| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Lipids, Lipoproteins, and Cardiovascular Risk Factors |
Departments of1
Pediatrics,
2 Clinical Biochemistry, and
3 Nutrition, Ste-Justine Hospital and Université de Montréal, Montreal, Quebec, Canada.
4 Department of Epidemiology and Biostatistics, McGill University, Montreal, Quebec, Canada.
aAddress correspondence to this author at: Medical Genetics Division, Ste-Justine Hospital, 3175 Côte-Sainte-Catherine, Montreal, Quebec, Canada, H3T 1C5. Fax 514-345-4766; e-mail marie.lambert{at}umontreal.ca.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: In 1999, we conducted a school-based survey of a representative sample of youths 9, 13, and 16 years of age in the province of Quebec, Canada. Standardized clinical measurements and fasting plasma lipid, glucose, insulin, and CRP concentrations were available for 2224 individuals.
Results: The distribution of CRP was positively skewed. The median and 95th percentile values by age and sex ranged from <0.2 to 0.56 mg/L and from 2.72 to 6.28 mg/L, respectively. A total of 7.7% of 9-year-olds, 5.5% of 13-year-olds, and 12.8% of 16-year-olds had CRP concentrations >3.0 mg/L, the threshold defining the adult high-risk category. We observed a strong relationship between CRP concentrations and both body mass index (BMI) and fasting insulin values. The association between CRP and insulin concentration was markedly attenuated after adjustment for BMI, whereas that between CRP and BMI remained unchanged after adjustment for insulin: a 1 SD increase in BMI was associated with a 52% increase in CRP concentration. An increased CRP concentration was independently associated with a worsening of the lipid profile, whereas the association between increased CRP values and high systolic blood pressure was no longer statistically significant after adjustment for BMI.
Conclusions: The metabolic correlates of excess weight, including a state of low-grade systemic inflammation, are detectable early in life. Their health impact in adults remains to be fully examined.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In adults (6)(7)(8)(9)(10)(11)(12)(13), plasma CRP concentrations are significantly associated with body fat as well as with specific components of the metabolic syndrome, including systolic blood pressure (BP) and fasting plasma concentrations of insulin, triglycerides (TGs), and HDL-cholesterol (HDL-C). However, the independent contributions of excess adiposity and its frequently associated metabolic abnormalities to variations in CRP values are poorly understood. Some studies suggest that chronic, low-grade inflammation might be an early event in the development of insulin resistance (IR) and the metabolic syndrome (14)(15). However, McLaughlin et al. (16) have demonstrated that in healthy obese women, CRP concentrations were increased predominantly among those who were insulin resistant and that the relationship between CRP concentrations and IR was independent of obesity, suggesting that IR and/or other abnormalities associated with the metabolic syndrome might induce inflammatory responses.
In children, data are limited, but associations between CRP concentrations and CVD risk factors similar to those reported for adults have been observed (17)(18)(19)(20). In contrast to adults, studies in youth are less likely to be confounded by chronic conditions such as bronchitis, arthritis, atherosclerosis, or undetected diseases. Therefore, studies in pediatric populations may help improve our understanding of the relationship between obesity, inflammation, IR, and the metabolic syndrome. Furthermore, because persistently increased CRP concentrations in youth may provide an early warning for later risk of CVD, it is important to precisely characterize the relationship between CRP values, excess weight, and CVD risk factors in pediatric populations. The objectives of this study thus were as follows: (a) to describe the distribution of high-sensitivity CRP concentrations in a population-based sample of children and adolescents; (b) to examine the association between CRP concentrations and a measure of obesity, body mass index (BMI), and between CRP and a surrogate measure of IR, fasting plasma insulin, and to evaluate the independent contributions of these two variables to variations in CRP concentrations; and (c) to assess the association between CRP concentrations and other components of the metabolic syndrome.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
clinical variables
BP, height, and weight were measured according to standardized protocols (21). Percentile cut point values used to categorize individuals with regard to metabolic risk factors were estimated from the study distributions. Cut points were age- and sex-specific. BP cut points were also height-specific. Increased TG concentrations and systolic BP were defined as values at or above the 75th percentile, and decreased HDL-C concentrations were defined as values at or below the 25th percentile. These thresholds are similar to those used in the Bogalusa Heart Study and the Cardiovascular Risk in Young Finns Study (22)(23). Current smokers were defined as those who answered yes to the following question: During the past 30 days, did you smoke cigarettes, even just a few puffs? This question was not asked of 9-year-olds; only 2.1% of this age answered yes to the following question: Have you ever smoked a whole cigarette? Ethnicity was defined as French Canadian or other.
