Abstract
Human studies suggest that oxidative stress is a risk factor for osteoporosis, but its relationship with fracture risk is poorly understood. The purpose of the present study was to investigate the association between biomarkers of oxidative stress and hip fracture in postmenopausal women. We conducted a prospective study in the Nurses’ Health Study among 996 women aged 60 years or older at baseline blood collection in 1989–1990. Plasma fluorescent oxidation products (FlOPs) were measured at three excitation/emission wavelengths (360/420 nm named as FlOP_360; 320/420 nm named as FlOP_320 and 400/475 nm named as FlOP_400). FlOPs are generated from many different pathways (lipid, protein and DNA) and reflect a global oxidation burden. FlOP assay is 10–100 times more sensitive than measurement of malondialdehyde. We used Cox proportional hazards regression model to investigate the association between baseline plasma FlOPs and risk of hip fracture, adjusting for multiple hip fracture risk factors such as age, history of osteoporosis, history of hypertension, prior fracture and smoking status. Forty four hip fractures (4.4%) were identified during the follow-up (Maximum = 23 years). In the multivariable model, the hazard ratios (HR) of hip fracture in the second and third tertiles of FlOP_320 were 2.11 (95% confidence interval [CI] = 0.88–5.10) and 2.67 (95% CI = 1.14–6.27), respectively, in comparison with the lowest tertile, and the risk increased linearly with increasing FlOP_320 (P for trend = 0.021). Neither FlOP_360 nor FlOP_400 was significantly associated with risk of hip fracture (Tertile 3 versus tertile 1: HR = 0.70, 95% CI = 0.32–1.54, P for trend = 0.386 for FlOP_360; and HR = 0.88, 95% CI = 0.40–1.96, P for trend = 0.900 for FlOP_400). In this prospective study, higher plasma FlOP_320 was an independent risk factor for hip fracture. Our results need further confirmation.
Keywords: Fluorescent oxidation products, oxidative stress, hip fracture, postmenopausal women
Introduction
Oxidative stress is generated as a result of insufficient activity of the endogenous antioxidant defense system against reactive oxygen species (ROSs). Excessive ROSs, such as superoxide anion, hydrogen peroxide and peroxynitrites are able to exert oxidative damage to lipids, DNA, carbohydrate and protein. This oxidative damage is involved in the development of many diseases, including coronary heart disease, cancer and neurodegenerative diseases (1–6).
Whether oxidative stress is related to fracture risk is poorly understood. Experimental studies in cell and animal models suggested that oxidative stress is an important factor in the regulation of bone remodeling (7–9). The results have been extended by cross-sectional and case-control studies, in which oxidative stress characterized by high level of F2-isoprostanes in urine and low level of antioxidant enzymes in blood is associated with reduced bone mineral density and increased risk of osteoporosis in the elderly (10,11). Thus, further investigation on the prospective relationship between oxidative stress and fragility fracture is of great importance to better understand bone health.
The level of oxidative stress can be assessed by the fluorescent oxidation products (FlOPs) as shown in our previous work (4,5,12). ROSs interact with proteins, phospholipids and nucleic acids to form FlOPs which produce coloring products with characteristic fluorescence spectra and can be measured at three excitation/emission wavelengths (360/420 nm named as FlOP_360; 320/420 nm named as FlOP_320 and 400/475 nm named as FlOP_400) (4–6,12). FlOPs measured at different wavelengths reflect different oxidation products (see method section). Our previous studies have demonstrated that these three types of FlOPs have differential associations with coronary heart disease and breast cancers (4–6), and therefore, it is necessary to measure three of them, especially when we do not know which type of FlOPs is the best predictor for hip fracture. As compared to malondialdehyde (MDA) which is a specific marker for lipid oxidation, FlOPs are generated from many different pathways (lipid, protein and DNA) and reflect a global oxidation burden. FlOP assay is 10–100 times more sensitive than measurement of MDA via colorimetric thiobarbituric acid assay (13). In large human epidemiologic studies, we found that FLOPs are associated with several oxidative indicators (i.e. smoking, hypertension and reduced renal function) (4,5,12,14). Further, most ROSs have a very short half-life in cells (15), but we have found that the levels of FlOP_360, FlOP_320, FlOP_400 are stable in blood samples with delayed processing up to 48 hours at 4°C, stable for more than 10 years in plasma samples in liquid-nitrogen freezers, and highly reproducible over 1–2 year among the same individuals (4,14,16). The well-established lipid oxidation markers (e.g. F2-isoprostanes and MDA) are not stable in blood samples that are not processed over 24 hours (16). Thus, Measurement of these FlOPs in plasma is a non-invasive and convenient approach to assess circulating oxidative stress in large epidemiologic studies.
We therefore hypothesized that elevated oxidative stress increases fracture risk in humans. The present prospective study sought to test the hypothesis by investigating the association between plasma levels of FlOP_360, FlOP_320, FlOP_400 and risk of hip fracture in a prospective study of postmenopausal women in the Nurse’s Health Study.
