Abstract
Background
Gouty arthritis (gout) is the most common inflammatory arthritis in the United States and several other countries. Some rare forms of gout have a known genetic basis, but the relative importance of genetic factors on the risk for the lifetime prevalence of gout is not clear.
Methods
We performed a heritability analysis for hyperuricemia and gout among 514 unselected, all-male twin pairs who were a part of the National Heart, Lung, and Blood Institute twin study, a prospective observational cohort study. Statistical analyses were performed using structural equation models and maximum likelihood methods. The covariates used for adjustment in the structural equation models were identified using bivariate logistic regressions.
Results
The study population included 253 monozygotic (MZ) and 261 dizygotic (DZ) twin pairs, aged 48 (±3) years at baseline and followed for a mean of 34 years. The lifetime prevalence of gout did not differ between MZ and DZ twins. The concordance of hyperuricemia was 53% in MZ and 24% in DZ twin pairs (P<.001). Models that quantified the relative contribution of genetic and environmental factors on phenotypic variance showed that individual variability in gout was substantially influenced by environmental factors shared between co-twins and not by genetic factors. In contrast, individual differences in hyperuricemia were influenced significantly by genetic factors.
Conclusion
Hyperuricemia is a genetic trait. Outside the context of rare genetic disorders, risk for gout is determined by the environment. This has implications for prevention and treatment approaches.
Keywords
Although there are specific dietary and lifestyle risk factors for hyperuricemia, their relative contributions are smaller than those entailed by genetic factors; genetic factors explain 40% to 80% of variation of serum urate concentration.
1
, 2
, 3
, 4
Hyperuricemia is required for the presence of gouty arthritis (gout), but only a small proportion of people with hyperuricemia develop clinical gout, suggesting that additional environmental or genetic factors are involved.5
, 6
Polymorphisms involving SLC2A9 are associated with the risk of gout, likely mediated through the protein product's ability to cause hyperuricemia.6
, 7
, 8
, 9
, 10
, 11
, 12
, 13
, 14
, 15
, 16
, 17
, 18
Other single-gene mutations, such as those associated with familial juvenile hyperuricemic nephropathy and medullary cystic kidney diseases, also are associated with severe hyperuricemia and phenotypic manifestations of urate overload, such as urate stones and gout.Clinical Significance
- •Hyperuricemia is mostly an inherited trait.
- •Hyperuricemia is necessary but not sufficient for occurrence of gout.
- •The risk for gout is a trait influenced more by environmental factors than any inherited factors.
- •This means many cases of gout can be preventable.
Few studies have examined the heritability of gout;
1
however, gout is known to cluster in families, with approximately 20% of those with gout reporting a family history of the disease.19
Such clustering could be due to shared genes or shared lifestyle and other environmental factors. Galen, the ancient Roman physician of the second century, astutely stated that gout is due to a “hereditary trait” and to “debauchery and intemperance.” The relative importance of nature and nurture in gout is unknown. We studied this question using data from a large twin cohort study conducted in North America. Twins are invaluable for studying this important question because they disentangle the sharing of genes and environments. The twin design compares the lifetime prevalence of gout among monozygotic (MZ) or identical twins, who share approximately 100% of their genetic polymorphisms, with that among dizygotic (DZ) or fraternal twins, who share approximately 50% of their genes. To our knowledge, this is the first twin study examining the relative genetic and environmental contributions to the lifetime prevalence of gout.Materials and Methods
All participants provided informed consent, and the original data collection was approved by institutional review boards of the participating sites. We used the data from the National Heart, Lung, and Blood Institute (NHLBI) twin study (NCT 00005124), presently housed at SRI International (Menlo Park, CA). The history and methodology of this cohort have been described.
