点击显示 收起
1 From the Department of Nutritional Sciences, University of California at Berkeley, and the Western Human Nutrition Research Center, US Department of Agriculture, Agricultural Research Service, University of California at Davis.
2 Supported in part by a gift from Mead Johnson Nutritionals, a Bristol-Myers Squibb company, and the University of California Agricultural Experimental Station, Department of Nutritional Sciences, University of California at Berkeley.
3 Address reprint requests to JC King, Western Human Nutrition Research Center, University of California, Davis, CA 95616. E-mail: jking{at}whnrc.usda.gov.
ABSTRACT
Background: Adjustment of gastrointestinal absorption is the primary means of maintaining zinc homeostasis; however, a precise, accurate method for measuring zinc absorption in humans has not been identified.
Objective: The purpose of this study was to compare the estimates of the fraction of dietary zinc absorbed (FZA) by using 4 stable isotopic tracer methods: mass balance (MB) corrected for endogenous secretion, fecal monitoring (FM), deconvolution analysis (DA), and the double isotopic tracer ratio (DITR) method.
Design: All 4 methods were applied to a single data set for each of 6 women. FZA was also determined for each subject by using a detailed compartmental model of zinc metabolism, and that value was used as the reference with which the simpler methods were compared.
Results: The estimates of FZA ( ± SD) determined by DA (0.27 ± 0.08) and the DITR technique in plasma (0.30 ± 0.10), 24-h urine samples (0.29 ± 0.09), and spot urine samples (0.291 ± 0.089) all compared well with the FZA reference value from the compartmental model (0.30 ± 0.10). The MB and FM methods tended to overestimate FZA compared with the reference value.
Conclusions: The determination of FZA by MB or FM is laborious, is sensitive to subject compliance, and may result in an overestimate. DA, although relatively accurate, has the disadvantage of requiring multiple blood drawings over several days. In contrast, the DITR technique applied to a spot urine specimen obtained 3 d after tracer administration provides an accurate measure of FZA and is easy to implement; therefore, it is the recommended method for determination of FZA.
Key Words: Zinc • stable isotopes • zinc absorption • compartmental modeling • stable isotopic tracer methods • mass balance • fecal monitoring • deconvolution analysis • double isotopic tracer ratio • women
INTRODUCTION
Zinc homeostasis is regulated primarily by changes in zinc absorption and endogenous excretion (1). A reduction in zinc intake from 5.5 to 0.8 mg/d caused a 2-fold increase in the fraction of dietary zinc absorbed (FZA) (2). Also, FZA appears to increase when zinc status is poor (3, 4). Therefore, the ability to measure FZA is essential to studies of zinc metabolism and homeostasis.
Previous studies of zinc absorption were limited by the lack of a precise, accurate measure of intestinal zinc uptake. The mass balance (MB) technique, which measures the difference between dietary zinc intake and fecal output, was used extensively in the past but, because unabsorbed dietary zinc and zinc derived from endogenous sources are both included in the fecal output value, MB measures only net absorption. To measure the homeostatic response to changes in dietary zinc intake, a measure of true absorption, ie, the uptake and transfer of dietary zinc across the mucosal cells, is needed. With the availability of stable isotopic tracers, it became possible to quantify the amount of fecal zinc derived from endogenous sources (5–7). With this information, the MB technique can be corrected to calculate FZA. Alternatively, after an oral stable isotope dose, FZA can be determined from the difference between the amount of isotope administered orally and the excretion of unabsorbed isotope in the stool (5). This approach is called the fecal monitoring (FM) method. However, some of the absorbed tracer is resecreted into the gut and excreted into the feces during the collection period, thereby causing an underestimate of FZA. Methods for correcting the FM technique for the amount of absorbed and resecreted tracer excreted in the feces were suggested by English et al (8) and by Rauscher and Fairweather-Tait (9).
Two other techniques available for determination of FZA are deconvolution analysis (DA) and the double isotopic tracer ratio (DITR) method. Both methods require administration of 2 stable isotopic tracers of zinc, 1 oral and 1 intravenous. With the DA method, both zinc tracers are measured in the plasma over several days; samples are taken frequently during the first 6 h after tracer administration. The fraction of the oral dose absorbed is then calculated from the plasma tracer concentrations by deconvolution (10–12). With the DITR method, FZA is estimated after simultaneous oral and intravenous administration of 2 different stable isotopic tracers of zinc. The plasma or urinary ratio of the oral to the intravenously administered tracer after correction for differences in dose provides an estimate of FZA (13).
