Abstract
Purpose:
Studies of the etiology of inflammatory breast cancer (IBC), a rare but aggressive breast cancer, have been hampered by limited risk factor information. We extend previous studies by evaluating a broader range of risk factors.
Methods:
Between 2009–2015 we conducted a case-control study of IBC at six centers in Egypt, Tunisia, and Morocco; enrolled were 267 IBC cases and for comparison 274 non-IBC cases and 275 controls, both matched on age and geographic area to the IBC cases. We administered questionnaires and collected anthropometric measurements for all study subjects. We used multiple imputation methods to account for missing values and calculated odds ratios (ORs) and 95% confidence intervals (CIs) using polytomous logistic regression comparing each of the two case groups to the controls, with statistical tests for the difference between the coefficients for the two case groups.
Results:
After multivariable adjustment, a livebirth within the previous two years (OR = 4.6; 95% CI 1.8 to 11.7) and diabetes (OR = 1.8;95% CI 1.1 to 3.0) were associated with increased risk of IBC, but not non-IBC (OR = 0.9; 95% CI 0.3 to 2.5 and OR = 0.9; 95% CI 0.5 to 1.6 for livebirth and diabetes, respectively). A family history of breast cancer, inflammatory-like breast problems, breast trauma, and low socioeconomic status were associated with increased risk of both tumor types.
Conclusions:
We identified novel risk factors for IBC and non-IBC, some of which preferentially increased risk of IBC compared to non-IBC. Upon confirmation, these findings could help illuminate the etiology and aid in prevention of this aggressive cancer.
Keywords: Inflammatory breast cancer, risk factor, Egypt, Tunisia, Morocco
Introduction
Inflammatory breast cancer (IBC) is a rare, poorly understood and particularly aggressive clinicopathologic form of breast cancer. Clinically, it is characterized by rapidly arising diffuse erythema and edema/peau d’orange of the breast thought to be due to the presence of tumor emboli in the breast dermal lymphatics [1]. There may or may not be an underlying tumor mass. Diagnosis of IBC is sometimes delayed because it can be difficult to distinguish clinically from an infectious process [2]. In the United States, IBC accounts for a disproportionate number of breast cancer deaths (7 percent) given that it constitutes only approximately 2 percent of all breast cancers [3]. The percentage of all breast cancer cases that are IBC is higher in North Africa (10 percent) than in the United States [4].
IBC has been called one of the “most perplexing problems in breast cancer” [5] and “in many ways an orphan disease for which adequate physician education is needed” [6]. The etiology of IBC and less aggressive forms of rapidly progressing breast cancer has been studied in Tunisia and France in the 1970s and 1980s [1, 7–11] and in some studies in the United States and elsewhere [1, 12–18], but all studies of IBC have had limited risk factor information.
In this case-control study in Egypt, Tunisia, and Morocco, we examined many under-studied potential risk factors for IBC and non-IBC, including time since last birth, breast problems during and outside lactation, breast injury, diabetes, agricultural exposures, and childhood socioeconomic status. Our study adds to the limited literature on IBC [10–11] and breast cancer in North Africa [4, 19–21]. Identifying risk factors, particularly those that differ for IBC and non-IBC, could provide clues to tumor biology, aid in prevention, and lead to earlier diagnosis and proper treatment.
Materials and Methods:
Between 2009–2015 we conducted a case-control study of IBC at 3 centers in Egypt (the Gharbiah Cancer Society and Tanta Cancer Center are adjacent to each other and for analyses are treated as one center), 1 center in Tunisia, and 2 centers in Morocco [22–23] (Table 1). The study was approved by ethical review boards at all study sites and all participants provided informed consent.
Table 1.
Study design factors
| IBC (N= 267) # (%) |
Non-IBC (N=274) # (%) |
Controls (N= 275) # (%) |
|
|---|---|---|---|
| Center | |||
| Tanta Cancer Center/Gharbiah Cancer Society, Tanta Egypt | 63 (23.6) | 64 (23.4) | 64 (23.3) |
| National Cancer Institute-Cairo, Cairo, Egypt | 60 (22.5) | 56 (20.4) | 60 (21.8) |
| Institut Salah Azaiz, Tunis, Tunisia | 57 (21.4) | 57 (20.8) | 49 (17.8) |
| Ibn Rochd Oncology Center, Casablanca, Morocco | 57 (21.4) | 70 (25.6) | 72 (26.2) |
| University Hospital Center Mohammed VI, Marrakech, Morocco | 30 (11.2) | 27 (9.9) | 30 (10.9) |
| Matching Factors | |||
| Age | |||
| 20–29 | 9 (3.4) | 8 (2.9) | 13 (4.7) |
| 30–34 | 13 (4.9) | 12 (4.4) | 13 (4.7) |
| 35–39 | 34 (12.7) | 30 (11.0) | 33 (12.0) |
| 40–44 | 35 (13.1) | 37 (13.5) | 37 (13.5) |
| 45–49 | 37 (13.9) | 44 (16.1) | 40 (14.6) |
| 50–54 | 34 (12.7) | 38 (13.9) | 35 (12.7) |
| 55–59 | 40 (15.0) | 37 (13.5) | 41 (14.9) |
| 60–64 | 28 (10.5) | 32 (11.7) | 31 (11.3) |
| 65–69 | 24 (9.0) | 18 (6.6) | 18 (6.6) |
| 70–74 | 9 (3.4) | 11 (4.0) | 9 (3.3) |
| 75–84 | 4 (1.5) | 7 (2.6) | 5 (1.8) |
| Geographic area* | |||
| Tanta Cancer Center/Gharbiah Cancer Society | |||
| Gharbiah governorate | 63 (100) | 64 (100) | 64(100) |
| NCI-Cairo | |||
| Cairo metropolitan area | 43 (71.7) | 40 (71.4) | 43 (71.7) |
| Other areas of Egypt | 15 (25.0) | 14 (25.0) | 12 (20.0) |
| Unknown | 2 (3.3) | 2 (3.6) | 5 8.3) |
| Institut Salah Azaiz | |||
| Tunis metropolitan area | 24 (42.1) | 17 (29.8) | 22 (44.9) |
| Other areas of Tunisia | 31 (54.4) | 40 (70.2) | 27 (55.1) |
| Foreign | 2 (3.5) | 0 (0.0) | 0 (0.0) |
| Ibn Rochd Oncology Center | |||
| Casablanca metropolitan area | 53 (93.0) | 49 (70.0) | 69 (95.8) |
| Unknown | 4 (7.0) | 21 (30.0) | 3 (4.2) |
| University Hospital Center Mohammed VI | |||
| Marrakech metropolitan area | 30 (100.0) | 27 (100.0) | 30 (100.0) |
We recruited 267 IBC cases aged 18 or older at diagnosis from the surgery departments at Tanta Cancer Center and the Gharbiah Cancer Society, and from medical oncology departments of the National Cancer Institute-Cairo, the Institut Salah Azaiz in Tunis, Tunisia, Ibn Rochd Oncology Center in Casablanca, Morocco, and University Hospital Center Mohammed VI in Marrakech, Morocco (Table 1). We also recruited 274 female non-IBC cases matched to the IBC cases on age (within 5-year categories, e.g. 20–24, 25–29, …, 75–79, 80–84; for analyses we collapsed the two youngest and two oldest age categories due to small numbers), and geographic area accrued over the same time period as the IBC cases (Table 1). Non-IBC patients were identified from outpatient surgery clinics at the Tanta Cancer Center/Gharbiah Cancer Society and the NCI-Cairo, and from the medical oncology departments at the Institut Salah Azaiz (selected from women having their initial consultation with the study medical oncologists), the Ibn Rochd Oncology Center and University Hospital Center Mohammed VI. Non-IBC cases at the Ibn Rochd Oncology Center were also selected from the gynecologic surgery department. At least 22 non-IBC cases were recruited post-mastectomy or post-chemotherapy (information on prior treatment was not available for 46 non-IBC cases). Pathology reports were retrieved for 247 (93%) IBC cases and 255 (93%) non-IBC cases. All IBC and non-IBC cases approached by study clinicians agreed to participate.
We also recruited 275 female visitor controls matched to the IBC cases on age category and geographic residence (as described for the non-IBC cases) from visitors or those accompanying outpatients to the participating hospitals (Table 1). First-degree relatives of cancer patients (mothers, sisters, daughters) and anyone visiting or accompanying breast, ovarian, endometrial, or nasopharyngeal cancer patients were considered not eligible during study implementation. Sixty-seven percent of visitors were friends, neighbors, or cousins of the person they were visiting or accompanying; 19.6 percent were aunts, nieces, granddaughters, grandmothers, mothers-in-law, sisters-in-law, or daughters-in-law to patients visited or accompanied; 11.6 percent had an unknown relationship. Ten visitors were included even though they did not strictly meet the eligibility criteria: 2 percent were first-degree relatives (cancer of patient they were visiting/accompanying was unknown) and four had visited a breast cancer patient (but none of these visitors were first-degree relatives). We had information on how many eligible visitors had to be approached for four of the six sites (NCI-Cairo, Tanta Cancer Center, Gharbiah Cancer Society, University Hospital Center Mohammed VI); 74 percent were the first eligible visitor approached. Of the 74 potential visitor controls who refused, 32 percent refused because the patient they were visiting was too sick, 61 percent did not have the time, and 6.8 percent had other reasons.
