Background: Paraoxonase may mitigate oxidative damage and thus lower risk of macrovascular disease.
Methods: DNA samples from 2252 participants in the Coronary Artery Risk Development in Young Adults (CARDIA) study were genotyped for the L55M and Q192R polymorphisms of the PON1 (paraoxonase 1) gene, and paraoxonase activity was measured in serum.
Results: The 192R (67.4% vs 29.7%) and 55L (80.0% vs 63.8%) alleles were more common in blacks vs whites. The Q192R locus was the strongest correlate of paraoxonase activity (100.4 nmol/mL/min greater in the 192RR than the 192QQ genotype). After adjustment for the Q192R locus, the L55M locus (12.7 nmol/mL/min difference between 55LL and 55MM) and race (6.6 nmol/mL/min difference between blacks and whites) were correlated with paraoxonase activity (P ≤0.0001), as were concentrations of HDL cholesterol (23.9 nmol/mL/min difference between highest and lowest quintiles), triglycerides (12.6 nmol/mL/min difference between highest and lowest quintiles), LDL cholesterol (8.2 nmol/mL/min difference between highest and lowest quintiles), smoking status (6.3 nmol/mL/min difference between current smokers of ≥15 cigarettes/day and never smokers), and glucose concentrations at the highest quintile (6.5 nmol/mL/min difference between highest and lowest quintiles in nondiabetic participants). There was no cross-sectional or longitudinal association between paraoxonase enzyme activity and coronary artery calcification (CAC), an early marker of cardiovascular disease, or its progression over 5 years.
Conclusions: Paraoxonase may not play an important role during the early pathogenesis of cardiovascular disease. However, associations with lipids and glucose suggest that paraoxonase may modify or react to macrovascular disease pathogenesis.
Paraoxonase, a high-density lipoprotein particle–associated A-oxonase initially characterized by its ability to hydrolyze organophosphates including the insecticide paraoxon(1), has been implicated in cardiovascular disease pathogenesis(1)(2)(3)(4). In experimental studies, paraoxonase may negate the damaging effects of organophosphates and oxidized lipids(5). Therefore, besides reverse cholesterol transport, high-density lipoprotein particles may prevent LDL oxidation through interaction with proteins such as paraoxonase(5).
The paraoxonase family (PON1, PON2, and PON3)1 is located on the long arm of human chromosome 7(6). Two coding region polymorphisms (leucine to methionine substitution at codon position 55 [L55M] and glutamine to arginine substitution at codon position 192 [Q192R]) and 5 promoter region polymorphisms, such as C-108T, are known in the PON1 gene, whereas 2 common polymorphisms are known in the PON2 gene coding region(2). Among these, the Q192R and L55M polymorphisms are most strongly associated with paraoxon hydrolysis(3)(7)(8). The 55L and 192R allele frequencies are higher in blacks vs whites(9), and Chinese have the highest frequency of 55L(10). A recent metaanalysis found a significant association between Q192R and coronary heart disease, but no parallel association for L55M(11). In addition to the genetic polymorphisms, numerous environmental factors such as smoking, atherogenic diet, polyphenols, and alcohol may affect paraoxonase activity and its potential to alter the course of macrovascular disease(2).
We hypothesized that greater paraoxonase activity would be associated with presence and progression of coronary artery calcification (CAC)2 , thus with early, subclinical cardiovascular disease pathogenesis. We studied this hypothesis in the prospective cohort Coronary Artery Risk Development in Young Adults (CARDIA) and its ancillary study Young Adult Longitudinal Trends in Antioxidants (YALTA). We also evaluated the associations of paraoxonase activity and its 2 coding region polymorphisms with cardiovascular disease risk factors.
Materials and Methods
CARDIA recruited 5115 black and white men and women aged 18–30 years in 1985–1986, balanced on race, sex, education (high school or less and more than high school), and age (18–24 and 25–30) in Birmingham, AL, Chicago, IL, Minneapolis, MN, and Oakland, CA. Follow-up examinations occurred during 1987–1988 (year 2, response rate 90%), 1990–1991 (year 5, 86%), 1992–1993 (year 7, 81%), 1995–1996 (year 10, 79%), 2000–2001 (year 15, 74%), and 2005–2006 (year 20, 72%).
Data included blood pressure, circulating lipids, anthropometrics (height and weight), use of tobacco and alcohol, dietary and exercise patterns, psychological variables, medical and family history, and laboratory tests. Height and weight formed body mass index (BMI, kg/m2).
