A Longitudinal Study of the Metabolic Syndrome and Risk of Postmenopausal Breast Cancer

The metabolic syndrome, or insulin resistance syndrome, which is associated with increased risk of diabetes and heart disease (1-3), has recently been suggested to play a role in breast carcinogenesis (4-6). This syndrome is characterized by abdominal obesity, high blood glucose levels, impaired glucose tolerance, dyslipidemia, and hypertension – conditions often associated with obesity, poor diet, and lack of physical activity (4). The prevalence of the metabolic syndrome has increased in the United States in recent years (7, 8), and one estimate indicates that roughly 47 million Americans currently have the syndrome (7). The metabolic syndrome could influence the risk of breast cancer through changes in a number of interrelated hormonal pathways, including those involving insulin, estrogen, cytokines, and growth factors (4, 6).

Numerous studies have examined the association of individual components of the metabolic syndrome with breast cancer risk, but their results have been inconclusive (9-41). To date, however, no study has assessed the metabolic syndrome per se in relation to breast cancer risk, and it is conceivable that the syndrome may show a stronger association with risk than its individual components. Furthermore, previous studies have had only baseline measurements of specific components of the syndrome. We therefore evaluated the association of the metabolic syndrome with the risk of postmenopausal breast cancer among women in the 6% sample of subjects in the Women's Health Initiative clinical trial and the 1% sample of women in the observational study who had repeated measurements of the components of the syndrome during follow-up, thereby allowing assessment of the association longitudinally.

The Women's Health Initiative is a large, prospective, multicenter study of factors affecting the health of postmenopausal women. It includes an observational study (N = 93,676) and three clinical trials (N = 68,132) of hormone therapy, dietary modification, and calcium plus vitamin D supplementation (42). Women were recruited at 40 clinical centers throughout the United States, largely via direct mailings, and were eligible to participate if they were postmenopausal, ages 50 to 79, likely to reside in their current residence for at least 3 y, and had provided written informed consent. Enrollment took place from October 1, 1993 to December 31, 1998. The clinical trials had a number of additional eligibility requirements. In general, eligible women were enrolled in the clinical trial component. Women who did not wish to be randomly assigned to an intervention or who were ineligible for the clinical trial component were invited to participate in the observational study.

The present analysis is based on a 6% random sample of women in the clinical trials (n = 4,396) who provided fasting blood samples at baseline and at years 1, 3, and 6 of follow-up, and a 1% sample of women in the observational study (n = 1,054) who provided fasting blood samples at baseline and in year 3 (43). Blood samples were analyzed for glucose, triglycerides, and high-density lipoprotein cholesterol (HDL-C; ref. 43). In addition, waist circumference, and systolic and diastolic blood pressure were measured by study staff using a standardized protocol at clinical visits. The 6% random sample was stratified by age, clinical center, and hysterectomy status, with oversampling of minority groups to increase the numbers of Black, Hispanic, and Asian-Pacific women.

In the clinical trial cancer outcomes were ascertained through self-administered questionnaires completed every 6 mo, and then confirmed by a centralized review of pathology reports, discharge summaries, operative and radiology reports, and tumor registry abstracts. In the observational study cancer outcomes were ascertained annually.

Fasting blood samples were collected with minimal stasis and maintained at 4°C until plasma/serum was separated. Plasma/serum aliquots were then frozen at −70°C and sent on dry ice to the central repository (Fisher BioServices), where storage at −70°C was maintained. Glucose was measured using the hexokinase method on the Hitachi 747 (Boehringer Mannheim Diagnostics; refs. 44, 45). An ongoing monthly quality assurance program is maintained with the Diabetes Diagnostic Laboratory at the University of Missouri. Monthly interassay coefficients of variation (CV) were <2% for mean concentrations of 84 and 301 mg/dL. Total cholesterol and triglycerides were analyzed by enzymatic methods on the Hitachi 747 analyzer (46). HDL-C was isolated using heparin manganese chloride (47). CVs for total cholesterol, triglycerides, and HDL-C were all ≤2%.

Waist circumference at the natural waist or narrowest part of the torso was measured to the nearest 0.1 cm. On each visit, two blood pressure measurements were obtained ≥30 s apart, and the average of the two measurements was used in the analysis. Values for waist circumference and blood pressure in the years corresponding to the blood analytes were used in the analysis.

