The Journals of Gerontology Series A: Biological Sciences and Medical Sciences 57:M660-M664 (2002)
© 2002 The Gerontological Society of America
Effect of Aging on Serum Uric Acid Levels
Longitudinal Changes in a Large Japanese Population Group
Masafumi Kuzuyaa,
Fujiko Andob,
Akihisa Iguchia and
Hiroshi Shimokatab
a Department of Geriatrics, Nagoya University Graduate School of Medicine, Japan
b Department of Epidemiology, National Institute for Longevity Sciences, Aich, Japan
Masafumi Kuzuya, Department of Geriatrics, Nagoya University Graduate School of Medicine, 65 Tsuruma-cho, Showa-ku, Nagoya 466-8550 Japan E-mail: kuzuya{at}med.nagoya-u.ac.jp.
 |
Abstract
|
|---|
Background. Serum uric acid (SUA) is related not only to an increased risk of gout, but also to an increased risk of cardiovascular diseases. However, real age-related changes in SUA remain unknown.
Methods. Longitudinal population-based study of epidemiological follow-up data of SUA, body mass index (BMI), and alcohol intake was conducted at a health examination center between 1989 and 1998. The subjects were 80,506 Japanese office workers or their families (50,157 men and 30,349 women) with an average age of 44.5 years for the men and 43.7 years for the women.
Results. SUA increased with age in all birth cohorts examined in men, and in women except for the youngest birth cohort (1960–1969). BMI and alcohol consumption positively contributed to the longitudinal changes of SUA. However, SUA also increased with age in the model controlled for BMI and alcohol consumption. There were birth cohort effects of SUA; at most ages, there were higher SUA levels in younger cohorts in men and lower SUA levels in younger cohorts in women, respectively.
Conclusions. SUA levels in men and women increased with advancing age, despite changes in drinking and in the BMI. There are birth cohort effects for SUA levels in the Japanese population.
HYPERURICEMIA is not only a risk factor for gout, but also for cardiovascular diseases (1)(2). However, there is controversy over whether or not serum uric acid (SUA) is an independent risk factor for cardiovascular diseases and cardiovascular mortality. Some studies have found hyperuricemia to be an independent risk factor for cardiovascular diseases (3)(4)(5)(6)(7)(8), while others have concluded that the association is confounded by the relationship of uric acid with established risk factors for cardiovascular diseases, such as hypertension, obesity, hyperlipidemia, or diabetes mellitus (9)(10)(11).
The serum level of uric acid is affected by aging and genetic and environmental factors. Obesity and alcohol consumption are well established as determinants of SUA levels (1)(2). Men have higher SUA levels than women. Based on cross-sectional studies, the age-related distributions of SUA levels are not always in agreement in different population-based studies (3)(4)(5)(6)(7)(8)(9)(10)(11). Cross-sectional age-related changes in SUA levels may represent cohort, period, and/or survivorship effects rather than a true aging effect. Although longitudinal studies are required to examine real age-related changes in SUA, to our knowledge, no study has examined the longitudinal changes in the SUA levels in individuals over time.
We studied longitudinal changes in SUA in a single cohort of individuals to observe the effect of the natural aging process on SUA.
|
Methods
|
|---|
Study Population
The study population consisted of office workers and their families residing in Aichi Prefecture in the central region of Japan (12)(13)(14). The subjects included 80,506 Japanese (50,157 men and 30,349 women) with an average age of 44.5 years for the men and 43.7 years for the women, who received annual examinations at a health examination center in Japan between 1989 and 1998 (Table 1 ). Subjects who were taking medication for hyperuricemia were excluded. Our cohort included more men than women, because the number of male workers in Japan is greater than the number of female workers. About 57% of the cohort attended at least one follow-up examination. The average number of visits made for the follow-up examinations were 3.0 times for the men and 2.6 times for the women.
Procedures
The examinations included a questionnaire, a physical examination, an anthropometric measurement, and laboratory analysis of blood samples, all taken on the same day. The anthropometric measurements included height and body weight, which were measured while the subject was wearing light clothing without shoes. The body mass index (BMI) was calculated as weight/height2 (kg/m2).
