Abstract
Background:
No epidemiological studies have directly assessed the association between dietary and urinary of isoflavonoids and risk of liver cancer in humans.
Methods:
A nested case-control study including 217 incident cases of liver cancer and 427 individually matched control subjects was conducted in Shanghai, China. Dietary isoflavonoids intakes were assessed through a validated food-frequency questionnaire and the Chinese Food Composition Tables. Urinary excretion levels of four major isoflavonoids were measured by the reversed-phase high-performance liquid chromatography. ORs and 95% confidence intervals (CIs) were derived using conditional logistic regression models.
Results:
The adjusted ORs (95% CIs) for liver cancer across increasing quartiles of urinary genistein levels were 1.00 (reference), 0.55 (95% CI, 0.22–1.36), 0.57 (95% CI, 0.23–1.43) and 0.19 (95% CI, 0.06–0.59) (Ptrend =0.008) in women and 1.00 (reference), 1.22(0.52–2.86), 1.17(0.47–2.90), 1.23 (0.55–2.76) in men, respectively. These associations were consistent by limiting the cases to primary malignant neoplasm of liver or malignant neoplasms of the intrahepatic bile ducts, or among participants without self-reported liver disease or cirrhosis at the baseline survey. No associations were found between dietary isoflavonoids and liver cancer risk.
Conclusions:
Our study suggests for the first time that urinary excretion of genistein may be associated with reduced risk of liver cancer in women.
Impact:
In this nested case-control study in China, we found that urinary excretion of genistein was associated with lower risk of liver cancer in women, and not in men.
Keywords: liver cancer, isoflavonoids, genistein, nested case-control study
Introduction
Isoflavonoids are a group of biologically-active phytochemicals, which humans are exposed to mainly through the intake of soy food [1]. After consumption, soy isoflavonoids undergo various metabolic conversions by intestinal bacteria. Then both the metabolites and parent compounds are absorbed into the blood and finally excreted, primarily in the urine [1]. Previous studies have observed that there is substantial interindividual heterogeneity in the absorption and metabolism of soy isoflavonoids [2]. The level of urinary isoflavonoids provided an aggregate measurement of the intake, absorption, and metabolism [3]. It might be more informative than estimates from food frequency questionnaire (FFQ) for the assessment of health effects. It should be the challenge in large epidemiologic studies to accurately evaluate intake of isoflavone over a long period of time [3]. FFQ as well as urinary level are regarded to be markers for soy exposure, although they are subject to measurement error.
Although evidence from in vitro and animal studies has strongly indicated that soy foods or isoflavonoids may protect against liver cancer [4–12], only a few epidemiologic studies have been published, and the results were inconsistent [13–15]. In a hospital-based case–control study conducted in Greece, no association was found between dietary isoflavonoids from fresh beans and whole-grain bread and HCC risk [13]. However, two Asian studies found higher consumption of soy food may reduce liver cancer risk [14,15]. Part of the inconsistency was mostly likely to be difficult in assessing soy foods intake, particularly among population that soy food intake is uncommon and intake level is low.
To our knowledge, no epidemiological studies have directly assessed the relationship between urinary isoflavonoids and liver cancer risk in humans. Here, we measured urinary isoflavonoids in prediagnostic urine samples of study participants. We aimed to examine their relationships with risk of liver cancer in a nested case–control study within the Shanghai Women’s Health Study (SWHS) and the Shanghai Men’s Health Study (SMHS).
Materials and Methods
Source Population
The current study is based on two ongoing, population-based, prospective cohort studies (SWHS and SMHS) in Shanghai, China. Both cohorts were approved by the Institutional Review Boards of all participating institutions. Written informed consent was obtained from all study participants. The designs and methods used in these studies are similar and have been described in detail previously [15–18].
Nested case-control design
Both cohorts were followed for cancer incidence and all-cause mortality by in-person follow-up surveys carried out every 3 years and annual record linkages with databases at the population-based Shanghai Cancer Registry, Shanghai Vital Statistics, and Shanghai Resident Registry [15–18]. By December 31, 2008, follow-up has led to the identification of 123 women and 181 men incident cases of liver cancer that were coded by the International Classification of Disease, Ninth Revision (ICD-9) [15,19]. Tumors were grouped as primary malignant neoplasms (ICD-9 155.0), malignant neoplasms of the intrahepatic bile ducts (ICD-9 155.1), and unspecified malignant neoplasms of the liver (ICD-9 155.2) [15]. Among them, 37 cases from the SWHS and 50 cases from the SMHS were excluded. The reasons are as follows: 18 female and 21 male subjects were excluded because of a lack of urine samples for assays; 17 female and 19 male subjects were excluded because of a lack of blood samples for assays; 2 female and 10 male subjects were not included because of the confirmation of liver cancer diagnosed after the sampling of lab. After these exclusions, 217 incident cases of liver cancer (86 cases from the SWHS and 131 cases from the SMHS) remained in this study. No difference was found between cases included in this study and those not included with regard to education level, family income, BMI, marriage status, occupation, and regular smoker.
The incidence-density sampling was applied to identify matching controls. For each case, two control subjects was randomly selected from the cohorts who donated a urine and blood sample at baseline and were alive and free of any cancer at the time the index case was diagnosed. Two controls were individually matched to the index case by sex, age at baseline (±2 years), date of biospecimen donation (±1 month), and the interval to the last meal. All cases had two matched control subjects, except for 7 women cases with only 1 control subject because of the sampling criteria. A total of 165 women and 262 men were included in the final analytic dataset.
Exposures and laboratory tests
The baseline survey was conducted in-person at participants’ home by trained interviewers using a validated FFQ[15–18]. Dietary total isoflavonoids and individual isoflavonoids intake for each food calculated by multiplying the amount of food consumed by isoflavonoids content per gram of the food, as obtained from the Chinese Food Composition Tables [20].
