Dysprosium as a nonabsorbable fecal marker in studies of zinc homeostasis
the Section of Nutrition, Department of Pediatrics, University of Colorado Health Sciences Center, Denver, CO
ABSTRACT
Background: Dysprosium is a nonabsorbable rare earth element that has had successful application as a marker for fecal excretion of unabsorbed zinc.
Objective: Our goals were 1) to evaluate the efficacy of administering dysprosium with all meals over several days as a method of determining the completeness of fecal collections, 2) to determine the similarity of gastrointestinal transit kinetics and excretion patterns of dysprosium and zinc tracer administered simultaneously over several days, and 3) to evaluate alternative methods of using the data for fecal excretion of orally administered zinc tracer and dysprosium to measure the fractional absorption of zinc.
Design: 70Zn and dysprosium were administered orally with all meals for 5 consecutive days to 7 healthy, free-living adults consuming a constant diet based on habitual intake. Additional tracers, 67Zn and 68Zn, were administered intravenously. Urine and fecal samples were collected during tracer administration and for 8 d after the last dose. Isotope ratios were measured in urine and feces, and total zinc and dysprosium were measured in fecal samples.
Results: The mean recovery of dysprosium was 101.3 ± 2.4%. The zinc oral tracer and dysprosium had similar fecal excretory patterns; the correlation coefficient for 70Zn and dysprosium in fecal samples exceeded 0.99 (P < 0.0001) for each subject. Fractional zinc absorption measurements using various dysprosium methods correlated well (r > 0.95) with those from the fecal monitoring and dual-isotope-tracer ratio methods.
Conclusion: Administration of dysprosium is a useful means of determining the completeness of fecal collections and of measuring zinc absorption.
Key Words: Zinc dysprosium fecal marker absorption stable isotopes dual-isotope-tracer ratio technique fecal monitoring method
INTRODUCTION
Dysprosium is a rare earth element that is nontoxic, nonabsorbable, present at no higher than trace amounts in the diet, and measured quite easily. It was first used as a quantitative fecal marker in zinc absorption studies in 1993 (1) and has subsequently had minimal reported application in studies of zinc metabolism. We are currently using dysprosium in studies ranging from measurements of the fractional absorption of zinc (FAZ) to complex studies of zinc homeostasis with the use of zinc stable-isotope-tracer techniques. Before using dysprosium in our research program, we undertook additional evaluation of the use of dysprosium and the interpretation of data derived from the application of this rare earth element in studies of zinc metabolism. The present article reports the results of this evaluation.
The first objective of the present study was to evaluate the efficacy of administering dysprosium with all meals over several days as a method for determining the completeness of fecal collections for free-living human metabolic studies. The second objective was to determine the similarity of gastrointestinal transit kinetics and excretion patterns of dysprosium and isotopically enriched zinc tracer after simultaneous oral administration in individual subjects and under the circumstances of this multiday metabolic study. The third objective was to evaluate alternative methods of using the data for fecal excretion of orally administered zinc tracer and dysprosium to determine FAZ and to compare the results from these methods with FAZ values derived simultaneously by using the dual-isotope-tracer ratio (DITR) technique and a fecal-monitoring (FM) method (2, 3). This comparison included an examination of the difference between results attributable to the fact that, unlike the DITR and FM methods, the dysprosium methods generally do not account for absorbed tracer that is subsequently secreted into the intestinal lumen and excreted in the feces, herein referred to as endogenous tracer.
SUBJECTS AND METHODS
Study design
This research was part of a larger protocol designed to evaluate selected zinc stable-isotope techniques for the investigation of human zinc homeostasis. Free-living, apparently healthy subjects were studied while consuming a constant daily diet that was based on their habitual diet and was provided daily by the diet kitchen of the General Clinical Research Center at the University of Colorado Hospital (day –6 to day 5). Four days before the start of dysprosium and oral tracer administration, a zinc stable-isotope tracer (68Zn) was administered intravenously. Then, starting at dinner on day 0, each meal was labeled with another zinc tracer (70Zn) together with dysprosium until dinner on day 5. A third tracer (67Zn) was administered intravenously on day 3 at 1230. This time was selected on the basis of the results of a simulation of the study protocol with the use of our published model of zinc metabolism (4), which identified 1230 as the optimal time for intravenous tracer administration for accurate calculation of FAZ by the DITR technique. Complete fecal samples were collected from the time of administration of the first dose of dysprosium until the end of day 13. Timed urine samples were collected on days 0–13. The study timeline is summarized in Figure 1.
Subjects
The participants were 7 healthy volunteers, 6 women and 1 man, aged 23–39 y ( ± SD: 34.1 ± 5.4 y) with a mean (±SD) body mass index (in kg/m2) of 22.1 ± 2.1. The study was approved by the Colorado Multiple Institutional Review Board, and all subjects gave their written consent to take part in the study.
Study diet
A 7-d diet record was analyzed by using the Nutrition Data System for Research, version 4.04-32 (Nutrition Coordinating Center, University of Minnesota, Minneapolis, MN). The subjects then met with a research dietitian to select a 3-d rotating diet with energy, zinc, and phytate intakes matching those calculated from the 7-d record. The constant daily diet was provided daily by the University of Colorado General Clinical Research Center diet kitchen from day –6 to day 5. During the metabolic study period, the subjects consumed all their meals under the supervision of one of the investigating team.
Isotope preparation and administration
Enriched stable isotopes of zinc were obtained from Trace Science International (Ontario, Canada). Accurately weighed quantities of each isotopically enriched preparation were dissolved in 0.5 mol H2SO4/L and then diluted in triply deionized water to prepare a stock solution. For preparation of orally administered doses, the stock solution of enriched 70Zn was diluted and titrated to pH 3.0 with metal-free ammonium hydroxide. This solution was filtered through a 0.22-μm filter to make a sterile solution. For preparation of intravenously administered doses of 67Zn and 68Zn, sterile techniques were used. The stock solution was diluted and adjusted to pH 6.0 and then filtered through a 0.22-μm filter to make a sterile solution. Concentrations of zinc in the isotope preparations were determined in triplicate by atomic absorption spectrophotometry, and concentration measurements were adjusted for the different atomic weights of the preparations.
