Impact of age and metabolic syndrome-like components on prostate cancer development: a nationwide population-based cohort study.

BACKGROUND: Because of the contradictory results, more epidemiologic data is needed to determine if metabolic syndrome is a risk factor for developing prostate cancer. This study investigated whether metabolic syndrome-like components affect the incidence of prostate cancer in a Korean population.
METHODS: Men over 50 years of age who underwent health examinations in 2009 were followed until December 2015 (n=1,917,430) using National Health Insurance System data. Subjects were divided into three groups according to the number of metabolic syndrome-like components. The predictive accuracy of age for prostate cancer was assessed by the Youden index and multivariate adjusted Cox regression analysis was used to analyze the effect of metabolic syndrome-like components on prostate cancer development.
RESULTS: The risk of prostate cancer increases with age, and the best cutoff age for prostate cancer detection was 62 years (the maximum value of the Youden index). When stratified by the number of metabolic syndrome-like components, the age with the highest Youden index of each group is still 61 or 62 years. In multivariate adjusted Cox regression analysis, there was no statistically significant difference in the incidence rate among the non-component group, the group with 1 or 2 components, and the group with ≥3 components.
CONCLUSIONS: The current study found that there was no statistically significant association between metabolic syndrome and prostate cancer development in a Korean population. However, results of this study should be interpreted with consideration due to several limitations including the diversity of definitions of metabolic syndrome components. 2021 Translational Andrology and Urology. All rights reserved.

Introduction

Prostate cancer is the second most commonly diagnosed cancer among men worldwide (1). The incidence of prostate cancer has also increased significantly in Korea, and is mainly due to rapid population aging, westernized dietary habits, and increased prostate-specific antigen (PSA) measurements in screening tests (2). Over the past several decades, the prevalence of metabolic syndrome has increased worldwide and has emerged as a public health problem (3). According to recent meta-analyses, westernized dietary patterns such as a greater intake of dietary fat and meat is associated with the incidence of metabolic syndrome (4,5). Additionally, many reports have shown the association between metabolic syndrome and development of various cancers such as colorectal, breast, endometrial, pancreas, and primary liver cancers (6).

In addition to these cancers, metabolic syndrome components could be risk factors for prostate cancer development. However, reports from previous studies about the link between prostate cancer and metabolic syndrome have been inconsistent (7). Although some studies have suggested that metabolic syndrome can increase the risk of developing prostate cancer (8-10), some studies have reported no association or a negative association between metabolic syndrome and prostate cancer (11-13). These inconsistencies have also been observed in the Asian population studies (14).

Because of the contradictory results, more epidemiologic data is needed to determine if metabolic syndrome is a risk factor for developing prostate cancer. Therefore, the objective of this study was to investigate whether metabolic syndrome affects the incidence of prostate cancer in a Korean population based on the National Health Insurance System (NHIS) data. We present the following article in accordance with the STROBE checklist (available at https://dx.doi.org/10.21037/tau-21-249).

Methods

Data base and demographic factors

This study used the NHIS database of Korea to identify patients with prostate cancer between January 2009 and December 2015 (NHIS-2017-1-222) (15). The NHIS database is composed of an eligibility database, medical treatment database, medical care institution database, and health examination database. Patients who were diagnosed with prostate cancer could be identified using diagnostic codes for the medical treatment database.

Information about metabolic syndrome was reported in the health examination database. In this manuscript a metabolic component refers to one of the following five metabolic syndrome-like components: (I) waist circumference ≥90 cm, (II) elevated triglycerides ≥150 mg/dL, (III) reduced HDL cholesterol <40 mg/dL, (IV) systolic blood pressure ≥130 mmHg or diastolic blood pressure ≥85 mmHg, and (V) fasting serum glucose level ≥100 mg/dL. Subjects were divided into three groups according to the number of metabolic syndrome-like components (a non-component group, a group with one or two components, and a group with ≥3 components) to analyze the effect of metabolic syndrome on prostate cancer development.

Lifestyle variables were also included in the health examination database. Smoking status was categorized into non-smokers, ex-smokers, and current smokers. Alcohol consumption status was categorized into four groups: non-drinker, light drinker (1–2 days/week), moderate drinker (3–4 days/week), and heavy drinker (≥5 days/week). Regular exercise over 20 minutes was also categorized into three groups: 0–1 days/week, 2–4 days/week, and ≥5 days/week.

