306
Research paper
Influence of factors on mammographic density in premenopausal Chinese women Yaping Yang*, Jieqiong Liu*, Ran Gu, Yue Hu, Fengtao Liu, Miaomiao Yun, Qiaozhen Xiao, Mei Wu, Qiang Liu and Fengxi Su Mammographic density is an independent strong risk factor for breast cancer. However, the influence of factors on mammographic density in premenopausal women remains unclear. In the Southern Professional Women Breast Cancer Screening Project, we assessed the associations between mammographic density and its influential factors using multivariate logistic regression in premenopausal women adjusting for BMI, age, duration of breastfeeding, number of live births, and breast size. A total of 1699 premenopausal women aged 27 to 57 years, who had been screened by mammography, were enrolled in this cross-sectional study. Overall, 85.2% were categorized as having dense breasts (BI-RADS density 3 and 4) and 14.8% as having fatty breasts (BI-RADS density 1 and 2). In multivariate and logistic regression analysis, only BMI and age were significantly negatively correlated with mammographic density in premenopausal women (P < 0.001). No significant associations between mammographic density and number of deliveries, breastfeeding duration, education level, family history of breast cancer, as well as breast size and sleep quality, were identified in the study. Age and BMI are
Introduction Breast density assessed by mammography is an independent, strong, and consistent risk factor for breast cancer (McCormack and dos Santos Silva, 2006; Boyd et al., 2007; Boyd et al., 2010). In a systematic metaanalysis of data for over 14 000 cases and 226 000 controls from 42 studies, the risk of developing breast cancer in women with extremely dense breasts was about 4.64 times greater than that in women with fatty or nondense breasts (McCormack and dos Santos Silva, 2006). Therefore, evaluation of mammographic density might be significant as a predictor of the risk of breast cancer, and factors influencing breast density may also contribute toward the risk of breast cancer. Thus, these influential factors are of public health importance for cancer prevention in the instances when the risk factors are modifiable. The association between mammographic density and influential factors has been extensively studied; however, these studies included limited numbers of premenopausal women (around 15 to 44.4% in the total study populations), and none of the studies evaluated the influence of breast size or sleep quality on mammographic density (Vachon et al., 2000; Wong et al., 2011; 0959-8278 Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
negatively associated with mammographic density in premenopausal Chinese women. Information on the influential factors of mammographic density in premenopausal women might provide meaningful insights into breast cancer prevention. European Journal of Cancer Prevention 25:306–311 Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved. European Journal of Cancer Prevention 2016, 25:306–311 Keywords: breast density, mammography, screening Guangdong Provincial Key Laboratory of Malignant Tumor Epigenetics and Gene Regulation, Department of Breast Surgery, Breast Tumor Center, Sun Yat-Sen Memorial Hospital, Sun Yat-Sen University, Guangzhou, China Correspondence to Fengxi Su, MD, Guangdong Provincial Key Laboratory of Malignant Tumor Epigenetics and Gene Regulation, Department of Breast Surgery, Breast Tumor Center, Sun Yat-Sen Memorial Hospital, Sun Yat-Sen University, Guangzhou, 510120, China Tel: + 0086 020 34071156; fax: + 0086 020 34071156; e-mail: [email protected] *Yaping Yang and Jieqiong Liu contributed equally to the writing of this article. Received 11 January 2015 Accepted 14 May 2015
Checka et al., 2012; Woolcott et al., 2012; Dai et al., 2014). Furthermore, the mean age at diagnosis of breast cancer in China is about 10 years earlier than that in Europe and the USA; thus, a large proportion of Chinese breast cancer patients are premenopausal (Zheng et al., 2012). Studies have reported that Asian women have a higher proportion of dense breasts [Breast Imaging Reporting and Data System (BI-RADS) density 3 and 4] than other races (Ziv et al., 2004; del Carmen et al., 2007). Therefore, studies addressing the influential factors on mammographic density in premenopausal women are urgently needed for breast cancer prevention in China. To determine the effects of lifestyle factors, age, BMI as well as breast size and sleep quality on mammographic density in premenopausal women, we carried out a crosssectional study in a Chinese population.