Materials and Methods
Study setting and participants
The Nurses’ Health Study (NHS) recruited 121,700 women ages 30 to 55 years in 1976 who responded to a mailed questionnaire on which they provided a medical history, personal characteristics and information on lifestyle factors (17). Follow-up questionnaires have been mailed every two years. In the NHS, blood samples were collected from 32,826 women through overnight mail in 1989–1990. The data for the present prospective study were derived from a selected group of women with FlOP measurements (n = 1766) who were previously sampled to prospectively examine the risk of breast cancer and coronary heart disease (4,6). The inclusion and exclusion criteria for identification of cases and controls in breast cancer (6) and coronary heart disease studies (4) were described previously. Briefly, in the breast cancer study (6), nearly all breast cancer cases (99%) between 2000 and 2006 were identified, and one control was selected randomly after matching for age, time and month of blood draw, fasting status, menopausal status and postmenopausal hormone use at the time of case diagnosis. In the coronary heart disease study (4), all coronary heart disease cases between 1989 and 2004 were identified, and two controls were selected randomly after matching for age, smoking and fasting status at the time of each case diagnosis. We used all original data of FlOP measurements from both studies in which we did not excluded any participants. All participants were free of cardiovascular diseases and cancers at the time of blood draw. We excluded the women in younger age (< 60 years) at baseline (n = 770) because most osteoporotic hip fractures (75%) occurred in women ≥ 60 years. Thus, 996 generally healthy postmenopausal women were included in our final analyses. This investigation was approved by the Institutional Review Board of Brigham, Women’s hospital and the University of Cincinnati.
Demographic data collection
Using the questionnaire closest to blood draw (1989–1990), we ascertained age, body weight, life style data (e.g. alcohol intake, physical activity, calcium intake, vitamin D, carotenoid intake and tobacco smoking), history of osteoporosis, history of hypertension and use of postmenopausal hormone therapy (HT). Body height was collected in 1976 questionnaire. Physical activity expressed as metabolic equivalent hours (MET-hours) was estimated from multiple individual activities (i.e., sitting, standing and walking) collected via 1988 questionnaire (18). All the other covariates were collected via 1990 questionnaire. Current and past smokers were defined as current and past tobacco use, respectively. Total calcium and vitamin D were intakes were calculated from food plus supplements (19,20). Total carotenoid intake was the sum of α-carotene, β-carotene, β-cryptoxanthin, lycopene, and lutein/zeaxanthin (6). History of physician-diagnosed osteoporosis and hypertension were reported by the participants.
Blood collection
Blood samples were collected in heparin anticoagulant tubes (4,21). Then the tubes were placed on ice packs, stored in Styrofoam containers and returned to central laboratory by overnight courier. The blood samples were centrifuged, and plasma, packed erythrocytes and buffy coats were divided into aliquots for storage in liquid-nitrogen freezers (−130°C or colder) within 36 hours. Time since last meal was collected by questionnaire during blood collection.
Measurement of FlOPs
Assay procedure
Measurement of FlOPs was performed with previously described procedures (14). Briefly, plasma was extracted with ethanol/ether (3:1, v/v) and centrifuged at 3,000 rpm for 10 min at 4°C, after which supernatant was moved to cuvettes for fluorescence measurement. Fluorescence was determined with a fluorescent spectrophotometer. The excitation/emission wavelengths were 360/420 nm for FlOP_360, 320/420 nm for FlOP_320 and 400/475 nm for FlOP_400. The level of fluorescence was expressed as relative fluorescent intensity units per milliliter of plasma. FlOP_360 represents oxidation products that are generated from oxidized phospholipids or from lipid oxidation products reacting with proteins, DNA and carbohydrates in presence of phospholipids. FlOP_320 is formed when oxidation products such as lipid hydroperoxides, aldehydes, and ketones react with DNA in the presence of metals. Finally, FlOP_400 reflects the interaction between MDA, proteins and phospholipids (22). Because coefficient of variations (CVs) of FlOP measurements across batch of some samples were relatively high (18 to 20%) (6), FlOPs were adjusted for batch-to-batch variation using the method described by Rosner et al.(23). After recalibration, the within-run average CVs for FlOP measurements were < 13%.
Assay stability in blood samples with delayed processing
The delay in processing blood samples up to 36 hours had minimal influence on the measurement of FlOPs. The overall intraclass correlation coefficients (ICCs) of FlOPs were all greater than 0.95 in the shorter- (0 to 24 hours) and longer-delayed processing (0 to 36 hours) (16).
Assay between- and within- person reproducibility
We conducted a pilot study in 40 NHS participants who donated two blood samples to assess the ICC of the between- and within- person variations of the FlOPs at an average of 1.4 years (range: 0.8–2.2 years) interval. After adjusting for fasting status, the ICC for repeated measurements over 1.4 year apart was 0.44 for FlOP_360, 0.55 for FlOP_320 and 0.70 for FlOP_400 (4). ICCs for FlOP_360, FlOP_320 and FlOP_400 over 10 years ranged from 0.14 to 0.30 (6). After classifying FlOP_360, FlOP_320 and FlOP_400 into tertiles, we found that 15%–27% of the 40 study participants changed FlOP tertiles over 1.4 years follow-up, and 53–57% of the 729 study participants changed FlOP tertiles over 10 years follow-up.
Fracture ascertainment
The circumstances and date of hip fracture were ascertained via the biennial questionnaires between 1989 and 2012. We included only the hip fractures incident after blood draw that were due to low or moderate trauma (e.g. tripping, slipping, falling from the height of a chair). Incidence of hip fracture was defined if it occurred after the time of blood draw, and was not due to high trauma events such as motor vehicle accidents, skiing and horseback riding. The validity of self-report hip fracture has been conducted in a previous study, in which all reported hip fractures (n=30) were confirmed via medical records (24). Prior fracture (a covariate) was defined as fractures ascertained before blood draw.
Statistical analysis
Baseline characteristics of women with and without hip fracture were compared by means (standard deviation, SD) or medians (inter-quartile range, IQR) for continuous variable, and percentages for categorical variable. We further examined the distribution of baseline characteristics according to the tertiles of FlOP_320, FlOP_360 and FlOP_400 in overall and fasting samples (≥ 8 hours).
