点击显示 收起
【摘要】 Despite recent interest in water intake, few data are available on water metabolism in adults. To determine the average and range of usual water intake, urine output, and total body water, we administered 2 H oxide to 458 noninstitutionalized 40- to 79-yr-old adults living in temperate climates. Urine was collected in a subset of individuals ( n = 280) to measure 24-h urine production using p -aminobenzoic acid to ensure complete collection. Preformed water intake was calculated from isotopic turnover and corrected for metabolic water and insensible water absorption from humidity. Preformed water intake, which is water from beverages and food moisture, averaged 3.0 l/day in men (range: 1.4-7.7 l/day) and 2.5 l/day in women (range: 1.2-4.6 l/day). Preformed water intake was lower in 70- to 79 (2.8 l/day)- than in 40- to 49-yr-old men and was lower in 70- to 79 (2.3 l/day)- than in 40- to 49- and 50- to 59-yr-old women. Urine production averaged 2.2 l/day in men (range: 0.6-4.9 l/day) and 2.2 l/day in women (0.9-6.0 l/day). There were no age-related differences in results in women, but 60- to 69-yr-old men had significantly higher urine output than 40- to 49- and 50- to 59-yr-old men. Only the 70- to 79-yr-old group included sufficient blacks for a racial analysis. Blacks in this age group showed significantly lower preformed water intake than did whites. Whites had significantly higher water turnover rates than blacks as well. Multivariate regression indicated that age, weight, height, and body mass index explained <12% of the gender-specific variance in water input or urine output, yet repeat measures indicated that within-individual coefficient of variation was 8% for preformed water intake ( n = 22) and 9% for 24-h urine production ( n = 222). These results demonstrate that water turnover is highly variable among individuals and that little of the variance is explained by anthropometric parameters.
【关键词】 water intake twentyfourhour urine preformed water insensible water
RECOMMENDATIONS FOR CONSUMPTION of water and nonalcoholic beverages have recently been questioned ( 20 ). The interest in this issue is widespread because water is among the most important nutrients for the maintenance of life. The body uses water for transporting nutrients and wastes, lubrication, temperature regulation, and tissue structure maintenance. In addition, plentiful fluid consumption may be protective against diverse medical conditions, including kidney stones ( 26 ), constipation ( 2 ), colorectal cancer, premalignant adenomatous polyps ( 35 ), and bladder cancer ( 23 ).
Water deprivation results in life-threatening dehydration within a few days. Loss of body water exceeding 5% of body weight leads to decreased endurance, culminating in heat exhaustion ( 20, 25 ). Older vs. younger individuals have been shown to have a higher risk of developing dehydration than younger adults, which may be attributed to decreased total body water (TBW) with age ( 33 ), impaired renal fluid conservation ( 3 ), and physiological hypodipsia ( 8 ).
Despite the physiological importance of water to life, little is known about water intake and excretion patterns in free-living individuals, because fluid intake, particularly from noncaloric, nonalcoholic, and noncaffeinated beverages is poorly documented. The 1977-1978 National Food Consumption Survey ( 16 ) is one of the few sources of information on water intake, but these data are limited by unaccounted water found in foods and the use of a single 24-h dietary recall ( 14, 20 ). Moreover, nonquantitative intake from water fountains and the likelihood that many people consume fluids with little thought leads to underreporting ( 19 ).
An alternative approach that does not depend on self-reported intake is the use of hydrogen-labeled water turnover, a method used by comparative animal physiologists for decades to objectively measure water turnover in wild animals ( 24 ). The procedure begins with a bolus administration of isotopically labeled water, such as nonradioactive 2 H oxide. Within 2-3 h, this tracer equilibrates with body water and provides a measure of the volume of the TBW pool ( 32 ). The labeled water is then excreted from the body through all routes of water loss and is diluted by unlabeled water through all routes of input. The time course of labeled water dilution provides a measure of water turnover (input and output) per unit of time ( 9, 24 ).
We combined data from two studies in healthy, free-living American adults across a broad age range to which 2 H-labeled water was administered to measure total energy expenditure (TEE) using the doubly labeled water (DLW) technique ( 32 ). In one of these studies, two 24-h urine collections were made from many of these same participants using p -aminobenzoic acid (PABA) to confirm completeness. These data are among the first objective assessment of water turnover in American adults and provide documentation of both the average and range of water input and urine production.
