Risk of second cancers cancer after a first primary breast cancer: A systematic review and meta-analysis

YGYNO-975688; No. of pages: 14; 4C: Gynecologic Oncology xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno

Review

Risk of second cancers cancer after a first primary breast cancer: A systematic review and meta-analysis Esther Molina-Montes a,b,c, Mar Requena a, Emilio Sánchez-Cantalejo a,b,c, Mariana F. Fernández b,c,d, Manuel Arroyo-Morales b,e, Jaime Espín a,b, Juan P. Arrebola b,c,d, María-José Sánchez a,b,c,⁎ a

Granada Cancer Registry, Andalusian School of Public Health, Granada Spain Instituto de Investigación Biosanitaria de Granada (Granada.ibs), Hospitales Universitarios de Granada/Universidad de Granada, Granada, Spain CIBER de Epidemiología y Salud Pública (CIBERESP), Spain d Radiology Department, University of Granada, Granada, Spain e Physical Therapy Department, University of Granada, Granada, Spain b c

H I G H L I G H T S • The scientific evidence and the risk of second primary cancers in women diagnosed with a first breast cancer was examined. • This is the first review and meta-analysis in regards to risk of second cancer after breast cancer. • Women diagnosed with breast cancer have a 17% higher second cancer risk, with this risk being higher in premenopausal women.

a r t i c l e

i n f o

Article history: Received 29 July 2014 Accepted 29 October 2014 Available online xxxx Keywords: Breast neoplasms Second primary Multiple primary Risk Systematic review Meta-analysis

a b s t r a c t Objective. To examine the scientific evidence and the risk of second primary cancers in women diagnosed with a first primary breast cancer. Methods. The literature was searched in Pubmed and Embase and included studies published up to June 2013, using population-based data and IARC/AICR codification rules for multiple primary cancers. A qualitative synthesis was carried out and the methodological quality of the studies evaluated. Standardised incidence ratios (SIRs) on second cancer risk, weighted by the standard error of each study, were pooled using fixed and random effects models. SIRs were also pooled by age at diagnosis (b50 and ≥50 years), and time since diagnosis of the first breast cancer (b 10 and ≥10 years). Results. 15 out of 710 articles fulfilled the inclusion criteria. All of them were retrospective cohort studies either population-based (13 studies) or hospital-based studies (2 studies). The studies varied with respect to number of cases, selection criteria, definition of multiple primary cancers, and the second cancer sites included. SIRs reported in these studies for all cancers combined varied from 1.0 to 1.4. The pooled SIR estimate for second cancer risk was 1.17 (95% CI: 1.10–1.25). By age groups, SIR estimates were 1.51 (95% CI: 1.35–1.70) for women younger than 50 years and 1.11 (95% CI: 1.02–1.21) for those who were older. Women with breast cancer are at risk of second cancers within the first 10 years after the first breast cancer diagnosis (SIR: 1.19; 95% CI: 1.06– 1.33), and thereafter (SIR: 1.26; 95% CI: 1.05–1.52). Conclusion. This higher risk of second cancers in women diagnosed with a first primary breast cancer with respect to the general population emphasises the importance of prevention and control policies aimed at reducing incidence of second cancers. © 2014 Elsevier Inc. All rights reserved.

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

0 0

⁎ Corresponding author at: Andalusian School of Public Health, Campus Universitario de Cartuja, Cuesta del Observatorio 4, Apartado de Correos 2070, E-18080 Granada, Spain. Fax: +34 958 027503. E-mail address: [email protected] (M.-J. Sánchez).

http://dx.doi.org/10.1016/j.ygyno.2014.10.029 0090-8258/© 2014 Elsevier Inc. All rights reserved.

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

2

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Search strategy . . . . . . . . . . Study inclusion and exclusion criteria Data extraction . . . . . . . . . . Synthesis of results . . . . . . . . Meta-analysis . . . . . . . . . . . Results . . . . . . . . . . . . . . . . Results of the bibliographical search. Description of the studies . . . . . Results of the meta-analyses . . . . Discussion . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . Conflict of interest . . . . . . . . . . Role of the funding source . . . . . . . Author's contributions are as follows . . References . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

Introduction

Methods

Breast cancer represents a major public health issue worldwide. It is the most commonly diagnosed cancer among women, with 1.38 million new cases estimated in 2012 [1]. In Europe, estimates of cancer incidence and mortality in 2012 show that it remains being the most common cancer and cause of cancer-related death in women [2]. Early detection through systematic screening, better access to care, and advances in treatments have been leading to a decline in mortality rates [3,4]. Thus, as the number of women who overcome a breast cancer is considerably increasing, the likelihood of developing subsequent cancers, i.e. Multiple Primary Tumours (MPT), becomes higher. Subsequent cancers after an initial breast cancer could be attributed either to common risk factors predisposing to both the first and second cancer, such as genetic predisposition or other identified risk factors, or to treatment-related side effects [5]. Several population-based cancer registry studies [6–11] as well as studies involving several cancer registries [12,13], have evaluated the risk of developing second primary cancers among women diagnosed with a first primary breast cancer with respect to the general population. Most of these studies were derived from European data [6–13] and from the National Cancer Institute's Surveillance, Epidemiology (SEER) cancer registries in the United States [14–16]. However, risk estimates provided by these studies are largely different, with an overall excess risk ranging between 15 and 45% for all cancer sites combined. Risk differences by age groups have also been examined in some of these studies [7, 10–13], showing that women diagnosed with breast cancer at premenopausal ages were at higher risk of developing a second cancer. In general, second primary cancers of the endometrium, ovary, melanoma, stomach and colon cancer have been reported to occur more frequently [6–10], although there is no consensus between studies. Some studies have also provided risk estimates of second cancers according to treatment of the breast cancer, such as radiotherapy [14, 17–20], chemotherapy and surgery [17–20] or hormonal therapy [18–20], to assess how treatment-related factors may influence this risk. However, as information on treatment is not systematically collected in population-based cancer registries, most of these studies reported risk estimates on a limited number of observed cases with information available on primary treatment for breast cancer. For this reason, calendar year and time since diagnosis of the first breast cancer have been used as a proxy for treatment in some studies [6,9,12,17]. However, results reported by these studies are rather inconsistent as some support an increased risk during the first years after the breast cancer diagnosis [6,12] whilst others report that risk increases or remains high over time [9,17]. The aim of the current study is to examine the scientific evidence related to the risk of developing a second primary cancer after a breast cancer diagnosis for all sites combined, by age at breast cancer diagnosis and by time since breast cancer diagnosis, and to further combine the results of these studies by using meta-analysis.

Search strategy

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

A search was carried out to find relevant studies and reviews published up to 30 June 2013 (no starting date was fixed). The databases used were Pubmed and Embase.

Table 1 Search strategy for Medline and Embase (30 June 2013). Search strategy using MeSH terms in Medline No. Search 1 “Neoplasms, Multiple Primary/Epidemiology” [MESH] OR “Neoplasms, Multiple Primary/Prevention and control” [MESH] OR “Neoplasms, Second Primary/ Epidemiology” [MESH] OR “Neoplasms, Second Primary/Prevention and control” [MESH] AND “Breast Neoplasms/epidemiology” [MESH] N = 632 2 Limits: female, adults N = 488 Search strategy using keywords in Medline 1 “Breast cancer”[subheading] N = 265,587 2 “Second cancer” [subheading] N = 70,611 3 “Second malignancies” [subheading] N = 26,324 4 “Multiple primary cancer” [subheading] N = 41,254 5 “Multiple primary malignancies” [subheading] N = 39,289 6 #2 OR #3 OR #4 OR #5 N = 108,982 7 #1 AND #6 N = 15,109 8 Population-based N = 58,126 9 Risk N = 1,509,093 10 #7 AND #8 AND #9 N = 235 11 Limits: female, adults N = 192 Search strategy using keywords in Embase 1 Second AND (‘cancer’/exp OR cancer) OR second AND malignancies OR multiple AND primary AND (‘cancer’/exp OR cancer) OR multiple AND primary AND malignancies N = 4390 2 ‘breast’/exp OR breast AND (‘cancer’/exp OR cancer) N = 338,811 3 #1 AND #2 N = 854 4 #3 AND ‘human’/exp AND (‘breast cancer’/exp OR ‘multiple cancer’/exp OR ‘second cancer’/exp) AND ‘population based’ AND [embase]/lim N = 30

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

3

The following MeSH terms related to “second cancers” and “multiple primary cancers” and related subcategories were selected: Neoplasms/ Multiple Primary, Neoplasms/Second Primary and epidemiology. A number of key words (“second cancers” and “population-based”) were also used and combined in different databases (Table 1). The reference lists of all relevant articles were also examined.

second primary cancer diagnosis, the SIR and 95% confidence interval for all second cancers combined, SIR by age group, SIR by time since diagnosis, and second primary site with significant SIRs. The information was extracted from all the articles by the same person and the results, i.e. standardised incidence ratios (SIR), were obtained directly from the studies.

Study inclusion and exclusion criteria

Synthesis of results

The following article types were included: 1. Original articles on risk of subsequent primary cancers in females after a first primary invasive breast cancer; 2. Article type: systematic reviews and meta-analyses, clinical trials, population-based studies, hospital-based studies, or prospective cohort studies; 3. Articles written in English language. 4. Population characteristics: females and adults; 5. Results: risk estimates as Standardized Incidence Ratio (SIR) with respective 95% Confidence Interval (CI) for all cancers combined, and/or by age groups and time since diagnosis of the first primary breast cancer.

We considered Standardized Incidence Ratios (observed divided by expected cases in the general population) of second primary cancers in women previously diagnosed with a primary breast cancer. Results with a p N 0.05 were deemed not significant. Meta-analysis

In the case of multiple studies on the same population, the largest or most updated study was included.