Methods Study population
This study was approved by the Sun Yat-Sen Memorial Hospital Institutional Review Board. All participants provided written informed consent in the study. The participants were identified from the Southern Professional Women Breast Cancer Screening Project, DOI: 10.1097/CEJ.0000000000000177
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Influence of factors on mammographic density Yang et al. 307
which was conducted between 1 August 2012 and 30 April 2014. A face-to-face interview questionnaire was performed. The questionnaire included information on socioeconomic status, family history of breast cancer, multiple lifestyle factors (such as smoking, alcohol intake, and use of oral contraceptives), and sleeping problems. Participants were asked to report their age at first birth, number of deliveries, breastfeeding history, menstrual status, and breast size. Age at screening, body weight (kg), and height (m) were determined by personal report. BMI was calculated as the weight divided by the height squared (kg/m2). For the present study, we included premenopausal women who had no history of breast cancer. Women with regular menses/periods without the use of hormone replacement therapy were classified as premenopausal. A total of 1699 participants were enrolled in this study; they were 27 to 55 years old at the time of first screening. Mammographic density
The mediolateral oblique and craniocaudal views, which are the routine views for breast cancer screening, were used to assess mammographic density. Two radiologists who had 3 years’ experience read the mammographic images separately to assess the breast density. If their assessments differed, the views were re-evaluated and the final judgment on breast density was made after discussion. When the density of the two breasts differed, the higher value was used as the final breast density. Breast density was classified into four categories following the BI-RADS: category 1, breast tissue that was less than 25% glandular or almost entirely fat; category 2, breast tissue that was ∼ 25–50% glandular or scattered fibroglandular; category 3, breast tissue that was 51–75% glandular or heterogeneously dense; and category 4, breast tissue that was more than 75% glandular or extremely dense. Statistical analyses
The software Epidata, version 3.1 (The EpiData Association, Jens Lauritsen, Enghavevej 34, DK5230 Odense M, Denmark) was used to create the study’s primary database. The χ2-test or Fisher’s exact test was used to compare the distribution of all factors (categorical) among four BI-RADS categories. These factors included age (< 45 and ≥ 45), BMI (< 18.5, 18.5–25, > 25), marital status (others and married), and education (high school or below and college). Reproductive factors were related to parity (parous and nulliparous), number of deliveries (≤ 1 and > 1), and age at first birth (< 30 and ≥ 30 years). Breastfeeding, life factors, family history of breast cancer, oral contraceptive use, and breast size were also dichotomized. The associations between factors and breast density were represented by odds ratios with 95% confidence intervals estimated by univariate logistic regression models and
stratified by age. In this model, breast density was analyzed as a binary outcome: dense breasts (BI-RADS category 3, 4) and fatty breasts (BI-RADS category 1, 2). Factors with statistical significance (P < 0.05) in univariate regression were used as covariates to other adjusted factors by multivariate logistic regression. Reported P values were two-sided and considered statistically significant at less than 0.05. All data analyses were carried out using SPSS, 19.0 for Windows (IBM Corporation, Armonk, New York, USA).
Results Baseline characteristics of the study population
A total of 1699 participants provided consent to participate in the current study. The entire study population was premenopausal, and the mean age at screening was 43.2 ± 0.2 years (range from 27 to 55 years). The participants had a mean BMI of 22.5 ± 0.1 (range from 15.6 to 38.5). In all, 52.6% of the study population had college/ university education, and very few smoked cigarettes (1.0%) or consumed alcohol (24.5%). Overall, 94.5% of the participants were parous, among whom 85.9% had ever breastfed a baby. About 4.8% of the participants reported a family history of breast cancer. Only 2.3% of women had ever taken oral contraceptive pills. Breast density distribution and factors influencing mammographic density