biochemical analyses
Blood was obtained by venipuncture after an overnight fast in a 1 g/L EDTA collection tube. High-sensitivity CRP concentrations were measured with the IMMAGE® immunochemistry system (Beckman Coulter). The lower limit of detection of this assay was 0.2 mg/L. Interassay CVs for controls at 0.84 and 13.8 mg/L were 4.2% and 3.5%, respectively. Fasting plasma insulin, glucose, total cholesterol, TG, and HDL-C concentrations were determined as described previously (21). The Homeostasis Model Assessment of IR (HOMA-IR) was calculated as insulin (mIU/L) x glucose (mmol/L)/22.5 (24).
statistical analyses
We used the sample quantiles to estimate population percentiles. Nonparametric confidence intervals (CIs) for the quantiles of interest were constructed by use of the algorithm described by Hutson (25). When comparing percentile values between sexes or across ages, we concluded that they were significantly different if their respective 95% CIs were nonoverlapping. Because a large proportion of CRP measurements were left-censored (lower detection limit = 0.2 mg/L), we used the LIFEREG procedure in SAS (a regression method for analysis of censored data) to study the association between CRP, BMI, and plasma insulin (26). This strategy avoids the biases and false precision induced when all concentrations below the detection limit are artificially set to the value of the detection limit (in our case, 0.2 mg/L). CRP concentrations were not gaussian distributed and were loge-transformed [undetectable values were considered censored at log(0.19)]. We assumed that loge-transformed CRP values had an underlying gaussian distribution (17). The regression coefficients for the 100 loge-transformed CRP concentrations represent the percentage change in CRP per unit change in independent variable (27). The associations between CRP category (age- and sex-specific upper quartile and other) and metabolic risk factors were examined in logistic regression analyses. Because we pooled age groups, age- and sex-specific Z-scores for BMI and insulin concentration were used in regression analyses. BMI and insulin concentration were not gaussian distributed and were loge-transformed. We estimated Z-scores from the study distributions. To take the complex sampling design into account, we estimated sampling weights and clustering effects and incorporated them into all of our computations except those done with the LIFEREG procedure. Statistical analyses were performed with SAS statistical software (SAS Institute, Inc.) and SUDAAN (Research Triangle Institute).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
. Cigarette smoking and infectious or inflammatory conditions could affect CRP concentrations. Among 13- and 16-year-olds, 12.7–39.9% were current smokers. Only 2.1% of 9-year-olds admitted to having ever smoked a whole cigarette. We considered that individuals who reported they were taking antibiotics, medications for pain/fever, cold/allergies, or respiratory problems in the 2 weeks before blood sampling might have had an infectious or inflammatory condition at the time of the blood sampling. Many individuals (35.0–69.1% according to age and sex) reported to have used these medications in the 2 weeks before blood sampling. We do not have information on the duration or amount of medications used; most of those who took these medications (24.6–54.7% according to age and sex) used only medicines for pain/fever and/or cold/allergies.
|
Forty-seven percent of 9-year-olds, 50.7% of 13-year-olds, and 32.3% of 16-year-olds had CRP concentrations below the detection limit (0.2 mg/L). Two percent of 9-year-olds, 1.1% of 13-year-olds, and 2.6% of 16-year-olds had CRP concentrations >10.0 mg/L, the conventional cut point for increased CRP (3). Selected age- and sex-specific percentile values of CRP concentrations are presented in Table 2
. These percentiles values were computed first including and then excluding individuals with possible infectious/inflammatory conditions and current smokers. Although there was a tendency toward lower values when individuals with possible infectious/inflammatory conditions and current smokers were excluded, for all ages, in both sexes, the CIs around each percentile value largely overlapped between estimations calculated with and without exclusions. Therefore, subsequent analyses were done with the whole sample.
|
In boys, the median and 75th percentile CRP values generally increased across ages and were significantly different between 9- and 16-year-olds. In girls, the median, the 75th, and the 95th percentile values were the lowest in 13-year-olds and were significantly different in 13-year-olds compared with 9- and 16-year-olds. Nine- and 16-year-old boys tended to have lower CRP concentrations than girls in the same age groups.