METHODS
To evaluate the water requirements of normal adults, we combined two data sets for which we measured water influx and efflux. The first was the Health, Aging and Body Composition (Health ABC) study conducted between July 1998 and August 2000 in Memphis, TN, and Pittsburgh, PA. The main objective of Health ABC was to establish relationships between the changes in body composition and the development of early disabilities and mortality in the elderly. The second was the Observing Protein and Energy Nutrition (OPEN) study, which took place in Rockville, MD, between August 1999 and March 2000. The main objective of the OPEN study was to assess the structure of dietary measurement error associated with food frequency questionnaires and 24-h dietary recalls in middle-aged adults. The experimental protocols for the Health ABC and OPEN studies have been previously described ( 6, 34 ). Data from these two studies were combined.
Participants
To be eligible for the Health ABC study, elderly persons had to be free of difficulties with activities of daily living and lower-extremity functional limitations, defined as difficulty walking one-quarter mile or climbing 10 steps without resting. Participants were recruited from a random sample of white Medicare beneficiaries and all age-eligible black community residents in designated zip codes in and around Pittsburgh, PA, and Memphis, TN. The Health ABC study involved 3,075 participants aged 70-79 yr stratified by race (black, white) and gender. Of these, a subgroup of 323 individuals, with the same race and gender stratification, participated in an energy expenditure study published elsewhere and thus were eligible for inclusion in the current analysis of water metabolism ( 6 ). Of the 323 participants, data from 40 were excluded, including 9 who did not have usable isotopic data, 16 who did not have different physical characteristics, and 15 who were over 79 yr old. Of the remaining 283 participants, 145 were white and 138 were black. Because the Health ABC study included far more blacks than the OPEN study and because the Health ABC study included only one age group (70-79 yr olds), age-related data analysis focused on white participants only.
The OPEN study involved 484 healthy participants aged 40-69 yr ( 34 ). Among the 399 white participants in the OPEN study, 27 participants had unusable TEE and were excluded from the analysis. Of the 372 remaining participants, 59 participants either did not have 24-h urine collected or their urine was deemed unusable due to incomplete PABA recovery.
Together, there were 458 white participants who had complete data and were included in the primary analysis, of which 145 participants were from the Health ABC study and 313 from the OPEN study. In addition, the 138 elderly black participants (70-79 yr olds) from the Health ABC study were included in a secondary analysis to examine race-related differences in preformed water intake in the elderly population.
The Health ABC and OPEN protocols were reviewed and approved by the institutions involved in each of the respective studies. All subjects provided informed, written consent regarding their participation in the respective study.
Protocol
In both the OPEN and Health ABC studies, TEE was measured using the DLW method according to protocols described elsewhere ( 6, 34 ). Briefly, participants made two visits approximately 2 wk apart. Before visit 1, participants fasted for 4 h and, in most instances, overnight. Body mass was measured either in a clinic setting or in a hospital, and baseline urine samples were collected. Then, a dose of DLW was given orally to each participant. The dose provided 1.9 g 10 atom% 18 O water and 0.12 g 99.9 atom% 2 H water/kg of estimated total body water ( 32 ). Urine collections were taken either at 2, 3, and 4 h postdose (OPEN study) or 4 and 6 h postdose (Health ABC study) to assess isotope equilibration in body water. Plasma samples were collected at 4 h postdose in participants above 60 yr of age in the OPEN study and in all participants in the Health ABC study. The participants came back for visit 2 14 days later and provided end-dose urine specimens. Urine samples were stored in cryogenic stable tubes at -20°C before analysis by isotope ratio mass spectrometry.
All participants in the OPEN study were asked to collect two 24-h urine sets between visits 1 and 2. The urine was analyzed for urinary nitrogen, potassium, and sodium. For 24-h urine collection, the participants were asked to take three PABA pills orally, one at each meal. The completeness of the 24-h urine collection was assessed using the amount of PABA excreted in urine as described by Bingham et al. ( 5 ). When PABA recovery 85%, the subject's urine was considered complete. Recoveries of less than 70% were removed from the analysis. When recoveries were between 70 and 85%, urine samples were considered usable after their adjustment to 93% recovery of PABA ( n = 51 of 935 24-h urine collections) ( 18 ). All samples in excess of 110% recovery by the colorimetric technique were analyzed by HPLC to distinguish between PABA and acetaminophen, a drug commonly taken by participants ( n = 123 of the 935 urine collections in the study).
Isotopic Analysis
The isotopic analyses have been previously described for Health ABC as well as for OPEN ( 6, 36 ). Briefly, 2 H analyses were performed by chromium reduction according to Schoeller et al. ( 30 ) using a dual-inlet isotope ratio mass spectrometer (Delta Plus Mass Spectrometer, Finnigan MAT, San Jose, CA). 18 O enrichments were measured by CO 2 equilibration on a Delta-S isotope ratio mass spectrometer (Finnigan MAT) through a continuous-flow inlet system developed in the laboratory ( 30 ). To protect against possible interference from postvoid residual volume, plasma specimens were collected from the participants who were in their seventh and eighth decades of life. Urine analyses were used 2%; plasma enrichment was used for calculating TBW. Agreement was observed in 87% of participants.