The main statistical model was Cox proportional hazards regression model, in which follow-up time was the time to hip fracture, death or last questionnaire follow-up from time of blood draw (median = 20.1 years; range = 0.1–22.8 years). We analyzed the association of FlOP_360, FlOP_320 and FlOP_400 separately with risk of hip fracture in the multivariate model. Covariates were defined if they are risk factors for fracture or may confound the association between oxidative stress and fracture. The covariates included age (continuous), history of hypertension (yes/no), body mass index (BMI; < 25, ≥ 25 and < 30, ≥ 30 and <35, and ≥ 35 kg/m2), fasting status (< 8 and ≥ 8 hours), time and month of blood draw, alcohol intake (In quartiles: 0, > 0 and <1, ≥ 1 and < 6.7, and ≥ 6.7 g/day), physical activity (In quartiles: < 5.2, ≥ 5.2 and < 12.7, ≥ 12.7 and < 25.9, and ≥ 25.9 MET-hours/week), total calcium intake (In quartiles: < 667, ≥ 667 and < 969, ≥ 969 and < 1347, and ≥ 1347 mg/day), total vitamin D intake (In quartiles: < 187, ≥ 187 and < 303, ≥ 303 and < 536, and ≥ 536 IU/day), history of osteoporosis (yes/no) and postmenopausal HT (yes/no). We tested for linear trend using the medians of FlOPs in each quartile as a continuous variable. We checked the proportional hazards assumption using Cox proportional hazards regression model and found that FlOP-hip fracture relationship did not interact with the time of follow-up. The association between FlOPs and hip fracture was also conducted in fasting samples (≥ 8 hours). When we found a significant association between FlOPs and hip fracture, we performed an interaction analysis between FlOPs and possible confounding factors (e.g. alcohol intake [< 1 and ≥ 1 g/day] and history of hypertension [yes/no]). To test the effect modification of total carotenoid intake, we analyzed the association between FlOPs (High versus low using median as cut-point) and hip fracture in the women with low and high total carotenoid intake (Using median as cut-point) separately. We derived data from two previous prospective studies which are nested case-control data of breast cancers (6) and coronary heart diseases (4). In the primary analysis, we did not exclude breast cancer and coronary heart disease cases in order to increase sample size and statistical power. To rule out the possibility that hip fracture incidence was due to breast cancers or coronary heart disease, we further analyzed the association between FlOPs and hip fracture in controls only (free of coronary heart disease and breast cancer incidence). All analyses were performed with Statistical Analysis System (Version 9, SAS Institute Inc., Cary, NC).
Results
Baseline characteristics and plasma FlOPs of women with and without hip fracture
Forty four hip fractures (4.4%) were identified during the follow-up (Table 1). Women with hip fracture had higher proportion of hypertension than those without hip fracture. The median level of FlOP_320 was higher in women with hip fracture than in those without hip fracture. However, the median levels of FlOP_360 and FlOP_400 of women with and without hip fracture were not significantly different.
Table 1.
Baseline characteristics and fluorescent oxidation products of postmenopausal women with and without hip fracture (N=996)
| Hip fracture (N=44) | Non-hip fracture (N=952) | P value | |
|---|---|---|---|
| Age (years) | 64.6 (2.6) | 64.4 (2.5) | 0.604 |
| Body mass index (kg/m2) | 25.3 (4.4) | 25.7 (4.6) | 0.578 |
| Alcohol intake (g/day)* | 1.7 (0, 5.4) | 1.0 (0, 6.7) | 0.746 |
| Physical activity (MET-hours/week)* | 12.7 (5.4, 21.5) | 10.4 (4.0, 22.2) | 0.533 |
| Total calcium intake (mg/day)* | 874 (657, 1400) | 976 (668, 1343) | 0.333 |
| Total vitamin D intake (IU/day)* | 471 (207, 641) | 299 (186, 533) | 0.197 |
| Total carotenoid intake (1000 IU/day)* | 10.0 (6.5, 12.8) | 9.0 (6.2, 12.8) | 0.518 |
| Fasting status (≥ 8 hours; n, %) | 32 (73%) | 718 (75%) | 0.616 |
| History of osteoporosis (n, %) | 3 (6.8%) | 49 (5.2%) | 0.626 |
| History of hypertension (n, %) | 17 (39%) | 239 (25%) | 0.045 |
| Smokers (n, %) | |||
| Current smokers | 6 (14%) | 167 (18%) | 0.504 |
| Past smokers | 16 (36%) | 379 (40%) | 0.648 |
| Hormone therapy (n, %)£ | 16 (36%) | 360 (38%) | 0.846 |
| Prior fracture (n, %) | 1 (2.3%) | 33 (3.5%) | 0.670 |
| Oxidative stress biomarkers (FI/ml) | |||
| FlOP_360* | 206 (187, 271) | 216 (177, 286) | 0.123 |
| FlOP_320* | 491 (351, 1307) | 402 (300, 664) | 0.031 |
| FlOP_400* | 56.8 (48.7, 79.3) | 61.1 (49.4, 81.1) | 0.355 |
Variables with normal distribution are shown in mean (standard deviation), unless otherwise specified.
Variables with skew distribution are shown in median (inter-quartile range).
Abbreviations: FlOP = Fluorescent oxidation products, FI = Fluorescent intensity units, METs = Metabolic equivalents.
Postmenopausal hormone therapy.
