Nipple Aspirate Fluid Collection, Related Factors and Relationship between Carcinoembryonic Antigen in Nipple Aspirate Fluid and Breast Diseases in Women in Harbin, PRC

Breast cancer continues to be a significant health threat, with an estimated 182,460 new cases and 40, 480 deaths occurring in United States in 2008 as reported by the American Cancer Society (1). In contrast to the situation in most developed countries, the breast cancer rate among Chinese women is much lower. Compared with the 101.1 per 100,000 people of the top-ranking United States, the breast cancer incidence rate was only 18.7 in China, which ranked Chinese native women 142nd in the world in 2002. The rate among Chinese American women, however, is 55 per 100,000 people, which could be alarming. Breast cancer among Chinese women who have been in the United States >10 years is 80% higher than their newly arrived peers (2). All of these have suggested that environmental factors such as diet may be of etiologic significance.

Physical examination and mammography have been the generally accepted screening tools available for breast cancer both in screening and in clinical practice. Thousands of women undergo these two methods each year to obtain optimal control of their breast disease. Unfortunately, physical examination does not identify a significant number of early breast cancers, and mammography misses 10% to 40% of early breast cancers (3, 4). Moreover, if a woman begins screening for breast cancer at the age of 40 years and her life expectancy is 75 to 80 years, she will have had mammography at least 17 to 20 times in her lifetime, and the potential harm of multiple mammography is unclear (5). Therefore, noninvasive and sufficiently sensitive and specific predictive markers or models for the prevention strategies to define the presence of malignancy in breast are needed.

Nipple aspiration has been reported as a quick, painless, and noninvasive method to obtain breast epithelial cells and extracellular fluid from the breast ductal and lobular epithelium (6). However, the ability to obtain adequate fluid has consistently been associated with four factors: age between 35 and 50 years, earlier age at menarche, non-Asian compared with Asian ethnicity, and a history of lactation (7). Several studies have investigated the epidemiologic, biochemical, and cytologic characteristics of nipple aspiration fluids in an effort to predict the risk of future development of carcinoma (8, 9). Previous studies with long-term follow-up showed that women with abnormal cytology in breast fluid had an increased relative risk of breast cancer compared with women from whom fluid was not obtained and with women whose fluid had normal cytology (10, 11).

Nipple aspirate fluid contains highly concentrated proteins secreted from the ductal and lobular epithelium that is the source of 99% of breast carcinoma (12). Nipple aspiration fluid has been identified as a potential source of biomarkers for the early diagnosis of breast cancer, such as the two human glandular kallikreins hK2 and hK3 (also known as prostate-specific antigen), basic fibroblast growth factors, urokinase-type plasminogen activator, plasminogen activator inhibitor, and urokinase-type plasminogen activator receptor (13-16). In addition, serum carcinoembryonic antigen (CEA), the first human cancer-associated antigen and serologic tumor biomarker, has been reported to assess and monitor tumor burden in patients with cancers of breast, colon, or other sites (17-19).

It has been reported that women with Asian ethnicity seem to have a lower success rate for nipple aspiration fluid collection. However, these studies were actually carried out among Chinese immigrants in American or European countries (20). Furthermore, no study has explored the association of CEA level in nipple aspiration fluids with breast disease in Chinese native women. This study is the first to examine the factors related to the ability to obtain nipple aspiration fluid in women in China and to compare the results with previous studies. We also compare CEA levels among normal women and those with breast cancer and with proliferative lesions, and further evaluate the diagnostic value of CEA in nipple aspiration fluids from patients with breast diseases.

Cross-sectional and case-control studies were used to examine the factors associated with the ability to obtain nipple aspiration fluid. The potential subjects included: (a) women who had a clinical breast examination in the Preventive and Health Care Hospital of Harbin Medical University from September to December in 2002; (b) women who had visited the Third Affiliated Hospital of Harbin Medical University for breast cancer or benign breast disease from January 2004 to January 2005; and (c) volunteers recruited from women living in Harbin and Jia Musi County from January 2004 to January 2005. All the participants were genetically unrelated. Lactating or pregnant women and those with bleeding tendencies, cicatricial nipples, local infections, or spontaneous nipple discharge were excluded from this study. Written informed consent was obtained from each participant. The Committee on Human Research of Harbin Medical University approved this study.

To evaluate the association of CEA levels in nipple aspiration fluid and breast disease, a matched case-control study was used. Patients who yielded nipple aspiration fluid and who were diagnosed with unilateral invasive breast cancer by pathologic diagnosis or with benign breast disease (including hyperplasia of mammary glands, breast adenofibroma, breast intracanalicular papilloma, and mammitis) diagnosed by clinical symptoms and auxiliary examination at the Third Affiliated Hospital of Harbin Medical University were enrolled. A woman with breast cancer was matched by age (±5 y) with both a patient with benign breast disease and a healthy control who came from the three study groups noted above.

