Abstract
Purpose
We assessed the association between diabetes and breast cancer and whether physical activity modified the effect of diabetes on breast cancer in Hispanic women.
Methods
We used data from a case-control study of breast cancer among Hispanic women aged 30 to 79 conducted between 2003 and 2008 on the Texas-Mexico border. In-person interviews were completed with 190 incident breast cancer cases ascertained through surgeons and oncologists, and 979 controls who were designated as both high-risk (n=511) and low-risk (N=468) for breast cancer (with respective response rates of 97%, 83% and 74%).
Results
After adjustment for menopausal status and mammography screening, there was no effect of diabetes on breast cancer risk (high-risk control group odds ratio [OR] 1.02, 95% confidence interval [CI] 0.71–1.48; low-risk control group OR 0.87, 0.58–1.30). Women who had a diabetes history and did not exercise were at no risk of breast cancer (OR 0.96, 95% CI 0.63–1.48) or a slightly reduced breast cancer risk (low-risk control group OR 0.72, 95% CI 0.46–1.15) depending on the control group used, while women with diabetes who did exercise had significantly reduced breast cancer risk (OR 0.41, 95% CI 0.21–0.83) regardless of the control group used (high-risk control group p-value for interaction=0.013, low-risk control group p-value for interaction 0.183).
Conclusions
Should other studies confirm our results, physical activity should be explored as a means of reducing breast cancer risk in diabetic women.
Keywords: breast cancer, diabetes, physical activity, case-control study
Introduction
Two recent meta-analyses of the association between diabetes and breast cancer reported summary relative risks of 1.20 (95% CI 1.12–1.28)1 and 1.15 (95% CI 1.12–1.19)2, respectively. Hispanic women, who have extremely high rates of diabetes and yet fairly low rates of breast cancer, were not well-represented in these meta-analyses. Diabetes was not associated with breast cancer risk in the Four Corners Breast Cancer Study, a study of Hispanic and Native American women combined (OR 0.92, 95% CI 0.71–1.20)3. Based on these findings, we hypothesized that we would not find an association between diabetes and breast cancer among Hispanic women in our study. Since physical activity is known to reduce the risk of breast cancer4 and diabetes5, we also hypothesized that physical activity would modify the effect of diabetes on breast cancer.
Patients and Methods
This clinic-based case-control study was conducted in the Lower Rio Grande Valley located at the southern tip of Texas on the Mexico border. Cases diagnosed with primary invasive breast cancer between November 2003 and August 2008 were identified through surgeons and oncologists shortly after diagnosis or treatment. Eligible cases were of Hispanic ethnicity based on self-report, aged 30–79, whose incident primary breast cancer was histologically confirmed and with no history of cancer, other than nonmelanoma skin cancer. A total of 190 breast cancer cases (97.0% of 196 eligible cases) completed a standardized in-person interview. Of potentially eligible cases, 3 refused (1.5%) and 3 were lost to follow-up (1.5%).
Control subjects were randomly selected from women receiving a diagnostic or screening mammogram at the mammography center where the case received her diagnostic mammogram. Two control groups were selected, a high-risk group of women receiving a diagnostic mammogram either due to inconclusive or abnormal results, and a low-risk group of women with no family history of breast cancer, no history of breast biopsy, and negative screening mammograms for the past two years. Two women from each control group were selected per case and frequency matched to the case on age. Eligible controls were Hispanic, aged 30–79 with no history of cancer, other than nonmelanoma skin cancer.
A total of 511 high-risk control subjects (83.0% of 616 eligible high-risk controls) and 468 low-risk control subjects (73.6% of 636 eligible low-risk controls) completed the interview. Of eligible high-risk controls, 61 refused (9.9%) and 44 (7.1%) were lost to follow-up. Of eligible low-risk controls, 127 refused (20.0%) and 41 (6.4 %) were lost to follow-up. For the present analysis we investigated the associations separately by type of control group.
