Abstract
Background
Circulating adipokine levels may be associated with endometrial cancer risk, yet few studies have evaluated these markers prospectively.
Methods
We conducted a nested case-control study of postmenopausal women in the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (n=78,216), including 167 incident endometrial cancer cases and 327 controls that were matched on age, study center, race, study year of diagnosis, year of blood draw, time of day of blood draw and menopausal hormone therapy (MHT) use. Adipokine and estradiol levels were categorized into tertiles (T). Odds ratios (ORs) and 95% confidence intervals (CIs) for the associations of adiponectin, leptin and visfatin with endometrial cancer risk were estimated by conditional logistic regression, adjusting for known endometrial cancer risk factors, including body mass index (BMI) and circulating estradiol levels.
Results
Adiponectin levels were inversely associated with risk of endometrial cancer [OR T3vsT1=0.48 (95%CI: 0.29-0.80); p-trend<0.01], whereas elevated leptin levels showed a positive association [2.77 (1.60-4.79); p-trend<0.01]. These results remained significant after adjustment for estradiol, but not after further adjustment for BMI. When analyses were restricted to non-MHT users, associations of adiponectin and leptin were stronger and remained significant after adjustment for estradiol and BMI [0.27 (0.09-0.80); p-trend=0.01 and 4.29 (1.07-17.15); p-trend=0.02, respectively]. Non-significant positive associations were observed for visfatin.
Conclusion
Adipokines may influence endometrial cancer risk through pathways other than estrogen-mediated cell growth in postmenopausal women not currently on MHT.
Impact
Understanding how adipokines influence endometrial cancer risk may help to elucidate biological mechanisms important for the observed obesity-endometrial cancer association.
Keywords: endometrial cancer risk, adiponectin, leptin, visfatin, obesity
INTRODUCTION
Adipose tissue produces and secretes many metabolically active molecules, including adipokines such as adiponectin, leptin and visfatin (1). Whereas obesity is a well-known risk factor for endometrial cancer, the relationship between these obesity-related factors and endometrial cancer risk remains largely unclear. To date, hypotheses regarding the mechanism by which obesity increases risk in postmenopausal women have largely centered around the aromatization of androgens in adipose tissue leading to increased circulating estradiol levels (2). Other mechanisms via inflammation, insulin resistance and adipokines, however, are thought to be important.
Adiponectin, the most abundant adipokine, has been suggested to have anti-angiogenic, anti-inflammatory and anti-apoptotic properties (3-5). In addition, increased adiponectin levels in serum have been shown to reduce blood glucose and insulin levels and thus are inversely correlated with obesity and type-2 diabetes (6-7). Results from the limited number of epidemiologic studies evaluating circulating adiponectin levels and endometrial cancer are inconsistent. Case-control studies have reported an inverse association; however, these studies used post-diagnostic serum and thus could not address temporality (8-11). To date, only three prospective studies have been conducted; one study reported an inverse association with pre-diagnostic serum levels (12), whereas the other two reported no association (13-14).
Leptin has contrasting biological functions to adiponectin; it has been shown to promote cell proliferation, angiogenesis and metastasis in certain cell lines (1, 5, 15). Serum leptin levels are positively correlated with obesity and function to regulate appetite, weight, metabolism and fertility (16-17). Results from case-control studies (8, 18-19) that suggested a positive association between leptin levels and endometrial cancer have recently been corroborated in a prospective study (13). Few studies of adipokines and endometrial cancer risk have included measures of both adiponectin and leptin (8, 10, 13) and only two have reported associations for their ratio (8, 13). The leptin:adiponectin ratio has been suggested to be a surrogate marker of insulin resistance in both diabetic and non-diabetic women (20-21) and has been shown to be positively associated with breast and colorectal cancer (22-23).
Visfatin, also known as nicotinamide phosphoribosyl-transferase (Nampt) or pre-B cell colony-enhancing factor, is an adipokine that was recently discovered in visceral fat. It has been implicated in a variety of metabolic and stress functions as well as cellular energy metabolism (24-26). Recent epidemiologic studies have shown direct associations of serum visfatin levels with gastric carcinoma, colorectal adenocarcinoma and postmenopausal breast cancer (27-29). In addition, a few studies have suggested a link between visfatin and polycystic ovary syndrome, a risk factor for endometrial cancer (30-32); however, to date, the association of visfatin with endometrial cancer risk has not been evaluated.
To shed further light on relationships of adiponectin, leptin and visfatin with endometrial cancer risk, we measured pre-diagnostic serum levels of these adipokines using a case-control study nested within the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. We adjusted for body mass index (BMI), estradiol and other factors known to influence endometrial cancer risk to determine whether serum adipokine levels are independently associated with endometrial cancer.
MATERIALS AND METHODS
Study population
The design of the Prostate, Lung, Colorectal and Ovarian (PLCO) cancer screening trial has been described previously (33). In brief, between 1993 and 2001, 78,216 women aged 55-74 years were recruited at 10 screening centers across the U.S. and randomized to either an intervention (screening) or control (usual care) arm. Incident cancers were ascertained primarily by self-report on an annual study update form and confirmed by medical record abstraction. The current study included women who were randomized to the screening arm of the trial and met the following criteria at baseline: intact uterus; no previous cancer diagnosis; available serum sample; completion of a detailed questionnaire, which captured basic demographics, lifestyle factors and reproductive history and informed consent for studies of cancer. Cancer cases were defined as women having a first primary diagnosis of an invasive epithelial tumor of the endometrium between their initial screening visit and April 1, 2006. Controls had to be alive at the time of the diagnosis of the case and were matched 2:1 on age (5-year intervals), study center, race (white/black/other), individual study year of diagnosis, individual year of blood draw, time of day of blood collection (AM/PM) and menopausal hormone therapy (MHT) use at baseline (never/former/current/unknown). Subjects with missing information on BMI, diabetes and smoking information (n=2 for cases and n=4 controls), one case with insufficient serum volume for assays and one case whose serum sample was later determined to have been drawn after cancer diagnosis were excluded from analyses. The final study population consisted of 167 cases and 327 matched controls; there were 7 cases with 1 matched control after applying exclusion criteria.