To evaluate alcohol consumption, we administered a self-reported questionnaire about alcohol intake that included items about the type of alcoholic beverage consumed, the weekly frequency of alcohol consumption, and the usual amount consumed daily. Alcohol intake was converted to an average weekly or daily consumption measured by the volume unit called go (a traditional Japanese unit of measurement for sake, corresponding to 23 g of ethanol), which was then converted to grams of ethanol per week or day. The subjects were classified into six categories according to their reported alcohol consumption: (i) almost no drinking, (ii) past drinker, (iii) less than 23 g of ethanol per week, (iv) about 23 g of ethanol every day, (v) about 46 g ethanol every day, (vi) about 69 g of ethanol or more every day. All serum samples were obtained following a 12- to 14-hour fast. The serum was separated promptly, and all analyses were conducted at the clinical laboratory in the health examination center. The SUA was measured by the Uricase method.
Data Analysis
Data were processed and analyzed using SAS version 6.12 (SAS Institute, Cary, NC) (15). Longitudinal changes in SUA levels were analyzed using a mixed-effect model (16)(17), which is a type of statistical analysis commonly used for repeated measurements. It is applied using the SAS procedure PROC MIXED, typically using the REPEATED statement. Responses from points close in time usually are more highly correlated with each other than responses from points far apart in time, and, therefore, special methods of analysis are usually needed to accommodate the correlation structure of the repeated measurements. This autoregressive structure was controlled in the model using the type = AR(1) option. The least-square means for SUA and BMI values at 50 years of age in each cohort were determined. The differences in the estimated SUA and BMI levels within the birth cohort were tested using the Tukey multiple comparison method (ADJUST option), and the trend in cohort difference was also tested using the CONTRAST option of the PROC MIXED.
|
Results
|
|---|
Fig. 1 shows the age-specific means and a 3-year moving average of the SUA levels and BMI at the initial examination of our cohort. The SUA levels in men declined slightly from age 20 up to the seventh decade of life, whereas in women, the levels declined slightly from 25 years to about age 40, after which they rose up to the seventh decade of life. BMI levels increased rapidly from 25 to 35 in men, and then reached a plateau, after which the curve declined slightly to the seventh decade. In women, however, after the decrease at age 25 to 30, the BMI levels rapidly increased up to 65, followed by a decline to the seventh decade.
View larger version (17K):
[in this window]
[in a new window]
|
Figure 1. The age-specific means and a 3-year moving average of the SUA levels and body mass index (BMI) at the initial examination of our cohort.
|
|
Longitudinal changes in the levels of SUA and BMI by birth cohort in men and women indicated the apparent birth cohort effect in the SUA levels and BMI for men (Fig. 2), because, at most ages, the SUA and BMI values of younger birth cohorts were higher than those of older birth cohorts. The shape of these curves is also informative with regard to the age effect. In this case, the SUA levels increased with age from the second decade of life in men and the fourth in women. With the BMI levels, a significant age effect was observed for men from the second up to the fourth decade of life. For women, however, the BMI levels increased with age, up to 60 years.
View larger version (38K):
[in this window]
[in a new window]
|
Figure 2. Longitudinal changes in the levels of serum uric acid (SUA) and body mass index (BMI) by birth cohort in men and women. Mean SUA and BMI levels are shown by age in five birth cohorts.
|
|
Table 2 examines the mean change in SUA levels for 1 year by birth cohort and gender using repeated measurements during 10 years on the same individuals. During the interval, SUA increased significantly with age in men from all the birth cohorts and in women from the birth cohorts between the 1920s and 1950s, while SUA levels decreased in women from the youngest birth cohort.
The SUA and BMI levels at 50 years of age in men and women were estimated for each birth cohort using the least-square means methods in the mixed effects model. Tukey multiple comparison showed significant differences in the estimated SUA and BMI levels among the birth cohorts (Table 3 ). In men, the younger cohorts showed the higher SUA and BMI values at 50 years of age (trend p < .0001). In contrast, in women, the younger cohorts showed lower SUA and BMI values at 50 years than the older cohorts (trend p < .0001).
A mixed effect model appropriate for repeated measures was used to explore the effects of birth cohort, aging in each birth cohort, BMI, and alcohol consumption on longitudinal changes in SUA levels (Table 4 ). SUA levels of individual subjects over time during 10 years were significantly associated with birth cohort, aging of most of the birth cohort except for the birth cohort of the 1950s for men in a mixed effect model adjusting for BMI and alcohol consumption. In addition, longitudinal changes in SUA levels were also significantly and positively associated with BMI, as well as the alcohol consumption of individual subjects.