Trained interviewers administered questionnaires, study personnel took anthropometric measurements for all study participants, and physicians conducted clinical examinations for IBC and non-IBC cases. We published an evaluation of the cases previously [23].
Statistical Analysis
We calculated odds ratios (ORs) and 95% confidence intervals (CIs) for risk factors in relationship to case types (IBC, non-IBC) compared with control subjects using polytomous logistic regression, adjusted for age and location (the matching factors), with categorical age (shown in Table 1) treated as a continuous variable. The location variable identified country, institute within country, and matching categories for referral (Table 1).
To account for missing data (which ranged from <1 percent missing to 21 percent missing for exposure variables [percents missing for key variables were 2 percent for family history of breast cancer, 6 percent for duration of breast feeding, 7 percent for body mass index (BMI) (weight in kilograms/square of height in meters), 11 percent for age at menarche, and 21 percent for chest size], and from 16 percent to 36 percent for hormone receptor status of cases – details are presented in supplementary materials), we implemented the sequential regression imputation method with IVEware (http://www.isr.umich.edu/src/smp/ive) [24]. We obtained five imputations from the models (details are presented in supplementary materials), which included interaction terms between the outcome variable and BMI, chest size, waist size, hip size, and age at menarche. Missing values for ER, PR and HER2 status were only imputed for the two case groups. ORs were obtained for each of the five imputed datasets and were then averaged over the five datasets with the overall variance estimated using SAS version 9.3 PROC MIANALYZE. P-values for the statistical significance of the difference in ORs for the two case groups compared to controls were obtained using the TEST statement in PROC MIANALYZE.
We first present results for individual risk factors adjusted only for the variables in Table 1. We then adjusted related variables, such as anthropometric measures, for each other and only included those that remained statistically significant in the final multivariate models.
We present analyses of selected IBC risk factors by menopausal status and ER-status (positive/borderline, negative). P-values associated with differences by menopausal status were obtained by including interaction terms in the logistic model. We used a two-sided p-value of 0.05 to determine statistical significance.
Socioeconomic status assessment
We created adult and childhood standard of living indices (household indices) based on 14 household assets. Household assets used to create household indices were as follows: (a) flush toilet inside residence, (b) connection to a sewage system, (c) water piped into the residence for drinking, (d) electric fan, (e) gas or electric stove, (f) automatic clothes washing machine, (g) refrigerator, (h) water heater, (i) home telephone, (j) windows with both glass and blinds, shades, or curtains, (k) electricity for lighting needs,(l) computer, (m) no dirt, sand or dung floor, (n) no livestock/poultry in/around the house. The weight for each item was derived from 100 minus the proportion of households in the study possessing that item in their current household. Thus, for example, because 89.69 percent of current households had a flush toilet inside the residence, having a flush toilet was given the weight of 10.31 (i.e., 100–89.69) [25]. We used the same weights for current and childhood residences and weights for each item were summed into a linear index. The weighted indices were:
For analysis, each index was divided into quartiles based on the distribution among the controls, with the highest quartile serving as the reference category. We also combined childhood and adult household indices to create a life course index as described in Table 2.
Table 2.
Odds ratios (ORs) and 95% confidence intervals (CIs) for demographic factors and inflammatory breast cancer (IBC) and non-IBC
| Variable | IBC (N= 267) |
Non-IBC (N=274) |
Controls (N= 275) |
IBC vs. controls ORs (95% CI)a |
Non-IBC vs. controls ORs (95% CI)a |
|---|---|---|---|---|---|
| Highest level of schooling | |||||
| Secondary/some college | 49 | 70 | 128 | 1.0 | 1.0 |
| Primary/preparatory/certificate | 67 | 65 | 56 | 3.5 (2.1–5.9) | 2.3 (1.4–3.8) |
| None | 150 | 139 | 88 | 5.6 (3.5–8.8) | 3.3 (2.1–5.1) |
| Unknown | 1 | 3 | |||
| Categories treated as continuous | 2.3 (1.9–2.9) | 1.8 (1.5–2.3) | |||
| Adult household index | |||||
| Highest quartile | 44 | 42 | 68 | 1.0 | 1.0 |
| Next quartile | 35 | 42 | 67 | 0.8 (0.5–1.4) | 1.0 (0.6–1.7) |
| Next quartile | 67 | 69 | 69 | 1.5 (0.9–2.5) | 1.5 (0.9–2.6) |
| Lowest quartile | 121 | 121 | 71 | 2.9 (1.8–4.8) | 2.9 (1.8–4.8) |
| Categories treated as continuous | 1.5 (1.3–1.7) | 1.4 (1.2–1.7) | |||
| Childhood household index | |||||
| Highest quartile | 31 | 37 | 68 | 1.0 | 1.0 |
| Next quartile | 38 | 61 | 68 | 1.3 (0.7–2.4) | 1.9 (1.4–3.4) |
| Next quartile | 76 | 87 | 70 | 3.1 (1.7–5.4) | 2.7 (1.5–4.6) |
| Lowest quartile | 122 | 89 | 69 | 5.5 (3.1–9.8) | 3.0 (1.7–5.3) |
| Categorical treated as continuous | 1.9 (1.5–2.2) | 1.4 (1.2–1.7) | |||
| Life course household index | |||||
| Quartiles 1,2,3 for child and adult | 101 | 131 | 170 | 1.0 | 1.0 |
| Lowest quartile child-3 higher quartiles adult | 44 | 54 | 36 | 2.3 (1.4–3.9) | 2.0 (1.2–3.3) |
| Lowest quartile adult-3 higher quartiles child | 45 | 22 | 34 | 2.5 (1.4–4.3) | 0.8 (0.5–1.5) |
| Lowest quartile child & adult | 77 | 67 | 35 | 4.5 (2.7–7.4) | 2.7 (1.7–4.5) |
| Categorical treated as continuous | 1.6 (1.4–1.9) | 1.3 (1.1–0.5) |
Based on multiply imputed data; frequencies are before imputation; adjusted for matching variables: age categories (20–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–84) as continuous, and location as shown in Table 1.
Results
The mean ages of the 267 IBC cases, 274 non-IBC cases, and 275 controls were 50.5, 50.8, and 49.9 years, respectively. Fifty-four percent of IBC cases (N=143), 53% (N= 146) of non-IBC cases, and 52% (n = 144) of visitors were postmenopausal.
Results from Table 2 indicate that those with lower education levels and childhood and adult socioeconomic status were at considerably higher risk of IBC and non-IBC than those at higher socioeconomic status. For all three groups the mean adult socioeconomic status index was higher than the childhood index (IBC cases: 195.6 vs 93.3; non-IBC cases:198.4 vs. 113.6; visitor controls: 236.5 vs. 140.3) (data not shown in Table).
As shown in Table 3, in univariate analyses, those with an age at menarche before age 12 were at higher risk of both IBC and non-IBC compared to those with menarche at ages 12–13. Giving birth to five or more children compared to 1–2 children was associated with increased risk of IBC, but not non-IBC. Age at first birth was not associated with risk of either tumor type, although nulliparous women were at lower risk of IBC than those with a first birth before age 20. Later age at last birth (≥35 years) was associated with increased risk of IBC (OR = 2.1; 95% CI 1.3 to 3.9). Compared to a last birth 10 or more years before, a birth in the last four years was associated with increased risk of IBC, but not non-IBC, particularly if the birth was in the last two years (OR = 4.5; 95% CI1.9 to 10.3). Longer breast-feeding duration, particularly exclusive breast feeding, was associated with increased risk of IBC, but not non-IBC. Lactational breast problems increased risk of both tumor types, particularly IBC (OR = 4.4; 95% CI 2.1 to 9.6). Difficulty becoming pregnant was not associated with risk of either tumor type. There were no dose-response patterns with oral contraceptive use.
Table 3.