After adjustment for age, sex, ethnicity, smoking, and possible inflammatory/infectious conditions, CRP concentrations remained strongly positively associated with BMI: a 1 SD increase in BMI was associated with a 56.9% increase in CRP value (Table 3
, model 1). This relationship was similar in boys and girls. CRP was also positively associated with fasting insulin: a 1 SD increase in insulin was associated with a 37.1% increase in CRP (Table 3
, model 2). Because BMI and insulin are correlated, we assessed their independent contributions to variations in CRP. The positive association of CRP with insulin was markedly attenuated after adjustment for BMI, whereas the relationship of CRP with BMI remained unchanged (Table 3
, model 3). Exclusion of individuals with CRP concentrations >10 mg/L and use of the HOMA-IR as a surrogate measure of IR did not affect the results (data not shown).
|
Because low-grade systemic inflammation may underlie, at least in part, the clustering of metabolic risk factors, we examined the relationship between increased CRP concentrations and high systolic BP, high TG and low HDL-C concentrations, and clustering of these risk factors. After adjustment for age, sex, ethnicity, smoking, and possible inflammatory/infectious diseases, individuals with CRP concentrations in the upper quartile for age and sex were 1.4, 1.7, and 2.3 times more likely to have high systolic BP, high TGs, and low HDL-C, respectively, than their counterparts with CRP concentrations below the age- and sex-specific 75th percentile values (Table 4
). There was also a significant association between increased CRP (upper quartile) and clustering of these risk factors. However, adjustment for insulin weakened the positive associations between increased CRP and both metabolic risk factors and risk factor clustering. Adjustment for BMI had an even greater effect such that the associations between increased CRP and high systolic BP and risk factor clustering were no longer statistically significant. The positive associations between increased CRP and high TGs and low HDL-C concentrations, however, were robust. We did not detect a significant association between high CRP and glucose 
6.1 mmol/L (data not shown); the number of individuals with glucose 6.1 mmol/L was small (n = 38). No heterogeneity of CRP effect by sex was detected (all P >0.1).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
0.1 mg/L (28). This contrasts with recently published data from large European and North American adult populations in which the 25th percentile value was 0.2 mg/L in all age, sex, and ethnic groups (29)(30). As in adults, the distribution of plasma CRP concentrations in youth was positively skewed in this study (17)(19)(28). The median and 95th percentile values by age and sex ranged, respectively, from <0.2 to 0.56 mg/L and from 2.72 to 6.28 mg/L in the Quebec survey and from 0.3 to 0.6 mg/L and from 5.9 to 8.6 mg/L in youth 3 to 19 years of age in the 1999–2000 US NHANES (28). Although CRP concentrations were lower in our pediatric population than typical values in adults, we observed that 7.7% of 9-year-olds, 5.5% of 13-year-olds, and 12.8% of 16-year-olds had CRP concentrations >3.0 mg/L, the proposed threshold to define the adult high-risk category (3). Therefore, a sizable number of children and adolescents might have already experienced the burden of chronic low-grade systemic inflammation. Its long-term clinical significance remains to be determined. CRP concentrations were higher in 9- and 16-year-old girls than in boys in the same age groups in the Quebec survey. Cook et al. (17) made similar observations in UK children 9–11 years of age: CRP concentrations were 47% higher in girls than in boys. However, CRP concentrations were similar by sex in the Taipei Children Heart Study (19). Ford et al.(28) reported significant differences by sex in the 1999–2000 US NHANES only for participants 16–19 years of age, among whom females had higher CRP concentrations than males. In adults, some studies have found differences by sex, whereas others have not (29)(30)(31). It has been suggested that sex differences in CRP concentrations in adults could be explained by estrogen use in women (32). Although no data on oral contraceptive use at the time of blood sampling were collected in our study, it is unlikely that sex differences at age 9 are attributable to estrogen use.
Several prospective studies have documented an association between markers of inflammation, such as CRP, and incident CVD and type 2 diabetes (1)(2)(3). However, the mechanisms responsible for the low-grade up-regulation of CRP production that predicts CVD and type 2 diabetes in general populations are poorly understood. Although cross-sectional studies cannot address this issue completely, we found a strong relationship between CRP concentration and BMI as early as age 9. Our observations concur with those reported in the Ten Towns Children’s Study, the Taipei Children Heart Study, the Third and the 1999–2000 US NHANES (17)(18)(19)(20), and several studies of adults (6)(7)(8)(9)(10)(11)(12)(13). In pediatric populations, it is unlikely that the association between CRP concentration and BMI was confounded by inflammatory processes associated with atherosclerosis, as has been hypothesized in adults (10)(11). In the present study as well as in adults (7)(11)(13), the relationship between CRP and BMI was independent of fasting insulin, which is closely associated with excess weight. Furthermore, the observation from population-based studies of a strong association between CRP and BMI is consistent with experimental and clinical studies. Mohammed-Ali et al. (14) demonstrated that 25–30% of circulating interleukin-6 (IL-6), a primary determinant of hepatic CRP production, was released from subcutaneous adipose tissue. The IL-6 concentration is increased in obesity (8)(33), and weight loss has been associated with significant decreases in IL-6 values in both adipose tissue and serum (15)(33) and with a decrease in plasma CRP concentrations (15)(34). These findings suggest that increased secretion of IL-6 by expanded adipose tissue mass up-regulates the production of CRP by the liver, thus providing a mechanism to explain how increased body fat is linked to increased CRP values.