The purpose of this study was to compare the estimates of FZA by using the MB, FM, DA, and DITR methods. All estimates were made simultaneously in 6 free-living women. The reference value against which these estimates were compared was calculated from a compartmental model of zinc metabolism. This model was derived by using the isotope enrichment data for plasma, urine, and feces (14). It was assumed that the model provided the most reliable estimation of FZA because more information was used to derive this estimate of FZA than was used for any of the other methods. Therefore, the overall aim of this study was to identify the method or methods for measuring FZA that agree best with the value derived from the detailed compartmental model of zinc metabolism in vivo.
SUBJECTS AND METHODS
Subjects
Six white women with no acute or chronic health problems aged 30 ± 11 y ( ± SD) were recruited for the study [weight: 54.2 ± 8.9 kg; body mass index (in kg/m2): 20.7 ± 2.6]. The usual dietary zinc intakes of each woman were assessed before the study by using 3-d weighed-food diet records. Dietary zinc intake was measured by using a computerized database (NUTRITIONIST III; N-Squared Computing, Salem, OR), which we updated by adding any missing information on the zinc content of foods. The women reported consuming 8.3 ± 3.4 mg Zn/d. The experimental design of this study was approved by the University of California at Berkeley Committee for the Protection of Human Subjects. All participants gave written, informed consent.
Experimental design
Details of the experimental design and preparation of stable isotopic tracers for administration were described elsewhere (14). In brief, subjects were free living and consumed a constant diet providing 7 mg Zn/d for a 7-d equilibration period before isotopic tracer administration. Energy was adjusted to match each subject's reported intake from the 3-d weighed-food diet records completed before the start of the study. The subjects were weighed on days 2, 5, and 7 of the equilibration period; adjustments were made to the diet if necessary to maintain a constant body weight. Baseline urine and fecal samples were collected on day 6. A fecal marker [1 g polyethylene glycol (PEG) (Sigma Chemical Company, St Louis) in 3 mL distilled water] was administered orally 2 h after the evening meal on day 7 to mark the beginning of the 6-d metabolic balance period.
On the morning of day 8, the subjects arrived at the metabolic unit having fasted since 2000 the previous evening. An indwelling catheter was placed in an arm vein and a fasting blood sample (8 mL) was drawn into a zinc-free plastic syringe (Monovette; Sarstedt, Hayward, CA). Fifteen minutes after a breakfast meal containing 1 mg Zn, each subject consumed 1.3 mg of the oral isotopic tracer highly enriched in 67Zn that had equilibrated in 213 g orange juice overnight. Immediately thereafter, 0.4 mg of another stable isotopic tracer highly enriched in 70Zn was infused intravenously over a period of 1 min into a vein in the arm without the indwelling catheter. Blood samples (8 mL) were taken via the catheter 5, 10, 15, 30, 45, 60, 75, 90, and 120 min and 3, 4, 5, 6, and 7 h postinfusion. The samples were placed on ice and the plasma was separated by centrifugation (13600 x g for 3 min at room temperature) within 1 h of collection.
Blood samples were taken daily for the next 7 d, during which the constant diet was consumed, and 24-h urine samples and total fecal samples were obtained. The urine samples were collected in 2 portions; the first morning void (spot urine sample) was collected separately from the rest of the day's output. A mock 24-h urine sample was prepared by combining appropriate volumes of the spot urine sample and the rest of the output from the previous day, such that the volumes of each combined were in their respective proportion to the total 24-h volume of urine collected. Urine samples were acidified with 1 mL concentrated HCl (Seastar; Chemicals, Inc, Seattle) per 125 mL urine before storage.
A second dose of PEG (1 g) was taken 2 h after the evening meal on day 13 to mark the end of the metabolic balance period. The subjects consumed a self-selected diet for the remaining 5 d and continued to collect all fecal output. All plasma, urine, and fecal samples were stored at -20°C.