Institutional Review Boards of the University of Texas at Brownsville and the University of Texas Health Science Center at Houston approved this study’s protocol. Trained interviewers conducted in-person interviews with subjects who provided consent. The questionnaire collected information on demographic characteristics, suspected breast cancer risk and protective factors, and medical history including diabetes. Exposures pertained to the period prior to a reference date, the date of diagnosis for the cases and an assigned date for controls comparable to the date of diagnosis for the cases.
With subjects’ permission we abstracted their medical records for information on breast cancer screening, diagnosis and treatment, and diabetes diagnosis and treatment. A 10-hour fasting blood draw was collected prior to treatment. Serum glucose was analyzed using the hexokinase/glucose-6-phosphsate dehydrogenase method. Since associations have been established between insulin resistance, a precursor of diabetes, and breast cancer, we grouped borderline diabetes and diabetes, hereafter termed diabetes. To define diabetes, we used serum glucose (available for 67 cases and 241 controls), followed by medical record information (available for an additional 25 cases and 116 controls), followed by information from the questionnaire (available for the remaining 98 cases and 622 controls). Women were considered diabetic if they had a fasting serum glucose concentration of ≥100 mg/dl, an indication of diabetes in their medical record, or they were told by their doctor they had diabetes. Diabetes was further categorized as occurring during pregnancy, other than pregnancy, or both during and other than pregnancy available from the questionnaire only. If women received both insulin and oral medications to treat diabetes, available from the medical record and then from the questionnaire, they were classified as having used insulin. Information on leisure-time physical activity, defined as having engaged in vigorous or moderate activity at least two hours a week for four months or more a year in the past three years, was available from the questionnaire. Examples of vigorous activity included basketball, jump rope, swimming laps, aerobic dance, running, jogging, bicycling on hills and some types of exercise equipment, while examples of moderate activity included brisk walking, golf, volleyball, bicycling on level ground, softball, dancing and gardening. Body mass index (BMI) at study entry was based on actual measurements of body weight and body height, while BMI at age 15 years was based on the questionnaire.
Statistical analyses were performed in SAS version 9.2. We assessed statistically significant (two-sided, p<0.05) differences between cases and controls for suspected breast cancer risk and protective factors using t-tests and chi-square tests. We used unconditional logistic regression to estimate the relative risk of breast cancer associated with the main effect of diabetes and the joint effects of diabetes and physical activity6. Interaction terms, the product of diabetes and putative effect modifiers (menopausal status and physical activity), were added to logistic regression models and likelihood ratio tests were performed to test for effect modification. Age, educational level, family history of breast cancer, age at menarche, menopausal status, number of full-term pregnancies, age at first pregnancy, breastfeeding history, BMI at study entry, BMI at age 15 years, use of oral contraceptives, use of hormone replacement therapy, alcohol intake (drinking at least one alcoholic beverage a month for 6 months or longer), mammography screening, comorbid conditions including heart disease, hypertension, myocardial infarction and stroke, and physical activity as categorized in Table 1 were evaluated as potential confounders. Missing data were not included in the variable percentages unless they contributed a large portion to the distribution as was the case for BMI at age 15 years. Variables were considered confounders if their addition to the model changed the unadjusted odds ratio by 10 percent or more. We performed a validation study of self-report of diabetes, including borderline diabetes, using both serum glucose and the medical record as gold standards. Sensitivities and specificities and their respective confidence intervals were calculated as measures of validity.
Table 1.
