the Stata statistical software (version 10, Stata Corporation, College Station, Texas, USA).
This study was approved by research and ethical committees of Tehran University of Medical Sciences.

Assessing publication bias
We used Egger's test to assess publication bias.The results of this test showed there were no publication biases for majority of reported risk factors (p>0.05).The results of Egger's test showed publication bias for history of using ovarian stimulation medication, age at first pregnancy, history of no live birth, education and age of menarche.
Three studies (Shaham et al., 2006;Saleh et al., 2008;Abdulbari et al., 2010) had evaluated the association between meat consumption more than three times per week and breast cancer in EMR countries.The pooled OR for meat consumption more than three times per week versus less than three times was 1.39 (95%CI: 1.03-1.87),with no heterogeneity (I² =0%; p=0.99).
We just found two studies in which alcohol as a risk factor of breast cancer was assessed in EMR.Therefore, we did not calculate pooled odds ratio for alcohol consumption.

Discussion
In this systematic review and meta-analysis we showed that obesity (BMI more than 30), over weight (BMI=25-30), meat consumption more than three times in a week, smoking, physical activity less than 5.5 h/w, first pregnancy more than 30 years old, not having live birth and education more than 12 years are the risk factors for breast cancer in women who live in EMRO.These findings are also supported by individual studies and other systematic reviews (Alothaimeen et al., 2004;Yavari et al., 2005;Saleh et al., 2008;Awatef et al., 2010).
Life style is one of the noteworthy risk factor for noncommunicable diseases such as cancers, cardiovascular diseases and diabetes (Javed et al., 2011).Life style contains four major subgroups: diet, physical activity, tobacco abuse and risky behaviours.Changing of the Mediterranean diet to westernized diet in the EMR may be a basic breast cancer risk factor in EMR and this situation is recognised as transient period in epidemiological studies (Henquin et al., 1994;Ilic et al., 2013).Body mass index is noticeably associated with hormone receptor status and higher risk of oestrogen receptor (ER) positive breast cancer (Ghiasvand et al., 2011).As a limitation of studies on risk factor of breast cancer in EMR, obesity and overweight was not considered in post-menopausal and premenopausal separately.In the current systematic review we considered just meat consumption and did not evaluate other aspects of diet such as pattern of food preparation and storing the food in the individual studies.
Moving to industries world was limited the physical activity in people of developing countries.This situation gets worst by inadequate physical, environmental and political support and family restrictions for women in EMR.Several biologic mechanisms have supported the protective effect of physical activity on breast cancer such as endogenous sex steroid hormone production, effect on immune and antioxidant system (Abdulbari et al., 2010;Awatef et al., 2010;Kruk, 2014).
Increasing the proportion of smoker women in developing countries is an important point that was seen in developed countries in past decades (Bidgoli et al., 2011;Ilic et al., 2013).About 70% of deaths attributed to tobacco use are likely to happen in developing countries (Maziak et al., 2004).Association of smoking and breast cancer is shown in current study also has supported by International Agency for Research on Cancer (IARC) evaluation (IARC 2004).There were no evaluation of some smoking related variables such as current/ex-smoker, pack-year of smoking, alcohol confounding effect, and passive/active smoker in studies carried in EMR on breast cancer and smoking (Henquin et al., 1994;Bidgoli et al., 2011;Javed et al., 2011).Evaluation of alcohol consumption and breast cancer was restricted too.Also, there was not any consensus between founded (Saadat et al., 2008;Kloong et al., 2011).Therefore, it is necessary to do studies considering this issue in EMR.However systematic reviews carried out mostly in developed and developing countries have found a positive association between alcohol consumption and breast cancer (Ellison et al., 2001;Collaborative group on Hormonal Factors in Breast cancer, 2002;Key et al., 2006;Ying et al., 2013) explanation for lack of studies on alcohol consumption and breast cancer in EMR could be religious beliefs and culture of people in this region that could be a strong source of publication bias which is seen in other eastern studies too (Key et al., 2006).Our findings about reproductive factors and breast cancer are supported by individual studies (Das et al., 2012;Shaukat et al., 2013;Toleutay et al., 2013).But, the relation between years of education and age of first pregnancy should be considered by caution and confirmed in future studies.
The strength point of current study was assessing risk factors of most common cancer in women in a specified region (EMR).Systematic review of observational studies, especially on risk factors of non-communicable diseases, could help policy makers to plan the best approach for prevention of these common modern diseases.This study was unique according this issue.We tried to account most common risk factors of breast cancer in current study.
This systematic review has several limitations.Some limitations were about included individual studies.Majority of included studies were hospital based.This point could limit the generalizability of the findings to all women.Although, studies on risk factors of cancers usually carried out on hospital cases and this limitation is inevitable.The other limitation was about the details of studied risk factors in EMR investigations.Majority of these studies have not included some special aspects of life style risk factors such as special pattern of food preparing (for example, smoked food) or other type of tobacco consumption except cigarette smoking ( for example, Bidi, chewing tobacco, betel nut, water pipe and snuff).Future studies with focusing on these aspects of risk factors recommended.
Finally, the most important limitation was about language of included studies.We considered language restriction to English; therefore, some relevant studies may be missed.Although, recently most investigators are interested in publishing their works in English journals that most of them indexed in searched databases in current study.However, this limitation could be account.
Figure 1.The Selection Process for the Studies Included in the Systematic Review-Meta Analysis Figure 2. Cumulative Forest Plot of Breast Cancer and Body Mass Index (25-30) Odds Ratio .05.1 1 2 3 4 5 10

Figure 3 .
Figure 3. Cumulative Forest Plot of Breast Cancer and Body Mass Index More than 30

Figure
Figure 4. Cumulative Forest Plot of Breast Cancer and Physical Activity More than 5.5 h/w