An accurately weighed quantity of 67Zn (1 mg) was administered intravenously on day 3 at 1230. The tracer was administered over a 10-min interval via a scalp vein needle in a superficial forearm vein with a 3-way closed stopcock system. This allowed us to rinse the delivery syringe twice with N-saline contained in a second sterile syringe.
A total of 1 mg 70Zn (accurately weighed) was administered orally divided between all meals for 5 days commencing with dinner on day 0 and continuing through lunch on day 5 of the metabolic period. The tracer was administered in water gradually during the second half of each meal. The oral tracer was distributed among the meals in approximate proportion to the dietary zinc content of the meals. Dietary zinc intake ranged from 5.9 to 27.0 mg/d with an average of 11.3 ± 7.4 mg/d. The quantity of oral 70Zn tracer administrated with meals was 0.2 mg/d or a maximum of 3.4% of zinc in the diet.
Dysprosium, in the form of DyCl3 · 6H2O, was obtained from Sigma Aldrich (Milwaukee, WI). The dysprosium dose was prepared in purified, filtered water and its concentration was measured by inductively coupled plasma mass spectrometry (ICP-MS) (VG Plasma Quad 3; VG Elemental, Cheshire, United Kingdom). A total oral dose of 1 mg Dy was divided in proportion to the 70Zn doses and was administered in 70Zn solution taken with all 5 d of meals.
Sample collection
All stools were collected from the time of the first 70Zn-labeled meal until 8 d after the last 70Zn-labeled meal. Feces were collected separately and quantitatively in plastic bags. A clean, midstream-void urine sample was collected into a zinc-free plastic container twice daily from day 0 to day 13. The times for each collection were noted on the specimen cup and log sheets. Baseline fecal and urine specimens were obtained before administration of any label. All samples were frozen at –20°C until analyzed.
Sample preparation and analyses
Accurately weighed aliquots of homogenized feces and whole-day food samples were dried separately to a constant weight in an electric oven. All samples were prepared in duplicate. The dried samples were ashed in a muffle furnace at 450°C for 24 h. A few drops of concentrated nitric acid were added to the ash, which was then dried before reheating again at 450°C for 24 h.
Ashed fecal samples were reconstituted quantitatively in 50 mL, 6 mol HCl/L. The concentration of total zinc in these reconstituted fecal samples was determined on a diluted aliquot with an atomic absorption spectrophotometer fitted with a deuterium arc background-correction lamp (Perkin-Elmer Corporation, Norwalk, CT.)
For the measurements of zinc stable-isotope ratios, the inorganic elements were removed from reconstituted ashed fecal samples by ion-exchange chromatography with AG-1 ion-exchange resin (Bio-Rad Laboratories, Richmond, CA).
Urine samples were digested by using an MDA-2000 microwave sample preparation system (CEM Corp, Mathews, NC). A 5-mL urine sample was placed into an Advanced Composite Vessel and combined with 1 mL of concentrated HNO3, and the pressure was gradually increased to a maximum of 120 psi. The total digestion time was 90 min. Digested samples were transferred to a beaker, evaporated to dryness on a hot plate, and reconstituted in 2 mL ammonia acetate buffer (pH = 5.6). Zinc in the sample was purified by its chelation with trifluoroacetylacetone and then extraction of the chelate with hexane (5).
Isotope enrichment was determined by measurement of the isotope ratios 67Zn/66Zn, 68Zn/66Zn, and 70Zn/66Zn by ICP-MS. Tracer enrichment was defined as all zinc in the sample from the isotopically enriched tracer preparation divided by the total zinc in the sample.
The dysprosium contents of the test meal, doses, and fecal samples were measured by ICP-MS. Ashed fecal and food samples were reconstituted quantitatively in 50 mL 10% (by vol) HNO3. Bismuth (Aldrich Chemical Company Inc, Milwaukee, WI) was used as an internal standard. A large amount of bismuth was found in the fecal samples of one subject, so indium was used as an internal standard for the samples from this subject. The sample solution and dysprosium standards (Aldrich Chemical Company Inc, Milwaukee, WI) were diluted by using 6 parts per billion bismuth solution in 2% (by vol) HNO3 (OPTIMA; Fisher Scientific, Pittsburgh, PA). A set of dysprosium standards was inserted every 10 samples to make the external drift corrections. Two spiked fecal samples with dysprosium concentrations of 100 and 200 parts per billion, respectively, were used to check the accuracy of our measurements, and their recovery was 99.5% ±1.6 (6 measurements).
Data processing
The content of 70Zn tracer and dysprosium in the fecal samples was quantified as a fraction of the administered dose, except in the case of method 1b, for which dysprosium was quantified as the fraction of total dysprosium recovered. The 70Zn and dysprosium fecal excretory patterns for each subject were plotted and inspected. In addition, ratios of 70Zn to Dy were plotted verses sample time to discern any temporal differences in excretion not evident with the other comparison techniques. All fecal samples having dysprosium concentrations greater than the limit of detection were included in the calculation of total dysprosium recovery.
FAZ was calculated from the dysprosium and zinc tracer data by using 3 methods. In addition, method 1 was modified to take into account the presence of endogenous 70Zn tracer (method 1a) and was further modified to quantify dysprosium relative to the total dysprosium recovered instead of total dysprosium dose (method 1b).
In method 1, FAZ was calculated for each fecal sample by using the following equation:
where 70Zn in the sample is equal to total zinc in the sample times the 70Zn enrichment of the sample. Then, a mean FAZ was calculated for each subject. To eliminate samples giving unreliable FAZ results because of low dysprosium and unabsorbed 70Zn concentrations, only samples having a dysprosium-to-zinc mass ratio >0.25 of (dysprosium intake per day/average total zinc in feces per day) were included in the calculation of the mean. It was almost always the case that the first or second stool sample after administration of the first Dy-70Zn doses and all subsequent stool samples up to and including the first or second stool sample after administration of the final dose met this criterion. One exception was the case of a subject having diarrhea, for whom the included samples were the fourth stool sample after the first dose to the fourth stool sample after the final dose. Method 1a used the following calculation:
Endogenous 70Zn in the sample was determined with this calculation:
where urine enrichments are estimated for the time at which the fecal sample content was transiting the upper gastrointestinal tract.
Method 1b used a modification of method 1a:
With method 2, the dysprosium and 70Zn contents of all samples having a dysprosium concentration greater than the limit of detection were summed to calculate a single FAZ by using an equation equivalent to that used for method 1.