Overall, 252 (14.8%) women in the current study showed nondense breasts (BI-RADS 1 + BI-RADS 2), whereas 1447 (85.2%) participants showed dense breasts (BIRADS 3 + BI-RADS 4). In all, 1020 women were younger than 45 years old, and among these, 897 (87%) had dense breasts. In all, 679 women were 45 years of age or older, among whom 560 (82.5%) participants had dense breasts. For women with a recommendation for tissue sampling during the screening, one woman in the nondense breasts group had breast cancer (0.39%), whereas nine women in the dense breasts group were diagnosed with breast cancer (0.62%). Lifestyle factors, age, BMI, education level, family history of breast cancer as well as breast size and sleep quality were compared between women with dense breasts and those with fatty breasts (data are shown in Table 1). In summary, age and BMI were associated negatively with mammographic density (P < 0.001). Higher education level was correlated positively with mammographic density (P < 0.05). More live births (for number > 1 compared with ≤ 1) and longer breastfeeding duration (for duration ≥ 8 months compared with < 8 months) were associated negatively with mammographic density (P < 0.05). Larger breast size (C cup or larger compared with B cup or smaller) was correlated negatively with mammographic density as well (P < 0.05). Mammographic density showed no significant associations with cigarette smoking or alcohol drinking habits, or
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
308
European Journal of Cancer Prevention 2016, Vol 25 No 4
Table 1
Association of potential influential factors and mammographic density BI-RADS density category
Characteristics Age < 45 ≥ 45 BMI < 18.5 18.5–25 > 25 Marital status Others Married Education High school or below College/university Parity Parous Nulliparous Number of deliveries ≤1 >1 Age at first birth < 30 ≥ 30 Ever breastfeeding Yes No Duration of breastfeeding (months) 25 Marital status Married vs. others Education University vs. high school or below Parity Nulliparous vs. parous Number of deliveries > 1 vs. 1 Age at first birth ≥ 30 vs. < 30 Ever breastfeeding No vs. yes Duration of breastfeeding (months) ≥ 8 vs. < 8 Oral contraceptive Never vs. ever Family history of breast cancer No vs. yes Ever alcohol drinking No vs. yes Ever smoking No vs. yes Sleeping problems (self-report) Yes vs. no Breast size ≥ C cup vs. ≤ B cup
1.92 (1.16–3.17) RE 0.46 (0.30–0.70)
P value 0.011
Age ≥ 45
P value 0.014
Total
P value < 0.001
< 0.001
2.67 (1.22–5.85) RE 0.39 (0.24–0.61)
< 0.001
2.18 (1.43–3.32) RE 0.39 (0.28–0.53)
0.77 (0.51–1.19)
0.275
1.00 (0.60–1.66)
0.997
0.90 (0.65–1.24)
0.516
0.97 (0.74–1.26)
0.800
1.09 (0.78–1.53)
0.626
1.17 (0.96–1.43)
0.126
0.92 (0.55–1.53)
0.742
0.86 (0.37–1.98)
0.719
0.96 (0.62–1.47)
0.844
0.66 (0.48–0.91)
0.011
0.98 (0.67–1.44)
0.924
0.75 (0.59–0.96)
0.021
1.11 (0.74–1.65)
0.618
0.77 (0.43–1.38)
0.374
1.01 (0.73–1.39)
0.954
0.84 (0.58–1.21)
0.349
1.50 (0.94–2.38)
0.086
1.05 (0.79–1.40)
0.722
0.78 (0.58–1.05)
0.096
0.86 (0.59–1.25)
0.416
0.75 (0.60–0.94)
0.012
0.99 (0.44–2.25)
0.996
0.84 (0.28–2.54)
0.757
0.92 (0.48–1.77)
0.811
1.02 (0.59–1.78)
0.939
0.88 (0.40–1.93)
0.756
0.95 (0.60–1.49)
0.813
1.04 (0.72–1.49)
0.836
1.19 (0.69–2.07)
0.528
0.99 (0.74–1.33)
0.936
2.25 (0.58–8.79)
0.231
NA
0.109
1.18 (0.34–4.06)
0.793
0.87 (0.68–1.11)
0.263
0.94 (0.67–1.30)
0.696
0.90 (0.74–1.10)
0.301
0.76 (0.53–1.08)
0.120
0.49 (0.30–0.79)
0.003
0.62 (0.47–0.82)
0.001
< 0.001
CI, confidence interval; NA, not available; RE, reference group.
Table 3
Influence of BMI and age on mammographic density Adjusted odds ratio (95% CI)
Characteristics
Model 1
BMI < 18.5 2.10 (1.21–3.64) 18.5–25 RE > 25 0.44 (0.30–0.64) Age ≥ 45 vs. < 45 0.53 (0.40–0.69) Number of deliveries > 1 vs. ≤ 1 0.82 (0.59–1.15) Duration of breastfeeding (months) ≥ 8 vs.
