Breast cancer (BC) is the most common cancer for women worldwide [1]. In the United States in 2018, there were an estimated 268,670 new BC cases and 40,920 deaths from BC causes [2]. BC prognosis is dependent on multiple factors including tumor stage, hormone receptor status, lymph node involvement, but also patient factors, like obesity, may also influence prognosis. Obesity has been on the rise globally [3], contributing to significant patient morbidity and increased risk of premature death in patients in the highest and lowest body mass index (BMI) categories [4]. Obesity increases the risk of developing multiple cancers, especially sex-hormone related cancers like BC. Obesity has been studied as a risk factor and poor prognosis indicator for BC. In multiple meta-analyses, patients with obesity had worse disease-free survival, overall survival [6, 7] and this persisted across all subtypes [8]. Higher BMI has also been associated with larger tumor size, greater regional/distant stage [5], more lymph node metastases [9] and increased risk of recurrence [10].

However, the relationship between BMI and BC remains complex, particularly with respect to menopausal status. In premenopausal women, an inverse relationship has been found between weight and BC development [11]. In postmenopausal women, women who were obese had over two-fold increased risk of BC development [12, 13]. Regardless, once diagnosed, women of all ages with obesity have been found to have worse BC specific and overall survival [14, 15]. The association between BMI with BC stage at diagnosis at all ages and subsequent survival has not been studied.

Therefore, in this study, we aimed to investigate how differences in BMI impact breast cancer stage at diagnosis and survival. Using a large population database, the California Cancer Registry (CCR), we analyzed associations between BMI at breast cancer diagnosis, stage of diagnosis and survival among women with breast cancer. The purpose of this study was to better understand the relationship of BMI with stage of breast cancer diagnosis and survival to identify high risk groups and improve care.

Our study consisted of 159,248 female patients diagnosed with a primary invasive BC. Selected characteristics are shown in Table 1. Among patients in our study, 2.2% were underweight, 34.6% normal weight, 30.5% overweight, 26.7% obesity class 1–2, and 6.0% obesity class 3. The distribution of BMI categories pre-and post-imputation were equivalent (Supplemental Table 1). In the cohort, 75.3% of women were age ≥ 51 years old, 39.8% lived in a high SES neighborhood and 56.6% had private insurance. 53.3% were non-Hispanic (nH) white, 22.1% Hispanic, 16.4% Asian/Pacific Islander, and 6.2% nH Black. 12.3% had a late stage BC diagnosis (AJCC stage III and IV). Most were treated surgically (89.9%), 31.9% received chemotherapy, and 48.3% radiation therapy. In total, 6.9% had died from BC and 5.6% from other causes.

Patients who were underweight (vs normal weight) had the highest risk of late stage at diagnosis [Odds Ratio (OR) 1.54, 95% CI 1.51–1.57] (Table 2). Obesity class 1–2 and class 3 were also associated with late stage at diagnosis [OR 1.06, 1.05–1.07 and OR 1.14, 1.12–1.16], but these associations were not observed among patients who were overweight (p = 0.11). These results were similar when looking at the pre-and post-imputation data (Supplemental Table 2).

In models stratified by age (Table 4), underweight (vs normal weight) patients had worse BCSS in the 40–50 age group (Hazard Ratio (HR) 1.43, 1.01–2.03) and ≥ 51 age group (HR 1.18, 1.03–1.36), but not the < 40 age group (p = 0.91). Overweight or class 3 obesity were not associated with worse BCSS in any of the age groups. Patients who were categorized as obesity class 1–2 had a better BCSS in the ≥ 51 age group [HR 0.93, 0.87–0.99], but not in the younger age categories. Results for BCSS prior to imputation were similar (Supplemental Table 3).

Unadjusted 5-year BCSS rates stratified by age identified that patients in the age group 40–50 are estimated to have better survival across all weight categories compared to ages 15–39 and ≥ 51 year-olds (Table 3; Fig. 1). Patients in the age group 15–39 were estimated to have worse survival if overweight, with obesity class 1–2, or class 3 compared to normal weight or underweight (Fig. 1). Patients with normal weight were shown to have better survival across all age groups, especially in the age group 40–50. Underweight patients were shown to have decreased 5-year BCSS in both the 40–50 [90.4% (CI 89.9%, 90.9%)] and the ≥ 51 [71.8% (CI 71.3%,72.2%)] age groups. Patients with obesity class 3 were also estimated to have worse 5-year survival across all age groups, but for age ≥ 51, underweight patients had worse survival than patients with obesity class 3.

In this large population-based study, we observed that patients with obesity, who comprised 33% of our cohort, were more likely to be diagnosed with a later stage of BC, adding to a growing literature that obesity is a risk factor for advanced BC. However, patients who were overweight or obesity class 3 did not have worse BCSS than normal weight women. Instead, patients 51 years of age or older with obesity class 1–2 had improved BCSS, supporting an “obesity paradox” in BC. Our study also identified a high-risk population of BC patients, approximately 2% of our cohort, who were underweight and more likely as well to be diagnosed at a late stage. In addition, underweight women over 40 years of age at diagnosis experienced worse BCSS, highlighting the need for interventions to improve outcomes in this population.