Method 3 derived an FAZ value from the slope of the regression of the fraction of the dysprosium dose in the sample versusthe fraction of the 70Zn dose in the sample by using the following equation:
As with method 2, all samples having dysprosium concentrations greater than the limit of detection were included in the regression analysis.
FAZ results from the dysprosium methods were compared with simultaneous calculations of FAZ made by using 2 established methods: one based on the measurement of dual (oral and intravenous)-isotope-tracer enrichment ratios in urine (DITR) (2) and a cumulative tracer in feces method that incorporates a correction for endogenous tracer (FM) (3). Because the oral tracer was administered over a much longer than usual period of time, optimum times for the administration of the intravenous tracer for the DITR method and for the extrapolation endpoint in the FM calculation had to be estimated to ensure accurate results. These were obtained from a simulation of the tracer administration and sampling protocols made with a compartmental model of human zinc metabolism (4). The simulation was performed by using WINSAAM software (WinSAAM Inc, University of Pennsylvania, Kennett Square, PA). In the FM method calculation, extrapolation to the midpoint of the oral tracer administration period was found to be satisfactory.
A compartmental model simulation of the dysprosium methods was also performed to predict the relation between the FAZ results from these methods and the DITR and FM methods attributable to the fact these dysprosium methods do not account for the endogenous oral tracer present in the feces.
Statistical analysis
All data are presented as means ± SDs unless otherwise stated. Comparison of FAZ results was performed by using repeated-measures analysis of variance (ANOVA) and Dunnett's multiple-comparison test. Statistical significance was defined as P < 0.05. Correlation analysis was used to compare the 70Zn and dysprosium content in fecal samples having dysprosium concentrations greater than the limit of detection and to compare the FAZ results from the dysprosium methods with those from the DITR and FM methods. Linear regression analysis was used to derive slope values for dysprosium method 3. All statistical analyses were performed by using GRAPHPAD PRISM version 3.01 (GraphPad Software Inc, San Diego, CA).
RESULTS
Objective #1: The individual quantities of dysprosium administered and the percentage recovered are given in Table 1. The mean recovery of dysprosium was 101.3 ± 2.4%. As expected, individual transit times varied; typically, however, most of the 70Zn and dysprosium had disappeared from the feces within 24 h of administration of the final doses. Dysprosium concentrations dropped below the limit of detection by 4.5 d after the last dose on average; the longest interval was 6.4 d. The limit of detection of dysprosium in fecal samples by ICP-MS was 2.5 parts per billion.
Objective #2: In all cases, the fecal excretory patterns of 70Zn and dysprosium were similar; an example for one subject is depicted in Figure 2. The fraction of zinc tracer excreted versus the fraction of dysprosium dose excreted in the fecal samples from the individual subjects is depicted in Figure 3. Linear regression lines are also shown. The correlation coefficient (r) exceeded 0.99 (P < 0.0001) for every subject. In addition, the linear regression analyses of these data indicated that the y intercepts were never significantly different from zero (the average y intercept was 0.00082). The number of samples from each subject included in the analysis, and plotted in Figure 3, ranged from 8 to 13. No consistent temporal pattern was observed in the plots of the ratio of zinc tracer to dysprosium versus sample time.
Objective #3: The results from the various methods of measuring FAZ are given in Table 2. Correlation coefficients for the dysprosium results compared with the 2 established methods are shown. In every case, r was >0.95 (P 0.0008). Repeated-measures ANOVA indicated no significant differences between FAZ results. Given this finding, Dunnett's test did not provide any additional information of value. The FAZ measurements from dysprosium methods 1, 1a, and 1b are plotted against the DITR measurements in Figure 4. The adjustment for endogenous 70Zn (method 1a) produced FAZ values 6.8% higher on average (range: 2.8–12.7%) than those of method 1, resulting in an average difference in FAZ of almost 0.02 in absolute terms. The model simulation predicted a difference of 4–5% between methods when one method does not take into account the secretion and excretion of endogenous tracer. The results from the modification of method 1 to account for endogenous tracer (method 1a) and to quantify dysprosium relative to that recovered (method 1b) both showed improved correlation and agreement with the DITR and FM results. Shown in Figure 5 is a Bland-Altman (6) plot of the results of methods 1a and 1b compared with DITR; the mean difference in FAZ was 0.012 and the limits of agreement were 0.036 and –0.013 for method 1b.
Analysis of the dysprosium content of the 5-d diet of subject number 1 showed very low concentrations of dysprosium, totaling little more than 0.1% of the total administered dose.
DISCUSSION
Dysprosium must be administered for several days in studies of zinc homeostasis that include quantitative fecal collections for, as an example, the measurement of excretion of endogenous zinc via the intestine, which is a measurement of cardinal importance in such studies. In the present study, our results provided important reassurance that the metabolic collections were complete in the free-living subjects.
Similar to other reports (1, 7), our mean recovery of dysprosium was slightly in excess of 100% of the quantity administered. Although we did not observe measurable concentrations of dysprosium in baseline fecal samples and found only minute concentrations of dysprosium when we analyzed the diet of one subject, others have measured more substantial amounts of dysprosium in baseline feces (8) and in test meals (7). Thus, it is possible that naturally occurring dysprosium in the diet may help explain the slight excess recovery in most of the subjects. In the only subject (number 3) having a dysprosium recovery of <100%, the dysprosium measurements may have been erroneously low due to the presence of bismuth and indium in this subject's diet, because these elements are used as internal standards in the ICP-MS analyses.
In considering the significance and possible causes of dysprosium recovery deviating from 100%, we noted that after the endogenous tracer was accounted for the magnitude and direction of the discrepancies between dysprosium values and DITR values exhibited a rough correspondence with the deviation of dysprosium recovery from 100%. In response, we tried an additional FAZ calculation method (1b) wherein sample dysprosium was quantified relative to the amount of dysprosium recovered instead of the total dysprosium dose, thereby improving the correlation and agreement with the DITR and FM value. This finding suggests that the quantity of administered dysprosium actually collected was generally closer to 100% than the recovery figures indicate and is consistent with several possible reasons for the discrepant recoveries, including those mentioned above (ie, dysprosium in the diet or analytic error in the case of subject 3). It also suggests that in carefully controlled studies such as this one, in which complete sample collection and rigorous quantitative processing are ensured, the calculation of FAZ by this method may be useful.