Baseline characteristics according to plasma FlOP levels
At baseline, higher levels of plasma FlOP_360, FlOP_320 and FlOP_400 were associated with greater proportion of current smokers (Table 2). Higher levels of plasma FlOP_360 and FlOP_320 were associated with lower proportion of women who had fasted ≥ 8 hours before blood draw. Higher levels of plasma FlOP_360 and FlOP_400 were associated with greater alcohol intake. Higher levels of FlOP_360 were correlated with lower BMI, greater proportion of history of hypertension. Higher levels of FlOP_320 were positively correlated with total vitamin D intake. Higher levels of FlOP_400 were associated with lower total calcium intake, total vitamin D intake and greater proportion of history of hypertension. Repeated analysis in fasting samples (≥ 8 hours) showed similar results as mentioned above.
Table 2.
Baseline characteristics according to the tertiles of fluorescent oxidation products
| Tertile 1 | Tertile 2 | Tertile 3 | P for trend | |
|---|---|---|---|---|
| FlOP_360 | ||||
| Range (FI/ml) | < 192 | ≥ 192; < 255 | ≥ 255 | --- |
| N | 332 | 331 | 333 | --- |
| Age (years) | 64.3 (2.5) | 64.3 (2.4) | 64.6 (2.5) | 0.189 |
| Body mass index (kg/m2) | 26.6 (4.9) | 25.4 (4.4) | 25.0 (4.5) | < 0.001 |
| Alcohol intake (g/day)* | 0 (0, 3.5) | 1.5 (0, 8.3) | 1.5 (0, 8.3) | < 0.001 |
| Physical activity (MET-hours/week)* | 10.2 (3.9, 20.2) | 10.2 (4.0, 23.5) | 11.8 (4.0, 24.6) | 0.054 |
| Total calcium intake (mg/day)* | 975 (675, 1327) | 963 (679, 1338) | 997 (650, 1411) | 0.207 |
| Total vitamin D intake (IU/day)* | 286 (184, 514) | 303 (197, 549) | 324 (170, 540) | 0.203 |
| Total carotenoid intake (1000 IU/day)* | 8.5 (6.1, 12.1) | 9.5 (6.6, 13.4) | 9.0 (6.1, 12.8) | 0.632 |
| Fasting status (≥ 8 hours; n, %) | 265 (80%) | 258 (78%) | 227 (68%) | 0.020 |
| History of osteoporosis (n, %) | 18 (5.4%) | 17 (5.1%) | 17 (5.1%) | 0.874 |
| History of hypertension (n, %) | 77 (23%) | 72 (22%) | 107 (32%) | 0.002 |
| Current smokers (n, %) | 24 (7%) | 54 (16%) | 95 (29%) | < 0.001 |
| Past smokers (n, %) | 120 (36%) | 146 (44%) | 129 (39%) | 0.840 |
| Hormone therapy (n, %)£ | 112 (34%) | 130 (39%) | 134 (40%) | 0.128 |
| Prior fracture (n, %) | 8 (2.4%) | 15 (4.5%) | 11 (3.3%) | 0.761 |
|
| ||||
| FlOP_320 | ||||
| Range (FI/ml) | < 335 | ≥ 335; < 546 | ≥ 546 | --- |
| N | 331 | 331 | 334 | --- |
| Age (years) | 64.6 (2.5) | 64.1 (2.5) | 64.5 (2.5) | 0.006 |
| Body mass index (kg/m2) | 25.9 (4.4) | 25.6 (4.7) | 25.4 (4.7) | 0.396 |
| Alcohol intake (g/day)* | 0.9 (0, 5.5) | 0.9 (0, 6.7) | 1.1 (0, 7.6) | 0.208 |
| Physical activity (MET-hours/week)* | 10.3 (5.2, 20.7) | 11.6 (3.5, 24.0) | 10.4 (3.9, 22.4) | 0.164 |
| Total calcium intake (mg/day)* | 951 (679, 1314) | 930 (664, 1317) | 1021 (669, 1415) | 0.983 |
| Total vitamin D intake (IU/day)* | 271 (179, 475) | 302 (192, 533) | 348 (193, 590) | 0.034 |
| Total carotenoid intake (1000 IU/day)* | 8.5 (6.1, 12.4) | 9.3 (6.6, 13.2) | 9.2 (6.5, 12.8) | 0.626 |
| Fasting status (≥ 8 hours; n, %) | 284 (86%) | 242 (73%) | 224 (67%) | 0.002 |
| History of osteoporosis (n, %) | 14 (4.2%) | 18 (5.4%) | 20 (6.0%) | 0.469 |
| History of hypertension (n, %) | 72 (22%) | 78 (24%) | 106 (32%) | 0.578 |
| Current smokers (n, %) | 28 (8%) | 68 (21%) | 77 (23%) | < 0.001 |
| Past smokers (n, %) | 139 (42%) | 133 (40%) | 123 (37%) | 0.636 |
| Hormone therapy (n, %)£ | 123 (37%) | 125 (38%) | 128 (38%) | 0.873 |
| Prior fracture (n, %) | 8 (2.4%) | 9 (2.7%) | 17 (5.1%) | 0.806 |
|
| ||||
| FlOP_400 | ||||
| Range (FI/ml) | < 52.9 | ≥52.9; < 72.9 | ≥ 72.9 | --- |
| N | 333 | 332 | 331 | --- |
| Age (years) | 64.5 (2.5) | 64.2 (2.5) | 64.4 (2.5) | 0.620 |
| Body mass index (kg/m2) | 26.2 (5.0) | 25.6 (4.4) | 25.2 (4.5) | 0.062 |
| Alcohol intake (g/day)* | 0.9 (0, 3.5) | 1.1 (0, 6.7) | 1.5 (0, 11.0) | < 0.001 |
| Physical activity (MET-hours/week)* | 10.7 (4.0, 20.4) | 10.2 (3.9, 21.6) | 11.7 (4.0, 25.0) | 0.263 |
| Total calcium intake (mg/day)* | 1045 (691, 1353) | 954 (665, 1357) | 924 (641, 1313) | 0.046 |
| Total vitamin D intake (IU/day)* | 313 (203, 535) | 303 (182, 548) | 293 (181, 529) | 0.015 |
| Total carotenoid intake (1000 IU/day)* | 9.5 (6.6, 14.1) | 9.0 (6.1, 12.8) | 8.5 (6.1, 11.9) | 0.842 |
| Fasting status (≥ 8 hours; n, %) | 275 (83%) | 238 (72%) | 237 (72%) | 0.155 |
| History osteoporosis (n, %) | 16 (4.8%) | 16 (4.8%) | 20 (6.0%) | 0.152 |
| History of hypertension (n, %) | 78 (23%) | 73 (22%) | 105 (32%) | 0.001 |
| Current smokers (n, %) | 10 (3%) | 44 (13%) | 119 (36%) | < 0.001 |
| Past smokers (n, %) | 129 (39%) | 146 (44%) | 120 (36%) | 0.210 |
| Hormone therapy (n, %)£ | 134 (40%) | 120 (36%) | 122 (37%) | 0.739 |
| Prior fracture (n, %) | 16 (4.8%) | 11 (3.3%) | 7 (2.1%) | 0.980 |
Abbreviations: FlOP = Fluorescent oxidation products, FI = Fluorescent intensity units, METs = Metabolic equivalent hours.
