| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
LETTERS TO THE EDITOR |
Danish Institute for Food and Veterinary Research
Soborg
Denmark
E-mail: lod{at}fdir.dk
Research Department of Human Nutrition
LMC Center for Advanced Food Studies
Royal Veterinary and Agricultural University
Frederiksberg
Denmark
Institute of Biochemistry and Food Chemistry
Technical University of Graz
Graz
Austria
Department of Geriatrics and Clinical Nutrition Research University of Uppsala
Uppsala
Sweden
Division of Human Nutrition and Epidemiology
Department of Food Technology and Nutritional Sciences
Wageningen University
Wageningen
Netherlands
Danish Research Institute for Agricultural Sciences
Foulum
Denmark
Department of Dairy end Food Science
LMC Center for Advanced Food Studies
Royal Veterinary and Agricultural University
Frederiksberg
Denmark
Institute of Public Health
University of Copenhagen
Copenhagen
Denmark
Dear Sir:
We appreciate the comments on our paper (1) made by Serafini et al, who highlight some important problems in the interpretation and power of biomarker-based human intervention studies. Serafini et al’s letter contains 2 major points of criticism. The first concerns the possibility that our intervention period of 25 d was insufficient to observe a change in fasting measures of antioxidant capacity without an added dietary oxidant stress, such as increased fat. Relatively few human intervention studies have actually been able to show differences in antioxidant capacity, and as far as we are aware, all of these found only postprandial effects. This is the case for studies of tea and chocolate, which have been shown to result in short-term increases in markers of antioxidant capacity equivalent to the increased plasma concentration of catechins (2–6). The tomato study mentioned by Serafini et al also came to this conclusion (7). In another study, the intervention of 20–25% changes in fat or total energy intake for 12 wk was insufficient to elicit observable changes in plasma antioxidant capacity (8).
Thus, we can speculate that prolonged dietary changes are necessary to affect antioxidant capacity. For example, the lifestyle factors leading to type 2 diabetes also result in chronic decreases in plasma antioxidant capacity, apparently as the result of changes in uric acid metabolism (9, 10). Whether fruit and vegetables would counteract this effect in the long run remains to be investigated. Therefore, our conclusion that a large intake of fruit and vegetables does not affect fasting plasma measures of antioxidant capacity seems valid and in accordance with the literature.
The second criticism concerns our method for measuring plasma antioxidant capacity. According to Serafini et al, an increase in the measurement error may have resulted in our failure to detect minor changes, such as the 10% change calculated from the drop in ascorbate concentrations. Our automated assay of the ferric-reducing ability of plasma (FRAP) and Trolox-equivalent antioxidant capacity (TEAC) was optimized within the boundaries of our equipment, eg, with the absorbance filters available. This decreased sensitivity offset the absolute values of TEAC and increased intra-assay variability compared with the same assays on other equipment. We agree that the intra-day CV of our standard plasma sample was high and understand the concerns of Serafini et al. We have reinvestigated the cause of this and found that other samples and our calibrators had much lower variability, indicating some unidentified problem with our standard plasma. In these other samples, our intra-assay variation was still higher (6.7% for TEAC and 3.9% for FRAP) than the reference values cited in the literature (11, 12). However, this is unlikely to have caused a type I error in our study because the interindividual variability in FRAP and TEAC was still much higher than the assay variability. The measurement error therefore has relatively little influence on the actual power to detect differences. We observed an overall interindividual CV at baseline of 11.2% for TEAC (
± SD: 885 ± 99 µmol/L) and 22.5% for FRAP ( ± SD: 693 ± 156 µmol/L) in the fasting samples (n = 43). In the postprandial samples, the variation was 17.0% and 26.7% (n = 28), respectively. In papers by others, including those cited by Serafini et al, the interindividual CVs for plasma antioxidant activities are variable but similar to ours, eg, 20.6% for FRAP [n = 141 (11)], 21.7% for total radical-trapping antioxidant potential [n = 11 (7)], 9.6% for TEAC [n = 312 (12)], and 18.3% for oxygen radical absorbance capacity [n = 60 (13)].
In our study (1), we tried to increase power by looking at the time course during the intervention with a repeated-samples analysis of covariance (ANCOVA) that used each volunteer’s value at baseline as a covariate. In this analysis, the analytic error becomes more important for the power because the interindividual differences are balanced out. However, it still depends on the intraindividual (interday) variation, which in our study was 9.3% for FRAP and 11% for TEAC. This leads to a power of 70% to detect a significant 10% change in TEAC or FRAP [determined by G-power (14) as a post hoc analysis]. In addition to the values at baseline and at the end of intervention (25 d), we measured plasma antioxidant capacity 3, 9, and 16 d after the start of the intervention and 8 and 29 d after the volunteers had resumed their habitual diet. As seen in Figure 1
, there is no indication of deviations from the initial or post-intervention values, as we also confirmed by repeated-samples ANCOVA. In the case of FRAP, the known difference of 25% between men and women (11) was readily observable at all time points, which indicates to us that we would have seen some indication of a 10% change in fasting blood samples. Moreover, the groups with higher initial values were stable throughout.
|
ACKNOWLEDGMENTS
None of the authors had any conflict of interest related to the results and discussion published in this letter.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
); a corresponding supplement containing nutrients, vitamins, and minerals (
); or a placebo pill plus an energy-balancing drink (
). The start and end of the intervention are marked with vertical arrows. None of the groups differed significantly at any time point by repeated-samples analysis of covariance.