OBJECTIVE—Type 1 diabetes is associated with endothelial dysfunction, arterial stiffness, and an increased risk of cardiovascular disease (CVD) events. We previously demonstrated increased arterial stiffness in children with type 1 diabetes compared with control subjects. However, traditional CVD risk factors did not explain the difference in arterial stiffness. Furthermore, children with type 1 diabetes displayed notable within-group variation in arterial stiffness. We hypothesized that polymorphisms in the NOS3 gene may be associated with the differences seen in arterial stiffness within the population of children with type 1 diabetes.
RESEARCH DESIGN AND METHODS—Thirty-six consecutively enrolled subjects aged 10–21 years with type 1 diabetes were studied. Subjects underwent radial tonometry in a fasting state. A corrected augmentation index (AI75) was the primary measure of arterial stiffness. Genotypes were determined for the NOS3 −786T→C and Glu298→Asp polymorphisms by pyrosequencing. AI75 values by genotype groups were compared by ANOVA and multivariate analysis.
RESULTS—Median (interquartile range) AI75 values for −786TT and −786C carriers were −3.5 (−8.8 to 2.3) and 11.0 (6.0 to 14.4), respectively (P = 0.01); AI75 values for Glu298Glu patients and Asp298 carriers were 2.3 (−4.0 to 13.0) and 7.3 (−2.0 to 11.5), respectively (P = 0.59). In univariate analysis, age, sex, BMI percentile, and −786T→C genotype were significantly associated with AI75. The multivariate model, which included these four variables, was significantly associated with AI75 (P = 0.002, R2 = 0.40).
CONCLUSIONS—This is the first reported association between −786T→C and arterial stiffness in type 1 diabetes. Larger studies are needed to confirm this observation for potential translation to risk assessment.
Type 1 diabetes is associated with endothelial dysfunction, premature atherosclerosis, and an increased risk of cardiovascular disease (CVD) events. Despite marked improvements in our understanding of the pathogenesis of diabetes-related complications, CVD event rates in patients with type 1 diabetes have shown little improvement over the last 30 years (1). Use of surrogate markers of CVD may help to stratify risk within the type 1 diabetic population and allow for more aggressive primary prevention of CVD in both adults and children. Several studies have already demonstrated the relationship between abnormal brachial reactivity and increased carotid intima-media thickness in children with type 1 diabetes (2–4). Furthermore, we have recently shown that arterial stiffness measured by radial artery tonometry is higher in children with type 1 diabetes than in age-, sex-, race-, and BMI-matched control subjects (5).
In addition to clear differences in the augmentation index (AI75), a measure of arterial stiffness, between children with and without type 1 diabetes, there is notable within-group variability in this measure among patients with type 1 diabetes. Furthermore, traditional cardiovascular risk factors do not predict AI75 well in these children. As such, it is important to identify other predictors of arterial stiffness to optimize risk stratification in type 1 diabetes.
Nitric oxide (NO) is an important regulator of endothelial function and vascular dilation. Endothelial NO synthase (eNOS) is the enzyme responsible for endothelial NO formation, and dysregulation of eNOS activity results in various pathophysiological vascular conditions. In fact, impairment of the NO pathway has been described in type 1 diabetic patients compared with control subjects (6,7). We previously investigated the relationship between arterial stiffness and direct measurements of NO activity and surprisingly found no significant correlation (8). Because of the instrumental role of eNOS in NO-mediated vascular dilation, we hypothesized that polymorphisms in the gene that encodes the eNOS protein, NOS3, may be associated with the differences seen in arterial stiffness in children with type 1 diabetes that are unexplained by direct measurement of NO activity in the systemic circulation or traditional CVD risk factors.
Two variant NOS3 alleles (−786C and Asp298) have been shown to be functional and occur with high frequency in the general population. The −786T→C promoter polymorphism appears to be functionally important, with the −786C allele showing diminished eNOS expression (9,10). The Glu298Asp nonsynonymous single nucleotide polymorphism (SNP) is associated with a risk of angiographic coronary disease, carotid atherosclerosis, and other pathological vascular phenotypes (11–13). Functionally, the Asp298 allele has been shown to result in dysregulation of NO production (14). In our current analysis, we investigated the single and joint (haplotype) impact of these two SNPs on AI75 in patients with type 1 diabetes.