Sample preparation and analysis
Fecal samples were lyophilized to constant weight and ground to homogeneity. The total zinc concentration of the plasma, urine, and fecal samples was measured by atomic absorption spectroscopy (Thermo Jarrell Ash, Franklin, MA) as described previously (14). Isotopic mass ratios based on a reference isotope of 66Zn were measured by using inductively coupled plasma mass spectroscopy and were converted first to tracer-to-tracee ratios and finally to tracer mass (in mg) according to equations 1–4, which were reported previously (14). The latter were then expressed as a percentage of administered tracer dose/L plasma [assuming plasma volume estimates for each subject as described previously (14)], percentage of dose per sample for urine, and cumulative percentage or fraction of administered tracer dose for feces. A Sciex ELAN 500 ICP mass spectrometer (Perkin-Elmer, Norwalk, CT) was used, equipped with a U-5000AT ultrasonic nebulizer (Cetac Technologies Inc, Omaha) and a 212B autosampler (Gilson Medical Electronics Inc, Middleton, WI). Mass bias drift was corrected by using gallium as an isotope ratio internal standard (15). Plasma and fecal samples were wet ashed in concentrated nitric acid by microwave digestion (MDS 2000; CEM Corporation, Matthews, NC) and purified zinc was separated by ion exchange chromatography (14). Macrominerals were removed from the urine samples by using a chelex resin before separating purified zinc by ion-exchange chromatography (14). Fecal PEG content was measured by using a modification of the method of Allen et al (16).
Estimation of fraction of dietary zinc absorbed
Mass balance corrected for endogenous zinc excretion
FZA was estimated by using equation 1, where D is the total dietary zinc intake, F is the cumulative fecal zinc output for the 6-d balance period, and S is the amount of zinc excreted endogenously during that period.
RESULTS
Dietary zinc intake, fecal zinc output, endogenous zinc excretion, and FZA calculated by using the corrected MB techniques are shown in Table 1. The uncorrected MB technique estimated that the average net absorption (dietary zinc intake minus fecal zinc output) of zinc over the 6-d metabolic balance period was 13.6% of that consumed, or 1 mg Zn/d. Net absorption is lower than true absorption (the uptake and transfer of dietary zinc across the mucosal cells) because fecal zinc is composed of both unabsorbed dietary zinc and zinc that has been absorbed and then excreted from endogenous pools into the stools. Endogenous fecal zinc excretion was estimated to be 2.29 and 1.87 mg/d by using the MB-J and MB-Y methods, respectively. These values were not significantly different from each other. When MB was corrected for endogenous zinc excretion, FZA was 0.46 and 0.40, respectively (Table 1). These values were not significantly different from each other.
View this table:
TABLE 1.. Fraction of dietary zinc absorbed (FZA), estimated by mass balance corrected for endogenous zinc excretion1
Estimates of FZA for the 6 subjects, determined from a detailed compartmental model (reference value; 14) using equation 13 and from the 7 other tracer methods, are summarized in Table 2. FZA estimated from the compartmental model averaged 0.301. When the FM method was used, 16 of the 17 estimates of FZA were higher than the respective values obtained by using the compartmental model. In subject 6, the FM-E method could not be applied because of an insufficient number of stool samples. All 6 estimates of FZA derived from the DA method were lower than the respective model values. FZA values derived from the DITR technique tended to bracket the respective values for the compartmental model. Each individual value in the table for the 3 DITR methods represents the average of 5 daily determinations from days 3 to 7 after administration of the dual tracer.
View this table:
TABLE 2.. Fraction of dietary zinc absorbed (FZA) estimated by various methods1
The differences between FZA estimated from each tracer method and the model-derived reference value based on the Bland and Altman approach (17) are shown in Figure 3. The 2 corrected MB methods substantially overestimated FZA compared with the reference value by 0.156 ± 0.063 and 0.097 ± 0.084 ( ± SD), respectively. Two of the 3 FM methods provided estimates of FZA closer to the reference values; the average overestimates of FZA by the 3 FM methods were 0.034 ± 0.045, 0.074 ± 0.035, and 0.111 ± 0.035 for FM-N, FM-E, and FM-R, respectively. The DA technique underestimated FZA compared with the reference value in all 6 subjects, but the magnitude of the error was generally smaller than that for FM, averaging -0.027 ± 0.015. The DITR methods produced average values closest to the reference FZA value, with differences of -0.003 ± 0.026, -0.015 ± .033, and -0.010 ± 0.037 for plasma, 24-h urine samples, and spot urine samples, respectively. The variability of the values produced by the DITR method was slightly larger than that for DA. A comparison of the FZA estimated from each method compared with that determined from the model using ANOVA showed that MB-J, MB-Y, and FM-R gave estimates of FZA that were significantly different from the reference value (P < 0.001, P < 0.001, and P < 0.01, respectively).
