Obesity, Age, and Oxidative Stress in Middle-Aged and Older Women

Abstract

Recent evidence suggests that urinary F₂-isoprostanes (F₂-IsoPs) are more accurate markers of oxidative stress than other available biomarkers. Most previous studies used unmetabolized F₂-IsoPs as a biomarker. Few previous studies measured 15-F_2t-IsoP-M, a metabolite of one of the most common F₂-IsoPs, 15-F_2t-IsoP. Unlike 15-F_2t-IsoP, 15-F_2t-IsoP-M is not subject to autoxidation and renal production. To our knowledge, no study has compared the associations of age and body mass index (BMI) with F₂-IsoPs to those with 15-F_2t-IsoP-M. Urinary levels of F₂-IsoPs and 15-F_2t-IsoP-M were measured using gas chromatography–mass spectrometry for 845 healthy women aged 40–70 years. Both F₂-IsoPs and 15-F_2t-IsoP-M were elevated among smokers. The level of 15-F_2t-IsoP-M increased with age, particularly after menopause, and with BMI. In contrast, F₂-IsoPs decreased with age, regardless of menopausal status, and was not related to BMI. The association of 15-F_2t-IsoP-M with age or menopausal status did not differ by BMI category, and the association with BMI was also independent of age or menopausal status. 15-F_2t-IsoP-M appears to be a valuable biomarker of oxidative stress in age- and obesity-related diseases. Antioxid. Redox Signal. 14, 2453–2460.

Introduction

Accumulating evidence from in vitro and in vivo studies indicates that normal basal levels of oxygen-derived free radicals or reactive oxygen species (ROS), acting as secondary messengers, play an essential role in the regulation of various normal physiologies (4, 18), including signal transduction, cell proliferation, homeostasis, microorganism defense, senescence, and apoptosis (1, 4, 22). However, overproduction of ROS may play a causative role in the development of numerous human diseases or conditions, such as cancer, neurodegenerative and cardiovascular diseases, and aging process (5, 6, 9, 32).

Since their first discovery in 1990 (24), F₂-isoprostanes (F₂-IsoPs), a series of free radical-catalyzed lipid peroxidation products of arachidonic acid in situ in phospholipids, have been widely used in animal or human studies to measure in vivo lipid peroxidation, a central feature of free radical damage (18, 31). Studies have shown that F₂-IsoPs in human biological fluids can be the most accurate marker in predicting oxidative stress (20, 26, 30). Urinary F₂-IsoPs have been used often because the urine is easy to sample and has a high presence of F₂-IsoPs (25). However, F₂-IsoPs could be artificially produced in vitro in fluids by autoxidation and its in vivo level in the human urine may be affected by local renal isoprostane production, which is age-dependent (20, 25). After β-oxidation, 15-F_2t-isoprostane (15-F_2t-IsoP), one of the most common F₂-IsoPs, converts to 2,3-dinor-5,6-dihydro-15-F_2t-IsoP (15-F_2t-IsoP-M), a metabolite not subject to autoxidation and renal production (19, 26). A method with both high sensitivity and accuracy has been developed to measure 15-F_2t-IsoP-M using gas chromatography–negative ion chemical ionization mass spectrometry (26). Normal levels of plasma or urinary excretion rates of F₂-IsoPs and 15-F_2t-IsoP-M in humans have been defined previously (19, 24). Although seldom examined in epidemiologic studies (13, 32), 15-F_2t-IsoP-M was suggested as another promising marker of in vivo oxidative stress (6).

Recently, we found that the correlation between urinary F₂-IsoPs and 15-F_2t-IsoP-M was only moderate (r = 0.31) (6). In a nested case–control study, while we observed no significant difference in urinary levels of F₂-IsoPs and 15-F_2t-IsoP-M between breast cancer cases and their matched controls overall, we found that levels of isoprostanes, particularly 15-F_2t-IsoP-M, were associated with a substantially increased risk of breast cancer among overweight/obese women, but not among those of normal weight (6). Several other recent studies have reported that higher levels of body mass index (BMI) were associated with an increased level of F₂-IsoPs (7, 14, 15), further suggesting that isoprostanes are promising biomarkers for obesity-related diseases.

