|Year : 2021 | Volume
| Issue : 5 | Page : 519-525
Lipid profile in breast cancer patients: A case-control study done at a public tertiary hospital in Ibadan Nigeria
Adeniyi Adedayo Olabumuyi1, Abbas Adesina Abdus-Salam2, Babatunde Olutoye Ogunnorin2, Modupe Akinrele Kuti3
1 Department of Radiation Oncology, University College Hospital, University of Ibadan, Ibadan, Nigeria
2 Department of Radiation Oncology, University of Ibadan, Ibadan, Nigeria
3 Department of Chemical Pathology, University of Ibadan, Ibadan, Nigeria
|Date of Submission||30-Jan-2021|
|Date of Decision||27-Jun-2021|
|Date of Acceptance||10-Jul-2021|
|Date of Web Publication||11-Oct-2021|
Dr. Adeniyi Adedayo Olabumuyi
Department of Radiation Oncology, University College Hospital, Ibadan
Source of Support: None, Conflict of Interest: None
Background: Breast cancer, the leading malignancy among women in Nigeria in terms of incidence and mortality, contributes a greater proportion of cancer burden compared with other cancers in Nigeria. A need to further identify and modify risk factors of breast cancer is necessary to apply preventive medicine and decrease its disease burden. Serum lipid profile is an objective index of fat metabolism, but its relationship with breast cancer is yet to be extensively investigated in our environment. Aim: To explore the relationship of serum lipid profile with breast cancer in the University College Hospital, Ibadan. Methods: The anthropometry, body fat percentage and lipid profile of 70 patients with breast cancer and 71 age-matched controls were studied. Statistical Analysis: Differences in the mean values of the lipid profile parameters were tested for significance using the Student's t-test and Mann–Whitney U test. Results: The breast cancer patients had statistically significantly higher serum triglyceride (TG) and very low-density lipoprotein cholesterol (P < 0.001). This was also an independent risk factor with multivariate analysis. Demonstrating the odds of breast cancer rose by 5.1% (odds ratio = 1.051, P < 0.001) with each 1 mg/dl rise in serum TG. Conclusion: There is potential for serum TG to be utilized as a marker of breast cancer. However, this needs to be determined by more elaborate studies.
Keywords: Cancer, lipid metabolism, Nigeria, triglyceride
|How to cite this article:|
Olabumuyi AA, Abdus-Salam AA, Ogunnorin BO, Kuti MA. Lipid profile in breast cancer patients: A case-control study done at a public tertiary hospital in Ibadan Nigeria. Niger J Med 2021;30:519-25
|How to cite this URL:|
Olabumuyi AA, Abdus-Salam AA, Ogunnorin BO, Kuti MA. Lipid profile in breast cancer patients: A case-control study done at a public tertiary hospital in Ibadan Nigeria. Niger J Med [serial online] 2021 [cited 2021 Dec 8];30:519-25. Available from: http://www.njmonline.org/text.asp?2021/30/5/519/327959
| Introduction|| |
The burden of breast cancer is high in Nigeria as it has an incidence rate of 41.7/100,000 and a mortality rate of 18.8/100,000 in the country, as estimated by GLOBOCAN 2018. In addition, studies on the tumour biology of breast cancer in Nigeria indicate a higher prevalence of aggressive variants like triple-negative and human epidermal growth factor receptor 2+ breast cancers.
The resource-limited circumstances in Nigeria imply a high cancer burden due to limited facilities for treatment (trained staff and equipment). Another factor that adds to the burden is the out-of-pocket payment system for healthcare services. Due to this high burden of breast cancer in resource-limited countries such as Nigeria, preventive efforts need to be adopted. In addition, means to determine an at-risk group of individuals need to be developed.
Genetic screening and testing have been developed. However, these are largely reserved for individuals with a strong family history. Furthermore, these genetic tests are not widely available in the Nigerian setting. The relationship between fat and energy metabolism and breast cancer is documented. Literature has shown that high body mass index (BMI), waist circumference (WC) and other anthropometric indices increase breast cancer risk. Yet, anthropometry has its drawbacks. High BMI is proven to be more of a risk factor in postmenopausal breast cancer as opposed to premenopausal breast cancer, where a negative association with breast cancer exists.,,, Furthermore, there also is a significant variation of anthropometry with race. In addition, there are individuals who are metabolically obese with normal weight. These individuals are usually insulin resistant, with altered inflammatory and adipokine profiles, adipose tissue inflammation, excess visceral adipose tissue and ectopic fat deposition and are at risk for Type 2 diabetes mellitus and cardiovascular diseases., Furthermore, a secondary data analysis by Iyengar et al. in 2018 on postmenopausal women with normal BMI concluded that normal BMI categorization might be inappropriate for the risk assessment of breast cancer in postmenopausal women. They noted metabolic factors such as insulin, C-reactive protein, triglyceride (TG), and truncal fat mass were associated with breast cancer.