Four major isoflavonoids (daidzein, genistein, glycitein, and equol) were measured by the reversed-phase high-performance liquid chromatography (RP-HPLC) using methods described previously [21]. Spot urine samples collected at any time of the day at baseline in each case-control set were arranged in random order, identified only by unique codes, and assayed in the same laboratory run to control batch-to-batch variability. All urine samples were assayed in duplicate. The average intra-assay and inter-assay coefficients of variation were 1.2% and 4.6% for daidzein, 1.5% and 3.9% for genistein, 2.5% and 7.2% for glycitein, and 2.4% and 4.9% for equol, respectively. The detection limits were 12.9 nmol/L for daidzein, 11.8 nmol/L for genistein, 13.9 nmol/L for glycitein, and 71.6 nmol/L for equol. The number of persons with undetectable urinary levels of isoflavones is 16 for daidzein, 21 for genistein, and 83 for glycitein in male subjects, and 10 for daidzein, 16 for genistein, and 45 for glycitein in female subjects, respectively. Undetectable values were replaced by a random number between 0 and the lower limit of detection. Urinary creatinine concentrations were measured by the Jaffé alkaline picrate procedure [22]. Concentrations of urinary isoflavonoids were adjusted for urinary creatinine concentration and expressed as nmol/mg creatinine.
Statistical analysis
Analyses were conducted separately for women and men. Of the total 644 study subjects, genistein and glycitein were missing for 1 subject, and equol was missing for 2 subjects due to peak overlap. Conditional logistic regression models were used to estimate ORs and their corresponding 95% confidence intervals (CIs) with adjustment for the potential confounders. Study subjects were grouped into quartiles based on the distributions of urinary isoflavonoids among all selected male or female control subjects. The linear trend test for the association between urinary concentrations of isoflavonoids and liver cancer risk was based on ordinal values.
Statistical analyses were performed using SAS software, version 9.2 (SAS Institute, Inc). All P values are two-sided. P < 0.05 was considered statistically significant.
Results
The baseline characteristics of study participants by case-control status are shown in Table 1. History of viral hepatitis or cirrhosis was higher in cases than in control in both women and men. In women, the prevalence proportions of regular alcohol drinker, higher fat intake, higher cholesterol intake, history of diabetes at baseline were higher among the individuals who developed liver cancer cases during follow-up compared with those who are not developed liver cancer cases. In men, family history of liver cancer in first-degree relatives and history of cholelithiasis or cholecystectomy at baseline were higher in cases than in control subjects, whereas higher education level and regular physical activity during past 5 years at baseline were lower in cases than in control subjects. No difference was found between cases and control subjects with respect to family income, marriage status, occupation, regular smoker, regular tea drinker, BMI, waist-hip-ratio, total energy intake, carbohydrate intake, protein intake, history of hypertension. Moreover, no difference was found for menopausal status and history of oral contraceptives in women. The correlation coefficient between dietary fat intake and cholesterol intake was 0.76, and adjustment for dietary cholesterol intake resulted in a much bigger change in risk estimates than adjustment for dietary fat energy. Therefore, dietary cholesterol intake was reported in the final multivariable model.
Table 1.
Baseline characteristics of study participants in the nested case-control study of liver cancer by gender
| characteristics | Men |
Women |
||||
|---|---|---|---|---|---|---|
| Cases(N=131) | Controls(N=262) | P | Cases(N=86) | Controls(N=165) | P | |
| Age at recruitment year (years, mean±SD) | 60.09±9.93 | 59.90±9.95 | P=0.859 | 58.99±8.98 | 58.88±8.87 | P=0.927 |
| Total energy intake (kcal/day, mean±SD) | 1911.8±553.17 | 1942.7±462.32 | P=0.560 | 1672.7±501.34 | 1620.9±319.80 | P=0.320 |
| Education level (%) | P=0.050 | P=0.341 | ||||
| Elementary school or less | 15.27 | 13.74 | 52.33 | 48.48 | ||
| Middle school | 41.22 | 39.69 | 17.44 | 26.67 | ||
| High school | 31.30 | 23.28 | 24.42 | 18.18 | ||
| College or above | 12.21 | 23.28 | 5.81 | 6.67 | ||
| Regular physical activity during past 5 years (%) | 37.40 | 47.33 | P=0.062 | 52.33 | 50.30 | P=0.761 |
| Ever had viral hepatitis (%) | 43.51 | 6.11 | P=0.000 | 19.77 | 5.45 | P=0.000 |
| Ever had chronic liver disease or cirrhosis (%) | 19.85 | 3.82 | P=0.000 | 6.98 | 0.61 | P=0.004 |
| Ever had diabetes (%) | 10.69 | 9.54 | P=0.721 | 12.79 | 6.06 | P=0.068 |
