We found that women with high fasting C-peptide levels (values > 2.5 ng/mL) collected 3years after diagnosis) had more than a two-fold increased risk of breast cancer death compared with women with low C-peptide levels. Healthy or normal C-peptide levels are between 0.5 ng/mL and 2.0 ng/mL, and high levels indicate that the pancreas is producing too much insulin.28
Women who reported having physician-diagnosed type 2 diabetes were at an even higher risk of breast cancer death than women without type 2 diabetes and low C-peptide levels. However, because of wide 95% CIs in these categorical analyses, these results should be interpreted with caution. Yet the 95% CIs were quite narrow when examining analyses between continuous levels of C-peptide and death owing to all causes and breast cancer, perhaps because of increased statistical power. Thus the statistically significant and clinically meaningful continuous C-peptide analyses suggest a strong association with death owing to all causes and death resulting from breast cancer.
Our results of a positive association between postdiagnosis C-peptide levels and breast cancer death are similar to the one other published report by Goodwin et al21
examining this relationship. However, our analyses also adjusted for (adjuvant treatment and race) and stratified by (BMI, disease stage, and estrogen receptor status) covariates known to be associated with insulin or C-peptide levels and breast cancer outcomes. More specifically, although our sample size, number of deaths, and greater than two-fold increased risk of breast cancer death is similar to that of the results of Goodwin et al21
in 512 women (45 deaths, 42 from breast cancer; HR = 3.3; 95% CI, 1.5 to 7.0; P
for trend = .002), Goodwin et al examined fasting insulin collected approximately 3 months after diagnosis and breast cancer death. C-peptide concentration may be a more stable measure of insulin production than circulating insulin levels and may have actions distinct from that of insulin.28
Whether collecting blood close to diagnosis rather than a couple years after diagnosis affects death differently is unknown. A strength of the 3-month postdiagnosis blood collection of Goodwin et al21
is the preadjuvant treatment assessment; however, a strength of a 3-year postdiagnosis blood collection is that it provides evidence that host factors long-term after diagnosis may play a role in determining risk of death in breast cancer survivors. Lastly, a limitation of both the study of Goodwin et al21
and our study that is women were enrolled and treated for breast cancer before 2000 and therefore before common use of aromatase inhibitors. Thus we are uncertain whether our results would be similar among women taking aromatase inhibitors; however, we did adjust for hormone receptor status.
When we examined C-peptide and breast cancer deaths in subgroup analyses by estrogen receptor status, a stronger association was observed among women with estrogen receptor–positive tumors. Stronger associations have been observed between physical activity, BMI, and breast cancer outcomes in women with estrogen receptor–positive tumors compared with women with estrogen receptor–negative tumors.5–11
Given that physical activity and BMI are associated with insulin and C-peptide levels in women with and without breast cancer,12–14
we hypothesized that associations between C-peptide and breast cancer death would be stronger in women with estrogen receptor–positive tumors. Fasting insulin and C-peptide levels may then explain the observed associations between BMI at diagnosis and postdiagnosis physical activity levels and death in women with breast cancer.
Associations between C-peptide and breast cancer death were hypothesized to be stronger in women with a higher versus lower BMI, but the opposite was observed in our study. This finding may be explained by the fact that, among leaner women (those with a BMI < 25 kg/m2), average C-peptide levels were significantly lower among women who were alive at follow-up (1.86 ng/mL, standard deviation [SD] = 0.70) than among those who had died (2.58 ng/mL, SD = 1.52). In contrast, average C-peptide levels were higher, less variable, and quite similar among women with a BMI more than 25 kg/m2 (2.59 ng/mL, SD = 1.20 and 2.57 ng/mL, SD = 1.10, for women alive and not alive at follow-up, respectively). Obesity is a strong predictor of insulin resistance and hyperinsulinemia, and we showed in our study that overweight and obese groups of women with breast cancer had higher C-peptide levels. It is possible that the influence of C-peptide on mortality could only be demonstrated in the lower BMI group because C-peptide levels in the overweight and obese groups were generally high and less variable.
Our study has several limitations. We collected one fasting blood draw and therefore cannot completely characterize the women's exposure to C-peptide. However, other studies indicate that a single fasting blood measure of this analyte is highly reproducible in fasting postmenopausal women. This limitation would bias the hazard ratio toward the null; thus we may have observed stronger associations between C-peptide and death with repeat measures of C-peptide. Also, we could not assess the effect of change in C-peptide on death, which would require testing interventions to change this analyte such as weight loss, physical activity, or medications to alter C-peptide or insulin resistance.
In summary, our study indicates that C-peptide is associated with increased risk of death among women with and without diabetes. Our findings emphasize the importance of maintaining a low C-peptide level. Physical activity is associated with C-peptide and insulin in women with or without breast cancer.12–14
Previous findings from the HEAL cohort indicated that moderate-intensity physical activity was associated with a lower risk of death.9
The current study complements this finding, indicating the importance of lower C-peptide levels and lower BMI. Thus participating in recommended levels of 2 to 3 hours/wk of moderate-intensity physical activity before and after a diagnosis of breast cancer may be beneficial to maintaining a low BMI and low C-peptide level, as well as decreased risk of death resulting from all causes and breast cancer. Randomized controlled trials examining changes in fasting C-peptide or insulin levels, via weight loss and physical activity, on breast cancer outcomes are needed. Although no trial has examined weight loss– or physical activity–induced changes in insulin or C-peptide on breast cancer outcomes, trials testing metformin on breast cancer outcomes are in progress.29,30