In adults, stronger associations between body fat and CRP values have been reported for women compared with men (6)(7). In the present study, before adjustment for fasting plasma insulin, we observed a tendency toward a stronger association in boys than girls. After adjustment for insulin, the strength of the association between BMI and CRP values was similar in both sexes. There are several possible explanations for the discrepant findings in studies of adults and children. The first explanation is that body composition and fat distribution are different between adults and children and that these differences may influence the relationship between obesity and CRP. The second possible explanation is that BMI is an indirect measure of adiposity and that the correlation between direct measure of obesity and BMI varies between sexes and across ages (35). The third possible explanation is that differences in sex steroid hormone concentrations between children and adults may affect the association between BMI and CRP; in fact, in adult women, estrogen use increases CRP.
We found a strong relationship between CRP and fasting insulin. However, this association was markedly weakened after adjustment for BMI to the extent that it was no longer statistically significant in girls. Use of HOMA as a measure of IR instead of fasting insulin yielded similar results. In the Ten Towns Children’s Study, Cook et al. (17) did not detect a significant association between CRP and fasting insulin. However, there was a significant relationship between CRP values and postload insulin that disappeared after adjustment for ponderal index. Several cross-sectional epidemiologic studies have examined the relationship between CRP and measures of IR in adults of various ages and both sexes. A consistent observation has been the finding of a positive association between CRP and measures of IR (8)(9)(10)(11)(12) that was variably attenuated after adjustment for adiposity (9)(10)(11). In a clinical study to determine whether CRP was associated with IR independently of obesity, Escobar-Morreale et al. (13) evaluated the relationship between inflammatory markers, IR, and obesity in women with polycystic ovary syndrome, a disease associated with IR independent of excess weight, and in controls. They concluded that obesity and not IR was the major determinant of CRP; only a weak association between insulin sensitivity index and CRP persisted after adjustment for BMI. Finally, Kopp et al. (15) studied the longitudinal relationship of CRP with features of the metabolic syndrome in 37 morbidly obese patients before and 14 months after gastric surgery. In a multivariable regression model including change in BMI, 1-h postload glucose, HOMA, and C-peptide as independent variables, only change in BMI was significantly related to the decrease in CRP. Together, adult and pediatric studies suggest that excess adipose tissue might be the common antecedent of both increased CRP and IR. If there is a biologically relevant independent association between CRP and IR, it appears minor compared with the association between CRP and BMI.
Other features of the metabolic syndrome have been related to markers of inflammation (8)(9)(10). In the present study, we observed a strong association between increased CRP and both high TG and low HDL-C concentrations that was only mildly attenuated after adjustment for BMI or fasting insulin. Conversely, the association between CRP and high systolic BP or clustering of risk factors was no longer statistically significant after adjustment for BMI. In the 1999–2000 US NHANES, Ford (20) reported positive correlations between CRP and both systolic BP and TG concentrations; however, in multivariate analyses that included BMI percentile, only systolic BP remained significantly associated with CRP and solely in girls. In the Ten Towns Children’s Study, Cook et al. (17) found that the negative association between quintiles of CRP and HDL-C remained statistically significant after adjustment for ponderal index, whereas associations between quintiles of CRP and TGs and systolic BP were no longer significant after adjustment for ponderal index. These results suggest that increases in CRP (or other inflammatory markers correlated with CRP) could contribute to a worsening of the lipid profile characteristic of the metabolic syndrome. Furthermore, each component of this syndrome related differently to CRP.
The limitations of this study include that the use of a single CRP measurement may not accurately reflect long-term inflammation status. In healthy adults, the biological variability of log high-sensitivity CRP has been estimated at 22% (36). However, nondifferential misclassification is unlikely to explain our significant findings and should affect similarly the strength of the different associations. Another limitation is that we used indirect measures of body fat and IR. The correlation between BMI and direct measures of adiposity is reasonably high (35). In individuals with normal glucose tolerance, fasting insulin correlates moderately with IR as measured by the euglycemic hyperinsulinemic clamp technique (37). Use of the clamp technique might have yielded a stronger association between CRP and IR in this study. However, it is unlikely that our conclusions are invalidated by the use of fasting insulin to measure IR.