The similar dysprosium and zinc tracer excretory patterns and the high correlation between zinc and dysprosium in the fecal samples provide assurance that orally administered dysprosium can be assumed to mimic the gastrointestinal transit behavior of simultaneously administered zinc tracer very closely, and, as a result, can be most useful in isotopic tracer studies of zinc.
The FAZ results produced by the various dysprosium methods all correlated well (r > 0.95) with the DITR and FM methods. The results of method 1 are probably the most useful because they provide information on the accuracy and variability of FAZ determinations from individual samples and, therefore, the feasibility of using fewer samples (incomplete collections) to measure FAZ. Although the data are generally reassuring in this regard, we observed several instances in which accuracy and variability were adversely affected by the presence of extreme data at the beginning or end of the sample collection period. Therefore, it may be possible to obtain more accurate results by applying a more restrictive sample inclusion criterion.
An accuracy issue inherent with the dysprosium methods is that they do not take into account the effect of tracer that is absorbed and then secreted back into the gut and excreted in feces. Method 1a incorporated a measurement of endogenous 70Zn in each sample and calculated FAZ accordingly. The calculation of endogenous 70Zn in method 1a used measurements of 68Zn intravenous tracer enrichment in a manner related to an established method for measuring endogenous fecal zinc (EFZ) (3). The magnitude of the resulting adjustment to FAZ is modest in relation to the variability in the measurements, exceeding little more than 1 SD in the worst cases. The FAZ increases from method 1a were generally larger than the 4–5% predicted by the model. This is likely the result of the simulation using a population value that was slightly lower than the average EFZ rate observed in these subjects. Regarding the feasibility of using a basic dysprosium method to measure FAZ with sufficient accuracy, we believe that in most cases, where FAZ, EFZ, the exchangeable zinc pool size, and metabolic balance conditions are expected to be typical, a fixed adjustment to the FAZ measurement would be adequate. On the basis of our experience thus far, we suggest that, in the case of healthy adults in metabolic balance, FAZ values from dysprosium methods be increased by 5% as an approximate correction for endogenous tracer.
As discussed above, method 1b was implemented to address the concern with dysprosium recovery. Correlation and agreement with the DITR and FM results improved as a consequence. The improvement of agreement with method 1b compared with method 1a is evident in the Bland-Altman plot (Figure 5). The limits of agreement for method 1b are at least as good as typically observed when comparing FAZ methods. The mean difference indicates a positive bias of 0.01 in the dysprosium method relative to the DITR results. We do not know the cause of this apparent difference.
Dysprosium method 2 was included because it is a simple way of calculating a single FAZ value when complete collections have been performed and is the method generally reported by others. Method 3, too, is a simple calculation possible when a sufficient number of samples with various concentrations of dysprosium and tracer have been collected.
The different calculation methods provided similar results using varied ways of averaging the information from all the samples. To a great extent, the differences between results reflect the inherent variability routinely encountered with biological measurements of this kind. Correcting for endogenous tracer by measurement or estimation can minimize bias in FAZ results. Furthermore, when complete collections are carefully performed and dysprosium recovery measured, the possibility that accuracy may be improved by quantifying dysprosium relative to total dysprosium recovered instead of dysprosium administered should be considered. Modifications equivalent to those applied to method 1 could be applied to methods 2 and 3 with a similar outcome.
Although all 3 FAZ measurement methods are feasible when using a multiday dysprosium-tracer administration protocol as we have done, administration over a single meal or meals in a single day would limit the options. In those cases, where there are few fecal samples containing appreciable quantities of dysprosium and tracer, method 2 may be the only reasonable method. When considering the determination of FAZ from incomplete fecal collections, the difficulty in predicting which stools will provide the most reliable information is affected by the dysprosium-tracer administration plan. If high value is placed on measuring FAZ with few fecal samples, a multiday dosing protocol that creates a wide sampling window, making sample collection timing less critical, should be considered.
In conclusion, the use of dysprosium provides a reliable means of determining the completeness of fecal collections for detailed studies of zinc homeostasis. Moreover, in circumstances in which other methods of measuring FAZ are impractical, eg, due to the unacceptability of intravenous isotope administration, the use of a dysprosium method provides a simple means of determining FAZ.
ACKNOWLEDGMENTS
We thank all the subjects who participated in the study. We gratefully acknowledge the contribution of Therese Ida and the dietitians, nurses, and kitchen staff of the General Clinical Research Center at the University of Colorado Hospital for their assistance with the conduct of this study
KMH, LVM, NFK, JEW, SL and X-YS were responsible for the conceptualization and the design of the study. X-YS and JEW implemented the clinical procedures. SL and X-YS completed the laboratory analyses. LVM supervised the data collection and the data analysis and statistical modeling procedures. LVM, KMH, and XYS drafted the manuscript, which was reviewed by all coauthors. None of the authors had any financial conflicts of interest.
REFERENCES
Schuette SA, Janghorbani M, Young VR, Weaver CM. Dysprosium as a nonabsorbable marker for studies of mineral absorption with stable isotope tracers in human subjects. J Am Coll Nutr 1993;12:307–15.
Friel JK, Naake VL, Miller LV, Fennessey PV, Hambidge KM. The analysis of stable isotopes in urine to determine the fractional absorption of zinc. Am J Clin Nutr 1992;55:473–7.
Krebs NF, Miller LV, Naake VL, et al. The use of stable isotope techniques to asses zinc metabolism. J Nutr Biochem 1995;6:292–301.
Miller LV, Krebs NF, Hambidge KM. Development of a compartmental model of human zinc metabolism: identifiability and multiple studies analyses. Am J Physiol Regul Integr Comp Physiol 2000;279:R1671–84.
Veillon C, Patterson KY, Moser-Veillon PB. Digestion and extraction of biological materials for zinc stable isotope determination by inductively coupled plasma mass spectrometry. J Anal Atom Spectrom 1996;11:727–30.
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;327:307–10.
Fairweather-Tait SJ, Minihane A-M, Eagles J, Owen L, Crews HM. Rare earth elements as nonabsorbable fecal markers in studies of iron absorption. Am J Clin Nutr 1997;65:970–6.