Laboratory assays
Serum concentrations of adiponectin, leptin, visfatin and estradiol were measured in the Reproductive Endocrine Research Laboratory of Dr. Frank Stanczyk (University of Southern California, Los Angeles, CA). Leptin and adiponectin were measured by radioimmunoassay (RIA) using reagents obtained from Millipore Linco Research (St. Charles, MO). The leptin assay sensitivity is 0.5 ng/ml, and the interassay coefficients of variation (CVs) are 6.2%, 4.7% and 3.6% at 4.9 ng/ml, 10.4 ng/ml and 25.6 ng/ml, respectively. The adiponectin assay sensitivity is 1 ng/ml, and the interassay CVs are 9.2%, 6.9% and 9.2% at 1.5 ng/ml, 3.0 ng/ml and 7.5 ng/ml, respectively. Visfatin was measured by an enzyme-linked immunosorbent assay (ELISA; ALPCO, Salem, NH). The assay sensitivity is 0.12 ng/ml, and the interassay CVs are 4.7%, 5.7% and 7.2% at 0.72 ng/ml, 1.01 ng/ml and 2.92 ng/ml, respectively. Estradiol was measured by RIA following organic solvent extraction and Celite column partition chromatography. The assay sensitivity is 2 pg/ml and the interassay CVs are 11%, 13% and 12% at 15, 36 and 101 pg/ml, respectively.
Matched sets were not retained within the same assay batch; however, the proportion of cases and controls was generally consistent across batches. We did not observe any systematic drifts in analyte levels based on batch, and adjustment for batch in subsequent analyses did not significantly change the results. In addition to the laboratory’s quality control (QC) samples, four blinded QC samples were included within each batch. Intra-batch coefficients of variation (CVs) from these masked samples were 4.1, 4.2, 7.9 and 5.3; the inter-batch CVs were 8.3, 4.5, 6.6 and 12.9; and the intraclass correlation values were 0.677, 0.971, 0.938 and 0.915 for adiponectin, leptin, visfatin and estradiol, respectively. Levels of visfatin were not available for 21 QC samples due to insufficient serum volume; therefore, CV and ICC values for visfatin were based on 43 QC samples. Samples that fell below the level of detection were set to the minimum detection level; 4 samples (3 controls and one case), were set to 1 pg/ml for estradiol and 2 samples (one case and one control), were set to 0.1 ng/ml for visfatin.
Statistical analyses
Differences in baseline characteristics and analyte levels were determined by Chi-square or Kruskall-Wallis tests as appropriate. Spearman correlations were estimated for the associations between analytes and BMI among controls. Odds ratios (OR) and 95% confidence intervals (CI) for the relationships between analyte levels and endometrial cancer were estimated using conditional logistic regression. Results from unconditional logistic regression models adjusted for matching factors were similar to those from conditional logistic regression models. Tertiles of analyte levels were determined based on the distribution in the control population. Multivariable models were adjusted for known endometrial cancer risk factors which were captured on the baseline questionnaire: oral contraceptive use (yes/no); family history of breast or endometrial cancer (yes/no); current smoking status (never/former/current); education level (less than high school (HS)/HS or equivalent/post-HS, no college/some college/college grad/post-graduate); parity (0, 1, 2, 3+ live births) and history of diabetes (yes/no). Additional adjustment was made for (i) BMI (continuous), (ii) estradiol (continuous) and (iii) BMI and estradiol. Linear trends were assessed by modeling categorical variables as ordinal variables. We conducted stratified analyses based on MHT status due to the potential affect of MHT use on adipokine levels. Due to the small number of former MHT users in the population, we combined never and former users into a single category for these analyses; however similar results were observed when restricting to women who had never used MHT only. Statistical tests for interaction were performed by likelihood ratio tests comparing unconditional logistic regression models that adjusted for all matching factors (excluding MHT status), BMI and estradiol with and without an interaction term for current MHT use and each adipokine. All statistical analyses were performed using STATA version 11.2 (STATA Corp, College Station, TX).
RESULTS
The average age of endometrial cancer cases in this study was 66.4 ± 5.7 years and the average time from blood draw to diagnosis was 3.4 ± 2.8 years (Table 1). Cancer cases were more likely to be obese (p<0.001) and less likely to have used oral contraceptives (p=0.03) than matched controls. Other known endometrial cancer risk factors such as parity, diagnosis of diabetes and smoking status did not differ significantly by case status. Furthermore, as expected, the distribution of matching factors did not differ substantially for cases compared with controls. Median levels of each adipokine differed by case status (Table 2). Levels of adiponectin were significantly lower among cases (p=0.002), whereas median levels of leptin and visfatin were higher among cases compared to controls (p<0.001 and p=0.05, respectively).
Table 1.