Calculations
TBW. The isotopic dilution spaces ( N ) of 2 H and 18 O were calculated according to Cole and Coward ( 7 )
where W is water (g) used to make a dilution of the dose water; A is the dose water (g) administered to the participant, and a is the dose water (g) used in the dilution; a is the isotopic abundance of the diluted dose water; t is the isotopic abundance of tap water used in dilution; s is the isotopic abundance of postdose specimen; and p is the isotopic abundance of the predose specimen. Isotopic abundances are measured in permil units [ = (R/R s - 1) 1,000, where R is the ratio of heavy to light hydrogen isotope in the sample (R) and standard (R s )]. The ratio of 2 H to 18 O dilution spaces averaged 1.034 ± 0.018.
TBW was calculated as an average of the 2 H dilution space ( N d ) and the 18 O dilution space ( N o ) after a correction for in vivo isotopic exchange using the equation ( 28 )
TEE. DLW-derived carbon dioxide production (rCO 2 ) was calculated according to Schoeller et al. ( 31 )
where TBW is total body water (moles), and k o and k d are the oxygen and 2 H elimination rates per day, respectively. TEE was derived from Wier's equation, assuming a respiratory quotient (RQ) of 0.86.
Water turnover. Water turnover (kg/day) in the body was calculated from the 2 H dilution space and elimination rate ( 9 )
N d is the dilution space measured using 2 H, and k d is the fractional turnover rate of 2 H in the body water after equilibration, where c f is the final enrichment of 2 H in the urine, c i is the initial enrichment of 2 H in the urine, t f is the final time point, and t i is the initial time point. In the calculation of water turnover, equality between the water influx and efflux was assumed.
Estimation of water influx. Water influx is accounted for by metabolic water, inspiratory water (moisture content of inhaled air), transcutaneous water intake (water absorbed by the skin), and preformed water (which is the water consumed orally from beverages and water in the food). Those variables were calculated using the following formulas described in Fjeld et al. ( 9 ).
METABOLIC WATER. Metabolic water production was calculated from the average macronutrient content of the diet of a normal healthy American according to the equation ( 9 )
where TEE is in kilocalories per day and %fat, %protein, and %carbohydrate are taken as 31, 14, 52, and 4%, respectively, based on the values obtained from 24-h dietary recall 1 from the OPEN study. The percentages do not total 100% due to a rounding error. The amount of water produced per kilocalorie of substrate oxidized as fat, protein, carbohydrates, and alcohol is 0.119, 0.103, 0.15, and 0.168 g, respectively.
INSPIRATORY WATER. Inspiratory water was calculated as
where respiratory volume is in liters per day and absolute humidity is in milligrams per liter, assuming average relative humidities of 35% for the winter period and 65% for the summer period at 24°C. Respiratory volume was calculated from the rCO 2 obtained from DLW, assuming that 3.5% of inspired air is CO 2.
TRANSCUTANEOUS WATER. Transcutaneous water influx was then calculated as
where 0.18 is the rate of transcutaneous absorption in grams per square meter of body surface area in an atmosphere saturated with water vapor (21.7 mg/l). BSA is the body surface area (m 2 ) estimated from the Dubois formula ( 9 ). A clothing factor of 50% was assumed, as clothing would decrease the rate of evaporation through the skin.
PREFORMED WATER INTAKE. Preformed water was calculated by the difference between water turnover (rH 2 O) and the sum of all the above-calculated values (metabolic water intake, inspiratory water intake, and transcutaneous water intake).
Estimation of water efflux. Water efflux includes urinary water, fecal moisture, and insensible water loss, i.e, water lost through expiration, transcutaneous loss, and sweat loss. Urinary losses were averaged for two 24-h urine collections ( n = 313) from participants in the OPEN study only. The mean difference between the two 24-h urine collections was 34.9 ml in women and 40.6 ml in men. We assumed average fecal moisture of 72% (±2%) and fecal weight based on the literature for a normal healthy individual consuming a typical American diet ( 1, 29 ). Insensible water loss was calculated as the sum of transcutaneous, expiratory, and sweat water. Expiratory water loss was calculated using the same formula as for inspiratory water influx but assuming a relative humidity of 97% in the expired air ( 13 ). Respiratory volume or tidal volume was calculated from the CO 2 production rate measured, assuming that 3.5% of tidal volume was composed of CO 2.
