Prognostic Impact of Pregnancy in Korean Patients with Breast Cancer

Breast cancer diagnosed during pregnancy or 1 year postpartum is referred to as pregnancy‐associated breast cancer (PABC). This article compares the outcomes of patients with PABC or pregnancy after breast cancer with outcomes of patients with breast cancer that was not associated with pregnancy in Korean patients diagnosed and treated at a single institution.

Abstract

Background.

Pregnancy concurrent with, shortly before, or after breast cancer poses unique challenges because hormonal changes in pregnancy potentially interact with breast cancer outcomes.

Materials and Methods.

We studied a cohort of 3,687 female patients of reproductive age (<50 years) with breast cancer, linking a large institutional database and the nationwide claims database to comprehensively capture exposure status and tumor characteristics. Exposures included breast cancer during pregnancy, postpartum breast cancer (<12 months after delivery), and pregnancy after breast cancer.

Results.

Forty‐five patients with postpartum breast cancer were significantly more likely to have advanced stage, hormone receptor‐negative tumor and to be younger than 35 years at diagnosis than those without postpartum breast cancer. This trend was not observed with 18 patients with breast cancer during pregnancy. The unadjusted 5‐year survival rates were 77% versus 96% for patients with postpartum breast cancer versus their counterparts, 89% versus 96% for patients with breast cancer during pregnancy versus their counterparts, and 98% versus 96% for patients with pregnancy after breast cancer versus their counterparts, respectively. In the multivariable analyses, postpartum breast cancer exhibited hazard ratios for death of 1.57 (95% confidence interval [CI], 0.82–2.99), whereas those for breast cancer during pregnancy and pregnancy after breast cancer were 1.09 (95% CI, 0.15–7.91) and 0.86 (95% CI, 0.26–2.83), respectively.

Conclusion.

Postpartum breast cancer, but not breast cancer during pregnancy, was associated with advanced stage, younger age at diagnosis (<35 years), hormone receptor‐negative disease, and poorer survival. Pregnancy after breast cancer did not compromise overall survival.

Implications for Practice.

Although pregnancy around the time of diagnosis of breast cancer is expected to become increasingly common with maternal age at first childbirth on the rise, data on the prognostic impact of pregnancy have been inconsistent and rare from Asian populations. In this investigation of a Korean patient cohort with breast cancer, pregnancy‐associated breast cancer was associated with advanced stage, younger age at diagnosis (<35 years), hormone receptor‐negative disease, and poorer survival. This adverse impact of pregnancy on the prognosis was apparent with postpartum breast cancer but not observed with breast cancer during pregnancy. Pregnancy after breast cancer did not compromise overall survival.

Introduction

Breast cancer, the leading cause of cancer‐related mortality in women in Korea [1], as well as worldwide [2], is the most common malignancy diagnosed during or around the time of pregnancy [3]. When breast cancer is diagnosed during pregnancy or 1 year postpartum, it is called pregnancy‐associated breast cancer (PABC) [4]. Both PABC and pregnancy following breast cancer pose unique challenges because hormonal changes occurring in pregnancy potentially interact with breast cancer outcomes. Furthermore, both of these conditions are expected to become increasingly common with maternal age at first childbirth on the rise [4], [5].

Numerous studies have reported that PABC tends to present with advanced stages and has a poor prognosis. Some researchers attributed this to diagnostic delay [6], [7] because pregnancy and lactation increase breast density, resulting in difficulty interpreting the results of a physical examination and a screening mammography [8]. Other studies suggested an independent effect of pregnancy on the prognosis, which was not diluted even after adjusting for the stage of disease [4]. In contrast, pregnancy after the diagnosis and treatment of breast cancer did not adversely affect survival and is generally deemed safe in breast cancer survivors [9].

However, most of the previous studies on PABC and pregnancy after breast cancer were from Western countries with only a few from Asia. Breast cancer in Asia has several distinct features compared with that in Western countries: younger age of onset, increasing incidence and mortality [10], and a higher proportion of patients under 35 years of age [11]. These differences possibly reflect the socioeconomic status, lifestyle and culture, genetic makeup, and tumor biology of Asians, which are different from those of Westerners, implying that the findings from Western populations may not directly apply to Asian patients [12]. In this regard, the impact of pregnancy on breast cancer outcome remains to be explored in Asian populations.