Study population

In Korea, a national health examination is semi-mandatory to local householder, company member and family member over the age of 40 and dependents of member once every 2 years. Therefore, during 1 year, about half of the population over the age of 40 undergo examination. And it is not possible to receive more than two national health examination during 1 year. In this study, male subjects that had undergone a national health examination in 2009 without a previous diagnosis of any cancer were included. Young men under 50 years of age were excluded because of the rare development of prostate cancer in this group. After excluding people with missing metabolic disease data from health examination databases, a total of 1,917,430 men were followed from January 2009 to December 2015 ().Subjects who the development of prostate cancer did not occur during the follow up period were censored and Cox regression analysis was used for these censored data. Sensitivity analysis was used to handle missing data.

Figure 1

Study design and subject disposition.

Statistical analysis

SAS software version 9.4 (SAS Institute, Cary, NC, USA) was used for all statistical analyses. Baseline demographic and clinical characteristics of subjects are presented as number (%). Incidence rate is expressed as the number of newly diagnosed cases of prostate cancer per 100,000 person-years of the follow-up period. One-way ANOVA used to test difference in the incidence rate among the non-component group, the group with 1 or 2 components, and the group with ≥3 components. The predictive accuracy of age for prostate cancer was assessed by calculating the c-index based on the receiver operating characteristics (ROC) curve. The optimum cut-off value was defined as the maximum value of the Youden index. Multivariate adjusted Cox regression analysis was conducted to examine the hazard ratio (HR) and 95% confidence interval (CI) for the development of prostate cancer by metabolic syndrome-like components. Calculations were adjusted for age, alcohol consumption, smoking status, and regular exercise. Statistical significance was considered when the P value was less than 0.05.

Ethical approval

This study protocol was approved by the Institutional Review Board of Bucheon St. Mary’s Hospital of Korea (HC20ZISI0062). The study was performed in accordance with the ethical principles of the Declaration of Helsinki (16). Informed consent was waived because anonymous and de-identified information was used for analysis.

Results

Demographic and clinical characteristics according to metabolic disease

summarizes the demographic and clinical characteristics of study subjects. Among a total of 1,917,430 participants, 22,584 (1.18%) were diagnosed with prostate cancer, including 3,387 (1.13%) subjects in the non-component group, 13,151 (1.19%) in the group with 1 or 2 components, and 6,046 (1.17%) in the group with ≥3 components. About 85% of the subjects had more than one metabolic syndrome-like component in this cohort. More than half of the subjects had hypertension, followed by hyperlipidemia and diabetes. The numbers of non-, ex-, and current smoker were similar among the groups, and the ratio of non-drunker was highest in the non-component group.

Table 1

Baseline demographic and clinical characteristics according to metabolic syndrome-like components

Metabolic disease	Non-component	1 or 2 components	≥3 components	Total
No. in population	301,030 (100.00)	1,101,800 (100.00)	514,600 (100.00)	1,917,430
No. of diagnosed prostate cancers	3,387 (1.13)	13,151 (1.19)	6,046 (1.17)	22,584
Age, years
50–54	109,428 (36.35)	350,969 (31.85)	161,137 (31.31)	621,533
55–59	60,838 (20.21)	222,893 (20.23)	108,289 (21.04)	392,020
60–64	53,612 (17.81)	208,056 (18.88)	101,231 (19.67)	362,899
65–69	34,890 (11.59)	143,595 (13.03)	67,911 (13.2)	246,396
70–74	27,332 (9.08)	114,695 (10.41)	52,207 (10.15)	194,234
≥75	14,930 (4.96)	61,593 (5.59)	23,825 (4.63)	100,348
BMI, kg/m2
<18.5	16,608 (5.52)	27,234 (2.47)	2,475 (0.48)	46,317
18.5–22.9	162,537 (53.99)	402,599 (36.54)	72,771 (14.14)	637,907
23.0–24.9	82,674 (27.46)	338,373 (30.71)	123,609 (24.02)	544,656
≥25.0	39,211 (13.03)	333,594 (30.28)	315,745 (61.36)	688,550
WC, cm
<90	301,030 (100)	918,346 (83.34)	206389 (40.11)	1,425,765
≥90	0	183,454 (16.66)	308,211 (59.89)	491,665
Triglycerides, mg/dL
<150	301,030 (100)	796,209 (72.26)	110,865 (21.54)	1,208,104
≥150	0	305,591 (27.74)	403,735 (78.46)	709,326
HDL cholesterol, mg/dL
≥40	301,030 (100)	989,384 (89.8)	322,583 (62.69)	1,612,997
<40	0	112,416 (10.2)	192,017 (37.31)	304,433
Hypertension
No	301,030 (100.00)	521,232 (47.31)	92,554 (17.99)	914,816
Yes	0	580,568 (52.69)	422,046 (82.01)	1,002,614
Diabetes
Normal	301,030 (100.00)	632,804 (57.43)	109,226 (21.23)	1,043,060
Pre-diabetes	0	361,921 (32.85)	278,724 (54.16)	640,645
Diabetes	0	107,075 (9.72)	126,650 (24.61)	233,725
Smoking status
Non	106,060 (35.45)	381,514 (34.81)	170,126 (33.25)	657,700
Ex	86,319 (28.85)	339,812 (31.01)	166,163 (32.47)	592,294
Current	106,816 (35.7)	374,522 (34.18)	175,440 (34.28)	656,778
Alcohol consumption status
Non	132,930 (45.96)	431,701 (41.05)	185,435 (37.99)	750,066
1–2 days/week	102,307 (35.37)	369,920 (35.17)	171,933 (35.22)	644,160
3–4 days/week	37,154 (12.85)	170,296 (16.19)	90,437 (18.53)	297,887
≥5 days/week	16,854 (5.83)	79,802 (7.59)	40,362 (8.27)	137,018
Regular exercise
0–1 day/week	196,434 (66.35)	717,216 (66.17)	340,579 (67.27)	1,254,229
2–4 days/week	68,394 (23.1)	249,366 (23.01)	115,215 (22.76)	432,975
≥5 days/week	31,241 (10.55)	117,316 (10.82)	50,462 (9.97)	199,019