Sweat loss was calculated as the difference between water turnover and the sum of expired, transcutaneous, fecal, and urinary water losses.
Statistical Analysis
Regression analyses were used to determine the effect of anthropometric variables [height, weight, age, and body mass index (BMI)] on the preformed water intake. Analysis of variance and post hoc Fischer's protected least significant difference test were performed to determine the effect of age on preformed water intake. Intraindividual variability was determined from the coefficient of variation between the two 24-h urine collections. Statistical analyses were performed using STATVIEW software, version 5.0.1. Results are expressed as means ± SD. A P value of
RESULTS
The anthropometric characteristics of 458 participants from the Health ABC and the OPEN study are presented in Table 1.
Table 1. Mean anthropometric and TBW values of 458 white participants: 251 men and 207 women
Water Influx
Percentile distributions of preformed water intake from 251 male and 207 female participants are shown in Fig. 1. Individuals varied widely in their preformed water intake, ranging from 1.4 to 7.7 l/day (mean 3.0 ± 0.9) for men and from 1.2 to 4.6 l/day (mean 2.5 ± 0.7) for women. Average estimated values for water influx components, i.e., metabolic water intake, inspired water intake, transcutaneous water intake, and preformed water, calculated by difference, are shown in Table 2 by age group and gender. Water influx values range from 1.9 to 8.6 l/day (mean 3.6 ± 0.9) in men and from 1.6 to 5.2 l/day (mean 3.0 ± 0.7) in women.
Fig. 1. Distribution of preformed water intake from 251 male and 207 female participants.
Table 2. Mean water turnover and influx values in men and women by decade of age
The percentile distribution of preformed water intake categorized by decade of age is shown in Fig. 2. In both men and women, preformed water intake did not differ significantly among age groups until the eighth decade of life. The 70- to 79-yr-old men had significantly lower preformed water intake than the 40- to 49-yr-old men ( P = 0.005), and the 70- to 79-yr-old women had lower preformed water intake than the 40- to 49- and 50- to 59-yr-old women ( P = 0.003 and 0.03, respectively). There were no other significant differences in preformed water intake by age and gender ( P 0.05, not detailed).
Fig. 2. Percentile distribution of preformed water intake for each age group within genders. The 90th, 75th, 25th, and 10th percentiles are indicated by the top error bar, the top of the box, the bottom of the box, and the bottom error bar, respectively. The median is indicated by the horizontal line dividing the box. y/o, Yr old. Age groups with the same letter differ significantly within genders ( P < 0.05).
Univariate regression analysis of preformed water intake by gender using weight, age, height, BMI, and urine output as independent variables are shown in Table 3. Univariate analysis of urine output showed a correlation with an r 2 of 0.51 and 0.47 with preformed water intake in men and women, respectively. None of the other variables assessed had a very biologically meaningful correlation with preformed water intake. Multiple regression analysis with the same independent variables explained 57 and 61% of the variance of preformed water in men and women, respectively.
Table 3. Univariate regression analysis of preformed water intake
Physical characteristics of the 70- to 79-yr-old black and white participants are presented in Table 4. Among all the 70-79 yr olds, preformed water intake was significantly lower in blacks than whites in both men and women ( P < 0.05). Whites had significantly higher water turnover rates than blacks as well ( P = 0.01). Gender differences among blacks in this limited group were similar to those of whites with respect to their weight, height, BMI, and TBW.
Table 4. Mean anthropometric and TBW values by race and gender among 70- to 79-yr-old participants
Body composition data were available for only the 70- to 79-yr-old participants. To ascertain whether these elderly black and white participants were subjected to chronic, hyperosmotic dehydration secondary to low water intake, we divided the elderly cohort into quintiles of preformed water intake and compared it with the hydration of fat free mass [FFM; measured using dual X-ray absorptiometry (DXA)] among 70-79 yr olds. The hydration of FFM (TBW/FFM) did not differ between quintiles of preformed water intake. Although the difference was not significant, the hydration of the lowest quintile of water intake averaged 0.3% greater than the highest quintile, with a group SD of 0.4%. As water intake was calculated from our primary measures, the above was not adjusted for body size, so we performed the identical analysis using fractional water turnover ( k 2 ) as the dependent variable. Again, there was no effect of water turnover on hydration status in the eighth decade of life.
Water Efflux
Average calculated values for water efflux and its components are shown in Table 5 by age and gender. Fecal losses were artifactually equal in all the participants (0.07 l/day) because we used the same assumptions of fecal weight and moisture content ( 1, 29 ). Insensible water losses (transcutaneous, expiratory, and sweat losses), calculated by difference, were highly variable among age groups. The 60- to 69-yr-old men showed significantly lower insensible water losses compared with the 40- to 49- and 50- to 59-yr-old men ( P < 0.0001 and 0.006, respectively), and the 60- to 69-yr-old women showed significantly lower insensible water losses than the 40- to 49-yr-old women ( P = 0.004).
