点击显示 收起
1 From the Unilever Health Institute, Unilever Research and Development Vlaardingen, Vlaardingen, Netherlands
2 Supported by Unilever Research and Development Vlaardingen. 3 Address reprint requests to JE Upritchard, Unilever Health Institute, Unilever Research and Development Vlaardingen, PO Box 114, 3130 AC Vlaardingen, Netherlands. E-mail: jane.upritchard{at}unilever.com.
ABSTRACT
Background: High doses of vitamin E have been shown to decrease lipid peroxidation in persons under oxidative stress. At
present, the data are insufficient to predict whether lower doses
offer the same benefit in healthy persons.
Objective: We studied the effect of moderate doses of a combination of vitamin E and carotenoids, incorporated into a food product, on markers of antioxidant status and lipid peroxidation in healthy persons.
Design: One hundred five healthy adults were randomly, evenly assigned in this double-blind, placebo-controlled, parallel, 11-wk intervention study. After a 2-wk stabilization period during which the subjects consumed a commercial unfortified spread, the subjects consumed 25 g/d of spread containing 43 mg -tocopherol equivalents (-TE; 2-3 fold the US dietary reference intake) and 0.45 mg carotenoids (spread A), 111 mg -TE and 1.24 mg carotenoids (spread B), or 1.3 mg RRR--tocopherol without carotenoids (spread C).
Results: In subjects consuming spread A, plasma -tocopherol concentrations increased 31% to 32 µmol/L, with small but significant increases in concentrations of -carotene and lutein. This resulted in LDL with significantly higher total antioxidant capacity (17%) and an increased resistance to oxidation, as determined by lag time (18%). These improvements were dose dependent: larger increases in these variables were observed in subjects consuming spread B. Furthermore, consumption of spread B significantly reduced concentrations of the plasma lipid peroxidation biomarker F2-isoprostane (15%).
Conclusion: The consumption of food products containing moderate amounts of vitamin E and carotenoids can lead to measurable and significant improvements in antioxidant status and biomarkers of oxidative stress in healthy persons.
Key Words: Antioxidant capacity oxidation resistance carotenoids F2-isoprostanes malondialdehyde spread oxidized LDL peroxidation vitamin E
INTRODUCTION
Oxidative stress is thought to play an important underlying
role in the development of several chronic diseases and in
age-related physiologic degeneration (1). In the body, vitamin
E can modulate some of the adverse effects that reactive
oxygen species have on normal physiologic function. High
doses of vitamin E have been shown to increase the resistance
of LDL to oxidation (2-5), enhance antioxidant capacity, and
reduce F2-isoprostane concentrations in persons under increased oxidative stress as a result of disease (6-9). In healthy
persons, however, little is known about the potential benefits of
vitamin E on in vivo markers of lipid oxidation, such as plasma
F2-isoprostanes. In particular, the effect on lipid peroxidation
of lower doses of vitamin E and of vitamin E in combination
with other dietary antioxidants is unknown.
The results of cross-sectional epidemiologic studies and experimental investigations have led to the suggestion that plasma -tocopherol concentrations > 28 µmol/L may be required to reduce cardiovascular disease risk (10). It is likely that average dietary vitamin E intakes of 15-30 mg/d would be required to achieve these plasma concentrations (11). Achieving dietary intakes of vitamin E in this range can be difficult. A recent dietary intervention reported that < 20% of subjects attained the target of 30 mg/d, despite intensive dietary instruction, provision of vitamin E-rich foods, and high motivation of the subjects (12). Furthermore, vegetable oils and other vitamin E-rich foods tend to also be high in energy, which may be incompatible with dietary advice to maintain or reduce energy intake. Therefore, the objective of the present study was to examine the effect of consuming moderate doses of a combination of vitamin E and selected dietary carotenoids, incorporated into a food product consumed daily, on markers of antioxidant status, antioxidant capacity, and biomarkers of lipid peroxidation in healthy persons.
SUBJECTS AND METHODS
Subjects
Participants were recruited from the Rotterdam area of the
Netherlands through newspaper advertisements. Eligible subjects for the study were aged 35-70 y, had a body mass index
(in kg/m2) between 18 and 32, and were assessed to be in good
health through a medical history questionnaire and biochemical
tests. Exclusion criteria were a plasma vitamin E concentration
> 36 µmol/L, smoking, or participation in the month before
the start of the study in medical treatments that could interfere
with the test results, such as a medically prescribed diet or
weight-loss regimen or the consumption of vitamin or mineral
supplements. All participants in the study were habitual spread
users. The protocol was approved by the Unilever Research
and Development Vlaardingen Medical Ethical Committeee,
and all subjects gave their written informed consent before
participation.