Baseline characteristics of study population
| Characteristic | Incident Endometrial Cases (n=167) |
Matched Controls (n=327) |
p-valuea |
|---|---|---|---|
|
| |||
| n (%) | |||
| Age (at randomization)b | |||
| <60 | 43 (25.8) | 86 (26.3) | 1.00 |
| 60-64 | 59 (35.3) | 115 (35.2) | |
| 65-69 | 48 (28.7) | 92 (28.1) | |
| 70+ | 17 (10.2) | 34 (10.4) | |
| Race b | |||
| White, non-hispanic | 157 (94.0) | 309 (94.5) | 0.94 |
| Black, non-hispanic | 5 (3.0) | 8 (2.5) | |
| Asian | 5 (3.0) | 10 (3.1) | |
| Time of blood draw b | |||
| AM | 103 (61.7) | 203 (62.1) | 0.93 |
| PM | 64 (38.3) | 124 (37.9) | |
| Menopausal hormone use (MHT) b | |||
| Never | 62 (37.1) | 125 (38.2) | 0.68 |
| Current | 79 (47.3) | 153 (46.8) | |
| Former | 24 (14.4) | 48 (14.7) | |
| Unknown | 2 (1.2) | 1 (0.3) | |
| BMI (kg/m2) | |||
| <25 | 46 (27.5) | 142 (43.4) | <0.001 |
| 25-30 | 56 (33.5) | 104 (31.8) | |
| 30-35 | 29 (17.4) | 59 (18.0) | |
| 35+ | 36 (21.6) | 22 (6.7) | |
| Education level | |||
| Less than high school | 7 (4.2) | 14 (4.3) | 0.38 |
| High school or equivalent | 34 (20.4) | 89 (27.2) | |
| Post high school, not college | 18 (10.8) | 41 (12.5) | |
| Some college | 50 (29.9) | 72 (22.0) | |
| College graduate | 33 (19.8) | 60 (18.4) | |
| Postgraduate | 25 (15.0) | 51 (15.6) | |
| Number of live births | |||
| 0 | 23 (13.8) | 34 (10.4) | 0.38 |
| 1-2 | 48 (28.7) | 86 (26.3) | |
| 3+ | 96 (57.5) | 207 (63.3) | |
| Smoking status | |||
| Never | 103 (61.7) | 178 (54.4) | 0.13 |
| Current | 8 (4.8) | 30 (9.2) | |
| Former | 56 (33.5) | 119 (36.4) | |
| Oral contraceptive use | 70 (41.9) | 171 (52.3) | 0.03 |
| Family history of breast cancer | 22 (13.3) | 47 (14.4) | 0.85 |
| Family history of endometrial cancer | 3 (1.8) | 3 (0.9) | 0.39 |
| History of hypertension diagnosis | 66 (39.5) | 107 (32.7) | 0.13 |
| History of diabetes diagnosis | 13 (7.8) | 15 (4.6) | 0.15 |
| Age at diagnosis (y), mean ± SD (10th, 90th) | 66.4 ± 5.7 (59, 74) | ||
| Time from blood draw to diagnosis (y), mean ± SD (10th, 90th) |
3.4 ± 2.8 (0.2, 7.3) | ||
| Estradiol (pg/ml), median (10th, 90th) | 20.3 (9.73, 47.2) | 15.5 (6.6, 40.2) | <0.001c |
p-values based on Chi-square statistic.
Matching factors. Other matching factors not included in the table were study center, study year of diagnosis, and year of blood draw; the distributions of these factors were not different between cases and controls.
p-value based on Wilcoxon rank sum test for differences in medians.
Table 2.
Median and interdecile range of serum adipokine levels by case status
| Analyte | Incident Endometrial Cases (n=167) |
Matched Controls (n=327) |
p-valuea |
|---|---|---|---|
|
| |||
| Median (10th, 90th) | |||
| Adiponectin (μg/ml) | 12.16 (6.79, 22.92) | 14.77 (7.49, 25.93) | 0.002 |
| Leptin (ng/ml) | 19.82 (7.55, 57.44) | 16.69 (5.98, 41.08) | <0.001 |
| Visfatin (ng/ml)b | 0.84 (0.41, 2.64) | 0.76 (0.33, 1.77) | 0.046 |
p-value calculated by Wilcoxon rank sum test for differences in medians.
Values based on 94 cases and 189 controls.
Levels of each adipokine were significantly correlated with BMI; a negative correlation was seen for adiponectin, whereas the association was positive for leptin and visfatin (unadjusted Spearman correlation coefficients for adiponectin, leptin and visfatin among controls: r=−0.40, 0.80, and 0.32, respectively; p<0.0001 for all correlations). Similarly, among controls who were not currently using MHT each adipokine was significantly correlated with estradiol (r=−0.25, 0.51, and 0.36, respectively; p<0.001 for all correlations). In addition, significant correlations were observed between the adipokines, except for between visfatin and adiponectin. Leptin was negatively correlated with adiponectin but positively with visfatin (r=−0.32, and 0.31, respectively; p<0.0001). Correlations between the adipokines and with estradiol and BMI were stronger among overweight/obese women compared with normal weight women (Supplementary Table 1).
Higher BMI (≥30 kg/m2 versus <25 kg/m2) and estradiol levels were associated with an increased risk of endometrial cancer after mutual adjustment in multivariable models (OR=2.69, [95% CI: 1.56-4.64]; p-trend<0.001, and ORT3vsT1=2.94 [1.41-6.12]; p-trend=0.02, respectively). Odds ratios for the association of adiponectin, leptin and visfatin with endometrial cancer, adjusted for (i) established endometrial cancer risk factors other than BMI and then additionally for (ii) BMI, (iii) estradiol, or (iv) BMI and estradiol, are presented in Table 3. In model (i), higher adiponectin levels were inversely associated with endometrial cancer risk (ORT3vsT1=0.48 [0.29-0.80]; p-trend=0.004), whereas elevated levels of leptin were associated with increased risk (ORT3vsT1=2.77 [1.60-4.79]; p-trend<0.001). In addition, a significant reduction in risk was observed with an increased ratio of adiponectin:leptin (ORT3vsT1=0.35 [0.20-0.61]; p-trend<0.001). Each of these associations remained significant after adjusting further for estradiol (p-trend≤0.01). However, after adjustment for BMI, the estimates and linear trends were no longer significant. Models that adjusted for BMI and estradiol gave similar results to those with BMI alone.