RESEARCH DESIGN AND METHODS—
Children were recruited from the Pediatric Atorvastatin in Diabetes Intervention Trial (PADIT), which is a randomized, double-blinded, placebo-controlled, cross-over study designed to determine the effects of 3 months of atorvastatin (20 mg daily) versus placebo on arterial stiffness, endothelial function, and lipids in children with type 1 diabetes. Subjects aged 10–21 years were eligible if they had diagnosed type 1 diabetes for a minimum of 1 year. Classification of type 1 diabetes was by history and autoantibody status. Exclusions were known CVD, hepatic disease, pregnancy, use of cholesterol-modifying agents, and other endocrinopathy. Subjects participated in two testing sessions before random assignment to a study drug. The initial testing was performed at enrollment, and repeat measurements were obtained 1 month later. A convenience sample of the first 36 type 1 diabetic subjects enrolled was used for this analysis. The study was approved by the University of Florida Institutional Review Board. All participants provided written assent, and parents provided informed consent.
Arterial stiffness measurement
Subjects underwent arterial stiffness testing by radial tonometry using the SphygmoCor Vx (version 7.01; Atcor Medical) as described previously (5). Tonometry data and blood samples were obtained between 6:00 and 10:00 a.m. after an overnight fast. Subjects were instructed to abstain from caffeine for 24 h before testing. Laboratory values were obtained in the fasting state. A corrected augmentation index (AI75) was the primary measure of arterial stiffness. AI75 was measured in duplicate at both the initial visit and the 1-month visit. The duplicate measures of AI75 were obtained 1–2 min after the initial measure. The average of the duplicate AI75 measures at each of the two time points was used for analysis. All arterial stiffness measurements used in this analysis were obtained before the initiation of the study medication.
Genotype determination and haplotype construction
Genomic DNA was isolated from peripheral blood using a standard protocol (QIAamp; QIAGEN, Valencia, CA). Genotypes were determined by PCR and pyrosequencing. The PCR included ∼50 ng of template DNA, 10 pmol each of forward and reverse primers, 1.5 μl of DMSO, 7 μl of H2O, and 12.5 of μl HotStarTaq Master Mix (QIAGEN).
For the −786T→C polymorphism, a forward 5′-CACCTGCATTCTGGGAACTGTA-3′ and biotinylated reverse 5′-GCCGCAGTAGCAGAGAGAC-3′ primer were used. For the Glu298Asp polymorphism, a forward 5′-CAGGAAACGGTCGCTTCGAC-3′ and biotinylated reverse 5′-CCATCCCACCCAGTCAATC-3′ primer were used. PCR conditions were identical for both reactions: 95°C for 15 min, followed by 45 cycles of 95°C for 30 s, 54°C for 45 s, and 72°C for 1 min, and a final extension at 72°C for 7 min.
Polymorphisms were detected using the PSQ HS 96A system. Between 4 and 7 μl of PCR product were used for sequencing reactions. Ten picomoles of sequencing primer 5′-ATCAAGCTCTTCCCTG-3′ and 5′-GCAGGCCCCAGAT-3′ were used in the −786T→C and Glu298Asp assays, respectively. Haplotypes were computationally reconstructed using PHASE software (version 2.0.9).
Statistical analyses
The primary end point for analysis was difference in AI75 by genotype group. Genotype frequency was determined by allele counting, and Hardy-Weinberg equilibrium was tested by χ2 analysis.
Baseline characteristics were compared between groups by χ2, Fisher’s exact, or unpaired t tests, as appropriate. AI75 values by genotype and haplotype groups were compared by ANOVA. Genetic analyses were performed, assuming a dominant model. Linear regression was performed to determine the joint effects of genetic and nongenetic factors on the AI75. Univariate analyses were performed for age, race, sex, BMI percentile, systolic and diastolic blood pressure, total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides, A1C, family history of heart disease, and both NOS3 polymorphisms. Any variable with P < 0.1 on univariate analysis was entered into the multivariate regression model. Multivariate regression was also performed to determine whether consideration of haplotypes improved model performance over single SNP analyses. Post hoc power calculations based on the available sample sizes for −786T→C indicated that we had 80% power with a two-sided α of 0.05 to detect a difference of 9.0 in AI75 between T/T and C carrier individuals. For Glu298→Asp, we had similar power to detect a difference of 9.2 in AI75 values between Glu298Glu and Asp298 carrier individuals. All statistical analyses were performed using SPSS (version 11.5; SPSS, Chicago, IL). P < 0.05 was considered statistically significant.