As such, a more objective index of fat and energy metabolism, which indicates activity at the molecular/cellular level and can readily be done in a resource-limited setting, is required. Serum lipid profile fits these criteria. There are reports of studies investigating the relationship between serum lipid profile and breast cancer. Abdelsalam et al., in a case-control study between 120 cases and 60 controls in Sudan, reported a nonstatistically significant increase in the serum total cholesterol (TC) values of cases compared to controls (P = 0.09). In another case-control study by Owiredu et al. in Ghana, among 100 cases and 100 controls, TC (P = 0.001) and serum TG (P = 0.026) were significantly raised. These findings were also noted in studies by Badid et al., Yadav et al., Hasija and Bagga and others.,,
We thus decided to ascertain the association between serum lipid profile and breast cancer in our environment to fill the knowledge gap, especially as it pertains to the Nigerian population.
| Methods|| |
Study design and setting
The study was done at the University College Hospital (UCH), Ibadan as a case-control study. Seventy cases of breast cancer were recruited from Radiation Oncology and Surgical Oncology Clinics, UCH, between August 2016 and January 2017. At the Family Medicine Clinic and Chief Tony Anenih Geriatric Centre UCH, Ibadan, 71 age- and gender-matched nonbreast cancer patients were recruited as controls.
The following inclusion criteria were applied to the cases: A histologically diagnosed invasive breast cancer patient with no previous cancer treatment (mastectomy, chemotherapy, radiation therapy, hormonal therapy). The exclusion criteria for the cases were: Patients with uncontrolled chronic comorbidities such as hypertension and diabetes mellitus, hypo/hyperthyroidism; patients on drugs known to alter lipid homeostasis (statins, bile acid sequestrants, nicotinic acid, fibrates, hormone replacement therapy, and other hormonal agents); patients with poor performance status (Eastern Cooperative Oncology Group [ECOG] ≥3); and patients with obvious nutritional impairment (severe mucositis, nasogastric tube feeding, parenteral feeding, or recent significant weight loss). For the controls, the inclusion criteria were the patient's gender and age (± one year) matches a case. The exclusion criteria for the controls were: Patients with breast disease; patients with uncontrolled chronic comorbidities such as hypertension and diabetes mellitus, hypo/hyperthyroidism, patients on drugs known to alter lipid homeostasis (statins, bile acid sequestrants, nicotinic acid, fibrates, hormone replacement therapy and other hormonal agents), patients with poor performance status (ECOG ≥3), patients with obvious nutritional impairment (severe mucositis, nasogastric tube feeding, parenteral feeding, or recent weight loss).
A questionnaire was used to retrieve sociodemographic and clinical data. The serum lipid profile tests on the participants were carried out on venous blood samples drawn after an overnight fast, lasting 9–12 h, into a gel and clot activator vacutainer. The samples were promptly transported to the Chemical Pathology Residents' Research Laboratory, UCH, for analysis on the same day. The serum lipid profile of all participants was determined using the enzymatic method on an automated analyzer (LANDWIND LW C100plus Auto Chemistry Analyzer). These included values of high-density lipoprotein-cholesterol (HDL-C), TG, and TC. From these values, Friedewald's equation was used to calculate low-density lipoprotein cholesterol (LDL-C) and very LDL-C (VLDL-C).
VLDL-C = TG/5
LDL-C = TC – (HDL-C + VLDL-C)
We arrived at a calculated total sample size of 140 through the following formula:
N = (Zα+ Zβ)2 × 2 × (SD2)/(μ1− μ2)2
N = Desired sample size
Zα = Two-sided percentage point of the normal distribution corresponding to 5% level of significance = 1.96
Zβ = Percentage point of the normal distribution corresponding to a power of 80% =0.84
SD = Standard deviation of LDL (the primary target of therapy in the ATP III guidelines) of a relevant previous study in Nigeria. The best estimate of the SD of LDL-C in women in southwest Nigeria is in a study by Ademuyiwa et al., in Abeokuta. The SD of the mean LDL-C of women in the study was 45.77 mg/dl.