Odds ratios (ORs) and 95% confidence intervals (CIs) associated with reproductive factors and inflammatory (IBC) and non-IBC
| Variable | IBC (N= 267) |
Non-IBC (N=274) |
Control (N= 275) |
IBC vs. controls OR (95% CI)a |
Non-IBC vs. Controls OR (95% CI)a |
|---|---|---|---|---|---|
| Age at Menarche | |||||
| <12 | 30 | 42 | 22 | 1.8 (0.9–3.6) | 2.1 (1.0–4.2) |
| 12–13 | 120 | 118 | 142 | 1.0 | 1.0 |
| ≥14 | 76 | 87 | 86 | 1.0 (0.7–1.5) | 1.2 (0.8–1.8) |
| Unknown | 41 | 27 | 25 | ||
| Number of livebirths | |||||
| 1–2 | 55 | 67 | 68 | 1.0 | 1.0 |
| 3–4 | 97 | 98 | 102 | 1.2 (0.8–1.9) | 0.9 (0.6–1.4) |
| ≥5 | 74 | 61 | 50 | 1.9 (1.1–3.3) | 1.1 (0.7–1.9) |
| Nulliparous | 41 | 48 | 55 | 1.0 (0.6–1.6) | 0.9 (0.5–1.5) |
| Age at first livebirth, years | |||||
| < 20 | 58 | 54 | 45 | 1.0 | 1.0 |
| 20–24 | 72 | 89 | 76 | 0.7 (0.4–1.2) | 0.9 (0.6–1.5) |
| ≥25 | 95 | 81 | 98 | 0.7 (0.4–1.2) | 0.6 (0.4–1.1) |
| Nulliparous | 41 | 48 | 55 | 0.6 (0.3–1.0) | 0.7 (0.4–1.3) |
| Unknown | 1 | 2 | 1 | ||
| Age at last livebirth, years | |||||
| <30 | 53 | 71 | 72 | 1.0 | 1.0 |
| 30–34 | 72 | 72 | 80 | 1.2 (0.4–1.9) | 0.9 (0.6–1.4) |
| ≥35 | 100 | 81 | 66 | 2.1 (1.3–3.9) | 1.2 (0.8–2.0) |
| Nulliparous | 41 | 48 | 55 | 1.1 (0.6–1.8) | 0.9 (0.6–1.6) |
| Unknown | 1 | 2 | 1 | ||
| Time since last live birth, years | |||||
| ≥10 | 148 | 171 | 164 | 1.0 | 1.0 |
| 5–9 | 29 | 31 | 30 | 1.4 (0.8–2.7) | 1.1 (0.6–1.9) |
| 3–4 | 20 | 10 | 13 | 2.7 (1.2–6.0) | 0.8 (0.3–2.1) |
| 0–2 | 28 | 12 | 11 | 4.5 (1.9–10.3) | 1.1 (0.5–2.9) |
| Nulliparous | 41 | 48 | 55 | 1.0 (0.6–1.6) | 0.9 (0.6–1.5) |
| Unknown | 1 | 2 | 1 | ||
| Duration breast feeding (months) -among parous women who breast fed | |||||
| 0 – < 24 | 31 | 41 | 35 | 1.0 | 1.0 |
| 24 – < 60 | 68 | 85 | 76 | 1.1 (0.6–2.0) | 1.0 (0.6–1.7) |
| 60 – < 84 | 41 | 33 | 45 | 1.2 (0.6–2.4) | 0.7 (0.3–1.4) |
| ≥84 | 57 | 49 | 33 | 2.4 (1.2–4.9) | 1.3 (0.6–2.6) |
| Unknown | 12 | 6 | 19 | ||
| Never breast fed-parous | 17 | 12 | 12 | 1.8 (0.7–4.4) | 1.0 (0.4–2.7) |
| Unknown whether breast fed | 1 | ||||
| Nulliparous | 41 | 48 | 55 | 1.0 (0.5–1.8) | 0.9 (0.5–2.6) |
| Duration exclusive breast feeding (months) among parous women who breast fed | |||||
| 0 – < 12 | 43 | 59 | 53 | 1.0 | 1.0 |
| 12 – < 24 | 65 | 72 | 74 | 1.4 (0.9–2.4) | 1.0 (0.6–1.7) |
| 24 – < 36 | 45 | 44 | 33 | 2.4 (1.3–4.4) | 1.4 (0.8–2.4) |
| ≥ 36 | 35 | 22 | 21 | 3.1 (1.6–6.1) | 1.1 (0.6–2.3) |
| Unknown | 21 | 17 | 27 | ||
| Never breast fed-parous | 17 | 12 | 12 | 2.2 (1.0–5.2) | 1.2 (0.5–2.8) |
| Unknown whether breast fed | 1 | ||||
| Nulliparous | 41 | 48 | 55 | 1.2 (0.7–2.1) | 1.0 (0.6–1.7) |
| Ever had a breast problemb that prevented lactation, occurred during lactation or within 6 months after stopping lactation | |||||
| No | 233 | 256 | 266 | 1.0 | 1.0 |
| Yes | 34 | 18 | 9 | 4.4 (2.1–9.6) | 2.1 (0.9–4.8) |
| Didn’t breast feed or stopped because of insufficient milk | |||||
| No among parous | 184 | 192 | 180 | 1.0 | 1.0 |
| Yes | 42 | 34 | 39 | 1.0 (0.6–1.7) | 0.8 (0.5–1.3) |
| Nulliparous | 41 | 48 | 55 | 0.8 (0.5–1.2) | 0.9 (0.6–1.4) |
| Ever tried to get pregnant for 2 years and/or visited a doctor for infertility | |||||
| No | 229 | 247 | 241 | 1.0 | 1.0 |
| Yes | 38 | 27 | 34 | 1.2 (0.7–1.9) | 0.8 (0.4–1.3) |
| Years of oral contraceptive use (yrs) | |||||
| No use | 156 | 164 | 186 | 1.0 | 1.0 |
| < 5 | 45 | 46 | 28 | 2.0 (1.2–3.5) | 2.0 (1.3–2.4) |
| 5–9 | 31 | 20 | 18 | 2.3 (1.2–4.3) | 1.4 (0.7–2.9) |
| ≥10 | 35 | 44 | 43 | 1.1 (0.6–1.8) | 1.3 (0.8–2.1) |
| Age at first oral contraceptive use | |||||
| Never | 156 | 164 | 186 | 1.0 | 1.0 |
| <25 | 40 | 55 | 35 | 1.5 (0.9–2.5) | 2.1 (1.2–3.4) |
| 25–34 | 52 | 37 | 48 | 1.4 (0.9–2.2) | 0.9 (0.6–1.5) |
| ≥ 35 | 19 | 18 | 6 | 4.1 (1.6–10.6) | 3.8 (1.4–9.9) |
| Age last first oral contraceptive use | |||||
| Never | 156 | 164 | 184 | 1.0 | 1.0 |
| <30 | 24 | 26 | 17 | 1.8 (0.9–3.5) | 1.9 (1.0–3.7) |
| 30–39 | 43 | 34 | 35 | 1.6 (1.0–2.7) | 1.2 (0.7–2.0) |
| ≥ 40 | 44 | 50 | 37 | 1.6 (0.9–2.6) | 1.7 (1.1–2.8) |
Based on multiply imputed data; frequencies are before imputation; adjusted for matching variables: age categories (20–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–84) as continuous, and location as shown in table 1.
Abscess, ulcer/sore, secretions, redness, swelling, pain, sore nipples, other
A first-degree family history of breast cancer, a non-lactational breast problem, and breast trauma were associated with increased risk of both IBC and non-IBC (Table 4). Chest radiotherapy for other than a breast problem was not associated with a significantly increased risk of either tumor type. Diabetes was associated with increased risk of IBC, particularly if diagnosed after age 40, regardless of treatment. Diabetes was not associated with non-IBC. Greater BMI was associated with IBC, but not non-IBC.
Table 4.