Ulusoy U, Whitley JE. Profiles of faecal output of rare earth elements and stable isotopic tracers in iron and zinc after oral administration. Br J Nutr 2000;84:605–17., 百拇医药(Xiao-Yang Sheng, K Michae)
ABSTRACT
Background: Dysprosium is a nonabsorbable rare earth element that has had successful application as a marker for fecal excretion of unabsorbed zinc.
Objective: Our goals were 1) to evaluate the efficacy of administering dysprosium with all meals over several days as a method of determining the completeness of fecal collections, 2) to determine the similarity of gastrointestinal transit kinetics and excretion patterns of dysprosium and zinc tracer administered simultaneously over several days, and 3) to evaluate alternative methods of using the data for fecal excretion of orally administered zinc tracer and dysprosium to measure the fractional absorption of zinc.
Design: 70Zn and dysprosium were administered orally with all meals for 5 consecutive days to 7 healthy, free-living adults consuming a constant diet based on habitual intake. Additional tracers, 67Zn and 68Zn, were administered intravenously. Urine and fecal samples were collected during tracer administration and for 8 d after the last dose. Isotope ratios were measured in urine and feces, and total zinc and dysprosium were measured in fecal samples.
Results: The mean recovery of dysprosium was 101.3 ± 2.4%. The zinc oral tracer and dysprosium had similar fecal excretory patterns; the correlation coefficient for 70Zn and dysprosium in fecal samples exceeded 0.99 (P < 0.0001) for each subject. Fractional zinc absorption measurements using various dysprosium methods correlated well (r > 0.95) with those from the fecal monitoring and dual-isotope-tracer ratio methods.
Conclusion: Administration of dysprosium is a useful means of determining the completeness of fecal collections and of measuring zinc absorption.
Key Words: Zinc dysprosium fecal marker absorption stable isotopes dual-isotope-tracer ratio technique fecal monitoring method
INTRODUCTION
Dysprosium is a rare earth element that is nontoxic, nonabsorbable, present at no higher than trace amounts in the diet, and measured quite easily. It was first used as a quantitative fecal marker in zinc absorption studies in 1993 (1) and has subsequently had minimal reported application in studies of zinc metabolism. We are currently using dysprosium in studies ranging from measurements of the fractional absorption of zinc (FAZ) to complex studies of zinc homeostasis with the use of zinc stable-isotope-tracer techniques. Before using dysprosium in our research program, we undertook additional evaluation of the use of dysprosium and the interpretation of data derived from the application of this rare earth element in studies of zinc metabolism. The present article reports the results of this evaluation.
The first objective of the present study was to evaluate the efficacy of administering dysprosium with all meals over several days as a method for determining the completeness of fecal collections for free-living human metabolic studies. The second objective was to determine the similarity of gastrointestinal transit kinetics and excretion patterns of dysprosium and isotopically enriched zinc tracer after simultaneous oral administration in individual subjects and under the circumstances of this multiday metabolic study. The third objective was to evaluate alternative methods of using the data for fecal excretion of orally administered zinc tracer and dysprosium to determine FAZ and to compare the results from these methods with FAZ values derived simultaneously by using the dual-isotope-tracer ratio (DITR) technique and a fecal-monitoring (FM) method (2, 3). This comparison included an examination of the difference between results attributable to the fact that, unlike the DITR and FM methods, the dysprosium methods generally do not account for absorbed tracer that is subsequently secreted into the intestinal lumen and excreted in the feces, herein referred to as endogenous tracer.
SUBJECTS AND METHODS
Study design
This research was part of a larger protocol designed to evaluate selected zinc stable-isotope techniques for the investigation of human zinc homeostasis. Free-living, apparently healthy subjects were studied while consuming a constant daily diet that was based on their habitual diet and was provided daily by the diet kitchen of the General Clinical Research Center at the University of Colorado Hospital (day –6 to day 5). Four days before the start of dysprosium and oral tracer administration, a zinc stable-isotope tracer (68Zn) was administered intravenously. Then, starting at dinner on day 0, each meal was labeled with another zinc tracer (70Zn) together with dysprosium until dinner on day 5. A third tracer (67Zn) was administered intravenously on day 3 at 1230. This time was selected on the basis of the results of a simulation of the study protocol with the use of our published model of zinc metabolism (4), which identified 1230 as the optimal time for intravenous tracer administration for accurate calculation of FAZ by the DITR technique. Complete fecal samples were collected from the time of administration of the first dose of dysprosium until the end of day 13. Timed urine samples were collected on days 0–13. The study timeline is summarized in Figure 1.
Subjects
The participants were 7 healthy volunteers, 6 women and 1 man, aged 23–39 y ( ± SD: 34.1 ± 5.4 y) with a mean (±SD) body mass index (in kg/m2) of 22.1 ± 2.1. The study was approved by the Colorado Multiple Institutional Review Board, and all subjects gave their written consent to take part in the study.
Study diet
A 7-d diet record was analyzed by using the Nutrition Data System for Research, version 4.04-32 (Nutrition Coordinating Center, University of Minnesota, Minneapolis, MN). The subjects then met with a research dietitian to select a 3-d rotating diet with energy, zinc, and phytate intakes matching those calculated from the 7-d record. The constant daily diet was provided daily by the University of Colorado General Clinical Research Center diet kitchen from day –6 to day 5. During the metabolic study period, the subjects consumed all their meals under the supervision of one of the investigating team.
Isotope preparation and administration
Enriched stable isotopes of zinc were obtained from Trace Science International (Ontario, Canada). Accurately weighed quantities of each isotopically enriched preparation were dissolved in 0.5 mol H2SO4/L and then diluted in triply deionized water to prepare a stock solution. For preparation of orally administered doses, the stock solution of enriched 70Zn was diluted and titrated to pH 3.0 with metal-free ammonium hydroxide. This solution was filtered through a 0.22-μm filter to make a sterile solution. For preparation of intravenously administered doses of 67Zn and 68Zn, sterile techniques were used. The stock solution was diluted and adjusted to pH 6.0 and then filtered through a 0.22-μm filter to make a sterile solution. Concentrations of zinc in the isotope preparations were determined in triplicate by atomic absorption spectrophotometry, and concentration measurements were adjusted for the different atomic weights of the preparations.
An accurately weighed quantity of 67Zn (1 mg) was administered intravenously on day 3 at 1230. The tracer was administered over a 10-min interval via a scalp vein needle in a superficial forearm vein with a 3-way closed stopcock system. This allowed us to rinse the delivery syringe twice with N-saline contained in a second sterile syringe.