RESULTS—
Baseline demographics for the overall population, as well as by genotype groups, are presented in Table 1. The average age was 15 ± 2.9 years, and the majority of patients were white of non-Hispanic origin (86%). Furthermore, 58% of the population was female. Median (interquartile range) of AI75 in the overall sample was 3.63 (−4.0 to 11.5). Frequencies for the −786C and Asp298 variant alleles were 0.31 and 0.25, respectively, and genotype distributions were in Hardy-Weinberg equilibrium. With respect to baseline clinical differences between genotype groups, total cholesterol was lower in Glu298Glu wild-type homozygotes compared with Asp298 variant carriers (4.01 ± 0.75 vs. 4.78 ± 0.96 mmol/l, respectively, P = 0.01). This result was largely driven by differences in LDL cholesterol (2.02 ± 0.57 vs. 2.66 ± 0.65 mmol/l, respectively, P = 0.003). There were no significant differences in clinical variables by −786T→C genotypes.
AI75 values by −786T→C genotypes are shown in Fig. 1. The median (interquartile range) AI75 values for −786TT and −786C carriers were −3.5 (−8.8 to 2.3) and 11 (6.0 to 14.4), respectively (P = 0.01). To further elucidate the gene-dose effect, AI75 was analyzed by −786T→C genotype groups (TT vs. TC vs. CC). AI75 values for −786TT, TC, and CC were −3.5 (−8.8 to 2.3), 10.8 (−0.3 to 20.8), and 11.0 (10 to 12.5), respectively (P = 0.039). AI75 was not significantly different by codon 298 genotype. AI75 values for Glu298Glu patients and Asp298 carriers were 2.3 (−4.0 to 13.0) and 7.3 (−2.0 to 11.5), respectively (P = 0.59).
To determine the joint effects of genetic and nongenetic variables on AI75, univariate analysis was performed to identify significant variables to be included in multivariate analysis. Of the variables included in the analysis, the following were significantly associated with AI75: age (R2 = 0.10, P = 0.06), sex (R2 = 0.12, P = 0.04), BMI percentile (R2 = 0.10, P = 0.07), and −786T→C genotype (R2 = 0.17, P = 0.01), with genotype being the most significant variable. The overall multivariate model that included these four variables (Table 2) was significantly associated with AI75 (P = 0.002, R2 = 0.40).
We performed haplotype analyses to determine whether consideration of both NOS3 genotypes was more informative than either genotype alone. The SNPs were in weak linkage disequilibrium (D′ = 0.29). Diplotype distributions for the children studied are shown in Table 3. Eighty percent of the sample displayed diplotypes that included at least one −786T/Glu298 allele. Comparative analysis of regression models containing haplotype information versus the model that contained only the −786T→C SNP demonstrated that consideration of haplotypes did not improve the overall four-variable model (Table 4).
CONCLUSIONS—
We hypothesized that polymorphisms in NOS3 are associated with arterial stiffness in children with type 1 diabetes. We demonstrated that the −786T→C SNP is associated with variable AI75 values in children with type 1 diabetes, such that variant −786C carriers have dramatically greater arterial stiffness than wild-type TT homozygotes. In fact, TT homozygotes displayed AI75 measurements comparable to those that had been described previously for children without type 1 diabetes (5). Of note, the codon 298 polymorphism was not significantly associated with AI75 variability, and consideration of haplotypes consisting of SNPs at both loci was not as informative as consideration of −786T→C genotype alone.
Although our data are the first to demonstrate an association between the −786T→C polymorphism and arterial stiffness in type 1 diabetes, this polymorphism has been linked to adverse vasculopathic phenotypes in patients with diabetes. For example, Taverna et al. (15) demonstrated that among individuals with long-standing type 1 diabetes, −786CC individuals experienced an average 4-year earlier onset of diabetic retinopathy compared with TT individuals. In line with these findings, Awata et al. (16) demonstrated a relationship between the −786C allele and macular edema in patients with type 2 diabetes. Furthermore, NOS3 polymorphisms other than the −786T→C promoter SNP have also been associated with the presence of diabetic retinopathy, strengthening the assertion that NOS3 variability is associated with multiple diabetes comorbidities (17).