The aim of this study was to compare the outcome of PABC and pregnancy after breast cancer with breast cancer not associated with pregnancy in Korean patients with breast cancer diagnosed and treated at a single institution. To comprehensively capture the exposure status to pregnancy, the nationwide claims database of Korea was linked to the large institutional database, because it covers nearly all the pregnancy events of the Korean people.

Materials and Methods

Data Sources

We obtained the required information for this retrospective cohort study from a large hospital‐based database and the nationwide claims database of the Korean Health Insurance Review and Assessment Service (HIRA). Korea's National Health Insurance is a government‐controlled, public medical insurance program that provides coverage to about 97% of its people, with the remaining 3%, mostly those in the low‐income bracket, covered by Medical Aid [13]. Hospitals and clinics submit claims to the HIRA for review to be reimbursed for costs of medical services provided within this system. These claims data include the relevant diagnostic codes from the Korean Classification of Disease (KCD), patient demographics, medical procedures used, and prescription records, with the KCD largely based on the 10th revision of the International Classification of Diseases. Although nearly all information on the use of medical services in Korea is indicated this way, records often lack details on the patients' survival and the clinicopathological features of the tumor, some of which are known to be highly relevant to the prognosis of patients with breast cancer [14], [15]. Data on the tumor characteristics and patients' survival were thus separately reviewed from the institutional database of Seoul National University Hospital (SNUH) and linked with the information on the diagnosis and treatment of breast cancer as well as the history of delivery or abortion from the HIRA database using the social security numbers of the affected individuals.

The study protocol was reviewed and approved by the institutional review board of SNUH (approval number, 1708‐149‐879). All patient records were deidentified prior to the study, and a waiver of informed consent was obtained for the study.

Patient Selection

Of a cohort of 11,565 adult female patients who had undergone surgery for breast cancer at SNUH from 2007 to 2015, 10,853 patients were identified with both the diagnostic and treatment codes for breast cancer recorded on the same day from the HIRA data. We excluded 1,526 patients who were diagnosed with breast cancer before January 1, 2008, to examine the medical history of the study population for a minimum of 1 year before the diagnosis date. Next, 4,658 patients were excluded who were aged at least 50 years at the diagnosis of breast cancer because we defined reproductive age as less than 50 years based on the previously reported mean and median age at menopause in Korean women [16]. Among the 4,669 remaining cases, 3,687 patients had data sufficient for the present analyses and were deemed eligible (Fig. 1).

Figure 1.

Selection of the study participants.

Abbreviation: SNUH, Seoul National University Hospital.

Exposures and Potential Confounders

Breast cancer during pregnancy was defined as a history of delivery or induced abortion within 9 months of the diagnosis of breast cancer and postpartum breast cancer as a breast cancer diagnosis within 12 months of delivery, and both were counted as PABC. Pregnancy after breast cancer was defined as delivery 9 months or more after the diagnosis of breast cancer. The definitions of the exposures are graphically depicted in supplemental online Figure 1.

Collected variables included the following: gender, date of the diagnosis, age at the diagnosis, comorbidities, histologic type, the presence or absence of tumor tissue expression of the estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2), along with the extent of the cancer spread at diagnosis classified in stages as tumor (T), lymph node (N), and metastasis (M). Diagnosis date was designated as the date of earliest treatment for breast cancer, namely the first date of chemotherapy for patients who had received chemotherapy before surgery or the date of surgery for the remaining cases. Comorbidities were graded using the Charlson comorbidity index as proposed previously [17], [18]. Histologic types were classified as invasive ductal carcinoma, invasive lobular carcinoma, and others, which included papillary carcinoma, mucinous carcinoma, Paget's disease, tubular carcinoma, medullary carcinoma, and mixed histology with presumed primary origin of breast. Disease stage was coded according to the American Joint Committee on Cancer staging, 6th edition, for cases diagnosed from 2007 to 2009, and 7th edition, for patients diagnosed between 2010 and 2015, respectively [19], [20].