We used the definition of the metabolic syndrome proposed by the Adult Treatment Panel III (ATP III) of the National Cholesterol Education Program (48, 49). An indicator variable was created for the presence of the metabolic syndrome (yes/no), defined as having ≥3 of the following characteristics: waist circumference ≥88 cm, fasting glucose ≥100 mg/dL, fasting HDL-C <50 mg/dL, fasting triglycerides ≥150 mg/dL, and blood pressure ≥130/85 mmHg.

Diabetes has been studied as a risk factor for breast cancer (4). We were interested in whether the criteria for the metabolic syndrome, including insulin resistance, predict breast cancer risk before a clinical diagnosis of diabetes. Therefore, women who reported taking diabetes medication at baseline or having a baseline fasting serum glucose level of ≥126 mg/dL were excluded from the analysis (25 breast cancer cases and 537 noncases). The results were not changed when women with diabetes were included in the analysis.

Cox proportional hazards models were used to estimate hazard ratios (HR) and 95% confidence intervals (95% CI) for the associations between the presence of the metabolic syndrome and its components and the risk of breast cancer, with duration of follow-up (days) as the time scale. For these analyses, study participants were considered to be at risk from their date of enrollment until the date of diagnosis of their breast cancer, termination of follow-up (September 12, 2005), loss to follow-up, withdrawal from the study, or death, whichever occurred first. Event times of participants who had not developed breast cancer by the end of follow-up, who had died, or who withdrew from the study before the end of follow-up, were censored.

In the first stage of the analysis we estimated the risk of breast cancer in association with the presence of the metabolic syndrome or its individual components at baseline. Presence of the metabolic syndrome was defined as having ≥3 of the individual components relative to having ≤2. In addition, the individual components of the metabolic syndrome were divided into three groups, using the ATP III cutoffs for the highest category (lowest for HDL-C), and the median for the remainder of the distribution. We also examined the association of “degree of metabolic syndrome” with risk, using each individual's score, ranging from 0 (reference group) to 5 and obtained by summing scores (1 = present, 0 = absent) for each of the individual components. Tests for trend were perform by assigning the median value to each category and modeling this variable as a continuous variable. Established breast cancer risk factors and potential confounding variables, obtained at baseline, were included in multivariable models as follows: age (continuous), education (less than high school, high school grad/some college, college graduate, postcollege), ethnicity (white, black, other), body mass index (continuous), oral contraceptive use (ever/never), postmenopausal hormone therapy (ever/never), age at menarche (continuous), age at first birth (<20, 20-29, ≥30, missing), age at menopause (<50, ≥50, missing), alcohol (servings per week - continuous), family history of breast cancer (yes/no), history of breast biopsy (ever, never), physical activity (metabolic equivalent tasks per week - continuous), energy intake (continuous), smoking status (never, former, current smoker), and randomization status (for women in the clinical trial) in the hormone therapy, calcium plus vitamin D, and dietary modification trials. All P values were two-sided.

Analyses were performed on all breast cancers (invasive and in situ) and on invasive cancers only. In addition, two sensitivity analyses were carried out: restricting the analysis to women who did not participate in any Women's Health Initiative clinical trial intervention, and excluding cases diagnosed during the first two years of follow-up. The results were not altered in either analysis, and we present the results for the total study population.

In the second stage of the analysis, the repeated measurements of the different components of the metabolic syndrome were analyzed by modeling them as time-dependent covariates in the Cox proportional hazards model (50). With this approach, we evaluated the predictive value of the most recent measurement; measurements obtained in the intervals 1 to 3 y, 2 to 4 y, and 3 to 5 y before the date of diagnosis of breast cancer; and the average of all available measurements. In all time-dependent analyses, measurements that were obtained within 1 y of diagnosis were excluded from all analyses, because these values may have been influenced by the presence of subclinical disease.

During a median follow-up of 8 years, a total of 165 breast cancer cases (131 invasive and 34 in situ) were ascertained. Of the total breast cancer cases, 99 were not randomized to any of the clinical trial intervention groups or were in the observational study, whereas the remaining 66 were randomized to at least one intervention group. The corresponding numbers for noncases were 2,620 and 2,103.