In addition, subclinical atherosclerosis was measured via computed tomography (CT) during years 15 and year 20. CAC was measured by electron beam CT (Imatron Inc.) in Oakland and Chicago and by multidetector CT in Birmingham (General Electric Lightspeed) and Minneapolis (Siemens S4+ Volume Zoom). A radiologist identified the courses of the coronary arteries, using specially developed image-processing software programmed to define a calcified focus, and average Agatston scores from 2 scans were used to obtain the total calcium score for each individual (year 15: Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center, Torrance, CA; year 20: Image Laboratory at Wake Forest University Baptist Medical Center, Winston-Salem, NC). Each scan set with at least 1 nonzero score and a random sample of those with zero scores were reviewed side by side by an expert investigator without knowledge of the Agatston scores to verify the presence of coronary calcium(12)(13). For enhanced interpretation of progression of CAC, a further side-by-side review was done of any scans that showed any calcium at either year 15 or year 20, plus a sample of those that showed no calcium at either examination. CT scans were obtained at both years 15 and 20 in those with paraoxonase activity, and single nucleotide polymorphism (SNP) was measured at year 15 in 1797 participants.
YALTA is an ancillary study to CARDIA, ongoing since 1995, with biochemical analyses performed on serum or plasma collected at the CARDIA year 0, 7, 10, 15, and 20 examinations, including paraoxonase activity at year 15 and the 2 PON1 polymorphisms.
Overnight fasting blood samples were collected to yield EDTA-plasma and serum, both processed within 90 min of blood collection and stored at −70 °C. DNA was extracted from blood collected at year 10 using a PureGene DNA extraction kit (Gentra Systems) and stored at −70 °C. Of 3672 year-15 participants, 2255 had both PON1 genotype and paraoxonase activity measurements. All assays were performed at the Molecular Epidemiology and Biomarker Research Laboratory (MEBRL) at the University of Minnesota (Minneapolis, MN), unless specified otherwise. Fifty-five individuals with serum creatinine concentrations >12 mg/L for women and 15 mg/L for men were excluded for possible renal dysfunction. Forty-seven participants treated with diabetes medications were also excluded. After excluding participants with other missing covariates, 2137 participants were included in the current analysis. Fifty-nine percent were white (47% male, 53% female), and 41% were black (42% male, 58% female).
pon1 genotype determination
Two common PON1 coding region polymorphisms, Q192R (rs662) and L55M (rs854560)(14), were detected using a PCR-Invader assay. PCR was performed using the following primers: PON 55 (forward) 5′- GAAGAGTGATGTATAGCCCCAG-3′ and (reverse) 5′-TTTAATCCAGAGCTAATGAAAGCC-3′ PON 192 (forward) 5′-TATTGTTGCTGTGGGACCTGA-3′ and (reverse) 5′-CACGCTAAACCCAAATACATCTC-3′.
The PCR mixture contained 5 μL DNA (20 ng/μL), 0.75 μL dNTPs, 1.25 μof MgCl2, 1.25 μL 10× buffer, 0.1 μL Taq polymerase, and 1 μL of the primers, in a final volume of 12.5 μL. PCR conditions were denaturation at 94 °C for 2 min and 20 cycles of amplification (94 °C for 30 s, 61 °C for 30 s, and 72 °C for 30 s) followed by elongation at 72 °C for 7 min. The PCR product, diluted 10 times with water, was used as a template for the monoplex Invader reaction(15). Genotypes for L55M polymorphism were obtained after a 20-min incubation at 65 °C, whereas genotypes for the Q192R polymorphism were obtained after a 5-min incubation at 65 °C.
In 42 randomly selected blindly regenotyped samples, reproducibility for L55M was 95%; for Q192R, 100%.
Three samples, with genotypes previously identified using PCR-RFLP methodology, were run with each sample run; genotypes for the sample batch were accepted only if all control samples yielded correct genotypes.
We measured paraoxonase activity as described by Furlong et al.(16) with slight modifications. Briefly, the mixture of 20 μL serum and 780 μL assay buffer (0.132 mol/L Tris-HCl, pH 8.5, 1.32 mmol/L CaCl2) was incubated at 37 °C and initiated by the addition of 200 μL of 6 mmol/L freshly prepared paraoxon substrate solution. The absorbance was continuously monitored spectrophotometrically at 405 nm (37 °C), with readings at 3 min and every minute thereafter to 8 min. A molar absorptivity of 18.05 × 103 was used to calculate activity, with 1 unit of paraoxonase activity defined as 1 nmol of p-nitro phenol formed per minute under these assay conditions (nmol/mL/min). A control sample was analyzed periodically (after every 10 study samples), and the values were monitored to ensure that it remained within 2 standard deviations of the established mean value.