Meta-analyses were performed for all second cancers combined, and separately by age at first breast cancer diagnosis (b50 years and ≥ 50 years), and by time since diagnosis of the first breast cancer (b10 years and ≥ 10 years). Although age groups and time intervals since diagnosis varied between the studies, the cut points set at 50 years for age at breast cancer diagnosis and 10 years for time since breast cancer diagnosis were the most commonly used. Only studies using the strata defined above were included in the stratified analyses. We extracted rate ratios and calculated standard errors (SEs), by applying the formula SE = √Observed/Expected2 [23]. Based on the SE, we estimated the weight of each study and pooled risk ratios in our analysis using both fixed-effects and random effects-models. We used the summary estimates from the random effects models of metaanalyses as the main results because they tend to give a more conservative estimate when between-study heterogeneity is present. Heterogeneity was estimated by calculation of the Cochran's Q statistic value. Sensitivity analysis consisted in excluding studies that did not report the MPT coding rules applied. Also, we evaluated the effect of the methodological quality of the studies on the estimates through subgroup analyses and meta-regression. Publication bias was investigated using funnel plots [24], and evaluated through Begg's [25] and Egger's [26] tests. We used Stata statistical software for the data analysis (release 12.0; College Station, TX). Finally, the PRISMA guidelines (27 item checklist and a four phase flow diagram) to report systematic reviews and meta-analyses were followed [27].

Data extraction

Results

Two independent reviewers (EMM and MRM) read, reviewed and selected the articles. In the case of disagreement, a third reviewer (MJS) decided whether or not an article would be included In order to evaluate the quality of the studies, the questionnaire proposed by the SIGN50 Scottish Intercollegiate Network 2013 was used [22]. This questionnaire assesses both the methodological and overall quality of the studies, classifying them as high, acceptable or low. In order to establish the levels of evidence of the studies, the methodology proposed in SIGN50, which sets out four main levels of evidence (1 to 4) was used. Sublevels are also distinguished depending on the methodological quality of the study. The overall quality of the evidence was assessed taking into account the design of the studies, their internal validity, consistency and the precision of results [22]. For each eligible study, we extracted the following data using a standardised form: first author's last name, publication year, country and centre were the data was derived from, design, study period and follow-up, study population with regard to inclusion and exclusion criteria, definition of second primary cancer, number of primary breast cancer cases and second cancer cases (observed and expected), person-years, mean age at first primary cancer diagnosis, mean age at

Results of the bibliographical search

Furthermore, we excluded studies that: 1. Used codification criteria other than the IARC/AICR coding rules for MPT, because countries following other coding rules (i.e. SEER) were not comparable. The IARC/AICR rules do not permit coding of more than 1 tumour in paired organs if the histology is identical. Thus, studies including contralateral breast cancer were excluded [21]. Studies that included second breast cancers of different histology with respect to the first primary breast cancer usually lacked specifications on whether the second breast cancer occurred in the same breast or was a contralateral breast cancer. Therefore, all studies including second breast cancers were excluded, unless risk estimates without second breast cancers were provided. Those studies that did not specify the coding rules were included if it was apparent that rules for accepting a second cancer as a new primary agreed with the IARC/AICR principle; 2. Examined risk only by breast cancer therapy, due to the nature of the multimodality treatment in cancer, except those that also reported non-treatment related risk estimates.

The defined search criteria in Pubmed and Embase retrieved 710 articles (Fig. 1). 16 articles, identified through manual review of reference lists, were added. 61 articles were ruled out because of duplication, leaving a total number of 665 articles for review. Based on the title and abstract we excluded 583 articles for different causes of exclusions. 82 articles were left for the “in extenso” analysis, of which 67 were further excluded: 11 studies did not use IARC/AICR coding rules for MPC, 10 studies only examined risk after a specific breast cancer treatment, 29 studies focused on one or a specific group of second primary cancer sites after the breast cancer diagnosis, 7 studies did not report SIR estimates and 3 studies included second primary cancers in the breast but did not clearly specify MPC rules nor SIR estimates without this second cancer site [28–30]. Moreover, of the studies identified, sample overlaps were noted in 7 studies. As shown in Table 2, the studies by Mellemkajer et al. [12,13] covered the same study populations (Norway, Finland and Denmark) but the study periods did not entirely overlap. The studies by Mellemkjaer et al. [12] and Prochazka et al. [31] included data from Sweden but the study periods covered by these studies did not coincide.

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

4

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Potentially relevant articles identified through PubMed/MEDLINE and EMBASE (n = 710)

Articles excluded based on title and abstract (n = 583):

Articles excluded due to duplication (n = 61)

Letters to the edictor and editorials (n = 10) Case-reports (n = 6) Narrative reviews (n = 38) Breast cancer (epidemiology/etiology) (n = 124) MPC after a first non breast cancer (n = 116) Breast cancer survival and prognosis (n= 74) Contraletaral BC (epidemiology) (n= 123) Second cancers after childhood cancer (n = 6) In situ first breast cancer cancer (n = 10) Treatment related second cancers (n = 16) Others (not related, irrelevant) (n = 45) Language (n = 15)

Secondary searches (manual review) (n = 16)

Articles selected for “in extenso” analysis (n = 82)

Articles ruled out based on text (n = 67) Overlapped population study/study period (n = 7) Not using MPT IARC/AICR rules (e.g. SEER) (n = 11) Not SIR or risk estimates (RR) reported (n = 7) Not SIR for total cancer reported, but for specific second cancer sites (n = 29) Included second cancer in the breast (n = 3) Treatment related second cancers (n = 10)

Articles selected (n= 15)

Fig. 1. Flow diagram showing the study search and selection procedure.

Overlaps were also detected between the studies by Mellemkjaer et al. [12,13] and the studies that included data from Slovenia [11], Finland [32], Denmark [33,34], Sweden [35], or from Denmark, Finland, Norway and Sweden [36]. All these 6 studies were therefore excluded [11,32–36]. The samples of the studies by Levi et al. [9,37], which included data from the Swiss cancer registries, also coincided; only the largest and most updated study was included [9]. The final selection was made up of 15 studies [6,7,9,10,12,13,17,18, 20,31,38–42], of which 13 were pooled in the meta-analysis [6,7,9,10, 12,17,18,20,38–42]. Thus, the results of two of the selected studies were not pooled [13,31] because their reported risk estimates were partly derived from the study populations (Sweden, Norway, Finland and Denmark) that were also included in the study by Mellemkjaer et al. [12]. The latter one had the largest sample size, as it included data from 13 cancer registries from Australia, Canada, Europe (including Sweden, Norway, Finland and Denmark) and Asia. Neither systematic reviews nor meta-analysis related to the topic in question were found in any of the databases. Description of the studies Characteristics of the selected studies are shown in Table 3. All of them were retrospective cohort studies. Most studies reported data from population-based cancer registries [6,7,9,10,12,13,18,20,31,

38,39,41,42]. Two studies were hospital-based [17,40] although cases were confirmed by cancer registry data in one of these studies [40]. The study population was always defined as women diagnosed with a first primary and invasive breast cancer. However, there were several differences in the definition of the first and second primary cancers: Only three studies detailed information on the proportion of histological confirmed first primary breast cancer cases [20,38, 39]. Previous primary cancers other than breast cancer were always excluded. Two studies specified that this was not applicable to nonmelanoma skin cancer as the first primary [18,20]. Second primary cancers were considered as those occurring after more than one month [9,13,31,41], two months [10,39], three months [12,38,42] or twelve months [6,7,17] since the first breast cancer diagnosis. Three studies did not mention any criteria to define (and exclude) synchronous primary cancers [18,20,40]. Cases dying within the first month since the breast cancer diagnosis were excluded in two studies [31,41]. Also, women dying in the first year after treatment for the first breast cancer were excluded in one study [17]. Second primary cancers included were either all cancer sites [12,41,42], all except those which are prone to metastases (liver, bone, skin, lung and brain and nervous system) [6], or all except non-melanoma skin cancer [7,9,10,13,17,18,20,31,38–40]. Thus, non-melanoma skin cancer as another second primary cancer site was considered in only three studies [12,41,42]. Nine studies mentioned the IARC/

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

1968–1992 (Singapore) 1970–1998 (Canada, British Columbia) 1970–1999 (Canada, Manitoba) 1967–1998 (Canada, Saskatchewan)

1943-1999 1953-2002

1953-2006

1953-1979

1943-2006

1943-1980 1943-1980

1972-1997 New South Wales 1943-1997 1953-1998

CR: Cancer Registry. FU: Follow-Up. Shaded studies were excluded because of overlaps by country and study period.