Comparison of cases and controls for suspected breast cancer risk and protective factors, South Texas Women’s Health Project, 2003–2008
| Characteristic | Cases (n=190) | High-risk controls (n=511) | Low-risk controls (n=468) | |||||
|---|---|---|---|---|---|---|---|---|
| N | % | N | % | p-valuea | N | % | p-valueb | |
| Mean age (years) (standard deviation) | 54.3 (11.1) | 49.2 (10.1) | <0.001 | 55.1 (8.4) | 0.371 | |||
| Educational level | ||||||||
| Less than high school | 111 | 58.4 | 261 | 51.2 | 0.160 | 242 | 51.8 | 0.304 |
| High school | 32 | 16.8 | 87 | 17.0 | 93 | 19.9 | ||
| More than high school | 47 | 24.7 | 162 | 31.8 | 132 | 28.3 | ||
| Missing | 0 | 1 | 1 | |||||
| Breast cancer among first-degree relatives | ||||||||
| No | 168 | 89.4 | 455 | 90.5 | 0.667 | 440 | 95.7 | 0.003 |
| Yes | 20 | 10.6 | 48 | 9.5 | 20 | 4.3 | ||
| Missing | 2 | 8 | 8 | |||||
| Age at menarche (years) | ||||||||
| ≤12 | 99 | 52.4 | 250 | 49.1 | 0.070 | 220 | 47.1 | 0.048 |
| 13 | 32 | 16.9 | 127 | 25.0 | 121 | 25.9 | ||
| >13 | 58 | 30.7 | 132 | 25.9 | 126 | 27.0 | ||
| Missing | 1 | 2 | 1 | |||||
| Menopausal status | ||||||||
| Premenopausal | 38 | 20.4 | 203 | 39.8 | <0.001 | 77 | 16.7 | 0.267 |
| Postmenopausal | 148 | 79.6 | 307 | 60.2 | 383 | 83.3 | ||
| Missing | 4 | 1 | 8 | |||||
| Number of full-term pregnancies | ||||||||
| 0 | 13 | 6.8 | 43 | 8.4 | 0.084 | 35 | 7.5 | 0.861 |
| 1–2 | 52 | 27.4 | 166 | 32.5 | 138 | 29.5 | ||
| 3–4 | 80 | 42.1 | 222 | 43.4 | 197 | 42.1 | ||
| ≥5 | 45 | 23.7 | 80 | 15.7 | 98 | 20.9 | ||
| Age at first pregnancy (years) | ||||||||
| <30 | 163 | 92.1 | 434 | 92.7 | 0.781 | 402 | 92.8 | 0.748 |
| ≥30 | 14 | 7.9 | 34 | 7.3 | 31 | 7.2 | ||
| Missing | 13 | 43 | 35 | |||||
| Breastfeeding history | ||||||||
| No | 86 | 45.3 | 187 | 36.7 | 0.036 | 193 | 41.2 | 0.344 |
| Yes | 104 | 54.7 | 324 | 63.4 | 275 | 58.8 | ||
| Body mass index at study entry | ||||||||
| <25 | 21 | 11.3 | 82 | 16.1 | 0.075 | 62 | 13.4 | 0.818 |
| 25–29.9 | 55 | 29.6 | 170 | 33.5 | 137 | 29.7 | ||
| 30–34.9 | 57 | 30.7 | 153 | 30.1 | 145 | 31.4 | ||
| ≥35 | 53 | 28.5 | 103 | 20.3 | 118 | 25.5 | ||
| Missing | 4 | 3 | 6 | |||||
| Body mass index at age 15 years | ||||||||
| <20 | 92 | 48.4 | 225 | 11.0 | 0.591 | 188 | 40.2 | 0.220 |
| 20–25 | 59 | 31.0 | 187 | 36.6 | 176 | 37.6 | ||
| ≥25 | 17 | 9.0 | 42 | 8.2 | 52 | 11.1 | ||
| Missing | 22 | 11.6 | 57 | 11.2 | 52 | 11.1 | ||
| Use of oral contraceptives | ||||||||
| No | 66 | 34.9 | 136 | 26.6 | 0.031 | 132 | 29.3 | 0.093 |
| Yes | 123 | 65.1 | 375 | 73.4 | 335 | 71.7 | ||
| Missing | 1 | 0 | 1 | |||||
| Use of hormone replacement therapy | ||||||||
| No | 130 | 68.4 | 348 | 68.1 | 0.936 | 222 | 47.8 | <0.001 |
| Yes | 60 | 31.6 | 163 | 31.9 | 242 | 52.2 | ||
| Missing | 0 | 0 | 4 | |||||
| Alcohol intake | ||||||||
| No | 155 | 81.6 | 415 | 81.2 | 0.912 | 385 | 82.4 | 0.793 |
| Yes | 35 | 18.4 | 96 | 18.8 | 82 | 17.6 | ||
| Missing | 0 | 0 | 1 | |||||
| Number of mammograms in past 6 years | ||||||||
| 0–1 | 39 | 20.5 | 91 | 17.9 | 0.392 | 6 | 1.3 | <0.001 |
| 2–3 | 55 | 29.0 | 131 | 25.7 | 55 | 11.8 | ||
| 4–5 | 34 | 17.9 | 85 | 16.7 | 101 | 21.8 | ||
| ≥6 | 62 | 32.6 | 202 | 39.7 | 302 | 65.1 | ||
