1. Introduction
In the United States, uterine cancer ranked as the fourth most common cancer and the seventh leading cause of cancer death in women [
1], whose incidence and mortality are increasing in contrast to other cancers [
2]. Uterine cancer is divided into uterine carcinoma and sarcoma based on histopathological characteristics. Uterine carcinoma accounts for the majority of uterine cancer, while uterine sarcoma accounts for only 3–7% of all uterine cancer [
3]. Uterine carcinoma, also known as endometrial cancer (EC) in some literature, has been classified into type 1 and type 2 based on histopathological characteristics and clinical outcomes. Among type 1 EC, endometrioid adenocarcinoma represents the most common cell type, and it is featured with a relatively favorable clinical outcome. Endometrioid adenocarcinoma is predisposed by prolonged or excessive estrogen exposure, such as the use of unopposed estrogen during hormone-replacement therapy, estrogen-producing tumors (e.g., ovarian granulosa-cell tumors) or associated with obesity [
4]. Instead, type 2 EC, including clear cell carcinoma, serous carcinoma, and carcinosarcoma, is typically poorly differentiated with a higher risk of metastasis and a poor overall outcome [
5]. In Taiwan, uterine cancer, especially endometrioid adenocarcinoma, was the most rapidly increasing malignancy at a younger age in the past 30 years [
6]. Several risk factors have been reported to be associated with an increased risk of EC development, including diabetes, obesity, old age, family history of Lynch syndrome, early menarche, nulliparity, prolonged or excessive unopposed estrogen stimulation, and late menopause. Among them, type 2 diabetes mellitus (DM) is significantly associated with a twofold higher risk of developing EC compared to individuals without DM. Moreover, DM itself has been considered as an independent risk factor of EC [
7,
8].
In Taiwan, the incidence and prevalence of DM are steadily increasing, especially in the young population [
9]. Similarly, the incidence of obesity in the young population gradually increases in parallel to the trend of DM incidence as well [
10]. Collectively, the trend of early onset of DM and obesity as well as increased incidence of EC in younger population were observed.
DM and obesity both belong to the spectrum of metabolic syndrome, which is a collection of several clinical conditions, including insulin resistance, hyperlipidemia (HPL), and hypertension. Among these, HPL is also associated with an increased risk of cancer [
11]; however, the association between HPL and EC is limited. A positive association between serum triglyceride level and risk of EC has been reported, but not observed in total cholesterol, low-density lipoprotein cholesterol (LDL) and high-density lipoprotein cholesterol (HDL) [
12].
Previously, Chen et al. reported that DM significantly increases the risk of EC in all age, and the highest relative risk was observed in women younger than 50 years of age [
13]. However, this finding should be interpreted with caution because several potential confounders, such as obesity and polycystic ovarian syndrome (PCOS), were not adjusted accordingly. PCOS is a common disease in young women that is associated with EC development [
14]. Additionally, the trend of early onset of obesity might also confound their analysis results.
Therefore, we proposed to utilize a representative nationwide database, the National Health Insurance (NHI) database, and the Taiwan Cancer Registry (TCR), to investigate the age-specific risks of uterine cancer in subjects with diabetes and related metabolic comorbidities.
4. Discussion
Uterine cancer is a worldwide public health issue. In the United States, the median age of uterine cancer diagnosis was 63 years from 2012 to 2016 [
20]. However, in Taiwan, the age of uterine cancer diagnosis was lower, as the incidence peak of uterine carcinoma was between 50 and 59 years, while between 40 and 49 years of uterine sarcoma from 1979 to 2008 [
6]. Since most uterine cancer was diagnosed before age 70, women older than 70 years were excluded in our study. Although the data were extracted from different time frames, our findings extracted from 2005 to 2015 also supported a previous report that the age group of 50–59 years had the highest risk of developing uterine cancer followed by the age group 40–49 years. Both findings suggest a trend of younger diagnosis of uterine cancer in Taiwan. The increase in uterine cancer during the last 30 years was mainly attributed to the increase in type 1 EC [
6], and our results also showed that the main subtype of uterine cancer was endometrioid adenocarcinoma. Our finding was consistent with the current global trend of carcinoma as the predominant cancer subtype in uterine cancer. In our report, we found no differences in uterine cancer subtypes between the non-DM and DM groups. This finding suggests that DM per se might not increase the incidence of a given subtype of uterine cancer in Taiwan. A similar finding was also reported in Australia that parity, oral contraceptive use, smoking, age at menarche, and DM were not associated with any given subtype of uterine cancer. However, BMI was specifically associated with type 1 tumors [
21]. However, we did not observe any associations between obesity or DM and uterine cancer subtype in Taiwanese populations.