FIGURE 3. . Comparison of 8 simple tracer methods for estimating the fraction of dietary zinc absorbed (FZA) with the value obtained from a compartmental model using the method of Bland and Altman (18). Plotted are the means ± SDs of the differences between the FZA values obtained by the simple techniques and that obtained from the compartmental model for all 6 subjects—mass balance corrected for endogenous zinc excretion determined by using the isotope dilution method (MB-J), mass balance corrected for endogenous fecal zinc determined by using the cumulative tracer excretion method (MB-Y), fecal monitoring without correction for resecretion of absorbed oral tracer (FM-N) (5; Equation 4), fecal monitoring using the corrections of English et al (FM-E) (8; Equation 5), and the method of Rauscher and Fairweather-Tait (FM-R) (9; equation 6) [deconvolution analysis (DA) (Equation 11) and dual isotopic tracer ratios averaged over days 3–7 in plasma (DI-Pl), 24-h urine samples (DI-24U), and spot urine samples (DI-spot) (13; Equation 12)].
DISCUSSION
Stable isotopic tracers are useful for measuring mineral absorption. In this article, we used 4 previously published stable isotopic tracer techniques that yielded 9 different estimates of FZA. The same data set from a group of 6 women who were fed a standard zinc diet was used for all comparisons. Estimated values for FZA based on these relatively simple measures of the data were compared with a reference value for FZA obtained from a detailed, physiologically based compartmental model fitted to all of the tracer and tracee data for each subject.
Estimates of FZA from the compartmental model averaged 0.301 (reference value) in these 6 women consuming a standard diet containing 7 mg Zn/d (Table 2). Both the DA technique, applied to plasma tracer data over the first 7 d after the isotope dose, and the DITR technique, whether estimated from plasma, 24-h urine samples, or spot urine samples averaged over days 3–7, provided reliable and comparable approximations of the reference value of FZA (Table 2).
The 2 corrected MB methods significantly overestimated FZA compared with the reference value. Endogenous zinc excretion estimated by using both the Jackson et al (6) and Yergey (7) methods agree well with values predicted by using the compartmental model, published previously, which averaged 2.01 ± 0.34 mg/d (14). Possible sources of error in the MB techniques include inaccurate measurements of dietary zinc intake and fecal zinc output.
All of the FM methods overestimated FZA compared with the reference value (Figure 3). Wastney and Henkin (19) showed that, in theory, the FM-N technique overestimates FZA for short fecal-collection periods and underestimates it for longer collection periods. The uncorrected technique gives an accurate value only over a narrow band of days. Because the collection period that gives accurate values differs among subjects (as a result of variance in rates of fractional absorption, fractional secretion, and gastrointestinal transit time) and because the best collection period is unknown, Wastney and Henkin (19) concluded that the FM technique should not be used to estimate zinc absorption.
A critical problem with the FM technique is that previously absorbed and resecreted oral tracer is excreted along with the unabsorbed tracer in the feces and there is no way to differentiate between the 2 sources of tracer in the feces. Several investigators attempted to provide a solution to this problem. One solution (FM-E), detailed in a review by Krebs et al (20), is based on the assumption that the rate of excretion of resecreted absorbed tracer is constant and can be determined from the slope of the fecal accumulation of the tracer plotted as a function of time after all of the unabsorbed oral tracer has passed through the gastrointestinal tract. On the basis of this assumption, extrapolation of this slope back to zero time should correct for all of the resecreted oral tracer. When applied to our data over a 12-d collection period, this correction leads to a larger overestimate of FZA than does no correction at all (Figure 3). Another potential solution to this problem (FM-R) was proposed by Rauscher and Fairweather-Tait (9), who attempted to correct for resecretion by monitoring the accumulation of an intravenous zinc tracer in the feces and used this information to correct the fecal secretion of orally administered tracer according to equation 6. When applied to our data, this correction leads to an even greater overestimate of FZA (Figure 3) compared with our reference estimate.
In summary, we compared 4 different techniques for estimating FZA—MB, FM, DA, and DITR—with a reference value derived from a compartmental model. The MB method substantially overestimates FZA. The FM technique, even when corrected for resecretion of absorbed oral tracer, also overestimates FZA. The DA and DITR techniques provide estimates of FZA that are close to the model derived value. The DA technique requires multiple blood draws over several days to define the response of the orally and intravenously administered tracers in plasma. In contrast, the DITR technique requires only a single plasma sample, a 24-h urine sample, or a spot urine sample obtained 2 d after tracer administration. The spot urine sample requires the least subject involvement of all the procedures. We therefore recommend the DITR technique with use of a spot urine sample collected 2 d after tracer administration (or the average of several spot urine samples) as the method of choice for estimation of FZA when detailed compartmental modeling and the extensive sampling it requires cannot be performed.
ACKNOWLEDGMENTS
We thank David Shames for his assistance with the mathematical modeling and data interpretation.
REFERENCES