Nipple aspiration fluids were collected using a modified breast pump consisting of a 20-mL syringe attached by a short plastic tube to a small round glass suction cup. In brief, after the nipples were cleansed with alcohol swabs to remove cellular debris, the subjects were asked to pressurize their breasts for several minutes following the guidelines of our researchers. With the cup placed over the nipple, the plunger was withdrawn and held for 5 to 15 s until fluid appeared. The fluid was collected by capillary tube and transferred to microcentrifuge tubes, which were stored at −60°C, within 8 h of collection. Nipple aspiration fluid was defined as obtained successfully at 2 μL to 20 μL of fluid from one breast.

The demographic characteristics, such as age, ethnicity, marital status, occupation, and educational achievement of all participants were determined by interview. The potential factors, such as age of menarche, age at first child, current menstrual status, reproductive history, presence of ear wax, fatty diet, and body mass index, which were related with the ability to obtain nipple aspiration fluid and with breast disease were administrated in one questionnaire.

Nipple aspiration fluids were diluted with FCS dependent on the volume and viscous degree, and centrifuged. Quantitative CEA assays were done using ELISA and conducted according to the protocol of the commercial assay kits, which were produced by American BPB Biological Company.

SPSS version 13.0 was used for all statistical analysis. Distributions of variables and rates of obtaining nipple aspiration fluid successfully for different groups were computed. The P value (0.05) of χ² test was used to determine the difference of rates among the groups of breast cancer, benign breast disease, and healthy controls. Univariate and multivariate logistic regression analyses were used to estimate odds ratios (OR) and 95% confidence intervals. The continuous variables (e.g. age) were categorized into dummy argument to improve the goodness-of-fit logistic models.

For the study of the relationship between CEA levels and breast disease, CEA levels from cancerous breasts, from breasts with benign disease, and from breasts of healthy controls were compared. When the participants with benign breast disease or healthy women had both breasts yielding nipple aspiration fluid, the means of CEA levels from both breasts were used in the analysis.

Means, medians, and full ranges of CEA levels of cases and controls were computed and compared with nonparametric test. Univariate (multivariate) conditional logistic regression analysis was used to estimate the ORs for the relationship between breast disease and CEA levels considering the potential confounders. Additionally, sensitivity, specificity, and area of receiver operating characteristic curve were used to evaluate the value of CEA levels in nipple aspiration fluids for breast disease. Pearson or Spearman correlation analysis was used to assess the association for the CEA concentration in nipple aspiration fluid from both breasts.

A total of 587 participants were enrolled in this study. Their age ranged from 20 to 76 years, with a mean of 43 years and a median of 42 years. Because most participants were of Han ethnicity (95.66%) and married (99.15%), these two variables were not used in logistic regressions reported in Table 1. Nipple aspiration fluid was obtained from both breasts in 59 (10.05%) women and from one breast of 142 (24.19%) women, giving a total collection rate of 34.24% (201 of 587; Table 1).

Among all the participants, 160 women were diagnosed with unilateral invasive breast cancer, 176 were diagnosed with benign breast disease (hyperplasia of mammary glands, 154; breast adenofibroma, 17; breast intracanalicular papilloma, 2; mammitis, 3), and the other 251 were healthy women. Nipple aspiration fluid was obtained from a little higher proportion of healthy women (35.86%) and benign breast disease (35.80%) than from women with a cancerous breast (30%), but these differences were not significantly different (χ² = 1.76; P = 0.42).

To examine the factors affecting the obtainment of nipple aspiration fluid, ORs and 95% confidence intervals were estimated by both univariate and multivariate logistic regression analyses (Table 1). Based upon univariate logistic regression analysis, the following factors were statistically significant: age 30 to 50 years (P < 0.01), wet-type ear wax (P < 0.01), postmenopausal (P < 0.01), earlier age at menarche (P = 0.04), lactation of >12 months (P = 0.01), and fatty diet (P = 0.02). Nipple aspiration fluid production was significantly higher in women ages 30 to 50 years compared with women >50 years old (OR, 2.99; 95% confidence interval, 1.78-5.04). The proportion of successful collection for women who had wet-type ear wax was 3.98 times greater than that in women who had dry-type ear wax (OR, 3.98; 95% confidence interval, 2.58-6.15).

In the multivariate logistic regression analysis, variables significant in the univariate analysis and that significantly associated with nipple aspiration fluid obtainment in previous studies were included in the full model by manual entry. Earlier age at menarche, lactating for >12 months, wet-type ear wax, and fatty diet were still significantly associated with a greater ability to obtain nipple aspiration fluid (P < 0.05) in this study. However, menstrual status was no longer statistically significant. In addition, compared with the result of the univariate logistic regression analysis, earlier age at menarche and lactating for >12 months are more likely to obtain nipple aspiration fluid.