| Missing | 0 | 2 | 4 | |||||
| Comorbid conditions | ||||||||
| No | 84 | 44.2 | 248 | 48.5 | 0.308 | 178 | 38.1 | 0.148 |
| Yes | 106 | 55.8 | 263 | 51.5 | 289 | 61.9 | ||
| Missing | 0 | 0 | 1 | |||||
| Physical activity | ||||||||
| No | 116 | 61.1 | 268 | 52.5 | 0.042 | 219 | 46.8 | <0.001 |
| Yes | 74 | 38.9 | 243 | 47.5 | 249 | 53.2 | ||
P-value for comparison of cases and high-risk controls.
P-value for comparison of cases and low-risk-controls.
There was no evidence of effect modification by menopausal status; however, physical activity did modify the effect of diabetes on breast cancer risk. We adjusted for menopausal status and mammography screening which were confounders in our data.
Results
Cases were more likely than high-risk controls to be older, to be postmenopausal, not to have breastfed, and not to have used oral contraceptives; cases were more likely than low-risk controls to have a first-degree relative with breast cancer, to have a younger age at menarche, to have used hormone replacement therapy, and to have had more mammograms in the past 6 years; cases were more likely than both control groups not to have engaged in physical activity (Table 1).
There was no effect of diabetes on risk of breast cancer among high-risk (OR 1.02, 95% CI 0.71–1.48) or low-risk controls (OR 0.87, 95% CI 0.58–1.30) after adjustment for menopausal status and mammography screening (Table 2). In high-risk controls, we found that women with type 2 diabetes treated with insulin had a significant increase in breast cancer risk (OR 2.23, 95% CI 1.18–4.19), while women treated with oral medications (OR 0.72, 95% CI 0.42–1.23) or with neither insulin nor oral medications (OR 0.55, 95% CI 0.18–1.70) had non-significant decreases in risk. As similar pattern existed in low-risk controls, with a significantly lower risk of breast cancer associated with no treatment (OR 0.31, 95% CI 0.10–0.96). There was little effect of type of diabetes, age at diabetes onset or of family history of diabetes on breast cancer risk.
Table 2.
Odds ratios and 95% confidence intervals for the associations of diabetes with incident invasive breast cancer, South Texas Women’s Health Project, 2003–2008
| Characteristic | Cases (n=190) | High-risk controls (n=511) | ORa | (95% CI) | Low-risk controls (n=468) | ORa | (95% CI) | |||
|---|---|---|---|---|---|---|---|---|---|---|
| N | % | N | % | n | % | |||||
| History of diabetes | ||||||||||
| No | 125 | 65.8 | 354 | 69.3 | 1.00 | (referent) | 306 | 65.5 | (referent) | |
| Yes | 65 | 34.2 | 157 | 30.7 | 1.02 | (0.71–1.48) | 161 | 34.5 | 0.87 | (0.58–1.30) |
| Missing | 0 | 1 | ||||||||
| Type of diabetes | ||||||||||
| Gestational | 3 | 1.7 | 19 | 3.9 | 0.21 | (0.03–1.64) | 8 | 1.8 | 0.26 | (0.03–2.31) |
| Type 2 | 47 | 26.9 | 112 | 23.1 | 1.00 | (0.66–1.51) | 129 | 29.1 | 0.77 | (0.49–1.23) |
| Missing | 15 | 26 | 24 | |||||||
| Age at diabetes onset (years)b | ||||||||||
| <35 | 3 | 1.7 | 12 | 2.6 | 0.79 | (0.21–2.96) | 11 | 2.6 | 0.66 | (0.17–2.61) |
| ≥35 | 44 | 25.6 | 99 | 21.3 | 0.96 | (0.63–1.47) | 115 | 26.6 | 0.80 | (0.51–1.26) |
| Missing | 0 | 1 | 3 | |||||||
| Diabetes treatmentb | ||||||||||