Oxidative stress may also play a critical role in age-related diseases (9, 10). Few studies have examined the associations of F₂-IsoPs with age (3, 17, 21, 29); however, the results are inconclusive. To the best of our knowledge, no study has compared the associations of age and BMI with F₂-IsoPs to those with 15-F_2t-IsoP-M. In the current study, we evaluated the associations of urinary F₂-IsoPs and 15-F_2t-IsoP-M levels with age and BMI using data from a subset of 845 cancer-free midlife women within a large population-based cohort study.

Materials and Methods

Subjects

This cross-sectional study was conducted among 845 women who were free of cancer diagnosis and participants of the Shanghai Women's Health Study (SWHS), a population-based cohort study. Details on the establishment of the cohort have been reported elsewhere (35). In brief, at baseline during March 1997 and May 2000, 74,942 Chinese women between 40 and 70 years of age were recruited in Shanghai, with a 92.7% participation rate. The study was approved by all relevant institutional review boards in the People's Republic of China and the United States. All participants provided informed written consent. Participants of the current study were identified as the control group for a nested case-control study of oxidative stress and breast cancer risk in the SWHS (6).

Data collection

Information on demographic characteristics, history of chronic diseases (e.g. cardiovascular diseases, diabetes mellitus, hypertension, and hepatitis), reproductive factors, and lifestyle factors was collected during in-person interviews conducted by trained interviewers. Smoking was defined as smoking ≥1 cigarette per day for >6 months continuously at some point during a woman's life. Drinking of alcohol (wine, beer, and/or liquor) or tea ≥3 times per week for >6 months continuously was defined as regular consumption of alcohol or tea, respectively. Use of ginseng was defined as taking ginseng and ginseng supplements at least five times per year in the past 3 years before enrollment. Menopausal status was defined as “cease of menstruation for 12 months or longer, excluding that caused by pregnancy and breastfeeding” (34). Participating in any exercise ≥1 time per week for >3 months continuously over the past 5 years preceding the interview was considered as regularly exercising (yes/no). During the interview, anthropometric parameters, including height, weight, waist circumference, and hip circumference were measured twice. BMI was calculated (weight/height², kg/m²).

Approximately 88% of the cohort members donated a spot urine sample at baseline or at the first follow-up 2 years later. The urine sample was collected into a sterilized 100-ml cup containing 125 mg of ascorbic acid and kept in a portable, insulated bag with ice packs (at about 0°C–4°C) and transported to the central laboratory for processing. Ascorbic acid was added to the sample container to prevent degradation of unstable compounds. Within 6 h after collection, samples were aliquoted and stored at −70°C until the laboratory analyses. The biologic sample repositories for the study are equipped with appropriate alarm systems and emergency electricity backup to prevent accidental thawing (35).

Measurement of urinary F₂-IsoPs and 15-F_2t-IsoP-M

Assays were performed for 845 samples in four batches. Urinary excretion of F₂-IsoPs (mainly 15-F_2t-IsoP or 8-iso-PGF_2α) and the major metabolite of 15-F_2t-IsoP, 15-F_2t-IsoP-M (2,3-dinor-5,6-dihydro-15-F_2t-IsoP or 2,3-dinor-5,6-dihydro-8-iso-PGF_2α), were measured by gas chromatography–negative ion chemical ionization mass spectrometry at Vanderbilt Eicosanoid Core Laboratory. The method has been reported in detail elsewhere (20, 25, 26). In brief, the metabolite (2,3-dinor-5,6-dihydro-15-F_2t-IsoP) was chemically synthesized and converted to an [¹⁸O₂]-labeled derivative for use as an internal standard. Both urinary F₂-IsoPs and 15-F_2t-IsoP-M levels were expressed as ng/mg after adjusting for urinary creatinine concentration. Precision of the assay was ± 4% and accuracy was 97%. The lower limit of sensitivity was ∼20 pg/mg (20).

Statistical analysis

Log-transformation was conducted to approach normal distribution for urinary F₂-IsoPs and 15-F_2t-IsoP-M. Geometric means and 95% confidence intervals (CIs) of F₂-IsoPs and 15-F_2t-IsoP-M were obtained based on the least square means estimated from general linear regression models. Potential confounding factors adjusted in regression models include age, menopausal status (pre-/post-), education, occupation, physical activity, BMI, tea drinking, cigarette smoking, vitamin supplement use, and batches for assays. Tests for trends were performed by entering the categorical variables as a continuous parameter in the model. Age was classified as six 5-year age groups from 40–45 to 65–70 years, and BMI was categorized according to the WHO cut points for international classification of BMI (i.e., BMI of ≥25 kg/m² for overweight and ≥30 kg/m² for obesity) as well as cut points for Asian populations (≥23 kg/m² for overweight/obesity), recommended by WHO expert consultation (33). We also performed stratified analyses by BMI (<25.0 vs. ≥25.0) or (normal weight vs. overweight/obesity), menopausal status (pre- vs. post-) and age (median, <51 vs. ≥51 years) to evaluate whether the associations of age and BMI with urinary isoprostanes differed according to these factors. Tests for interaction were performed using the likelihood ratio test. All statistical tests were two-sided and were performed using SAS statistical software, version 9.2 (SAS Institute, Cary, NC).