Research paper
Influence of factors on mammographic density in premenopausal Chinese women Yaping Yang*, Jieqiong Liu*, Ran Gu, Yue Hu, Fengtao Liu, Miaomiao Yun, Qiaozhen Xiao, Mei Wu, Qiang Liu and Fengxi Su Mammographic density is an independent strong risk factor for breast cancer. However, the influence of factors on mammographic density in premenopausal women remains unclear. In the Southern Professional Women Breast Cancer Screening Project, we assessed the associations between mammographic density and its influential factors using multivariate logistic regression in premenopausal women adjusting for BMI, age, duration of breastfeeding, number of live births, and breast size. A total of 1699 premenopausal women aged 27 to 57 years, who had been screened by mammography, were enrolled in this cross-sectional study. Overall, 85.2% were categorized as having dense breasts (BI-RADS density 3 and 4) and 14.8% as having fatty breasts (BI-RADS density 1 and 2). In multivariate and logistic regression analysis, only BMI and age were significantly negatively correlated with mammographic density in premenopausal women (P < 0.001). No significant associations between mammographic density and number of deliveries, breastfeeding duration, education level, family history of breast cancer, as well as breast size and sleep quality, were identified in the study. Age and BMI are
Introduction Breast density assessed by mammography is an independent, strong, and consistent risk factor for breast cancer (McCormack and dos Santos Silva, 2006; Boyd et al., 2007; Boyd et al., 2010). In a systematic metaanalysis of data for over 14 000 cases and 226 000 controls from 42 studies, the risk of developing breast cancer in women with extremely dense breasts was about 4.64 times greater than that in women with fatty or nondense breasts (McCormack and dos Santos Silva, 2006). Therefore, evaluation of mammographic density might be significant as a predictor of the risk of breast cancer, and factors influencing breast density may also contribute toward the risk of breast cancer. Thus, these influential factors are of public health importance for cancer prevention in the instances when the risk factors are modifiable. The association between mammographic density and influential factors has been extensively studied; however, these studies included limited numbers of premenopausal women (around 15 to 44.4% in the total study populations), and none of the studies evaluated the influence of breast size or sleep quality on mammographic density (Vachon et al., 2000; Wong et al., 2011; 0959-8278 Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.
negatively associated with mammographic density in premenopausal Chinese women. Information on the influential factors of mammographic density in premenopausal women might provide meaningful insights into breast cancer prevention. European Journal of Cancer Prevention 25:306–311 Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved. European Journal of Cancer Prevention 2016, 25:306–311 Keywords: breast density, mammography, screening Guangdong Provincial Key Laboratory of Malignant Tumor Epigenetics and Gene Regulation, Department of Breast Surgery, Breast Tumor Center, Sun Yat-Sen Memorial Hospital, Sun Yat-Sen University, Guangzhou, China Correspondence to Fengxi Su, MD, Guangdong Provincial Key Laboratory of Malignant Tumor Epigenetics and Gene Regulation, Department of Breast Surgery, Breast Tumor Center, Sun Yat-Sen Memorial Hospital, Sun Yat-Sen University, Guangzhou, 510120, China Tel: + 0086 020 34071156; fax: + 0086 020 34071156; e-mail: [email protected] *Yaping Yang and Jieqiong Liu contributed equally to the writing of this article. Received 11 January 2015 Accepted 14 May 2015
Checka et al., 2012; Woolcott et al., 2012; Dai et al., 2014). Furthermore, the mean age at diagnosis of breast cancer in China is about 10 years earlier than that in Europe and the USA; thus, a large proportion of Chinese breast cancer patients are premenopausal (Zheng et al., 2012). Studies have reported that Asian women have a higher proportion of dense breasts [Breast Imaging Reporting and Data System (BI-RADS) density 3 and 4] than other races (Ziv et al., 2004; del Carmen et al., 2007). Therefore, studies addressing the influential factors on mammographic density in premenopausal women are urgently needed for breast cancer prevention in China. To determine the effects of lifestyle factors, age, BMI as well as breast size and sleep quality on mammographic density in premenopausal women, we carried out a crosssectional study in a Chinese population.
