| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
Division of Gastroenterology and Nutrition, Department of Pediatrics, University of Zurich, Switzerland.
2
Medical Statistics Department, University of Nijmegen,
The Netherlands.
a Address for correspondence: Institute of Biochemistry, Karl-Franzens University of Graz, Schubertstr. 1, A-8010 Graz, Austria. Fax 43-316-380-9845.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
- and
-tocopherol, - and ß-carotene (cis and trans isomers),
lycopene, and retinol were determined by reversed-phase HPLC, and
ratios of plasma -tocopherol to cholesterol were calculated in 208
Swiss individuals ages 0.4–38.7 years. The influence of age, sex, and
season of sampling was studied. Age was a significant predictor of all
plasma concentrations except -carotene. No sex-related differences
were observed. Season of sampling affected -tocopherol and retinol
(higher in winter) and -tocopherol and cholesterol concentrations
(higher in winter and spring than in the other seasons). After
correction for seasonal influences, age differences were 0.24 µmol/L
per year for -tocopherol, 0.04 µmol/L per year for retinol, and
0.04 µmol/L per year for cholesterol concentrations; ratios of plasma
-tocopherol to cholesterol were not affected by age. We constructed
age-specific reference intervals from the regression line and a
multiple of the standard deviation. Separate regression equations are
presented for seasons with low and high values.
Key Words: indexing terms: tocopherol • lycopene • retinol • carotene • cholesterol • age-related effects • pediatric chemistry • variation, source of • epidemiology
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-tocopherol
concentrations and mortality from ischemic heart disease are inversely
related (1)(2)(3), as are serum carotenoid concentrations and
the incidence of several types of cancer (4). Additional
significance regarding several aspects of immune function has also been
recognized (5)(6). Antioxidant status evidently varies considerably among populations (1). Moreover, changes over time in dietary habits affect not only the intake of fruits and vegetables but also the consumption of polyunsaturated fatty acids, all of which influence the oxidant–antioxidant balance.
Impaired antioxidant status has been identified in several disorders, such as cholestatic liver disease (7), exocrine pancreatic insufficiency (8), nutritional deprivation due to protein calorie malnutrition (9), and acquired immunodeficiency syndrome (AIDS) (10). Intervention studies for correction of these deficiencies have been conducted or are still in progress (11)(12)(13). In these situations, definition of threshold values for initiation of intervention and successful treatment is critical. Preferably, reference values obtained from the general healthy population living in the same area should be used to define this threshold.
Optimal nutrition in childhood and adolescence is becoming increasingly
a general concern because development of certain diseases may be
determined very early in life. Autopsy studies, for instance, confirm
the presence of atherosclerotic lesions in the aortas and coronary
arteries of teenagers and young adults (14). In general,
vitamin status is believed to differ between children and adults, but
data comparing concentrations of different antioxidant vitamins and
micronutrients between childhood and adulthood in the same population
are limited to French (15) and Japanese individuals
(16). Additional data have focused only on -tocopherol
(17)(18) and retinol concentrations
(19) in North American (17) and Hispanic
individuals (18)(19).
The purpose of this study was to determine cross-sectional age trends and reference intervals for plasma concentrations of tocopherols, carotenoids, and retinol in a Swiss pediatric and adult population living in the Zurich area.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
analytical methods
Blood was drawn into EDTA (1.6 g/L)-containing tubes (Sarstedt
Monovette; Nürmbrecht, Germany) from fasted individuals,
protected from light with aluminum foil, and centrifuged without delay
at 2000g for 8 min at room temperature. Plasma was separated
and kept at -20 °C for no more than 4 days before analysis. Plasma
- and -tocopherol, - and ß-carotene, lycopene, and retinol
concentrations were determined by reversed-phase HPLC according to the
method of Hess et al. (20) in the laboratory of the
Vitamin Research Department of F. Hoffmann-La Roche, Basel,
Switzerland. Within-day reproducibility (CV) was 2.3% for
-tocopherol, 2.4% for -tocopherol, 9.4% for -carotene, 4.3%
for ß-carotene, 2.1% for lycopene, and 2.0% for retinol as
described (20). Plasma cholesterol concentrations were
determined in the same laboratory with a COBAS-BIO instrument (F.
