MASAKI MORIYAMA, HIROSHI SAITO, ATSUHIRo NAKANO,* Department of Preventive Medicine and Health Promotion,

Transcription

1 Tohoku J. exp. Med., 1988, 156, Estimation of Urinary 24-hr Creatinine Excretion by Body Size and Dietary Protein Level : A Field Survey Based on Seasonally Repeated Measurements for Residents Living in Akita, Japan MASAKI MORIYAMA, HIROSHI SAITO, ATSUHIRo NAKANO,* SHOETSU FUNAKI õ and SABURO KOJIMA õ Department of Preventive Medicine and Health Promotion, Nagasaki University School of Medicine, Nagasaki 852 *Department of Epidemiology, National Institute for Minamata Disease, Minamata 867 õakita Prefectural Institute of Public Health, Akita 010 MORIYAMA, M., SAITO, H., NAKANO, A., FUNAKI, S. and KOJIMA, S. Estima tion of Urinary 24-hr Creatinine Excretion by Body Size and Dietary Protein Level : A Field Survey Based on Seasonally Repeated Measurements for Residents Living in Akita, Japan. Tohoku J. exp. Med., 1988, 156 (1), We aimed to investigate the modifying effect of dietary protein level on the estimation of 24-hr creatinine excretion using anthropometric measurements, in a group of 40 healthy Japanese adults, during usual daily life. To evaluate the variability of each measurement during non-restrictive life, we assayed 24-hr urinary excretions of creatinine and urea-n and measured heights and weights once in every three months in a total of 4 times during a year. Urinary urea-n was used as a marker of dietary protein level. Seasonal variability of anthropometrics and urinary excretions was observed on the individual basis. In comparison to the rather small individual variability of weight (mean of CV=1.2 for males, 1.7 for females), variability of urinary creatinine (CV=11.7, 8.4) and urea-n (CV=18.9, 17.3) took larger values. The creatinine excretion was estimated by the following two sets of predictor variables ; (1) age, height and weight, (11) age, height, weight and urea-n. Although, a statistically significant estimation was obtained by (I), estimation (II) resulted higher regression coefficient than (I), and contribution of urea-n to the estimation of creatinine was shown. Average level of seasonal variability of each predictor variable was further applied to the regression equation (II). Therefore, the possible difference of 24-hr creatinine excretion caused by the seasonal variability of dietary protein level was estimated to be 9.2% for males and 3.6% for females in the given population. Thus, when one predicting formula of 24-hr creatinine by the use of anthropometrics is applied to other populations, the difference of protein nutritional status should be carefully considered. ereatinine : urea-n : dietary protein level ; seasonal variability The twenty four-hour (24-hr) urinary excretion of creatinine of a given Received June 27, 1988; revision accepted for publication August 18,