Postmenopausal hormone therapy.
Variables with normal distribution are shown in mean (standard deviation), unless otherwise specified.
Variables with skew distribution are shown in median (inter-quartile range).
Association between plasma FlOPs and hip fracture
In the multivariable model, the hazard ratios (HR) of hip fracture in the second and third tertiles of FlOP_320 were 2.11 (95% confidence interval [CI] = 0.88–5.10) and 2.67 (95% CI = 1.14–6.27), respectively, in comparison with the lowest tertile, and the risk increased linearly with increasing FlOP_320 (P for trend = 0.021) (Table 3). After stratifying our analysis by time of follow-up (<10 and ≥ 10 years), we found similar risk estimates of hip fracture in both time intervals as the overall association between FlOP_320 and hip fracture (data not shown). We did not find significant interaction of FlOP_320 with alcohol intake (P = 0.126) and history of hypertension (P = 0.854). When we analyzed the association between FlOP_320 and hip fracture in the women with low and high total carotenoid intake, the positive association between FlOP_320 (High versus low using median as cut-point) and hip fracture appeared to be stronger in women with low carotenoid intake (HR = 2.41, 95% CI = 0.98–5.92) than in those with high total carotenoid intake (HR = 1.81, 95% CI = 0.76–4.30). Neither FlOP_360 nor FlOP_400 was significantly associated with risk of hip fracture (Tertile 3 versus tertile 1: HR = 0.70, 95% CI = 0.32–1.54, P for trend = 0.386 for FlOP_360; and HR = 0.88, 95% CI = 0.40–1.96, P for trend = 0.900 for FlOP_400; Table 3).
Table 3.
Association between plasma fluorescent oxidation products and risk of hip fracture
| Tertile 1 | Tertile 2 | Tertile 3 | P for trend | |
|---|---|---|---|---|
| FlOP_360 | ||||
| Range (FI/ml) | < 192 | ≥ 192; < 255 | ≥ 255 | --- |
| Median (FI/ml) | 163 | 215 | 338 | --- |
| Incidence of hip fracture (n, %) | 15 (4.5%) | 17 (5.1%) | 12 (3.6%) | --- |
| Adjusted for age | 1 (ref) | 1.13 (0.56, 2.26) | 0.85 (0.40, 1.81) | 0.638 |
| Adjusted for age and hypertension | 1 (ref) | 1.16 (0.58, 2.32) | 0.79 (0.37, 1.69) | 0.540 |
| Adjusted for multiple risk factors* | 1 (ref) | 1.03 (0.50, 2.14) | 0.70 (0.32, 1.54) | 0.386 |
|
| ||||
| FlOP_320 | ||||
| Range (FI/ml) | < 335 | ≥ 335; < 546 | ≥ 546 | --- |
| Median (FI/ml) | 270 | 404 | 1337 | --- |
| Incidence of hip fracture (n, %) | 9 (2.7%) | 15 (4.5%) | 20 (6.0%) | --- |
| Adjusted for age | 1 (ref) | 1.81 (0.79, 4.14) | 2.51 (1.14, 5.52) | 0.012 |
| Adjusted for age and hypertension | 1 (ref) | 1.78 (0.78, 4.06) | 2.35 (1.07, 5.18) | 0.015 |
| Adjusted for multiple risk factors* | 1 (ref) | 2.11 (0.88, 5.10) | 2.67 (1.14, 6.27) | 0.021 |
|
| ||||
| FlOP_400 | ||||
| Range (FI/ml) | < 52.9 | ≥ 52.9; < 72.9 | ≥ 72.9 | --- |
| Median (FI/ml) | 45.6 | 61.0 | 93.4 | --- |
| Incidence of hip fracture (n, %) | 15 (4.5%) | 17 (5.1%) | 12 (3.6%) | --- |
| Adjusted for age | 1 (ref) | 1.12 (0.56, 2.25) | 0.89 (0.42, 1.91) | 0.973 |
| Adjusted for age and hypertension | 1 (ref) | 1.15 (0.57, 2.30) | 0.84 (0.39, 1.81) | 0.875 |
| Adjusted for multiple risk factors* | 1 (ref) | 1.21 (0.58, 2.52) | 0.88 (0.40, 1.96) | 0.900 |
Values are hazard ratios (95% confidence interval), unless otherwise specified. FlOP = Fluorescent oxidation products, FI = Fluorescent intensity units.