Results

The average age of the study population at study enrollment was 52.9 ± 8.9 standard deviation (SD) years, and the median age was 51 years. The log-transformed mean levels of urinary F₂-IsoPs and 15-F_2t-IsoP-M were 1.62 (SD 1.51) and 0.56 (SD 0.57) ng/mg of urinary creatinine (cr), respectively. The correlation coefficient was 0.34 between log-transformed F₂-IsoPs and log-transformed 15-F_2t-IsoP-M in all subjects (p < 0.01, data not shown). Table 1 presents the adjusted geometric means and 95% CI for F₂-IsoPs and 15-F_2t-IsoP-M levels (ng/mg cr) according to lifestyle factors (adjusted for age and batches for assays). The associations of isoprostanes with education and occupation were significant (p for trend <0.05 for both, data not shown). Smokers had elevated levels of F₂-IsoPs (p = 0.007) or 15-F_2t-IsoP-M (p = 0.09). Physically active women tended to have lower mean levels of F₂-IsoPs and 15-F_2t-IsoP-M than women who were inactive, but the associations were not significant. Tea drinking (primarily green tea) was associated with an elevated mean level of 15-F_2t-IsoP-M (p = 0.03), but not with F₂-IsoPs. Women who took any vitamin supplement regularly had nonsignificantly decreased F₂-IsoPs, but significantly reduced 15-F_2t-IsoP-M levels (p = 0.008). This significant association was observed mainly with the use of multivitamin or vitamin E supplement, but not of vitamin A, C, or B- complex (data not shown). Women who took ginseng supplements regularly (26% of cohort) had low levels of F₂-IsoPs (p = 0.03); however, there was no appreciable difference in 15-F_2t-IsoP-M levels between ginseng users and nonusers. Regular alcohol consumption, exposure to environmental tobacco smoke, use of fish oil supplement, and self-reported chronic disease history were not significantly related to either biomarker. Therefore, education, occupation, tea drinking, cigarette smoking, and vitamin supplement use were adjusted in subsequent analyses as potential confounding factors.

Table 1.

Urinary F₂-IsoPs and 15-F_2t-IsoP-M Levels (ng/mg cr) in Middle-Aged and Older Women According to Lifestyle-Related Factors, the Shanghai Women's Health Study

		F2-IsoPs		15-F2t-IsoP-M
	n = 845	Geometric mean (95% CI)a	p-valueb	Geometric mean (95% CI)a	p-valueb
Cigarette smoking regularly
No	822	1.60 (1.54–1.66)		0.56 (0.54–0.58)
Yes	23	2.25 (1.77–2.89)	0.007	0.68 (0.54–0.86)	0.09
Environmental tobacco smoke
No	146	1.68 (1.52–1.86)		0.54 (0.54–0.59)
Yes	699	1.62 (1.54–168)	0.46	0.55 (0.50–0.60)	0.44
Alcohol consumption regularly
No	819	1.63 (1.57–1.70)		0.57 (0.54–0.68)
Yes	26	1.49 (1.18–1.88)	0.45	0.56 (0.54–0.70)	0.94
Physically active past 10 years
No	564	1.64 (1.57–1.73)		0.58 (0.54–0.60)
Yes	281	1.58 (1.48–1.70)	0.42	0.54 (0.51–0.58)	0.11
Tea drink regularly
Never	605	1.62 (1.54–1.70)		0.55 (0.53–0.58)
Ever	240	1.63 (1.52–1.77)	0.77	0.60 (0.56–0.64)	0.03
Fish oil supplement use
Never	633	1.60 (1.51–1.68)		0.57 (0.54–0.59)
Ever	165	1.66 (1.52–1.82)	0.47	0.56 (0.52–0.61)	0.72
Ginseng product use
Never	624	1.66 (1.58–1.75)		0.56 (0.53–0.58)
Ever	221	1.51 (1.39–1.63)	0.03	0.58 (0.54–0.63)	0.09
Vitamin supplement use
Never	716	1.65 (1.57–1.72)		0.58 (0.55–0.60)
Ever	129	1.51 (1.36–1.66)	0.13	0.50 (0.45–0.60)	0.008
History of chronic diseases
No	533	1.66 (1.58–1.75)		0.56 (0.53–0.59)
Yes	312	1.55 (1.45–1.75)	0.12	0.57 (0.54–0.61)	0.56

Missing value was (<5%) not included in the percentage estimation.