Hoffmann-La Roche) and a diagnostic kit (Merckotest 14350; Merck,
Darmstadt, Germany) for which the CV was 4%. Plasma
-tocopherol:cholesterol ratios were calculated as the index of
choice for vitamin E status (21).
statistical analysis
To obtain approximately gaussian distributions, we had to
log-transform the measurement data for plasma - tocopherol,
ß-carotene, and lycopene concentrations and
-tocopherol:cholesterol ratios. Analysis of covariance was used to
study the effects of age, age2, sex, and season of sampling
on plasma concentrations of the antioxidant vitamins and
micronutrients, assuming an absence of interaction. The influence of
age2 was considered so we could study deviations from
linearity with age. For all variables that were dependent on age, the
relationship with age was best fitted by a linear regression line. The
data obtained in this study were not sufficient to determine whether
the standard deviation (SD) would change with age; therefore, we
assumed a constant SD over the whole age range. Because no sex-related
differences were evident, we combined the data for males and females.
The raw data for variables that showed significant seasonal influences
were corrected accordingly. For those variables for which age was an
influential factor, we constructed age-specific reference intervals
from the regression lines plus or minus 1.645 SD to obtain the central
90% prediction limits (corresponding to the 5th through 95th
percentiles). For analytes that did not show an age-dependency, the 5th
and 95th percentiles were calculated for the results over the whole age
range. To construct the age-specific reference intervals, we used the
methodology described recently by Royston (22) as far as
was applicable. Partial correlations were used to study the effects of
age and cholesterol concentrations on the different variables adjusted
for seasonal influences. SPSS-X release 4.1 (SPSS, Chicago, IL) and
Statgraphics Plus for Windows (Manugistics, Rockville, MD) were used
for statistical procedures. P <0.05 was considered
significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-carotene,
but no effect of sex. Therefore, age-specific but not sex-specific
reference intervals were calculated.
Influences of season of sampling.
Plasma -tocopherol
and retinol concentrations were highest in winter, and
-tocopherol and cholesterol concentrations were higher in winter
and spring than in the other seasons; the other variables did not show
significant seasonal influences. The data points presented in Fig. 1
are corrected for these seasonal influences and represent the
values from the seasons with the lower values. The plasma
-tocopherol:cholesterol ratios were higher in summer and in winter
than in the other seasons; however, the numerator and the denominator
of this ratio each has its own pattern of seasonal changes, making any
differences in the combined data difficult to interpret. We therefore
decided not to correct for seasonal influences the raw data for the
ratio.
|
Influences of age.
Plasma concentrations of all
antioxidant vitamins and micronutrients except -carotene showed a
linear increase with age. The strongest influence of age was observed
on retinol concentrations (r = 0.70, P
<0.0001), followed by -tocopherol (r = 0.50,
P <0.0001) and cholesterol concentrations
(r = 0.43, P <0.0001) (Table 1
, Fig. 1
). The estimated age differences were 0.24 µmol/L per
year for -tocopherol, 0.037 µmol/L per year for retinol, and 0.035
mmol/L per year for cholesterol concentrations. In contrast to its two
components, the plasma -tocopherol:cholesterol ratio did not change
throughout the whole age range. Plasma -tocopherol, ß-carotene,
and lycopene concentrations increased by 1.6%, 1.3%, and 1.6% per
year, respectively.