Data are presented as number (%). BMI, body mass index; WC, waist circumference; Hypertension, systolic blood pressure ≥130 mmHg or diastolic blood pressure ≥85 mmHg; Pre-diabetes, fasting serum glucose level ≥100 mg/dL; Diabetes, fasting serum glucose level ≥126 mg/dL.

Incidence rate of prostate cancer according to age

shows the incidence rate of prostate cancer according to age. The number of new cases of prostate cancer increased with age. The total incidence rate of prostate cancer increased from 28.07 per 100,000 person-years in the 51-year-old age group to 366.01 per 100,000 person-years in the 75-year-old age group. A similar increasing trend was found when subjects were analyzed by groups, according to the number of metabolic syndrome-like components. There was no statistically significant difference in the incidence rate among the non-component group, the group with 1 or 2 components, and the group with ≥3 components (Table S1).

Figure 2

Incidence rate of prostate cancer: (A) total incidence rate of prostate cancer, (B) incidence rate of prostate cancer according to the number of metabolic syndrome-like components.

Age cut-off value for predicting prostate cancer according to metabolic diseases

Sensitivity, specificity, and Youden index for predicting the development of prostate cancer with different age cut-off values are shown in Table S2. The age with the highest Youden index was 62 years for all subjects. The cut-off value for age was identified based on the highest Youden index. When subjects were stratified by the number of metabolic syndrome-like components, the age with the highest Youden index of each group was also 62 years ().

Figure 3

Youden index for predicting the development of prostate cancer with different age cut-off values. (A) All subjects, (B) based on the number of metabolic syndrome-like components.

Risk for prostate cancer according to metabolic diseases when stratified by the number of metabolic diseases

The HRs (95% CI) for prostate cancer according to metabolic syndrome-like components were: 0.992 (0.954–1.031) for the group with 1 or 2 components and 0.974 (0.933–1.018) for the group with ≥3 components in a model adjusted for age, alcohol consumption, smoking status, and regular exercise (). There was no statistically significant difference in the incidence rate among the non-component group, the group with 1 or 2 components, and the group with ≥3 components.

Table 2

Age- and multivariable-adjusted HRs for prostate cancer according to metabolic syndrome-like components

Metabolic disease	Event	Person-years	Incidence*	HR (95% confidence interval)			HR (95% confidence interval)
				Model 1†	P value		Model 2‡	P value
Non-component	3,387	2,327,477	145.52	Ref.			Ref.
1 or 2 components	13,151	8,459,308	155.46	0.997 (0.960–1.035)	0.874		0.992 (0.954–1.031)	0.679
≥3 components	6,046	3,954,513	152.74	0.986 (0.946–1.029)	0.526		0.974 (0.933–1.018)	0.241

*, all rates are expressed as number per 100,000 person-years; †, adjusted for age; ‡, adjusted for age, alcohol consumption, smoking status, and regular exercise. HR, hazard ratio.