| Ever had cholelithiasis or cholecystectomy (%) | 16.79 | 9.16 | P=0.027 | 22.09 | 16.36 | P=0.266 |
| Family history of liver cancer in first-degree relatives (%) | 15.27 | 3.82 | P=0.000 | 9.30 | 4.85 | P=0.171 |
| Regular alcohol drinkera (%) | 32.06 | 37.02 | P=0.333 | 3.49 | 0.61 | P=0.084 |
| Menopausal Status (%) | 74.42 | 72.12 | P=0.698 | |||
| Ever had oral contraceptives (%) | 19.77 | 21.21 | P=0.789 | |||
| Body mass indexb (%) | P=0.107 | P=0.276 | ||||
| Q1 | 35.11 | 25.57 | 31.40 | 25.45 | ||
| Q2 | 23.66 | 24.43 | 23.26 | 24.85 | ||
| Q3 | 16.03 | 25.19 | 15.12 | 24.85 | ||
| Q4 | 25.19 | 24.81 | 30.23 | 24.85 | ||
| Dietary fat intakec (%) | P=0.469 | P=0.025 | ||||
| Q1 | 28.24 | 25.19 | 32.56 | 25.45 | ||
| Q2 | 18.32 | 24.81 | 12.79 | 24.85 | ||
| Q3 | 24.43 | 25.19 | 17.44 | 24.85 | ||
| Q4 | 29.01 | 24.81 | 37.21 | 24.85 | ||
| Dietary cholesterol intaked (%) | P=0.539 | P=0.030 | ||||
| Q1 | 26.72 | 25.19 | 32.56 | 25.45 | ||
| Q2 | 18.32 | 24.81 | 15.12 | 24.85 | ||
| Q3 | 29.01 | 25.57 | 15.12 | 24.85 | ||
| Q4 | 25.95 | 24.43 | 37.21 | 24.85 | ||
| Family income, per person per year (%) | P=0.254 | P=0.580 | ||||
| Low | 70.99 | 63.74 | 38.37 | 32.12 | ||
| Middle | 22.90 | 25.95 | 41.86 | 47.88 | ||
| High | 6.11 | 10.31 | 19.77 | 20.00 | ||
| Marriage status (%) | P=0.297 | P=0.290 | ||||
| Married | 83.97 | 87.79 | 83.72 | 88.48 | ||
| Other | 16.03 | 12.21 | 16.28 | 11.52 | ||
| Occupation (%) | P=0.732 | P=0.125 | ||||
| Professional | 24.43 | 27.86 | 17.44 | 23.78 | ||
| Clerical | 18.32 | 18.70 | 26.74 | 16.46 | ||
| Manual workers | 57.25 | 53.44 | 55.81 | 59.76 | ||
| Regular smoker (%) | P=0.206 | P=0.384 | ||||
| No | 31.30 | 37.79 | 96.51 | 93.94 | ||
| Yes | 68.70 | 62.21 | 3.49 | 6.06 | ||
| Regular tea drinker (%) | P=0.309 | P=0.576 | ||||
| No | 36.64 | 41.98 | 74.42 | 77.58 | ||
| Yes | 63.36 | 58.02 | 25.58 | 22.42 | ||
| Waist-hip-ratioe (%) | P=0.625 | P=0.772 | ||||
| Q1 | 22.90 | 25.19 | 36.05 | 32.12 | ||
| Q2 | 20.61 | 25.19 | 22.09 | 20.61 | ||
| Q3 | 29.77 | 25.95 | 23.26 | 23.03 | ||
| Q4 | 26.72 | 23.66 | 18.60 | 24.24 | ||
| Carbohydrate intakef (%) | P=0.640 | P=0.924 | ||||
| Q1 | 31.30 | 25.19 | 26.74 | 25.45 | ||
| Q2 | 23.66 | 24.81 | 22.09 | 24.85 | ||
| Q3 | 22.90 | 25.19 | 27.91 | 24.85 | ||
| Q4 | 22.14 | 24.81 | 23.26 | 24.85 | ||
| Protein intakeg (%) | P=0.233 | P=0.754 | ||||
| Q1 | 32.06 | 25.19 | 26.74 | 25.45 | ||
| Q2 | 17.56 | 24.81 | 20.93 | 24.85 | ||
| Q3 | 22.14 | 25.19 | 22.09 | 24.85 | ||
| Q4 | 28.24 | 24.81 | 30.23 | 24.85 | ||
| Ever had hypertension (%) | P=0.406 | P=0.329 | ||||
| No | 69.47 | 65.27 | 74.42 | 68.48 | ||
| Yes | 30.53 | 34.73 | 25.58 | 31.52 | ||
The definition for regular alcohol drinker is “Have you drunk alcohol at least 3 times per week, for more than 6 months, continuously”.
Body mass index (kg/m2, cut-off points for the quartiles in the SWHS were ≤22.37, ≤24.67, ≤26.77, and >26.77 and in the SMHS were ≤21.75, ≤23.82, ≤25.56, and >25.56).
Dietary fat intake (g/day, cut-off points for the quartiles in the SWHS were ≤20.84, ≤26.07, ≤32.73, and >32.73 and in the SMHS were ≤24.20, ≤31.50, ≤41.37, and >41.37).
Dietary cholesterol intake (mg/day, cut-off points for the quartiles in the SWHS were ≤167.34, ≤253.05, ≤346.04, and >346.04, and in the SMHS were ≤182.81, ≤294.12, ≤419.68, and >419.68).
Waist-hip-ratio (cut-off points for the quartiles in the SWHS were ≤0.79, ≤0.82, ≤0.86, and >0.86 and in the SMHS were ≤0.85, ≤0.89, ≤0.94, and >0.94).
Dietary carbohydrate intake (g/day, cut-off points for the quartiles in the SWHS were ≤244.61, ≤272.51, ≤317.19, and >317.19 and in the SMHS were ≤274.19, ≤320.14, ≤383.54, and >383.54).
Dietary protein intake (g/day, cut-off points for the quartiles in the SWHS were ≤50.98, ≤62.25, ≤73.36, and >73.36, and in the SMHS were ≤62.22, ≤74.80, ≤90.37, and >90.37).
Adjusted ORs and 95% CIs of liver cancer associated with dietary total and individual isoflavonoids, urinary excretion levels of total isoflavonoids, and four major isoflavonoids are presented in Tables 2 and 3 for women and men, respectively. Total isoflavone intake and three individual isoflavonoids (including daidzen, genistein, and glycitein) either from diet or urine excretion were inversely associated with liver cancer risk in women, although trend tests and point estimates were only significant for higher urinary excretion of genistein. The adjusted ORs (95% CIs) for liver cancer across increasing quartiles of urinary genistein levels were 1.00 (reference), 0.55 (95% CI, 0.22–1.36), 0.57 (95% CI, 0.23–1.43) and 0.19 (95% CI, 0.06–0.59) (Ptrend =0.008), respectively. No association was found between urinary equol level and liver cancer risk in women. In men, none of the individual isoflavonoids from diet or urine excretion showed significant associations with liver cancer risk.