Methods Study population
This study was approved by the Sun Yat-Sen Memorial Hospital Institutional Review Board. All participants provided written informed consent in the study. The participants were identified from the Southern Professional Women Breast Cancer Screening Project, DOI: 10.1097/CEJ.0000000000000177
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
Influence of factors on mammographic density Yang et al. 307
which was conducted between 1 August 2012 and 30 April 2014. A face-to-face interview questionnaire was performed. The questionnaire included information on socioeconomic status, family history of breast cancer, multiple lifestyle factors (such as smoking, alcohol intake, and use of oral contraceptives), and sleeping problems. Participants were asked to report their age at first birth, number of deliveries, breastfeeding history, menstrual status, and breast size. Age at screening, body weight (kg), and height (m) were determined by personal report. BMI was calculated as the weight divided by the height squared (kg/m2). For the present study, we included premenopausal women who had no history of breast cancer. Women with regular menses/periods without the use of hormone replacement therapy were classified as premenopausal. A total of 1699 participants were enrolled in this study; they were 27 to 55 years old at the time of first screening. Mammographic density
The mediolateral oblique and craniocaudal views, which are the routine views for breast cancer screening, were used to assess mammographic density. Two radiologists who had 3 years’ experience read the mammographic images separately to assess the breast density. If their assessments differed, the views were re-evaluated and the final judgment on breast density was made after discussion. When the density of the two breasts differed, the higher value was used as the final breast density. Breast density was classified into four categories following the BI-RADS: category 1, breast tissue that was less than 25% glandular or almost entirely fat; category 2, breast tissue that was ∼ 25–50% glandular or scattered fibroglandular; category 3, breast tissue that was 51–75% glandular or heterogeneously dense; and category 4, breast tissue that was more than 75% glandular or extremely dense. Statistical analyses
The software Epidata, version 3.1 (The EpiData Association, Jens Lauritsen, Enghavevej 34, DK5230 Odense M, Denmark) was used to create the study’s primary database. The χ2-test or Fisher’s exact test was used to compare the distribution of all factors (categorical) among four BI-RADS categories. These factors included age (< 45 and ≥ 45), BMI (< 18.5, 18.5–25, > 25), marital status (others and married), and education (high school or below and college). Reproductive factors were related to parity (parous and nulliparous), number of deliveries (≤ 1 and > 1), and age at first birth (< 30 and ≥ 30 years). Breastfeeding, life factors, family history of breast cancer, oral contraceptive use, and breast size were also dichotomized. The associations between factors and breast density were represented by odds ratios with 95% confidence intervals estimated by univariate logistic regression models and
stratified by age. In this model, breast density was analyzed as a binary outcome: dense breasts (BI-RADS category 3, 4) and fatty breasts (BI-RADS category 1, 2). Factors with statistical significance (P < 0.05) in univariate regression were used as covariates to other adjusted factors by multivariate logistic regression. Reported P values were two-sided and considered statistically significant at less than 0.05. All data analyses were carried out using SPSS, 19.0 for Windows (IBM Corporation, Armonk, New York, USA).
Results Baseline characteristics of the study population
A total of 1699 participants provided consent to participate in the current study. The entire study population was premenopausal, and the mean age at screening was 43.2 ± 0.2 years (range from 27 to 55 years). The participants had a mean BMI of 22.5 ± 0.1 (range from 15.6 to 38.5). In all, 52.6% of the study population had college/ university education, and very few smoked cigarettes (1.0%) or consumed alcohol (24.5%). Overall, 94.5% of the participants were parous, among whom 85.9% had ever breastfed a baby. About 4.8% of the participants reported a family history of breast cancer. Only 2.3% of women had ever taken oral contraceptive pills. Breast density distribution and factors influencing mammographic density