Plasma total cholesterol, HDL cholesterol, and triglycerides were measured enzymatically within 6 weeks of collection(17) at the Northwest Lipid Research Laboratory at the University of Washington (Seattle, WA). HDL cholesterol was determined after precipitation of LDL-containing lipoproteins with dextran sulfate/magnesium chloride(18); LDL cholesterol was calculated in participants with triglyceride concentrations <4000 mg/L by the Friedwald equation(19). Glucose was measured using a Cobas Mira Plus chemistry analyzer (Roche Diagnostic Systems) with the hexokinase method(20). High-sensitivity ELISAs were used to measure serum C-reactive protein (CRP) at the Department of Pathology, University of Vermont(21). The test-retest correlation, in 448 masked duplicate samples, was 0.98–0.99 for total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, and CRP.
All analyses were carried out with PC SAS version 9.0. We tabulated genotype frequencies for L55M and Q192R polymorphisms and determined race-specific departure from Hardy-Weinberg equilibrium by the χ2 test. We evaluated linkage disequilibrium between the 2 polymorphisms using Haploview(24). Haplotypes using the 2 SNPs were constructed using PHASEv2 software(25). Both year 15 and year 20 CAC were analyzed dichotomously. Year 15 F2-isoprostane and lipid concentrations were normally distributed, whereas glucose concentrations and CRP concentrations had skewed distributions. Year 15 F2-isoprostane concentration was analyzed continuously, as was year 15 CRP concentration after log transformation. Year 15 fasting lipid concentrations (HDL cholesterol, LDL cholesterol, triglycerides, and total cholesterol), and year 15 glucose were analyzed as quintiles. Smoking status at year 15 had 4 categories: never smokers, former smokers, current smokers of <15 cigarettes/day, and current smokers of 15 or more cigarettes/day (thus allowing for possible understatement by a smoker of a pack a day).
We used linear regression to model combined and individual genotypes to determine the associations of L55M and Q192R with paraoxonase activity after adjustment for race, sex, age, and clinical center. In addition, we used linear regression to evaluate the cross-sectional association of paraoxonase activity with year 15 smoking status, fasting lipids, CRP, F2-isoprostanes, and glucose after adjustment for the above potential confounders. Unconditional logistic regression was used to analyze the association of year 15 and year 20 CAC presence with year 15 paraoxonase enzyme activity and genotypes of L55M and Q192R polymorphisms after adjustment for race, sex, age, and clinical center.
Mean age was 25 years at study entry, with 26% of blacks and 16% of whites current smokers at year 15 (P <0.0001) (Table 1⇓ ). Blacks vs whites had significantly higher CRP concentrations, systolic and diastolic blood pressure, and BMI (P <0.0001). Fasting glucose concentrations, isoprostane concentrations, and serum HDL and LDL cholesterol were similar across racial groups, whereas whites had significantly higher triglycerides and total cholesterol vs blacks (P <0.003).
pon1 genotypes and paraoxoanase activity
Both PON1 polymorphism frequencies differed between whites and blacks (P <0.0001) (Table 2⇓ ). The 55M allele was less common and the 192R allele was more common among blacks vs whites. The 2 polymorphisms were in Hardy-Weinberg equilibrium, except for the Q192R genotype in whites (P = 0.02). The L55M and Q192R polymorphisms were in weak linkage disequilibrium (r2 = 0.03, P <0.0001 in blacks; r2 = 0.19, P <0.0001 in whites).