In conclusion, in a large representative sample of youth from the province of Quebec, Canada, a high proportion (8.9%) of individuals had CRP concentrations >3.0 mg/L, the threshold defining the adult high-risk category. Body fat was the major determinant of CRP, whereas fasting insulin might have had a small independent effect. We observed a strong association between increased CRP and both high TG and low HDL-C concentrations that was independent of BMI and fasting insulin, suggesting that increases in CRP contributed to a worsening of the lipid profile.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, in vivo. J Clin Endocrinol Metab 1997;82:4196-4200.The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  J. Pirkola, M. Vaarasmaki, M. Ala-Korpela, A. Bloigu, D. Canoy, A.-L. Hartikainen, M. Leinonen, S. Miettola, M. Paldanius, T. H. Tammelin, et al. Low-Grade, Systemic Inflammation in Adolescents: Association With Early-Life Factors, Gender, and Lifestyle Am. J. Epidemiol., November 16, 2009; (2009) kwp320v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Corvalan, R. Uauy, A. D Stein, J. Kain, and R. Martorell Effect of growth on cardiometabolic status at 4 y of age Am. J. Clinical Nutrition, September 1, 2009; 90(3): 547 - 555. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. B. Lande, T. A. Pearson, R. P. Vermilion, P. Auinger, and I. D. Fernandez Elevated Blood Pressure, Race/Ethnicity, and C-Reactive Protein Levels in Children and Adolescents Pediatrics, December 1, 2008; 122(6): 1252 - 1257. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Nagel, K. Rapp, M. Wabitsch, G. Buchele, A. Kroke, I. Zollner, S. K. Weiland, and W. Koenig Prevalence and Cluster of Cardiometabolic Biomarkers in Overweight and Obese Schoolchildren: Results from a Large Survey in Southwest Germany Clin. Chem., February 1, 2008; 54(2): 317 - 325. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Herder, S. Schneitler, W. Rathmann, B. Haastert, H. Schneitler, H. Winkler, R. Bredahl, E. Hahnloser, and S. Martin Low-Grade Inflammation, Obesity, and Insulin Resistance in Adolescents J. Clin. Endocrinol. Metab., December 1, 2007; 92(12): 4569 - 4574. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Nelson, N. Sattar, D. J. Freeman, J. D. Walker, and R. S. Lindsay Inflammation and Endothelial Activation Is Evident at Birth in Offspring of Mothers With Type 1 Diabetes Diabetes, November 1, 2007; 56(11): 2697 - 2704. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Aeberli, L. Molinari, G. Spinas, R. Lehmann, D. l'Allemand, and M. B Zimmermann Dietary intakes of fat and antioxidant vitamins are predictors of subclinical inflammation in overweight Swiss children. Am. J. Clinical Nutrition, October 1, 2006; 84(4): 748 - 755. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Groner, M. Joshi, and J. A. Bauer Pediatric Precursors of Adult Cardiovascular Disease: Noninvasive Assessment of Early Vascular Changes in Children and Adolescents Pediatrics, October 1, 2006; 118(4): 1683 - 1691. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Juonala, J. S.A. Viikari, T. Ronnemaa, L. Taittonen, J. Marniemi, and O. T. Raitakari Childhood C-Reactive Protein in Predicting CRP and Carotid Intima-Media Thickness in Adulthood: The Cardiovascular Risk in Young Finns Study Arterioscler Thromb Vasc Biol, August 1, 2006; 26(8): 1883 - 1888. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. D. de Ferranti, K. Gauvreau, D. S. Ludwig, J. W. Newburger, and N. Rifai Inflammation and Changes in Metabolic Syndrome Abnormalities in US Adolescents: Findings from the 1988-1994 and 1999-2000 National Health and Nutrition Examination Surveys Clin. Chem., July 1, 2006; 52(7): 1325 - 1330. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. M. Crimmins and C. E. Finch Infection, inflammation, height, and longevity PNAS, January 10, 2006; 103(2): 498 - 503. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. S. Ford, U. A. Ajani, and A. H. Mokdad The Metabolic Syndrome and Concentrations of C-Reactive Protein Among U.S. Youth Diabetes Care, April 1, 2005; 28(4): 878 - 881. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