20
, 21
, 22
The NHLBI twin study was initiated in 1955 by the National Academy of Sciences/National Research Council, which developed a twin registry of US Armed Forces veterans. Sixteen thousand white male twin pairs born between 1917 and 1927 identified in the Veterans Affairs Master Index were invited to participate in this registry, and 7000 expressed interest. The NHLBI Twin cohort with 514 twin pairs was assembled from the above group according to residence in California, New England States, or within 200 miles of Indianapolis, Indiana. There were no specific exclusion criteria, and all those who volunteered from these states were accepted. These individuals completed a study questionnaire, physically appeared for a study visit, and provided blood samples. Twins living in the New England area were examined at the Framingham Heart Study facilities from mid-1969 to mid-1970. The California twins were examined at the University of California, Davis facility, the Rancho Los Amigos Hospital, and the Mount Zion-Harold Braun Research Institute. Those in Indiana were examined at the University of Indiana School of Medicine. Subsequent to the baseline visit, there were 5 follow-up visits during the years 1981-1982, 1986-1987, 1995-1997, 1999-2001, and 2001-2003. Further follow-up and vital status assessment were performed in the year 2010.Zygosity was determined by combining information from 22 blood group antigens and the twins' self-reports. Antigens tested included A, A1, B, M, N, S, s, P, C, Cw, D, Du, E, c, e, K, Kpa, Kpb, k, Lea, Fya, and Jka. Reverse grouping on antigens A and B also was done to confirm ABO blood type. If twins were found discordant in any of these, they were deemed to be DZ. Of the 264 self-classified DZ pairs, 9.8% had identical blood group antigens; the statistically expected rate was approximately 3%.
20
Standardized anthropometry and blood pressure measurements were performed on all participants at the first examination conducted between 1969 and 1972. Fasting blood samples were obtained for lipid measures (low-density lipoprotein, high-density lipoprotein, triglycerides, total cholesterol) and urate. Samples from the first 3 study visits (1969-1972, 1981-1982, 1986-1987) were analyzed for serum urate. These data were not available for 6 twin pairs, and thus only 508 twin pairs were included in the analyses that used serum urate. The uricase method was used for all urate assays. Glucose tolerance was assessed by analysis of blood glucose levels 1 hour after a 50-g oral glucose load. A food frequency questionnaire was used to assess nutritional intake. By using the results from the Health Professional Follow-up study,
23
we calculated a high-risk diet score at the study start based on the number of servings of food items thought to increase the risk for incident gout (sum of the frequency of 10 food items, including meat, fish, and seafood; range 0-8 times/week). Lifetime prevalence of gout was assessed over the observation period by self-reported physician diagnosis or use of gout medications, such as allopurinol, probenecid, colchicine, and sulfinpyrazone. This method has been reported to have excellent sensitivity.24
Hyperuricemia was defined as a serum urate level greater than 416 μmol/L (7 mg/dL) at any of the 3 occasions serum urate was measured. Other data collected included use of diuretics, Thurstone physical activity index,25
and alcohol use. Bivariate logistic regression models were created using STATA version 10.0 (College Station, Tex) to identify covariates for use in the genetic models.By using data from MZ and DZ twin pairs reared together, information can be obtained about the degree to which behavioral traits are influenced by the relative contribution of genetic and environmental influences. The twin method rests on several theoretic assumptions. First, MZ twins are, on average, 100% genetically identical, whereas DZ twins share, on average, 50% of their genes. Therefore, the genetic relationship between MZ twins is perfectly correlated (r=1.0), and that between DZ twins is correlated one half (r = 0.5). Second, the prenatal and postnatal (family, social, cultural) environment for co-twins is more similar than that for siblings of different ages. It is therefore assumed that both MZ and DZ twin pairs are perfectly correlated vis-à-vis their environmental experiences. By comparing the degree to which MZ co-twins are similar for a given trait, relative to DZ co-twins, and using statistical methods of maximum likelihood, the proportion of the total phenotypic variance can be attributed to the additive effects of genes (A), which contribute to twin similarity, to shared environmental effects in common to co-twins (C), which also contribute to twin similarity, and to non-shared environmental effects or experiences unique to each twin that contribute to twin dissimilarity, which also includes measurement error variance (E). The expected correlation between MZ twins is rMZ=A+C, and that between DZ twins is rDZ=0.5 A+C. The rMZ/rDZ ratio should equal 2 if a trait is heritable with no significant influence of shared environmental effects (“AE” models). If both genetic and shared environmental effects contribute to trait similarity, and considering that shared environment is assumed to contribute equally to similarity of MZ and DZ twins, this would result in greater similarity (concordance) of DZ co-twins relative to MZ co-twins and the rMZ/rDZ ratio would be less than 2 (“ACE” models). If only shared environmental effects contribute to trait similarity, MZ and DZ twin pair correlations will be similar to each other and the rMZ/rDZ ratio will be approximately 1 (“CE” models). The relative contribution of genetic and environmental factors was quantified using the structural equation modeling software Mx.