View this table:
[in this window]
[in a new window]
|
Table 4. Longitudinal Analysis of the Effect of Aging, Body Mass Index (BMI), and Alcohol Consumption on Serum Uric Acid Levels by Mixed Effect Model
|
|
|
Discussion
|
|---|
Cross-sectional age-related distributions of SUA are controversial in different cohorts and decades. There was a clear increase in SUA level with age in French men between the ages of 20 and 55, based on a survey conducted between 1967 and 1970 (18). The study conducted in the United States from 1959 to 1960 reported that the SUA values in men rose rapidly to a peak level at 20 to 24 years, after which they declined slightly, and then reached a plateau, except for a slight later rise at 55 to 59 years (19). In women from the same study, however, SUA levels rose to a minor peak at 15 to 19 years, after which they declined slightly, and leveled out to about age 40. From about 40 to 54 years, the curves rose slowly again to peak levels at 50 to 54, and later at 60 to 64 years. In the Framingham Heart Study, no significant age-related changes of SUA levels were observed between ages 20 and 80 in men (10). In women, the SUA levels gradually increased from the fourth to the seventh decades of life.
In our cohorts, the SUA levels in men declined slightly from age 20 up to the seventh decade of life, whereas in women, the levels declined slightly from 25 years and leveled out to about age 40, after which they rose up to the seventh decade of life. These inconsistent, age-related distributions of SUA levels in different population-based samples seem to be due to the fact that SUA levels in individuals are affected by genetic, biological, and environmental factors. In fact, in tri-partite, comparative epidemiological studies of cardiovascular diseases among Japanese men, SUA levels were significantly higher in Japanese-American individuals living in Hawaii and California than in the indigenous Japanese population, although the genetic backgrounds of these Japanese populations were essentially the same (20). This indicates that environmental changes are likely to be responsible for such differences.
Cross-sectional age-related changes in the SUA level may represent cohort, period, and/or survivorship effects rather than a true aging effect. Longitudinal study is necessary to show true age-related changes.
In contrast to a cross-sectional study, we demonstrated that longitudinal data produced inconsistent results regarding age-associated changes in SUA levels, especially in men. The increase in SUA levels with age was observed in all the birth cohorts of men and in women, except for the youngest birth cohort. In contrast, the obvious increase in SUA levels with age for women in the longitudinal analysis was similar to the cross-sectional study showing that SUA levels gradually increased from the fourth to the seventh decades of life. This increase seems to be associated with menopause, as suggested by other investigators (8)(19).
A mixed effect model analysis suggested that BMI values and alcohol consumption positively contributed to the longitudinal changes in the SUA levels of individual subjects. However, we clearly showed that a significant positive association of SUA levels and aging exists in older birth cohorts for men and women, even adjusted for BMI values and alcohol consumption. These results indicated that the longitudinal change in the SUA level in individual subjects was positively related to BMI as well as to alcohol consumption. However, aging still has a positive association with SUA levels for men and women in older cohorts, even after controlling for BMI and alcohol intake. It is uncertain whether aging itself is an independent correlate of the longitudinal change of SUA levels, or if other confounding factors may be related to SUA levels.
Our longitudinal analysis also indicated a birth cohort effect of SUA and BMI levels in men as well as in women, although the effect was opposite between the two. Higher estimated SUA and BMI levels at 50 years old were observed in younger birth cohorts than in older cohorts in men. In contrast, higher SUA and BMI levels were observed in older cohorts in women than in younger cohorts. Because BMI is well established as a determinant of the SUA level, the birth cohort effect of BMI may explain at least partially the birth cohort effect of SUA levels. This may suggest that, with men, younger birth cohorts tend to pursue a more westernized lifestyle than do older birth cohorts. However, this is not true for women. These observations were supported by our recent observation, using the same baseline cohort, which found that birth cohort effects regarding serum triglyceride exist in men, but no effect was observed in women (14).
It should be noted that our results may not be indicative of the Japanese population as a whole, because the subjects in this study represent an urban population. Some selection bias, such as a healthy worker bias, may exist in our study, because most of the subjects were healthy office workers. In addition, the subjects may be aware of their SUA levels, because they had received annual examinations at a health examination center. Nevertheless, their SUA levels increased with age.
The main findings of this study include an observed longitudinal increase in SUA levels with age in men, and in most of the birth cohorts of women, in a large Japanese population group observed from 1989-1998. Birth cohort effects and variations in SUA levels were observed in this population.
|
Acknowledgments
|
|---|
This work was funded by a Grant in Aid for the Comprehensive Research on Aging and Health from the Ministry of Health and Welfare of Japan.