Discussion

The main findings of this study are as follows: (I) the risk of prostate cancer increases with age; (II) the age of 62 years (with the maximum value of the Youden index) may be a point of inflection; (III) when stratified by the number of metabolic syndrome-like components, the age with the highest Youden index of each group was 61 or 62 years; and (IV) there was no statistically significant difference in the incidence rate among the non-component group, the group with 1 or 2 components, and the group with ≥3 components.

The prevalence of components of metabolic syndrome including obesity, hypertension, dyslipidemia, and diabetes has sharply increased worldwide (3). Similar to other developing or newly developed countries, this increasing trend is also observed in Korea due to lifestyle changes associated with westernization such as a greater intake of dietary fat and meat. Lee et al. recently reported that the prevalence of metabolic syndrome increased significantly from 28% in 2009 to 30% in 2013, and this increase was more pronounced in men than in women using the Korean NHIS data (17). Prevention of metabolic syndrome is important, because numerous data have shown the association between metabolic diseases with development of various cancers (6).

Components of metabolic syndrome could also be risk factors for prostate cancer development. However, conclusions of previous studies about the link between prostate cancer and metabolic syndrome have been inconsistent (7). Beebe-Dimmer et al. (8) reported that features of metabolic syndrome are associated with prostate cancer in African-American men. Hypertension and abdominal obesity have been reported to be more common among men with prostate cancer, although diabetes is not associated with the risk of prostate cancer. These authors have also found that the association between metabolic syndrome and prostate cancer risk differs by race. There is a significant association in African-American men, but no association in white men (9). Additionally, Lund Haheim et al. (10) reported that metabolic syndrome could predict the incidence of prostate cancer in a cohort of middle-aged Norwegian men.

However, some studies have reported no association or a negative association between metabolic syndrome and prostate cancer. Wallner et al. (12) found that metabolic syndrome is only minimally and inversely associated with prostate cancer and that there is no monotonic association between the number of metabolic components and prostate cancer in Caucasian men. Tande et al. (13) even reported that metabolic syndrome is associated with a decreased incidence of prostate cancer using data from the Atherosclerosis Risk in Communities (ARIC) Study, a multicenter prospective cohort in the United States. Recently, Xu et al. reported that components of metabolic syndrome were not associated with biochemical recurrence through a retrospective analysis of a Chinese cohort (18). Like these studies, in our study, there was no statistically significant association between metabolic syndrome and prostate cancer development. Additionally, the age with the maximum Youden index value for each group was similar when subjects were stratified by the number of metabolic syndrome-like components.

Differences in factors such as the size of the cohort, follow-up period, and the cut-off value of components of metabolic syndrome could provide rational for why previous studies have inconsistent conclusions. Numerous definitions of metabolic syndrome have been proposed by various organizations like the World Health Organization (WHO), the European Group for Study of Insulin Resistance (EGIR), the American Association of Clinical Endocrinologists (AACE), the National Cholesterol Education Program (NCEP), among others (19). Therefore, definitions of metabolic syndrome components could differ from those used in other previous studies. For example, for the definition of abdominal obesity, the WHO defines central obesity as waist/hip ratio >0.9 in men; however, the EGIR and the NECP use waist circumference. In this study, the waist circumference cut-off value for abdominal obesity was ≥90 cm for men, according to Korean guidelines (20).

In addition, there are several limitations that must be considered. The present study only analyzed the effect of metabolic syndrome-like components on prostate cancer risk in one geographic site, though the use of large-scale nationwide cohort data is obviously a strength of this study which had a high statistical power. Second, detection bias could have occurred. Compared with the non-component group, groups with metabolic syndrome-like components could be more likely to visit a hospital for any health problems. These participants are consequently more likely to have a PSA test. Finally, since metabolic syndrome-like components of an individual can change during the follow-up period, the results of this study should be interpreted with consideration. Our study used single measurements of waist circumference, triglycerides, cholesterol, blood pressure, and fasting serum glucose at baseline. Therefore, future studies should also consider effects of changes in components of metabolic syndrome on prostate cancer risk.

Conclusions

The current study found that there was no statistically significant association between metabolic syndrome and prostate cancer development. The ages with the maximum Youden index value of each group were similar when subjects were stratified by the number of metabolic syndrome-like components. However, results of this study should be interpreted with consideration due to several limitations. Therefore, further study is needed to analyze the association between metabolic syndrome and prostate cancer development.

The article’s supplementary files as