^aANOVA estimates (geometric means and standard error of the log-transformed isoprostanes [F₂-IsoPs, 15-F_2t-IsoP-M]) adjusted for age and batches assays for F₂-IsoPs or 15-F_2t-IsoP-M (categories).

^bp-value obtained from the ANOVA of the log-transformed values adjusted for the same variables as above.

ANOVA, analysis of variance; CI, confidence interval.

Table 2 presents the association between mean levels of F₂-IsoPs or 15-F_2t-IsoP-M by age or menopausal status at urine collection, and in analyses stratified by BMI (<25.0 vs. ≥25.0). Overall, there was a statistically significant inverse association between urinary F₂-IsoPs levels and age (1.70 ng/mg of creatinine for women aged <45 years and 1.51 for those aged ≥65 years, p for trend = 0.02) independent of menopausal status (p for interaction 0.36, data not shown) and BMI (p for interaction = 0.60). Conversely, urinary level of 15-F_2t-IsoP-M was significantly increased with age (0.50 ng/mg of creatinine for women aged <45 years and 0.62 for those aged ≥65 years, p for trend <0.01). This association was seen regardless of BMI status (p for interaction = 0.29). Figure 1A and B shows the univariate associations of age with F₂-IsoPs and 15-F_2t-IsoP-M, respectively. The mean level of F₂-IsoPs did not differ significantly between premenopausal and postmenopausal women (p = 0.46), regardless of BMI status (p for interaction = 0.55). However, 15-F_2t-IsoP-M levels were significantly elevated among postmenopausal women (0.62, 95% CI: 0.58–0.65) compared to premenopausal women (0.51, 95% CI: 0.48–0.54) (p < 0.01). This pattern of higher 15-F_2t-IsoP-M level in post- than premenopausal women occurred in both normal and overweight/obese women (<25.0 vs. ≥25.0 kg/m²) (p for interaction = 0.48).

Table 2.

Association Between Urinary F₂-IsoPs and 15-F_2t-IsoP-M Levels (ng/mg cr), and Age and Menopausal Status in Middle-Aged and Older Women, the Shanghai Women's Health Study

		F2-IsoPs		15-F2t-IsoP-M
	n	Geometric mean (95% CI)a	p trendb	Geometric mean (95% CI)a	p trendb
Age, years	All women (n = 845)
<45	213	1.70 (1.57–1.84)		0.50 (0.45–0.53)
45–49	159	1.72 (1.55–1.88)		0.55 (0.50–0.59)
50–54	109	1.70 (1.52–1.90)		0.54 (0.49–0.60)
55–59	117	1.60 (1.43–1.79)		0.66 (0.60–0.74)
60–64	121	1.48 (1.32–1.65)		0.59 (0.53–0.64)
≥65	126	1.51 (1.35–1.62)	0.02	0.62 (0.56–0.69)	<0.01
	BMI <25 kg/m2 (n = 517)
<45	164	1.70 (1.54–1.86)		0.48 (0.44–0.53)
45–49	115	1.73 (1.55–1.95)		0.54 (0.49–0.60)
50–54	62	1.65 (1.42–1.91)		0.51 (0.44–0.59)
55–59	63	1.57 (1.35–1.84)		0.62 (0.54–0.73)
60–64	63	1.54 (1.32–1.80)		0.57 (0.49–0.66)
≥65	50	1.55 (1.31–1.84)	0.15	0.66 (0.56–0.77)	0.006
	BMI ≥25 kg/m2 (n = 328)
<45	49	1.68 (1.42–1.99)		0.51 (044–0.59)
45–49	44	1.57 (1.32–1.88)		0.54 (0.47–0.63)
50–54	47	1.79 (1.51–2.12)		0.59 (0.51–0.68)
55–59	54	1.65 (1.40–1.93)		0.72 (0.62–0.82)
60–64	58	1.40 (1.21–1.65)		0.66 (0.54–0.72)
≥65	76	1.49 (1.30–1.65)	0.14	0.62 (0.55–0.70)	0.02
p for interaction			0.60		0.29
Menopausal status	All women (n = 845)
Premenopausal	404	1.65 (1.55–1.75)		0.51 (0.48–0.54)
Postmenopausal	441	1.60 (1.51–1.70)	0.46c	0.62 (0.58–0.65)	<0.01c
	BMI <25 kg/m2 (n = 517)
Premenopausal	289	1.66 (1.55–1.79)		0.50 (0.46–0.53)
Postmenopausal	228	1.63 (1.51–1.77)	0.65c	0.61 (0.56–0.65)	<0.01c
	BMI ≥25 kg/m2 (n = 328)
Premenopausal	115	1.58 (1.42–1.79)		0.53 (0.48–0.58)
Postmenopausal	213	1.57 (1.45–1.72)	0.90c	0.64 (0.60–0.69)	0.002c
p for interaction			0.55		0.48