Table 2.
ORs for the association of liver cancer with dietary and urinary excretion of isoflavonoids (women)
| Case (N=86) | Controls (N=165) | OR (95%CI)a | OR (95%CI)b | OR (95%CI)c | |
|---|---|---|---|---|---|
| Dietary isoflavone intake(mg/day) | |||||
| <19.26 | 26 | 42 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 19.26–31.71 | 21 | 41 | 0.93(0.39–2.21) | 0.98(0.39–2.48) | 0.87(0.35–2.16) |
| 31.72–45.16 | 22 | 41 | 0.80(0.32–2.02) | 1.02(0.37–2.82) | 0.80(0.30–2.16) |
| ≥45.17 | 17 | 41 | 0.47(0.18–1.26) | 0.63(0.20–1.97) | 0.51(0.18–1.46) |
| Ptrend | P=0.136 | P=0.504 | P=0.229 | ||
| Dietary daidzein intake(mg/day) | |||||
| <7.93 | 26 | 42 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 7.93–13.80 | 21 | 41 | 0.92(0.38–2.22) | 0.96(0.37–2.51) | 0.83(0.33–2.11) |
| 13.81–18.38 | 21 | 41 | 0.80(0.32–1.96) | 0.99(0.37–2.69) | 0.83(0.32–2.15) |
| ≥18.39 | 18 | 41 | 0.52(0.20–1.36) | 0.68(0.22–2.08) | 0.55(0.20–1.53) |
| Ptrend | P=0.185 | P=0.569 | P=0.289 | ||
| Dietary genistein intake(mg/day) | |||||
| <10.82 | 27 | 42 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 10.82–18.77 | 20 | 41 | 0.80(0.33–1.92) | 0.82(0.32–2.12) | 0.73(0.29–1.85) |
| 18.78–25.17 | 21 | 41 | 0.70(0.28–1.74) | 0.81(0.30–2.20) | 0.66(0.25–1.74) |
| ≥25.18 | 18 | 41 | 0.49(0.19–1.27) | 0.64(0.21–1.96) | 0.52(0.19–1.45) |
| Ptrend | P=0.142 | P=0.465 | P=0.213 | ||
| Dietary glycitein intake(mg/day) | |||||
| <1.66 | 26 | 42 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 1.66–2.43 | 20 | 41 | 0.78(0.32–1.94) | 0.92(0.34–2.50) | 0.88(0.34–2.29) |
| 2.44–3.56 | 19 | 41 | 0.40(0.15–1.07) | 0.37(0.12–1.15) | 0.36(0.13–1.04) |
| ≥3.57 | 21 | 41 | 0.48(0.17–1.35) | 0.63(0.18–2.26) | 0.58(0.19–1.77) |
| Ptrend | P=0.087 | P=0.254 | P=0.152 | ||
| Urinary isoflavones (nmol/mg Cr) | |||||
| <5.03 | 30 | 41 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 5.03–12.05 | 18 | 41 | 0.71(0.29–1.71) | 0.65(0.25–1.65) | 0.64(0.25–1.61) |
| 12.06–25.92 | 23 | 41 | 0.67(0.27–1.68) | 0.83(0.32–2.20) | 0.75(0.28–1.98) |
| ≥25.93 | 14 | 40 | 0.37(0.15–0.93) | 0.43(0.16–1.14) | 0.48(0.18–1.27) |
| Ptrend | P=0.041 | P=0.140 | P=0.183 | ||
| Urinary daidzein(nmol/mg Cr) | |||||
| <2.36 | 28 | 42 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 2.36–6.22 | 23 | 41 | 0.63(0.27–1.49) | 0.57(0.22–1.44) | 0.58(0.23–1.46) |
| 6.23–14.49 | 19 | 41 | 0.43(0.17–1.06) | 0.53(0.20–1.41) | 0.50(0.19–1.32) |
| ≥14.50 | 16 | 41 | 0.35(0.14–0.90) | 0.41(0.15–1.10) | 0.44(0.17–1.17) |
| Ptrend | P=0.018 | P=0.083 | P=0.092 | ||
| Urinary genistein(nmol/mg Cr) | |||||
| <0.85 | 30 | 41 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 0.85–2.57 | 22 | 41 | 0.61(0.25–1.46) | 0.57(0.23–1.45) | 0.55(0.22–1.36) |
| 2.58–6.08 | 23 | 41 | 0.56(0.23–1.32) | 0.57(0.22–1.45) | 0.57(0.23–1.43) |
| ≥6.09 | 11 | 41 | 0.17(0.06–0.49) | 0.20(0.06–0.61) | 0.19(0.06–0.59) |
| Ptrend | P=0.002 | P=0.007 | P=0.008 | ||
| Urinary glycitein (nmol/mg Cr) | |||||
| <0.37 | 23 | 41 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 0.37–1.04 | 26 | 41 | 0.95(0.43–2.10) | 0.88(0.38–2.05) | 0.95(0.41–2.19) |
| 1.05–2.92 | 23 | 41 | 0.76(0.33–1.74) | 0.93(0.39–2.22) | 0.99(0.41–2.39) |
| ≥2.93 | 14 | 41 | 0.51(0.21–1.24) | 0.68(0.27–1.72) | 0.75(0.29–1.94) |
| Ptrend | P=0.132 | P=0.480 | P=0.619 | ||
| Urinary equol | |||||
| negative | 63 | 122 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| positive | 22 | 42 | 0.89(0.43–1.85) | 0.98(0.45–2.12) | 1.01(0.46–2.21) |
Adjusted for total energy intake(kcal/day, continuous variable), education level (elementary school or less, middle school, high school, college or above), regular physical activity during past 5 years (yes or no), history of viral hepatitis (yes or no), history of chronic liver disease or cirrhosis (yes or no), history of diabetes (yes or no), history of cholelithiasis or cholecystectomy (yes or no), and family history of liver cancer (yes or no).