Paraoxonase activity was normally distributed, with mean (SD) 93.8 (53.2) nmol/mL/min, although there was skewing toward larger values within some of the race and genotype subgroups. Both L55M and Q192R were strongly and independently associated with paraoxonase activity (mean for QQ 45.5 nmol/mL/min vs RR 154.5 nmol/mL/min; MM 52.5 nmol/mL/min vs LL 113.1 nmol/mL/min, both P <0.0001). The association between PON haplotypes and paraoxonase activity was similar to that observed for the Q192R polymorphism (data not shown). After adjustment for the Q192R polymorphism (Table 3⇓ ), mean difference in paraoxonase activity in the L55M polymorphism was substantially attenuated (MM 86.0 nmol/mL/min vs LL 98.7 nmol/mL/min, P <0.0001). The unadjusted difference in paraoxonase activity between blacks (121.0 nmol/mL/min) and whites (75.0 nmol/mL/min, P <0.0001) was greatly attenuated by adjustment for the 2 polymorphisms (97.1 nmol/mL/min in blacks vs 91.5 nmol/mL/min in whites, P <0.0001) (Table 3⇓ ).
paraoxonase activity, pon1 gene polymorphisms, and cac
CAC (Agatston score >0) was observed in 202 individuals at year 15, and 319 individuals had CAC at year 20. A significantly higher percentage of whites had CAC at year 15 compared with blacks (11.0% vs 7.2%; P = 0.004) (Table 1⇑ ), and the difference between races persisted at year 20 (19.8% for whites vs 14.5% for blacks; P = 0.004) (Table 4⇓ ). Among the 1797 participants who underwent CT examination at both years 15 and 20, all but a few participants remained stable or progressed to a higher Agatston score. Specifically, 1468 (81.7%) participants showed no evidence of CAC at both year 15 and year 20 exams, 9% had incident CAC at year 20, 8% increased CAC over the 5 years, and only 1% decreased or lost CAC (Table 4⇓ ). The median Agatston scores at year 15 (21.0 U [range 0.8–752.1]) and year 20 (25.7 U [range 0.6–2026.0]) were low, reflecting the early stage of atherosclerosis among CARDIA participants.
Neither PON1 polymorphism nor year 15 paraoxonase activity was associated with year 15 CAC or year 20 CAC (Table 5⇓ ). PON1 haplotypes were also not associated with year 15 CAC or year 20 CAC (data not shown). In addition, year 15 paraoxonase activity was not associated with progression of CAC (defined as an increase in Agatston score >10 U) over a 5-year period (odds ratio for 53 nmol/mL/min increase in paraoxonase activity 0.86, CI 0.70–1.06, P = 0.15). The lack of association between paraoxonase activity, PON1 gene polymorphisms, and year 15 or year 20 CAC did not differ across racial groups.
paraoxonase activity, pon1 gene polymorphisms, and other major cardiovascular risk factors
Fasting glucose and lipid concentrations were significantly associated with paraoxonase activity after adjustment for age, race, sex, clinical center, and the Q192R and L55M polymorphisms. Year 15 paraoxonase activity was higher in the highest quintile of fasting glucose concentrations vs the lower 4 quintiles (98.8 nmol/mL/min vs 92.3 nmol/mL/min, P = 0.006) (Table 6⇓ ). However, paraoxonase activity levels were not different among the lower 4 quintiles (P >0.05). These associations remained unchanged even after excluding people with untreated diabetes (n = 41), whose paraoxonase activity was 99.8 nmol/mL/min.
Paraoxonase activity was significantly higher in the highest vs lowest HDL cholesterol quintile, with an interaction between Q192R and HDL cholesterol (P = 0.005). Within each level of the PON1 Q192R polymorphism, paraoxonase activity was significantly higher at the highest HDL cholesterol quintile vs lowest HDL cholesterol quintile, the difference being greatest in the 192RR genotype (P <0.001) (Table 7⇓ ). Paraoxonase activity was also significantly higher in the highest quintiles of LDL cholesterol and triglyceride concentrations vs their lowest quintiles after adjusting for HDL concentrations (lowest and highest triglyceride quintiles: 87.4 nmol/mL/min vs 100.0 nmol/mL/min, P <0.0001; lowest and highest LDL quintiles: 89.5 nmol/mL/min vs 97.7 nmol/mL/min, P = 0.0003) (Table 6⇑ ). There was no interaction between paraoxonase genotype and either LDL cholesterol or triglycerides in estimating paraoxonase activity after adjustment for HDL cholesterol and its interaction with paraoxonase genotype. Current smoking of 15 or more cigarettes/day was associated with a significantly lower paraoxonase activity vs current smokers who smoked fewer than 15 cigarettes/day, former smokers, or never smokers (87.9 nmol/mL/min vs 92.0 nmol/mL/min, 96.3 nmol/mL/min, or 94.2 nmol/mL/min, respectively) (P = 0.01) (Table 6⇑ ). In contrast, paraoxonase activity was not associated with other cardiovascular risk factors, including year 15 CRP concentrations, alcohol intake, BMI, and physical activity (data not shown). Furthermore, paraoxonase activity was not associated with year 15 carotenoid, vitamin C, or F2-isoprostane concentrations (data not shown). Evaluations between paraoxonase activity and other variables in people with diabetes and people with CAC at years 15 and 20 revealed similar associations as noted in disease-free individuals (data not shown).