Levi et al. 1989 [37]

Volk and Pompe 1997 [11]

Prochazka et al. 2006 [31]

Teppo et al. 1985 [32] Storm et al. 1985 [33] Ewertz et al. 1985 [34] Dong et al. [35] Brown et al. 2007 [36] Levi et al. 2003 [9]

1974-1998 (Vaud and Neuchatel CR) 1974-1994 (Vaud CR)

1958-2000

1958-1996 1958-2002

1961-1985 (FU: 1994)

1953-2000

1953-2003 Mellemkjaer et al. 2011 [13]

Mellemkjaer et al. 2006 [12]

5

AICR coding rules for MPC when defining what was regarded as a second primary cancer [9,10,12,13,20,39–42], while the remaining studies did not. However, for some of them it could be established that these rules were followed [6,7,18,38,31]. A total number of five studies considered breast cancer as a second cancer site but provided risk estimates for all second cancers except breast cancer in the same site, and were therefore included [6,7,10,38,39]. The study population was confined to women aged over 15 years at the first breast cancer diagnosis in one study [10], or to women older than 20 years in two studies [13,40]. The other studies did not define any entry criteria by age at diagnosis. Mean age at first breast cancer and at second primary cancer diagnosis was about 50 years [17,18,40, 41] or about 60 years [10,39,42]. However, most of the studies did not report this data. Five studies addressed how treatment of the initial breast cancer affects second cancer risk [13,17,18,38,40]. There was one study examining the effect of stage at diagnosis of the first breast cancer on the risk of developing a second primary breast cancer [17]. No further studies focused on the impact of other clinical variables on the occurrence of second cancers. The methodological quality of the selected studies was deemed to be high in all except in those that were more prone to selection [10,17,18, 38] and/or misclassification bias [17,20]. Since all the studies were retrospective, their level of evidence was classified as “Level 2” Based on their methodological quality we judged most of them to be of low risk bias. The overall level of evidence was rated as “moderate”. Table 4 shows SIR estimates for second cancer risk, by age group at first breast cancer diagnosis and time since the diagnosis. Statistically significant increased second cancer risks were reported by eleven studies [9,10,12,13,17,20,30,38,40–42], with increased risks ranging from 9% to 40%. All except one study [9] estimated risks of second cancers by age. In general, SIR estimates were higher in younger women at breast cancer diagnosis, with risks increases between 13% and more than twofold. By time since diagnosis of the first primary breast cancer, it was reported that risk increased over time in six studies [9,12,17,18,20,38], but an opposite trend was observed in the other six studies that evaluated risk by this variable [13,20,39–42]. Of these, some studies supported that risk of second cancers was highest and statistically significant within the first year [13,20,40] while other supported an increased risk within the first five years [41,42] since the diagnosis of the first breast cancer. Second cancer sites with statistically significant increased SIR estimates differed largely between the studies. There was also no distinct pattern of second cancer sites by age groups.

1955-2000 1953-1999 1961-1998 1978–1998 (Zaragoza CR) 1960-1996 1961-1998

Spain Scotland Sweden Switzerland Author, publication year [reference]

Table 2 Overlapping studies by study period and country of origin.

Slovenia

Norway

Iceland

Finland

Denmark

Australia

North America

Asia

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Results of the meta-analyses Thirteen studies were included in the meta-analysis of second cancers [6,7,9,10,12,17,18,20,38–42]. Pooled estimates are shown in Fig. 2. The summary SIR estimated from the random effects model was 1.15 (95% CI: 1.04–1.27). Evidence of study heterogeneity was found based on the Q value (669.4, p b0.001). The influence of individual studies on the summary effect estimate showed that heterogeneity was driven by the smaller studies and most notably by the study by Evans [7]. The pooled estimate after excluding this study was 1.17 (95% CI: 1.10– 1.25) and 1.23 (95% CI: 1.22–124) for the random and fixed effects model, respectively. Heterogeneity decreased but was still apparent (Q = 136.8, p b0.001). The same results were observed when the studies with low methodological quality were excluded from the pooled analyses [10,17] (data not shown). Meta-regression analysis revealed that study quality did not influence the risk estimates (p = 0.38). Ten studies were pooled for the stratified analyses by age groups [6, 10,12,17,18,20,38,40–42]. Those studies with age groups set at 45 years were considered as of 50 years [12,38], while age groups set at 65 years were deemed not comparable and were therefore not included in these sub-analyses [39]. The summary SIR estimate was 1.51 (95% CI: 1.351.70) for women younger than 50 years (Q = 24.3, p = 0.004) and

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

6

N first breast cancer/N second cancer

Mean age at breast cancer diagnosis/at second cancer

Study population: Definition and inclusion criteria

9503/254

–

Women diagnosed with a first breast cancer (ICD-9: 174). 93.6% of the cases were histologically confirmed. Cases were identified through hospital reports and death certificates.

Retrospective cohort study 1968–1987

9678 7/206

–

Italy (Cancer Registries of Toscana, Ragusa, Romago)

Retrospective cohort study 1981–1992

5237/144

61.3 years

Rubino et al. 2000 [17]

France (Institute Gustave Roussy IGR)

Retrospective cohort study 1973–1992

4416/193

55 years

Tanaka et al. 2001 [40]

Japan (Osaka Medical Center for Cancer and Cardiovascular Diseases)

Retrospective cohort study 1970–1994 Follow-up until 1995

2786/117

50.9 years

Evans et al. 2001 [7]

England (The Thames Cancer Registry)

Retrospective cohort study 1961–1995 Follow-up until 1996

Author, publication year [Reference]

Country (Centers where the data was derived from)

Design Study period Follow-up

Murakami et al. 1987 [38]

Japan (Osaka Cancer Registry)

Retrospective cohort study 1965–1982 Follow-up until 1983

Brenner et al. 1993 [6]

Germany (Cancer Registry of Saarland)

Buiatti et al. 1997 [39]

145,677/4470

–

Definition of second cancers: inclusion criteria

Methdological qualitya and level of evidenceb

Second cancers diagnosed after more than 3 months since the first Acceptable breast cancer diagnosis. 72.7% of the second cancers had Selection bias: Women were histological confirmation. assumed to be alive until the end All second cancers except non-melanoma skin cancer were of the study period. included. Level: 2+ SIRs reported for: Bucal cavity, oesophagus, stomach, colon, rectum, liver, pancreas, lung, cervix uteri, corpus uteri, ovary, urinary bladder, thyroid gland, leukaemia. MPT coding rules: not specified, but the Osaka cancer registry follow the IARC/AICR coding rules for MPC. High Women diagnosed with a first primary Second cancers diagnosed after more than 12 months since the first breast cancer diagnosis to reduce misclassification of second Level: 2++ breast cancer (ICD-9: 174). cancers and metastases. Cases were identified through notifications made by physicians. All second cancers (ICD-9: 140–208), except skin cancer (172– 173), cancers of unspecified location (195–199) and cancers of preferred locations of breast cancer metastases (liver (155), lung (162), bone (170), brain and nervous system (191–192)), were included. SIRs reported for: Stomach, colon, rectum, gallbladder and bile conducts, pancreas, cervix uteri, corpus uteri, ovaries, urinary bladder, kidney, lymphomas and leukaemias. MPT coding rules: not specified, but the German cancer registries follow the IARC/AICR coding rules for MPC. Women diagnosed with a first primary Second cancers diagnosed after more than 2 months since the first High Level: 2++ breast cancer and residing in the areas breast cancer diagnosis. covered by the three registries. All second cancers except non-melanoma skin cancer were 87.3% of the cases were histologically included. confirmed. SIRs reported for: Rectum, stomach, colon, lung, cervix uteri, corpus uteri, bladder, kidney, lymphomas, ovary, skin melanoma, thyroid gland. MPT coding rules: IARC/AIRC 92% of the second cancer cases were histologically confirmed. Women initially treated at the IGR for a Second cancers diagnosed after more than 1 year since the first Low breast cancer. Women dying in the first breast cancer diagnosis. Misclassification bias might be year after treatment were excluded. All second cancers except non-melanoma skin cancer were possible as it is not specified included. whether first primary cancers SIRs reported for: Oral cavity, digestive tract, oesophagus, were included. Selection bias: a stomach, colon, rectum, liver, gall bladder, pancreas, larynx, lung high number of women were lost and bronchus, uterus, ovaries, bladder, kidney, melanoma, to follow-up. Level: 2nervous system, thyroid gland, bone, soft tissue. MPT coding rules: not specified. All second cancer cases were histologically confirmed. Women aged between 20 and 75 years, It is not mentioned whether synchronous cancers were excluded. High diagnosed with a first primary breast All second cancers except non-melanoma skin cancer were Level: 2++ cancer (ICD-9: 174) at the Osaka Medical included. Center for Cancer and Cardiovascular SIRs reported for: Stomach, colon, rectum, liver, biliary tract, Diseases. pancreas, lung, uterus, cervix, uterus corpus, ovary, bladder, Cases were identified through the hospital thyroid gland, non-hodgkin's lymphoma. MPT coding rules: IARC/AICR cancer registry, and confirmed by the Osaka Cancer Registry. All second cancer cases were identified by the Osaka Cancer Registry. Women diagnosed with a first primary Second cancers diagnosed after more than 1 year since the first Acceptable breast cancer. breast cancer diagnosis. Those diagnosed on the same day of the Selection bias: Exclusions applied Women with missing date of birth, date of first breast cancer diagnosis were excluded. to women with unknown diagnosis or on residence at the time of All second cancers except non-melanoma skin cancer were information on date of diagnosis or diagnoses were excluded. included. on residence. SIRs reported for: Tongue, mouth, oesophagus, stomach, colon, Level: 2+

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

Table 3 Characteristics of the selected studies.

Women diagnosed with a first primary breast cancer and residing in the area covered by the two cancer registries.

152,586/9758

–

Women diagnosed with a histologically confirmed invasive first primary breast cancer. Women dying within 1 month after the breast cancer diagnosis were excluded.

Retrospective cohort study 1943–2000

525,527/31,399

–

Women diagnosed with a first primary breast (ICD-9: 174) cancer residing in areas covered by the thirteen populationbased cancer registries.

Taiwan (Taiwan National Cancer Registry)

Retrospective cohort study 1979–2003

53,783/1085

50.35 years/ 60.56 years

Women diagnosed with a first primary breast cancer residing in Taiwan. Women dying within 1 month after the breast cancer diagnosis were excluded.

The Netherlands (Comprehensive Cancer Centers of the North (Groningen), Amsterdam and South (Eindhoven) Netherlands, within the Dutch Network of cancer registries).

Retrospective cohort study 1989–2003 Follow-up until 2004 or 2005, depending on the center.

