Serum lipids, lipoproteins, and risk of breast cancer: a nested case-control study using multiple time points

Abstract

Background: There is strong evidence that breast cancer risk is influenced by environmental factors. Blood lipid and lipoprotein levels are also influenced by environmental factors and are associated with some breast cancer risk factors. We examined whether serial measures of serum lipids and lipoproteins were associated with breast cancer risk.

Methods: We carried out a nested case-control study within a randomized long-term dietary intervention trial with 4690 women with extensive mammographic density followed for an average of 10 years for breast cancer incidence. We measured lipids in an average of 4.2 blood samples for 279 invasive breast cancer case subjects and 558 matched control subjects. We calculated subaverages of lipids for each subject based on menopausal status and use of hormone replacement therapy (HRT) at blood collection and analyzed their association with breast cancer using generalized estimating equations. All statistical tests were two-sided.

Results: High-density lipoprotein-cholesterol (HDL-C) (P = .05) and apoA1 (P = .02) levels were positively associated with breast cancer risk (75(th) vs 25(th) percentile: HDL-C, 23% higher; apoA1, 28% higher) and non-HDL-C (P = .03) and apoB (P = .01) levels were negatively associated (75(th) vs 25(th) percentile: non-HDL-C, 19% lower; apoB, 22% lower). These associations were observed only when lipids were measured when HRT was not used. Total cholesterol and triglyceride levels were not statistically significantly associated with breast cancer risk.

Conclusions: These results demonstrate that serum lipids are associated with breast cancer risk in women with extensive mammographic density. The possibility that interventions for heart disease prevention, which aim to reduce non-HDL-C or raise HDL-C, may have effects on breast cancer risk merits examination.

Figure 1.

Examples of calculation of lipid subaverages. A maximum of three subaverages of serum lipid measurements were calculated for each subject depending on menopausal status and hormone replacement therapy (HRT) use at the time of blood collection: 1) premenopausal (no HRT), 2) postmenopausal and taking HRT, and 3) postmenopausal and not taking HRT. The figure shows examples of how the lipid subaverages would be calculated for subjects with measurements at six time points: (A) premenopausal, no HRT, (B) premenopausal, no HRT and postmenopausal, HRT, (C) premenopausal, no HRT and postmenopausal, no HRT, (D) premenopausal, no HRT, postmenopausal, HRT, and postmenopausal, no HRT. Sixty percent of subjects had only one subaverage (all of their samples in one category), 33% had two subaverages, and 7% had three subaverages. HRT = hormone replacement therapy.

Figure 2.

Serum lipids (subaverages) and risk of breast cancer: interquartile odds ratio (95% confidence interval). Results are adjusted for study group (low-fat dietary intervention or comparison), parity at baseline (parous, nonparous), if smoked ever at baseline (yes, no), if had first degree relatives with breast cancer at baseline (yes, no), study site (Toronto, London, Hamilton, Windsor, Vancouver, Surrey), age at menarche (years), age at birth of first child (years), number of live births, average weight (kg), average age (years), date of random assignment, and menopausal status and homone replacement therapy use (three categories). Data missing for estrogen receptor status for 13 subjects. The interquartile range for each lipid was determined based on all subaverages for all subjects, and then the odds ratios calculated corresponding to that difference using the beta estimate for each lipid. All statistical tests were two-sided. ApoA1 = apolipoprotein A1; ApoB = apolipoprotein B; HDL-C = high density lipoprotein cholesterol; HRT = hormone replacement therapy.

Figure 3.

Serum lipids (subaverages) and risk of breast cancer stratified by menopausal status and hormone replacement therapy use: interquartile odds ratio (95% confidence interval). Results are adjusted for study group (low-fat dietary intervention or comparison), parity at baseline (parous, nonparous), if smoked ever at baseline (yes, no), if had first degree relatives with breast cancer at baseline (yes, no), study site (Toronto, London, Hamilton, Windsor, Vancouver, Surrey), age at menarche (years), age at birth of first child (years), number of live births, average weight (kg), average age (years), and date of random assignment. The interquartile range for each lipid was determined based on all subaverages for all subjects, and then the odds ratios calculated corresponding to that difference using the beta estimate for each lipid. All statistical tests were two-sided. ApoA1 = apolipoprotein A1; ApoB = apolipoprotein B; HDL-C = high density lipoprotein cholesterol; HRT = hormone replacement therapy.