| Insulin | 22 | 12.8 | 25 | 5.4 | 2.23 | (1.18–4.19) | 37 | 8.5 | 1.24 | (0.66–2.31) |
| Oral med. | 21 | 12.2 | 71 | 15.2 | 0.72 | (0.42–1.23) | 64 | 14.7 | 0.77 | (0.43–1.38) |
| No insulin or oral med. | 4 | 2.3 | 16 | 3.4 | 0.55 | (0.18–1.70) | 28 | 6.4 | 0.31 | (0.10–0.96) |
| Family history of diabetes | ||||||||||
| No | 59 | 31.6 | 160 | 31.6 | 1.00 | (referent) | 154 | 33.3 | 1.00 | (referent) |
| Yes | 128 | 68.4 | 346 | 63.4 | 1.02 | (0.71–1.49) | 309 | 66.7 | 1.16 | (0.77–1.74) |
| Missing | 3 | 5 | 5 | |||||||
Odds ratio (OR) and 95% confidence interval (95% CI) adjusted for menopausal status and mammography screening.
Among women with type 2 diabetes.
Relative to women who had no history of diabetes and did not engage in physical activity, women who had a diabetes history and did not exercise were at slightly reduced breast cancer risk (high-risk controls OR 0.96, 95% CI 0.63–1.48; low-risk controls OR 0.72, 95% CI 0.46–1.15) (Table 3). While women with diabetes who did exercise had significantly reduced breast cancer risk in both control groups (high-risk controls OR 0.40, 95% CI 0.21–0.79; low-risk controls OR 0.41, 95% CI 0.21–0.83) in comparison with women without diabetes who did engage in physical activity. This effect modification was significant for the high-risk control group (p-value for interaction=0.013), but not for the low-risk control group (p-value for interaction=0.183).
Table 3.
Odds ratios for breast cancer associated with joint effects of diabetes and physical activity, South Texas Women’s Health Project, 2003–2008
| Characteristic | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Did not engage in physical activity | Cases (n=116) | High-risk controls (n=268) | ORa | (95% CI) | Low-risk controls (n=219) | ORa | (95% CI) | |||
| N | % | N | % | n | % | |||||
| History of diabetes | ||||||||||
| No | 69 | 59.5 | 185 | 69.0 | 1.00 | (referent) | 136 | 62.4 | 1.00 | (referent) |
| Yes | 47 | 40.5 | 83 | 31.0 | 0.96 | (0.63–1.48) | 82 | 37.6 | 0.72 | (0.46–1.15) |
| Missing | 0 | 1 | 1 | |||||||
| Did engage in physical activity | Cases (n=74) | High risk-controls (n=243) | ORa | (95% CI) | Low-risk controls (n=249) | |||||
| N | % | N | % | |||||||
| History of diabetes | ||||||||||
| No | 56 | 75.7 | 169 | 69.6 | 1.00 | (referent) | 170 | 68.3 | 1.00 | (referent) |
| Yes | 18 | 24.3 | 74 | 30.4 | 0.40 | (0.21–0.79) | 79 | 31.7 | 0.41 | (0.21–0.83) |
| P-value for interaction | 0.013 | 0.183 | ||||||||
Odds ratio (OR) and 95% confidence interval (95% CI) adjusted for menopausal status and mammography screening.
Table 4 presents results of our validation study which indicated very high sensitivities of self-report of diabetes in comparison with serum glucose (ranging from 88.4 to 92.3) and the medical record (ranging from 86.5 to 94.8), which were higher among cases than controls. Specificities for serum glucose were much lower (ranging from 51.3 to 67.3), while those for the medical record were comparable across study groups (ranging from 86.8 to 89.6). After restricting our analysis to diabetes only (data not shown), specificities were considerably stronger for serum glucose (cases 75.0, high-risk controls 90.9, low-risk controls 82.4).