Additionally adjusted for regular alcohol drinker (yes or no), body mass index (kg/m2, cut-off points for the quartiles in the SWHS were ≤22.37, ≤24.67, ≤26.77, and >26.77), dietary fat intake (g/day, cut-off points for the quartiles in the SWHS were ≤20.84, ≤26.07, ≤32.73, and >32.73).
Replace dietary fat intake by dietary cholesterol intake (mg/day, cut-off points for the quartiles in the SWHS were ≤167.34, ≤253.05, ≤346.04, and >346.04) as one of adjustment factors.
Table 3.
ORs for the association of liver cancer with dietary and urinary excretion of isoflavonoids (men)
| Case (N=131) | Controls (N=262) | OR (95%CI)a | OR (95%CI)b | OR (95%CI)c | |
|---|---|---|---|---|---|
| Dietary isoflavone intake(mg/day) | |||||
| <17.13 | 34 | 66 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 17.13–32.25 | 39 | 65 | 1.59(0.75–3.37) | 1.61(0.74–3.52) | 1.64(0.77–3.50) |
| 32.26–47.02 | 17 | 66 | 0.67(0.28–1.56) | 0.62(0.26–1.50) | 0.72(0.30–1.69) |
| ≥47.03 | 41 | 65 | 1.70(0.78–3.71) | 1.36(0.57–3.21) | 1.71(0.76–3.81) |
| Ptrend | P=0.527 | P=1.000 | P=0.515 | ||
| Dietary daidzein intake(mg/day) | |||||
| <7.21 | 35 | 66 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 7.21–13.19 | 36 | 65 | 1.24(0.58–2.63) | 1.21(0.55–2.65) | 1.26(0.59–2.68) |
| 13.20–19.20 | 18 | 66 | 0.69(0.30–1.57) | 0.66(0.28–1.56) | 0.75(0.32–1.72) |
| ≥19.21 | 42 | 65 | 1.60(0.75–3.42) | 1.27(0.55–2.96) | 1.63(0.74–3.58) |
| Ptrend | P=0.458 | P=0.917 | P=0.437 | ||
| Dietary genistein intake(mg/day) | |||||
| <9.27 | 37 | 66 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 9.27–17.51 | 35 | 65 | 1.22(0.58–2.58) | 1.25(0.58–2.72) | 1.28(0.60–2.72) |
| 17.52–26.44 | 19 | 66 | 0.66(0.29–1.48) | 0.64(0.28–1.47) | 0.72(0.32–1.63) |
| ≥26.45 | 40 | 65 | 1.56(0.71–3.41) | 1.24(0.52–2.93) | 1.59(0.71–3.57) |
| Ptrend | P=0.610 | P=0.923 | P=0.571 | ||
| Dietary glycitein intake(mg/day) | |||||
| <1.60 | 42 | 66 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 1.60–2.65 | 30 | 65 | 0.74(0.34–1.64) | 0.72(0.32–1.60) | 0.79(0.35–1.77) |
| 2.66–3.81 | 25 | 66 | 0.71(0.33–1.57) | 0.62(0.27–1.43) | 0.76(0.34–1.71) |
| ≥3.82 | 34 | 65 | 0.97(0.44–2.14) | 0.72(0.30–1.74) | 0.95(0.41–2.17) |
| Ptrend | P=0.824 | P=0.382 | P=0.818 | ||
| Urinary isoflavones (nmol/mg Cr) | |||||
| <3.59 | 37 | 66 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 3.59–8.87 | 25 | 65 | 0.42(0.17–1.01) | 0.39(0.16–0.96) | 0.39(0.16–0.96) |
| 8.88–18.24 | 27 | 66 | 0.74(0.33–1.67) | 0.63(0.27–1.49) | 0.65(0.28–1.51) |
| ≥18.25 | 42 | 65 | 1.08(0.50–2.32) | 0.99(0.46–2.16) | 1.01(0.47–2.17) |
| Ptrend | P=0.396 | P=0.560 | P=0.504 | ||
| Urinary daidzein(nmol/mg Cr) | |||||
| <1.49 | 36 | 66 | 1.00(ref) | 1.00 (ref) | 1.00 (ref) |
| 1.49–5.04 | 29 | 65 | 0.58(0.23–1.44) | 0.53(0.21–1.34) | 0.51(0.20–1.28) |
| 5.05–10.82 | 25 | 66 | 0.85(0.38–1.88) | 0.69(0.30–1.57) | 0.73(0.33–1.64) |
| ≥10.83 | 41 | 65 | 1.08(0.50–2.33) | 0.98(0.44–2.14) | 0.99(0.46–2.13) |
| Ptrend | P=0.489 | P=0.700 | P=0.621 | ||
| Urinary genistein(nmol/mg Cr) | |||||
| <0.54 | 27 | 66 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 0.54–1.82 | 34 | 65 | 1.30(0.56–2.98) | 1.16(0.49–2.77) | 1.22(0.52–2.86) |
| 1.83–4.09 | 35 | 66 | 1.29(0.54–3.08) | 1.12(0.46–2.74) | 1.17(0.47–2.90) |
| ≥4.10 | 35 | 65 | 1.31(0.59–2.92) | 1.17(0.52–2.63) | 1.23(0.55–2.76) |
| Ptrend | P=0.577 | P=0.767 | P=0.690 | ||
| Urinary glycitein (nmol/mg Cr) | |||||
| <0.26 | 31 | 66 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| 0.26–1.11 | 31 | 65 | 0.74(0.31–1.77) | 0.69(0.29–1.66) | 0.68(0.28–1.65) |
| 1.12–2.33 | 23 | 66 | 0.66(0.28–1.54) | 0.55(0.23–1.35) | 0.58(0.25–1.38) |
| ≥2.34 | 46 | 65 | 1.62(0.76–3.44) | 1.38(0.63–3.04) | 1.44(0.67–3.11) |
| Ptrend | P=0.220 | P=0.425 | P=0.354 | ||
| Urinary equol | |||||
| negative | 99 | 199 | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| positive | 32 | 63 | 0.89(0.46–1.71) | 0.87(0.44–1.72) | 0.87(0.45–1.68) |
Adjusted for total energy intake(kcal/day, continuous variable), education level (elementary school or less, middle school, high school, college or above), regular physical activity during past 5 years (yes or no), historyof viral hepatitis (yes or no), history of chronic liver disease or cirrhosis (yes or no), history of diabetes (yes or no), history of cholelithiasis or cholecystectomy (yes or no), and family history of liver cancer (yes or no).