Odds ratios (ORs) and 95% confidence intervals (CIs) for family history of breast cancer, medical conditions, and anthropometry and inflammatory breast cancer (IBC) and non-IBC
| Variable | IBC (N= 267) |
Non-IBC (N=274) |
Controls (N= 275) |
IBC vs. controls OR (95% CI)a |
Non-IBC vs. controls OR (95% CI)a |
|---|---|---|---|---|---|
| 1st degree family history of breast cancer | |||||
| No | 245 | 246 | 261 | 1.0 | 1.0 |
| Yes | 18 | 23 | 9 | 2.1 (0.9–4.8) | 2.7 (1.2–5.9) |
| Unknown | 4 | 5 | 5 | ||
| Parents related outside marriage | |||||
| No | 194 | 218 | 222 | 1.0 | 1.0 |
| Yes | 73 | 56 | 53 | 1.5 (1.0–2.3) | 1.0 (0.7–1.6) |
| Ever had a breast procedureb | |||||
| No | 259 | 263 | 265 | 1.0 | 1.0 |
| Yes | 8 | 11 | 10 | 0.8 (0.3–2.0) | 1.0 (0.4–2.5) |
| Non-lactation breast problemc | |||||
| No | 253 | 263 | 270 | 1.0 | 1.0 |
| Yes | 14 | 11 | 5 | 2.8 (1.0–8.2) | 1.8 (0.6–5.4) |
| Trauma to the breast | |||||
| No | 243 | 260 | 272 | 1.0 | 1.0 |
| Yes | 24 | 14 | 3 | 9.2 (2.7–31.0) | 5.1 (1.4–18.2) |
| Ever received radiation therapy to the chest for condition other than breast problem and excluding diagnostic x-rays | |||||
| No | 260 | 266 | 269 | 1.0 | 1.0 |
| Yes | 7 | 8 | 6 | 1.2 (0.4–3.8) | 1.5 (0.5–4.6) |
| Diabetes | |||||
| No | 203 | 234 | 240 | 1.0 | 1.0 |
| Yes | 64 | 40 | 35 | 2.2 (1.4–3.6) | 1.1 (0.7–1.9) |
| Age dx diabetes | |||||
| Never | 203 | 234 | 240 | 1.0 | 1.0 |
| <40 | 10 | 5 | 7 | 1.5 (0.6–3.7) | 0.5 (0.2–1.6) |
| ≥40 | 48 | 33 | 25 | 2.5 (1.5–4.3) | 1.4 (0.8–2.5) |
| Unknown | 6 | 2 | 3 | ||
| Ever prescribed medication for diabetes | |||||
| No diabetes | 203 | 234 | 240 | 1.0 | 1.0 |
| Yes diabetes-no medication | 12 | 8 | 7 | 2.2 (0.8–5.6) | 1.2 (0.4–3.4) |
| Yes diabetes-medication | 52 | 32 | 27 | 2.3 (1.3–3.8) | 1.1 (0.6–2.0) |
| Unknown | 1 | ||||
| BMI (bmicat) | |||||
| < 25 | 41 | 52 | 51 | 1.0 | 1.0 |
| 25–<30 | 71 | 86 | 95 | 0.9 (0.5–1.4) | 0.7 (0.4–1.2) |
| 30–<35 | 53 | 60 | 60 | 1.1 (0.6–1.9) | 0.8 (0.5–1.4) |
| ≥ 35 | 78 | 55 | 60 | 1.5 (0.8–2.8) | 0.7 (0.3–1.4) |
| Unknown | 24 | 21 | 9 | ||
| Increase per 5 BMI unit | 1.2 (1.0–1.4) | 1.0 (0.8–1.2) | |||
| Waist/hip ratio | |||||
| ≤ .8 | 37 | 45 | 49 | 1.0 | 1.0 |
| ≤.9 | 85 | 83 | 104 | 1.1 (0.6–1.9) | 1.0 (0.6–1.6) |
| >.9 | 91 | 89 | 88 | 1.3 (0.6–3.1) | 1.4 (0.9–2.4) |
| Unknown | 54 | 57 | 34 | ||
| Chest size (centimeters) | |||||
| 84–94 | 38 | 52 | 56 | 1.0 | 1.0 |
| 95–101 | 38 | 44 | 58 | 1.0 (0.5–2.0) | 0.8 (0.5–1.3) |
| 102–111 | 56 | 57 | 55 | 1.4 (0.7–2.7) | 1.0 (0.6–1.7) |
| >111 | 76 | 56 | 60 | 1.6 (0.5–5.1) | 1.1 (0.5–2.4) |
| Unknown | 59 | 65 | 46 | ||
| Increase per 5 breast size unit | 1.0 (0.9–1.2) | 1.0 (0.9–1.1) |
Based on multiply imputed data; frequencies are before imputation; adjusted for matching variables: age categories (20–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–84) as continuous, and location as shown in Table 1.
Needle aspiration, biopsy, removal of a lump, augmentation or reduction other than for current breast condition.
Abscess, ulcer/sore, secretions, redness, swelling, pain, sore nipples, other.
Ever having lived/worked on a farm was associated with increased risk of IBC, but not non-IBC (Table 5). Longer duration, older age at first exposure and more recent exposure to farm life were all associated with increased risk. Risk was increased regardless of exposure to field work or pesticides, and among those involved with animal care, but not among those not involved. Soap production was the only home business activity related to risk.
Table 5.
Odds ratios (ORs) and 95% confidence intervals (CIs) for agricultural and other occupation-related exposures and inflammatory breast cancer (IBC) and non-IBC
| Variable | IBC (N= 267) |
Non-IBC (N=274) |
Control (N= 275) |
IBC vs. controls OR (95% CI)a |
Non-IBC vs. controls OR (95% CI)a |
|---|---|---|---|---|---|
| Ever lived/ worked on a farm | |||||
| No | 199 | 242 | 238 | 1.0 | 1.0 |
| Yes | 68 | 32 | 37 | 2.3 (1.5–3.6) | 0.8 (0.5–1.3) |
| Years lived/worked on a farm | |||||
| Never | 199 | 242 | 238 | 1.0 | 1.0 |
| ≤15 | 23 | 7 | 17 | 1.7 (0.9–3.3) | 0.4 (0.2–0.9) |
| >15 | 45 | 25 | 20 | 2.8 (1.6–5.0) | 1.2 (0.6–2.2) |
| Age first worked/lived on farm | |||||
| Never | 199 | 242 | 238 | 1.0 | 1.0 |
| < 20 | 41 | 22 | 29 | 1.9 (1.1–3.1) | 0.7 (0.4–1.3) |
| ≥20 | 22 | 9 | 5 | 4.0 (1.7–9.5) | 1.1 (0.4–3.0) |
| Unknown | 5 | 1 | 3 | ||
| Years since worked on a farm | |||||
| Never | 199 | 242 | 238 | 1.0 | 1.0 |
| ≤ 2 | 29 | 17 | 12 | 2.6 (1.3–5.1) | 1.1 (0.5–2.3) |
| 3–19 | 21 | 13 | 8 | 3.3 (1.5–7.5) | 1.3 (0.5–3.3) |
| ≥20 | 13 | 1 | 13 | 1.3 (0.6–2.9) | 0.1 (0.0–0.6) |
| Unknown | 5 | 1 | 4 | ||
| Ever work in the fields | |||||
| Never worked on farm | 199 | 242 | 238 | 1.0 | 1.0 |
| No | 17 | 5 | 6 | 3.7 (1.4–9.6) | 0.8 (0.2–2.8) |
| Yes | 51 | 27 | 31 | 2.0 (1.2–3.3) | 0.8 (0.4–1.4) |
| Heavy exposure to pesticidesb | |||||
| Never on farm | 199 | 242 | 238 | 1.0 | 1.0 |
| No | 24 | 13 | 13 | 2.2 (1.1–4.6) | 0.8 (0.4–1.9) |
| Yes | 44 | 19 | 24 | 2.3 (1.3–4.0) | 0.8 (0.4–1.5) |
| Ever take care of livestock or other animalsc | |||||
| Never on farm | 199 | 242 | 238 | 1.0 | 1.0 |
| No | 15 | 5 | 15 | 1.2 (0.6–2.5) | 0.3 (0.1–0.9) |
| Yes | 53 | 27 | 22 | 3.1 (1.8–5.3) | 1.1 (0.6–2.1) |
| Made soap for family business | |||||
| No | 247 | 261 | 270 | 1.0 | 1.0 |
| Yes | 20 | 13 | 5 | 6.2 (2.2–17.6) | 3.7 (1.3–11.1) |
| Non-agricultural job outside home | |||||
| No | 209 | 210 | 202 | 1.0 | 1.0 |
| Yes | 58 | 64 | 73 | 0.9 (0.6–1.3) | 0.7 (0.5–1.1) |
Based on multiply imputed data; frequencies are before imputation; adjusted for matching variables: age categories (20–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–84) as continuous, and location as shown in Table 1.
Prepared, mixed, or applied pesticides; worked in the fields or took food to those working in the fields while the field was being sprayed with pesticides or soon after pesticides had been applied.
Livestock care, milking, pasteurization, herding, barn cleaning, breeding poultry, other animal husbandry.
Results from multivariable models are shown in Table 6. Other correlated reproductive factors and anthropometric factors from tables 3 and 4 were not included in these multivariable models because they were not associated with risk after adjustment for the variables that are included in the models. Results without adjustment for life course household index and education level are very similar to univariate results. With adjustment for life course household index and education, ORs for IBC for breast feeding and ever lived/worked on a farm were attenuated and no longer significant. Notably, education and life course household index were associated with risk even after adjustment for the other factors.
Table 6.