Table 3.
Multivariable adjusted associations of adipokine levels and endometrial cancer risk in the full study population
| Case/Control |
Multivariable (MV) Adjusted |
MV & Estradiol | MV & BMI | MV & BMI & Estradiol |
|
|---|---|---|---|---|---|
|
|
|||||
| N | OR (95% CI) | ||||
| Adiponectin (μg/ml) | |||||
| ≤ 11.89 | 79/107 | Ref (1.00) | Ref (1.00) | Ref (1.00) | Ref (1.00) |
| 11.90 – 18.10 | 47/109 | 0.54 (0.34-0.87) | 0.57 (0.35-0.92) | 0.64 (0.39-1.05) | 0.65 (0.40-1.08) |
| > 18.10 | 41/111 | 0.48 (0.29-0.80) | 0.53 (0.32-0.88) | 0.71 (0.41-1.23) | 0.73 (0.43-1.27) |
| p-trend | 0.004 | 0.012 | 0.185 | 0.231 | |
| Leptin (ng/ml) | |||||
| ≤ 11.03 | 34/109 | Ref (1.00) | Ref (1.00) | Ref (1.00) | Ref (1.00) |
| 11.04 – 22.55 | 52/109 | 1.61 (0.93-2.78) | 1.59 (0.91-2.77) | 1.20 (0.67-2.12) | 1.21 (0.67-2.16) |
| > 22.55 | 81/109 | 2.77 (1.60-4.79) | 2.55 (1.46-4.46) | 1.31 (0.65-2.64) | 1.29 (0.64-2.61) |
| p-trend | <0.001 | 0.001 | 0.454 | 0.483 | |
| Visfatin (ng/ml) | |||||
| ≤ 0.54 | 22/62 | Ref (1.00) | Ref (1.00) | Ref (1.00) | Ref (1.00) |
| 0.55 – 0.93 | 31/63 | 1.29 (0.55-3.01) | 1.27 (0.54-2.97) | 0.92 (0.37-2.30) | 0.93 (0.37-2.30) |
| > 0.93 | 41/64 | 1.67 (0.71-3.95) | 1.68 (0.70-3.99) | 1.25 (0.50-3.09) | 1.24 (0.50-3.08) |
| p-trend | 0.197 | 0.239 | 0.582 | 0.598 | |
| Adiponectin/Leptin (μg) | |||||
| ≤ 0.55 | 82/108 | Ref (1.00) | Ref (1.00) | Ref (1.00) | Ref (1.00) |
| 0.56 – 1.43 | 50/109 | 0.53 (0.32-0.89) | 0.57 (0.34-0.96) | 0.87 (0.48-1.56) | 0.89 (0.50-1.61) |
| > 1.43 | 35/110 | 0.35 (0.20-0.61) | 0.39 (0.22-0.69) | 0.74 (0.36-1.51) | 0.78 (0.38-1.60) |
| p-trend | <0.001 | 0.001 | 0.409 | 0.504 | |
Multivariable models: adjusted for family history of breast or endometrial cancer, education level, parity, history of diabetes diagnosis, oral contraceptive use, and current smoking status. Odds ratios (OR) and 95% CIs were estimated by conditional logistic regression (matched sets: 167 cases and 327 controls). Tertiles of adipokines based on the control distribution. Estradiol and BMI were included as continuous terms in the model.
Women currently using MHT were more likely to have a BMI <25 than those not on MHT (50% vs. 30%, respectively, p<0.0001), suggesting MHT use may modify the association between adipokines and endometrial cancer risk. Therefore, associations for each adipokine, stratified by MHT use, are presented in Table 4. For these analyses, we combined never and former MHT users; however, similar results were observed when the population was restricted to never users only. Among women who were not using MHT at baseline, we observed stronger magnitudes of association than those for the overall population for adiponectin (ORT3vsT1=0.25 [0.08-0.75], p-trend=0.01), leptin (OR T3vsT1=4.72 [1.15-19.38], p-trend=0.02) and the ratio of adiponectin:leptin (OR T3vsT1=0.15 [0.03-0.61], p-trend=0.01). Furthermore, the associations and linear trends remained statistically significant after adjustment for BMI and estradiol. No significant associations for adipokines were seen among women who were using MHT at baseline (Table 4).
Table 4.