A total of 1 mg 70Zn (accurately weighed) was administered orally divided between all meals for 5 days commencing with dinner on day 0 and continuing through lunch on day 5 of the metabolic period. The tracer was administered in water gradually during the second half of each meal. The oral tracer was distributed among the meals in approximate proportion to the dietary zinc content of the meals. Dietary zinc intake ranged from 5.9 to 27.0 mg/d with an average of 11.3 ± 7.4 mg/d. The quantity of oral 70Zn tracer administrated with meals was 0.2 mg/d or a maximum of 3.4% of zinc in the diet.
Dysprosium, in the form of DyCl3 · 6H2O, was obtained from Sigma Aldrich (Milwaukee, WI). The dysprosium dose was prepared in purified, filtered water and its concentration was measured by inductively coupled plasma mass spectrometry (ICP-MS) (VG Plasma Quad 3; VG Elemental, Cheshire, United Kingdom). A total oral dose of 1 mg Dy was divided in proportion to the 70Zn doses and was administered in 70Zn solution taken with all 5 d of meals.
Sample collection
All stools were collected from the time of the first 70Zn-labeled meal until 8 d after the last 70Zn-labeled meal. Feces were collected separately and quantitatively in plastic bags. A clean, midstream-void urine sample was collected into a zinc-free plastic container twice daily from day 0 to day 13. The times for each collection were noted on the specimen cup and log sheets. Baseline fecal and urine specimens were obtained before administration of any label. All samples were frozen at –20°C until analyzed.
Sample preparation and analyses
Accurately weighed aliquots of homogenized feces and whole-day food samples were dried separately to a constant weight in an electric oven. All samples were prepared in duplicate. The dried samples were ashed in a muffle furnace at 450°C for 24 h. A few drops of concentrated nitric acid were added to the ash, which was then dried before reheating again at 450°C for 24 h.
Ashed fecal samples were reconstituted quantitatively in 50 mL, 6 mol HCl/L. The concentration of total zinc in these reconstituted fecal samples was determined on a diluted aliquot with an atomic absorption spectrophotometer fitted with a deuterium arc background-correction lamp (Perkin-Elmer Corporation, Norwalk, CT.)
For the measurements of zinc stable-isotope ratios, the inorganic elements were removed from reconstituted ashed fecal samples by ion-exchange chromatography with AG-1 ion-exchange resin (Bio-Rad Laboratories, Richmond, CA).
Urine samples were digested by using an MDA-2000 microwave sample preparation system (CEM Corp, Mathews, NC). A 5-mL urine sample was placed into an Advanced Composite Vessel and combined with 1 mL of concentrated HNO3, and the pressure was gradually increased to a maximum of 120 psi. The total digestion time was 90 min. Digested samples were transferred to a beaker, evaporated to dryness on a hot plate, and reconstituted in 2 mL ammonia acetate buffer (pH = 5.6). Zinc in the sample was purified by its chelation with trifluoroacetylacetone and then extraction of the chelate with hexane (5).
Isotope enrichment was determined by measurement of the isotope ratios 67Zn/66Zn, 68Zn/66Zn, and 70Zn/66Zn by ICP-MS. Tracer enrichment was defined as all zinc in the sample from the isotopically enriched tracer preparation divided by the total zinc in the sample.
The dysprosium contents of the test meal, doses, and fecal samples were measured by ICP-MS. Ashed fecal and food samples were reconstituted quantitatively in 50 mL 10% (by vol) HNO3. Bismuth (Aldrich Chemical Company Inc, Milwaukee, WI) was used as an internal standard. A large amount of bismuth was found in the fecal samples of one subject, so indium was used as an internal standard for the samples from this subject. The sample solution and dysprosium standards (Aldrich Chemical Company Inc, Milwaukee, WI) were diluted by using 6 parts per billion bismuth solution in 2% (by vol) HNO3 (OPTIMA; Fisher Scientific, Pittsburgh, PA). A set of dysprosium standards was inserted every 10 samples to make the external drift corrections. Two spiked fecal samples with dysprosium concentrations of 100 and 200 parts per billion, respectively, were used to check the accuracy of our measurements, and their recovery was 99.5% ±1.6 (6 measurements).
Data processing
The content of 70Zn tracer and dysprosium in the fecal samples was quantified as a fraction of the administered dose, except in the case of method 1b, for which dysprosium was quantified as the fraction of total dysprosium recovered. The 70Zn and dysprosium fecal excretory patterns for each subject were plotted and inspected. In addition, ratios of 70Zn to Dy were plotted verses sample time to discern any temporal differences in excretion not evident with the other comparison techniques. All fecal samples having dysprosium concentrations greater than the limit of detection were included in the calculation of total dysprosium recovery.
FAZ was calculated from the dysprosium and zinc tracer data by using 3 methods. In addition, method 1 was modified to take into account the presence of endogenous 70Zn tracer (method 1a) and was further modified to quantify dysprosium relative to the total dysprosium recovered instead of total dysprosium dose (method 1b).
In method 1, FAZ was calculated for each fecal sample by using the following equation:
where 70Zn in the sample is equal to total zinc in the sample times the 70Zn enrichment of the sample. Then, a mean FAZ was calculated for each subject. To eliminate samples giving unreliable FAZ results because of low dysprosium and unabsorbed 70Zn concentrations, only samples having a dysprosium-to-zinc mass ratio >0.25 of (dysprosium intake per day/average total zinc in feces per day) were included in the calculation of the mean. It was almost always the case that the first or second stool sample after administration of the first Dy-70Zn doses and all subsequent stool samples up to and including the first or second stool sample after administration of the final dose met this criterion. One exception was the case of a subject having diarrhea, for whom the included samples were the fourth stool sample after the first dose to the fourth stool sample after the final dose. Method 1a used the following calculation:
Endogenous 70Zn in the sample was determined with this calculation:
where urine enrichments are estimated for the time at which the fecal sample content was transiting the upper gastrointestinal tract.
Method 1b used a modification of method 1a:
With method 2, the dysprosium and 70Zn contents of all samples having a dysprosium concentration greater than the limit of detection were summed to calculate a single FAZ by using an equation equivalent to that used for method 1.