Multivariable odds ratios (ORs) and 95% confidence intervals (CIs) associated with reproductive factors and inflammatory (IBC) (N = 267) and non-IBC (N = 274) compared to controls (N = 275)
| Variable | IBC vs. Controls OR (95% CI)a |
Non-IBC vs. Controls OR (95% CI)a |
P-valueb | IBC vs. Controls OR (95% CI)c |
Non-IBC vs. Controls OR (95% CI)c |
P-valued |
|---|---|---|---|---|---|---|
| Age at Menarche | ||||||
| <12 | 1.4 (0.7–2.9) | 2.2 (1.2–3.9) | .26 | 1.5 (0.7–3.4) | 2.3 (1.1–4.3) | .28 |
| 12–13 | 1.0 | 1.0 | 1.0 | 1.0 | ||
| ≥14 | 1.1 (0.7–1.6) | 1.3 (0.9–1.9) | .33 | 1.1 (0.7–1.7) | 1.4 (0.9–2.1) | .35 |
| Age at first livebirth, years | ||||||
| < 20 | 1.0 | 1.0 | 1.0 | 1.0 | ||
| 20–24 | 0.8 (0.5–1.4) | 0.9 (0.5–1.5) | .78 | 1.0 (0.6–1.8) | 1.1 (0.6–2.0) | .76 |
| ≥25 | 1.0 (0.5–1.7) | 0.6 (0.3–1.1) | .12 | 1.2 (0.7–2.2) | 0.8 (0.5–1.5) | .14 |
| Nulliparous | 0.7 (0.2–1.9) | 0.7 (0.2–2.0) | .99 | 1.1 (0.4–3.2) | 1.1 (0.4–3.1) | .99 |
| Time since last live birth, years | ||||||
| ≥10 | 1.0 | 1.0 | 1.0 | 1.0 | ||
| 5–9 | 1.7 (0.9–3.3) | 1.0 (0.6–1.9) | .13 | 1.4 (0.7–2.9) | 0.9 (0.5–1.7) | .15 |
| 3–4 | 2.6 (1.1–6.4) | 0.8 (0.3–2.1) | .01 | 2.2 (0.9–5.6) | 0.7 (0.3–1.8) | .01 |
| 0–2 | 4.6 (1.9–11.2) | 0.9 (0.4–2.5) | .0005 | 4.5 (1.8–11.3) | 0.9 (0.3–2.5) | .004 |
| Nulliparous | 1.1 (0.4–2.9) | 1.0 (0.4–2.6) | .74 | 1.4 (0.5–3.6) | 1.2 (0.4–3.2) | .74 |
| Duration only breast feeding (months) among parous women who breast fed | ||||||
| 0 – < 12 | 1.0 | 1.0 | 1.0 | 1.0 | ||
| 12 – < 24 | 1.2 (0.7–2.1) | 1.0 (0.6–1.6) | .38 | 1.0 (0.6–1.8) | 0.8 (0.5–1.3) | .44 |
| 24 – < 36 | 1.7 (0.9–3.3) | 1.2 (0.7–2.3) | .29 | 1.3 (0.7–2.5) | 0.9 (0.5–1.7) | .30 |
| ≥ 36 | 2.2 (1.1–4.5) | 1.0 (0.5–2.1) | .03 | 1.3 (0.6–2.8) | 0.6 (0.3–1.3) | .05 |
| Nulliparous+never breast fed | 1.7 (0.7–4.2) | 1.1 (0.5–2.8) | .36 | 1.2 (0.5–3.1) | 0.8 (0.3–2.1) | .36 |
| Ever had a breast probleme that prevented lactation, occurred during lactation or within 6 months after stopping lactation | ||||||
| No | 1.0 | 1.0 | 1.0 | 1.0 | ||
| Yes | 3.0 (1.3–6.7) | 1.9 (0.8–4.6) | .21 | 3.2 (1.4–7.4) | 2.1 (0.9–5.1) | .23 |
| Years of oral contraceptive use (yrs) | ||||||
| No use | 1.0 | 1.0 | 1.0 | 1.0 | ||
| < 10 | 1.8 (1.1–3.0) | 1.8 (1.1–2.9) | .96 | 1.9 (1.1–3.2) | 1.9 (1.1–3.1) | .93 |
| ≥10 | 1.2 (0.7–2.1) | 1.4 (0.8–2.4) | .59 | 1.1 (0.6–1.9) | 1.2 (0.7–2.1) | .64 |
| 1st degree family history of breast cancer | ||||||
| No | 1.0 | 1.0 | 1.0 | 1.0 | ||
| Yes | 2.0 (0.9–4.9) | 2.7 (1.2–6.1) | .46 | 2.2 (0.9–5.5) | 2.9 (1.3–6.8) | .45 |
| Non-lactation breast probleme | ||||||
| No | 1.0 | 1.0 | 1.0 | 1.0 | ||
| Yes | 1.9 (0.6–6.0) | 1.6 (0.5–4.8) | .64 | 2.2 (0.7–6.9) | 1.8 (0.6–5.6) | .65 |
| Trauma to the breast | ||||||
| No | 1.0 | 1.0 | 1.0 | 1.0 | ||
| Yes | 8.3 (2.4–29.2) | 4.9 (1.4–17.8) | .16 | 9.3 (2.6–33.9) | 5.4 (1.4–20.1) | .14 |
| Diabetes | ||||||
| No | 1.0 | 1.0 | 1.0 | 1.0 | ||
| Yes | 2.0 (1.2–3.4) | 1.1 (0.6–1.8) | .008 | 1.8 (1.1–3.1) | 0.9 (0.5–1.6) | .009 |
| BMI | ||||||
| Increase per 5 BMI unit | 1.1 (1.0–1.3) | 1.0 (0.8–1.2) | .12 | 1.1 (1.0–1.3) | 1.0 (0.8–1.1) | .11 |
| Ever lived/worked on a farm | ||||||
| No | 1.0 | 1.0 | 1.0 | 1.0 | ||
| Yes | 2.4 (1.5–3.9) | 0.7 (0.4–1.3) | <.001 | 1.2 (0.7–2.1) | 0.4 (0.2–0.7) | <.001 |
| Highest level of schooling – categorical variable treated as continuous | N/A | N/A | 2.0 (1.5–2.6) | 2.0 (1.5–2.5) | .88 | |
| Life course household index – categorical variable treated as continuous | N/A | N/A | 1.3 (1.0–1.5) | 1.2 (1.0–1.5) | .54 |
Based on multiply-imputed data; adjusted for matching variables: age categories (20–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–84) treated as continuous, and location as shown in table 1, and other variables in the table except for level of schooling and life course SES; age at first birth and time since last birth were not included in the same model. ORs and p-values for variables other than age at first birth are from the model with time since last birth.
P-value for testing linear hypotheses that coefficients for IBC and non-IBC are equal based on the model without level of schooling and life course SES.
Based on multiply-imputed data; adjusted for matching variables: age categories (20–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–84) treated as continuous, and location as shown in Table 1, and other variables in the table including level of schooling and life course SES; age at first birth and time since last birth were not included in the same model. ORs and p-values for variables other than age at first birth are from the model with time since last birth.
P-value for testing linear hypotheses that coefficients for IBC and non-IBC are equal based on the model including level of schooling and life course SES.
Abscess, ulcer/sore, secretions, redness, swelling, pain, sore nipples, other.
The main IBC findings did not differ significantly by menopausal and ER status, except data on time since last birth was lacking for post-menopausal women (Tables 7 and 8). Non-IBC associations with BMI and age at first livebirth also did not differ significantly by menopausal or ER status (data not shown in tables).
Table 7.
Odds ratios (ORs) and 95%confidence intervals (CIs) for selected risk factors for inflammatory breast cancer (IBC) by pre-/post-menopausal status
| Variable | Premenopausal | Postmenopausal | Pre-Menopausal IBC vs. controls OR (95% CI)a |
Post-menopausal IBC vs. controls OR (95% CI)a |
P-valueb | ||
|---|---|---|---|---|---|---|---|
| IBC N = 128 |
Controls N = 137 |
IBC N = 139 |
Controls N = 138 |
||||
| Age at first livebirth, years | |||||||
| < 20 | 20 | 19 | 38 | 26 | 1.0 | 1.0 | |
| 20–24 | 34 | 37 | 38 | 39 | 0.9 (0.4–2.0) | 0.6 (0.3–1.3) | .43 |
| ≥25 | 52 | 48 | 43 | 50 | 1.1 (0.5–2.5) | 0.6 (0.3–1.3) | .25 |
| Nulliparous | 22 | 33 | 19 | 22 | 0.7 (0.3–1.8) | 0.8 (0.3–1.8) | .99 |
| Unknown | 1 | 1 | |||||
| Time since last live birth, years (parous women only) | |||||||
| ≥10 | 30 | 52 | 118 | 112 | 1.0 | N/A | |
| 5–9 | 28 | 28 | 1 | 3 | 3.6 (1.6–8.1) | ||
| 3–4 | 20 | 13 | 0 | 0 | 6.0 (2.2–16.7) | ||
| 0–2 | 28 | 11 | 0 | 0 | 13.9 (4.8–40.4) | ||
| Nulliparous | 22 | 33 | 19 | 22 | 2.6 (1.1–6.2) | ||
| Unknown | 1 | 1 | |||||
| Diabetes | |||||||
| No | 110 | 129 | 93 | 111 | 1.0 | 1.0 | .23 |
| Yes | 18 | 8 | 46 | 27 | 4.1 (1.5–11.5) | 2.2 (1.2–3.9) | |
| BMI | |||||||
| Increase per 5 BMI unit | 1.1 (0.9–1.4) | 1.3 (1.0–1.5) | .34 | ||||
| Ever lived/ worked on a farm | |||||||
| No | 93 | 117 | 106 | 121 | 1.0 | 1.0 | |
| Yes | 35 | 20 | 33 | 17 | 3.3 (1.6–6.6) | 2.5 (1.3–4.9) | .72 |
Based on multiply imputed data; frequencies are before imputation; adjusted for matching variables: age categories (20–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–84) as continuous, and location as shown in Table 1, and other variables in the table except that age at first birth and time since last birth were not included in the same model. ORs and p-values for variables other than age at first birth are from the model with time since last birth.