One hundred thirty persons volunteered for the study. Twelve persons were not eligible because of plasma vitamin E concentrations > 36 µmol/L (n = 8) or abnormalities in biochemical variables (n = 4). One person failed to attend the screening appointment. From the remaining 117 volunteers, 105 were randomly assigned to 1 of 3 equal intervention groups on the basis of sex and screening plasma vitamin E concentration. Eight subjects withdrew during the study for personal reasons that were unrelated to the intervention regimen.
Experimental design
The study was a randomized, double-blind, placebo-controlled, 11-wk intervention study. After a 2-wk stabilization
period during which the subjects consumed a commercial
spread (containing 1.3 mg RRR--tocopherol without carotenoids), the subjects consumed 25 g/d of spread containing
43 mg -tocopherol equivalents (-TE) and 0.45 mg carotenoids (50% lutein, 25% ß-carotene, 15% lycopene, and 10%
-carotene; spread A), 111 mg -TE and 1.24 mg carotenoids (50% lutein, 25% ß-carotene, 15% lycopene, and 10%
-carotene; spread B), or 1.3 mg RRR--tocopherol without
carotenoids (spread C). The daily amounts of -TE and carotenoids equaled 3-7 times the recommended daily intake for
vitamin E (11) and 5-40% of the usual intake of dietary carotenoids (13). For practical reasons, the study was conducted in 3
cohorts starting 1-2 d apart. Blood samples were collected after
the subjects had fasted overnight at the end of the 2-wk run-in
phase and during the intervention at weeks 5 and 11.
Intervention regimens
The antioxidant composition of the spreads is presented in
Table 1. The 4 spreads were prepared from a single low-fat
(38% fat) base spread. Spreads A and B contained vitamin E
(Roche Products Ltd, Derbyshire, United Kingdom), /ß-carotene (palm carotenoids suspended in vegetable oil; CHR
Hansen A/S, Hørsholm, Denmark), and encapsulated lutein and
lycopene (Lyc-o-Lut beadlets; Hoffmann-La Roche, Basel,
Switzerland). These spreads were the color of commercial
spread, with the encapsulated lutein and lycopene visible as
very small red beads. The spread used during the run-in and for
the control group (spread C) did not contain any carotenoids.
For masking reasons, spread C was colored orange with synthetic colorants (E102/E110). Participants were instructed to
consume 25 g/d of their assigned spread in place of their usual
spread. Compliance was evaluated every 2-3 wk by weighing
the returned sample containers and by having the subjects
complete an interview with a dietitian.
View this table:
TABLE 1. . Measured antioxidant content of the study products1
During the 13-wk study, the participants were asked to
maintain their usual diet but with the following restrictions:
< 15 mg dietary vitamin E/d (excluding the amounts supplied
in the test products) and minimal fluctuations in other antioxidant-rich foods, such as vegetables, tea, and fruit juices.
Vitamin E intake was monitored by using a dietary checklist,
and additional oils were supplied during the study (low-fat
spread, deep-frying products, and salad oil). Selected high-carotenoid foods (eg, foods containing =" BORDER="0"> 10 mg lycopene, 25
mg lutein, or 30 mg ß-carotene per serving) were forbidden
during the study.
Clinical data and laboratory analyses
Plasma and serum were prepared within 60 min of collection
and were immediately divided into aliquots and stored at
-70 °C. LDL was isolated in duplicate from thawed EDTA-treated plasma (stored with 0.6% sucrose) by using an Optima
TLX tabletop ultracentrifuge (162 000 x g, 24 h, 10 °C;
Beckman Instruments Inc, Palo Alto, CA). Antioxidant and
oxidation assays were conducted on LDL immediately after
isolation. All biochemical measurements were carried out in
duplicate. For each individual, the samples from all of the
appointments were measured on the same day.