μ1− μ2 = Mean difference of LDL between cases and controls the study aims to discern assumed to be half of the SD.
The estimated sample size was 63 (for each group of cases and controls). To adjust for a nonresponse of 10%, the estimated sample size was divided by 0.9, yielding a sample size of 70. We recruited 70 cases and 71 controls for the study, yielding a total sample size of 141.
The data were analyzed with the IBM Statistical Package for Social Sciences (Version 25.0. IBM Corp. Released 2017. Armonk, NY, USA). The quantitative variables were expressed with means and SD, the qualitative data were expressed with frequencies and percentages, and appropriate tables and charts were used. Normality was tested for using the Shapiro–Wilk test. The mean serum TG and TC of the breast cancer patients and their controls were analyzed using the Mann–Whitney U test as they were not normally distributed. The difference in the mean values of all other qualitative variables was analyzed with the Student's t-test.
To further eliminate confounders in the comparison of outcomes between the breast cancer patients and their controls, the significant variables were analyzed by multiple logistic regression analyses. The level of significance was set at 5%.
Ethical approval for this study was granted by the joint ethical review committee of the University of Ibadan/UCH, Ibadan (approval number: UI/EC/15/0460).
| Results|| |
All the participants of the study were women. The age range of the cases was 23–82 years, and their mean age was 52.1 ± 12.0 years. The controls had ages ranging between 24 and 84 years, and their mean age was 52.5 ± 12.6 years. The cases and controls were similar with regard to marital status, ethnicity, education, religion, and occupation. Over 83% of the controls were employed compared to 68.6% of cases; this difference was statistically significant [Table 1].
Of the 70 cases of breast cancer majority (60%) had left breast cancer. The majority of the patients had T4 sized tumors (71.4%) and N2 nodal status (52.9%). Over 60% of the patients presented with Stage III breast cancer, while 27.2% of the patients presented with distant metastases [Table 2]. Invasive carcinoma no special type (NST) was the histological type with the highest proportion [Table 3], noted in 93.0% of the cases. Majority of the patients presented with Scarff-Bloom-Richardson Grade 2 [Table 4].
|Table 2: Tumour, node, metastasis distribution of breast malignancy in the study population|
Click here to view
|Table 4: Distribution of cases of breast malignancies by Scarff-Bloom-Richardson grade|
Click here to view
The serum lipid profile values indicate serum TG was significantly higher among cases with a mean of 125.9 mg/dl compared to controls with mean serum TG of 84.0 mg/dl (P = <0.001). Both the premenopausal and postmenopausal groups also had significantly higher mean serum TG values among cases compared to controls. Consequentially, the mean serum VLDL-C was also significantly higher among the cases than controls in all the patients and in the premenopausal and postmenopausal groups. The difference in the mean values of TC, HDL-C, and LDL-C between the cases of breast cancer and the controls among the entire participants and even when grouped into premenopausal and postmenopausal was not significant [Table 5].
A previous study on the same participants as this study has been earlier published. It revealed a statistically significant association between unemployment and low percentage body fat (estimated by skinfold thicknesses [SFTs] at four body sites according to the Durnin and Womersley formula) and breast cancer (P = 0.010). Another publication on a study on the same participants also indicated the cases of breast cancer had lower anthropometric indices (weight, BMI, WC and hip circumference) compared to their controls. Thus, regression models to prove independent risks and eliminate confounding factors were done using uncorrelated, statistically significant variables (including percentage body fat from the previously published study). The logistic regression analysis was done in a series of 3 models [Table 6]. The first model utilised only employment status. This showed statistically significant higher odds for breast cancer among unemployed women compared to the employed, with an odds ratio (OR) of 3.687 (P = 0.018). In model 2, percentage body fat (estimated from SFTs) was added. In this model (model 2), unemployment remained a statistically significant factor with an OR of 3.707 (P = 0.019). Yet, breast cancer was significantly associated with low percentage body fat (estimated from SFTs) (P = 0.013). With an OR of 0.934, every percent increase in percentage body fat reduced the odds of breast cancer by 6.6%. The combined predictive nature of employment status, percentage body fat (estimated by SFT) and serum TG on breast cancer in the study population was analyzed in the 3rd logistic regression model. This proved unemployment was not a statistically significant predictor of breast cancer when analyzed alongside serum TG and percentage body fat by SFT in multivariate analysis. The association between breast cancer and low percentage body fat by SFT remained statistically significant (P = 0.001). With an OR of 0.9, every percent increase in body fat conferred a reduction of the odds of breast cancer by 10%. In model 3, serum TG was a statistically significant risk factor for breast cancer (P < 0.001). An OR of 1.051 indicated a 5.1% increase in the odds of breast cancer for each 1 mg/dl rise in serum TG.
