• 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • 2021-03
  • br As stage was such a strong predictor of


    As stage was such a strong predictor of survival for women with non-endometrioid cancers (Stage 3, HR 7.1; 95% CI 4.9), the models were further stratified by this factor. In women with non-endometrioid histology and extra-uterine disease at presentation (Stage 3 and 4), age was no longer significantly associated with DSS. The only remaining factors independently associated with a decreased hazard of DSS were treatment with both chemotherapy and radiation and statin use (Table 4). Although BMI overall was non-significant, the 2 highest BMI groups (BMI 35–39.9 and N40) had a decreased hazard of death compared to women with a normal BMI (p = 0.01 and p = 0.08, respectively).
    In our study, the sub-cohort most likely to be effected by high risk features was identified to be endometrioid type lesions, clinically con-fined to the uterus (Stage I). The high risk features that were evaluated included depth of invasion by percentage myometrial invasion, lymph vascular space invasion and tumor grade. Of 998 women identified to have endometrioid type lesions confined to the uterus, data were evaluable for 934 (94%). In univariable analysis, for women with Stage 1 endometrioid histology, body mass index was associated with depth
    Table 3
    Univariable and multivariable Cox proportional hazards regression models stratified by histology.
    Variable Endometrioid histology
    Non-endometrioid histology
    Univariable Hazard Ratio Multivariable Hazard Ratio
    Univariable Hazard Ratio Multivariable Hazard Ratio
    White/Caucasian 1.00 (reference)
    Statin use
    of invasion, grade, and lymph vascular invasion (p b 0.05) Higher body mass index was associated with lower grade, less invasion, and absence of lymph vascular space invasion. After controlling for all three high risk factors, only grade was independently associated with body mass index (p b 0.05), with higher body mass index exhibiting Grade 1 disease. Race was also independently associated with body mass index in this cohort of women with Stage I endometrioid histology, with black and hispanic women having on average higher body mass index than white women. Race, however, did not alter the associations of high-risk variables with obesity. There was no association seen for Cyclosporin H or progesterone receptor positivity and body mass index. This association was not influ-enced by race. Only 21 (2%) of 934 patients had Lynch testing, and therefore no meaningful associations to obesity or race could be concluded.
    In our diverse population there were differences in histologies asso-ciated with race. Black women represented only 24% of the women with endometrioid histologies and 56% of the women with non-endometrioid histologies. In the overall cohort, endometrioid only and 
    non-endometrioid histologies, there was a significant association with race and body mass index. Women of black race were more likely to have body mass indices in the highest categories (Table 1). For women with endometrioid type cancer 75% of black women had body mass indices over 30 mg/kg2 in comparison to 64% of white women. For non-endometrioid cancers, 62% of black women had body mass in-dices over 30 kg/m2 as compared to 44% of white women. When data for the entire cohort were analyzed in a univariable fashion black race was significantly associated with increased hazard of death, (HR 1.95, 95% CI 1.53–2.5 p b 0.001), but when other variables were included in multi-variable analysis black race was not shown to be associated with sur-vival (HR 0.99, 95% CI 0.75–1.31, p = 0.96).(Table 2) Similarly there was no increased hazard for death associated with race when histolo-gies of endometrioid versus non-endometrioid were considered sepa-rately. (Table 3) There was no alteration of the association of high risk factors of depth of invasion, tumor grade and lymph vascular space invasion associated with race for early stage endometrioid cancers con-fined to the uterus.
    Table 4
    Cox regression analysis limited to non-endometrioid cancer stratified stage.
    Variable Type 2 Endometrial Cancers Stage 1/2
    Type 2 Endometrial Cancers Stage 3/4
    Univariable Hazard Ratio Multivariable Hazard Ratio
    Univariable Hazard Ratio Multivariable Hazard Ratio
    White/Caucasian 1.00 (reference)
    Other –
    Statin use
    Beta blocker
    Treatment Overall Wald p = 0.17
    4. Discussion
    Age at diagnosis was younger for patients in the highest BMI catego-ries for the entire cohort and subgroups of endometrioid and non-endometrioid cancers. In the entire cohort and subgroups of endometrioid and non-endometrioid histologies this relationship was linear. In multivariable analysis obesity was not associated with overall or disease specific survival, but factors of age, stage, grade, statin use, anti-hyperglycemic, and adjuvant therapy remained significant. How-ever, in a subgroup of non-endometrioid patients, with Stage 3 and 4 disease, age was no longer associated with survival and patient in the obese category had decreased hazard of death. In our dataset, obese patients did not have a higher overall hazard of death from co-morbidities frequently associated with obesity such as cardiovascular death. Additionally, when the presence of high-risk features was evalu-ated for association with obesity only Grade 1 disease was associated with higher BMI, which may suggest metabolic pathways important in pathogenesis. These data, in particular the impact of statin and anti-hyperglycemic medications, suggest the potential importance of anti-inflammatory and metabolic pathways as well as estrogen pathways in endometrial cancer, whereby obese women have high levels of en-dogenous estrogen due to the conversion of androsteedione to estrone and the aromatizaion of androgens to estradiol in peripheral adipose tissue.