Antioxidant status of the subjects
Plasma tocopherol and carotenoid concentrations were measured by using HPLC. Extracts were prepared with heptane
from deproteinated EDTA-treated plasma (400 µL) containing
the internal standards -tocopherol acetate (2.6 µmol/L) and
retinyl acetate (2.6 µmol/L). For the -tocopherol determinations, the extract was dissolved in 100 µL dichloromethane:isopropanol (1: 3, by vol), and 20 µL was injected onto a 50 x
4.6 mm Chromolith SpeedROD RP-18e column (Merck,
Darmstadt, Germany) at 20 °C. This was eluted with methanol:water (95:5, by vol) with a flow rate of 3.5 mL/min. -Tocopherol was detected by ultraviolet-visible light at wavelengths of 292 nm (-tocopherol) and 284 nm (-tocopherol
acetate) with a CV of 4.1%.
For the carotenoid determinations, the extract was dissolved in 100 µL eluent solutions A and B (1:1, by vol). Eluent A contained methanol:tertiary butylmethyl ether (TBME):1.5% ammonium acetate in water (830:150:20, by vol), and eluent B contained methanol:TBME:1.0% ammonium acetate in water (80:900:20, by vol). The sample (20 µL) was injected onto a 150 x 4.6 mm YMC C30 column (YMC Inc, Wilmington, DE) at 20 °C and was eluted by a gradient of A and B with a flow rate of 0.8 mL/min. The gradient separation started with A:B of 95:5 (by vol), which was maintained for 9 min; over 24 min, a gradient was run to A:B of 5:95 (by vol), which was maintained for 4 min, and then returned to A:B of 95:5 (by vol) by 44 min, with 5 min between injections. Retinyl acetate was detected by ultraviolet-visible light at wavelengths of 325 and 450 nm for lutein, lycopene, -carotene, and ß-carotene. The interassay CVs were as follows: lutein, 9.3%; lycopene, 12.9%; -carotene, 10.6%; and ß-carotene, 11.4%.
Plasma vitamin C was measured enzymatically in heparin-treated plasma samples (stored with 4.5% metaphosphoric acid). The analysis was conducted according to Vuilleumier et al (14) and was adapted for use on a Packard Multiprobe II HT analyzer (Packard Instrument Company, Merides, CT) with a CV of 3.7%.
Antioxidant capacity
The ferric-reducing abilities of EDTA-treated plasma and of
native LDL were measured according to Benzie and Strain (15)
but adapted for use on a Packard Multiprobe II HT analyzer.
The CVs for the plasma and LDL assays were 3.7% and 2.7%,
respectively.
The susceptibility of native LDL to copper ion-stimulated oxidation was assessed by monitoring conjugated diene formation (16). In brief, LDL (50 µg protein) was added to physiologic phosphate-buffered saline (final volume 1 mL), and oxidation was stimulated with copper chloride (50 µmol/L) at 30 °C with a CV of 10%.
Serum arylesterase activity with the substrate phenylacetate was used to estimate paraoxonase 1 activity (17). Change in absorbance at wavelength 270 nm was monitored for 3 min by using the Spectra-max 190 microplate reader (Molecular Devices, Sunnyvale, CA) at 25 °C with a CV of 3.5%.
Biomarkers of oxidative damage
Plasma concentrations of total F2-isoprostanes were measured by using HPLC electron spray ionization mass spectrometry and stable isotope dilution mass spectrometry (18). First,
an internal standard (deuterated F2-isoprostane, 8-iso-prostaglandin F2-d) was mixed with 1000 µL EDTA-treated plasma and 80
µL butylated hydroxytoluene (10 mmol/L):triphenylphosphine
(1 mmol/L) in ethanol. Esterified isoprostanes were hydrolyzed in
ethanolic potassium hydroxide at 40 °C for 45 min. The sample
was loaded onto a solid-phase extraction column (tC18 Sep-pak
Vac, 3 mL, 200 mg; Waters Chromatography BV, Etten-Leur,
Netherlands) and washed with water followed by heptane and was
subsequentially eluted with ethyl acetate:heptane:methanol (50:40:10, by vol). The residue was dissolved in 60 µL acetonitrile:water (60:40, by vol, with 0.05% acetic acid), filtered by using
micro spin filter tubes (0.2 µm nylon; Alltech Associates Inc,
Deerfield, IL), and centrifuged (13400 x g, 3 min, 25 °C).