26
Models were fit to raw data, and model fit was assessed using maximum likelihood methods.27
MZ and DZ twin pair concordance for gout lifetime prevalence and hyperuricemia were computed using tetrachoric correlations for dichotomous data in SAS software (SAS Institute Inc, Cary, NC).Results
The study subjects were twins enrolled in the NHLBI twin study (NCT 00005124). Table 1 provides the descriptive measures of the study cohort; 253 MZ and 261 DZ twin pairs were included. The baseline age range was 42 to 55 years. These individuals had been followed for a mean of 34 years. The last examination included 174 individuals (age range 76-86 years). The lifetime prevalence of gout (ie, prevalence anytime during observation in the study) did not differ significantly between MZ and DZ twins (11.9% and 11.5%, respectively). The tetrachoric correlations were 0.42 for MZ twin pairs (P=.01) and 0.48 for DZ twin pairs (P <.001 ). At the baseline examination, the mean serum urate in the gout group (7.05±1.4 mg/dL) was higher than in the group without gout (6.21±1.1 mg/dL).
Lifetime prevalence assessed at the end of follow-up as assessed by study physician or use of gout medications.
Table 1Descriptive Features of the Study Population
| Features | Overall Lifetime Prevalence (%) or Measured Levels (Mean±SD) | P Value for Difference between MZ and DZ Twins | ||
|---|---|---|---|---|
| All Subjects (n=1028) | MZ Twins (n=506) | DZ Twins (n=522) | ||
| Lifetime prevalence of gout | 11.60% | 11.9% | 11.5% | .8560 |
| Serum urate (mg/dL) | 6.31 (1.2) | 6.34 (1.2) | 6.28 (1.2) | .4447 |
| Hyperuricemia (serum urate ≥416 mmol/dL; 7 mg/dL) | 27.0% | 28.3% | 25.7% | .3493 |
| BUN, mg/dL | 16.6 (3.9) | 16.6 (3.9) | 16.5 (3.9) | .8116 |
| Renal dysfunction (BUN≥19) | 26.9% | 26.7% | 27.0% | .9045 |
| Age, y | 47.9 (3.2) | 47.8 (3.1) | 47.9 (3.2) | .5670 |
| Body mass index (kg/m2) | 25.7 (3.3) | 25.7 (3.2) | 25.8 (3.4) | .7831 |
| Systolic blood pressure (mm Hg) | 127.7 (16.7) | 128.6 (17.7) | 126.8 (15.7) | .0912 |
| Diastolic blood pressure (mm Hg) | 81.5 (10.9) | 82.3 (11.1) | 80.8 (10.7) | .0350 |
| Nondiuretic antihypertensive use | 5.5% | 5.9% | 5.2% | .5962 |
| Daily alcohol use | 28.6% | 31.6% | 25.7% | .0348 |
| Diuretic use | 3.3% | 3.4% | 3.3% | .9265 |
| Diabetes lifetime prevalence | 8.6% | 9.9% | 7.3% | .1361 |
| High daily protein consumption (≥88 g/d) | 22.0% | 22.9% | 21.1% | .4735 |
| High-risk gout diet | 27.5% | 28.7% | 26.4% | .4258 |
BUN=blood urea nitrogen; SD=standard deviation.
† Serum urate was available for only 508 twin pairs (n=1016).
‡ Defined as physician assessment during the study examination or adherence to a diabetic diet.
§ High-risk diet defined as consuming an average of 5 servings/week of any of the 5 food items (meats, fish, and seafood) known to be associated with higher risk of gout per the Health Professionals Study.