Table 5. Mean water turnover and efflux values by gender and age
Distributions of urine volume in 134 women and 180 men from the OPEN data are shown in Fig. 3. The urine output in men was 0.62 4.94 l/day (mean 2.18 ± 0.94) and in women was 0.91-4.87 l/day (mean 2.19 ± 0.75).
Fig. 3. Distribution of urinary volume in 133 women and 180 men 40-69 yr of age. The urine output (mean ± SD) of the men varied between 0.62 and 4.94 l/day (2.18 ± 0.94) and that of women varied between 0.91 and 4.87 l/day (2.19 ± 0.75).
Intraindividual variability of 24-h urine output was calculated using two 24-h urine collections from a subset of 90 women and 133 men from the OPEN study. The mean difference between the two 24-h urine collections was 34.9 ml in women and 40.6 ml in men ( P < 0.001). Urine output did not vary significantly with age among women ( P 0.05), but the 60- to 69-yr-old men had significantly higher urine output than the 40- to 49- and 50- to 59-yr-old men, as seen in Fig. 4 ( P = 0.04 and 0.02, respectively).
Fig. 4. Percentile distribution of 24-h urine output for each age group within genders. The 90th, 75th, 25th, and 10th percentiles are indicated by the top error bar, the top of the box, the bottom of the box, and the bottom error bar, respectively. The median is indicated by the horizontal line dividing the box. Age groups with the same letter differ significantly within genders ( P < 0.05).
DISCUSSION
This is an unusually large set of objective data on water intake and urine production for American adults. Data on water turnover have proven to be elusive in the past because objective and accurate measures of water intake and urine production were lacking. Virtually no large population data are available regarding individual consumption of drinking water in the United States since the 1977-1978 National Food Consumption Survey data ( 6 ), and these data are based on self-reports as against the objective data presented here. Our analysis showed that preformed water intake varied from as low as 1.2 l/day to as high as 7.7 l/day among 458 individuals 40-79 yr old. There was a 15 ± 5 ml decrease in preformed water intake with every decade increase in age among both men and women. The decrease, however, was small and thus only discernable because of the large sample size. Interestingly, the preformed water intake in these individuals did not show a major change when the values were adjusted for their FFM, indicating that in this large sample older subjects are not prone to any greater risk of chronic low water intake than younger subjects. Similarly, the 24-h urinary output values did not show any age-related differences in women. The 24-h urinary output of 40- to 49- and 50- to 59-yr-old men, however, was significantly lower than in the 60- to 69-yr-old men.
Even though there were statistically significant age effects on water turnover and preformed water intake, the differences between the age groups were small, and thus the age-independent mean values for water intake in adults between the ages of 40 and 79 yr can be compared with the National Research Council daily recommendations for water intake ( 10 ). For this comparison, water input was expressed in milliliters per kilocalorie of energy expended. The water input per kilocalorie of energy expended in the current study was 1.1 ml·kcal -1 ·day -1 in men (TEE =2,746 ± 488 kcal/day) and 1.2 ml·kcal -1 ·day -1 in women (TEE = 2,138 ± 392 kcal/day), which is slightly higher than the National Research Council's recommendation of 1 ml/kcal of energy expended ( 11 ). Many (38%) of our participants, however, had water intakes that were lower than recommended. Despite this, hyperosmotic dehydration was not observed. Unfortunately, our primary study design was not directed toward the determination of water requirements, and hence we do not have data on health indicators such as urine osmolality, stone formation, or urinary tract infection rates, all of which are good predictors of hyperosmotic conditions and dehydration ( 12 ). These observations therefore do not directly refute the NRC recommendation.
It has often been recommended that individuals consume at least eight 8-oz glasses of water each day ( 1.9 l/day) ( 37 ); however, it is not clear what data formed the basis for this recommendation. Our data indicate that it is not based on an averaged actual intake. If an intake of 64 oz of water were combined with the typical water content of food ( 1 l/day) ( 10 ) and individuals drank no other beverages, intake would be about 3 l/day or 1.2 ml·kcal -1 ·day -1 (as for our participants). If they drank other beverages in addition to the 64 oz of water, intake would be even greater. Preformed water intake at or above this level was observed in only 36% of our participants. Therefore, it is clear from these data that the recommendation that individuals consume eight 8-oz glasses of water daily is not consistent with observed water intake in these healthy adults living in the U.S.