The supernatant fluid was analyzed by using HPLC electron spray ionization mass spectrometry, and stable isotope dilution mass spectrometry was used to quantify the 8-isoprostanes in plasma. The sample (30 µL) was injected onto a reversed-phase HPLC C18 Symmetry column (dp 3.5 µm, 150 x 2.0 mm internal diameter; Waters Chromatography BV) at 20 °C attached to a Waters Alliance 2790 liquid chromatographic system. A gradient separation was used starting with water and acetonitrile (60:40, by vol), with a flow rate of 0.2 mL/min. This system was maintained for 4 min. Next, a gradient was run over 1 min to 100% acetonitrile, which was maintained for 8 min before being returned to the original eluent composition. The HPLC system was connected to a Micromass Quatro Ultima Mass Spectrometer (Micromass, Manchester, United Kingdom) with an electrospray interface that was operated in the negative ionization mode. In the multiple reaction mode of the mass spectrometer the dwell time was 400 ms; the pause time was 100 ms. Target ions were selected at a mass-to-charge ratio (m/z) of 353/193 for F2-isoprostane and at m/z 357/197 for the deuterium-labeled internal standard. The mass spectrometer was set as follows: capillary voltage, 3 kV; cone voltage, 70 V; collision energy, 25 eV; source temperature, 120 °C; desolvation temperature, 200 °C; cone gas, 175 L/h (nitrogen); and desolvation gas, 525 L/h (nitrogen). The CV was 8.8%.
Malondialdehyde was measured in EDTA-treated plasma (stored with glutathione and butylated hydroxytoluene) with a CV of 17% (19). Antibody titres specific to oxidatively modified LDL were measured in triplicate in EDTA-treated plasma by use of a commercial enzyme-linked immunosorbent assay (Mercodia AB, Uppsala, Sweden). The intraassay CV was 8.5%.
Other measures
Plasma total cholesterol and triacylglycerols were measured
spectrophotometrically by using enzymatic methods with commercially available test kits (CHOD-PAP; Boehringer Mannheim, Mannheim, Germany). HDL cholesterol was measured
as described by Lopes-Virella et al (20). The Friedewald equation was used to calculate LDL (21).
Statistical methods
Power calculations performed before the start of the study
estimated that 32 persons per group would be required to
observe with statistical significance a potential change in
plasma total F2-isoprostanes of 10%. This number was also
judged to be sufficient to detect changes in plasma concentrations of -tocopherol and carotenoids, the resistance of LDL to
oxidation, and the ferric-reducing ability of plasma and LDL.
The data were analyzed by three-way analysis of variance with
treatment, sex, and cohort as factors. Change from baseline was
calculated and differences in change compared with the control
group were established by using Dunnett's multiple-comparison
test. Treatment effects are presented as differences between the
changes as a result of the intervention with SEMs unless stated
otherwise. Data were analyzed with SAS version 8.2 (SAS
Institute Inc, Cary, NC).
RESULTS
General characteristics of the subjects
The baseline characteristics of the subjects are shown in
Table 2. Age, BMI, and the distribution of men and women in
each group were not significantly different at baseline. All
subjects had blood cholesterol concentrations within the normal range; however, at baseline, those assigned to spread A had
higher plasma total cholesterol (11%) and LDL (15%) concentrations than did those in group C. Body weights and blood
lipid concentrations did not change significantly during the
study.
View this table:
TABLE 2. . Clinical characteristics of the subjects1
General health of the subjects and compliance with the
intervention
A small number of participants reported having influenza,
cold symptoms, or headaches during the study, but these persons were excluded only if they were unable to continue the
intervention regimen or required medication. The incidence of
illness did not differ significantly between the groups, nor were
any of these events considered serious as defined by good
clinical practice guidelines. All volunteers consumed > 90% of
their assigned spread (
Antioxidant status
At baseline, plasma concentrations of vitamin E, vitamin C,
and carotenoids were within the range previously reported for
the Dutch population (13). Plasma -tocopherol concentrations
increased significantly in subjects consuming spreads A (31%)
and B (73%) but did not change significantly in those consuming spread C (2%; Table 3). By 11 wk, all of the subjects
consuming spreads A and B had plasma -tocopherol concentrations > 28 µmol/L. Lipid-standardized plasma -tocopherol
concentrations paralleled the increases observed for plasma
-tocopherol. Subjects assigned to spreads A and B had small
but significant increases in plasma carotenoid concentrations,
reflecting the changes predicted relative to dietary intakes
(Table 3). Plasma -carotene, lutein, and ß-carotene concentrations also increased in subjects consuming the antioxidant-fortified spreads. Plasma lycopene did not change significantly
during the study. Subjects consuming the control spread had
stable carotenoid concentrations during the study. Plasma vitamin C increased 18-25% in all groups during the study. It is
probable that this increase was due to changes in fruit intake
associated with seasonal changes, because the entire increase
occurred in the second half of the intervention and the magnitude was not significantly different between the groups.