|
reference intervals
Reference intervals for the two variables that did not show
significant age differences (plasma -carotene and
-tocopherol:cholesterol ratio) were calculated for the whole age
range (Table 2
). For variables with significant age differences, the mean and
90% prediction limits (5th, 50th, and 95th percentiles) by age are
shown in Fig. 1
. The data points plotted are either raw data
(ß-carotene and lycopene) or data corrected for seasonal influences
(- and -tocopherol, retinol, and cholesterol). As stated earlier,
the data presented are those for the seasons with low values;
correction factors for the other seasons are presented in the figure
legend. The percentages of values below and above the reference
intervals are shown in Table 1
. The percentage of values within the
90% prediction limits for the various analytes ranged from 89.1% to
91.9%, or well close to the nominal 90%. Furthermore, Fig. 1
does not indicate any clustering of (extreme) values. In general,
extreme values were seen in younger ages, probably because the number
of observations were greater for younger subjects.
|
Table 1
provides the information necessary for calculating the
reference intervals for these analytes in subjects of a given age from
the regression equation y = intercept + slope x
age and the residual SDs. Where indicated, separate intercepts for
seasons with low and high values are presented. These data can further
be used to calculate z-scores for comparing individual
measurement data obtained in clinical and research settings with the
reference values presented here. An example for the calculations is
shown in the Appendix.
correlation with cholesterol
As a constitutional component of the lipoproteins, cholesterol
exerts a carrier function for the lipophilic antioxidants in blood. We
therefore wondered whether the increase in cholesterol concentrations
might have determined the increase of these analytes with age. Partial
correlations between age, cholesterol, and plasma concentrations of the
various antioxidants and vitamins (corrected for seasonal influences)
are compared with Pearson correlations in Table 3
, which shows that all analytes except -carotene are
correlated with cholesterol concentrations. Correlations with
cholesterol were stronger than the correlations with age for
-tocopherol and ß-carotene but not for retinol.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
- and -tocopherol, retinol, and cholesterol
concentrations are presented for the different seasons (winter vs
nonwinter; winter and spring vs summer and fall). (b)
Age-related effects were documented, with plasma concentrations of
-tocopherol, retinol, and cholesterol increasing markedly from
infancy to adulthood; concentrations of -tocopherol, lycopene, and
ß-carotene also increased with age but to a lesser extent. Because of
its cross-sectional nature, this study does not provide information
about the evolution of these plasma concentrations in individual
subjects. (c) In contrast to other reports
(23)(24), participants of this study did not
take vitamin supplements and all were nonsmokers. (d) The
age-specific reference intervals were constructed from the results of
linear regression analysis. This methodology has the advantage that the
age range under consideration is not broken down into arbitrary age
groups. The researcher who uses the data of this study can easily
calculate z-scores from the information in Table 1
; the
clinician can further quickly compare individual measurement values
with the reference intervals shown in Fig. 1
. Finally, (e)
in focusing on differences among populations, all comparisons were
limited to reports on determinations of these antioxidant vitamins and
micronutrients within the past 10 years; older studies might not be
comparable regarding dietary habits and analytical methods,
particularly because methods for quantitative determinations of
specific carotenoids have only recently become available. As mentioned above, the age-specific reference intervals were constructed from the results of linear regression analysis. This approach followed essentially the method of curve-fitting described recently by Royston (22), which allows the use of statistical software to construct time-specific reference intervals. Depending on the reference values being considered, more or less complex and sophisticated curve-fitting methodologies have been proposed. To ensure that the linear model was the one that best fit these data, we checked several characteristics: e.g., the effect of age2, the gaussian distribution of the sample, and the validity of the calculated 90% reference intervals.
Sex-related differences were not observed for any of the antioxidant
vitamins and micronutrients under investigation—which agrees well with
some other studies (18)(25)(26)
but contrasts with those that reported higher plasma retinol
concentrations in males and higher - and ß-carotene concentrations
in females
(15)(16)(23)(24)(27).
Whether this reflects a lack of differences in dietary habits between
men and women in the Swiss population, in contrast to the case in other
populations, remains to be investigated.