Forty-eight women with breast cancer and 48 matched patients with benign breast disease and 48 matched healthy volunteers were included in the case-control study to determine the correlation of CEA levels and breast disease (Table 2). CEA concentrations in nipple aspiration fluids of all subjects ranged from 24.52 ng/mL to 420.63 ng/mL. Medians of CEA were 211.38 ng/mL, 137.59 ng/mL, and 129.47 ng/mL for breast cancer, benign breast disease, and healthy controls, respectively. For statistical analysis, the CEA levels for the three groups were normalized by log₁₀ transformations. CEA levels in the nipple aspiration fluid of cancerous breasts were significantly higher than those from breasts with benign disease (P < 0.05) and healthy controls (P < 0.01). No significant difference between breasts with benign disease and healthy controls was observed (P_adj = 1.00).

Comparing cancerous breasts and healthy controls, the result of univariate conditional logistic regression indicated that CEA levels in nipple aspiration fluid were significantly related with breast cancer (OR, 4.24; P < 0.01). Adjustment for potential confounding factors (age at menarche and lactating time) showed a stronger association with OR_adj at 5.39 (P < 0.01; Table 3). When the CEA cutoff value was 307.61 ng/mL, the sensitivity of CEAs was 35.42% and specificity was 85.42%. With the cutoff value set at 102.32 ng/mL, the corresponding sensitivity and specificity were 81.25% and 29.17%, respectively (Table 4). The area under the receiver operating characteristic curve, which indicated the diagnosed value of CEA in nipple aspirate fluid, was 0.65 (Fig. 1).

Analysis for the correlations of CEA from bilateral nipple aspiration fluids were conducted among 33 women (including 6 cancerous breasts, 12 breasts with benign disease, and 15 healthy controls). CEA levels in nipple aspiration fluids from breasts with carcinoma were significantly higher than those from opposite side tumor-free breasts for the 6 women (P < 0.01). No significant differences were observed for CEA levels from the bilateral breasts for the groups of both breasts with benign disease and healthy controls. However, no correlations for the CEA concentrations from the bilateral breasts were determined by the coefficient correlation of Spearman for the three groups (r = 0.60, P = 0.21 for women with breast cancer; r = 0.46, P = 0.14 for women with benign breast disease; r = −0.11, P = 0.69 for tumor-free controls).

Additionally, the mean value of log₁₀CEA (2.41) from the tumor-free breasts of 6 breast cancer patients was significantly higher than that from the breasts of 15 healthy women (log₁₀CEA = 2.13; P = 0.03).

Studies conducted on the collection of nipple aspiration fluid and related factors found that several factors were associated with the ability to collect nipple aspiration fluid by breast pump, including age, early menarche, history of lactation, high dietary fat consumption, dietary intake of lactose (21, 22), and ear wax. Ear wax is a normal product of the ear that protects the skin of the ear from water and infection. Asians have mostly dry ear wax; in the United States, however, 85% of whites and virtually all blacks have wet ear wax. It has been reported that wet ear wax is related with the ability of yielding nipple aspiration fluid, because both the ceruminous glands and the breasts are exocrine glands (23).

A striking ethnic difference for the proportion of women who yielded nipple aspiration fluid by breast pump was observed. Wrensch et al. (7) reported a lower proportion of nipple aspiration fluid obtainment for the 25- to 49-year-old Asian-descendant premenopausal nulliparous women (17%) compared with White American women (38%). In United States–born Chinese women compared with Asia-born Chinese women, the proportion of yielding nipple aspiration fluid was 46.3% versus 31.2% (20). In contrast, however, Sauter et al. (24) reported a high rate of 94.4% in American women.

In the present study, nipple aspiration fluid was obtained from 201 (34.24%) women of 587 participants. The highest yield of nipple aspiration fluid was from women ages 30 to 50 years with early onset of menarche. The same results have been reported by Wrensch et al. (7). A lower proportion of nipple aspiration fluid obtained from the women who had not lactated (26.67%; 8 of 30) was observed compared with the women who had experienced lactation (35.15%; 193 of 549) with a marginal significant difference (P = 0.08). Earlier age at menarche, lactating >12 months, wet-type ear wax, and fatty diet remained significantly associated with higher successful collection rates, even after adjustment for age. Although menstrual status has been suggested to be highly significantly associated with increased ability to obtain fluid in previous studies (22), it was not found to be significant in this study.

