Breast cancer in the south of Iran.

To the Editor: Breast cancer arises from a multifactorial process. Recent attention has focused on genetic predispositions to breast cancer.1,2 Furthermore, the effect of reproductive factors strongly supports a hormonal role in its etiology.3,4 Earlier age at menarche5,6,7 and later age at first full term pregnancy6,7,8,9 have shown a significant increase in the risk of the disease. While numerous studies have been conducted in western counties to assess the epidemiology of breast cancer, there have been few studies on Middle East populations. Although breast cancer in Iranian women is the most common form of cancer, 10 few epidemiological studies have been conducted on its risk factors in this country. Age-adjusted incidence of the disease in Iran is estimated to be 22.4 per 100 000,10 and epidemiological studies have found a lower age in Iranian patients compared with their western counterparts,11 and a moderately rapid increase in incidence rate of the disease in recent years.10 Now, the question arises as to whether or not breast cancer in the south of Iran is influenced by some specific risk factors established for high or moderate incidence areas. This case-control study was conducted from April 2002 to March 2004 in Bandar Abbas, Hormozgan, Iran to investigate this subject.

Eligible cases were incident (diagnosed within 2 years before the interview) breast cancer patients living in the region, and were entered into the study if they had a confirmed pathological primary breast cancer diagnosis from the pathology department of Bandar Abbas Shahid Mohammadi Hospital, which is the leading university-based hospital in the region. For each case three age-matched (±3 years) women were recruited from patients with no history of breast problems or neoplastic diseases who attended outpatient ophthalmology or dermatology clinics in the same hospital. After receiving informed consent, a structured questionnaire was completed at the time of recruitment. All interviews were carried out by two interviewers who had been thoroughly familiarized with the study protocol.

Odds ratios from univariate logistic regression were used to estimate the relative risk of breast cancer associated with the various factors and their predictive effects. Based on the univariate analysis, the odds ratios were adjusted for potential confounding variables; 95% confidence intervals (CIs) for each of the odds ratios were calculated. A forward multivariate logistic regression model was used for significant associated risk factors and P<0.05 was considered statistically significant.

Of 173 women with breast cancer who were newly diagnosed, 168 patients entered the study as cases and 504 women were selected as controls. There were no significant differences between cases and controls with regard to parity, history of breast feeding, history of induced or spontaneous abortion, oral contraceptive use, menopausal status, age at menopause, HRT use, history of previous benign breast disease and tobacco exposure. Breast cancer history in first-degree relatives was a significant risk factor (OR 7.09, 95%CI 4.06–12.26). Women with a younger age at menarche (<13 years old) were found to be at higher risk for breast cancer than women with older age of menarche (OR 4.00, 95%CI 1.82–9.84).The results show that never-married women demonstrated a higher risk of breast cancer than the others (OR 2.69, 95%CI 1.38–7.12). Breast cancer risk was significantly greater in women whose first full-term pregnancy was at age 30 years or more in comparison with the others with a first full-term pregnancy at a lower age (OR 7.79, 95%CI 4.25–9.12). Furthermore, it was shown that more than five full-term pregnancies would be expected to correlate with an increase in the risk of breast cancer (χ2 = 111.12, P<0.05).

In the forward multivariate logistic regression analysis, in addition to those factors which were significantly associated with breast cancer, parity and breast feeding, factors with relatively high but not statistically significant odds ratios, were included in the model. The final model revealed that in addition to those factors, which were significant in univariate logistic regression analysis, a history of breast feeding is a significant factor in decreasing risk of breast cancer (OR 0.68, 95%CI 0.12–0.97) but nulliparity is still not significant.

Although the results cannot be generalized, the findings suggest that the associations between some known risk factors for breast cancer may differ in the south of Iran as compared with other populations. Intensive study of developing countries might reveal other important risk factors in these populations.