Additionally adjusted for regular alcohol drinker (yes or no), body mass index (kg/m2, cut-off points for the quartiles in the SMHS were ≤21.75, ≤23.82, ≤25.56, and >25.56), dietary fat intake (g/day, cut-off points for the quartiles in the SMHS were ≤24.20, ≤31.50, ≤41.37, and >41.37).
Replace dietary fat intake by dietary cholesterol intake (mg/day, cut-off points for the quartiles in the SMHS were ≤182.81, ≤294.12, ≤419.68, and >419.68) as one of adjustment factors.
We conducted subgroup analyses by limiting the cases to primary malignant neoplasm of liver or malignant neoplasms of the intrahepatic bile ducts. Results were similar to those observed in the entire study participants. The ORs (95% CIs) for the second to the fourth quartiles of the urinary genistein excretion in women were 0.67 (95% CI, 0.25–1.80), 0.66 (95% CI, 0.24–1.81), and 0.31 (95% CI, 0.10–0.96) (Ptrend=0.052) for primary malignancies of the liver and 0.63 (95% CI, 0.17–2.37), 0.58 (95% CI, 0.15–2.23), and 0.18 (95% CI, 0.04–0.90) (Ptrend=0.043) for malignancies of the intrahepatic bile ducts. Because of small sample size of unspecified malignancies of the liver, we did not conduct this subgroup analysis.
When excluding liver cancer cases diagnosed within first year or first 2 years of study enrollment, the ORs (95% CIs) for the fourth quartiles of the urinary genistein excretion in women were 0.21 (95%CI, 0.07–0.66) (Ptrend=0.010) and 0.23 (95% CI, 0.07–0.73) (Ptrend=0.021). Further analyses excluding individuals with self-reported liver disease or cirrhosis at the baseline survey showed that urinary excretion of genistein was consistently associated with reduced risk of liver cancer [second to the fourth quartiles: OR=0.68 (95% CI, 0.24–1.91), OR= 0.65(95% CI, 0.24–1.78), and OR=0.30(95% CI, 0.09–0.99); Ptrend=0.057] among female participants without self-reported liver disease.
Furthermore, we carried out the analysis on dietary intake of isoflavones and the risk of liver cancer based on all members from our two cohorts (see supplemental table S1 and table S2). No statistically significant associations were observed between dietary total and individual isoflavonoids and risk of liver cancer.
Discussion
In this nested case-control study conducted within two prospective, population-based cohort studies, we found an inverse association between genistein excretion in prediagnostic urine and risk of developing liver cancer in women but not in men. To our knowledge, this is the first report of epidemiological study assessing the correlation between urinary isoflavonoids excretion and risk of liver cancer in humans. Our results provide evidence for the hypothesis that genistein may be the key to the potential health effects of soy in liver cancer prevention.
Genistein (4',5,7-trihydroxyisoflavone) is a small, biologically active flavonoid, which is abundant in soy products. Both epidemiologic and experimental studies have suggested that genistein may play an important role in the prevention and progression of tumors including colon cancer, prostate cancer, breast cancer, lung cancer, gastric cancer, leukemia, and melanoma, etc. [23, 24].There is growing evidence from in vitro and animal studies suggesting that genistein also plays a role in human hepatocellular carcinoma (HCC) [2, 4–12]. Genistein was observed to inhibit tumor growth, attachment and invasion in human HCC cell lines (MHCC97-H[4] and BEL-7402[5]) by blocking the FAK signaling process. Genistein induced apoptosis of human Hep3B cells through increasing intracellular reactive oxygen species (ROS), and inducing endoplasmic reticulum (ER) stress and mitochondrial damage[6]. It also induced apoptosis in human SK-Hep-1 cells via both Fas- and mitochondria-mediated pathways [7]. Previous studies provided evidence that genistein had anti-invasive and antimetastatic effects against TPA-mediated metastasis via downregulation of MMP-9 and EGFR and subsequent suppression of NF-κB and activator protein(AP)-1 transcription factors though inhibition of MAPK [7–9], IκB [8, 9], and PI3K/Akt[8, 10] signaling pathways in human hepatocellular carcinoma cells (HepG2, Huh-7, HA22T). There are also in vitro or animal studies that showed the effect of genistein in combination with other agents on human HCC[11, 12]. Genistein synergized with a low dose of ATO to significantly inhibit the growth of HepG2 tumors, and suppress cell proliferation and induce apoptosis in situ [11]. Genistein potentiated the effect of arsenic trioxide against human hepatocellular carcinoma by suppression of ATO-induced Akt and NF-κB activation, and subsequent suppression of NF-κB regulated gene products, including COX-2, c-myc, Bcl-2, Bcl-xL, cyclin D1, and VEGF [12].