Our findings indicate that preformed water intake and urinary output were highly variable among individuals. Regression analysis of preformed water with urinary output and anthropometric variables (age, weight, BMI, and height) showed that only 4-8% of the variability was explained by anthropometric variables. Thus it is likely that individual behavior and not the physiological differences we investigated account for most of these large interindividual variations. In addition, we found that urinary output accounted for 66% of the total water efflux, rather than the 50% assumed in the past ( 10 ). This percentage, however, is likely to be dependent on preformed water intake and ambient conditions.
It is unlikely that the large variability results from measurement error. The methods used herein to measure preformed water intake and urinary output are more accurate than methods used in previous research. The urine collection employed PABA recovery to confirm complete collections and thus reduced errors associated with incomplete collections. The intake data are derived not from self-report but from the 2 H technique, and this has been validated for accuracy in animal models and humans ( 7, 23, 31 ).
An important caveat is that the 2 H method does not directly measure water intake. Preformed water intake constitutes only 80-85% of water turnover volume. Because of this, we corrected water turnover values for metabolic water, inspired water, and transcutaneous water, which are estimated to constitute 16-18% of water influx volume. The assumptions made in these calculations are not perfectly accurate, but even if our assumption were in error by 25% for these minor routes of water metabolism, the relative error in preformed water intake would be <5%. As such, the accuracy of our estimates of preformed water intake is probably very high.
Our urinary output data should also be considered accurate. Normally, 24-h urine collections are prone to error due to incomplete collection; however, we reduced this error by employing PABA as a tracer for complete collection ( 4, 5, 17, 18 ). The average PABA recovery was 103 ± 14% for the data reported herein, indicating little mean collection error.
Perhaps the greatest limitation in our methods was the use of DXA to assess %hydration of FFM. DXA is a widely used method for measurement of body composition in humans ( 15 ). However, there is a limit in the use of DXA in estimating the hydration of FFM. Changes in fluid balance in the body cause a small systematic and predictable error in DXA soft tissue composition analysis and can result in a misassessment of dehydration in the body. In a validation study by Lukaski et al. ( 21 ), Sprague-Dawley rats were exposed to a variety of dietary stressors, and changes in body composition were measured using DXA, decrease in body mass, and hydration status. DXA underestimated the body mass and FFM by 3% and significantly overestimated the fat mass, the greatest errors occurring in treatment groups in which body mass was diminished and body hydration was decreased ( 21 ). In addition, %hydration of FFM is, by its very nature, an insensitive measure of dehydration. A decrease in the TBW by 2% of body weight in a person with 20% body fat will result in a 2.5% decrease in FFM and a 3.4% decrease in TBW. The combined effect is a decrease in FFM hydration of only 1%. Whether this amount of decrease in hydration is detectable by DXA and TBW in humans is questionable. Also, DXA is known to be sensitive to changes in electrolytes. If a hyperosmotic condition exists, as in thermal dehydration where there is an increased loss of fluid relative to electrolytes, the changes in X-ray absorptions can lead to an overestimation of FFM by DXA ( 27 ), thereby masking the hydration status differences among age groups. The potential impact of this confounder cannot be ascertained in our data set.
Other limitations of this study include the fact that the participants were selected from only three geographical areas (suburban Washington, DC, Pittsburgh, PA, and Memphis, TN) and that most were studied in fall or winter. Thus our data are limited both geographically and seasonally. Individuals living in hot climates would be expected to have increased sweat loss, decreased urine loss, and/or increased preformed water intake. Thus the data presented here hold only for adults doing low to moderate exercise with little exposure to extremes of temperature and humidity. Racial differences were reported only in the elderly subjects (70-79 yr olds) due to nonavailability of data in the younger age groups. Differences could not be explained by body size, but we did not investigate other causes. These differences could be due to cultural differences with regard to fluid intake or differences in socioeconomic status. However, the above are purely speculations, and further studies need to be done to identify the racial effect on fluid intake.
One of our aims of this analysis was to test whether the elderly had low intakes of water that might predispose them to chronic dehydration. We found that, on average, the oldest group of individuals had a preformed water intake that was 98% that of the younger group of individuals when expressed per kilocalorie of energy expended. Although our methods had limited sensitivity, we did not find any evidence of dehydration in the 70- to 79-yr-old group, despite the majority of the individuals having intakes less than the commonly used suggestion of eight 8-oz glasses of water each day. Furthermore, recommendations to increase fluid intake to eight 8-oz glasses of water in the elderly may not be prudent because the elderly have an elevated risk of overhydration due to an attenuated osmoregulatory mechanism ( 22 ). Instead, it may be better to concentrate on recommendations for increasing fluid intake during periods of acute thermal stress ( 8 ).