Statistical Analysis

The baseline characteristics of the patients were described with a mean, median, SD, minimum, and maximum for continuous variables and a frequency and proportion for categorical variables. The percentages of the categorical variables were compared using the chi square or Fisher's exact test and the means of continuous variables using the Wilcoxon rank sum test between patients with a different exposure status. The primary outcome was overall survival, which was defined as the time from the diagnosis date of breast cancer to the date of death of any cause or last follow‐up. The cutoff point for exposure was June 30, 2017, and January 5, 2018, for death events. Survival curves were plotted using the Kaplan‐Meier methods and compared between groups using log‐rank tests. Cox proportional hazards models were applied to examine the impact of pregnancy on overall survival after adjusting for potential confounders. Among the exposures, pregnancy after breast cancer was incorporated into a Cox model as a time‐varying covariate with transition of the exposure state from nonexposure to exposure at the time of conception. Because the number of patients with an unknown HER2 expression status was greater than those with other variables missing, the comparison of the baseline characteristics and the multivariable analyses were performed separately with and without HER2 as a covariate. All p values were two‐sided, and values of p < .05 were considered statistically significant. SAS Enterprise Guide 6.1 was used for all analyses.

Results

Baseline Characteristics of the Patients

Among the 3,867 patients, 18 and 45 patients were classified as having breast cancer during pregnancy and postpartum breast cancer, respectively, comprising 63 cases of PABC. Fifty‐seven patients were identified as having given birth after the diagnosis of breast cancer. Supplemental online Figure 2 shows the prevalence of delivery and abortion in the overall 3,867 patients, in patients younger than 35 years, and in patients who were aged at least 35 years. The majority of exposures were observed in patients younger than 35 years.

Patients with PABC were significantly more likely to have advanced stage and/or hormone receptor‐negative tumor and to be younger than 35 years at diagnosis of breast cancer than those without it (p value for T stage, .0001; N stage, .0011; ER negativity, <.0001; PR negativity, <.0001; age at breast cancer diagnosis <35 years, <.0001; Table 1). A similar distribution pattern of the clinical characteristics was observed with postpartum breast cancer compared with their counterparts (p value for T stage, <.0001; N stage, .0001; M stage, .0421; ER negativity, <.0001; PR negativity, <.0001; age at breast cancer diagnosis <35 years, <.0001; supplemental online Table 1), whereas these characteristics were comparable between patients with and without breast cancer during pregnancy (supplemental online Table 2). A higher proportion of patients who had experienced pregnancy following breast cancer was noted to be younger than 35 years at diagnosis of breast cancer and to have a hormone receptor‐negative tumor compared with their counterparts (p value for ER negativity, .0225; PR negativity, .0372; age at breast cancer diagnosis <35, <.0001; supplemental online Table 3).

Table 1. Characteristics of patients with and without pregnancy‐associated breast cancer.

^aChi‐square test.

^bFisher's exact test.

^cWilcoxon rank‐sum test.

Abbreviations: CCI, Charlson comorbidity index; ER, estrogen receptor; HER2, human epidermal growth factor receptor 2; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; M, metastasis; N, node; PR, progesterone receptor; T, tumor.

When the comparison was restricted to 2,438 patients who had HER2 data available, the distribution of the clinical characteristics largely mirrored the previous results (supplemental online Tables 4–7).

Survival Curves

Figure 2 represents the unadjusted survival curves of patients, compared between patients with and without PABC (Fig. 2A), postpartum breast cancer (Fig. 2B), breast cancer during pregnancy (Fig. 2C), and pregnancy after breast cancer (Fig. 2D). Survival was significantly shorter in the group of 63 patients with PABC than in the 3,624 patients without PABC with a 5‐year survival rate of 80.11% (95% confidence interval [CI], 65.37–89.07) versus 95.99 % (95% CI, 95.26‐ 96.61; p value <.0001; Fig. 2A). Likewise, the postpartum breast cancer group (n = 45) had an inferior survival time compared with their counterparts (n = 3,642) with a 5‐year survival rate of 77.15% (95% CI, 60.12–87.61) and 95.97% (95% CI, 95.24–96.60; p value <.0001), respectively (Fig. 2B).

Figure 2.

Survival curves of the patients. Survival is analyzed by presence or absence of pregnancy‐associated breast cancer (A), postpartum breast cancer (B), breast cancer during pregnancy (C), and pregnancy after breast cancer (D).

Abbreviation: PABC, pregnancy‐associated breast cancer.