Table 4.
Validation of self-report of diabetes by serum glucose and medical record, South Texas Women’s Health Project, 2003–2008
| Sensitivity (95% CI)a | Specificity (95% CI) | Sensitivity (95% CI) | Specificity (95% CI) | Sensitivity (95% CI) | Specificity (95% CI) | |
|---|---|---|---|---|---|---|
| Cases (n=67) | High-risk controls (n=120) | Low-risk controls (n=121) | ||||
| Serum glucose | 92.3 (78.0–98.0) | 57.1 (37.4–75.0) | 90.1 (81.0–95.3) | 51.3 (35.0–67.3) | 88.4 (77.9–94.5) | 67.3 (52.8–79.3) |
| Cases (n=92) | High-risk controls (n=157) | Low-risk controls (n=200) | ||||
| Medical record | 94.8 (84.7–98.7) | 88.2 (71.6–96.2) | 86.5 (77.2–92.5) | 86.8 (75.9–93.4) | 91.9 (85.2–95.8) | 89.6 (80.0–95.1) |
95% confidence interval (95% CI)
Discussion
The Four Corners Breast Cancer Study, the one study that included a sufficient number of Hispanic women to stratify by ethnicity3, grouped Hispanic and Native American women so our studies are not strictly comparable. Nevertheless, we saw a similar association between diabetes and breast cancer risk (combined control groups OR 1.00) as the Four Corners Study (OR 0.92). Nor could we directly compare our findings on diabetes treatment (combined control groups insulin OR 1.82, oral medication OR 0.81, neither OR 0.43) with the Four Corners Study which grouped insulin and oral medications (either OR 0.91, neither OR 1.84), thus ours is the first study to report a significantly increased breast cancer risk associated with use of insulin among diabetics. We additionally adjusted our analyses for age, BMI at age 15 years, and number of full-term pregnancies which were confounders in the Four Corner Study to ensure our differing results were not due to confounding, but our findings did not materially change (data not shown).
The increased risk of breast cancer associated with insulin resistance and diabetes seen in most studies1, 2 is thought to be due to elevated levels of insulin that promote proliferative and antiapoptotic effects through the insulin receptor7. Metformin, an oral medication used to treat type 2 diabetes, has recently been proposed as a breast cancer chemopreventive agent8 as it may reduce insulin levels9 or activate AMP-dependent protein kinase thereby suppressing protein synthesis10. While the slightly reduced risk we saw for use of oral medications to treat type 2 diabetes provides support for the metformin mechanism, there was a greater reduction in risk among women who reported they took neither insulin nor oral medications. This latter finding may have been due to hyperinsulinemia induced chronic anovulation11 which would result in lower cumulative estrogen exposure over time thereby placing a woman at lower risk of breast cancer. Alternatively this latter finding may have been due to chance since we had very few cases (n=4) who took neither insulin nor oral medications for diabetes treatment.
One of the meta-analyses of the association between diabetes and breast cancer stratified by physical activity, and found nearly identical relative risks among women who did (relative risk [RR] 1.16) and did not exercise (RR 1.20)1. In contrast, we found a substantially lower breast cancer risk among diabetic women who exercised (combined control groups OR 0.42) than among diabetic women who did not exercise (combined control groups OR 0.87) and this effect modification was significant (combined control groups p-value for interaction=0.034). A plausible biological mechanism for our finding of a substantially reduced breast cancer risk resulting from the joint effect of diabetes and physical activity may be the reduction in insulin resistance that accompanies physical activity5. A randomized clinical trial of the effect of metformin or lifestyle intervention (which included moderate intensity physical activity for at least 150 minutes per week) on subsequent diabetes among persons at high risk of diabetes reported reductions of 31% and 58% compared to placebo after 2.8 years of follow-up 12. The investigators surmised that these interventions could also delay or prevent complications from diabetes.