Season of sampling had a significant effect on plasma concentrations of
- and -tocopherol and retinol but not on carotenoids. For all
three, the highest concentrations were found in winter, and
-tocopherol was also high in spring. The effects of season (winter
vs nonwinter) were estimated to be 2.48 µmol/L for -tocopherol
(~16% of mean -tocopherol concentrations) and 0.23 µmol/L
(21%) for retinol; the effects for -tocopherol and cholesterol
(winter and spring vs summer and fall) were 25% of mean values for
-tocopherol and 9% for cholesterol concentrations. Further
interpretation of these seasonal trends is limited by the
cross-sectional nature of this study. Only a longitudinal study,
combined with dietary intake data, will be able to elucidate possible
determinants of these trends. The influence of season on
-tocopherol concentrations found in this study contrasts to Finnish
(28) and Spanish (27) observations of stable
-tocopherol concentrations. That season of sampling did not
influence plasma carotenoid concentrations is in agreement with a North
American study (29) but contrasts with a Spanish
investigation showing highest - and ß-carotene concentrations in
summer (27) and a Finnish study with the highest values in
fall (28).
Cholesterol concentrations increased from childhood to adulthood as shown before (30). In a longitudinal study a decrease in cholesterol concentrations during the early teenage years and a subsequent increase have been reported, with these changes occurring ~2 years earlier in girls than in boys (31). A recent analysis of data from adults (ages 27 to 62) showed an increase not only in the mean values but also in the spread (SD) of cholesterol concentrations, particularly after the age of 40 (22). Except for French children with serum cholesterol concentrations as high as 4.5 mmol/L (26), values in other populations, e.g., Italian (25), English (25), and North American adults (24)(32), were comparable with those found in this study.
Plasma -tocopherol concentrations were strongly correlated with
cholesterol concentrations, confirming the findings of other studies
(11)(17)(21). Partial
correlations, controlling for the age effect, showed correlations
between cholesterol and all variables except -carotene. The
correlations for -tocopherol and ß-carotene were clearly stronger
than their correlations with age, whereas the opposite was true for
retinol concentrations. These results suggest that the increase of
almost all variables with age can at least in part be attributed to the
age-related increase in cholesterol concentrations.
Because of the close correlation between plasma - tocopherol and
lipid concentrations, both plasma -tocopherol:cholesterol and
-tocopherol:total lipids ratios have been proposed as desirable
indexes of vitamin E status (21). Using the former ratio
alleviated the changes with age observed in both numerator and
denominator; the ratio was practically constant over the whole age
range. This finding was unexpected because (a)
supplementation studies have shown that lipoproteins can be loaded
quite extensively with vitamin E, resulting in plasma
-tocopherol:cholesterol ratios that exceeded 10 mmol/mol, with only
minor indications of a plateau effect (11); (b)
a close relation of the intake of -tocopherol and cholesterol cannot
be expected; and (c) the mechanisms for regulation of blood
concentrations of cholesterol are not the same as those for
-tocopherol. As for all other vitamins and micronutrients included
in this study, we did not attempt to investigate any possible
determinants of their plasma concentrations except for correlations
with cholesterol concentrations.
Comparison with other populations was restricted to plasma
-tocopherol concentrations because plasma
-tocopherol:cholesterol ratios were not presented in most
studies. Swedish (33) and Japanese children
(16) had clearly lower plasma -tocopherol
concentrations than did French (15)(26) and
Swiss children of this study. Spanish (27) and Southern
Italian adults (25) showed higher concentrations than our
Swiss adult population and than English (25), Austrian
(34), and Finnish adults (35). Thus, a
European north–south gradient was evident, with the Swiss population
showing intermediate concentrations. Data from North American adults
showed relatively low concentrations, even though 18% took vitamin E
supplements (23). Similar concentrations were found for
nonusers in another study, whereas those for users of vitamin E
supplements were substantially higher (24).