Received February 27, 2002
Accepted April 16, 2002
|
References
|
|---|
- Hall AP, Barry PE, Dawber TR, McNamara PM, 1967. Epidemiology of gout and hyperuricemia. A long-term population study. Am J Med 42:27-37. [Medline]
- Campio ED, Glynn RJ, DeLabry LO, 1987. Asymptomatic hyperuricemia. Risks and consequences in the normative aging study. Am J Med 82:421-426. [Medline]
- Klein R, Klein BE, Cornoni J, Maready J, Cassel JC, Tyroler HA, 1973. Serum uric acid, Georgia. Arch Intern Med 132:401-410. [Abstract/Free Full Text]
- Bengtsson C, Lapidus L, Stendahal C, Waldenstrom J, 1988. Hyperuricemia and risk of cardiovascular disease and overall death. Acta Med Scand 224:549-555. [Medline]
- Freedman DS, Williamson DF, Gunter EW, Byers T, 1995. Relation of serum uric acid to mortality and ischemic heart disease: the NHANES I Epidemiologic Follow-up Study. Am J Epidemiol 141:637-644. [Abstract/Free Full Text]
- Liese AD, Hense H-W, Lowel H, Doring A, Tietze M, Keil U, 1999. Association of serum uric acid with all-cause and cardiovascular disease mortality and incident myocardial infarction in the MONICA Augsburg cohort. Epidemiology. 10:391-397. [Medline]
- Fang J, Alderman MH, 2000. Serum uric acid and cardiovascular mortality. JAMA 283:2404-2410. [Abstract/Free Full Text]
- Levine W, Dyer AR, Shekelle RB, Schoenberger JA, Stamler J, 1989. Serum uric acid and 11.5-year mortality of middle-aged women. J Clin Epidemiol. 42:257-267. [Medline]
- The Coronary Drug Project Research Group1976. Serum uric acid: its association with other risk factors and with mortality in coronary heart disease. J Chron Dis 29:557-569. [Medline]
- Culleton BF, Larson MG, Kannel WB, Levy D, 1999. Serum uric acid and risk for cardiovascular disease and death: the Framingham Heart Study. Ann Intern Med 131:7-13. [Abstract/Free Full Text]
- Moriarity JT, Folsom AR, Iribarren C, Nieto FJ, Rosamond WD, 2000. Serum uric acid and risk of coronary heart disease: Atherosclerosis Risk in Communities (ARIC) study. Ann Epidemiol. 10:136-143. [Medline]
- Nomura H, Shimokata H, Ando F, Miyake Y, Kuzuya F, 1999. Age-related changes in intraocular pressure in a large Japanese population: a cross-sectional and longitudinal study. Ophthalmology. 106:2016-2022. [Medline]
- Mori K, Ando F, Nomura H, Sato Y, Shimokata H, 2000. Relationship between intraocular pressure and obesity in Japan. Int J Epidemiol. 29:661-666. [Abstract/Free Full Text]
- Kuzuya M, Ando F, Iguchi A, Shimokata H, 2002. Changes in serum lipid levels during a 10-year period in a large Japanese population: a cross-sectional and longitudinal study. Atherosclerosis 163:313-320. [Medline]
- SAS Institute. SAS Language Guide for Personal Computers, Version 6.03. Cary, NC: SAS Institute; 1988.
- Ward MM, Leigh JP, 1993. Pooled time series regression analysis in longitudinal studies. J Clin Epidemiol. 46:645-659. [Medline]
- Littell RC, Milliken GA, Stroup WW, Wolfinger RD. 1996 SAS System for Mixed Models. Cary, NC: SAS Institute; 1989.
- Zalokar J, Lellouch J, Claude JR, Kuntz D, 1974. Epidemiology of serum uric acid and gout in Frenchmen. J Chron Dis. 27:59-75. [Medline]
- Mikkeksen WM, Dodge HJ, Valkenburg H, 1965. The distribution of serum uric acid values in a population unselected as to gout or hyperuricemia. Tecumseh, Michigan, 1959–1960. Am J Med. 39:242-251. [Medline]
- Kagan A, Harris BR, Winkelstein W, Jr et al. 1974. Epidemiologic studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii and California: demographic, physical, dietary and biochemical characteristics. J Chron Dis. 27:345-364. [Medline]
Top
Abstract
Methods