P-value for interaction term between variable and menopausal status in model; p-values for other than age at first birth are based on a model with time since last birth.
Table 8.
Odds ratios (ORs) and 95%confidence intervals (CIs) for selected risk factors and estrogen receptor (ER) positive (+) and ER-negative (−) inflammatory breast cancer (IBC).
| Variable | ER+ IBC N = 96 |
ER− IBC N = 81 |
Control N = 275 |
ER+ IBC vs. controls OR (95% CI)a |
ER− IBC vs. controls OR (95% CI)a |
P-valueb |
|---|---|---|---|---|---|---|
| Age at first livebirth, years | ||||||
| < 20 | 25 | 17 | 45 | 1.0 | 1.0 | |
| 20–24 | 21 | 27 | 76 | 0.6 (0.3–1.2) | 0.8 (0.4–1.6) | |
| ≥25 | 39 | 27 | 98 | 0.8 (0.5–1.5) | 0.9 (0.4–1.6) | |
| Nulliparous | 11 | 10 | 55 | 0.6 (0.3–1.2) | 0.9 (0.4–2.0) | |
| Unknown | 1 | |||||
| Categories treated as continuous | 0.9 (0.7–1.1) | 1.0 (0.8–1.2) | .49 | |||
| Time since last live birth, years (parous women only) | ||||||
| ≥10 | 59 | 41 | 164 | 1.0 | 1.0 | |
| 5–9 | 7 | 11 | 31 | 1.6 (0.7–3.7) | 2.2 (1.0–5.1) | .51 |
| 3–4 | 8 | 8 | 13 | 3.0 (1.1–8.6) | 3.3 (1.1–7.8) | .88 |
| 0–2 | 11 | 11 | 11 | 6.8 (2.3–20.4) | 6.9 (2.2–21.3) | .99 |
| Nulliparous | 1.0 (0.5–2.1) | 1.7 (0.8–3.6) | .29 | |||
| Unknown | 1 | |||||
| Diabetes | ||||||
| No | 73 | 67 | 240 | 1.0 | 1.0 | |
| Yes | 23 | 14 | 35 | 2.5 (1.3–4.8) | 2.3 (1.1–4.8) | .88 |
| BMI | ||||||
| Increase per 5 BMI unit | 1.2 (1.0–1.5) | 1.1 (1.0–1.4) | .50 | |||
| Ever lived/ worked on a farm | ||||||
| No | 69 | 56 | 238 | 1.0 | 1.0 | |
| Yes | 27 | 25 | 37 | 2.5 (1.4–4.5) | 3.1 (1.7–5.6) | .58 |
Based on multiply imputed data; frequencies are before imputation; adjusted for matching variables: age categories (20–29, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60–64, 65–69, 70–74, 75–84) as continuous, and location as shown in Table 1, and other variables in the table except that age at first birth and time since last birth were not included in the same model. ORs and p-values for variables other than age at first birth are from the model with time since last birth.
P-value for using the TEST statement in PROC MIANALYZE.
Discussion
In this case-control study in North Africa, a birth within the previous four years and diabetes were associated with increased risk of IBC, but not non-IBC. Although power was limited, associations did not differ by menopausal or ER-status. There was a trend of increasing risk with higher BMI for IBC but not non-IBC, but differences were not statistically significant. Low childhood and adult socioeconomic status, an early age at menarche, a family history of breast cancer, breast trauma, and lactational and non-lactational breast problems were associated with increased risk of both tumor types. The increased risk of IBC, but not non-IBC, with ever having lived/worked on a farm and with longer duration of breast feeding was attenuated after adjustment for socioeconomic status and education level.
Many, but not all, early case reports suggested an association between pregnancy and lactation and IBC risk [1]. A French study reported a significantly higher proportion of IBC among cases diagnosed during pregnancy or in the first six-months post-partum than a matched group of non-pregnancy associated cancers [26], and early studies in Tunisia showed associations with pregnancy/lactation [7–8]. A pooled analysis of 15 prospective studies in women younger than 55 years concluded that recent childbirth increased risk of all breast cancer by about 80 percent compared to nulliparous women [27], a considerably smaller increase in risk than for IBC in our study (although we used a different reference category—last birth 10 or more years before—which was equivalent in risk to nulliparous women
Higher BMI was associated with increased IBC risk in other studies [11, 12, 14–15] regardless of tumor subtype [14–15] or menopausal status [14]; two studies reported greater risk for IBC than non-IBC [11–12, 14]. We did not find the increased risk among post-menopausal women, as reported in many studies [e.g. 28].
To our knowledge, only one prior study has examined comorbidities associated with obesity and IBC risk [16]. In that study, diabetes was associated with a modest increase in risk of IBC and other advanced tumor types in women over age 66, but there was inadequate adjustment for BMI [16]. Diabetes has been associated with a small increase in risk of breast cancer in general [29].
Lower socioeconomic status was also associated with increased IBC risk in case-case, case-control and tumor registry studies [11, 14, 17–18]. Most studies in high-resource countries [30], and a previous study in Tunisia [20], have reported increased risk for all breast cancer with higher socioeconomic status. Rural residence was associated with increased risk of rapidly progressing breast cancer in earlier Tunisian studies [6], but in a geospatial analysis in the United States, high rate cluster centers of IBC had a lower proportion of those who lived in rural areas [18].
Family history of breast cancer had similar associations with IBC and non-IBC in a previous study [14]. A study which reported reduced risk of IBC with family history was most likely compromised by referral bias [15]. To our knowledge, no other studies have examined breast trauma and lactational and non-lactational breast problems and IBC risk. In one study of breast cancer in general, cases were more likely to report physical trauma to the breast than controls [31] and in another, breast cancer was slightly elevated in women with mastitis [32].
Previous studies found an association between an early age at first birth and increased risk of ER- IBC [14] or triple negative IBC (ER-, PR-, HER2-) [15], which we did not confirm in our data. We also did not find the established associations between nulliparity and older age at first livebirth reported in other studies of breast cancer in North Africa and elsewhere [4, 19–21].
Longer duration of breast feeding, even after adjustment for socioeconomic status, was associated with increased risk of IBC in a French study [11], whereas we report increased risk only before adjustment for socioeconomic status. Another study reported reduced risk of luminal and triple negative IBC among those who ever breast fed [15]. Consistent with other analyses [21, 33], we found some evidence that longer duration of breast feeding was associated with reduced non-IBC risk.
Compared to non-IBC, IBC has a higher proportion of aggressive molecular subtypes (HER2+ and triple negative) [34], is characterized by over-expression of e-cadherin [35–36], is more angiogenic and angioinvasive [35–36] and has high levels of vascular endothelial growth factor (VEGF) [37–39]. In addition, high levels of inflammatory cell infiltration and deregulated inflammatory signaling pathways have been identified in IBC tumors [2]. However, there are currently no molecular factors that reliably distinguish IBC from non-IBC tumors [35]; therefore, others have hypothesized that environmental factors that cause changes in the breast parenchyma before cancer development may prompt an aggressive subset of non-IBC cells to function like IBC cells [35].
Our findings, in general, support the hypothesis that inflammatory processes create a microenvironment more conducive to IBC than non-IBC development. The angiogenesis characteristic of IBC may be related to inflammation and inflammatory cytokines, which upregulate vascular endothelial growth factor (VEGF) [40], a major factor that stimulates new blood vessel formation. Obesity and type 2 diabetes are associated with macrophage infiltration into adipose tissue, and increased production of pro-inflammatory cytokines and chemokines [41, 42]. Mammary gland involution following pregnancy/lactation is also associated with an enhanced inflammatory microenvironment [43]. Pregnancy-associated changes have been shown to persist in human breast tissue for as long as five to 10 years [43]. Tissue injury can also initiate an inflammatory process [44]. Our data suggest that breast trauma may be associated with greater risk of IBC than non-IBC, although differences were not statistically significant.
Other possible carcinogenic mechanisms include comorbidities of obesity and diabetes, including hyperinsulinemia, insulin resistance and other metabolic disorders, the enhanced estrogen production associated with obesity, and the cell proliferation due to hormonal stimulation that occurs in pregnancy and lactation [43]. Human breast milk also contains a variety of potentially influential growth factors, including VEGF [45].
The persistence of an association with socioeconomic status after adjustment for other risk factors in our study suggests that low socioeconomic status may be a surrogate for unmeasured adverse environmental characteristics, possibly poor nutrition, exposure to infectious agents, and lack of access to medical care.