We found no association between urinary genistein excretion and liver cancer risk in men. It is unclear why the associations between genistein and liver cancer differ between men and women. Although we tried to adjust some gender-related lifestyle factors such as smoking and alcohol drinking in our models, we cannot completely rule out all gender-specific affect. The differences in sex hormones, dietary habits and other lifestyle factors between men and women may contribute to some of the divergent results observed in men and women. Further biological study are needed. There were several strengths in our study. The design of nested case-control study and the measurement of prediagnostic urine specimens minimized the possible influence of disease symptoms on intake of dietary and other lifestyle factors. Moreover, similar results were observed in our subgroup and sensitivity analyses. Assessment ways including both dietary questionnaire and urinary measure are also a noticeable strength. But our study also had several limitations. It was based on two cohorts of more than 136,500 Chinese women and men; however, the number of patients with incident liver cancer was relatively small due to limited follow-up times. Hence, CIs between isoflavonoids and liver cancer risk were wide. Although we took into consideration participants’history of liver diseases or cirrhosis in our analyses, we cannot thoroughly rule out unmeasured confounding factors such as aflatoxin exposure and hepatitis C virus infection. But the prevalence proportions of aflatoxin exposure and hepatitis C virus infection are very low in Shanghai [15, 25].
It is worth to indicate that the assessment of urinary isoflavonoids was only conducted in baseline spot urine samples. In addition, flavonoid half-lives are short. Our previous study suggested that single spot urine could provide a reasonable measure of the long-term exposure of isoflavone in SMHS [3, 26]. The correlations of isoflavone intake from the second FFQ with those from the multiple 24-HDR ranged from 0.38 for genistein to 0.44 for glycitein, and the correlations with urinary isoflavone levels were 0.48, 0.44, 0.42, 0.54 for total isoflavones, daidzein, genistein, and glycitein. When isoflavone intake and urinary isoflavone levels were categorized by quartile, disagreement (in the opposite quartiles) ranged from 3 to 6% for isoflavones intake and 2 to 4% for urinary isoflavone levels. Mean urinary isoflavone levels were similar across the four seasons. The correlation between individual measurements and the mean measurement across four seasons was reasonably high, with the Spearman correlation coefficients ranging from 0.72 to 0.89. Considering Shanghai is an area with abundant soy foods supplies, the intake levels for most individuals are likely to be relatively stable over time [1]. In our nested case-control study, the correlation coefficients between dietary isoflavone intake and urinary isoflavone level were 0.13 in women and 0.14 in men. On the one hand, we evaluated the effect of dietary total and individual isoflavonoids using the database obtained from Chinese Food Composition Table. Given the marked variability of isoflavonoids in various foods, the statistical analysis might not be completely accurate without a standard chemical analysis of isoflavonoid content in food. On the other hand, urinary isoflavone level is a comprehensive reflection of intake, absorptions, and metabolism, which have been shown substantial interindividual heterogeneity. Although soy exposure measured by FFQ or urinary level (especially spot urine) is subject to measurement error, the association direction between dietary individual isoflavonoids and liver cancer risk is consistent with urinary individual isoflavonoids levels in our study.
In summary, this study showed that higher prediagnostic urinary level of genistein was associated with statistically significantly reduced risk of liver cancer in middle-aged or older women in Shanghai, China. The findings are biologically plausible and consistent with results from recent studies in vitro and animals, suggesting a potential beneficial effect of genistein in the prevention of liver cancer in women. However, a similar association was not observed in men. The reason for such gender-specific association is unknown. Future multicenter or different population studies are warranted to confirm our results.
Supplementary Material
acknowledgments
We would like to thank the participants and the staffs from the Shanghai Women’s and Men’s Health Studies for their contribution to this research.
Y-BX, JW designed and conducted research; W Zhang, JW and Y-BX analyzed data and interpreted results; W Zhang wrote the first draft; All authors read, review and approved the final manuscript. Y-BX obtained funding and had primary responsibility for final content.
Funding: This work was supported by the funds of National Key Project of Research and Development Program of China [2016YFC1302503]; National Key Basic Research Program “973 project” [2015CB554000]; State Key Project Specialized for Infectious Diseases of China [2008ZX10002-015, 2012ZX10002008-002]; and parents cohorts were supported by the grants from the NIH [UM1 CA182910, UM1 CA173640]. The funding organizations had no role in the design and conduct of the study; the collection, analysis, and interpretation of the data; or the preparation, review, or approval of the manuscript.
Abbreviations
- HCC
hepatocellular carcinoma
- SWHS
the Shanghai Women’s Health Study
- SMHS
the Shanghai Men’s Health Study
- ICD-9
International Classification of Disease, Ninth Revision
- BMI
body mass index
- RP-HPLC
reversed-phase high-performance liquid chromatography
- ORs
odds ratios
- CIs
confidence intervals
- FFQ
food-frequency questionnaire
Footnotes
Disclosure: The authors declare no potential conflicts of interest.