The −786T→C polymorphism has been studied in relation to other microvascular complications of diabetes. For example, using a case-control design of type 1 diabetic patients with and without nephropathy, Zanchi et al. (18) found that CC homozygotes were nearly three times more likely to develop advanced nephropathy. The putative association was strengthened by family linkage analysis that revealed higher than expected haplotype transmission of the −786C/4a allele from parents to offspring with advanced nephropathy. In the sibling Diabetes Heart Study, the −786C allele was independently associated with both the albumin-to-creatinine ratio and risk for albuminuria (19). Of interest, and consistent with our study, the Glu298→Asp polymorphism was not significantly associated with outcomes.
Taken in sum, the above data add plausibility to our findings that the −786T→C SNP is associated with vascular dysfunction and specifically arterial stiffness in children with type 1 diabetes. Dysregulation of the NO pathway has been hypothesized to contribute to adverse metabolic processes in diabetes (6). Umbilical vein endothelial cells obtained from women with type 1 diabetes express ∼20% less eNOS than endothelial cells from women without type 1 diabetes (7). NO homeostasis appears to be tenuous in individuals with diabetes, and the functional consequence of the −786C allele may thus predispose these individuals to endothelial dysfunction and diminished vasorelaxation compared with those who have −786T wild-type alleles. In addition, it is possible that the −786C allele causes impaired insulin-mediated glucose uptake, resulting in poorer glycemic control (20). All of these factors could potentially contribute to the vasculopathies observed in genetic association studies of this SNP to date.
Despite the novel findings provided by this study, several important limitations require further discussion. First, continued debate over the validity of radial artery tonometry as a surrogate marker for arterial stiffness necessitates some degree of caution when interpreting our results. Although some groups continue to question the validity of radial tonometry as a surrogate marker for arterial stiffness, data from other groups support use of the technique (21–28). The transfer function used to calculate the aortic pressure wave has been validated using directly measured aortic and radial pressure waves in adults. However, a large set of data are not yet available to confirm the validity of the transfer function in children. In a small group (n = 3) of children aged 5–10 years who underwent cardiac catheterization, we recently demonstrated that simultaneously performed radial tonometry provided excellent derivations of the aortic pulse wave and accurately estimated the augmentation index (M.J.H., unpublished observations). Clearly the completion of this and other larger studies will be essential to validating the use of radial tonometry as a noninvasive surrogate marker of arterial stiffness.
Second, although our previous studies failed to demonstrate an association between arterial stiffness and surrogate markers of NO activity (serum superoxide dismutase and nitrite) in children between type 1 diabetes, this study demonstrated significant associations between NOS3 polymorphisms and arterial stiffness. Whereas these data may initially appear contradictory, an explanation for this seemingly discordant observation is that gene polymorphisms may in fact be better surrogates of NO activity at the endothelial level than a single measurement of serum superoxide dismutase activity or serum nitrate. Because NO is rapidly degraded and may be influenced by multiple factors (acute stress, hyperglycemia, hypoglycemia, and others), polymorphisms in NOS3 may be more robustly associated with arterial stiffness.
Third, due to the narrow scope of this study, we did not attempt to evaluate the relationship between arterial stiffness and other novel CVD risk markers such as advanced glycation end products, isoprostanes, C-reactive protein, or adhesion molecules. These analyses were beyond the scope of our study question and should be evaluated in a larger cohort. Finally, inclusion of only children with type 1 diabetes in this analysis does not allow us to determine whether NOS3 polymorphisms are associated with variations in arterial stiffness in the general population as well as in the subset with type 1 diabetes. Additional studies with normal control subjects are needed to determine whether NOS3 polymorphisms are associated with differences in vascular function across entire populations.