Our study has several strengths, including a detailed evaluation of IBC cases by photographs [23], a comparison group of non-IBC cases, and extensive risk factor information. In interpreting our results, we must consider, however, that geopolitical events or resource shortages precluded or reduced case ascertainment during certain periods of time at some of the study sites. Furthermore, our non-IBC case series was matched to the age and geographic distributions of the IBC cases, and thus is a non-representative sample. We also did not have enough cases to adequately investigate risk factors for pre- and post-menopausal women by ER, PR and HER2 status, factors which are known to affect associations with reproductive factors and BMI [28, 46, 47]. It is reassuring that we did find expected associations for some established risk factors, including age at menarche, family history of breast cancer, and breastfeeding.
In summary, we identified some novel potential risk factors for IBC and non-IBC, some of which preferentially increased risk of IBC compared to non-IBC. Although generally consistent with limited previous research on IBC, our findings require replication. Further research is also needed to identify the relevant biologic processes (e.g. inflammatory, hormonal, metabolic, infectious) associated with identified risk factors for IBC as well as other breast cancer types.
Supplementary Material
Acknowledgements:
We are very grateful to the women who participated in this study. We also thank the following individuals for their important contributions to the study, including identifying, enrolling, and interviewing study subjects, study management, data keying, logistical support, and translation: Tanta Cancer Center/Gharbiah Cancer Society: Dr. Salwas Samir, Dr. Eman Sobhy, Dr. Mohammed El Kholy, Mr. Khalid Dabboos, Ms. Hanaa Elmenshawy; NCI:Cairo: Dr. Gamal Amira, Dr. Heba Mohamed El Leethy, Dr. Usama Farouk El Nagar, Ms. Walaa Emara, Ms. Magda Hassan, Ms. Hanan Mabrouk; Tunisia: Dr. Karima Mrad, Dr. Dalenda Hentati, Dr. Houda Belfekih, Dr. Nesrine Chraiet, Dr. Amira Daldoul, Dr. Mouna Ayadi, Dr. Bassem Alani, Ms. Leila Fadhlaoui, Ms. Neila Ben Hassen; Ibn Rochd Oncology Center: Prof. Nadia Benchakroun, Dr. Hoda Eddakaoui, Dr. Myriam Hatime, Dr. Soufya Majdoul, Dr. Fadwa Qachach, Dr. Basma Billal, Dr. Hajar Kouss, Dr. Noureddine Matar, Dr. Iman Meziane, Dr. Sara Belkheiri, Dr. Karima Bendahhou, Dr. Majdouline Khounigere, Dr. Ibrahim Khalil Ahmadaye; University Hospital Center Mohammed VI, Marrakech: Dr. Mariam Affane, Dr. Sana Zabroug, Dr. Mouna Darfaoui, Mr. Hicham Jabraoui; IMS (data management and keying): Ms. Leslie Carroll, Mr. Bob Banks, Ms. Bette Griffith, Dr. Carol Giffen, Mr. Michael Spriggs; NCI-USA (logistical support): Mr. David Check, Ms. Prisca N. Fall; Westat (translation): Dr. Susan Crystal-Mansour, Ms. Jerelyn Bouic. The National Institutes of Health National Library of Medicine Interaction Tool used in this work was developed by the Communications Engineering Branch of the National Library of Medicine (Mr. Rodney Long and Mr. Leif Neve). We would also like to thank the CEESP students who assisted in different parts of the study during their CEESP projects.
Funding:
This work was supported by the National Cancer Institute at the National Institutes of Health Intramural Research Program (C.S., R.M.P., S.M.G.), the Breast Cancer Research Foundation (S.D.M.), the Metavivor Foundation (S.D.M.), and the Cancer Epidemiology Education in Special Populations (CEESSP) Program (Grant R25 CA112383) (A.S.S.).
Conflict of Interest:
Sandra Swain reports receiving honoraria from Novartis, personal fees from Cardinal Health, Daiichi-Sankyo, Eli Lilly & Co., Genentech/Roche, Genomic Health, Inivata, Peiris Pharmaceuticals, Tocagen; research support from Genentech; travel and accommodations from Caris Centers of Excellence, Daiichi-Sankyo, Eli Lilly & Co., Genentech/Roche, NanoString Technologies; and remuneration from AstraZeneca for participation on OlympiA IDMC. The other authors declare they have no conflicts of interest.
Abbreviations:
- IBC
inflammatory breast cancer
- ER
estrogen receptor
- BMI
body mass index
- OR
odds ratio
- CI
confidence interval
- PR
progesterone receptor
- HER2
human epidermal growth factor receptor 2
Footnotes
Ethical approval
All procedures performed in this study involving human participants were in accordance with the ethical standards of the institutional research committees.
Informed consent
Informed consent was obtained from all study participants.
Data availability statement:
Data are not available in a public database, but could be made available upon reasonable request.
References
- 1.Anderson WF, Schairer C, Chen BE, Hance KW, Levine PH (2006) Epidemiology of inflammatory breast cancer (IBC). Breast Disease 22:9–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Morrow RJ, Etemadi N, Yeo B, Ernst M (2017) Challenging a misnomer? The role of inflammatory pathways in inflammatory breast cancer. Mediators Inflamm 2017:4754827. doi: 10.1155/2017/4754827. Epub 2017 May 14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hance KW, Anderson WF, Devesa SS, Young HA, Levine PH (2005) Trends in inflammatory breast carcinoma incidence and survival: the surveillance, epidemiology, and end results program at the National Cancer Institute. J Natl Cancer Inst 97:966–975. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Corbex M, Bouzbid S, Boffetta P (2014) Features of breast cancer in developing countries, examples form North-Africa. Eur J Cancer 50:1808–1818. [DOI] [PubMed] [Google Scholar]
- 5.Amiri-Kordestani L, Kamangar F, Zujewski JA. Inflammatory breast cancer: yet another risk of the obesity epidemic? (2013) J Natl Cancer Inst 105:1340–1342. [DOI] [PubMed] [Google Scholar]
- 6.Cristofanilli M (2016). Inflammatory breast cancer: a new approach. Lancet Onco 17:544–546. [DOI] [PubMed] [Google Scholar]
- 7.Mourali N, Muenz L, Tabbane F, Belhassen S, Bahi J, Levine PH (1980) Epidemiologic features of rapidly progressing breast cancer in Tunisia. Cancer 46:2741–2746. [DOI] [PubMed] [Google Scholar]
- 8.Tabbane F, El May A, Hachiche M, Bahi J, Jaziri M, Cammoun M, Mourali N (1985) Breast cancer in women under 30 years of age. Breast Cancer Res and Treat 6:137–144. [DOI] [PubMed] [Google Scholar]
- 9.Boussen H, Bouzaiene H, Ben Hassouna J, Gamoudi A, Benna F, Rahal K (2008) Inflammatory breast cancer in Tunisia: reassessment of incidence and clinicopathological features. Semin Oncol 35:17–24. [DOI] [PubMed] [Google Scholar]
- 10.Boussen H, Bouzaiene H, Ben Hassouna J, Dhiab T, Khomsi F, Benna F, Gamoudi A, Mourali N, Hechiche M, Rahal K, Levine PH (2010) Inflammatory breast cancer in Tunisia: epidemiologic and clinical trends. Cancer 116(suppl 11):2730–2735. [DOI] [PubMed] [Google Scholar]
- 11.Lê MB, Arriagada R, Bahi J, Pfeiffer F, Cammoun M, Tabbane F, Rubino C (2006) Are risk factors for breast cancer similar in women with inflammatory breast cancer and in those with non-inflammatory breast cancer? Breast 15:355–362. [DOI] [PubMed] [Google Scholar]
- 12.Chang S, Buxdar AU, Hursting SD (1998) Inflammatory breast cancer and body mass index. J Clin Oncol 16:3731–3735. [DOI] [PubMed] [Google Scholar]