23
Genetic analyses for gout are shown in Table 2. Analyses using an alternate case definition of gout, excluding medication use as a criterion, did not change the results. The rMZ/rDZ ratio was approximately 1, suggesting that individual differences in gout lifetime prevalence are influenced by shared and non-shared environmental factors and that genetic factors are not important contributors. As the environmental “CE” model shows, 45.1% of the phenotypic variance was attributable to shared environmental effects with the remaining variance attributable to individual-specific factors. The shared environmental component was significant, because dropping it from the model (ie, equating it to zero) resulted in significant deterioration of model fit (Appendix). After adjustment for the effects of environmental covariates and hyperuricemia, the contribution of shared environmental factors did not change, suggesting that these covariates do not explain the environmental variation in gout. Statistical power computations showed that to detect a magnitude of C of 20%, 30%, or 40%, given the prevalence of gout and the sample size, power was 44%, 77.6%, and 97.3%, respectively.
Gouty arthritis assessed by study physician or use of gout medications.
Table 2Unadjusted and Adjusted Estimates of the Relative Contribution of Genetic (A), Shared Environmental (C), and Non-shared Environmental (E) Influences on Gout Incidence Among 514 Male Twin Pairs
| Model | A (95% CI) | C (95% CI) | E (95% CI) | −2 LL | df | AIC | −2 LL diff | df diff | P diff |
|---|---|---|---|---|---|---|---|---|---|
| Gout | |||||||||
| ACE | 0 (0-58.1) | 45.1 (0-61.4) | 54.9 (35.2-73.8) | 720.723 | 1025 | −1329.277 | |||
| CE | 45.1 (26.2-61.4) | 54.9 (38.6-73.8) | 720.723 | 1026 | −1331.277 | 0.000 | 1 | − | |
| AE | 50.7 (27.6-69.5) | 49.3 (30.5 (72.4) | 724.059 | 1026 | −1327.941 | 3.336 | 1 | .068 | |
| E | 100 | 740.904 | 1027 | −1313.096 | 20.181 | 2 | .000 | ||
| Gout+covariates | |||||||||
| ACE | 0 (0-61.6) | 45.9 (0-62.8) | 54.1 (33.5-73.9) | 689.431 | 1020 | −1350.569 | |||
| CE | 45.9 (26.1-62.8) | 54.1 (37.2-73.9) | 689.431 | 1021 | −1352.569 | 0 | 1 | − | |
| AE | 51.6 (27.5-70.8) | 48.4 (29.2-72.5) | 692.395 | 1021 | −1349.605 | 2.964 | 1 | .085 | |
| E | 100 | 708.419 | 1022 | −1335.581 | 18.989 | 2 | .000 | ||
| Gout + covariates + hyperuricemia | |||||||||
| ACE | 0 (0-61.8) | 40.5 (0-58.8) | 59.5 (36.3-80.7) | 669.652 | 1019 | −1368.348 | |||
| CE | 40.5 (19.3-58.8) | 59.5 (41.2-80.7) | 669.652 | 1020 | −1370.348 | 0 | 1 | − | |
| AE | 46.1 (20.1-67.3) | 53.9 (32.7-79.9) | 671.549 | 1020 | −1368.451 | 1.897 | 1 | .168 | |
| E | 100 | 682.958 | 1021 | −1359.402 | 13.305 | 2 | .001 |
−2 LL=log likelihood goodness-of-fit test; df=degrees of freedom; AIC=Akaike Information Criterion; −2 LL diff=difference of the log likelihoods between the full ACE model and the genetic (AE) or environmental (CE, E) submodels; df diff=degrees of freedom difference; P diff=P value associated with the log likelihood difference between the full and a nested model; uncalc=P value could not be calculated for a log likelihood difference of zero.
† Covariates emerging significant in bivariate logistic regression models: age, body mass index, antihypertensive or diuretic medications, renal dysfunction, and diabetes.
‡ Serum urate was available for only 508 twin pairs (n=1016); hyperuricemia: serum urate concentration≥7.0 mg/dL.