|Table 6: Logistic regression of employment status, percentage body fat (skinfold thicknesses) and serum triglyceride on breast cancer in the entire sample|
Click here to view
| Discussion|| |
This study was designed to detect an association between serum lipid profile and breast cancer. The mean age of the patients (52 years) indicates breast cancer is still a disease that affects middle-aged women.
The breast cancer patients in this study had presenting findings in conformity with other studies conducted in this environment. Majority of the breast cancer cases in this study presented in advanced stages. In a study by Ogundiran et al., over 50% of the patients presented with axillary lymph nodes and 17% with distant metastases. While a study by Kene et al. in Zaria revealed 15.5% of patients presented with distant metastases and 46.6% with Stage III disease. This study had over 80% of the cases presenting with axillary lymphadenopathy, and over 27% of patients had distant metastases. This keeps with the pattern of late presentation of malignancies in this environment. For instance, Abdus-Salam et al., in their study on cervical cancer patients treated at this same center, also noted the late presentation.
In Nigeria, studies indicate a higher proportion of breast cancer patients are premenopausal, ranging from 50% to 70%., However, in this study, postmenopausal breast cancer accounted for 60% of the cases recruited. The most prevalent histological type, as noted by previous studies in Nigeria, is the invasive carcinoma NST., In this study, invasive carcinoma NST accounted for 93% of the histological types. A minority (5.7%) of the tumours in this study were Grade 1 tumors. Previous studies in the environment agree with this. Gukas et al. in Jos revealed Grade 1 breast cancer comprised approximately 6% of breast cancer in their study. Similarly, a study by Adebamowo et al. done in Ibadan, at the same centre where this study was done, showed that only a little over 9% of breast cancers presented with Grade 1 tumours.
The mean serum TC in this study was higher among cases than controls. Similarly, the mean serum TG levels were also higher among cases. However, with respect to significance, the differences in the mean serum TG in both the pre and postmenopausal women were significant. The differences remained significant on multivariate analysis (OR = 1.051 95% confidence interval [CI] 1.030–1.071; P < 0.001). This compares well to the study by Barkat et al. in Morocco in which breast cancer patients were found to have elevated, but not significantly, TC. Just like this study, there was a significantly elevated TG in the breast cancer patients, which remained in multivariate analysis (OR = 4.5, 95% CI 2.94–6.88; P = 0.0094). In another study by Abdelsalam et al. in Sudan, a nonstatistically elevated TC was noted among breast cancer cases compared to controls. There are studies that showed statistically elevated TC and TG. These include studies by Owiredu et al., Gupta et al., Badid et al. and Yadav et al.,, However, Peela et al., in a similar case-control study carried out in Libya, reported significantly elevated TC but no significant difference in TG. The significant TC elevation in the Libyan study remained in both postmenopausal and premenopausal breast cancer. This can be explained by the relationship between oxidative stress, which results in dyslipidemia, and malignancy. Increased values of TG, TC, and LDL-C in combination with decreased values of HDL-C are associated with increased levels of pro-inflammatory cytokines such as tumour necrosis factor-α and interleukin-6. These factors have pro-carcinogenic effects by altering mitosis, apoptosis, angiogenesis and senescence.