Our study was limited by potential detection bias since women identified through mammography centers as controls may have been more likely than cases to have been screened for both diabetes and breast cancer. Three studies conducted in the U.S.13–15 and one study conducted in Canada16 found lower mammography screening rates among diabetics than among non-diabetics. However, when McBean et al.15 examined the association in women of races/ethnicities other than white or black, there was no difference in mammography screening rates by diabetes status. Additional limitations of our study were self-report of physical activity for leisure-time only which did not include physical activity for housework or work outside the home and is prone to misclassification, our lack of study power to detect some main effects, and the higher percentage of cases than controls with additional sources of information on diabetes which may have resulted in differential misclassification.
To minimize detection bias and misclassification we used serum glucose available for 48% of cases and 37% of controls, followed by medical records available for 35% of cases and 25% of controls, followed by self-report available for the remaining 17% of cases and 38% of controls to define diabetes. We conducted a validation study of diabetes reporting utilizing serum glucose and medical records and found that sensitivity was quite high relative to specificity, and that reporting was slightly more accurate among cases than among controls. Additional strengths of the study were the focus on Hispanic women who have been understudied with regard to breast cancer, high response rates, adjustment for mammography screening to further reduce the likelihood of detection bias, and assessment of confounding for established breast cancer risk and protective factors.
Hispanic women possess a number of breast cancer risk factors including diabetes and yet have a relatively low incidence of the disease. Very few breast cancer studies have focused on Hispanic women; however, the identification of protective factors against breast cancer may enlighten our understanding of the biological mechanisms of the disease. Our finding that physical activity modified the effect of diabetes on breast cancer in Hispanic women contributes to the sparse body of knowledge in this area and suggests hypotheses for further investigation. Should other studies confirm our results, physical activity should be explored as a means of reducing breast cancer risk in diabetic women.
Acknowledgments
This research was supported in part by grant numbers DAMD-17-03-1-0274 and DAMD-17-00-1-0340 from the Department of Defense, U.S. Army Medical Research and Materiel Command, and by grant number 5 P20 MD000170 from the National Center on Minority Health and Health Disparities.
The authors wish to thank the subjects, providers (Drs. Karen Brooks-Searle, Osvaldo Cantu, William Elkins, Carol Erwin, Ashraf Hilmy, Ruben Lopez, Roselle Pettorino, Todd Shenkenberg, Lonnie Stanton, Hector Salcedo-Dovi, She Ling Wong, Brownsville Community Health Center, Brownsville Doctors Hospital, Central Imaging, Clinica Santa Maria, Harlingen Medical Center, Planned Parenthood of Cameron and Willacy Counties, South Texas Cancer Center, South Texas Hospital, Su Clinica Familiar, Valley Baptist Medical Center-Harlingen), and study staff (Adela Rodriguez, Elena Garcia, Margarita Ramirez, Lydia Melendez, Dr. Alberto Diaz de Leon, Varun Gupta, Patty Hernandez, Terry Aguirre, Iris Cantu, Sandra Tipton, Laura Barrera, Tammy Witerski, Linda Camacho, Maria Sanchez, Janie Castillo, Adriana Ramos, Connie Hayes) for their invaluable assistance with the project.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Contributor Information
Maureen Sanderson, Email: msanderson@mmc.edu.
Gerson Peltz, Email: Gerson.peltz@utb.edu.
Adriana Perez, Email: Adriana.perez@uth.tmc.edu.
Matthew Johnson, Email: matthew.johnson@utb.edu.
Sally W. Vernon, Email: sally.w.vernon@uth.tmc.edu.
Maria E. Fernandez, Email: maria.e.fernandez@uth.tmc.edu.
Mary K. Fadden, Email: mkfadden@yahoo.com.