58,068/2578

–

Women diagnosed with a first primary breast cancer, with no previous cancer other than non-melanoma skin cancer, in the regions of the Comprehensive Cancer Centers. All unknown primary cancers were excluded. 99% of the cases were histologically confirmed. Synchronous or metachronous cancer cases were not excluded.

Switzerland (The Vaud and Neuchâtel Cancer Registries)

Retrospective cohort study 1974–1998

Prochazka et al. 2006 [31]

Sweden (Swedish Cancer Registry)

Retrospective cohort study 1958–2000

Mellemkjaer et al. 2006 [12]

13 Cancer Registries: Europe (Denmark, Finland, Iceland, Norway, Slovenia, Sweden) Canada, Australia, Singapur

Lee et al. 2008 [41]

Shaapveld et al. 2008 [20]

9729/443

High Level: 2++

High Level: 2++

High Level: 2++

High Level: 2++

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

–

Levi et al. 2003 [9]

rectum, liver, gallbladder, pancreas, larynx, lung and bronchus, bone, connective tissue, skin melanoma, cervix uteri, corpus uteri, ovary, bladder, kidney, brain/nervous system, thyroid gland, NHL, multiple myeloma, myeloid leukaemia. MPT coding rules: new primary tumour occurring at a different anatomical site and of a different histological type from the first tumour, or to be stated explicitly as being a new tumour by the treating clinician. Second cancers diagnosed after more than 1 month since the first breast cancer diagnosis. All second cancers except non-melanoma skin cancer were included. SIRs reported for: Mouth or pharynx, oesophagus, stomach, colorectum, gallbladder, pancreas, lung, soft tissue, skin melanoma, cervix uteri, corpus uteri, ovary other female genital organs, bladder, kidney, thyroid gland, non-hodgkin's lymphomas, multiple myelomas, leukaemias. MPT coding rules: IARC/AICR. Second cancers diagnosed after more than 1 month since the first breast cancer diagnosis. All second cancers except non-melanoma skin cancer were included. SIRs reported for: Upper aerodigestive tract, salivary glands, oesophagus, stomach, small intestine, colorectum, liver, pancreas, lung, cervix, endometrium, ovary, kidney, urinary organs, melanoma, nervous system, thyroid gland, endocrine glands, bone, connective tissue, non-hodgkin's lymphoma, hodgkin's disease, multiple myeloma, plasmocytoma, leukaemia (Acute lymphatic leukaemia, chronic lymphatic leukaemia, acute myeloid leukaemia, chronic myeloid leukaemia). MPT coding rules: not specified, but the Swedish cancer registry follow the IARC/AICR coding rules for MPC. Second cancers diagnosed after more than 3 months since the first breast cancer diagnosis. All second cancers including non-melanoma skin cancer (partially collected in the participating cancer registries) were included. SIRs reported for: Oral cavity, pharynx, oesophagus, stomach, small intestine, colorectal, liver, pancreas, larynx, lung, bone, soft tissue, sarcoma, melanoma, non-melanoma skin, corpus uteri, ovary, bladder, kidney, brain, nervous system, thyroid gland, nonhodgkin's lymphoma, leukaemia, myeloid leukaemia. MPT coding rules: IARC/AICR Second cancers diagnosed after more than 1 month since the first breast cancer diagnosis. All second cancers including non-melanoma skin cancer were included. SIRs reported for: Bone, corpus uteri, ovary, non melanoma skin, thyroid gland, head and neck, small intestine, colon and rectum, liver, pancreas, lung, thymus, sarcoma, cervix uteri, urinary bladder, kidney and other urinary organs. MPT coding rules: IARC/AICR It is not mentioned whether synchronous cancers were excluded. All second cancers except non-melanoma skin cancer were included. SIRs reported for: Head and neck, thyroid gland, esophagus, stomach, pancreas, gall bladder, extrahepatic bile ducts, colon, rectum, anus, lung, soft tissue, sarcomas, melanoma of skin, ovary, uterus, cervix, vulva, kidney, bladder, brain, leukaemia, acute myeloid leukaemia, non-Hodgkin's lymphoma, multiple myeloma. Meningiomas, myelodysplastic syndrome and polycythemia as second cancers were excluded. If a subsequent cancer was diagnosed after a non-melanoma skin cancer, this cancer was included as a first occurring second cancer. MPT coding rules: IARC/AICR

Acceptable Misclassification bias might be possible because either synchronous or metachronous, was considered as the first primary cancer. Level: 2+

7

(continued on next page)

8

Author, publication year [Reference]

Country (Centers where the data was derived from)

Design Study period Follow-up

N first breast cancer/N second cancer

Anderson et al. 2008 [18]

Denmark (populationbased Danish Breast Cancer Cooperative Group)

Retrospective cohort study 1977–2001 Follow-up: until 2002

53,418/1993

Cluze et al. 2009 [10]

France (Cancer Registry of Isère)

Retrospective cohort study 1989–1997

5663/144

Mellemkjaer et al. 2011 [13]

Cancer Registries: Denmark (1943–2006), Norway (1953–2003), Finland (1953–2006)

Molina-Montes et al. 2013 [42]

Spain (Granada Cancer Registry)

Retrospetive cohort study 1943–2006 Follow-up until 2006 for Denmark and Finland and 2007 for Norway. Retrospective cohort study 1985–2007

Mean age at breast cancer diagnosis/at second cancer

Study population: Definition and inclusion criteria

Definition of second cancers: inclusion criteria

Methdological qualitya and level of evidenceb

56 years

Women diagnosed with a first primary breast cancer, treated according to the guidelines of DBCG (Danish Breast Cancer Cooperative Group). Women with a previous cancer, excluding nonmelanoma skin cancer, were excluded. Cases were identified through the DBCG, and confirmed by the Danish Cancer Registry.

Acceptable Selection bias because only women who were treated and followed by the DBCG were included. Level: 2+

59.9 years

Women older than 15 years, diagnosed with an invasive primary breast cancer. Women lost to follow-up were excluded at the date of their last contact.

Follow-up of second cancers started one year after the first breast cancer diagnosis (second cancers evaluated in one year survivors of breast cancer). It is not mentioned whether synchronous cancers were excluded. All second cancers except non-melanoma skin cancer were included. SIRs reported for: Lip, tongue, salivary glands, mouth, pharynx, oesophagus, stomach, small intestine, colon, rectum, liver, gallbladder, pancreas, nose, sinuses, larynx, lung, pleura, cervix, corpus uteri, ovary, kidney, bladder, melanoma of skin, eye, brain, nervous system, thyroid gland, bone, soft tissues, non-hodgkins lymphoma, hodgkins disease, multiple myeloma, acute leukaemia, other leukaemia. MPT coding rules: not specified, but MPT were ascertained through the Danish Cancer Registry and NORDCAN database, which follow the IARC/AICR coding rules for MPC. All second cancer cases were identified by the Danish Cancer Registry. Second cancers diagnosed after more than 2 months since the first breast cancer diagnosis. Only second cancers occurred within 5 years since the breast cancer diagnosis were considered. All second cancers except non-melanoma skin cancer were included. SIRs reported for: Upper aerodigestive tract, oesophagus, stomach, small intestine, colon, rectum, liver, pancreas, other digestive organs, lung, other thoracic organs, bone, cartilage, skin, soft tissues, female genitals, kidney, bladder, central nervous system, eye, endocrine, acute myeloid leukaemia, other haemopathies. MPT coding rules: IARC/AICR. Second cancers diagnosed after more than 1 month since the first breast cancer diagnosis. All second cancers except non-melanoma skin cancer were included. SIRs reported for: Salivary glands, esophagus, lung, pleura, bone, connective tissue, thyroid gland, leukaemia, endometrial, ovarian, mouth, pharynx, liver, larynx colorectum, gallbladder and bile ducts, pancreas, kidney. MPT coding rules: IARC/AICR. Second cancers diagnosed after more than 3 months since the first breast cancer diagnosis. All second cancers including non-melanoma skin cancer were included. SIRs reported for: Endometrium, colon, rectum, stomach, ovary, thyroid gland, skin non-melanoma, kidney, bladder, hematologic malignancies. MPT coding rules: IARC/AICR.

304,703/23,304 –

5897/314

58.59 years/ 60.91 years

Women older than 20 years, diagnosed with invasive first primary breast cancer and residing in the areas covered by the three registries.

Woman diagnosed with an invasive first primary breast cancer (ICD-O3: C50). Women whose first primary cancer diagnosis and death were recorded simultaneously and synchronous first primary cancers were excluded.

Low Acceptable Selection bias: Women lost to follow-up were excluded. Level: 2-

High Level: 2++

High Level: 2++

Study inclusion criteria were: 1) Studies on second primary cancers after a diagnosis of a first primary breast cancer; 2) Studies using the IARC/AICR coding rules for Multiple Primary Cancers (MPT). 3) Studies not examining risk solely by a specific breast cancer therapy. a Although the methodological quality of some studies was classified as high, it can be considered as acceptable according to the SIGN50 methodology checklist, because retrospective cohort studies are generally regarded as weaker designs [22]. b Level of evidence is set to “2” for cohort studies. Depending on the risk of bias, three sublevels are distinguished: low risk (++), moderate risk (+), and high risk (−) of bias.