2 56 M. Moriyama et al. subject is relatively constant from day to day, and proportionate to muscle mass (Heymsfield et al ; Suzuki et al. 1984). Bucause of constancy of renal excretion, urinary creatinine has been used to standardize the excretion of various substances in a spot urine sample by representing them as ratios to creatinine (Hodgkinson and Thompson 1982; Hojo 1982). Thus, the development of simple method to estimate the individual 24-hr creatinine excretion is important in dealing with health related information obtained from urine samples, and Kawasaki et al. (1985) already indicated regression equation to estimate 24-hr excretion of Japanese adults by age, height and weight. The authors indicated such equations for Japanese children, aged 2 to 18 years (Moriyama et al. 1987). However, it is not yet fully confirmed that one predicting formula, which is obtained in a given population at a certain cross sectional study, can satisfactorily predict the excretion of the same population at the different occasion or the excretion of other populations living in different ecological settings. Except the body size, one of the major factor which modify the creatinine excretion is diet, and three of influential dietary constituents, such as protein, creatine, and creatinine, are known (Heymsfield et al. 1983). In the present study, the authors aimed to evaluate the modifying effect of dietary protein level on the estimation of 24-hr urinary creatinine using anthropometric measurements. In order to evaluate the variability of both of dependent variable (creatinine) and independent variables (dietary protein level, anthropometrics) at different occa sions, we assayed 24-hr urinary excretions of creatinine and urea-n, as a marker of protein intake, and measured heights and weights once in every three months in a total of 4 times during a year. The dependent variable, creatinine excretion, was estimated by the following two sets of independent variables ; (I) age, height and weight, (II) all variables in (I) and urea-n. Two sets of results are compared, and effect of dietary protein level was evaluated. SUBJECTS AND METHODS Forty residents (22 males ; age=46.6 }7.8 (S.D.) ; 18 females, age=44.5 }8.3) living in Akita city and its neighboring town of Ikawa, northern part of Honshu island volunteered to provide 24-hr urine specimens. Subjects consisted of professional/technical workers (13 males, 8 females), own-account farm managers (5 males), Japan,, their co-operatives (9 females), and others (4 males, 1 female). No restrictions were placed on the activity and diet during the day of collection. The 24-hr urines were collected at four occasions, at around the 20th day of February, May, August and November, in the year The 24-hr collection started and ceased at 6 : 30 A.M. Every time when urine was voided, the time was recorded by the subject. On the day collection was completed, samples from each subject were placed into 10 ml sample tubes and stored at -30 Ž until analysis. Urinary creatinine and urea-n was determined by the Jaffe reaction (Bonsnes and Taussky 1945) and diacetylmonoxime reaction (Coulombe and Favreau 1963). Height and weight were measured on the day of urine collection.

3 Urinary Creatinine and Dietary Protein Level 57 Seasonal variability of measurements RESULTS In Table 1 are presented means of height, weight, 24-hr urinary urea-n and creatinine obtained at each season (1, February ; 2, May ; 3, August ; 4, Novem- TABLE 1. Mean values of four measurements at each season Ryan-Einot-Gabriel-Welsch (REGWQ) multiple range test for means, (a=0.05); ª significantly larger than means of other seasons, «significantly smaller than means of other seasons. TABLE 2. Mean values of coefficients of variation (CV) and their distributions, based on the seasonally repeated measurements for each subject

4 58 M. Moriyama et al. ber) for each sex. Four seasonal means were compared among each other by the use of multiple range test. In both sexes, means of weight and urea-n at season 3 were smaller than those of other seasons, and in males, the observed differences were statistically significant. There were no statistically significant differences in creatinine excretion between seasons. In order to observe the seasonal variability of measurements on the individ- TABLE 3. season Correlations of creatinine with height, weight and urea-n at each *p < **P<0.01. TABLE 4. Multiple regression analysis to estimate 24-hr creatinine excretion at each season UN, R, multiple correlation coefficient ; Int, intercept ; Ht, height ; Wt, weight ; urea-n. ALL, means of four seasonally repeated measurements are used for the calculation.

5 TABLE 5. Urinary Creatinine and Dietary Protein Level 59 Estimated change of 24-hr creatinine excretion resulted from seasonal variability of body size and urea-n excre tion (A), Overall mean including four seasonally repeated measure ments. (B), Average level of individual seasonal variability of each predictor variable calculated using the following formula ; Seasonal variability= mean of CV (in Table 2) x overall mean (A) (C), Provoked change of creatinine excretion (C) was estimated by applying (B) to the regression equation of ALL (seasons were combined) in Table 4. ual basis, four measurements were averaged and a coefficient of variation (CV (%) : S.D./meanx 100) was calculated for each subject. Table 2 shows means and 0 distributions of CV for height, weight, urea-n and creatinine in the given popula tion. In both sexes, means of CV of weight are below 2%, and means of CV of height are much smaller than those of weight. In comparison to this rather small seasonal variation of body size, creatinine and urea-n show much larger variation (urea-n=18.9 for males, 17.2 for females ; creatinine-11.7 for males, 8.4 for females). Relationships of creatinine with body size and urea-n In Table 3 are shown bivariate correlations of creatinine with height, weight or urea-n calculated for each season (1-4), and for the average value (ALL). For males, any of height, weight or urea-n correlated significantly with creatinine in the group ALL. However, the correlation between urea-n and creatinine in ALL was not significant for females. Even in the same sex, levels of correlations were not constant among different seasons. For example, in females, the correlation between urea-n and creatinine was significant only at season 2. Multiple regression analysis was applied to estimate creatinine for each season, and for ALL (Table 4). In the estimation H, urea-n was added to other