References
- 1.Larsson SC, Mantzoros CS, Wolk A. Diabetes mellitus and risk of breast cancer: a meta-analysis. Int J Cancer. 2007;121:856–862. doi: 10.1002/ijc.22717. [DOI] [PubMed] [Google Scholar]
- 2.Michels KB, Xue F. Diabetes, metabolic syndrome, and breast cancer: a review of the current evidence. Am J Clin Nutr. 2007;86(suppl):823S–835S. doi: 10.1093/ajcn/86.3.823S. [DOI] [PubMed] [Google Scholar]
- 3.Rollison DE, Giuliano AR, Sellers TA, Laronga C, Sweeney C, Risendal B, et al. Population-based case-control study of diabetes and breast cancer risk in Hispanic and non-Hispanic white women living in the US southwestern states. Am J Epidemiol. 2008;167:447–456. doi: 10.1093/aje/kwm322. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Friedenreich CM, Orenstein MR. Physical activity and cancer prevention: etiologic evidence and biologic mechanisms. J Nutr. 2002;132:3456S–3464S. doi: 10.1093/jn/132.11.3456S. [DOI] [PubMed] [Google Scholar]
- 5.Kriska AM, Edelstein SL, Hamman RF, Otto A, Bray GA, Mayer-Davis EJ, et al. Physical activity in individuals at risk of diabetes: Diabetes Prevention Program. Med Sci Sports Exerc. 2006;38:826–832. doi: 10.1249/01.mss.0000218138.91812.f9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Breslow NE, Day NE. The Analysis of Case-Control Studies. Vol. 1. IARC; Lyon, France: 1980. Statistical Methods in Cancer Research. [PubMed] [Google Scholar]
- 7.Papa V, Belfiore A. Insulin receptors in breast cancer: biological and clinical role. Endocrinol Invest. 1996;19:324–333. doi: 10.1007/BF03347871. [DOI] [PubMed] [Google Scholar]
- 8.Cazzaniga M, Bonanni B, Guerrieri-Gonzaga A, Decensi A. Is it time to test metformin in breast cancer clinical trials? Cancer Epidemiol Biomark Prev. 2009;18:701–705. doi: 10.1158/1055-9965.EPI-08-0871. [DOI] [PubMed] [Google Scholar]
- 9.Goodwin PJ, Pritchard KI, Ennis M, Clemons M, Graham M, Fantus IG. Insulin-lowering effects of metformin in women with early stage breast cancer. Clin Breast Cancer. 2008;8:501–505. doi: 10.3816/CBC.2008.n.060. [DOI] [PubMed] [Google Scholar]
- 10.Zakikhani M, Dowling R, Fantus IG, Sonenberg N, Pollak M. Metformin is an AMP kinase-dependent growth inhibitor for breast cancer cells. Cancer Res. 2006;66:10269–10273. doi: 10.1158/0008-5472.CAN-06-1500. [DOI] [PubMed] [Google Scholar]
- 11.Kaaks R. Nutrition, hormones and breast cancer: is insulin the missing link? Cancer Causes Control. 1996;7:605–625. doi: 10.1007/BF00051703. [DOI] [PubMed] [Google Scholar]
- 12.Knowler WC, Barrett-Connor E, Fowler SE, Hamman RE, Lachin JM, Walker EA, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. New Engl J Med. 2002;346:393–404. doi: 10.1056/NEJMoa012512. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Fontana SA, Baumann LC, Helberg C, Love RR. The delivery of preventive services in primary care practices according to chronic disease status. Am J Public Health. 1997;87:1190–1196. doi: 10.2105/ajph.87.7.1190. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Beckman TJ, Cuddihy RM, Scheitel SM, Naessens JM, Killian JM, Pankratz VS. Screening mammogram utilization in women with diabetes. Diabetes Care. 2001;24:2049–2053. doi: 10.2337/diacare.24.12.2049. [DOI] [PubMed] [Google Scholar]
- 15.McBean AM, Yu X. The underuse of screening services among elderly women with diabetes. Diabetes Care. 2007;30:1466–1472. doi: 10.2337/dc06-2233. [DOI] [PubMed] [Google Scholar]
- 16.Lipscombe LL, Hux JE, Booth GL. Reduced screening mammography among women with diabetes. Arch Intern Med. 2005;165:2090–2095. doi: 10.1001/archinte.165.18.2090. [DOI] [PubMed] [Google Scholar]