Risk factors included age (continuous), history of hypertension (yes/no), body mass index (< 25, ≥ 25 and < 30, ≥30 and <35, and ≥ 35 kg/m2), fasting status (< 8 and ≥ 8 hours), time and month of blood draw, alcohol intake (In quartiles: 0, > 0 and <1, ≥ 1 and < 6.7, and ≥ 6.7 g/day), physical activity (In quartiles: < 5.2, ≥ 5.2 and < 12.7, ≥ 12.7 and < 25.9, and ≥ 25.9 MET-hours/week), total calcium intake (In quartiles: < 667, ≥ 667 and < 969, ≥ 969 and < 1347, and ≥ 1347 mg/day), total vitamin D intake (In quartiles: < 187, ≥ 187 and < 303, ≥ 303 and < 536, and ≥ 536 IU/day), history of osteoporosis (yes/no) and postmenopausal hormone therapy (yes/no).
When we analyzed the relationship between plasma FlOPs and hip fractures in fasting samples only (≥ 8 hours; n = 597 subjects; n = 31 hip fractures), higher levels of FlOP_320 were associated with increased risk of hip fracture (Tertile 3 versus tertile 1: HR = 3.16, 95% CI = 1.18–8.46, P for trend = 0.018). Neither FlOP_360 nor FlOP_400 was associated with risk of hip fracture in the fasting samples (Tertile 3 versus tertile 1: HR = 0.90, 95% CI = 0.32–2.54, P for trend = 0.849 for FlOP_360 and HR = 1.47, 95% CI = 0.57–3.77, P for trend = 0.523 for FlOP_400).
When we analyzed the relationship between plasma FlOPs and hip fractures in controls only (free of coronary heart disease and breast cancer incidence; n = 628 subjects; n = 31 hip fractures), FlOP_320 remained significantly associated with increased risk of hip fracture (Tertile 3 versus tertile 1: HR = 6.94, 95% CI = 1.88–25.54, P for trend = 0.005). Again, FlOP_360 and FlOP_400 were not associated with risk of hip fracture (Tertile 3 versus tertile 1: HR = 0.85, 95% CI = 0.32–2.28, P for trend = 0.566 for FlOP_360 and HR = 0.99, 95% CI = 0.36–2.77, P for trend = 0.548 for FlOP_400).
Discussion
To our knowledge, this is the first prospective study among postmenopausal women demonstrating that oxidative stress, quantified by FlOP_320, was a significant predictor for hip fracture. Further research is required to determine if FlOP_320 would be useful for prediction of hip fracture in a clinical setting.
Unlike FlOP_320, we did not find FlOP_360 or FlOP_400 to be independent risk factors for hip fracture. This is possibly due to the fact that FlOP_320 reflects oxidation products which are generated in presence of metals, whereas FlOP_360 and FlOP_400 can be generated without metals. With increased bone loss, heavy metals in the blood are commonly released from bones, which will promote the production of FlOP_320 in blood. Indeed, excessive ROSs increase bone loss by inhibiting bone formation and promoting bone resorption (7–9,25), and an increased level of heavy metals has been found to be associated with increased risk of osteoporosis and fractures (26–28). Thus, FlOP_320 is superior to other types of FlOPs in relation to bone loss and fracture as FlOP_320 reflects the coexisting effects of ROSs and heavy metals.
The present study has several strengths. This study has a long follow-up period. In addition, the prospective study design reduces the possibility of the influence from existing hip fracture.
A limitation of this study is that a single measurement of FlOPs may not accurately reflect the average levels of the biomarker over a prolonged period of time, especially after 10 years. However, we have assessed their reproducibility over approximately one-year period, and high ICCs suggest that this marker can be used as a marker for chronic exposure. Even though the ICCs of FlOP measurements across 10 years were relatively low, we still found significant association between plasma FlOP_320 at baseline and hip fracture. Our results suggest that past high FlOP_320 exposure may play significant role in the development of hip fracture occurs long after baseline exposure, even though FlOP_320 may not be consistently high in long term. Future studies with multiple and repeated measurements of FlOPs over several years interval are warranted to confirm our results. The number of hip fractures in women was relatively low, and the positive relationship between FlOP_320 and hip fracture was observed from multiple statistical tests and adjustments, which may cause type I error. The study is also limited because the data on bone markers, bone mineral density and other fractures were not available.
Hip fracture is regarded as the most severe manifestation of osteoporotic fracture, because it is associated with substantial cost, as well as higher risk of disability, comorbidities and mortality than any other fractures (29–33). Current fracture risk assessment is far from optimal by using traditional risk factors such as age and osteoporosis (34). If our findings are confirmed in other studies, adding this marker into existing fracture assessment model may improve the prediction of hip fracture in postmenopausal women.
In summary, we found an independent and positive relationship between FlOP_320 and risk of hip fracture among postmenopausal women. FlOP_360 and FlOP_400 were not the predictors for risk of hip fracture. Since this is the first prospective study examining the relationship of oxidative stress with fracture risk, our findings warrant further investigation and validation. In addition, the underlying mechanisms for FlOP_320 as a risk factor for hip fracture merit additional studies.