Our findings align with prior work showing women with obesity are more likely to be diagnosed with BC and have more advanced BC [5, 9, 10, 15]. The reasons for this are multifactorial. First, women with obesity have excessive adipose tissue. Adipose tissue contains aromatase, which converts androstenediol into estradiol and produces estrogen. High levels of estrogen stimulate the growth of breast tumor cells [20]. Also, individuals with obesity produce more leptin than those of normal weight. Leptin induces expression of aromatase in breast cancer cells, leading to more estrogen and proliferation [21]. In addition, obese adipose tissue over-expresses inflammatory mediators like interleukin-6, tumor necrosis factor, and intercellular adhesion molecule 1 [22], a pro-inflammatory state that stimulates tissue growth [23]. Patients with obesity frequently have imbalances in insulin and insulin-like growth factor, which also stimulate tumor growth and inhibit apoptosis [24]. Finally, patients with obesity typically have larger breast size, which can reduce the accuracy and sensitivity of ultrasound or other examinations, potentially leading to missed BC diagnoses [25]. Overall, our finding of women with obesity being diagnosed with a later stage of BC [26] adds to the growing body of literature that being obese is a risk factor for the development of advanced BC.

However, unlike prior studies, [5–8] we found that patients with obesity class 1–2 had better BCSS in patients over 50 years of age. One reason for this could be the “obesity paradox” seen in other cancers, including gynecological and colorectal cancers [27], but previously not seen with BC [28]. This paradox refers to patients who are over-weight or obese being more likely to develop a disease, but also more likely to have better survival than their normal counterparts [28]. The mechanisms underlying these associations are still unknown, but potential theories have been explored. Lee and Giovannucci touch on the theory of “cancer cachexia,” which describes an inaccurate categorization of patients into a normal or underweight BMI category based on the observation that cancer patients with advanced disease have often already lost a lot of weight prior to diagnosis. While these patients might have been overweight or obese before, due to this cachectic state, they are now labeled as normal or even underweight and have a higher risk of mortality. This supports the paradox that obese patients are more likely to have a better survival, even though this may not be true [29]. Also, patients with a high BMI have an increased metabolic buffer that can protect them from this catabolic state [28] and provide supportive nutrients during cancer treatments leading to a better tolerance of cancer cachexia and subsequent treatments [30]. Prior literature has shown that the “obesity paradox” is a more beneficial in older individuals [31]. That is, older individuals who have obesity may have survival benefits due to protection against protein-energy wasting and the nutritional insult of a cancer diagnosis [31].

Another consideration is the use of BMI. BMI is the most widely used tool to assess adiposity in obesity with clinical outcomes [32]; however, it is limited by its inability to discriminate body composition, adipose distribution, and cardiovascular health [33, 34]. Prior work has shown that people with the same BMI measurement can have vastly different body compositions which leads to different metabolic profiles and the negative health associations commonly attributed to BMI and weight [35]. The different distributions of adipose tissue are variably tied to adverse effects due to differences in adipokine synthesis [36]. Additionally, being cardiovascular unfit, regardless of one’s BMI, is associated with twice the risk of death compared to normal weight patients [37, 38]. Using solely BMI, and not considering leanness and adipose distribution, likely contributes to less accurate results and contributing to an “obesity paradox” [29]. Future studies should incorporate other methods to assess body composition, adipose distribution and metabolic health, like waist-to-height ratio or waist circumference [39].

Patients who were underweight had the highest risk of late-stage disease. Prior work has shown that underweight women have lower odds of complying with regular screening mammograms [40, 41], potentially contributing to this finding. We also found underweight women had worse BCSS, consistent with prior studies [42, 43]. Underweight may be associated with sarcopenia, cancer cachexia, or undernourishment, which are conditions that decrease the efficacy and potency of chemotherapy [44, 45]. Sarcopenia also carries a high risk of morbidity and mortality in a variety of health conditions [35], potentially contributing to the worse survival observed in our study. We found for patients over the age of 39 years, being underweight is associated with worse survival, indicating that this high-risk group may benefit from screening for sarcopenia and malnourishment for early physical therapy or registered dietician interventions to improve outcomes.

This study has some important limitations. We had a high percentage of missing BMI data (36.5%) but were able to impute the missing information using a validated method [19]. In addition, BMI data was only recorded at BC diagnosis, which does not account for BMI changes in the time prior to cancer diagnosis or during treatment, an important consideration for future studies. With the increased use of glucagon-like peptide-1 (GLP-1) receptor agonists to treat obesity, their effects on BC stage and BCSS remain unknown. While our dataset did not include the use of GLP-1 receptor agonists, further exploration into the potential association of weight loss via the use of GLP-1 receptor agonists or bariatric surgery in BC outcomes may be useful. Despite these limitations, our large population-based study has added to the complex picture concerning BMI, stage of BC diagnosis, and BCSS.

Patients with obesity class 1–2 were more likely to be diagnosed with a later stage of diagnosis but had improved BCSS, potential evidence of an “obesity paradox” in BC. Other measures besides BMI are needed to better assess body composition, adipose distribution and metabolic health of patients. Patients who were underweight were more likely to be diagnosed late stage BC and have worse survival, indicating a high-risk group that should be assessed for early interventions to improve outcomes.

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/s10549-025-07678-7.