Neither L55M nor Q192R PON1 polymorphism was associated with cardiovascular risk factors, including year 15 carotenoid concentrations, BMI, lipids, glucose, CRP, F2-isoprostanes, and ascorbic acid (data not shown).
In the present study, Q192R polymorphism was the strongest correlate of paraoxonase activity, whereas L55M and race were weaker correlates. Neither paraoxonase activity nor PON1 polymorphisms were associated with coronary artery calcification, an early marker of cardiovascular disease. They were not associated with plasma F2-isoprostane concentrations, a marker of systemic oxidative stress, or other cardiovascular risk factors, including BMI and CRP concentration. However, paraoxonase activity was associated with smoking status, increased concentrations of HDL cholesterol, triglycerides, and LDL cholesterol, and the highest quintile of glucose, even after exclusion of untreated diabetics.
The 55L and the 192R alleles are more common in blacks vs whites, similar to other studies(26)(27). Both 55L and 192R alleles were associated with paraoxonase activity (with paraoxon used as the substrate) in both races, consistent with findings from previous studies(28)(29). The Q192R polymorphism was by far the strongest predictor of paraoxonase activity, and this observation is consistent with findings from a previous study that indicated that 88% of variability in paraoxonase activity across individuals was explained by 4 PON1 polymorphisms with the Q192R polymorphism accounting for most of the variation(3). Since L55M and Q192R were not highly correlated in this study, the marked attenuation in the strength of association between L55M polymorphism and paraoxonase activity after adjustment for the Q192R polymorphism suggests that changes in the paraoxonase enzyme structure induced by the Q192R polymorphism may be more important in determining levels of activity using paraoxon as substrate vs the changes induced by the L55M polymorphism.
Paraoxonase activity was higher in blacks than whites, consistent with the observation that the 192R and 55L alleles (associated with increased paraoxonase activity), were more common in blacks than whites. However, race remained a predictor of paraoxonase activity even after adjustment for the Q192R and L55M polymorphisms. To our knowledge, this is the first study to report ethnic differences in paraoxonase activity.
Paraoxonase activity was positively associated with lipid concentrations in this study. The physical association between HDL particle and paraoxonase probably accounts for the strong positive association of paraoxonase activity with the HDL cholesterol particle (difference of 24 nmol/mL/min between highest and lowest HDL cholesterol quartiles) and the significant interaction between Q192R genotype and HDL cholesterol concentrations in determining paraoxonase activity. This finding is consistent with prior studies(30) and is consistent with decreased paraoxonase activity being associated with increased cardiovascular disease risk. The weaker positive association between paraoxonase activity and the atherogenic lipid fraction, LDL cholesterol (difference of 12 nmol/mL/min between highest and lowest LDL cholesterol quartiles) probably reflects association of paraoxonase with other factors such as SREBP-2, a cholesterol transcription factor, which regulates PON1 expression(31)(32)(33). The positive association between paraoxonase activity and triglycerides (difference of 8 nmol/mL/min between highest and lowest triglyceride quartiles) reflects the small fraction of paraoxonase enzyme that is carried on the triglyceride fraction (<5%)(34).