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

Table 3 (continued)

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

1.11 (95% CI: 1.02–1.21) for women aged ≥ 50 years (Q = 115.6, p b0.001) at breast cancer diagnosis. The fixed effects model provided SIR estimates of similar magnitude (b50 years: 1.59; 95% CI: 1.52–1.67 and ≥50 years: 1.18; 95% CI: 1.16–1.19, respectively). Only five studies could be combined for the stratified analyses by time since diagnosis [12,17,18,38,40], either because time intervals differed or due to lack of data. With respect to the general population, women diagnosed with breast cancer have a 19% higher risk (95% CI: 1.06–1.33) of developing a second cancer within the first 10 years after their first breast cancer diagnosis (Q = 37.6, p b 0.001). After more than 10 years, this risk remained increased (SIR: 1.26; 95% CI: 1.05–1.52; Q = 35.8, p b0.001). Similar estimates were obtained when fixed effects models were considered (SIR b10 years = 1.16; 95% CI: 1.14–1.18 and SIR ≥10 years = 1.39; 95% CI: 1.35–1.42). No visible changes in the pooled risk estimates were noted after excluding studies not specifying the use of MPT coding rules (data not shown). We found no evidence of publication bias of any of our estimates based on funnel plots or on the Begg and Egger test statistics (data not shown). Discussion To the best of our knowledge, this systematic review and metaanalysis, conducted with all the studies available, is the first one to assess the risk of second primary cancers among women previously diagnosed with a first primary breast cancer. Our results show that these women have a 17% increase in the risk of developing a new primary non-breast cancer in comparison with women without cancer of the general population. This excess in risk is likely to be associated with treatment modalities for breast cancer, shared genetic predisposition or shared environmental risk factors [5]. All of the selected studies were retrospective cohort studies, either derived from cancer registries (population-based) or from hospitalbased settings. Although the methodological quality of the studies was overall high, it can be considered as intermediate according to the SIGN50 methodology checklist [22], due to the well known associated biases of retrospective cohort studies, which is driven by losses to follow-up and reliability of data collected about events that occurred in the past [43]. Women included in the studies were representative of the general population, at least for the studies that used populationbased cancer registry data [6,7,9,10,12,13,18,20,31,38,39,41,42]. The studies varied with respect to number of cases, selection criteria, definition of multiple primary cancers applied to ascertain second primary cancers, and the second cancer sites included. Despite a restrictive definition for second primary cancers in most studies, the definition of synchronous tumours varied between the studies because the IARC/ AICR coding rules do not consider time since initial diagnosis to identify multiple primary cancers [21]. Few studies [9,10,13,31,39,41] adopted the definition of synchronous cancers as those diagnosed within 2 months of the first primary breast cancer [39,43]. Moreover, some studies did not account for the synchronous development of two or more primary adenocarcinomas [20,40]. Because some of the studies included in this analysis were limited in sample size and statistical power, their effect estimates for second cancer risk were large but not statistically significant, suggesting that a meta-analysis and presentation of summary statistics was appropriate. The study by Mellemkjaer et al. [12] contributed with 81.2% to the total weight of the included studies. This was the largest one, including over 30 thousand second cancer cases provided by 13 population-based cancer registries worldwide. The risk estimates reported by this study (SIR = 1.25; 95% CI: 1.24–1.26) could be therefore considered the most reliable ones. Nevertheless, it is important to assess the methodological quality concerning risk of bias, regardless of the results reported or sample size of all the selected studies [44]. The results of our meta-analysis suggest a 17% increase in second cancer risk in women who have been diagnosed with a first primary

9

breast cancer with respect to the general population when a random effects model is considered, or a 23% increased risk when the fixed effects model estimate is taken into account. As such, random and fixed effects models varied to some extent depending on the weight given to smaller studies (random effects) or to the biggest study (fixed effects) [45]. Due to the large heterogeneity observed between the selected studies, we considered estimates from random effects models as our main result. Despite this heterogeneity, the consistency of an increased risk of second cancers in women diagnosed with a first primary breast cancer across studies confirms our results. The differences among studies with regard to the definition of new primary cancers, inclusion criteria, patterns of treatment of breast cancer and follow-up care program for breast cancer patients, hindered their comparability and may have driven heterogeneity between the studies. Moreover, the studies included are drawn from a range of populations in which the effect size also varies. This variability in the estimate of effects was driven by all smaller studies, although one study contributed more notably to it [7]. The source of this heterogeneity can not readily be explained but it is possible that it was caused by the overall significant decrease in second cancers reported in this study [7]. As Evans and co-authors reported, under-ascertainment of second cancers might be plausible in women diagnosed with a first primary breast cancer at older ages. For instance, breast cancer diagnosed at age 50–84 years resulted in a statistically significant decreased risk of second cancers in this study. This study was therefore excluded from the pooled analyses [7]. The study by Mellemkjaer et al. [12] reported data with a high level of precision due to the large number of first and second primary cancer cases. Moreover, similar criteria and definitions were applied to the populations included in this study. Despite the consistency of risk estimates reported in this study, the results were related to women diagnosed with breast cancer from different geographical areas and over a long period of time. Differences in breast cancer treatment protocols between these areas and over time may have consequently affected survival and occurrence of second cancers. This study assumed that differences in the pattern of breast cancer treatment and surveillance did not affect the overall pattern of second cancers. Since metastastic sites of breast cancer, such as liver and brain, were not frequent second cancer sites, it was considered that misclassifications of metastases as new primary cancers were unlikely [12]. Among the selected studies, only one of them excluded the most common sites of breast cancer metastases [6]. In the remaining ones, it was presumed that second cancers were all new primary tumours. Nonetheless, misclassification of second cancers being metastases from the primary cancer can be considered implausible given that all the selected studies follow the IARC/AICR coding rules for MPT [21]. The selected studies were derived from different populations and study periods, with different treatment regimens for breast cancer. Therefore, the estimated risk of second primary cancers in women with a first primary cancer may also not correspond to that of women who are currently diagnosed with breast cancer. We did not include studies considering second cancers in the breast, unless it was manifest that the study used the IARC/AICR rules for coding of MPT and that estimates without second cancers in the breast could be derived from the studies. The risk of second cancers in the breast after a first primary cancer in the same location has been already addressed in several studies [8,28–30,46,47]. It is well established that second primary breast cancer after a first primary breast tumour is elevated, especially among women with family history of breast cancer [48, 49], and among hormone receptor-negative first tumours [49]. Most of these studies are derived from the SEER database, as contralateral breast cancer is accounted for in SEER registry system. On the other hand, several studies have evaluated the risk of a second primary cancer in a specific site, such as the ovary, thyroid gland, colorectal, lung, among others [50–53]. They were not selected for the current review because they did not provide SIR estimates for cancer risk overall. Among the studies included in this review, there were also many cancer sites reported to

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

10

Study [Reference]

Observed Second Cancer/Expected Person-years (PY)

SIR (95% CI)

Standard SIR by age group error SE (Observed/Expected)

Murakami et al. [38]

254/190 (excluding second breast cancer) PY = 53,738

1.3 **

0.084

b45 y: 2.0 ** (57/28) 45–54 y: 1.5 ** (75/52) ≥55 y: 1.1 (122/110)

Brenner et al. [6]

206/246 (excluding second breast cancer) PY = 43,642.25

0.84

0.058

b50 y: 1.13 (29/26) ≥50 y: 0.8 (177/221)

Buiatti et al. [39]

144/138 PY = 20,537

1.04 (0.88–1.22)

0.087

b65 y: 1.58 * (63/40) ≥65 y: 0.82 (81/99)

Rubino et al. [17]

193/136 PY = 33,044

1.40 (1.20–1.60) ***

0.102

N50 y: 2.1 (1.6–2.6) *** (71/34) ≥50 y: 1.2 (1.0–1.4) * (122/102)

Tanaka et al. [40]

117/90 PY = 24,025

1.30 (1.10–1.60) *

0.120

b50 y: 1.4 (0.9–1.9) (37/26) ≥50 y: 1.3 (1.0–1.6) * (80/61)

Evans et al. [7]

4470/5010 PY = 1,850,098

0.89

0.013

Levi et al. [9]

443/389 PY = 61,834

1.14 (1.04–1.25) *

0.054

b50 y: 1.21 (1.13–1.31) *** (710/587) ≥50 y: 0.85 (0.83–0.88) *** (3760/4423) –

Prochazka et al. [31]

9758/7998 PY = 1,599,680

1.22 (1.20–1.24) *

0.012

b50 y: 1.64 (1.56–1.72) * (1692/1032) ≥50 y: 1.20 (1.17–1.23) * (6405/5337)

Mellemkjaer 31,399/25,119 et al. [12] PY = 3,784,660

1.25 (1.24–1.26) *

0.007

≤45 y: 1.68 (1.63–1.73) * (3682/2192)

SIR by time since diagnosis (Observed/Expected)

Statistically significant SIRs for second cancer sites

b1 y: 1.4 (38/27)) 1–4 y: 1.4 ** (111/79) 5–9 y: 1.1 (56/51) ≥10 y: 1.6 ** (49/31) b1 y 1–5 y N5 y SIR and Number of observed or expected cases were not provided.

All ages: buccal cavity, stomach, colon and thyroid gland. 50 y: ovary and skin non-melanoma. ≥50 y: endometrium, skin non-melanoma. The effect of chemotherapy and radiotherapy (in the first 4 months after the diagnosis) on the risk of second cancers was also examined. Only radiation was associated with an increased risk of second thyroid cancer.

b1 y: 1.09 (30/27) 1–5 y: 1.06 (101/95) N5 y: 0.84 (13/15) 1–10 y: 1.3 (1.1–1.6) *** (124/92) N10 y: 1.5 (1.2–1.9) *** (69/45)

0 y: 3.0 ** (22/7) 1–4 y: 1.3 (35/27) 5–9 y: 1.2 (32/27) ≥10 y: 0.9 (28/31) Provided for intervals of 0–4, 5–9, 10–14 and ≥10 years and specific cancer second sites, but not for all cancers. b5 y: 1.04 (0.90–1.19) (212/204) ≥5 y: 1.25 (1.09–1.42) * (231/185) –

b1 y: 1.04 (1.01–1.08) * (2929/2816)

All ages: SIR for total second cancers, or for a specific second cancer site, was not statistically significant. b50 y: ovary ≥50 y: SIR for total second cancers, or for a specific second cancer site, was not statistically significant. Lack of information on treatment. All ages: rectum Lack of information on treatment.