Method 3 derived an FAZ value from the slope of the regression of the fraction of the dysprosium dose in the sample versusthe fraction of the 70Zn dose in the sample by using the following equation:
As with method 2, all samples having dysprosium concentrations greater than the limit of detection were included in the regression analysis.
FAZ results from the dysprosium methods were compared with simultaneous calculations of FAZ made by using 2 established methods: one based on the measurement of dual (oral and intravenous)-isotope-tracer enrichment ratios in urine (DITR) (2) and a cumulative tracer in feces method that incorporates a correction for endogenous tracer (FM) (3). Because the oral tracer was administered over a much longer than usual period of time, optimum times for the administration of the intravenous tracer for the DITR method and for the extrapolation endpoint in the FM calculation had to be estimated to ensure accurate results. These were obtained from a simulation of the tracer administration and sampling protocols made with a compartmental model of human zinc metabolism (4). The simulation was performed by using WINSAAM software (WinSAAM Inc, University of Pennsylvania, Kennett Square, PA). In the FM method calculation, extrapolation to the midpoint of the oral tracer administration period was found to be satisfactory.
A compartmental model simulation of the dysprosium methods was also performed to predict the relation between the FAZ results from these methods and the DITR and FM methods attributable to the fact these dysprosium methods do not account for the endogenous oral tracer present in the feces.
Statistical analysis
All data are presented as means ± SDs unless otherwise stated. Comparison of FAZ results was performed by using repeated-measures analysis of variance (ANOVA) and Dunnett's multiple-comparison test. Statistical significance was defined as P < 0.05. Correlation analysis was used to compare the 70Zn and dysprosium content in fecal samples having dysprosium concentrations greater than the limit of detection and to compare the FAZ results from the dysprosium methods with those from the DITR and FM methods. Linear regression analysis was used to derive slope values for dysprosium method 3. All statistical analyses were performed by using GRAPHPAD PRISM version 3.01 (GraphPad Software Inc, San Diego, CA).
RESULTS
Objective #1: The individual quantities of dysprosium administered and the percentage recovered are given in Table 1. The mean recovery of dysprosium was 101.3 ± 2.4%. As expected, individual transit times varied; typically, however, most of the 70Zn and dysprosium had disappeared from the feces within 24 h of administration of the final doses. Dysprosium concentrations dropped below the limit of detection by 4.5 d after the last dose on average; the longest interval was 6.4 d. The limit of detection of dysprosium in fecal samples by ICP-MS was 2.5 parts per billion.
Objective #2: In all cases, the fecal excretory patterns of 70Zn and dysprosium were similar; an example for one subject is depicted in Figure 2. The fraction of zinc tracer excreted versus the fraction of dysprosium dose excreted in the fecal samples from the individual subjects is depicted in Figure 3. Linear regression lines are also shown. The correlation coefficient (r) exceeded 0.99 (P < 0.0001) for every subject. In addition, the linear regression analyses of these data indicated that the y intercepts were never significantly different from zero (the average y intercept was 0.00082). The number of samples from each subject included in the analysis, and plotted in Figure 3, ranged from 8 to 13. No consistent temporal pattern was observed in the plots of the ratio of zinc tracer to dysprosium versus sample time.
Objective #3: The results from the various methods of measuring FAZ are given in Table 2. Correlation coefficients for the dysprosium results compared with the 2 established methods are shown. In every case, r was >0.95 (P 0.0008). Repeated-measures ANOVA indicated no significant differences between FAZ results. Given this finding, Dunnett's test did not provide any additional information of value. The FAZ measurements from dysprosium methods 1, 1a, and 1b are plotted against the DITR measurements in Figure 4. The adjustment for endogenous 70Zn (method 1a) produced FAZ values 6.8% higher on average (range: 2.8–12.7%) than those of method 1, resulting in an average difference in FAZ of almost 0.02 in absolute terms. The model simulation predicted a difference of 4–5% between methods when one method does not take into account the secretion and excretion of endogenous tracer. The results from the modification of method 1 to account for endogenous tracer (method 1a) and to quantify dysprosium relative to that recovered (method 1b) both showed improved correlation and agreement with the DITR and FM results. Shown in Figure 5 is a Bland-Altman (6) plot of the results of methods 1a and 1b compared with DITR; the mean difference in FAZ was 0.012 and the limits of agreement were 0.036 and –0.013 for method 1b.
Analysis of the dysprosium content of the 5-d diet of subject number 1 showed very low concentrations of dysprosium, totaling little more than 0.1% of the total administered dose.
DISCUSSION
Dysprosium must be administered for several days in studies of zinc homeostasis that include quantitative fecal collections for, as an example, the measurement of excretion of endogenous zinc via the intestine, which is a measurement of cardinal importance in such studies. In the present study, our results provided important reassurance that the metabolic collections were complete in the free-living subjects.
Similar to other reports (1, 7), our mean recovery of dysprosium was slightly in excess of 100% of the quantity administered. Although we did not observe measurable concentrations of dysprosium in baseline fecal samples and found only minute concentrations of dysprosium when we analyzed the diet of one subject, others have measured more substantial amounts of dysprosium in baseline feces (8) and in test meals (7). Thus, it is possible that naturally occurring dysprosium in the diet may help explain the slight excess recovery in most of the subjects. In the only subject (number 3) having a dysprosium recovery of <100%, the dysprosium measurements may have been erroneously low due to the presence of bismuth and indium in this subject's diet, because these elements are used as internal standards in the ICP-MS analyses.
In considering the significance and possible causes of dysprosium recovery deviating from 100%, we noted that after the endogenous tracer was accounted for the magnitude and direction of the discrepancies between dysprosium values and DITR values exhibited a rough correspondence with the deviation of dysprosium recovery from 100%. In response, we tried an additional FAZ calculation method (1b) wherein sample dysprosium was quantified relative to the amount of dysprosium recovered instead of the total dysprosium dose, thereby improving the correlation and agreement with the DITR and FM value. This finding suggests that the quantity of administered dysprosium actually collected was generally closer to 100% than the recovery figures indicate and is consistent with several possible reasons for the discrepant recoveries, including those mentioned above (ie, dysprosium in the diet or analytic error in the case of subject 3). It also suggests that in carefully controlled studies such as this one, in which complete sample collection and rigorous quantitative processing are ensured, the calculation of FAZ by this method may be useful.