Overall, 252 (14.8%) women in the current study showed nondense breasts (BI-RADS 1 + BI-RADS 2), whereas 1447 (85.2%) participants showed dense breasts (BIRADS 3 + BI-RADS 4). In all, 1020 women were younger than 45 years old, and among these, 897 (87%) had dense breasts. In all, 679 women were 45 years of age or older, among whom 560 (82.5%) participants had dense breasts. For women with a recommendation for tissue sampling during the screening, one woman in the nondense breasts group had breast cancer (0.39%), whereas nine women in the dense breasts group were diagnosed with breast cancer (0.62%). Lifestyle factors, age, BMI, education level, family history of breast cancer as well as breast size and sleep quality were compared between women with dense breasts and those with fatty breasts (data are shown in Table 1). In summary, age and BMI were associated negatively with mammographic density (P < 0.001). Higher education level was correlated positively with mammographic density (P < 0.05). More live births (for number > 1 compared with ≤ 1) and longer breastfeeding duration (for duration ≥ 8 months compared with < 8 months) were associated negatively with mammographic density (P < 0.05). Larger breast size (C cup or larger compared with B cup or smaller) was correlated negatively with mammographic density as well (P < 0.05). Mammographic density showed no significant associations with cigarette smoking or alcohol drinking habits, or
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
308
European Journal of Cancer Prevention 2016, Vol 25 No 4
Table 1
Association of potential influential factors and mammographic density BI-RADS density category
Characteristics Age < 45 ≥ 45 BMI < 18.5 18.5–25 > 25 Marital status Others Married Education High school or below College/university Parity Parous Nulliparous Number of deliveries ≤1 >1 Age at first birth < 30 ≥ 30 Ever breastfeeding Yes No Duration of breastfeeding (months) 25 Marital status Married vs. others Education University vs. high school or below Parity Nulliparous vs. parous Number of deliveries > 1 vs. 1 Age at first birth ≥ 30 vs. < 30 Ever breastfeeding No vs. yes Duration of breastfeeding (months) ≥ 8 vs. < 8 Oral contraceptive Never vs. ever Family history of breast cancer No vs. yes Ever alcohol drinking No vs. yes Ever smoking No vs. yes Sleeping problems (self-report) Yes vs. no Breast size ≥ C cup vs. ≤ B cup
1.92 (1.16–3.17) RE 0.46 (0.30–0.70)
P value 0.011
Age ≥ 45
P value 0.014
Total
P value < 0.001
< 0.001
2.67 (1.22–5.85) RE 0.39 (0.24–0.61)
< 0.001
2.18 (1.43–3.32) RE 0.39 (0.28–0.53)
0.77 (0.51–1.19)
0.275
1.00 (0.60–1.66)
0.997
0.90 (0.65–1.24)
0.516
0.97 (0.74–1.26)
0.800
1.09 (0.78–1.53)
0.626
1.17 (0.96–1.43)
0.126
0.92 (0.55–1.53)
0.742
0.86 (0.37–1.98)
0.719
0.96 (0.62–1.47)
0.844
0.66 (0.48–0.91)
0.011
0.98 (0.67–1.44)
0.924
0.75 (0.59–0.96)
0.021
1.11 (0.74–1.65)
0.618
0.77 (0.43–1.38)
0.374
1.01 (0.73–1.39)
0.954
0.84 (0.58–1.21)
0.349
1.50 (0.94–2.38)
0.086
1.05 (0.79–1.40)
0.722
0.78 (0.58–1.05)
0.096
0.86 (0.59–1.25)
0.416
0.75 (0.60–0.94)
0.012
0.99 (0.44–2.25)
0.996
0.84 (0.28–2.54)
0.757
0.92 (0.48–1.77)
0.811
1.02 (0.59–1.78)
0.939
0.88 (0.40–1.93)
0.756
0.95 (0.60–1.49)
0.813
1.04 (0.72–1.49)
0.836
1.19 (0.69–2.07)
0.528
0.99 (0.74–1.33)
0.936
2.25 (0.58–8.79)
0.231
NA
0.109
1.18 (0.34–4.06)
0.793
0.87 (0.68–1.11)
0.263
0.94 (0.67–1.30)
0.696
0.90 (0.74–1.10)
0.301
0.76 (0.53–1.08)
0.120
0.49 (0.30–0.79)
0.003
0.62 (0.47–0.82)
0.001
< 0.001
CI, confidence interval; NA, not available; RE, reference group.
Table 3
Influence of BMI and age on mammographic density Adjusted odds ratio (95% CI)
Characteristics
Model 1
BMI < 18.5 2.10 (1.21–3.64) 18.5–25 RE > 25 0.44 (0.30–0.64) Age ≥ 45 vs. < 45 0.53 (0.40–0.69) Number of deliveries > 1 vs. ≤ 1 0.82 (0.59–1.15) Duration of breastfeeding (months) ≥ 8 vs.