For analyses of hyperuricemia, data were available on 508 twin pairs (n=1016). There were 411 participants with hyperuricemia (40.5%). The mean serum rate concentrations of MZ and DZ twin pairs were 6.34 (±1.2) and 6.28 (±1.2) mg/dL, respectively. The tetrachoric correlations for hyperuricemia were 0.53 for MZ twin pairs (P <.0001) and 0.24 for DZ twin pairs (P=.0257). Genetic analyses for hyperuricemia are shown in Table 3. The rMZ/rDZ ratio of the tetrachoric correlations was just more than 2, suggesting that shared environmental factors are not important contributors to the phenotypic variance. As the genetic “AE” model shows, 60.0% of the phenotypic variance in hyperuricemia was attributable to heritable factors, with the remaining variance attributable to non-shared environmental influences. The genetic component was significant because dropping it from the model resulted in significant deterioration of model fit. The heritability estimate diminished somewhat to 49.6% after covariate adjustment, but remained significant. The slight decrease in the heritability magnitude suggests that some of the covariates account for some of the genetic factors influencing hyperuricemia, possibly because of common genetic risk. Post-hoc statistical power computations showed that to detect a magnitude of A of 20%, 30%, or 40%, given the prevalence of hyperuricemia and the sample size in the present study, power was 63.6%, 93.8%, and 99.8%, respectively.
Serum urate was available for only 508 twin pairs (n=1016); hyperuricemia: serum urate concentration≥7.0 mg/dL.
Table 3Unadjusted and Adjusted Estimates of the Relative Contribution of Genetic (A), Shared Environmental (C), and Non-shared Environmental (E) Influences on Hyperuricemia
| Model | A (95% CI) | C (95% CI) | E (95% CI) | −2 LL | df | AIC | −2 LL diff | df diff | P diff |
|---|---|---|---|---|---|---|---|---|---|
| Hyperuricemia | |||||||||
| ACE | 60.0 (0-65.7) | 0 (0-41.0) | 49.0 (34.4-67.1) | 1157.566 | 1025 | −882.434 | |||
| CE | 38.7 (24.6-51.7) | 61.3 (48.4-75.4) | 1171.425 | 1026 | −880.575 | 3.859 | 1 | .049 | |
| AE | 60.0 (34.0-65.7) | 49.0 (34.4-66.0) | 1167.566 | 1026 | −884.434 | 0.000 | 1 | − | |
| E | 100 | 1198.065 | 1027 | −855.935 | 30.498 | 2 | .000 | ||
| Hyperuricemia + covariates | |||||||||
| ACE | 49.6 (2.3-65.2) | 0 (0-37.1) | 50.4 (34.8-69.1) | 1108.731 | 1020 | −931.269 | |||
| CE | 36.4 (21.5-50.0) | 63.6 (50.0-78.5) | 1112.932 | 1021 | −929.068 | 4.200 | 1 | .040 | |
| AE | 49.6 (31.4-65.2) | 50.4 (34.8-68.6) | 1108.73 | 1021 | −933.269 | 0.000 | 1 | − | |
| E | 100 | 1134.473 | 1022 | −909.527 | 25.742 | 2 | .000 |
−2 LL=log likelihood goodness-of-fit test; CI=confidence interval; df=degrees of freedom; AIC=Akaike Information Criterion; −2 LL diff=difference of the log likelihoods between the full ACE model and the genetic (AE) or environmental (CE, E) submodels; df diff=degrees of freedom difference; P diff=P value associated with the log likelihood difference between the full and a nested model; uncalc=P value could not be calculated for a log likelihood difference of zero.
† Covariates emerging significant in bivariate logistic regression models: age, body mass index, antihypertensive or diuretic medications, renal dysfunction, and diabetes.
Discussion
Our results confirm the previously reported strong heritability of hyperuricemia but suggest that environmental factors are more important in the phenotypic expression of gout. The difference in heritability between hyperuricemia and gout is not entirely surprising. Hyperuricemia, a critical predisposing factor for gout, is invariably present in gout but not vice versa. Vitart et al
6
reported that only 10% of those with hyperuricemia eventually develop gout. The Normative Aging Study suggests that even with a high urate concentration of more than 9 mg/dL, the 5-year cumulative incidence rate of gout is only 22%.5
Alcohol and obesity have been reported to be risk factors for gout independently of hyperuricemia.28
Hyperuricemia has a strong genetic basis, although additional factors such as insulin resistance/diabetes, the obesity-inactivity dyad, hypertension, and renal dysfunction are known to be independent risk factors or pathway factors.