Acknowledgments
The current study was funded by Dr. Wu’s American Heart Association grant 0430202N, K07 award (CA138714), start-up funds and P30-ES006096 and by research grants including HL34594, CA49449, CA131218 and CA87969 from the National Institute of Health, Bethesda, MD.
Footnotes
No supplemental data have been included with the submission.
Author’s roles: Study design: SY and TW. Study conduct: SY and TW. Data collection: WW, AHE, DF and TW. Data analysis: SY. Data interpretation: SY and TW. Drafting manuscript: SY and TW. Revising manuscript content: SY, TW, WW, DF and AHE. Approving final version of manuscript: all authors. SY and TW take responsibility for the integrity of the data analysis and data interpretation.
References
- 1.Holvoet P. Oxidized LDL and coronary heart disease. Acta Cardiol. 2004;59(5):479–84. doi: 10.2143/AC.59.5.2005219. [DOI] [PubMed] [Google Scholar]
- 2.Koenig W, Karakas M, Zierer A, Herder C, Baumert J, Meisinger C, Thorand B. Oxidized LDL and the risk of coronary heart disease: results from the MONICA/KORA Augsburg Study. Clin Chem. 2011;57(8):1196–200. doi: 10.1373/clinchem.2011.165134. [DOI] [PubMed] [Google Scholar]
- 3.Southorn PA, Powis G. Free radicals in medicine. II. Involvement in human disease. Mayo Clin Proc. 1988;63(4):390–408. doi: 10.1016/s0025-6196(12)64862-9. [DOI] [PubMed] [Google Scholar]
- 4.Jensen MK, Wang Y, Rimm EB, Townsend MK, Willett W, Wu T. Fluorescent oxidation products and risk of coronary heart disease: a prospective study in women. J Am Heart Assoc. 2013;2(5):e000195. doi: 10.1161/JAHA.113.000195. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wu T, Rifai N, Willett WC, Rimm EB. Plasma fluorescent oxidation products: independent predictors of coronary heart disease in men. Am J Epidemiol. 2007;166(5):544–51. doi: 10.1093/aje/kwm120. [DOI] [PubMed] [Google Scholar]
- 6.Fortner RT, Tworoger SS, Wu T, Eliassen AH. Plasma florescent oxidation products and breast cancer risk: repeated measures in the Nurses’ Health Study. Breast Cancer Res Treat. 2013;141(2):307–16. doi: 10.1007/s10549-013-2673-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bai XC, Lu D, Bai J, Zheng H, Ke ZY, Li XM, Luo SQ. Oxidative stress inhibits osteoblastic differentiation of bone cells by ERK and NF-kappaB. Biochem Biophys Res Commun. 2004;314(1):197–207. doi: 10.1016/j.bbrc.2003.12.073. [DOI] [PubMed] [Google Scholar]
- 8.Mody N, Parhami F, Sarafian TA, Demer LL. Oxidative stress modulates osteoblastic differentiation of vascular and bone cells. Free Radical Biology and Medicine. 2001;31(4):509–519. doi: 10.1016/s0891-5849(01)00610-4. [DOI] [PubMed] [Google Scholar]
- 9.Almeida M, Han L, Martin-Millan M, Plotkin LI, Stewart SA, Roberson PK, Kousteni S, O’Brien CA, Bellido T, Parfitt AM, Weinstein RS, Jilka RL, Manolagas SC. Skeletal involution by age-associated oxidative stress and its acceleration by loss of sex steroids. J Biol Chem. 2007;282(37):27285–97. doi: 10.1074/jbc.M702810200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Sanchez-Rodriguez MA, Ruiz-Ramos M, Correa-Munoz E, Mendoza-Nunez VM. Oxidative stress as a risk factor for osteoporosis in elderly Mexicans as characterized by antioxidant enzymes. Bmc Musculoskeletal Disorders. 2007;8 doi: 10.1186/1471-2474-8-124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Basu S, Michaelsson K, Olofsson H, Johansson S, Melhus H. Association between oxidative stress and bone mineral density. Biochemical and Biophysical Research Communications. 2001;288(1):275–279. doi: 10.1006/bbrc.2001.5747. [DOI] [PubMed] [Google Scholar]
- 12.Rebholz CM, Wu T, Hamm LL, Arora R, Khan IE, Liu Y, Chen CS, Mills KT, Rogers S, Kleinpeter MA, Simon EE, Chen J. The association of plasma fluorescent oxidation products and chronic kidney disease: a case-control study. Am J Nephrol. 2012;36(4):297–304. doi: 10.1159/000342330. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Frankel EN. In: Lipid oxidation. 2. Frankel EN, editor. The Oily Press LTD; Dundee, Scotland: 2005. [Google Scholar]
- 14.Wu TY, Willett WC, Rifai N, Rimm EB. Plasma fluorescent oxidation products as potential markers of oxidative stress for epidemiologic studies. American Journal of Epidemiology. 2007;166(5):552–560. doi: 10.1093/aje/kwm119. [DOI] [PubMed] [Google Scholar]
- 15.Uttara B, Singh AV, Zamboni P, Mahajan RT. Oxidative Stress and Neurodegenerative Diseases: A Review of Upstream and Downstream Antioxidant Therapeutic Options. Current Neuropharmacology. 2009;7(1):65–74. doi: 10.2174/157015909787602823. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wu TY, Rifai N, Roberts LJ, Willett WC, Rimm EB. Stability of measurements of biomarkers of oxidative stress in blood over 36 hours. Cancer Epidemiology Biomarkers & Prevention. 2004;13(8):1399–1402. [PubMed] [Google Scholar]