Multivariable adjusted associations of adipokine levels and endometrial cancer risk stratified by MHT use
| Non-Current Usersa |
Current Users |
||||
|---|---|---|---|---|---|
| Case/Control | MV & BMI & Estradiol |
Case/Control | MV & BMI & Estradiol |
p-valueb | |
| Adiponectin (μg/ml) | |||||
| ≤ 11.89 | 53/57 | Ref (1.00) | 26/50 | Ref (1.00) | 0.005 |
| 11.90 – 18.10 | 23/57 | 0.56 (0.25-1.23) | 22/51 | 0.67 (0.32-1.41) | |
| > 18.10 | 10/59 | 0.25 (0.08-0.75) | 31/52 | 1.19 (0.56-2.53) | |
| p-trend | 0.011 | 0.645 | |||
| Leptin (ng/ml) | |||||
| ≤ 11.03 | 6/46 | Ref (1.00) | 28/63 | Ref (1.00) | 0.003 |
| 11.04 – 22.55 | 21/61 | 2.00 (0.57-7.05) | 30/47 | 1.18 (0.56-2.45) | |
| > 22.55 | 59/66 | 4.72 (1.15-19.38) | 21/43 | 0.70 (0.24-2.03) | |
| p-trend | 0.019 | 0.640 | |||
| Visfatin (ng/ml) | |||||
| ≤ 0.54 | 10/29 | Ref (1.00) | 12/33 | Ref (1.00) | 0.842 |
| 0.55 – 0.93 | 14/30 | 0.58 (0.08-4.06) | 17/33 | 0.94 (0.16-5.37) | |
| > 0.93 | 23/35 | 1.71 (0.15-20.12) | 18/29 | 2.84 (0.53-15.11) | |
| p-trend | 0.585 | 0.163 | |||
| Adiponectin/Leptin (μg) | |||||
| ≤ 0.55 | 62/65 | Ref (1.00) | 19/43 | Ref (1.00) | 0.0002 |
| 0.56 – 1.43 | 17/56 | 0.30 (0.10-0.84) | 32/52 | 2.16 (0.86-5.43) | |
| > 1.43 | 7/52 | 0.15 (0.03-0.61) | 28/58 | 1.92 (0.62-5.91) | |
| p-trend | 0.006 | 0.350 | |||
MV=multivariable: adjusted for family history of breast or endometrial cancer, education level, parity, history of diabetes diagnosis, oral contraceptive use, and current smoking status. Odds ratios (OR) and 95% CIs were estimated by conditional logistic regression (matched sets: 86 cases and 173 controls for never/former users and 79 cases and 153 controls for current users). Tertiles of adipokines based on the control distribution. Estradiol and BMI were included as continuous terms in the model.
Includes never and former MHT users.
p-value calculated from a likelihood ratio test comparing unconditional multivariable logistic regression models with and without an interaction term for the adipokine and current MHT use adjusting for matching factors and the variables included in the MV conditional logistic regression models.
DISCUSSION
In this case-control study nested within the PLCO Trial, we prospectively evaluated the relationship of pre-diagnostic serum levels of adiponectin, leptin and visfatin with postmenopausal endometrial cancer risk, while also considering the role of circulating estradiol and BMI in these associations. In the overall population, elevated adiponectin and the ratio of adiponectin-to-leptin were associated with a significant decrease in endometrial cancer risk, whereas higher levels of leptin conferred an increased risk. These relationships remained significant after adjustment for estradiol, but not BMI. In addition, we observed evidence that MHT use modified the association of adiponectin and leptin with endometrial cancer. Stronger magnitudes of association were seen in women not currently using MHT, compared with the overall population. These estimates remained significant after adjustment for estradiol levels and BMI. However, no significant associations were observed in current MHT users.
Previous studies evaluating adiponectin and leptin levels and endometrial cancer risk have had various inclusion criteria regarding menopausal status and MHT use; one study included only postmenopausal women (13), whereas the others stratified by menopausal status (12, 14) and women were either matched on MHT use (14) or current users were excluded (12-13). Overall, these studies suggest an inverse association of adiponectin levels with endometrial cancer risk, which is consistent with the results from the current study, although some differences are noted. Specifically, in our stratified analysis, we observed significant associations of adiponectin and leptin with endometrial cancer after adjusting for estradiol and BMI. The two previous studies that excluded MHT users, the European Prospective Investigation into Cancer and Nutrition (EPIC) and the Breast and Bone Follow-up of the Fracture Intervention Trial (B~FIT) reported a non-significant inverse association or no association among postmenopausal women who were not MHT users, respectively, after adjusting for BMI and/or estradiol (12-13). Potential reasons for the observed discrepancies include differences in study population demographics, sample size and assay limitations. Although the sample size for the current study was similar to that of EPIC, the average level of adiponectin was much higher in our population (12 vs 8 μg/ml, for cases in PLCO and EPIC, respectively). Similar results are seen when comparing tertiles with similar ranges of adiponectin levels between the two studies (second tertile of PLCO to third of EPIC). Reasons for differences in adiponectin levels between the study populations could be related to the assays used to measure adipokines (RIA vs. ELISA) or differences in population demographics such as BMI levels or other characteristics. Similarly, differences in observed associations reported from B~FIT could be related to sample size, only 62 cases were included in B~FIT (compared with 86 in the stratified models for the current study) or population characteristics such as the older age of diagnosis for B~FIT cases compared with the current study (74 vs. 66 years). The only other prospective study of adiponectin, the Nurses’ Health Study (NHS), did not report associations for postmenopausal women who were not MHT users; however, the results from the overall population in the current study are similar to those reported for postmenopausal women in NHS.
Our observed association of leptin with endometrial cancer risk is generally consistent with the only other prospective study to evaluate this adipokine (13), which reported a non-significant positive association, after adjusting for BMI among postmenopausal women not on MHT. However, we observed a stronger elevated risk among women not on MHT, with over a 4-fold increased risk among women with leptin levels in the third tertile compared with the first. The median levels of leptin reported in B~FIT were higher than those in the current population (25 vs. 17 ng/ml, respectively) which could be related to the different assays used or population characteristics. When estimates from comparable leptin levels are compared between the two studies, similar results are observed. In addition, associations for the ratio of adiponectin:leptin with endometrial cancer reported in the current study are similar with those reported in B~FIT (13), however continued evaluation is needed to determine if this ratio is indeed a more sensitive measure of insulin resistance, as has been suggested (20-21), and its relationship with endometrial cancer risk.