29
We previously observed that hyperuricemia runs as a “trait” among the participants of the Coronary Artery Development in Young Adults cohort (ie, serum urate concentrations track over years).30
Prior twin studies in age groups ranging from young adults to octogenarians support the notion that hyperuricemia also has a major genetic basis. For example, urinary urate excretion has a heritability coefficient of 50% to 96%.31
, 32
, 33
Genetic factors can alter purine metabolism (HPRT and PRPS enzyme mutations) associated with excessive cell death (glycogen storage disease and hemoglobinopathies) or reduced renal urate clearance (uromodulin mutations). The genes that are responsible for hyperuricemia caused by reduced renal excretion of urate are localized on chromosome 16.34
The important caveats to our observations are as follows. First, the prevalence of gout in this population might appear high; however, this rate is consistent with recent estimates from the National Health and Nutrition Examination Survey in the corresponding age group (65+ years, 9.8%).
35
Second, estimates for the genetic and environmental contribution to hyperuricemia and gout, respectively, are likely to be underestimates because of the potential biochemical measurement errors engendered by wide intervals between study examinations. A significant part of the unexplained variance is likely to arise from factors other than those adjusted for in our models. Third, the role of gender could not be assessed in this all-male study. The fractional urate excretion is high among children and similar in both boys and girls.36
Post-puberty, urate excretion decreases substantially especially among men. A study of data from the Danish Twin Registry suggested that the relative contributions of genes and environment for hyperuricemia may be different between men and women,37
and this conclusion was supported by a genome-wide association study.6
Potential for misclassification of gout diagnosis exists, but this is likely to be non-differential between MZ and DZ twins. The case definition we used has been noted to be sensitive and has been validated in other epidemiologic studies.23
, 24
, 38
Some degree of misclassification error is possible, but it is likely to be non-differential across twin pairs. Competing causes of mortality can become a significant factor in long observational studies such as ours. Hyperuricemia is known to be associated with higher cardiovascular risk, and many individuals might have died, precluding development of clinical gout.Conclusions
Our analysis of data collected on 514 male twin pairs through the NHLBI twin study suggests that the phenotype of gout is determined primarily by environmental factors. This has implications for prevention and treatment of the disease. Future larger twin studies must be performed to assess whether there is a detectable heritable component to gout and whether our observations are valid among women.
Acknowledgements
This work was supported by grant HL51429 from the National Heart, Lung, and Blood Institute. We acknowledge Dorit Carmelli, PhD (formerly of SRI International), Terry Reed, PhD (Indiana University Medical Center), Philip A. Wolf, MD (Boston University Medical Center) and Bruce L. Miller, MD (Harbor/UCLA Medical Center) for their rigor in overseeing data collection at their research sites.
References
- Recent insights into the pathogenesis of hyperuricaemia and gout.Hum Mol Genet. 2009; 18: R177-R184
- Genetics of gout.Curr Opin Rheumatol. 2010; 22: 144-151
- Pathogenesis of gout.Ann Intern Med. 2005; 143: 499-516
- New insights into gout epidemiology.Curr Opin Rheumatol. 2006; 18: 199-203
- Asymptomatic hyperuricemia.Am J Med. 1987; 82: 421-426
- SLC2A9 is a newly identified urate transporter influencing serum urate concentration, urate excretion and gout.Nat Genet. 2008; 40: 437-442
- Plasma urate level is directly regulated by a voltage-driven urate efflux transporter URATv1 (SLC2A9) in humans.J Biol Chem. 2008; 283: 26834-26838
- A genome-wide association study of serum uric acid in African Americans.BMC Med Genomics. 2011; 4: 17