The WHO estimated that the worldwide prevalence of DM was 171 million in 2000, and would be approximately 366 million by 2030 [
22]. In Taiwan, the prevalence of DM gradually increased and was 9.82% in subjects aged 18 years and older [
23]. From 2005 to 2014, the total population of DM in Taiwan increased by 66%, and the age-standardized prevalence increased by 41% in the population aged 20 to 79 years [
24]. The change was largely caused by the increase in the young population [
9]. Type 2 DM has been shown to be significantly associated with a higher risk of developing EC [
25]. A meta-analysis showed that women with DM had a 72% increased risk of EC compared to those without DM [
26]. The highest relative risk was seen especially in women under 50 years of age [
13]. In our results, we found that age was positively associated with uterine cancer of all age groups; in particular, the 50–59 age group was associated with the highest risk followed by age groups 40–49 and 60–69. In 2017, an association between type 2 DM and cancer incidence in Taiwan was reported, showing that younger participants (<55 years) with DM had a higher risk of all-cause cancers compared to those in the same age group without DM [
27]. Chen et al. in Taiwan also reported a similar result—that DM significantly increased the risk of endometrial cancer in women of all ages, and the highest risk of endometrial cancer was observed in those younger than 50 years [
13]. However, they did not adjust for possible confounders such as BMI, obesity, and unopposed estrogen exposure, which may also affect the incidence of uterine cancer. We observed an elevated risk of uterine cancer in women with DM of all ages. After adjusting confounders, the association between DM and uterine cancer remained in all age groups, but the effect (HR) was greater in the <50 years group than the ≥50 years group. Accordingly, DM is the most important risk factor for uterine cancer compared to other metabolic comorbidities in younger patients in our study.
We also found that obesity was associated with an increased risk of uterine cancer in both the DM and non-DM groups. Obese women were at two to four times greater risk of developing uterine cancer than those with normal body weight in the United States [
28]. In a recent umbrella review, a positive association was found between high BMI, waist-to-hip ratio, and endometrial carcinomas in premenopausal women [
29]. Plausible mechanisms were proposed that obesity increases uterine cancer risk by augmenting aromatase activity, which converts androgen to estrogen to promote endometrial proliferation [
30]. Although we and others found that obesity was positively associated with uterine cancer, this result should be interpreted with caution. First, obesity might be underestimated, since the coding of obesity other than morbid obesity tended to be missed in the medical record. Second, obesity was defined by the value of BMI; however, Asians tend to have a lower BMI but a higher percentage of body fat than the white population [
31]. In South-East Asia, waist circumference was associated with metabolic abnormalities and is more correlated with visceral adiposity than BMI [
32]. Therefore, waist circumference or waist-to-hip ratio might be a more appropriate surrogate than BMI for Asians in the diagnosis of obesity. Thus, the risk of obesity on uterine cancer might be underestimated in our population.
In addition to the associations between DM, obesity, and uterine cancer development, there was also increased cancer mortality in patients with DM after age and comorbidities were adjusted [
33]. This phenomenon could be explained by a variety of metabolic abnormalities including hyperinsulinemia, insulin resistance, increased insulin-like growth factor-1 (IGF-1) level, dyslipidemia, augmented inflammatory cytokines, elevated leptin, as well as decreased adiponectin [
33]. These metabolic abnormalities are commonly seen in PCOS, which is a common endocrine disorder in young women. A meta-analysis showed that women with PCOS were at a three times higher risk of developing endometrial cancer [
34]. In our study, the diagnosis of PCOS was not different between the DM and non-DM groups; however, PCOS was significantly associated with uterine cancer after age, DM, obesity, HPL, statin use, and HRT were adjusted. However, this association was only found in subjects <50 years since PCOS is a premenopausal disease. Specifically, the presence of PCOS would further increase risk of uterine cancer; however, the effect was modest in subjects with DM. This might suggest that PCOS also behaves as a risk factor, but it was overshadowed by DM if both diagnoses were present in analysis. Alternatively, it can be explained that the small sample sizes in the DM subjects with PCOS render the non-significance. The effects of PCOS on uterine cancer might be explained by the aforementioned metabolic abnormalities PCOS exhibits that lead to an unopposed estrogen and increased risk of uterine cancer [
35].