Age and menopausal status were highly correlated with each other (r = 0.83); thus, they were not included in the same model.

^aANOVA estimates (geometric means and standard error of the log-transformed isoprostanes [F₂-IsoPs, 15-F_2t-IsoP-M]) adjusted for age or menopausal status (pre-/post-), education, occupation, physical activity, BMI, tea drinking, cigarette smoking, vitamin supplement use, and batches assays for F₂-IsoPs or 15-F_2t-IsoP-M (categories).

^bp for trend (^cp value) obtained from the ANOVA of the log-transformed values adjusted for the same variables as above.

BMI, body mass index.

FIG. 1.

Associations between urinary F₂-IsoPs and 15-F_2t-IsoP-M, age and BMI.

Further, we examined the levels of F₂-IsoPs and 15-F_2t-IsoP-M across BMI categories, stratified by age (<51 vs. ≥51 years old) and menopausal status (pre- vs. postmenopause) (Table 3). Overall, the mean levels of F₂-IsoPs tended to decrease with increasing BMI, but not statistically significant (p for trend = 0.35). This inverse association was consistent across categories of age and menopausal status (p for interaction = 0.96 and 0.70, respectively). Conversely, 15-F_2t-IsoP-M level was significantly positively associated with age (0.53 ng/mg of creatinine, 95% CI: 0.49–0.56 for BMI <23.0 and 0.61 ng/mg of creatinine, 95% CI: 0.53–0.72 for BMI ≥30.0, p for trend = 0.02). This positive association appeared in both younger and older age groups (p for interaction = 0.36) and in pre- and postmenopuasal women (p for interaction = 0.40), although the association reached statistical significance only among younger (p for trend = 0.01) or premenopausal women (p for trend = 0.02). Figure 1C and D shows the univariate correlations of BMI with F₂-IsoPs and 15-F_2t-IsoP-M, respectively.

Table 3.

Association Between Urinary F₂-IsoPs and 15-F_2t-IsoP-M Levels (ng/mg cr) and Body Mass Index in Middle-Aged and Older Women, the Shanghai Women's Health Study

		F2-IsoPs		15-F2t-IsoP-M
BMI, kg/m2	n	Geometric mean (95% CI)a	p trendb	Geometric mean (95% CI)a	p trendb
	All women (n = 845)
<23.0	292	1.66 (1.55–1.79)		0.53 (0.49–0.56)
23.0–24.9	225	1.62 (1.49–1.73)		0.59 (0.54–0.63)
25.0–29.9	278	1.60 (1.49–1.72)		0.58 (0.54–0.62)
≥30.0	50	1.57 (1.34–1.86)	0.35	0.61 (0.53–0.72)	0.02
	Age <51 years (n = 391)
<23.0	171	1.72 (1.58–1.86)		0.48 (0.44–0.52)
23.0–24.9	117	1.66 (1.51–1.84)		0.56 (0.51–0.62)
25.0–29.9	93	1.70 (1.51–1.90)		0.57 (0.50–0.64)
≥30.0	10	1.51 (1.10–2.12)	0.63	0.53 (0.37–0.75)	0.01
	Age ≥51 years (n = 454)
<23.0	121	1.65 (1.48–1.84)		0.58 (0.53–0.64)
23.0–24.9	108	1.57 (1.39–1.77)		0.60 (0.54–0.66)
25.0–29.9	185	1.51 (1.38–1.65)		0.62 (0.56–0.65)
≥30.0	40	1.55 (1.28–1.90)	0.30	0.66 (0.56–0.78)	0.27
p for interaction			0.96		0.36
	Premenopausal (n = 404)
<23.0	173	1.68 (1.54–1.82)		0.47 (0.43–0.51)
23.0–24.9	116	1.63 (1.46–1.80)		0.56 (0.51–0.62)
25.0–29.9	102	1.68 (1.51–1.90)		0.55 (0.50–0.62)
≥30.0	13	1.43 (1.06–1.95)	0.73	0.51 (0.37–0.69)	0.02
	Postmenopausal (n = 441)
<23.0	119	1.65 (1.47–1.84)		0.59 (0.54–0.66)
23.0–24.9	109	1.60 (1.43–1.80)		0.61 (0.55–0.67)
25.0–29.9	176	1.54 (1.40–1.68)		0.62 (0.56–0.66)
≥30.0	37	1.62 (1.34–1.97)	0.48	0.68 (0.57–0.80)	0.30
p for interaction			0.70		0.40