In contrast, groups with and without breast cancer during pregnancy (n = 18 and 3,669, respectively) had a comparable survival time with a 5‐year survival rate of 88.89% (95% CI, 43.30–98.36) and 95.76% (95% CI, 95.02–96.40; p = .6353), respectively (Fig. 2C). When the survival curves were compared between 57 patients who had pregnancy after breast cancer and 3,630 patients who did not, there was no significant survival difference between the two groups with a 5‐year survival rate of 98.11% (95% CI, 87.35–99.73) for the pregnancy after breast cancer group and 95.71% (95% CI, 94.96–96.36; p = .6881) for those without it (Fig. 2D).

Survival curves of the patients with HER‐2 data available showed a similar pattern with those of the original study population (supplemental online Fig. 3).

Univariable and Multivariable Analyses of Survival

In the Cox proportional hazards model of survival with PABC, the hazard ratios (HRs) of death for patients with PABC were 4.53 (95% CI, 2.53–8.13; p <.0001) and 1.52 (95% CI, 0.82–2.83; p = .1841) from the univariable and multivariable analyses, respectively. Similarly, postpartum breast cancer was associated with a worse survival with a univariable HR of 5.34 (95% CI, 2.91–9.82; p <.0001) and a multivariable HR of 1.57 (95% CI, 0.82–2.99; p = .1708), respectively. In contrast, breast cancer during pregnancy was not associated with a worse survival with a univariable HR of 1.60 (95% CI, 0.22–11.45; p = .6384) and a multivariable HR of 1.09 (95% CI, 0.15–7.91; p = .9355), respectively (Table 2).

Table 2. Cox proportional hazards model of survival for patients with and without pregnancy‐associated breast cancer.

Abbreviations: CCI, Charlson comorbidity index; CI, confidence interval; ER, estrogen receptor; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; M, metastasis; N, node; PR, progesterone receptor; T, tumor.

With pregnancy after breast cancer as the time‐varying covariate, pregnancy after breast cancer was not associated with a worse survival with a univariable HR of 1.48 (95% CI, 0.47–4.66; p = .5012) and a multivariable HR of 0.86 (95% CI, 0.26–2.83; p = .8004), respectively (Table 3).

Table 3. Cox proportional hazards model of survival for patients with and without pregnancy after breast cancer as time‐varying covariate.

Abbreviations: CCI, Charlson comorbidity index; CI, confidence interval; ER, estrogen receptor; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; M, metastasis; N, node; PR, progesterone receptor; T, tumor.

In all the above‐mentioned multivariable analyses, T2 or T3 versus T1, N2 or N3 versus N0, and M1 versus M0 were independently associated with a worse survival, whereas ER positivity versus ER negativity, PR positivity versus PR negativity, and ≥35 years of age versus <35 years of age at diagnosis were associated with a longer survival (Tables 2, 3). When limited to the 2,438 patients with HER2 data available, the results of multivariable analyses mostly followed the above findings, with HER2 positivity being an additional independent factor for a favorable survival (supplemental online Tables 8–11).

Discussion

This study showed that cases with postpartum breast cancer (<12 months after delivery) were significantly more likely to be of advanced stage, younger age at diagnosis (<35 years), hormone receptor‐negative disease, and worse survival, whereas breast cancer during pregnancy exhibited similar distribution of stage, age, and ER positivity, as well as survival, compared with their counterparts. Pregnancy following breast cancer had no adverse effect on survival in our patient cohort with breast cancer.

Although PABC has been conventionally defined to include breast cancer diagnosed during pregnancy and the postpartum period, previous studies in Western populations have demonstrated that these two have a different impact on the prognosis, as in the current study. A meta‐analysis by Azim et al. compared 3,628 patients with PABC and 37,100 controls and found that the poor prognosis of PABC was primarily driven by patients with postpartum breast cancer rather than PABC diagnosed during pregnancy [4]. Similarly, Amant et al. reported a comparable survival of patients with breast cancer diagnosed during pregnancy compared with those without PABC [21].

This survival difference may indicate that the biologic behavior of postpartum breast cancer differs from that of breast cancer diagnosed during pregnancy. Indeed, mammary gland involution following pregnancy has been suggested to explain early dissemination and poor prognosis of postpartum breast cancer [22]. Breast involution occurs as a tissue remodeling process to the prepregnant state through wound healing and inflammation, which supposedly contributes to a tumor‐promoting microenvironment [23], [24]. Other factors that have been postulated to play a role in the tumorigenesis of PABC include immune tolerance during and following pregnancy [25] as well as the effects of gestational hormones on the tumor stroma [26].