King EB et al. suggested that an optimum collection method for nipple aspiration fluid cytology should consist of at least three or four separate fluid aspiration attempts instead of one. Nipple aspiration fluid was obtained from 178 women (66.67%) at the first visit and from an additional 15, 10, 2, and 4 women at visits 2, 3, 4, and 5, respectively, for a cumulative total of 78.27% by visit 5 (25). However, all the women included in our study were at their first visit. The proportion of women yielding nipple aspiration fluid will be improved in future studies through a variety of modifications in the original technique and giving breasts hot fomentation before collection.

Many investigators have suggested that several chemical constituents in nipple aspiration fluids, such as immunoglobulin, lipid, lactose, estrogen, prostate-specific antigen, and CEA are associated with breast disease (26-28). Foretva et al. evaluated CEA in 215 nipple aspiration fluids in 1998. Although CEA levels in nipple aspiration fluid exceed normal serum CEA levels by approximately 200-fold, no significant differences were observed (29) due to small sample size (four nipple aspiration fluids from cancerous breasts were available in Foretva's study). Zhao et al. further examined this relation with the expanded sample size, which included 44 women with invasive breast cancers, 67 women with hyperplasia of mammary glands, and 277 healthy controls (30). Significant differences were found for the medians of CEA levels in nipple aspiration fluid from the three groups, but when specificity was 90%, the corresponding sensitivity for cancer detection was 32%.

In our study, medians of CEA titers were 211.38 ng/mL, 137.59 ng/mL, and 129.47 ng/mL for breast cancer, benign breast disease, and healthy controls, respectively. No significant difference was observed between CEA levels from benign breast disease and healthy controls, which suggested that the concentration of CEA in nipple aspiration fluid cannot be used as a diagnostic biomarker for benign breast disease. In this study, CEA levels were significantly associated with breast cancer in both univariate (OR, 4.24; P < 0.01) and multivariate conditional logistic regression (OR, 5.39; P < 0.01). This result was the same as that reported by Zhao et al. (30). However, when the sensitivity was 35.42%, the specificity was 85.42% (CEA = 307.61 ng/mL). These findings suggested that the test of CEA levels in nipple aspiration fluid was comparable with noninvasive clinical examinations (when sensitivity is 50%, the corresponding specificity is 90%), and were inferior to mammography (when sensitivity is 80%, the corresponding specificity is 91%). Furthermore, the area under the receiver operating characteristic curve was 0.65, which indicated that the test of CEA cannot be used as diagnostic tool (general standard for diagnosis is ≥0.7). Accordingly, a CEA test in nipple aspiration fluid can only be treated as auxiliary diagnosis of mammograms for early cancer detection.

Published studies have suggested that CEA levels ranged from undetectable to 8,400 ng/mL (mean, 1,100 ng/mL). The concentration of CEA level in nipple aspiration fluids was 100 times higher than that in corresponding serum, including both cancerous and healthy subjects (29). This elevation in nipple aspiration fluid might be caused by secretion and reabsorption with the intraductal of mammary gland repeatedly for many years. Compared with the CEA concentration in the study of Zhao et al. (30), which was conducted among American women, it was a little lower in this study. Factors associated with this difference were assumed as techniques for assaying, ethnicity of subjects, experimental condition, etc. (31).

Nipple aspiration fluids were obtained from both breasts of six breast cancer patients in this study. The levels of CEA (256.33 ng/mL) in nipple aspiration fluids from tumor-free breasts of cancer patients were much higher than that from healthy controls (134.71 ng/mL; P = 0.03). The explanation for this significant difference might be as follows: the cancerous patients have more possibility to be exposed to risk factors, which caused the elevation of CEA levels. In addition, the CEA may accumulate in the tumor-free breasts of cancerous patients along with the blood circulation and lymphatic return. Furthermore, the lymphatic metastasis and hematogenous metastasis occurring during the advanced stage of breast cancer may bring more risk of adenogenesis for tumor-free breasts compared with healthy women. If this is right, monitoring CEA level in nipple aspiration fluid from tumor-free breasts of cancerous patients would be meaningful for predicting the occurrence of cancer in the healthy breast and prognosis.

In summary, the nipple aspiration fluid collection rate of Chinese native women is comparably lower than that of non-Asian-descendant women. Consistent with prior studies, earlier age at menarche, lactating >12 months, and wet-type ear wax were significantly associated with higher successful collection rates. The significant relationship between CEA levels and breast cancer risk suggested that CEA levels in nipple aspiration fluid may be a potential biomarker for breast cancer detection. However, the low sensitivity and specificity indicated that it cannot be as a diagnosis criterion independently. CEA level might be used in conjunction with other assays for early cancer detection. Further studies to explore better nipple aspiration collection technique, as well as to clarify the association between breast cancer with cytologic and chemical constituents in nipple aspiration fluid are necessary. It is proposed that more sensitive and specific biomarkers be researched and the value of nipple aspiration fluid as a diagnosis biomarker by either dependent or combination be reappraised.

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.