Age is grouped according to median age (<51 and ≥51 years).

^aANOVA estimates (geometric means and standard deviation of the log-transformed isoprostanes [F₂-IsoPs, 15-F_2t-IsoP-M]) adjusted for age or menopausal, status, education, occupation, physical activity, tea drinking, cigarette smoking, vitamin supplement use, and batches assays for F₂-IsoPs or 15-F_2t-IsoP-M (categories).

^bp for trend obtained from ANOVA of the log-transformed values adjusted for the variables as above.

Discussion

In this Chinese population, urinary F₂-IsoPs levels were similar to the urinary levels of F₂-IsoPs in the U.S. population, as defined previously (1.6 ± 0.6 ng/mg of creatinine) (26). However, urinary isoprostane metabolite (15-F_2t-IsoP-M) levels were higher in our population than in the U.S. population (0.39 ± 0.18 ng/mg cr) (26), but race/ethnicity, age, and gender of study subjects were not reported. Although few women smoked in our population, we found that women who smoked had elevated levels of both F₂-IsoPs and 15-F_2t-IsoP-M than nonsmoking women, a finding that is consistent with the results of previous studies on F₂-IsoPs (3, 11, 23), with the exception of the small nested case–control study with 26 pairs of diabetes and controls in which current smokers tended to have a reduced level of 15-F_2t-IsoP-M compared with nonsmokers or ex-smokers (13). To our knowledge, no study has compared the associations of age and BMI with F₂-IsoPs to those with 15-F_2t-IsoP-M. We found the level of F₂-IsoPs decreased with age, whereas 15-F_2t-IsoP-M significantly increased with age and postmenopausal status. Further, the urinary level of 15-F_2t-IsoP-M was significantly positively associated with BMI level, whereas the level of urinary F₂-IsoPs was not significantly associated with BMI status. These findings suggest that, in addition to F₂-IsoPs, using 15-F_2t-IsoP-M as a biomarker will provide independent information, particularly in obesity- and age-related diseases (5, 29, 32).

ROS are continuously formed and degraded in normal cellular processes. It was hypothesized that ROS including oxygen-containing free radicals play a role in the changes associated with aging or in age-induced oxidative stress (9, 10). In this study, we found that a marker of oxidative stress, urinary isoprostane metabolite (15-F_2t-IsoP-M), but not F₂-IsoPs, was positively associated with age or postmenopausal status. Our finding on F₂-IsoPs was consistent with that in two large-scale epidemiologic studies, the Framingham Heart Study (15) and the European Union Study (3), in which urinary concentrations of F₂-IsoPs decreased with age, particularly among nonsmokers in the latter study. Other small-scale studies found a positive (23) or no association (17) between age and F₂-IsoPs. In addition to smoking status, it is possible that the association between age and urinary F₂-IsoPs may be confounded by age-dependent kidney function (29), which might explain in part the inconsistencies in previous studies, whereas 15-F_2t-IsoP-M is not affected by local kidney production (12). Further investigations of the associations between age and both parent compound and metabolite isopostanes are needed to clarify these scientific issues.