This is a large set of objective data on water intake and urine production in adults using methods to assess preformed water input and urine production that are more accurate than those used to assess these parameters in the past. We found no evidence of chronic hyperosmotic dehydration among the elderly with lower than average preformed water intake, although the sensitivity of DXA to small changes in hydration is limited. Furthermore, these data do not address the observation that elderly individuals are more prone to dehydration under acute stresses of reduced fluid intake and excessive fluid losses ( 8 ). Although our findings are not applicable to all climates and regions, these data provide the largest sample to date for preformed water intake from beverages and foods and urinary output for adults between 40 and 79 yr of age.
GRANTS
This work was supported by National Institutes of Health Grants N01-AG-6-2106, N01-AG-6-2102, and N01-AG-6-2103 from the Health Aging and Body Composition study and also by the Observable Protein and Energy Nutrition study. A. Raman was supported by National Institutes of Health Grant DK-30031.
ACKNOWLEDGMENTS
We thank the staff at the Health ABC and OPEN field centers for subject recruitment and specimen collections. We also thank Dr. Stephane Blanc for expert advice on the statistical analysis.
【参考文献】
Allinger UG, Johansson GK, Gustafsson JA, and Rafter JJ. Shift from a mixed to a lactovegetarian diet: influence on acidic lipids in fecal water-a potential risk factor for colon cancer. Am J Clin Nutr 50: 992-996, 1989.
Anti M, Pignataro G, Armuzzi A, Valenti A, Iascone E, Marmo R, Lamazza A, Pretaroli AR, Pace V, Leo P, Castelli A, and Gasbarrini G. Water supplementation enhances the effect of high-fiber diet on stool frequency and laxative consumption in adult patients with functional constipation. Hepatogastroenterology 45: 727-732, 1998.
Beck LH. Changes in renal function with aging. Clin Geriatr Med 14: 199-209, 1998.
Bingham SA and Cummings JH. Creatinine and PABA as markers for completeness of collection of 24-hour urine samples. Hum Nutr Clin Nutr 40: 473-476, 1986.
Bingham SA, Williams R, Cole TJ, Price CP, and Cummings JH. Reference values for analytes of 24-h urine collections known to be complete. Ann Clin Biochem 25: 610-619, 1988.
Blanc S, Colligan AS, Trabulsi J, Harris T, Everhart JE, Bauer D, and Schoeller DA. Influence of delayed isotopic equilibration in urine on the accuracy of the (2)H(2)(18)O method in the elderly. J Appl Physiol 92: 1036-1044, 2002.
Cole TJ and Coward WA. Precision and accuracy of doubly labeled water energy expenditure by multipoint and two-point methods. Am J Physiol Endocrinol Metab 263: E965-E973, 1992.
De Castro JM. Age-related changes in natural spontaneous fluid ingestion and thirst in humans. J Gerontol 47: 321-330, 1992.
Fjeld CR, Brown KH, and Schoeller DA. Validation of the deuterium oxide method for measuring average daily milk intake in infants. Am J Clin Nutr 48: 671-679, 1988.
Food and Nutrition Board. Recommended Dietary Allowances (9th ed.). Washington, DC: National Academy Press, 1980.
Food and Nutrition Board. Recommended Dietary Allowances (10th ed.). Washington, DC: National Academy Press, 1989.
Francesconi RP, Hubbard RW, Szlyk PC, Schnakenberg D, Carlson D, Leva N, Sils I, Hubbard L, Pease V, Young J, and et al. Urinary and hematologic indexes of hypohydration. J Appl Physiol 62: 1271-1276, 1987.
Guyton AC. Textbook of Medical Physiology. Philadelphia, PA: Saunders, 2000.
Heller KE, Sohn W, Burt BA, and Eklund SA. Water consumption in the United States in 1994-96 and implications for water fluoridation policy. J Public Health Dent 59: 3-11, 1999.
Heymsfield SB, Smith R, Aulet M, Bensen B, Lichtman S, Wang J, and Pierson RN Jr. Appendicular skeletal muscle mass: measurement by dual-photon absorptiometry. Am J Clin Nutr 52: 214-218, 1990.
Human Nutrition Information Service. Nationwide Food Consumption Survey (NFCS). Nutrient Intakes: Individuals in 48 States, Year 1977-78. Hyattsville, MD: Consumer Nutrition Division, US Department of Agriculture, 1984.
Jakobsen J, Ovesen L, Fagt S, and Pedersen AN. Para-aminobenzoic acid used as a marker for completeness of 24 hour urine: assessment of control limits for a specific HPLC method. Eur J Clin Nutr 51: 514-519, 1997.