From our stratified analyses, we observed evidence that MHT use significantly modified the association of adiponectin and leptin with endometrial cancer risk, with significant associations observed only in non-MHT users. These results are similar to what has been reported for the relationship of BMI and MHT with endometrial cancer. The affect of BMI on endometrial cancer risk are restricted to women who have not used MHT (34) and conversely, MHT use has been shown to increase the risk of endometrial cancer among women with a normal body weight, whereas no additional risk was observed among obese women (35-36). This suggests that MHT and BMI can influence endometrial cancer risk through similar pathways, namely estrogen-driven proliferation. Furthermore, these data indicate that including MHT users when evaluating hormonally-related factors may mask potential associations.
Although the leading hypothesis for the obesity-related increase in endometrial cancer risk in postmenopausal women centers around elevated circulating estradiol levels through the aromatization of androgens in adipose tissue, other non-hormonal pathways may also be important (2). After accounting for estradiol levels and obesity (BMI), we observed a significant association of adiponectin and leptin with endometrial cancer risk among non-MHT users, suggesting adipokines may influence risk through mechanisms other than estrogen-mediated proliferation. Results from these stratified analyses should be interpreted with caution due to the limited range of BMI among the two strata of MHT use and the small number of non-MHT users in the lowest (leptin) and highest (adiponectin) tertiles, which is not surprising given the interrelatedness of BMI, MHT and adipokines. These small numbers could lead to unstable estimates, which may not reflect the true association. However, the magnitudes of association for adiponectin and leptin were greater than that of BMI alone among non-MHT users (OR=4.00, OR=4.72 and OR=3.07, respectively), further supporting potential roles for adipokines in non-hormonal pathways associated with cancer risk. Cell culture studies have implicated adiponectin in anti-inflammatory and insulin resistance pathways, whereas leptin has been shown to act in carcinogenic pathways such as cell proliferation and metastasis (1-2). Future research into these functions could elucidate additional mechanisms of action for adipokines in endometrial carcinogenesis. It is also possible that adiponectin and leptin levels may serve as better markers of the biological mechanisms that underlie the obesity-related associations for endometrial cancer risk, especially among non-MHT users. Future studies are needed to distinguish between these possibilities.
Few epidemiological studies have evaluated the relationship of visfatin with cancer risk and none have examined endometrial cancer. Our analyses with visfatin were limited by inadequate serum volume to measure levels in all samples, which could have affected our power to detect associations. Nevertheless, median levels of visfatin were significantly higher among cases compared with controls. In addition, point estimates from regression analyses suggest increased cancer risk among the highest tertile compared to the lowest, which is in agreement with previous epidemiologic studies of visfatin and other obesity-related cancers (27-29). Continued evaluation is needed to fully understand the potential role of visfatin in endometrial carcinogenesis and its relationship to risk.
The main strengths of this study include the prospective nature of the PLCO trial, the measurement of multiple adipokines and adjustment for key covariates for endometrial cancer such as BMI and estradiol. In addition, cases and controls were matched on time of day of blood draw, age, race and several other important variables, limiting the possibility that these factors influenced the observed associations. Finally, this is the first study to report the association of visfatin with endometrial cancer risk.
Despite these strengths, we lacked information on other anthropometric measures of obesity, such as waist-to-hip ratio, which may be a more accurate measure of central adiposity than BMI. Other studies have not seen further attenuation of associations for adipokines and endometrial cancer when including both BMI and waist circumference in the model (12). We also did not have information on fasting status at the time of blood draw, which could influence adipokine levels. However, previous studies have shown that adiponectin levels are not significantly affected by fasting status (12, 37) and we matched on time of day of blood draw to minimize these effects. Additionally, approximately 25% of the serum samples in this study were collected within one year of cancer diagnosis, making it difficult to establish temporality. However, sensitivity analyses excluding these samples resulted in similar patterns. Finally, this study included a large portion of women who were current MHT users at baseline (46%), leading to small numbers of women in the stratified models and potentially unstable estimates of association.
In summary, this study supports the notion that adiponectin, leptin and the ratio of the two may be important predictors of endometrial cancer risk. In contrast, in this first-time evaluation, visfatin did not appear related. Our study also expanded on previous knowledge by specifically assessing the potential interactions between MHT use and these analytes, showing that adiponectin and leptin were only related to risk among non-users of MHT. Together, these studies provide evidence for potential biological mechanisms that underlie the obesity-related endometrial cancer risk and suggest that in addition to estradiol, adiponectin and leptin play an important role in endometrial carcinogenesis. Given the inter-relatedness of adipokines with BMI and MHT, larger sample sizes from pooling efforts or consortia may be necessary to fully evaluate the roles of these adipokines in endometrial cancer development.
Supplementary Material
Acknowledgements
This research was supported by contracts from the Division of Cancer Prevention, National Cancer Institute, NIH, DHHS. The authors thank Drs. Christine Berg and Philip Prorok, Division of Cancer Prevention, National Cancer Institute, the Screening Center investigators and staff of the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial, Mr. Tom Riley and staff, Information Management Services, Inc., Ms. Barbara O’Brien and staff, Westat, Inc. Most importantly, we acknowledge the study participants for their contributions to making this study possible. We would also like to thank Dr. Britton Trabert for her participation in scientific discussions.
Funding: This work was supported by the Intramural Research Program of the National Cancer Institute
Footnotes
Conflict of interest: The authors have no conflicts of interest to disclose.