In our cohort, nearly 40% of the subjects in the DM group had HPL, whereas only 8% in the non-DM group. A case-control study suggested that lower levels of serum cholesterol and low-density lipoprotein cholesterol (LDL) were associated with a higher risk of endometrial cancer [
36]. However, no association of total cholesterol, LDL cholesterol, and triglyceride (TG) levels with the risk of endometrial cancer was found after the adjustment of BMI [
37]. A Norway study in 2009 found that only serum TG level, but not total cholesterol, LDL, or high-density lipoprotein cholesterol (HDL), was positively associated with the risk of endometrial cancer [
12]. However, we found that HPL was associated with uterine cancer specifically in the age group <50 years after adjustment. Similar to HPL, statin use was associated with a greater risk of uterine cancer in the <50 years age group. Since statins were prescribed in HPL, this phenomenon of statins was possibly driven by HPL itself. Interestingly, in those with an age ≥50 years, HPL and statin use were associated with a reduced risk of uterine cancer. The statin use in the post-menopause group seems protective against uterine cancer even though HPL was not present. This age-specific effect of statins was not previously mentioned [
38]. A meta-analysis also found that statin use was associated with a suggestive reduction in the risk of uterine cancer in Asian populations [
38]. Lavie et al. reported that statins are associated with decreased cancer risk and improved survival in endometrial cancer by blocking cancer cell growth pathways and induction of tumor cell apoptosis [
39]. Currently, the age-specific risk of HPL and uterine cancer remains unclear. Our findings suggest HPL and statin use might act as either a risk factor or a protective factor of uterine cancer according to specific ages.
In our study, HRT was positively associated with uterine cancer development after age, DM, and other comorbidities were adjusted. No significance was found in DM subjects with HRT and uterine cancer, which could be explained by the small sample size of the DM with HRT group. Different regimens of HRT (estrogen only or estrogen plus progestin) might also affect cancer risk. In subjects diagnosed with uterine cancer, HRT was associated with higher cancer risk in both premenopausal and postmenopausal groups. Endometrial cancer was strongly associated with the use of unopposed estrogen in postmenopausal women without hysterectomies [
40]. This risk, however, can be mitigated if progestins were added. However, it remains debatable whether continuous combined therapy provides a greater protection of the uterus than placebo or never-use. A review concluded that the use of estrogen alone, tibolone, and sequential combined therapy could increase the risk of endometrial cancer, even when treatment lasts less than five years, but continuous combined therapy could present a lower risk than never use even for more than 10 years [
41].
There were some strengths as well as limitations of our study. Regarding the strengths, the study cohort was collected from the NHI database with links to the TCR database, which is a representative nationwide database with a large sample size and high completeness. Second, multivariate Cox proportional hazard regression models with adjustment for potential confounding factors were performed. Third, age-stratification was carried out to analyze the risk in premenopausal and postmenopausal populations. There are some limitations in this study; first, we cannot differentiate type 1 or type 2 DM in our cohort, but we excluded DM subjects younger than 30 years old to ensure that most of the DM were type 2. Second, information on the duration or severity of DM and treatment regimens was lacking in our cohort. Studies have reported that a significantly increased incidence of uterine cancer was seen within the first month after type 2 DM was diagnosed [
42], or even in the pre-diabetes phase [
43,
44]. However, a recent study suggested that the highest risk of uterine cancer was 4 to 8 years after the diagnosis of type 2 DM and remains at a plateau afterward [
45]. Third, information of BMI, waist-to-hip ratio, or body weight to determine the degree of obesity, potential confounders including parity, age of menarche, or menopause were not available. Fourth, treatment duration of HRT and statin use or some given medications such as metformin cannot be extracted from the database.