Johansson G, Bingham S, and Vahter M. A method to compensate for incomplete 24-hour urine collections in nutritional epidemiology studies. Public Health Nutr 2: 587-591, 1999.
Johansson G, Wikman A, Ahren AM, Hallmans G, and Johansson I. Underreporting of energy intake in repeated 24-hour recalls related to gender, age, weight status, day of interview, educational level, reported food intake, smoking habits and area of living. Public Health Nutr 4: 919-927, 2001.
Kleiner SM. Water: an essential but overlooked nutrient. J Am Diet Assoc 99: 200-206, 1999.
Lukaski HC, Hall CB, Marchello MJ, and Siders WA. Validation of dual x-ray absorptiometry for body-composition assessment of rats exposed to dietary stressors. Nutrition 17: 607-613, 2001.
Mack GW, Stachenfeld NS, and DiPietro L. Altered osmotic and volume control of thirst with aging. In: Hydration Throughout Life, edited by Arnaud MJ. France: John Libbey Eurotext, 1998, p. 127-136.
Michaud DS, Spiegelman D, Clinton SK, Rimm EB, Curhan GC, Willett WC, and Giovannucci EL. Fluid intake and the risk of bladder cancer in men. N Engl J Med 340: 1390-1397, 1999.
Nagy KA and Costa DP. Water flux in animals: analysis of potential errors in the tritiated water method. Am J Physiol Regul Integr Comp Physiol 238: R454-R465, 1980.
Noakes TD. Fluid and electrolyte disturbances in heat illness. Int J Sports Med 19: S146-S149, 1998.
Pearle MS. Prevention of nephrolithiasis. Curr Opin Nephrol Hypertens 10: 203-209, 2001.
Pietrobelli A, Wang Z, Formica C, and Heymsfield SB. Dual-energy X-ray absorptiometry: fat estimation errors due to variation in soft tissue hydration. Am J Physiol Endocrinol Metab 274: E808-E816, 1998.
Racette SB, Schoeller DA, Luke AH, Shay K, Hnilicka J, and Kushner RF. Relative dilution spaces of 2 H- and 18 O-labeled water in humans. Am J Physiol Endocrinol Metab 267: E585-E590, 1994.
Reddy BS, Engle A, Simi B, O'Brien LT, Barnard RJ, Pritikin N, and Wynder EL. Effect of low-fat, high-carbohydrate, high-fiber diet on fecal bile acids and neutral sterols. Prev Med 17: 432-439, 1988.
Schoeller DA, Colligan AS, Shriver T, Avak H, and Bartok-Olson C. Use of an automated chromium reduction system for hydrogen isotope ratio analysis of physiological fluids applied to doubly labeled water analysis. J Mass Spectrom 35: 1128-1132, 2000.
Schoeller DA, Ravussin E, Schutz Y, Acheson KJ, Baertschi P, and Jequier E. Energy expenditure by doubly labeled water: validation in humans and proposed calculation. Am J Physiol Regul Integr Comp Physiol 250: R823-R830, 1986.
Schoeller DA and van Santen E. Measurement of energy expenditure in humans by doubly labeled water method. J Appl Physiol 53: 955-959, 1982.
Steen B. Body composition and aging. Nutr Rev 46: 45-51, 1988.
Subar AF, Kipnis V, Troiano RP, Midthune D, Schoeller DA, Bingham S, Sharbaugh CO, Trabulsi J, Runswick S, Ballard-Barbash R, Sunshine J, and Schatzkin A. Using intake biomarkers to evaluate the extent of dietary misreporting in a large sample of adults: the OPEN study. Am J Epidemiol 158: 1-13, 2003.
Tang R, Wang JY, Lo SK, and Hsieh LL. Physical activity, water intake and risk of colorectal cancer in Taiwan: a hospital-based case-control study. Int J Cancer 82: 484-489, 1999.
Trabulsi J and Schoeller DA. Evaluation of dietary assessment instruments against doubly labeled water, a biomarker of habitual energy intake. Am J Physiol Endocrinol Metab 281: E891-E899, 2001.
Valtin H. "Drink at least eight glasses of water a day." Really? Is there scientific evidence for "8 x 8"? Am J Physiol Regul Integr Comp Physiol 283: R993-R1004, 2002.
作者单位：1 Department of Nutritional Sciences, University of Wisconsin-Madison, Madison, Wisconsin 53706; 2 Division of Cancer Control and Population Sciences, Applied Research Program, and 3 Division of Cancer Epidemiology and Genetics, Nutritional Epidemiology Branch, National Cancer Institute, 4 National