All ages: uterus, ovary, kidney, melanoma, soft tissue, leukaemia. The effect of radiotherapy combined with chemotherapy or of radiotherapy alone on the risk of second cancers was also examined. Only radiation was associated with an increased risk of all second cancers and specifically with leukemia, lymphoma, soft tissue, melanoma and uterus. Chemotherapy was associated with an increased risk of kidney cancer only. All ages: Ovary, thyroid and non-Hodgkin's lymphoma (NHL). b50 y: thyroid, NHL ≥50 y: all second cancers, but not specific second cancer sites. The effect of radiotherapy, chemotherapy or of hormonal therapy on the risk of second cancers was also examined. Statistically significant increases in risk were observed for NHL and chemotherapy, for ovary and hormonal therapy, and for bladder and radiotherapy. All ages: not provided. b50 y: oesophagus, stomach, lung and bronchus, corpus uteri, ovary, myeloid leukaemia. ≥50 y: Corpus uteri, myeloid leukaemia. Lack of information on treatment. All ages: soft tissue, corpus uteri

All: salivary glands, oesophagus, stomach, small intestine, colorectum, pancreas, lung, endometrium, ovary, Kidney, melanoma, nervous system, thyroid, endocrine glands, bone, connective tissue, non-Hodgkin's lymphoma, leukaemia. b50y: oesophagus, stomach, colorectum, pancreas, lung, endometrium, ovary, kidney, melanoma, thyroid, endocrine glands, connective tissue, thorax and upper limbs, nonHodgkin's lymphoma, acute myeloid leukaemia, chronic myeloid leukaemia. ≥50 y: oesophagus, stomach, colorectum, lung, endometrium, kidney, melanoma, nervous system, thyroid, connective tissue, thorax and upper limbs, acute lymphatic leukaemia, acute myeloid leukaemia, chronic myeloid leukaemia Lack of information on treatment. All ages: oral cavity and pharynx, esophagus, stomach, small intestine, colorectal, pancreas, larynx, lung, bone, soft tissue, melanoma, non-melanoma, corpus uteri, ovary,

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

Table 4 Characteristics of studies on second cancers after a first primary breast cancer and results according to SIRs.

1–9 y: 1.20 (1.18–1.22) * (17,489/14,574) ≥10 y: 1.42 (1.39–1.44) * (10,981/7733)

Lee et al. [41]

1085/993 PY = 290,966

1.09 (1.03–1.16) *

0.033

b50 y: 1.43 (1.29–1.58) * (394/275) ≥50 y: 0.96 (0.89–1.04) (691/718)

Higher risk in first 5 y: b50 y: Bone, corpus uteri, ovary, thyroid, leukemia or lymphoma ≥50 y: Corpus uteri, ovary

Shaapveld et al. [20]

2,578/2,100 PY = 362,470

1.22 (1.17–1.27) *

0.024

b50 y: 1.75 * (384/219) ≥50 y:1.2 * (2194/1828)

Andersson et al. [18]

1993/1916 PY = 256,563

1.04 (0.99–1.08)

0.023

b50 y: 1.3 (1.2–1.5) * (420/323) 50–59 y: 1.1 (1.0–1.1) * (604/549) 60–69 y: 1.0 (0.9–1.0) (762/762) 70–89 y: 0.9 (0.7–1.0) (207/230)

b1 y: highest, about 1.6 * 1–4 y: about 1.2 * 5–9 y: about 1.2 * ≥10 y: not statistically significant Number of observed or expected cases was not provided. 1–9 y: 1.0 (1.0–1.1) * (1370/1370) 10–19 y: 1.1 (1.0–1.2) * (568/516) ≥20 y: 1.2 (0.9–1.5) (55/46)

Cluze et al. [10]

144/110 PY = 22,400.5

1.30 *

0.109

–

Mellemkjaer et al. [13]

23,304/20,330 PY = 4,069,283

1.15 (1.13–1.16) *

0.007

b50 y: 2.6 (20/7.5) 50–74 y: 1.35 (84/62) ≥75 y: 1.0 (40/40) b40 y: 1.83 (1.73–1.94) * (1219/666) 40–49 y: 1.38 (1.34–1.42) * (4707/3411) 50–59 y: 1.19 (1.16–1.22) * (5674/4768) 60–69 y: 1.09 (1.07–1.12) * (6114/5609) ≥70 y: 0.95 (0.92–0.97) (5,590/5,884)

MolinaMontes et al. [42]

314/225.8 PY = 37,605

1.39 (1.23–1.54) *

0.078

b50 y: 1.96 (1.48–2.44) * (63/32) ≥50 y: 1.29 (1.13–1.45) * (251/195)

≤5 y: 3.45 (2.92–3.98) * (162/50) N5 y: 0.85 (0.71–0.98) (152/179)

b1 y: 1.30 (1.24–1.36) * 1–4 y: 1.05 (1.02–1.07) * 5–9 y: 1.17 (1.14–1.20) * ≥10 y: 1.17 (1.15–1.20) * Number of observed or expected cases was not provided.

other female organs, bladder, kidney, thyroid, non-hodgkin lymphoma, leukaemia, myeloid leukaemia ≤45 y: esophagus, stomach, colorectal, pancreas, lung, bone, soft tissue, melanoma, nonmelanoma, corpus uteri, ovary, other female organs, bladder, kidney, thyroid, leukaemia, myeloid leukaemia 46–55 y: oral cavity and pharynx, esophagus, stomach, colorectal, pancreas, lung, bone, soft tissue, melanoma, non-melanoma, corpus uteri, ovary, other female organs, bladder, kidney, thyroid, non-hodgkin lymphoma, leukaemia, myeloid leukaemia ≥56 y: oral cavity and pharynx, esophagus, stomach, small intestine, colorectal, larynx, lung, soft tissue, melanoma, non-melanoma, corpus uteri, ovary, other female organs, bladder, kidney, thyroid, non-hodgkin lymphoma, leukaemia, myeloid leukaemia Lack of information on treatment. All: bone, corpus uteri, ovary, non-melanoma, and thyroid b50 y: bone, corpus uteri, ovary, esophageal, kidney, lung, non-melanoma, leukemia or lymphoma ≥50 y: corpus uteri, ovary, cervix uteri, non-melanoma Lack of information on treatment. All ages: esophagus, stomach, colon, rectum, lung, soft tissue, melanoma, uterus, ovary, bladder, kidney, NHL, AML. b50 y: esophagus, rectum, ovary, uterus, soft tissue, melanoma, and AML 50–69 y: melanoma, uterus, ovary, bladder, kidney, AML ≥70 y: esophagus, soft tissue, melanoma, uterus, NHL Lack of information on treatment. All ages: leukaemia, soft tissues, ovary, corpus uteri, rectum b50 y: stomach, lung, ovary, bladder, leukaemia 50–59 y: ovary, leukaemia 60–69 y: pharynx, corpus uteri, soft tissues 70–89 y: none The effect of adjuvant treatment (chemotherapy, tamoxifen and radiotherapy) on second cancers was also examined. Statistically significant increases in risk were observed for ovary, soft tissue and leukaemia and radiotherapy, corpus uteri and tamoxifen, and ovary, bladder and leukaemia and chemotherapy. All ages: Kidney, acute myeloid leukaemia b50 y: Female genitals 50–74 y: Kidney, acute myeloid leukaemia ≥75 y: none Lack of information on treatment. All ages: esophagus, lung, bone, connective tissue, thyroid, leukaemia, endometrium, ovary, liver, colon and rectum, pancreas, kidney, stomach, melanoma, lymphoma. b40 y: esophagus, lung, bone, connective tissue, thyroid, leukaemia, endometrium, ovary, colon and rectum, pancreas, kidney, stomach, eye. 40–49 y: esophagus, lung, connective tissue, leukaemia, endometrium, ovary, stomach 50–59 y: esophagus, connective tissue, endometrium, ovary, colon and rectum, stomach 60–069 y: esophagus, connective tissue, endometrium, kidney, stomach ≥70 y: endometrium The effect of treatment (chemotherapy, tamoxifen and radiotherapy) on second cancers was also examined. Statistically significant increases in risk were observed for: Radiation: thyroid, esophagus, lung Chemotherapy: leukaemia Tamoxifen: endometrium BRCA-gene related sites (ovary) SIR estimates were adjusted for calendar period, length of follow-up and country. All ages: endometrium, non-melanoma skin. b50 y: ovary and skin non-melanoma. ≥50 y: endometrium, skin non-melanoma. Lack of information on treatment.

11

Study inclusion criteria were: 1) Studies on second primary cancers after a diagnosis of a first primary breast cancer; 2) Studies using the IARC/AICR coding rules for Multiple Primary Cancers (MPT). 3) Studies not examining risk solely by a specific breast cancer therapy. Statistically significant SIR estimates set at: *** p b0.001; ** p b0.01; * p b0.05.