Plasma -tocopherol accounts for only a minor proportion of vitamin E
in the blood, most likely because of preferential incorporation of
-tocopherol into lipoproteins (36). In the Swiss
population we studied, -tocopherol concentrations increased with
age, but to a lesser extent than did -tocopherol. Appropriate data
for comparison of plasma -tocopherol concentrations in other
populations were not available.
Among all vitamins and micronutrients investigated, retinol concentrations showed the strongest dependency on age: a positive slope of 0.04 µmol/L per year. This increase is in agreement with other studies (15)(16)(26). Malvy et al. (26) showed that postpubertal values tend to be slightly higher than prepubertal values and that fluctuations with age are more pronounced in males than in females. They also showed that concentrations of the main carrier for retinol in plasma, retinol-binding protein, increase throughout childhood. In contrast to studies conducted in Finland (28), Spain (27), France (15)(26), and Japan (16), which showed generally higher plasma retinol concentrations in males, no sex-related difference was observed in our study population. Values in the Swiss pediatric age group of this study were lower than those in Swedish (33) children and one group of French children (26), but were comparable with those in another group of French children (15). Those in adults of this study were lower than in Finnish (35) and Japanese individuals (16) but comparable with those in Spanish adults (27). A survey in a North American population showed plasma retinol concentrations of 2.18 µmol/L for men and 1.84 µmol/L for women; roughly a third of these individuals were taking vitamin A supplements (23).
A recent study compared English and Southern Italian adults and found exceptionally high plasma concentrations of ß-carotene in both groups (25). In contrast, Austrian (34), Spanish (27), and North American adults (23), Japanese individuals over the whole age range (16), French children (26) and adults (15), and the Swiss population of this study showed considerably lower values—which raises the question of analytical differences.
A few investigations have shown that all-trans-ß-carotene is predominant in human plasma (37)(38)(39)(40). Cis isomers are present in various fruits and vegetables, and it is not clear whether isomerization reactions, including potential isomer discrimination at the uptake or transport level, might explain why 9- and 13-cis- isomer concentrations are generally very small in plasma compared with trans-ß-carotene but are present in higher concentrations in tissues (37). This study, which comprises the most comprehensive data currently available on total cis and trans isomers of ß-carotene in human plasma, shows a 5% contribution of the cis form to total plasma ß-carotene. Of the studies that differentiated between 9- and 13-cis isomers of ß-carotene, some identified 9-cis (10% of all-trans-ß-carotene) (38)(39) and some 13-cis (5% of total ß-carotene) (37)(40) as the main cis isomer of ß-carotene in plasma.
Another carotenoid, lycopene, exerts only one-twelfth the vitamin A
activity of ß-carotene but is a more efficient quencher of singlet
oxygen than is ß-carotene (41). Comparison with the few
data available showed that the Swiss study population had higher plasma
concentrations of lycopene than did Spanish (27), Japanese
(16) and Austrian individuals (34), whereas
the concentrations in North American adults were even slightly higher
(23). Plasma -carotene concentrations in the Swiss
population were comparable with those in Japanese (16),
about twice as high as in Spanish (27), and three times
those in North American adults (23).
In summary, season-adjusted actual reference intervals for ages 0–40 years were established in a Swiss population living in the German-speaking part of Switzerland. These values may be useful for comparison of plasma concentrations both in healthy subjects and in patients with various disorders who are at risk for developing deficiencies in these vitamins and micronutrients.
Appendix: Comparison of Individual Measurement Data with
the Reference intervals by Using z-Scores; Example for
Log-Transformed Data
To obtain the z-score for a given individual
measurement, e.g., 0.20 µmol/L lycopene in a 9-year-old subject,
(a) calculate the log of the mean value for the reference
population at age 9 years, according to y = intercept +
slope x age (shown in Table 1
): y = -1.00 +
0.016 x 9 = -0.856. (b) Take the log of the
individual measurement value: log 0.20 = -1.609. (c)
Calculate the z-score: z-score = (log of
individual measurement value - log of the mean value for the reference
population)/residual SD (shown in Table 1
); that is, (-1.609 +
0.856)/0.51 = -1.476. Thus the individual measurement value is
1.48 SD below the mean value for the reference population, or
z-score = -1.48.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-tocopheryl polyethylene glycol-1000 succinate. Gastroenterology 1987;93:975-985.