- Genome-wide scan identifies a quantitative trait locus at 4p15.3 for serum urate.Eur J Hum Genet. 2010; 18: 1243-1247
- Association of three genetic loci with uric acid concentration and risk of gout: a genome-wide association study.Lancet. 2008; 372: 1953-1961
- Homozygous SLC2A9 mutations cause severe renal hypouricemia.J Am Soc Nephrol. 2010; 21: 64-72
- SLC2A9 influences uric acid concentrations with pronounced sex-specific effects.Nat Genet. 2008; 40: 430-436
- Meta-analysis of 28,141 individuals identifies common variants within five new loci that influence uric acid concentrations.PLoS Genet. 2009; 5: e1000504
- Mutations in glucose transporter 9 gene SLC2A9 cause renal hypouricemia.Am J Hum Genet. 2008; 83: 744-751
- Uric acid transport and disease.J Clin Invest. 2010; 120: 1791-1799
- A ‘complexity’ of urate transporters.Kidney Int. 2010; 78: 446-452
- Association of a common nonsynonymous variant in GLUT9 with serum uric acid levels in old order amish.Arthritis Rheum. 2008; 58: 2874-2881
- Genome-wide association study identifies genes for biomarkers of cardiovascular disease: serum urate and dyslipidemia.Am J Hum Genet. 2008; 82: 139-149
- Heredity of gout and hyperuricemia.Arthritis Rheum. 1965; 8: 627-647
- The NHLBI twin study of cardiovascular disease risk factors: methodology and summary of results.Am J Epidemiol. 1977; 106: 284-285
- Correlations of alcohol consumption with related covariates and heritability estimates in older adult males over a 14- to 18-year period: the NHLBI Twin Study.Alcohol Clin Exp Res. 1994; 18: 702-710
- Heritability of clinical chemistries in an older twin cohort: the NHLBI Twin Study.Genet Epidemiol. 1987; 4: 1-11
- Purine-rich foods, dairy and protein intake, and the risk of gout in men.N Engl J Med. 2004; 350: 1093-1103
- Reliability and sensitivity of the self-report of physician-diagnosed gout in the campaign against cancer and heart disease and the atherosclerosis risk in the community cohorts.J Rheumatol. 2011; 38: 135-141
- The structured interview and questionnaire methods of assessing coronary-prone behavior in male and female college students.J Behav Med. 1979; 2: 71-83
- (Neale MC BS)Mx: Statistical Modeling.6th ed. Department of Psychiatry, Virginia Commonwealth University, Richmond VA2002
- The genetical analysis of covariance structure.Heredity. 1977; 38: 79-95
- Community based epidemiological study on hyperuricemia and gout in Kin-Hu, Kinmen.J Rheumatol. 2000; 27: 1045-1050
- Gout in 2006: the perfect storm.Bull NYU Hosp Jt Dis. 2006; 64: 82-86
- Fluctuations in serum urate concentrations over 10 years.Ann Rheum Dis. 2010; 69: 651
- Genetic heritability of urinary stone risk in identical twins.J Urol. 2006; 175: 2125-2128
- Genetic control of the renal clearance of urate: a study of twins.Ann Rheum Dis. 1992; 51: 375-377
- Heritabilities for fifteen routine biochemical values: findings in 215 Swedish twin pairs 82 years of age or older.Scand J Clin Lab Invest. 2009; 69: 562-569
- Localization of a gene for familial juvenile hyperuricemic nephropathy causing underexcretion-type gout to 16p12 by genome-wide linkage analysis of a large family.Arthritis Rheum. 2000; 43: 925-929
- Prevalence of gout and hyperuricemia in the US general population.Arthritis Rheum. 2011; 63: 3136-3141
- Presymptomatic detection of familial juvenile hyperuricaemic nephropathy in children.Pediatr Nephrol. 1998; 12: 357-364
- Heritability of biochemical kidney markers and relation to survival in the elderly–results from a Danish population-based twin study.Clin Chim Acta. 2004; 349: 143-150
- Gout and risk for subsequent coronary heart disease.Arch Intern Med. 1997; 157: 1436-1440
Article info
Publication history
Published online: February 23, 2012
Footnotes
Funding: This study was supported by NHLBI grant HL51429.
Conflict of Interest: None.
Authorship: All authors had access to the data and played a role in writing this manuscript.
Clinical trials.gov: NCT 00005124.
Identification
Copyright
© 2012 Elsevier Inc. Published by Elsevier Inc. All rights reserved.