At baseline, cases and noncases were similar with respect to age and anthropometric and reproductive variables (Table 1). Cases had significantly lower levels of physical activity and were more likely to be non-Hispanic white compared with noncases.

The presence of the metabolic syndrome at baseline was not associated with altered risk of total breast cancer (invasive plus in situ) or of invasive breast cancer alone, with multivariable HR of 1.12 (95% CI, 0.78-1.62) and 1.19 (95% CI, 0.79-1.79), respectively (Table 2). Of the individual components of the metabolic syndrome measured at baseline, diastolic blood pressure was associated with increased risk of total breast cancer (multivariable HR for highest versus lowest tertile, 1.55; 95% CI, 1.02-2.36) but not of invasive breast cancer (HR, 1.43; 95% CI, 0.90-2.29). None of the other components was associated with altered risk of the end points of interest. When the number of components of the syndrome was treated as an ordinal variable (relative to a reference group of 0), women with ≥4 components had a HR of 1.56 (95% CI, 0.86-2.85). These results were not affected by restriction of the sample to women who were not randomized to any of the clinical trial intervention groups or were in the observational study, or by exclusion of cases diagnosed within the first two years of follow-up (data not shown).

In the time-dependent covariate analyses, presence of the metabolic syndrome 3 to 5 years prior to diagnosis was associated with increased risk of total breast cancer and of invasive breast cancer, with HR of 1.84 (95% CI, 1.12-3.01) and 1.77 (95% CI, 1.01-3.12), respectively (Table 3). A borderline positive association was also seen for metabolic syndrome 2 to 4 years prior to diagnosis but not for presence of the syndrome 1 to 3 years prior to diagnosis. When presence of the metabolic syndrome (≥3 components satisfying the ATP III cutoff values) was based on the average of all individual components, the HR for all breast cancer was 1.57 (95% CI, 1.09-2.26), and that for invasive breast cancer was 1.59 (95% CI, 1.06-2.41).

Because the time-dependent analysis suggested that metabolic syndrome status measured earlier in time is more predictive of breast cancer than more recent measurements, we also evaluated the association of cumulative exposure to the metabolic syndrome with breast cancer. For a subject at risk at time t, cumulative exposure was estimated as the number of visits up to time t at which the subject was determined to have the syndrome and was analyzed as a time-dependent covariate. The HR associated with ≥2 prior diagnoses of the syndrome versus <2 diagnoses was 1.59 (95% CI, 1.05-2.41) for total breast cancer and 1.68 (95% CI, 1.05-2.71) for invasive breast cancer only.

Of the individual components of the metabolic syndrome, positive associations were seen for serum glucose (average and 1-3 years prior to diagnosis) with all breast cancer, serum triglycerides (3-5 years prior to diagnosis) with total breast cancer and invasive breast cancer, and average diastolic blood pressure with total breast cancer and invasive breast cancer (Table 3).

In this longitudinal study, presence of the metabolic syndrome at baseline was not associated with breast cancer risk. Furthermore, of the individual components of the syndrome, only baseline diastolic blood pressure showed any suggestion of an association. However, in some time-dependent analyses (particularly, presence of the syndrome 3-5 years prior to diagnosis and presence of the syndrome based on the average value of its components), presence of the metabolic syndrome showed a positive association with breast cancer. In addition, after adjustment for other components of the metabolic syndrome, serum glucose and triglycerides and diastolic blood pressure were associated with increased risk in the time-dependent analyses. Although numerous studies have examined individual components of the metabolic syndrome in relation to breast cancer risk (9-41), no study to date has assessed the association of the metabolic syndrome per se with breast cancer.