References
- 1.van Kruijsdijk RC, van der Wall E, Visseren FL. Obesity and cancer: the role of dysfunctional adipose tissue. Cancer Epidemiol Biomarkers Prev. 2009;18:2569–78. doi: 10.1158/1055-9965.EPI-09-0372. [DOI] [PubMed] [Google Scholar]
- 2.Calle EE, Kaaks R. Overweight, obesity and cancer: epidemiological evidence and proposed mechanisms. Nat Rev Cancer. 2004;4:579–91. doi: 10.1038/nrc1408. [DOI] [PubMed] [Google Scholar]
- 3.Fantuzzi G. Adipose tissue, adipokines, and inflammation. J Allergy Clin Immunol. 2005;115:911–9. doi: 10.1016/j.jaci.2005.02.023. quiz 20. [DOI] [PubMed] [Google Scholar]
- 4.Lihn AS, Pedersen SB, Richelsen B. Adiponectin: action, regulation and association to insulin sensitivity. Obes Rev. 2005;6:13–21. doi: 10.1111/j.1467-789X.2005.00159.x. [DOI] [PubMed] [Google Scholar]
- 5.Roberts DL, Dive C, Renehan AG. Biological mechanisms linking obesity and cancer risk: new perspectives. Annu Rev Med. 2010;61:301–16. doi: 10.1146/annurev.med.080708.082713. [DOI] [PubMed] [Google Scholar]
- 6.Arita Y, Kihara S, Ouchi N, Takahashi M, Maeda K, Miyagawa J, et al. Paradoxical decrease of an adipose-specific protein, adiponectin, in obesity. Biochem Biophys Res Commun. 1999;257:79–83. doi: 10.1006/bbrc.1999.0255. [DOI] [PubMed] [Google Scholar]
- 7.Weyer C, Funahashi T, Tanaka S, Hotta K, Matsuzawa Y, Pratley RE, et al. Hypoadiponectinemia in obesity and type 2 diabetes: close association with insulin resistance and hyperinsulinemia. J Clin Endocrinol Metab. 2001;86:1930–5. doi: 10.1210/jcem.86.5.7463. [DOI] [PubMed] [Google Scholar]
- 8.Ashizawa N, Yahata T, Quan J, Adachi S, Yoshihara K, Tanaka K. Serum leptin-adiponectin ratio and endometrial cancer risk in postmenopausal female subjects. Gynecol Oncol. 2010;119:65–9. doi: 10.1016/j.ygyno.2010.07.007. [DOI] [PubMed] [Google Scholar]
- 9.Dal Maso L, Augustin LS, Karalis A, Talamini R, Franceschi S, Trichopoulos D, et al. Circulating adiponectin and endometrial cancer risk. J Clin Endocrinol Metab. 2004;89:1160–3. doi: 10.1210/jc.2003-031716. [DOI] [PubMed] [Google Scholar]
- 10.Petridou E, Mantzoros C, Dessypris N, Koukoulomatis P, Addy C, Voulgaris Z, et al. Plasma adiponectin concentrations in relation to endometrial cancer: a case-control study in Greece. J Clin Endocrinol Metab. 2003;88:993–7. doi: 10.1210/jc.2002-021209. [DOI] [PubMed] [Google Scholar]
- 11.Soliman PT, Wu D, Tortolero-Luna G, Schmeler KM, Slomovitz BM, Bray MS, et al. Association between adiponectin, insulin resistance, and endometrial cancer. Cancer. 2006;106:2376–81. doi: 10.1002/cncr.21866. [DOI] [PubMed] [Google Scholar]
- 12.Cust AE, Kaaks R, Friedenreich C, Bonnet F, Laville M, Lukanova A, et al. Plasma adiponectin levels and endometrial cancer risk in pre- and postmenopausal women. J Clin Endocrinol Metab. 2007;92:255–63. doi: 10.1210/jc.2006-1371. [DOI] [PubMed] [Google Scholar]
- 13.Dallal CM, Brinton LA, Bauer DC, Buist DS, Cauley JA, Hue TF, et al. Obesity-related hormones and endometrial cancer among postmenopausal women: a nested case-control study within the B~FIT cohort. Endocr Relat Cancer. 2012 doi: 10.1530/ERC-12-0229. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Soliman PT, Cui X, Zhang Q, Hankinson SE, Lu KH. Circulating adiponectin levels and risk of endometrial cancer: the prospective Nurses’ Health Study. Am J Obstet Gynecol. 2011;204:167, e1–5. doi: 10.1016/j.ajog.2010.08.045. PMCID: 3032826. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Garofalo C, Surmacz E. Leptin and cancer. J Cell Physiol. 2006;207:12–22. doi: 10.1002/jcp.20472. [DOI] [PubMed] [Google Scholar]
- 16.Paracchini V, Pedotti P, Taioli E. Genetics of leptin and obesity: a HuGE review. Am J Epidemiol. 2005;162:101–14. doi: 10.1093/aje/kwi174. [DOI] [PubMed] [Google Scholar]
- 17.Friedman JM, Halaas JL. Leptin and the regulation of body weight in mammals. Nature. 1998;395:763–70. doi: 10.1038/27376. [DOI] [PubMed] [Google Scholar]
- 18.Cymbaluk A, Chudecka-Glaz A, Rzepka-Gorska I. Leptin levels in serum depending on Body Mass Index in patients with endometrial hyperplasia and cancer. Eur J Obstet Gynecol Reprod Biol. 2008;136:74–7. doi: 10.1016/j.ejogrb.2006.08.012. [DOI] [PubMed] [Google Scholar]
- 19.Petridou E, Belechri M, Dessypris N, Koukoulomatis P, Diakomanolis E, Spanos E, et al. Leptin and body mass index in relation to endometrial cancer risk. Ann Nutr Metab. 2002;46:147–51. doi: 10.1159/000063081. [DOI] [PubMed] [Google Scholar]