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

46–55 y: 1.36 (1.33–1.40) * (7,035/5,173) ≥56 y: 1.16 (1.15–1.18) * (20,682/17,829)

12

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

Murakami Brenner Buatti Rubino Tanaka Levi Mellemkjaer Lee Schaapveld Andersson Cluze Molina-Montes

Combined

.8

1

2.5 SIR

A SIR = 1.17; 95% CI: 1.10-1.25 Test for heterogeneity: Q = 136.827 (p=<0.001)

Murakami

Murakami

Brenner

Brenner

Rubino

Rubino

Tanaka

Tanaka

Mellemkjaer

Mellemkjaer

Lee

Lee

Schaapveld

Schaapveld

Andersson

Andersson

Cluze

Cluze

Molina-Montes

Molina-Montes

Combined

Combined

.8

1

2.5 SIR

1

.8

2.5 SIR

< 50 years

≥ 50 years

B

SIR = 1.11; 95% CI: 1.02-1.21 Test for heterogeneity: Q = 115.644 (p=<0.001)

SIR = 1.51; 95% CI: 1.35-1.70 Test for heterogeneity: Q = 24.328 (p=0.004)

Murakami

Murakami

Rubino

Rubino

Tanaka

Tanaka

Mellemkjaer

Mellemkjaer

Anderson

Anderson

Combined

Combined

.8

1

2.5 SIR

< 10 years

.8

1

2.5 SIR

≥ 10 years

C SIR = 1.19; 95% CI: 1.06-1.33 Test for heterogeneity: Q = 37.641 (p=<0.001)

SIR = 1.26; 95% CI: 1.05-1.52 Test for heterogeneity: Q = 35.813 (p=<0.001)

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

occur in excess. There was, however, no consistent pattern of increased second cancer risks except for gynaecological cancers. As such, ovarian and endometrial cancers were consistently associated with an increased risk after the first primary breast cancer in most studies [10,12,13,18,20, 31,38,41,42], which is further supported by some studies that specifically evaluated this risk [28,53]. According to the selected studies, women diagnosed with breast cancer at premenopausal ages (b50 years) exhibit a higher risk of second cancers with respect to the general population than women diagnosed at postmenopausal ages (≥ 50 years). It is therefore likely that pre- and postmenopausal breast cancer predispose differently to a second cancer. In premenopausal women, the higher risk might be consistent with an inherited cancer predisposition, as second cancers in the ovary are predominant in these women. Almost all of the selected studies reported an increased second cancer in the ovary in these women [6, 7,10,12,13,18,20,31,38,41,42]. This association might be related to germline mutations in BRCA1/2 genes, which increase the risk of both types of cancer [54], or to shared reproductive or environmental risk factors [55,56]. Moreover, it has been reported that an excess risk for second cancers remains in pre-menopausal women with breast cancer having not received any treatment [57]. By contrast, in postmenopausal women, the overall increased risk of second cancers with respect to the general population was lower when compared to premenopausal women. Second cancers in the endometrium were more commonly reported in this age group [7,12,13,18,31,38,41,42]. Hormonal therapy with tamoxifen used to treat estrogen-receptor positive cancer has been proposed as an explanation [58], but it might be also attributable to common risk factors, such as reproductive, environmental and/or genetic factors [54,55]. Obesity has been also associated with an increased risk of endometrial cancer after breast cancer [59]. Risk estimates by time since diagnosis of the first tumour are rather inconsistent between studies. We were not able to harmonise equal time intervals and pooled therefore a smaller set of the selected studies. Women diagnosed with breast cancer may be at higher risk of second cancers during the first 10 years after diagnosis and thereafter, although this result has to be taken with caution. Our data agrees with the finding of increased second cancer risk with increasing time since breast cancer diagnosis reported by Mellemkjaer et al. [12], although it still contradicts results of other studies [20,39–42]. Although an increased diagnostic attention following breast cancer could partly explain the excess risk of second cancers observed within a shorter length of time after the first breast cancer diagnosis, the increasing risk of second cancers with time since diagnosis of the first breast cancer could be attributed to treatment related factors [12]. Patterns of second primary cancer development after a first primary cancer are poorly understood. Treatment related factors or shared risk factors have been proposed to explain this association but their effect on second cancer risk is still uncertain [5]. To our knowledge, no large population-based study has yet investigated second primary cancer incidence according to tumour subtype, or other clinical and anatomical characteristics, which may also influence second primary cancer development. Only one study has examined the effect of tumour size at diagnosis on the risk although no differences on second cancer risk were associated to this variable [17]. Only few of the studies evaluated the effect of treatment on the risk of second tumours [13,17,18,38,40], although the treatment modalities were not comparable among the studies. There are several studies exploring the effect of different breast cancer treatment strategies on second cancer risk [19,57], but it was not possible to combine them for the same reasons. Consequently, more studies investigating this issue are warranted. Some limitations and strengths of the present study warrant consideration. Our study includes different study populations, from both

13

hospital-based and population-based studies; however, these criteria also may contribute to the substantial heterogeneity observed and therefore may be a limitation of the study design. Stratification of results according to etiological factors may reduce this heterogeneity. However, this could not be performed because we lacked the data necessary to evaluate the effects of birth cohort, treatment of the breast cancer, or other factors that may influence the risk estimates associated with a first primary breast cancer. Furthermore, the articles found by carrying out a bibliographic search may be a biased selection of all studies carried out. Although we did not found evidence for publication bias, it is impossible to rule out the presence of this bias. Conclusions In this review we observed varying results on the magnitude of risk for subsequent cancers after a diagnosis of breast cancer that could be explained by various differences between the published studies, either methodological, such as the definition of second cancers, or inherent to the study populations, such as varying breast cancer treatment protocols between the studies and over time. The qualitative analysis of studies included in this review has revealed that these and other variables may have affected the results. By pooling the risk estimates reported in these studies, in a quantitative analysis of their results, it has been found that women diagnosed with a first primary breast cancer have a 17% higher risk of developing a second primary cancer with respect to the general population. However, a number of unanswered questions remain with regard to the modifying effect of breast cancer treatments (chemotherapy, surgery, radiotherapy or hormonal agents) or pathological and clinical characteristics of the tumours (stage, hormone receptors and biomarkers) on this risk. Further studies addressing these issues are warranted to accumulate evidence on this issue. This higher risk emphasises the importance of prevention and control policies targeted to these women and aimed at reducing incidence of second cancers. Primary and secondary prevention efforts are both highly important to lessen the burden of second cancers. Conflict of interest The authors declare that they have no conflict of interest. Role of the funding source This study has been supported by the Spanish Regional Government of Andalucia: Consejería Economía, Innovación, Ciencia y Empleo, Junta de Andalucia (CTS-3935, CTS-177). Author's contributions are as follows EMM, MR and MJS carried out the review and drafted the manuscript. They also conceived of the study. MJS participated in its coordination. EMM, MFF and ESC conducted the data analysis. MFF, MAM and JE and JPA helped to extract the data and reviewed critically all the studies selected. All authors read and approved the final manuscript. References [1] Ferlay J, Soerjomataram I, Ervik M, Dikshit R, Eser S, Mathers C, et al. GLOBOCAN 2012 v1.0, Cancer incidence and mortality worldwide: IARC CancerBase No. 11, [Internet]. Lyon, France: International Agency for Research on Cancer; 2013 [Available from http://globocan.iarc.fr]. [2] Ferlay J, Steliarova-Foucher E, Lortet-Tieulent J, Rosso S, Coebergh JW, Comber H, et al. Cancer incidence and mortality patterns in Europe: estimates for 40 countries in 2012. Eur J Cancer 2013;49(6):1374–403.

Fig. 2. Meta-analysis of second cancers after a first primary breast cancer. Random effects models are shown. A) SIR for all second cancers combined. B) SIR by age at diagnosis (b50 years and ≥50 years). C) SIR by time since diagnosis (b10 years and ≥10 years). The dashed vertical line represents the combined estimate, and the shaped box the size of the squares in terms of weight of each study in the meta-analysis. Error bars indicate 95% confidence interval.