6 60 M. Moriyama et al. TABLE 6. Simulated value of 24-hr urinary creatinine at diver gent dietary protein levels An estimation for a male with the following presupposi tions : Age -50 years, height=170 cm, weight= 70 kg. (A), Hypothesized value of protein intake. (B), Urinary urea-n corresponding to the protein intake of (A). Urinary urea-n was estimated by the use of a regression equation [(B)= ~(A)+1.41], which was presented by Yukawa (1981). (C), Urinary creatinine was estimated by a regression equation in Table 4 ; (C)= ~ age x height x weight+63.1 x (B) three variables (age, height and weight), which were already used in the estima tion I. In both sexes, estimation II resulted higher correlation coefficient than estimation I. By adding the urea-n as a fourth independent variable, the contribution of height and weight somewhat decreased, and apparent contribution of urea-n was added, especially for males. In the case of females, much smaller contribution of urea-n on the creatinine was observed. When the obtained equations were compared among different seasons, levels of regression coefficients were not the same. For example, largest regression coefficient of urea-n (males, 88.1 at season 3 ; females, 37.9 at season 2) was two to three times larger than the smallest one (males, 31.0 at season 2 ; females, 12.4 at season 1). An estimation of difference of creatinine caused by the seasonal variation of body size and dietary protein level In Table 5, the average difference of individual creatinine excretion originat ed from each of seasonal variations of body size and/or dietary protein intake was further estimated, by applying means of CV (Table 2) to the multiple regression equation (Table 4, estimation II for group ALL). Differences of creatinine caused by the seasonal variability of height, weight and/or urea-n were estimated to be 0.1, 0.7 and 9.2% for males, and 0.1, 0.8 and 3.6% for females, respectively.

7 Urinary Creatinine and Dietary Protein Level 61 DISCUSSION According to the review work by Heymsfield et al. (1983), a normal daily variation of urinary creatinine ranges from 4 to 8% in reliable studies, the subjects of which are mostly healthy Caucasian adults who were maintained in the metabolic ward. For Japanese, the same order of variation is reported for two adults under the experimental diets with medium protein level (Tomita 1961), and the range of 3.6 to 8.5% (expressed in terms of CV) is reported for five adults under the free diet (Yoneyama 1984). In both of these studies for Japanese, 24-hr urine specimens were collected for successive days only during summer season ; i.e., for 20 days during August and September (Tomita 1961), and/or days during July and August (Yoneyama 1984). However, in order to assess the reliability and stability of a predicted value of creatinine, an analysis of variabil ity throughout seasons is also necessary. In the first time, we collected 24-hr urine samples once in every three months in a total of 4 times during a year for each of 40 subjects. The observed variations (CV) of 24-hr urinary creatinine, 11.7 }6.9% (mean }s.d.) for males and 8.4 }4.7% for females, appear to be larger in comparison to the former reported value of 4 to 8% on the daily basis. Ingestion of creatine and creatinine free diet not only lowered creatinine excretion by as much as 30%, but also lowered the correlation between 24-hr excretion and weight (Bleiler and Schedl 1962). However, except for these experimental conditions, in which the amount of protein intake is switched experimentally from medium to high and/or low conditions, it has been accepted that day to day variability of physical activity and diet during the usual daily life does not explain the variability of 24-hr urinary creatinine (Heymsfield et al. 1983). Yoneyama (1984), however, analyzed the relationship between 24-hr creatinine excretion and daily protein intake, and found the significant positive coy-relations between protein intake and creatinine excretion for three out of five subjects. In our sample of males, urea-n also correlated significantly with creatinine throughout four successive observations, and at seasons 2 and 4, the urea-n showed even higher correlation in comparison to anthropometrics (Table 3). In the females, urea-n showed significant correlation at season 2. Therefore, our results seem to strengthen the contribution of dietary protein level to the 24-hr urinary excretion of creatinine in the usual daily life of Japanese. In the present study, seasonally repeated observations resulted in varying leves of variability (Table 2) according to the types of measures ; among which, smallest variability was observed for height (mean of CV=0.09 for males, 0.07 for females) and largest one was observed for urinary urea-n (18.9, 17.2). In Table 5, resulted values of CV were further applied to a regression equation, and differences of 24-hr urinary creatinine caused by the seasonal variation of urea-n were estimated at 9.2% for males and 3.6% for females. Thus, in the usual daily life, it is not practically necessary to consider the modifying effect of dietary