In this study, the mean serum VLDL-C was significantly higher among breast cancer cases than controls. A similar pattern was noted for TG; it remained significant with multivariate analysis (OR = 1.280 CI 1.162–1.410; P < 0.001). The higher mean values were seen in both premenopausal and postmenopausal cases. The study had a nonsignificant difference between the serum LDL-C of the cases and controls, though the breast cancer patients had lower mean values. There was also a nonsignificant difference between serum HDL-C of the cases and controls, though cases had higher mean values. Even though Peela's et al.'s study also had (significantly) elevated HDL-C among cases of breast cancer, most large studies indicate low HDL-C is the actual risk factor for breast cancer. For instance, the Norwegian cross-sectional study by Furberg et al. indicated low HDL-C was associated with a hormonal profile in healthy subjects that puts them at an increased risk of breast cancer. While the Atherosclerosis Risk in Communities (ARIC)-cohort study reported a modest association between low baseline HDL-C and incidence of premenopausal breast cancer (hazard ratio = 1.67 CI 1.06–2.63). It is difficult to place a finger on the reason this study's finding regarding HDL-C differs. A similar study was done by Owiredu et al. in Ghana using 100 breast cancer patients and 100 controls. Unlike in this study, significantly higher levels of LDL-C were found in the breast cancer patients compared to the controls. Furthermore, in that study, no significant difference existed between the serum levels of serum HDL-C between the breast cancer patients and their controls.
The true association between deranged serum lipid profile and breast malignancy merits further investigation. Which is the possible risk/etiological factor for which? Cross-sectional studies are limited by assumptions of cause. As findings noted were those seen in the participants of the study on the day they were recruited. One can achieve this through larger cohort and longitudinal studies. These studies could also be done in multiple centers across the nation or the West African region. These proposed studies would also be more informative if other metabolic markers are assessed. For example, Iyengar et al. also noted a positive association between breast cancer risk and serum leptin, baseline insulin and C-reactive protein, apart from TG. They alluded this to dysregulation of insulin signaling, which can lead to activation of PI3K/Akt/mTOR and Ras/Raf/MAPK pathways; these promote cell growth and division and increase the risk of neoplasia.
Among the controls, breast cancer was ruled out by clinical means. However, mammography was not done. As such, there is a remote possibility a control could have had undiscovered breast cancer.
| Conclusion|| |
Given these findings, the results of this study conform to previous studies in revealing higher TG and VLDL-C as probable risk factors of breast cancer. However, findings in the literature contradict that of this study pertaining to breast cancer in relation to LDL-C and HDL-C. More elaborate studies would be needed to definitively ascertain the pattern in our environment.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Agboola AJ, Musa AA, Wanangwa N, Abdel-Fatah T, Nolan CC, Ayoade BA, et al.
Molecular characteristics and prognostic features of breast cancer in Nigerian compared with UK women. Breast Cancer Res Treat 2012;135:555-69.
Holmes MD, Willett WC. Does diet affect breast cancer risk? Breast Cancer Res 2004;6:170-8.
Picon-Ruiz M, Morata-Tarifa C, Valle-Goffin JJ, Friedman ER, Slingerland JM. Obesity and adverse breast cancer risk and outcome: Mechanistic insights and strategies for intervention. CA Cancer J Clin 2017;67:378-97.
Baer HJ, Tworoger SS, Hankinson SE, Willett WC. Body fatness at young ages and risk of breast cancer throughout life. Am J Epidemiol 2010;171:1183-94.
Feigelson HS, Jonas CR, Teras LR, Thun MJ, Calle EE. Weight gain, body mass index, hormone replacement therapy, and postmenopausal breast cancer in a large prospective study. Cancer Epidemiol Biomarkers Prev 2004;13:220-4.
Amadou A, Hainaut P, Romieu I. Role of obesity in the risk of breast cancer: Lessons from anthropometry. J Oncol 2013;2013:906495.
Chandran U, Hirshfield KM, Bandera EV. The role of anthropometric and nutritional factors on breast cancer risk in African-American women. Public Health Nutr 2012;15:738-48.
Conus F, Rabasa-Lhoret R, Péronnet F. Characteristics of metabolically obese normal-weight (MONW) subjects. Appl Physiol Nutr Metab 2007;32:4-12.
Park YM, Fung TT, Steck SE, Zhang J, Hazlett LJ, Han K, et al
. Diet Quality and Mortality Risk in Metabolically Obese Normal-Weight Adults. Mayo Clin Proc 2016;91:1372-83.
Iyengar NM, Arthur R, Manson JE, Chlebowski RT, Kroenke CH, Peterson L, et al.
Association of body fat and risk of breast cancer in postmenopausal women with normal body mass index: A secondary analysis of a randomized clinical trial and observational study. JAMA Oncol 2019;5:155-63.