View this table:
TABLE 3. . Markers of antioxidant status and lipid peroxidation1
Antioxidant capacity ex vivo
In line with the increases in plasma concentrations of
-tocopherol and some carotenoids, the total antioxidant capacity of plasma tended to increase (NS). The ferric-reducing
ability of LDL increased significantly in subjects consuming
spreads A (11%) and B (24%) compared with a reduction (3%)
in subjects consuming spread C. LDL lag time was significantly increased in subjects consuming spreads A (18%) and B
(18%) compared with no significant change in subjects assigned to spread C (Figure 1).
FIGURE 1.. Mean (± SEM) lag time of isolated LDL oxidized ex vivo
in subjects assigned to spread A (n = 33), B (n = 32), or C (n = 31) at
baseline () and at 5 wk (). Data were analyzed by three-way ANOVA
with sex and cohort as factors. Dunnett's test was used to establish post hoc
differences from group C. *Change in lag time significantly different from
that for group C, P < 0.001.
Markers of oxidative stress in vivo
Plasma total F2-isoprostane concentrations were significantly reduced in subjects assigned to spread B (Figure 2). Total plasma F2-isoprostanes decreased by 7% after 5 wk and
by 15% after 11 wk. In contrast, there were no significant
changes (< 2%) in plasma total F2-isoprostane concentrations
in the subjects assigned to spread A or C. Subjects who
consumed spread B had significantly lower plasma malondialdehyde concentrations at the end of the study than did those in
the control group. However, this difference was due to a small
(< 10%) increase in plasma malondialdehyde concentrations in
the control group rather than to a reduction in the subjects
assigned to spread B. Antibody titres to oxidatively modified
LDL and serum arylesterase activity did not change significantly during the study (Table 3).
FIGURE 2.. Mean (± SEM) plasma concentrations of total F2-isoprostanes in subjects assigned to spread A (; n = 31), B (; n = 33), or C (;
n = 29) from baseline to week 11. Data were analyzed by three-way
ANOVA with sex and cohort as factors. Dunnett's test was used to
establish post hoc differences from group C. *Changes in F2-isoprostanes
in group B after 5 wk (-7%; P < 0.05) and 11 wk (-15%; P < 0.05)
significantly different from the changes in group C. There were no significant differences between groups A and C, and there was no significant
time-by-diet interaction.
DISCUSSION
In the present study, plasma concentrations of -tocopherol
and carotenoids increased significantly in the subjects who
consumed the antioxidant-fortified spreads. The increase in
plasma -tocopherol was dose dependent and in line with
increases previously observed in healthy persons after supplementation with vitamin E (16). Interestingly, despite the relatively modest concentrations of carotenoids in the antioxidant-fortified spreads, significant dose-dependent increases in
-carotene, lutein, and ß-carotene were observed. These increases are unlikely to have been a result of background dietary
changes because concentrations did not increase in those subjects consuming the control spread. The bioavailability of carotenoids dispersed in oils was previously reported to be better
than that from vegetables, and this may account for the high
percentage increases in plasma concentrations (22-25). Plasma
lycopene concentrations did not increase in subjects consuming
the antioxidant-fortified spreads; however, it is likely this
2-4% increase in total dietary intake could not be distinguished
from the background diet. Plasma vitamin C concentrations
increased in all groups during the study, most likely in response
to seasonal variations in dietary intake.
An important finding of the present study was the increase in both plasma -tocopherol and carotenoids after consumption of 25 g antioxidant spreads daily. Increasing plasma -tocopherol and carotenoid concentrations through diet alone was previously reported to be difficult, with substantial increases in vegetable oils, fruit, and vegetables required to achieve relatively modest increases in plasma concentrations of -tocopherol and carotenoids (12, 26, 27). Dietary supplements are effective at increasing plasma concentrations of vitamin E and carotenoids, but compared with food sources, the risk of overconsumption is higher with these compounds (11). Therefore, a low-fat spread fortified with moderate amounts of vitamin E and carotenoids may provide a safe and convenient way to achieve desirable antioxidant status.