Prior studies reported that higher plasma or urinary levels of F₂-IsoPs were associated with higher BMI (7, 14, 15). Both BMI and waist circumference were positively related to urinary F₂-IsoPs in women and men in Japan (7). Our inverse, but not statistically significant, association between F₂-IsoPs and BMI did not support these previous findings. One possible explanation for this inconsistency is that previous studies mainly used the immunoassay method to quantify F₂-IsoPs, in which a cross-reactivity was found between the immunoassay antibody of F₂-IsoPs and structurally related isoprostane isomers (28), including 15-F_2t-IsoP, the major precursor of 15-F_2t-IsoP-M (27). Thus, it is possible that 15-F_2t-IsoP-M is the underlying factor for the positive association between urinary F₂-IsoPs and BMI in most previous studies. Our results of a significant positive association of 15-F_2t-IsoP-M with BMI support this possibility. In addition, findings of this study is consistent with our recent report that an increased risk of breast cancer was strongly related to urinary 15-F_2t-IsoP-M level than F₂-IsoPs level, among overweight/obese women (6). It is possible that the positive association between BMI and isoprostane metabolite may be due to their links to estrogen. Still, replication is needed for our findings on the association between urinary 15-F_2t-IsoP-M and BMI.

The present study has a number of notable strengths. Urinary levels of F₂-IsoPs were measured together with its major metabolite 15-F_2t-IsoP-M using a newly developed, more sensitive method. Another major strength of our study was its relatively large sample size. The parent population-based cohort study had remarkably high rates for baseline participation and follow-up, which minimized selection bias. We were able to adjust for many potential confounding factors, including smoking and vitamin supplement use.

There are several limitations of the present study. One limitation is that a single urine sample was used. However, reliability studies found that at a group level, F₂-IsoPs measured in a single urine sample were comparable to that measured using multiple samples or a 24-h urine sample (2). Previous studies generated inconsistent results on the inter-day variation of urinary isoprostanes levels (2), whereas our previous data in the same study population suggest that the major contributor to intra-person variation is seasonal fluctuation (16). Since inter-day variation is random, any residual inter-day variation may lead to nondifferential misclassification, which usually biases the result to the null. To the extent that residual inter-day variation levels exist in our data, the true associations could be stronger than those we observed. Moreover, urinary isoprostanes, age, and BMI were measured at the same time. The temporal sequence thus is unclear for the observed associations. However, it is unlikely that the dose–response associations between age and BMI with urinary concentrations of F₂-IsoPs and 15-F_2t-IsoP-M are a result of selection biases, as measurement of biomarkers was blinded to subject characteristics. Another limitation is that we did not collect information on plasma lipids and estradiol, and we were therefore unable to investigate the effects of these biomarkers on levels of urinary isoprostanes. However, we have calculated the dietary intake of lipids and found no apparent associations between dietary intakes of arachidonic acid and other lipids and F₂-IsoPs or 15-F_2t-IsoP-M. Further, we have investigated the associations between urinary isoprostane levels and erythrocyte membrane phospholipid polyunsaturated fatty acid concentrations in our validation study among 48 Chinese men with measurements of blood lipids and urinary isoprostanes for each season over a period of year, and found no significant associations. Further investigations are needed to study how tissue levels of lipids affect isoprostane levels (8).

One of the most critical needs in free radical research in humans has been the development of a reliable noninvasive method to assess oxidative stress (25, 30). High levels of F₂-IsoPs have been linked to a number of obesity- or age-related diseases (9, 15, 29). Our findings suggest that in addition to F₂-IsoPs, 15-F_2t-IsoP-M is another valuable biomarker of oxidative stress in age- and obesity-related diseases. Further studies are warranted to confirm our findings.

Abbreviations Used

15-F_2t-IsoP

isoprostane 8-iso-prostaglandin F2α

15-F_2t-IsoP-M

isoprostane 8-iso-prostaglandin F2α metabolite

BMI

body mass index

CI

confidence interval

F₂-IsoPs

F₂-isoprostanes

ROS

reactive oxygen species

SD

standard deviation

SWHS

Shanghai Women's Health Study

WHO

World Health Organization

Acknowledgments

Dr. Jason Morrow, our beloved colleague, long-term collaborator, and key co-investigator of the project, passed away. Without him, our research work would have been impossible. This article is dedicated to his memory. The authors thank the Shanghai residents who participated in the study and the research staff of the SWHS for their dedication and contributions to the study. This work was supported by research grant R01CA106591 as well as parent studies R37CA70867 and N02 CP1101066 from the National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the National Institutes of Health.

Author Disclosure Statement

No competing financial interests exist.