[Web of Science][Medline]
[Order article via Infotrieve]
-tocopherol in healthy persons determined by high-performance liquid chromatography. Clin Chim Acta 1990;194:131-144.
[Web of Science][Medline]
[Order article via Infotrieve]
-tocopherol levels of Mexican Americans, Cubans, and Puerto Ricans aged 4–74. Am J Clin Nutr 1989;50:491-496.
-tocopherol, and cholesterol among elderly people and young adults: increased plasma retinyl esters among vitamin A-supplement users. Am J Clin Nutr 1989;49:112-120.
-tocopherol and cholesterol in healthy French children. Int J Epidemiol 1993;22:237-246.
-tocopherol. Am J Clin Nutr 1994;60:106-110.
-tocopherol. Am J Clin Nutr 1993;57:551-556.
-tocopherol on the vitamin E content of human low-density lipoproteins and resistance to oxidation. J Lipid Res 1991;32:1325-1332.
[Abstract]
-tocopherol vs -tocopherol in human lipoproteins. Am J Clin Nutr 1989;49:517-526.
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  M. Hogstrom, A. Nordstrom, and P. Nordstrom Retinol, retinol-binding protein 4, abdominal fat mass, peak bone mineral density, and markers of bone metabolism in men: the Northern Osteoporosis and Obesity (NO2) Study Eur. J. Endocrinol., May 1, 2008; 158(5): 765 - 770. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Jordan, A. J. De Roos, J. B. Renner, G. Luta, A. Cohen, N. Craft, C. G. Helmick, M. C. Hochberg, and L. Arab A Case-Control Study of Serum Tocopherol Levels and the Alpha- to Gamma-Tocopherol Ratio in Radiographic Knee Osteoarthritis: The Johnston County Osteoarthritis Project Am. J. Epidemiol., May 15, 2004; 159(10): 968 - 977. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B.-L. Lee, A.-L. New, and C.-N. Ong Simultaneous Determination of Tocotrienols, Tocopherols, Retinol, and Major Carotenoids in Human Plasma Clin. Chem., December 1, 2003; 49(12): 2056 - 2066. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Feki, S. Belal, H. Feki, M. Souissi, M. Frih-Ayed, N. Kaabachi, F. Hentati, M. Ben Hamida, and A. Mebazaa Serum Vitamin E and Lipid-adjusted Vitamin E Assessment in Friedreich Ataxia Phenotype Patients and Unaffected Family Members Clin. Chem., March 1, 2002; 48(3): 577 - 579. [Full Text] [PDF]
|
|
|
|
 |
|
 S Renner, R Rath, P Rust, S Lehr, T. Frischer, I Elmadfa, and I Eichler Effects of {beta}-carotene supplementation for six months on clinical and laboratory parameters in patients with cystic fibrosis Thorax, January 1, 2001; 56(1): 48 - 52. [Abstract] [Full Text]
|
|
|
|
|
|
M. Feki, M. Souissi, E. Mokhtar, M. Hsairi, N. Kaabachi, H. Antebi, L. G. Alcindor, R. Mechmeche, and A. Mebazaa Vitamin E and Coronary Heart Disease in Tunisians Clin. Chem., September 1, 2000; 46(9): 1401 - 1405. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. ROOB, G. KHOSCHSORUR, A. TIRAN, J. H. HORINA, H. HOLZER, and B. M. WINKLHOFER-ROOB Vitamin E Attenuates Oxidative Stress Induced by Intravenous Iron in Patients on Hemodialysis J. Am. Soc. Nephrol., March 1, 2000; 11(3): 539 - 549. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