- 17.Colditz GA, Manson JE, Hankinson SE. The Nurses’ Health Study: 20-year contribution to the understanding of health among women. Journal of Womens Health. 1997;6(1):49–62. doi: 10.1089/jwh.1997.6.49. [DOI] [PubMed] [Google Scholar]
- 18.Khalili H, Ananthakrishnan AN, Konijeti GG, Liao XM, Higuchi LM, Fuchs CS, Spiegelman D, Richter JM, Korzenik JR, Chan AT. Physical activity and risk of inflammatory bowel disease: prospective study from the Nurses’ Health Study cohorts. Bmj-British Medical Journal. 2013:347. doi: 10.1136/bmj.f6633. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Ascherio A, Hennekens C, Willett WC, Sacks F, Rosner B, Manson JA, Witteman J, Stampfer MJ. Prospective study of nutritional factors, blood pressure, and hypertension among US women. Hypertension. 1996;27(5):1065–1072. doi: 10.1161/01.hyp.27.5.1065. [DOI] [PubMed] [Google Scholar]
- 20.Feskanich D, Willett WC, Colditz GA. Calcium, vitamin D, milk consumption, and hip fractures: a prospective study among postmenopausal women. American Journal of Clinical Nutrition. 2003;77(2):504–511. doi: 10.1093/ajcn/77.2.504. [DOI] [PubMed] [Google Scholar]
- 21.Zhang XH, Tworoger SS, Eliassen AH, Hankinson SE. Postmenopausal plasma sex hormone levels and breast cancer risk over 20 years of follow-up. Breast Cancer Research and Treatment. 2013;137(3):883–892. doi: 10.1007/s10549-012-2391-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Farankel EN. Lipid oxidation. The Oily Press; Dundee: 1998. [Google Scholar]
- 23.Rosner B, Cook N, Portman R, Daniels S, Falkner B. Determination of blood pressure percentiles in normal-weight children: Some methodological issues. American Journal of Epidemiology. 2008;167(6):653–666. doi: 10.1093/aje/kwm348. [DOI] [PubMed] [Google Scholar]
- 24.Colditz GA, Martin P, Stampfer MJ, Willett WC, Sampson L, Rosner B, Hennekens CH, Speizer FE. Validation of Questionnaire Information on Risk-Factors and Disease Outcomes in a Prospective Cohort Study of Women. American Journal of Epidemiology. 1986;123(5):894–900. doi: 10.1093/oxfordjournals.aje.a114319. [DOI] [PubMed] [Google Scholar]
- 25.Stanojkovic I, Kotur-Stevuljevic J, Spasic S, Milenkovic B, Vujic T, Stefanovic A, Ivanisevic J. Relationship between bone resorption, oxidative stress and inflammation in severe COPD exacerbation. Clinical Biochemistry. 2013;46(16–17):1678–1682. doi: 10.1016/j.clinbiochem.2013.08.003. [DOI] [PubMed] [Google Scholar]
- 26.Alfven T, Elinder CG, Hellstrom L, Lagarde F, Jarup L. Cadmium exposure and distal forearm fractures. J Bone Miner Res. 2004;19(6):900–5. doi: 10.1359/JBMR.040202. [DOI] [PubMed] [Google Scholar]
- 27.Friberg L, Elinder C, Kjellstrom T, Nordberg GF. Cadmium and health: A toxicological and epidemiological appraisal Volume II: Effects and response. 1985:320. Medium: X. [Google Scholar]
- 28.Khalil N, Cauley JA, Wilson JW, Talbott EO, Morrow L, Hochberg MC, Hillier TA, Muldoon SB, Cummings SR. Relationship of blood lead levels to incident nonspine fractures and falls in older women: The study of osteoporotic fractures. Journal of Bone and Mineral Research. 2008;23(9):1417–1425. doi: 10.1359/JBMR.080404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Randell AG, Nguyen TV, Bhalerao N, Silverman SL, Sambrook PN, Eisman JA. Deterioration in quality of life following hip fracture: a prospective study. Osteoporosis International. 2000;11(5):460–6. doi: 10.1007/s001980070115. [DOI] [PubMed] [Google Scholar]
- 30.Osnes EK, Lofthus CM, Meyer HE, Falch JA, Nordsletten L, Cappelen I, Kristiansen IS. Consequences of hip fracture on activities of daily life and residential needs. Osteoporosis International. 2004;15(7):567–74. doi: 10.1007/s00198-003-1583-0. [DOI] [PubMed] [Google Scholar]
- 31.Haentjens P, Autier P, Barette M, Boonen S Belgian Hip Fracture Study G. The economic cost of hip fractures among elderly women. A one-year, prospective, observational cohort study with matched-pair analysis. Belgian Hip Fracture Study Group. The Journal of Bone and Joint Surgery American. 2001;83-A(4):493–500. [PubMed] [Google Scholar]
- 32.de Luise C, Brimacombe M, Pedersen L, Sorensen HT. Comorbidity and mortality following hip fracture: a population-based cohort study. Aging Clinical and Experimental Research. 2008;20(5):412–8. doi: 10.1007/BF03325146. [DOI] [PubMed] [Google Scholar]
- 33.Center JR, Nguyen TV, Schneider D, Sambrook PN, Eisman JA. Mortality after all major types of osteoporotic fracture in men and women: an observational study. Lancet. 1999;353(9156):878–82. doi: 10.1016/S0140-6736(98)09075-8. [DOI] [PubMed] [Google Scholar]
- 34.Kanis JA. Diagnosis of osteoporosis and assessment of fracture risk. Lancet. 2002;359(9321):1929–1936. doi: 10.1016/S0140-6736(02)08761-5. [DOI] [PubMed] [Google Scholar]