8 62 M. Moriyama et al. protein level for the estimation of urinary creatinine using height/or weight. However, if one's dietary protein level is extremely deviated from the average level of the population, we can no longer overlook the bias of estimation. For example, daily availability of protein per person in Africa South of Sahara (51 g) and Southeast Asia (53 g) is nearly half in comparison to Industrialized countries (103 g) (UNICEF 1987). In Table 6, we tried to simulate the amount of urinary creatinine at divergent levels of protein intakes. In this simulation, urinary levels of urea-n corresponding to divergent dietary protein levels were firstly estimated by the use of a regression equation of Yukawa (1981). Urinary creatinine was, then, estimated by a regression equation (Table 4) obtained in the present study. According to this result, if we continue to use a same equation under the different settings, differences of protein intakes in the divergent socio economic and developmental status might cause as much as 40% difference of urinary creatinine. Thus, when one predicting formula of creatinine by the use of anthropometrics is applied to other populations living in different ecological settings, the difference of protein nutritional status should be carefully examined before the application. References 1) Bleiler, R. & Schedl, H. (1962) Creatinine excretion : Variability and relationship to diet and body size. J. Lab. clin. Med., 59, ) Bonsnes, R.W. & Taussky, II.H. (1945) On the colorimetric determination of creatinine by the Jaffe reaction. J. biol. Chem., 158, ) Coulombe, J.J. & Favreau, L. (1963) A new simple semi-micromethod for color imetric determination of urea. Clin. Chem., 9, ) Heymsfield, S.B., Argeaga, C., McManus, C., Smith, J. & Moffitt, S. (1983) Measure ment of muscle mass in humans : Validity of the 24-hour urinary creatinine method. Amer. J. clin. Nutr., 37, ) Hodgkinson, A. & Thompson, T. (1982) Measurement of the fasting urinary hydroxy proline : Creatinine ratio in normal adults and its variation with age and sex. J. clin. Path., 35, ) Hojo, Y. (1982) Single-void urine selenium level expressed in terms of creatinine content as an effective and convenient indicator of human selenium status. Bull. environm. Contam. Toxicol., 29, ) Kawasaki, T., Uezuno, K., Kikkawa, K., Utsunomiya, H. & Imamura, K. (1985) Studies of urinary creatinine excretion in clinically healthy subjects. J. Hlth Sci. Kyushu Univ., 7, (in Japanese with English abstract) 8) Moriyama, M., Saito, H. & Kabuto, M. (1987) Relationship between twenty-four hour urinary creatinine excretion and weight, or weight and height of Japanese children. J. nutr. Sci. Vitaminol., 33, ) Suzuki, T., Inaoka, T. & Kawabe, T. (1984) Creatinine height index in a sample of Japanese adults under sedentary activities. J. Nutr. Sci. Vitaminol., 30, ) Tomita, Y. (1961) A study on the influence of the dietary protein level upon the urinary creatinine excretion. Eiyo to Shokuryo, 14, (in Japanese with English abstract) 11) United Nations Children's Fund (1987) Availability of food per person. In : The State of the World's Children 1987, Oxford University Press, Oxfordshire, U.K., p.

9 Urinary Creatinine and Dietary Protein Level ) Yoneyama, K. (1984) The day to day variations of urinary hydroxyproline and creatinine excretions, and dietary protein intake. Jap. J. Hygiene, 39, ) Yukawa, K. (1981) Studies on estimation of dietary protein intake from urinary nitrogen ingredients excretion. Jap. J. pub. Health, 28, (in Japanese with English abstract)