Despite these limitations, the potential clinical relevance of our findings should be considered. Our data demonstrate that traditional risk factors for endothelial dysfunction and CVD such as blood pressure and lipoprotein profiles may not contribute meaningfully to the AI75. Furthermore, it is becoming increasingly evident that traditional risk factors do not adequately predict CVD risk in all patient populations (e.g., children, women, and African Americans), and the paucity of algorithms to stratify risk in children mandates a need for better prognostic tools. We found that typing of the NOS3 promoter polymorphism, along with consideration of age, sex, and BMI percentile, explained 40% of the observed variability in arterial stiffness in children with type 1 diabetes. Although the association of NOS3 polymorphisms with the AI75 does not establish a cause-effect relationship, these findings should be tested in a larger cohort with additional genetic and protein biomarkers. In terms of clinical application, it is foreseeable that genotyping of the NOS3 SNP could add to the predictive power of risk assessment tools currently available and could perhaps guide clinicians’ approaches to the medical management of individuals with type 1 diabetes.
AI75 by NOS3 −786T→C genotype. Horizontal line, median; upper box edge, 75th percentile; lower box edge, 25th percentile; box height, interquartile range; whiskers, 1.5 times the interquartile range. P = 0.01 by ANOVA.
AI75 by NOS3 −786T→C genotype. Horizontal line, median; upper box edge, 75th percentile; lower box edge, 25th percentile; box height, interquartile range; whiskers, 1.5 times the interquartile range. P = 0.01 by ANOVA.
Baseline Characteristics
Clinical Variable . | Overall (n = 36) . | −786TT (n = 17) . | −786C Carriers (n = 19) . | Glu298Glu (n = 19) . | 298Asp Carriers (n = 17) . |
---|---|---|---|---|---|
Age (years) | 15 ± 2.9 | 15 ± 2.2 | 15 ± 3.4 | 15 ± 2.6 | 14 ± 3.2 |
Female (%) | 58 | 47 | 68 | 63 | 53 |
White (%) | 86 | 88 | 84 | 84 | 88 |
BMI percentile | 69 ± 25 | 74 ± 22 | 64 ± 26 | 68 ± 26 | 71 ± 24 |
Systolic blood pressure (mmHg) | 110 ± 10 | 112 ± 8 | 109 ± 12 | 110 ± 11 | 111 ± 11 |
Diastolic blood pressure (mmHg) | 70 ± 10 | 71 ± 9 | 68 ± 10 | 69 ± 9 | 70 ± 10 |
Total cholesterol (mmol/l) | 4.38 ± 0.93 | 4.27 ± 0.93 | 4.47 ± 0.93 | 4.01 ± 0.75 | 4.78 ± 0.96* |
LDL cholesterol (mmol/l) | 2.33 ± 0.70 | 2.17 ± 0.67 | 2.46 ± 0.70 | 2.02 ± 0.57 | 2.66 ± 0.65* |
HDL cholesterol (mmol/l) | 1.58 ± 0.36 | 1.58 ± 0.41 | 1.60 ± 0.28 | 1.55 ± 0.47 | 1.63 ± 0.31 |
Triglycerides (mmol/l) | 1.03 ± 0.70 | 1.13 ± 0.73 | 0.93 ± 0.68 | 0.98 ± 0.62 | 1.07 ± 0.79 |
A1C (%) | 8.4 ± 1.3 | 8.2 ± 1.5 | 8.6 ± 1.0 | 8.1 ± 1.5 | 8.7 ± 0.8 |
AI75 | 4.0 ± 13 | -1.7 ± 14 | 9.1 ± 11 | 5.1 ± 14 | 2.8 ± 12 |
Clinical Variable . | Overall (n = 36) . | −786TT (n = 17) . | −786C Carriers (n = 19) . | Glu298Glu (n = 19) . | 298Asp Carriers (n = 17) . |
---|---|---|---|---|---|
Age (years) | 15 ± 2.9 | 15 ± 2.2 | 15 ± 3.4 | 15 ± 2.6 | 14 ± 3.2 |
Female (%) | 58 | 47 | 68 | 63 | 53 |
White (%) | 86 | 88 | 84 | 84 | 88 |
BMI percentile | 69 ± 25 | 74 ± 22 | 64 ± 26 | 68 ± 26 | 71 ± 24 |
Systolic blood pressure (mmHg) | 110 ± 10 | 112 ± 8 | 109 ± 12 | 110 ± 11 | 111 ± 11 |
Diastolic blood pressure (mmHg) | 70 ± 10 | 71 ± 9 | 68 ± 10 | 69 ± 9 | 70 ± 10 |
Total cholesterol (mmol/l) | 4.38 ± 0.93 | 4.27 ± 0.93 | 4.47 ± 0.93 | 4.01 ± 0.75 | 4.78 ± 0.96* |
LDL cholesterol (mmol/l) | 2.33 ± 0.70 | 2.17 ± 0.67 | 2.46 ± 0.70 | 2.02 ± 0.57 | 2.66 ± 0.65* |
HDL cholesterol (mmol/l) | 1.58 ± 0.36 | 1.58 ± 0.41 | 1.60 ± 0.28 | 1.55 ± 0.47 | 1.63 ± 0.31 |
Triglycerides (mmol/l) | 1.03 ± 0.70 | 1.13 ± 0.73 | 0.93 ± 0.68 | 0.98 ± 0.62 | 1.07 ± 0.79 |
A1C (%) | 8.4 ± 1.3 | 8.2 ± 1.5 | 8.6 ± 1.0 | 8.1 ± 1.5 | 8.7 ± 0.8 |
AI75 | 4.0 ± 13 | -1.7 ± 14 | 9.1 ± 11 | 5.1 ± 14 | 2.8 ± 12 |
Data presented as means±SD unless otherwise indicated;
P ≤ 0.01 between Glu298Glu homozygotes and 298Asp carriers.