The results of previous studies that reported on the individual components of the metabolic syndrome have been inconsistent. Increased central adiposity (as measured by waist circumference and waist-hip ratio) has been associated with increased risk of postmenopausal breast cancer in some (9-13) but not all cohort studies (15-17; see ref. 18 for review). An inverse association between HDL-C and breast cancer risk has been reported in several case-control studies (19, 20, 23, 24) and in two cohort studies (25, 26), but not in other studies (27, 28). A nested case-control study (30) found that HDL-C was inversely associated with breast cancer among premenopausal women but positively associated with disease in postmenopausal women. A number of studies have found that serum triglyceride levels were positively associated with breast cancer risk (19-22), whereas other studies have found no association (26, 27). In several studies, hypertension has been linked to higher breast cancer risk (31-33); however, two of these studies (32, 33) were case-control studies that relied on self-reports of history of hypertension or treatment for hypertension. Among cohort studies that measured blood pressure at baseline (31, 34-36), only one study (31) reported a positive association with systolic blood pressure. Finally, of six studies (34, 37-41) that examined the association of fasting blood glucose level with breast cancer risk, three studies (37, 39, 41) showed evidence of a positive association with breast cancer, whereas the results of the remaining studies were null. Of the two cohort studies indicating a positive association, one (37) showed a significant association in premenopausal women, whereas the other (41) reported a significant association in all women, the largest increase in risk being among women of age ≥65.

Hyperinsulinemia may provide the unifying mechanism by which the metabolic syndrome might be associated with increased breast cancer risk (4, 16, 51). Insulin has mitogenic activity in addition to metabolic effects and can promote cell proliferation in normal mammary epithelial cells and breast cancer cell lines (52, 53). Insulin may also contribute to tumor promotion by up-regulating the secretion of ovarian hormones (54, 55). In theory, high glucose levels could additionally increase the risk of postmenopausal breast cancer by conferring a selective growth advantage on malignant cells (56), as high rates of glucose uptake and glycolysis are a common feature of malignant growth (57). In addition to its association with insulin resistance and the metabolic syndrome, abdominal obesity is associated with the release of nonesterified fatty acids from adipose tissue and their accumulation in muscle and liver, leading to dyslipidemia (16). Furthermore, adipose tissue in obese individuals exhibits abnormalities in the production of several adipokines, including increased production of inflammatory cytokines and plasminogen activator inhibitor-1 and reduced production of adiponectin, which may affect insulin resistance (16). Thus, in addition to the effects of insulin and glucose, low-grade chronic inflammatory effects associated with the metabolic syndrome may be relevant to breast carcinogenesis (58). Although hypertension associated with the metabolic syndrome seems to be secondary to the effects of insulin resistance and compensatory hyperinsulinemia on the sympathoadrenal system (59), breast cancer and hypertension may share common pathways involving inflammation and hormone synthesis and metabolism (4, 60, 61).

Our results provide some support for an association between the metabolic syndrome and breast cancer risk, but need to be interpreted with caution. Certain time-dependent analyses showed significant associations of both the metabolic syndrome per se and glucose, triglycerides, and diastolic blood pressure with breast cancer risk. However, only baseline diastolic blood pressure showed any association with risk. The time-dependent measures provide a more reliable estimate of exposure over the time period relevant to the development of breast cancer and, therefore, perhaps should carry greater weight than those derived from the baseline measures alone. However, given the many comparisons performed and the limited sample size, some of these differences could also be due to chance. In addition, waist circumference of ≥88 cm relative to waist circumference of <79 cm was not associated with breast cancer risk in any of the analyses, whereas excess abdominal adiposity is an important determinant of insulin resistance (22) and is thought to play a key role in the metabolic syndrome (22). Our findings of positive associations of breast cancer with diastolic blood pressure (in both the baseline and the repeated measures analyses) and with serum glucose and serum triglycerides (in the time-dependent analyses) have some support in the literature but require further elucidation.

Given the limited number of cases in this study, it would be premature to draw definitive conclusions about whether the associations with individual components or with the composite metabolic syndrome are more informative in our data or how convincing a biological rationale exists for the observed associations with individual components. Larger studies with repeated measurements are needed to determine whether the metabolic syndrome per se and/or particular components show a consistent association with breast cancer risk.

The strengths of the present study include its prospective nature, the availability of detailed information on breast cancer risk factors and other potential confounders, and availability of repeated measurements of components of the metabolic syndrome. Limitations include the relatively small sample size and the lack of information on circulating estrogen levels.

In conclusion, the present study, which is the first to assess the association of the metabolic syndrome per se with breast cancer risk, provides some evidence of a modest positive association of postmenopausal breast cancer with the metabolic syndrome per se (in repeated measures analyses), diastolic blood pressure (baseline and repeated measures analyses), and serum glucose and triglycerides (repeated measures analyses).

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.