references
- 1.Zheng W, Dai Q, Custer LJ, Shu XO, Wen WQ, Jin F, et al. Urinary Excretion of Isoflavonoids and the Risk of Breast Cancer, Cancer Epidemiol Biomarkers Prev, 1999; 8: 35–40. [PubMed] [Google Scholar]
- 2.Franke AA, Halm BM, Kakazu K, Li X, Custer LJ. Phytoestrogenic isoflavonoids in epidemiologic and clinical research. Drug Test Anal 2009; 1: 14–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lee SA, Wen W, Xiang YB, Barnes S, Liu D, Cai Q, et al. Assessment of dietary isoflavone intake among middle-aged Chinese men. J Nutr 2007; 137:1011–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Gu Y, Zhu CF, Dai YL, Zhong Q, Sun B. Inhibitory effects of genistein on metastasis of human hepatocellular carcinoma. World J Gastroenterol 2009; 15:4952–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Gu Y, Zhu CF, Iwamoto H, Chen JS. Genistein inhibits invasive potential of human hepatocellular carcinoma by altering cell cycle, apoptosis, and angiogenesis. World J Gastroenterol 2005; 11:6512–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Yeh TC, Chiang PC, Li TK, Hsu JL, Lin CJ, Wang SW, et al. Genistein induces apoptosis in human hepatocellular carcinomas via interaction of endoplasmic reticulum stress and mitochondrial insult. Biochem Pharmacol 2007; 73:782–92. [DOI] [PubMed] [Google Scholar]
- 7.Fang SC, Hsu CL, Lin HT, Yen GC. Anticancer effects of flavonoid derivatives isolated from Millettia reticulata Benth in SK-Hep-1 human hepatocellular carcinoma cells. J Agric Food Chem 2010; 58:814–20. [DOI] [PubMed] [Google Scholar]
- 8.Wang SD, Chen BC, Kao ST, Liu CJ, Yeh CC. Genistein inhibits tumor invasion by suppressing multiple signal transduction pathways in human hepatocellular carcinoma cells. BMC Complement Altern Med 2014; 14: 26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lee KH, Yeh MH, Kao ST, Hung CM, Liu CJ, Huang YY, et al. The inhibitory effect of hesperidin on tumor cell invasiveness occurs via suppression of activator protein 1 and nuclear factor-kappa B in human hepatocellular carcinoma cells. Toxicol Lett 2010; 194:42–9. [DOI] [PubMed] [Google Scholar]
- 10.Pan MH, Chiou YS, Chen WJ, Wang JM, Badmaev V, Ho CT. Pterostilbene inhibited tumor invasion via suppressing multiple signal transduction pathways in human hepatocellular carcinoma cells. Carcinogenesis 2009; 30:1234–42. [DOI] [PubMed] [Google Scholar]
- 11.Ma Y, Wang J, Liu L, Zhu H, Chen X, Pan S, et al. Genistein potentiates the effect of arsenic trioxide against human hepatocellular carcinoma: role of Akt and nuclear factor-κB. Cancer Lett 2011; 301:75–84. [DOI] [PubMed] [Google Scholar]
- 12.Jiang H, Ma Y, Chen X, Pan S, Sun B, Krissansen GW, et al. Genistein synergizes with arsenic trioxide to suppress human hepatocellular carcinoma. Cancer Sci 2010; 101: 975–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Lagiou P, Rossi M, Lagiou A, Tzonou A, La Vecchia C, Trichopoulos D. Flavonoid intake and liver cancer: a case-control study in Greece. Cancer Causes Control 2008; 19:813–8. [DOI] [PubMed] [Google Scholar]
- 14.Sharp GB, Lagarde F, Mizuno T, Sauvaget C, Fukuhara T, Allen N, et al. Relationship of hepatocellular carcinoma to soya food consumption: a cohort-based, case-control study in Japan. Int J Cancer 2005; 115:290–5. [DOI] [PubMed] [Google Scholar]
- 15.Zhang W, Shu XO, Li HL, Yang G, Cai H, Ji BT, et al. Vitamin intake and liver cancer risk: a report from two cohort studies in China. J Natl Cancer Inst 2012; 104:1174–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Zhang W, Xiang YB, Li HL, Yang G, Cai H, Ji BT, et al. Vegetable-based dietary pattern and liver cancer risk: results from the Shanghai Women’s and Men’s Health studies. Cancer Sci 2013; 104: 1353–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Zheng W, Chow WH, Yang G, Jin F, Rothman N, Blair A, et al. The Shanghai Women’s Health Study: rationale, study design, and baseline characteristics. Am J Epidemiol 2005; 162:1123–31. [DOI] [PubMed] [Google Scholar]
- 18.Shu XO, Li H, Yang G, Gao J, Cai H, Takata Y, et al. Cohort Profile: The Shanghai Men’s Health Study. Int J Epidemiol 2015;44:810–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.World Health Organization. International Classification of Diseases, ninth revision (ICD-9). Geneva, Switzerland: World Health Organization; 1977. [Google Scholar]
- 20.Yang G, Wang G, Pan X. Chinese Food Composition Tables. Beijing, China: Peking University Medical Press; 2002. p.24–193. [Google Scholar]
- 21.Wang J, Han LH, Zhu LF, Gao YT. Determination of isoflavones in human urine by high performance liquid chromatograohy. Chinese J Analyt Chem 2006; 34: 569–72. [Google Scholar]
- 22.Slot C. Plasma creatinine determination. A new and specific Jaffe reaction method. Scand J Clin Lab Invest 1965, 17:381–7 [DOI] [PubMed] [Google Scholar]
- 23.Pavese JM, Farmer RL, Bergan RC. Inhibition of cancer cell invasion and metastasis by genistein. Cancer Metastasis Rev 2010; 29:465–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Russo Maria, Spagnuolo Carmela, Tedesco Idolo, Russo Gian Luigi. Phytochemicals in cancer prevention and therapy: truth or dare? Toxins 2010; 2: 517–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Yuan JM, Ross RK, Stanczyk FZ, Govindarajan S, Gao YT, Henderson BE, et al. A cohort study of serum testosterone and hepatocellular carcinoma in Shanghai, China. Int J Cancer 1995; 63:491–3. [DOI] [PubMed] [Google Scholar]
- 26.Franke AA, Hebshi SM, Pagano I, Kono N, Mack WJ, Hodis HN. Urine accurately reflects circulating isoflavonoids and ascertains compliance during soy intervention. Cancer Epidemiol Biomarkers Prev 2010; 19: 1775–83. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.