Multivariable analysis of AI75
Variable . | Coefficient . | SE . | P value . |
---|---|---|---|
Constant | 17.325 | 11.905 | 0.156 |
Age | −1.392 | 0.637 | 0.037 |
Sex | 8.384 | 3.777 | 0.034 |
BMI percentile | −0.144 | 0.077 | 0.072 |
NOS3 −786C carrier | 7.237 | 3.796 | 0.066 |
Variable . | Coefficient . | SE . | P value . |
---|---|---|---|
Constant | 17.325 | 11.905 | 0.156 |
Age | −1.392 | 0.637 | 0.037 |
Sex | 8.384 | 3.777 | 0.034 |
BMI percentile | −0.144 | 0.077 | 0.072 |
NOS3 −786C carrier | 7.237 | 3.796 | 0.066 |
Model P = 0.002; R2=0.40.
−786T→C and Glu298Asp diplotype frequencies
Diplotype . | N (36) . | % . |
---|---|---|
1/1 | 12 | 33 |
1/4 | 8 | 22 |
1/2 | 5 | 14 |
1/3 | 4 | 11 |
2/2 | 2 | 5.6 |
2/4 | 2 | 5.6 |
4/4 | 1 | 2.8 |
3/4 | 1 | 2.8 |
3/3 | 1 | 2.8 |
Diplotype . | N (36) . | % . |
---|---|---|
1/1 | 12 | 33 |
1/4 | 8 | 22 |
1/2 | 5 | 14 |
1/3 | 4 | 11 |
2/2 | 2 | 5.6 |
2/4 | 2 | 5.6 |
4/4 | 1 | 2.8 |
3/4 | 1 | 2.8 |
3/3 | 1 | 2.8 |
1, −786T/Glu298; 2, −786C/Glu298; 3, −786T/Asp298; 4, −786C/Asp298.
Comparison of single SNP- versus haplotype-containing regression models
SNP/haplotype . | R2 . | P value . |
---|---|---|
−786T→C* | 40% | 0.002 |
Glu298/−786T | 35% | 0.008 |
Glu298/−786C | 37% | 0.005 |
Asp298/−786T | 36% | 0.007 |
Asp298/−786C | 35% | 0.009 |
SNP/haplotype . | R2 . | P value . |
---|---|---|
−786T→C* | 40% | 0.002 |
Glu298/−786T | 35% | 0.008 |
Glu298/−786C | 37% | 0.005 |
Asp298/−786T | 36% | 0.007 |
Asp298/−786C | 35% | 0.009 |
Reference model; all models include age, sex, and BMI percentile in addition to the SNP or haplotype.
Article Information
This study was supported by grants from the Diabetes Action Research and Education Foundation (DARE), the Children’s Miracle Network, Pfizer, the American Heart Association Florida/Puerto Rico Affiliate, and the National Institutes of Health (CO6 Grant RR17568, and GCRC MO1-RR00082).
We thank Jennifer Stein for obtaining radial tonometry data and Amy DeBella and Gregory Welder for laboratory assistance. We thank Dr. Michael A. Province for his thoughtful comments.
References
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.