- 20.Finucane FM, Luan J, Wareham NJ, Sharp SJ, O’Rahilly S, Balkau B, et al. Correlation of the leptin:adiponectin ratio with measures of insulin resistance in non-diabetic individuals. Diabetologia. 2009;52:2345–9. doi: 10.1007/s00125-009-1508-3. PMCID: 2759015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Oda N, Imamura S, Fujita T, Uchida Y, Inagaki K, Kakizawa H, et al. The ratio of leptin to adiponectin can be used as an index of insulin resistance. Metabolism. 2008;57:268–73. doi: 10.1016/j.metabol.2007.09.011. [DOI] [PubMed] [Google Scholar]
- 22.Chen DC, Chung YF, Yeh YT, Chaung HC, Kuo FC, Fu OY, et al. Serum adiponectin and leptin levels in Taiwanese breast cancer patients. Cancer Lett. 2006;237:109–14. doi: 10.1016/j.canlet.2005.05.047. [DOI] [PubMed] [Google Scholar]
- 23.Guadagni F, Roselli M, Martini F, Spila A, Riondino S, D’Alessandro R, et al. Prognostic significance of serum adipokine levels in colorectal cancer patients. Anticancer Res. 2009;29:3321–7. [PubMed] [Google Scholar]
- 24.Galli M, Van Gool F, Rongvaux A, Andris F, Leo O. The nicotinamide phosphoribosyltransferase: a molecular link between metabolism, inflammation, and cancer. Cancer Res. 2010;70:8–11. doi: 10.1158/0008-5472.CAN-09-2465. [DOI] [PubMed] [Google Scholar]
- 25.Garten A, Petzold S, Korner A, Imai S, Kiess W. Nampt: linking NAD biology, metabolism and cancer. Trends Endocrinol Metab. 2009;20:130–8. doi: 10.1016/j.tem.2008.10.004. PMCID: 2738422. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Moschen AR, Gerner RR, Tilg H. Pre-B cell colony enhancing factor/NAMPT/visfatin in inflammation and obesity-related disorders. Curr Pharm Des. 2010;16:1913–20. doi: 10.2174/138161210791208947. [DOI] [PubMed] [Google Scholar]
- 27.Dalamaga M, Karmaniolas K, Papadavid E, Pelekanos N, Sotiropoulos G, Lekka A. Elevated serum visfatin/nicotinamide phosphoribosyl-transferase levels are associated with risk of postmenopausal breast cancer independently from adiponectin, leptin, and anthropometric and metabolic parameters. Menopause. 2011;18:1198–204. doi: 10.1097/gme.0b013e31821e21f5. [DOI] [PubMed] [Google Scholar]
- 28.Nakajima TE, Yamada Y, Hamano T, Furuta K, Gotoda T, Katai H, et al. Adipocytokine levels in gastric cancer patients: resistin and visfatin as biomarkers of gastric cancer. J Gastroenterol. 2009;44:685–90. doi: 10.1007/s00535-009-0063-5. [DOI] [PubMed] [Google Scholar]
- 29.Nakajima TE, Yamada Y, Hamano T, Furuta K, Matsuda T, Fujita S, et al. Adipocytokines as new promising markers of colorectal tumors: adiponectin for colorectal adenoma, and resistin and visfatin for colorectal cancer. Cancer Sci. 2010;101:1286–91. doi: 10.1111/j.1349-7006.2010.01518.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Jongwutiwes T, Lertvikool S, Leelaphiwat S, Rattanasiri S, Jultanmas R, Weerakiet S. Serum visfatin in Asian women with polycystic ovary syndrome. Gynecol Endocrinol. 2009;25:536–42. doi: 10.1080/09513590903015478. [DOI] [PubMed] [Google Scholar]
- 31.Ozkaya M, Cakal E, Ustun Y, Engin-Ustun Y. Effect of metformin on serum visfatin levels in patients with polycystic ovary syndrome. Fertil Steril. 2010;93:880–4. doi: 10.1016/j.fertnstert.2008.10.058. [DOI] [PubMed] [Google Scholar]
- 32.Plati E, Kouskouni E, Malamitsi-Puchner A, Boutsikou M, Kaparos G, Baka S. Visfatin and leptin levels in women with polycystic ovaries undergoing ovarian stimulation. Fertil Steril. 2010;94:1451–6. doi: 10.1016/j.fertnstert.2009.04.055. [DOI] [PubMed] [Google Scholar]
- 33.Prorok PC, Andriole GL, Bresalier RS, Buys SS, Chia D, Crawford ED, et al. Design of the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. Control Clin Trials. 2000;21:273S–309S. doi: 10.1016/s0197-2456(00)00098-2. [DOI] [PubMed] [Google Scholar]
- 34.Crosbie EJ, Zwahlen M, Kitchener HC, Egger M, Renehan AG. Body mass index, hormone replacement therapy, and endometrial cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev. 2010;19:3119–30. doi: 10.1158/1055-9965.EPI-10-0832. [DOI] [PubMed] [Google Scholar]
- 35.Beral V, Bull D, Reeves G. Endometrial cancer and hormone-replacement therapy in the Million Women Study. Lancet. 2005;365:1543–51. doi: 10.1016/S0140-6736(05)66455-0. [DOI] [PubMed] [Google Scholar]
- 36.Trabert B, Wentzensen N, Yang HP, Sherman ME, Hollenbeck AR, Park Y, et al. Is estrogen plus progestin menopausal hormone therapy safe with respect to endometrial cancer risk? Int J Cancer. 2013;132:417–26. doi: 10.1002/ijc.27623. PMCID: 3427719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Gavrila A, Chan JL, Yiannakouris N, Kontogianni M, Miller LC, Orlova C, et al. Serum adiponectin levels are inversely associated with overall and central fat distribution but are not directly regulated by acute fasting or leptin administration in humans: cross-sectional and interventional studies. J Clin Endocrinol Metab. 2003;88:4823–31. doi: 10.1210/jc.2003-030214. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.