Interestingly, higher stage, younger age at diagnosis, and hormone receptor negativity, but not pregnancy per se, were independently associated with a worse prognosis in the multivariable analyses in our study. Association with pregnancy may exert its detrimental effect via these known risk factors of breast cancer, with or without yet unknown factors. However, it remains largely controversial whether pregnancy has an independent prognostic impact in patients with PABC.

We also found that pregnancy after breast cancer did not confer a worse prognosis on patients with breast cancer, in agreement with increasing evidence supporting the safety of pregnancy in this setting. A meta‐analysis of 1,244 cases versus 18,145 controls reported that childbearing did not compromise overall survival in patients with a history of breast cancer [9]. Nevertheless, studies addressing this topic, including our own, are inherently prone to the “healthy mother effect,” in which patients without evidence of disease recurrence are more likely to conceive [27]. Subsequent studies thus matched the disease‐free interval of cases and controls and demonstrated that pregnancy following breast cancer had no detrimental effect on recurrence or survival irrespective of the ER status [28], [29]. Of note, younger age at diagnosis (<35 years), hormone receptor negativity, and HER‐2 positivity were more frequent in patients who experienced pregnancy following breast cancer than in those who did not, although tumor stages were comparable among them. This seems plausible as younger patients are more likely to experience motherhood after treatment of breast cancer than older patients and breast cancer at a young age is often associated with endocrine‐unresponsive tumors [30]. In addition, patients with hormone receptor‐positive disease would have a lower chance of pregnancy after completion of the planned tamoxifen treatment [31].

Our study has several limitations. Most of all, this is a retrospective single center study with unavoidable selection bias and may not be completely representative of the whole Korean population. Moreover, only patients with surgically resected breast cancer were included in the current analyses, rendering patients who had not received surgery underrepresented in this study. Secondly, subtle differences may have gone undetected because of the limited number of cases. Lastly, this analysis did not address parity because of the lack of relevant data. Parity, however, would have influenced the outcome because pregnancy is known to have a dual effect on the development of breast cancer, promoting and then suppressing breast cancer depending on the length of time following childbirth [32]. Despite these limitations, the present study is one of the rare studies on pregnancy and breast cancer outcomes conducted in Asian countries and the first to report separate analyses for breast cancer during pregnancy and postpartum breast cancer among these. Furthermore, we integrated an institutional database with nationwide claims data to acquire comprehensive information on exposure as well as tumor characteristics, which is often lacking in registry‐based large‐scale studies.

Conclusion

PABC was associated with advanced stage, younger age at diagnosis (<35 years), hormone receptor‐negative disease, and poorer survival. This adverse impact of pregnancy on the prognosis was apparent with postpartum breast cancer but not observed with breast cancer during pregnancy. Pregnancy after breast cancer did not compromise overall survival in our patient cohort with breast cancer. Further study is warranted to confirm our findings and to further explore the biologic background behind them.

See http://www.TheOncologist.com for supplemental material available online.

Author Contributions

Conception/design: Kyung‐Hun Lee

Provision of study material or patients: Miso Kim, Tae‐Yong Kim, Seock‐Ah Im, Han‐Byoel Lee, Hyeong‐Gon Moon, Wonshik Han, Dong‐Young Noh, Kyung‐Hun Lee

Collection and/or assembly of data: Mihong Choi, Jiyeon Han, Bo Ram Yang, Myoung‐jin Jang, Kyung‐Hun Lee

Data analysis and interpretation: Mihong Choi, Jiyeon Han, Bo Ram Yang, Myoung‐jin Jang, Kyung‐Hun Lee

Manuscript writing: Mihong Choi, Kyung‐Hun Lee

Final approval of manuscript: Mihong Choi, Jiyeon Han, Bo Ram Yang, Myoung‐jin Jang, Miso Kim, Tae‐Yong Kim, Seock‐Ah Im, Han‐Byoel Lee, Hyeong‐Gon Moon, Wonshik Han, Dong‐Young Noh, Kyung‐Hun Lee

Disclosures

Seock‐Ah Im: AstraZeneca, Amgen, Eisai, Hanmi, Roche, Novartis, Pfizer (C/A), AstraZeneca, Pfizer (RF). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board