Paraoxonase activity levels and PON1 gene polymorphisms were not associated with CAC or other cardiovascular risk factors such as BMI, CRP, F2-isoprostanes, and circulating dietary antioxidants, including year 15 ascorbic acid and year 7 carotenoids. These findings are consistent with a recent study that showed no association between homocysteine concentrations and paraoxonase activity in the general population(35). Early markers of cardiovascular disease, such as CAC, may not necessarily be representative of all pathological changes that occur in development of cardiovascular disease. Furthermore, since CAC was observed in 319 participants at year 20, this study had 80% power at α = 0.05 to detect an odds ratio of 1.48 between paraoxonase activity and CAC pooled across both races. This study also had 80% power at α = 0.05 to detect an odds ratio of 1.43 between the 2 PON1 polymorphisms, Q192R and L55M, and CAC when the heterozygous and homozygous minor variant genotypes of each polymorphism were combined into a single category and compared to the homozygous major variant. Thus modest associations between paraoxonase activity, Q192R and L55M polymorphisms, and CAC lower than 1.43 may have been missed. This is the first study to find higher paraoxonase activity to be significantly associated with the highest quintile of glucose concentration vs other quintiles even after excluding people with diabetes, and this finding needs to be confirmed in other populations. Consistent with our hypothesis that lower paraoxonase activity would be associated with high oxidative stress, current smokers who smoked >15 cigarettes/ day had lower paraoxonase activity compared with former or never smokers. However, this hypothesis is not supported by the lack of association of the PON1 genotypes and paraoxonase activity with F2-isoprostanes, a marker of systemic oxidation. It is possible that the role of paraoxonase in protection against oxidative stress to specific substrates may be masked when using F2-isoprostanes as a measure of global oxidative stress. The poor correlation between estimated paraoxonase enzyme activity and its antioxidant role when paraoxon is used as a substrate(36)(37) may be a possible explanation for the lack of association between F2-isoprostanes and paraoxonase activity. However, some of the studies linking paraoxonase activity to coronary heart disease did use a paraoxon substrate in their assessment of activity(1). Thus, though a previous study has shown F2-isoprostanes to be associated with CAC, the results of this study do not support an important role for paraoxonase enzyme in predicting CAC or the early pathogenesis of cardiovascular disease(12). The exact role of paraoxonase in accelerating or modulating the pathogenesis of cardiovascular disease remains controversial(2).
CARDIA, a study of healthy young adults who have been followed for the last 20 years, is ideally suited to evaluate the effect of paraoxonase enzyme early in the pathogenesis of cardiovascular disease. Further follow-up of the CARDIA cohort will provide valuable information regarding longitudinal associations between paraoxonase activity and cardiovascular risk factors.
Grant/Funding Support: The study was funded by National Heart, Lung, and Blood Institute contracts N01-HC-48047, N01-HC-48048, N01-HC-48049, N01-HC-48050, N01-HC 95095 (CARDIA) and R01-HL-53560 (YALTA).
Financial Disclosures: None declared.
Data are mean (SD) or %.
1 Statistically significant difference between whites and blacks (all P <0.007)
Data are n (%).
The table is based on a linear regression model including age, clinical center, Q192R and L55M polymorphisms, race, sex, smoking status, glucose concentrations (quintiles), triglycerides (quintiles), LDL cholesterol (quintiles), and HDL cholesterol (quintiles).
1 Highest vs lowest category.
Data are n (%).
Data are odds ratio (95% CI). A separate model was run for year 15 CAC presence and for year 20 CAC presence, each with adjustment for age, race, sex, and clinical center. Further adjustment for the variables in Table 6⇑ did not alter conclusions.
1 Calculated for 1 standard deviation (53 nmol/mL/min) in paraoxonase activity.
Data in the table are based on a linear regression including all tabulated variables, plus age, race, sex, Q192R and L55M polymorphisms, and clinical site. Instead of quintiles, smoking categories are never, former, current <15 cigarettes/day, and current ≥15 cigarettes/day. Quintiles (in mg/L) for glucose: Q1, 510–760; Q2, 770–810; Q3, 820–850; Q4, 860–910; Q5, 920–1230; for LDL cholesterol: Q1, 170–860; Q2, 870–1020; Q3, 1030–1180; Q4, 1190–1380; Q5, 1390–2540; for triglycerides: Q1, 190–540; Q2, 550–720; Q3, 730–950; Q4, 960–1360; Q5, 1370–3960; and for HDL cholesterol: Q1, 140–380; Q2, 390–450; Q3, 460–520; Q4, 530–620; Q5, 630–1140.
1 Quintile terminology is used even though the categories presented are not 5 equally sized groups.
2 Highest vs lowest category.
Data in the table are based on a linear regression including smoking, glucose, LDL cholesterol, triglycerides, HDL cholesterol, age, race, sex, Q192R and L55M polymorphisms, and clinical site. P for interaction <0.0001.
1 Highest vs lowest category.
↵1 Human genes: PON1, PON2, PON3, paraoxonase 1, 2, 3
↵2 Nonstandard abbreviations: CAC, coronary artery calcification; BMI, body mass index; CT, computed tomography; CRP, C-reactive protein.
- © 2008 The American Association for Clinical Chemistry