- 13.Moslehi R, Freedman E, Zeinomar N, Veneroso C, Levine PH (2016) Importance of hereditary and selected environmental risk factors in the etiology of inflammatory breast cancer: a case-comparison study. BMC Cancer 16:334. doi: 10.1186/s12885-016-2369-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Schairer C, Li Y, Frawley P, Graubard BI, Wellman RD, Buist DS, Kerlikowske K, Onega TL, Anderson WF, Miglioretti DL (2013) Risk factors for inflammatory breast cancer and other invasive breast cancers. J Natl Cancer Inst 105:1373–1384. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Atkinson RL, El-Zein R, Valero V, Lucci A, Bevers TB, Fouad T, Liao W, Ueno NT, Woodward WA, Brewster AM (2016) Epidemiological risk factors associated with inflammatory breast cancer subtypes. Cancer Causes Control 27:359–366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Schairer C, Gadalla SM, Pfeiffer RM, Moore SC, Engels EA (2017) Diabetes, abnormal glucose, dyslipidemia, hypertension, and risk of inflammatory and other breast cancer. Cancer Epidemiol Biomarkers Prev 26:862–868. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Schlichting JA, Soliman AS, Schairer C, Banerjee M, Rozek LS, Schottenfeld D, Harford JB, Merajver SD (2012) Association of inflammatory and noninflammatory breast cancer with socioeconomic characteristics in the Surveillance, Epidemiology, and End Results database, 2000–2007. Cancer Epidemiol Biomarkers Prev 21:155–165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Scott L, Mobley LR, Il’yasova D (2017). Geospatial analysis of inflammatory breast cancer and associated community characteristics in the United States. Int J Environ Res Public Health 14:404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Khalis M, Charbotel B, Chajès V, Rinaldi S, Moskal A, Biessy C, Dossus L, Huybrechts I, Fort E, Mellas N, Elfakir S, Charaka H, Nejjari C, Romieu I, El Rhazi K. Menstrual and reproductive factors and risk of breast cancer: A case-control study in the Fez region, Morocco (2018) PLoS ONE 13(1):e0191333. tps://doi.org/ 10.1371/journal.pone.0191333 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Awatef M, Olfa G, Slim BA (2013) Impact of menstrual and reproductive factors on breast cancer risk in Tunisia: a case-control study. Med Oncol 30:480. [DOI] [PubMed] [Google Scholar]
- 21.Awatef M, Olfa G, Imed H, Kacem M, Imen C, Rim C, Mohamed B, Slim BA (2010) Breastfeeding reduces breast cancer risk: a case-control study in Tunisia. Cancer Causes Control 21:393–397. [DOI] [PubMed] [Google Scholar]
- 22.Soliman AS, Schairer C (2012) Considerations in setting up and conducting epidemiologic studies of cancer in middle- and low-income countries: the experience of a case-control study of inflammatory breast cancer in North Africa in the past 10 years. Cancer Med 1:338–349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Schairer C, Hablas A, Seif Eldein IAS, Gaafar R, Rais H, Mezlini A, Ayed FB, Ayoub WB, Benider A, Tahri A, Khouchani M, Aboulazm D, Karkouri M, Eissa S, Pfeiffer RM, Gadalla SM, Swain SM, Merajver SD, Brown LM, Soliman AS (2019) Clinico-pathologic and mammographic characteristics of inflammatory and non-inflammatory breast cancer at six centers in North Africa Breast Cancer Res Treat 176:407–417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Raghunathan TE, Lepowski JM, van Hoewyk M, Solenberger Peter (2001) A multivariate technique for multiply imputing missing values using a sequence of regression models. Survey Methodol 27:85–95. [Google Scholar]
- 25.Subramanian SV, Nandy S, Irving M, Gordon D, Davey Smith G (2005) Role of socioeconomic markers and state prohibition policy in predicting alcohol consumption among men and women in India: a multilevel statistical analysis. Bull World Health Organ 83:829–836. [PMC free article] [PubMed] [Google Scholar]
- 26.Bonnier P, Romain S, Dilhuydy JM, Bonichon F, Julien JP, Charpin C, Lejeune C, Martin PM, Piana L (1997) Influence of pregnancy on the outcome of breast cancer: A case-control study. Int J Cancer 72:720–727. [DOI] [PubMed] [Google Scholar]
- 27.Nichols HB, Schoemaker MJ, Cai J, Xu J, Wright LB, Brook MN, Jones ME, Adami HO, Baglietto L, Bertrand KA, Blot WJ, Boutron-Ruault MC, Dorronsoro M, Dossus L, Eliassen AH, Giles GG, Gram IT, Hankinson SE, Hoffman-Bolton J, Kaaks R, Key TJ, Kitahara CM, Larsson SC, Linet M, Merritt MA, Milne RL, Pala V, Palmer JR, Peeters PH, Riboli E, Sund M, Tamimi RM, Tjønneland A, Trichopoulou A, Ursin G, Vatten L, Visvanathan K, Weiderpass E, Wolk A, Zheng W, Weinberg CR, Swerdlow AJ, Sandler DP (2019) Breast cancer risk after recent childbirth: a pooled analysis of 15 prospective studies. Ann Intern Med 170:22–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Cheraghi Z, Poorolajal J, Hashem T, Esmailnasab N, Doosti Irani A (2012) Effect of body mass index on breast cancer during premenopausal and postmenopausal periods: a meta-analysis. PLoS One 7(12):e51446: doi: 1. 1371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.De Bruijn KM, Arends LR, Hansen BE, Leeflang S, Ruiter R, van Eijck CH (2013) Systematic review and meta-analysis of the association between diabetes mellitus and incidence and mortality in breast and colorectal cancer. Br J Surg 100:1421–1429. [DOI] [PubMed] [Google Scholar]
- 30.Klassen AC, Smith KC (2011) The enduring and evolving relationship between social class and breast cancer burden: A review of the literature. Cancer Epidemiol 35(3):217–234. [DOI] [PubMed] [Google Scholar]
- 31.Rigby JE, Morris JA, Lavelle J, Stewart M, Gatrell AC (2002) Can physical trauma cause breast cancer? Eur J Cancer Prev 11:307–311. [DOI] [PubMed] [Google Scholar]
- 32.Lambe M, Johansson ALV, Altman D, Eloranta S (2009) Mastitis and the risk of breast cancer. Epidemiology 20:747–751. [DOI] [PubMed] [Google Scholar]
- 33.Zhou Y, Chen J, Li Q, Huang W, Lan H, Jiang H (2015) Association between breastfeeding and breast cancer risk: evidence from a meta-analysis. Breastfeed Med 10:175–182. [DOI] [PubMed] [Google Scholar]
- 34.Van Laere SJ, Van den Eynden GG, Van der Auwera I, Vandenberghe M, van Dam P, Van Marck EA, van Golen KL, Vermeulen PB, Dirix LY (2006) Identification of cell-of-origin breast tumor subtypes in inflammatory breast cancer by gene expression profiling. Breast Cancer Res Treat 95:234–255. [DOI] [PubMed] [Google Scholar]
- 35.Woodward WA (2015) Inflammatory breast cancer: unique biological and therapeutic considerations. Lancet Oncol 16(15):e568–76. doi: 10.1016/S1470-2045(15)00146-1. [DOI] [PubMed] [Google Scholar]
- 36.Alpaugh ML, Tomlinson JS, Kasraeian S, Barsky SH (2002) Cooperative role of E-cadherin and sialyl-Lewis X/A-deficient MUC1 in the passive dissemination of tumor emboli in inflammatory breast carcinoma. Oncogene 21:3631–3643. [DOI] [PubMed] [Google Scholar]
- 37.Vermeulen PB, van Golen KL, Dirix LY (2010) Angiogenesis, lymphangiogenesis, growth pattern, and tumor emboli in inflammatory breast cancer: a review of current knowledge. Cancer 116(suppl 11):2748–2754. [DOI] [PubMed] [Google Scholar]
- 38.Van der Auwera I, Van den Eynden GG, Colpaert CG, Van Laere SJ, van Dam P, Van Marck EA, Dirix LY, Vermeulen PB (2005) Tumor lymphangiogenesis in inflammatory breast carcinoma: a histomorphometric study. Clin Cancer Res 11:7637–7642. [DOI] [PubMed] [Google Scholar]
- 39.Kleer CG, van Golen KL, Merajver SD (2000) Molecular biology of breast cancer metastasis. Inflammatory breast cancer: clinical syndrome and molecular determinants. Breast Cancer Res 2:423–429. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Angelo LS, Kurzrock (2007) Vascular endothelial growth factor and its relationship to inflammatory mediators. Clin Cancer Res 13:2825–2830. [DOI] [PubMed] [Google Scholar]
- 41.Kolb R, Sutterwala FS, Weizhou Zhang (2016) Obesity and cancer: inflammation bridges the two. Curr Opin Pharmacol 29:77–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Donath MY, Shoelson SE (2011) Type 2 diabetes as an inflammatory disease. Nat Rev| Immunol 11:98–107. [DOI] [PubMed] [Google Scholar]
- 43.Ruiz R, Herrero C, Strasser-Weippl K, Touya D, St Louis J, Bukowski A, Goss PE (2017) Epidemiology and pathophysiology of pregnancy-associated breast cancer: A review. The Breast 35:136–141. [DOI] [PubMed] [Google Scholar]
- 44.Chen L, Deng H, Cui H, Fang J, Zuo Z, Deng J, Li Y, Wang X, Zhao L (2017) Inflammatory responses and inflammation-associated diseases in organs. Oncotarget 9:7204–7218. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Nishimura S, Maeno N, Matsuo K, Nakajima T, Kitajima I, Saito H, Maruyama I (2002) Human lactiferous mammary gland cells produce vascular endothelial growth factor (VEGF) and express the VEGF receptors, Flt-1 and KDR/Flk-1. Cytokine 18:191–198. [DOI] [PubMed] [Google Scholar]
- 46.Anderson KN, Schwab RB, Martinez ME (2014) Reproductive risk factors and breast cancer subtypes: a review of the literature. Breast Cancer Res Treat 144:1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Suzuki R, Orsini N, Saji S, Key TJ, Wolk A (2009) Body weight and incidence of breast cancer defined by estrogen and progesterone receptor status – A meta-analysis. Int J Cancer 124:698–712. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data are not available in a public database, but could be made available upon reasonable request.