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029

14

E. Molina-Montes et al. / Gynecologic Oncology xxx (2014) xxx–xxx

[3] De Angelis R, Sant M, Coleman MP, Francisci S, Baili P, Piearannunzio D. EUROCARE5 Working Group, et al., Cancer survival in Europe 1999–2007 by country and age: results of EUROCARE-5—a population-based study. Lancet Oncol 2014;15(1):23–34. [4] Verdecchia A, Guzzinati S, Francisci S, De Angelis R, Bray F, Allemani C, et al. Survival trends in European cancer patients diagnosed from 1988 to 1999. Eur J Cancer 2009; 45(6):1042–66. [5] Hemminki K, Boffetta P. Multiple primary cancers as clues to environmental and heritable causes of cancer and mechanisms of carcinogenesis. IARC Sci Publ 2004; 157:289–97. [6] Brenner H, Siegle S, Stegmaier C, Ziegler H. Second primary neoplasms following breast cancer in Saarland, Germany, 1968–1987. Eur J Cancer 1993;29A(10):1410–4. [7] Evans HS, Lewis CM, Robinson D, Bell CM, Moller H, Hodgson SV. Incidence of multiple primary cancers in a cohort of women diagnosed with breast cancer in southeast England. Br J Cancer 2001;84(3):435–40. [8] Soerjomataram I, Louwman WJ, de Vries E, Lemmens VE, Klokmanand WJ, Coebergh JW. Primary malignancy after primary female breast cancer in the South of the Netherlands, 1972–2001. Breast Cancer Res Treat 2005;93(1):91–5. [9] Levi F, Te VC, Randimbison L, La Vecchia C. Cancer risk in women with previous breast cancer. Ann Oncol 2003;14(1):71–3. [10] Cluze C, Delafosse P, Seigneurin A, Colonna M. Incidence of second cancer within 5 years of diagnosis of a breast, prostate or colorectal cancer: a population-based study. Eur J Cancer Prev 2009;18(5):343–8. [11] Volk N, Pompe-Kirn V. Second primary cancers in breast cancer patients in Slovenia. Cancer Causes Control 1997;8(5):764–70. [12] Mellemkjaer L, Friis S, Olsen JH, Scelo G, Hemminki K, Tracey E, et al. Risk of second cancer among women with breast cancer. Int J Cancer 2006;118(9):2285–92. [13] Mellemkjaer L, Christensen J, Frederiksen K, Pukkala E, Weiderpass E, Bray F, et al. Risk of primary non-breast cancer after female breast cancer by age at diagnosis. Cancer Epidemiol Biomarkers Prev 2011;20(8):1784–92. [14] Berrington dG, Curtis RE, Gilbert E, Berg CD, Smith SA, Stovall M, et al. Second solid cancers after radiotherapy for breast cancer in SEER cancer registries. Br J Cancer 2010;102(1):220–6. [15] Raymond JS, Hogue CJ. Multiple primary tumours in women following breast cancer, 1973–2000. Br J Cancer 2006;94(11):1745–50. [16] Yu GP, Schantz SP, Neugut AI, Zhang ZF. Incidences and trends of second cancers in female breast cancer patients: a fixed inception cohort-based analysis (United States). Cancer Causes Control 2006;17(4):411–20. [17] Rubino C, de Vathaire F, Diallo I, Shamsaldin A, Le MG. Increased risk of second cancers following breast cancer: role of the initial treatment. Breast Cancer Res Treat 2000;61(3):183–95. [18] Andersson M, Jensen MB, Engholm G, Henrik SH. Risk of second primary cancer among patients with early operable breast cancer registered or randomised in Danish Breast Cancer cooperative Group (DBCG) protocols of the 77, 82 and 89 programmes during 1977–2001. Acta Oncol 2008;47(4):755–64. [19] Kirova YM, De Rycke Y, Gambotti L, Pierga JY, Asselain B, Fourquet A. Second malignancies after breast cancer: the impact of different treatment modalities. Br J Cancer 2008;98(5):870–4. [20] Schaapveld M, Visser O, Louwman MJ, de Vries EG, Willemse PH, Otter R, et al. Risk of new primary nonbreast cancers after breast cancer treatment: a Dutch population-based study. J Clin Oncol 2008;26(8):1239–46. [21] International Agency for Research on Cancer (IARC), World Health Organization (WHO), International Association of Cancer Registries (IACR), European Network of Cancer Registries (ENCR). International Rules for Multiple Primary Cancers (ICD-O-Third Edition). Lyon: IARC; 2000. [22] Scottish Intercollegiate Guidelines Network. SIGN 50: a guidelines developers' handbook. Edinburgh: SIGN; February 2013. [23] Megdal SP, Kroenke CH, Laden F, Pukkala E, Schernhammer ES. Night work and breast cancer risk: a systematic review and meta-analysis. Eur J Cancer 2005; 41(13):2023–32. [24] Tang JL, Liu JL. Misleading funnel plot for detection of bias in meta-analysis. J Clin Epidemiol 2000;53(5):477–84. [25] Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50(4):1088–101. [26] Egger M, Davey SG, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315(7109):629–34. [27] Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, Ioannidis JP, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate healthcare interventions: explanation and elaboration. BMJ 2009; 339:b2700. [28] Cortesi LD, De Matteis E, Rashid I, Cirilli C, Proietto M, Rivasi F, et al. Distribution of second primary malignancies suggests a bidirectional effect between breast and endometrial cancer a population-based study. Int J Gynecol Cancer 2009;19(8): 1358–63. [29] Okamoto N, Morio S, Inoue R, Akiyama K. The risk of a second primary cancer occurring in five-year survivors of an initial cancer. Jpn J Clin Oncol 1987;17(3):205–13. [30] Schenker JG, Levinsky R, Ohel G. Multiple primary malignant neoplasms in breast cancer patients in Israel. Cancer 1984;54(1):145–50.

[31] Prochazka MH, Hall P, Granath F, Czene K. Family history of breast cancer and young age at diagnosis of breast cancer increase risk of second primary malignancies in women: a population-based cohort study. Br J Cancer 2006;95(9):1291–5. [32] Teppo L, Pukkala E, Saxen E. Multiple cancer—an epidemiologic exercise in Finland. J Natl Cancer Inst 1985;75(2):207–17. [33] Storm HH, Jensen OM, Ewertz M, Lynge E, Olsen JH, Schou G, et al. Summary: multiple primary cancers in Denmark, 1943–80. Natl Cancer Inst Monogr 1985;68: 411–30. [34] Ewertz M, Mouridsen HT. Second cancer following cancer of the female breast in Denmark, 1943–80. Natl Cancer Inst Monogr 1985;68:325–9. [35] Dong C, Hemminki K. Second primary neoplasms in 633,964 cancer patients in Sweden, 1958–1996. Int J Cancer 2001;93(2):155–61. [36] Brown LM, Chen BE, Pfeiffer RM, Schairer C, Hall P, Storm H, et al. Risk of second nonhematological malignancies among 376,825 breast cancer survivors. Breast Cancer Res Treat 2007;106(3):439–51. [37] Levi F, Randimbison L, Te VC, La Vecchia C. Second primary cancers in breast cancer patients in Vaud, Switzerland. Cancer Causes Control 1998;9(4):463–4. [38] Murakami R, Hiyama T, Hanai A, Fujimoto I. Second primary cancers following female breast cancer in Osaka, Japan—a population-based cohort study. Jpn J Clin Oncol 1987;17(4):293–302. [39] Buiatti E, Crocetti E, Acciai S, Gafa L, Falcini F, Milandri C, et al. Incidence of second primary cancers in three Italian population-based cancer registries. Eur J Cancer 1997;33(11):1829–34. [40] Tanaka H, Tsukuma H, Koyama H, Kinoshita Y, Kinoshita N, Oshima A. Second primary cancers following breast cancer in the Japanese female population. Jpn J Cancer Res 2001;92(1):1–8. [41] Lee KD, Chen SC, Chan CH, Lu CH, Chen CC, Lin JT, et al. Increased risk for second primary malignancies in women with breast cancer diagnosed at young age: a population-based study in Taiwan. Cancer Epidemiol Biomarkers Prev 2008; 17(10):2647–55. [42] Molina-Montes E, Pollan M, Payer T, Molina E, Davila-Arias C, Sanchez MJ. Risk of second primary cancer among women with breast cancer: a population-based study in Granada (Spain). Gynecol Oncol 2013;130(2):340–5. [43] Crocetti E, Arniani S, Buiatti E. Synchronous and metachronous diagnosis of multiple primary cancers. Tumori 1998;84(1):9–13. [44] Higgins JPT, Green S, editors. Cochrane handbook for systematic reviews of interventions version 5.1.0 [updated March 2011]. The Cochrane Collaboration; 2011 [Available from www.cochrane-handbook.org]. [45] Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in metaanalyses. BMJ 2003;327(7414):557–60. [46] Bernstein JL, Lapinski RH, Thakore SS, Doucette JT, Thompson WD. The descriptive epidemiology of second primary breast cancer. Epidemiology 2003;14(5):552–8. [47] Gao X, Fisher SG, Emami B. Risk of second primary cancer in the contralateral breast in women treated for early-stage breast cancer: a population-based study. Int J Radiat Oncol Biol Phys 2003;56(4):1038–45. [48] Hemminki K, Vaittinen P. Familial risks in second primary breast cancer based on a family cancer database. Eur J Cancer 1999;35(3):455–8. [49] Bouchardy C, Benhamou S, Fioretta G, Verkooijen HM, Chappuis PO, Neyroud-Caspar I, et al. Risk of second breast cancer according to estrogen receptor status and family history. Breast Cancer Res Treat 2011;127(1):233–41. [50] Prochazka M, Granath F, Ekbom A, Shields PG, Hall P. Lung cancer risks in women with previous breast cancer. Eur J Cancer 2002;38(11):1520–5. [51] Schoen RE, Weissfeld JL, Kuller LH. Are women with breast, endometrial, or ovarian cancer at increased risk for colorectal cancer? Am J Gastroenterol 1994;89(6): 835–42. [52] Srinivasan R, Yang YX, Rubin SC, Morgan MA, Lewis JD. Women with a prior diagnosis of breast cancer are not at an increased risk for subsequent colorectal cancer. Am J Gastroenterol 2005;100(12):2759–64. [53] Gulhan I, Eser S, Yakut C, Bige O, Ilhan E, Yildirim Y, et al. Second primary gynecologic cancers after breast cancer in Turkish women. Int J Gynecol Cancer 2009;19(4): 648–50. [54] Welcsh PL, King MC. BRCA1 and BRCA2 and the genetics of breast and ovarian cancer. Hum Mol Genet 2001;10(7):705–13. [55] Trentham-Dietz A, Newcomb PA, Nichols HB, Hampton JM. Breast cancer risk factors and second primary malignancies among women with breast cancer. Breast Cancer Res Treat 2007;105(2):195–207. [56] Zografos GC, Panou M, Panou N. Common risk factors of breast and ovarian cancer: recent view. Int J Gynecol Cancer 2004;14(5):721–40. [57] Langballe R, Olsen JH, Andersson M, Mellemkjaer L. Risk for second primary non-breast cancer in pre- and postmenopausal women with breast cancer not treated with chemotherapy, radiotherapy or endocrine therapy. Eur J Cancer 2011;47(6):946–52. [58] Curtis RE, Boice Jr JD, Shriner DA, Hankey BF, Fraumeni Jr JF. Second cancers after adjuvant tamoxifen therapy for breast cancer. J Natl Cancer Inst 1996;88(12):832–4. [59] Druesne-Pecollo N, Touvier M, Barrandon E, Chan DS, Norat T, Zelek L, et al. Excess bodyweight and second primary cancer risk after breast cancer: a systematic review and meta-analysis of prospective studies. Breast Cancer Res Treat 2012;135(3): 647–54.

Please cite this article as: Molina-Montes E, et al, Risk of second cancers cancer after a first primary breast cancer: A systematic review and metaanalysis, Gynecol Oncol (2014), http://dx.doi.org/10.1016/j.ygyno.2014.10.029