Increases in plasma antioxidant concentrations may contribute to enhanced antioxidant defense. This is indicated by the fact that ex vivo measures of LDL antioxidant capacity and resistance to oxidation showed consistent and significant increases after consumption of the antioxidant-fortified spreads. This observation is consistent with published data showing that dosages of 25-200 mg -tocopherol/d increased LDL lag time in a dose-dependent manner in healthy persons (16, 28). LDL is vulnerable to oxidative damage, and lag time has been suggested to provide an ex vivo measure of LDL peroxidation resistance. Paraoxonase is associated with HDL, but it can protect HDL and LDL by deactivating certain oxidized fatty acids (17, 29). Preliminary evidence suggests that paraoxonase 1 activity is modulated by oxidative stress (30) and antioxidant status (31). However, in the present intervention study, we did not observe a significant change in paraoxonase 1 activity after consumption of a combination of vitamin E and carotenoids.
LDL that has undergone oxidative modification is immunogenic, allowing antibody titers specific to malondialdehyde on oxidized LDL to be quantified in the plasma (32). To date, most of the published research that used this method was collected from patients with established CVD (32-34), and the effect of antioxidants on this variable is unknown. In the present study, concentrations of oxidized LDL were low at baseline and did not change significantly during the study, despite changes in other variables of lipid peroxidation. More data are required on the sensitivity and specificity of antibodies to oxidized LDL in healthy persons.
Malondialdehyde and F2-isoprostanes are produced in the body as a consequence of the peroxidation of PUFAs that contain > 2 double bonds, or in the case of F2-isoprostanes, mainly arachidonic acid (1). The validity of the malondialdehyde method has frequently come under scrutiny because malondialdehyde is not exclusively formed via lipid peroxidation (35). Although we observed a reduction in malondialdehyde in subjects assigned to spread B relative to the control group, this was most likely due to differences at baseline.
F2-isoprostanes in plasma and associated urinary metabolites are recognized as valid markers of oxidative stress in vivo and are the preferred biomarker of lipid peroxidation (36). Persons consuming the spread that provided 111 mg -tocopherol and 1.24 mg carotenoids/d had a 15% reduction in plasma total F2-isoprostane concentrations during the 11-wk study. This finding shows that in healthy persons with normal concentrations of basal F2-isoprostanes, improvements in the antioxidant defense system can result in measurable reductions in lipid oxidation.
The present study had several strengths, including a randomized, placebo-controlled design and a sample size sufficient to detect small changes with statistical confidence. Furthermore, the use of HPLC with tandem mass spectrometry to measure plasma F2-isoprostane concentrations allowed the less volatile components to be measured without derivatization, resulting in lower CVs without compromising specificity and sensitivity (37). Our findings agree with the results of several vitamin E intervention studies that showed reductions in F2-isoprostane metabolites in the urine (34-58%) after supplementation with 67-900 mg -tocopherol/d (6-9, 38). However, not all intervention studies have recorded a benefit in F2-isoprostane concentrations with comparable doses of vitamin E or a combination of dietary tocopherols and carotenoids (39-45). The findings of these studies may vary because they were conducted in subjects with differences in health (high cholesterol, diabetes, cystic fibrosis, or prothrombotic disorder) or smoking status or used different forms of vitamin E (2R--tocopherol, RRR--tocopherol, tocopherols, and tocotrienols), different dosages (7-2000 mg/d), or different time periods (5-60 d).
In conclusion, using a controlled, 11-wk intervention study with sufficient statistical power, we showed that the consumption of food products containing moderate amounts of vitamin E and selected dietary carotenoids can lead to measurable and significant improvements in antioxidant status and biomarkers of oxidative stress, such as F2-isoprostane concentrations and resistance of LDL to oxidation ex vivo.
ACKNOWLEDGMENTS
We thank LA Akerboom-Voogd, SY Gielen, M Jäkel, A Porcu, PC
Remmerswaal, VIO Ringelberg-van Eerdenburg, M Slotboom, C van Tuijl,
and RLC Weterings for recruitment and clinical procedures and our colleagues who produced, packed, and analyzed the spreads. We are indebted
to J Gerrits, L van Buren, M van der Ham, and WGL van Nielen for their
careful analytic work and to JNJJ Mathot for the quality control and
coordination of the laboratory analyzes. We acknowledge the expert help
of FHM van de Put in automating several key assays and JJ Schilt and A
van Unnik for data management. Last, we thank the volunteers who
participated in the study.
The study was conceived and designed by SAW, LBMT, JEU, and PJR. CRWCS coordinated the trial and JEU supervised the analytic aspects. SAJC gave significant advice on the methods used in the study. AW performed all statistical testing. JEU wrote the manuscript, and all authors were involved in interpreting the results and in critical revision of the paper. No authors had any advisory board affiliations.
REFERENCES