Abdelsalam KE, Hassan IK, Sadig IA. The role of developing breast cancer in alteration of serum lipid profile. J Res Med Sci 2012;17:562-5.
Owiredu WK, Donkor S, Addai BW, Amidu N. Serum lipid profile of breast cancer patients. Pak J Biol Sci 2009;12:332-8.
Badid N, Ahmed FZ, Merzouk H, Belbraouet S, Mokhtari N, Merzouk SA, et al.
Oxidant/antioxidant status, lipids and hormonal profile in overweight women with breast cancer. Pathol Oncol Res 2010;16:159-67.
Yadav NK, Poudel B, Thanpari C, Chandra Koner B. Assessment of biochemical profiles in premenopausal and postmenopausal women with breast cancer. Asian Pac J Cancer Prev 2012;13:3385-8.
Hasija K, Bagga HK. Alterations of serum cholesterol and serum lipoprotein in breast cancer of women. Indian J Clin Biochem 2005;20:61-6.
Kirkwood BR, Sterne JA. Calculation of required sample size. In: Essential Medical Statistics. 2nd
ed. Oxford: Blackwell Publishing Ltd.; 2003. p. 413-28.
Ademuyiwa O, Ugbaja RN, Rotimi SO. Plasma lipid profile, atherogenic and coronary risk indices in some residents of Abeokuta in south-western Nigeria. Biokemistri 2008;20:85-91.
Durnin JV, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: Measurements on 481 men and women aged from 16 to 72 years. Br J Nutr 1974;32:77-97.
Abdus-Salam AA, Babatunde OO, Adenipekun AA, Adeniyi AO, Olabumuyi O. Percentage body fat in breast cancer patients at the University College Hospital, Ibadan: A case-control study. Niger J Med 2019;28:237-47.
Abdus-Salam AA, Ogunnorin OB, Olabumuyi AA, Orekoya AA. Anthropometry parameters in breast cancer patients at the University College Hospital Ibadan: A case-control study. Afr J Med Med Sci 2020;49:287-95.
Ogundiran TO, Ayandipo OO, Ademola AF, Adebamowo CA. Mastectomy for management of breast cancer in Ibadan, Nigeria. BMC Surg 2013;13:59.
Kene TS, Odigie VI, Yusufu LM, Yusuf BO, Shehu SM, Kase JT. Pattern of presentation and survival of breast cancer in a teaching hospital in north Western Nigeria. Oman Med J 2010;25:104-7.
Abdus-Salam AA, Eriba LO. Histopathological patterns of cervical carcinoma seen at a radiotherapy centre in Ibadan, Nigeria. Niger Q J Hosp Med 2013;23:334-7.
Adebamowo CA, Famooto A, Ogundiran TO, Aniagwu T, Nkwodimmah C, Akang EE. Immunohistochemical and molecular subtypes of breast cancer in Nigeria. Breast Cancer Res Treat 2008;110:183-8.
Gukas ID, Jennings BA, Mandong BM, Igun GO, Girling AC, Manasseh AN, et al.
Clinicopathological features and molecular markers of breast cancer in Jos, Nigeria. West Afr J Med 2005;24:209-13.
Laamiri FZ, Otmani A, Ahid S, Barkat A. Lipid profile among Moroccan overweight women and breast cancer: A case-control study. Int J Gen Med 2013;6:439-45.
Peela J, Jarari A, El Saiety S, El Busaifi S, El Awamy H, Srikumar S. The relationship between serum lipids and breast cancer in Libya. Biochem Anal Biochem 2012;1:117.
Haddy N, Sass C, Droesch S, Zaiou M, Siest G, Ponthieux A, et al
. IL-6, TNF-α and atherosclerosis risk indicators in a healthy family population: the STANISLAS cohort. Atherosclerosis 2003;170:277-83.
Furberg AS, Jasienska G, Bjurstam N, Torjesen PA, Emaus A, Lipson SF, et al.
Metabolic and hormonal profiles: HDL cholesterol as a plausible biomarker of breast cancer risk. The Norwegian EBBA Study. Cancer Epidemiol Biomarkers Prev 2005;14:33-40.
Kucharska-Newton AM, Rosamond WD, Mink PJ, Alberg AJ, Shahar E, Folsom AR. HDL-cholesterol and incidence of breast cancer in the ARIC cohort study. Ann Epidemiol 2008;18:671-7.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]