The similar dysprosium and zinc tracer excretory patterns and the high correlation between zinc and dysprosium in the fecal samples provide assurance that orally administered dysprosium can be assumed to mimic the gastrointestinal transit behavior of simultaneously administered zinc tracer very closely, and, as a result, can be most useful in isotopic tracer studies of zinc.
The FAZ results produced by the various dysprosium methods all correlated well (r > 0.95) with the DITR and FM methods. The results of method 1 are probably the most useful because they provide information on the accuracy and variability of FAZ determinations from individual samples and, therefore, the feasibility of using fewer samples (incomplete collections) to measure FAZ. Although the data are generally reassuring in this regard, we observed several instances in which accuracy and variability were adversely affected by the presence of extreme data at the beginning or end of the sample collection period. Therefore, it may be possible to obtain more accurate results by applying a more restrictive sample inclusion criterion.
An accuracy issue inherent with the dysprosium methods is that they do not take into account the effect of tracer that is absorbed and then secreted back into the gut and excreted in feces. Method 1a incorporated a measurement of endogenous 70Zn in each sample and calculated FAZ accordingly. The calculation of endogenous 70Zn in method 1a used measurements of 68Zn intravenous tracer enrichment in a manner related to an established method for measuring endogenous fecal zinc (EFZ) (3). The magnitude of the resulting adjustment to FAZ is modest in relation to the variability in the measurements, exceeding little more than 1 SD in the worst cases. The FAZ increases from method 1a were generally larger than the 4–5% predicted by the model. This is likely the result of the simulation using a population value that was slightly lower than the average EFZ rate observed in these subjects. Regarding the feasibility of using a basic dysprosium method to measure FAZ with sufficient accuracy, we believe that in most cases, where FAZ, EFZ, the exchangeable zinc pool size, and metabolic balance conditions are expected to be typical, a fixed adjustment to the FAZ measurement would be adequate. On the basis of our experience thus far, we suggest that, in the case of healthy adults in metabolic balance, FAZ values from dysprosium methods be increased by 5% as an approximate correction for endogenous tracer.
As discussed above, method 1b was implemented to address the concern with dysprosium recovery. Correlation and agreement with the DITR and FM results improved as a consequence. The improvement of agreement with method 1b compared with method 1a is evident in the Bland-Altman plot (Figure 5). The limits of agreement for method 1b are at least as good as typically observed when comparing FAZ methods. The mean difference indicates a positive bias of 0.01 in the dysprosium method relative to the DITR results. We do not know the cause of this apparent difference.
Dysprosium method 2 was included because it is a simple way of calculating a single FAZ value when complete collections have been performed and is the method generally reported by others. Method 3, too, is a simple calculation possible when a sufficient number of samples with various concentrations of dysprosium and tracer have been collected.
The different calculation methods provided similar results using varied ways of averaging the information from all the samples. To a great extent, the differences between results reflect the inherent variability routinely encountered with biological measurements of this kind. Correcting for endogenous tracer by measurement or estimation can minimize bias in FAZ results. Furthermore, when complete collections are carefully performed and dysprosium recovery measured, the possibility that accuracy may be improved by quantifying dysprosium relative to total dysprosium recovered instead of dysprosium administered should be considered. Modifications equivalent to those applied to method 1 could be applied to methods 2 and 3 with a similar outcome.
Although all 3 FAZ measurement methods are feasible when using a multiday dysprosium-tracer administration protocol as we have done, administration over a single meal or meals in a single day would limit the options. In those cases, where there are few fecal samples containing appreciable quantities of dysprosium and tracer, method 2 may be the only reasonable method. When considering the determination of FAZ from incomplete fecal collections, the difficulty in predicting which stools will provide the most reliable information is affected by the dysprosium-tracer administration plan. If high value is placed on measuring FAZ with few fecal samples, a multiday dosing protocol that creates a wide sampling window, making sample collection timing less critical, should be considered.
In conclusion, the use of dysprosium provides a reliable means of determining the completeness of fecal collections for detailed studies of zinc homeostasis. Moreover, in circumstances in which other methods of measuring FAZ are impractical, eg, due to the unacceptability of intravenous isotope administration, the use of a dysprosium method provides a simple means of determining FAZ.
ACKNOWLEDGMENTS
We thank all the subjects who participated in the study. We gratefully acknowledge the contribution of Therese Ida and the dietitians, nurses, and kitchen staff of the General Clinical Research Center at the University of Colorado Hospital for their assistance with the conduct of this study
KMH, LVM, NFK, JEW, SL and X-YS were responsible for the conceptualization and the design of the study. X-YS and JEW implemented the clinical procedures. SL and X-YS completed the laboratory analyses. LVM supervised the data collection and the data analysis and statistical modeling procedures. LVM, KMH, and XYS drafted the manuscript, which was reviewed by all coauthors. None of the authors had any financial conflicts of interest.
REFERENCES
Schuette SA, Janghorbani M, Young VR, Weaver CM. Dysprosium as a nonabsorbable marker for studies of mineral absorption with stable isotope tracers in human subjects. J Am Coll Nutr 1993;12:307–15.
Friel JK, Naake VL, Miller LV, Fennessey PV, Hambidge KM. The analysis of stable isotopes in urine to determine the fractional absorption of zinc. Am J Clin Nutr 1992;55:473–7.
Krebs NF, Miller LV, Naake VL, et al. The use of stable isotope techniques to asses zinc metabolism. J Nutr Biochem 1995;6:292–301.
Miller LV, Krebs NF, Hambidge KM. Development of a compartmental model of human zinc metabolism: identifiability and multiple studies analyses. Am J Physiol Regul Integr Comp Physiol 2000;279:R1671–84.
Veillon C, Patterson KY, Moser-Veillon PB. Digestion and extraction of biological materials for zinc stable isotope determination by inductively coupled plasma mass spectrometry. J Anal Atom Spectrom 1996;11:727–30.
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;327:307–10.
Fairweather-Tait SJ, Minihane A-M, Eagles J, Owen L, Crews HM. Rare earth elements as nonabsorbable fecal markers in studies of iron absorption. Am J Clin Nutr 1997;65:970–6.
Ulusoy U, Whitley JE. Profiles of faecal output of rare earth elements and stable isotopic tracers in iron and zinc after oral administration. Br J Nutr 2000;84:605–17., 百拇医药(Xiao-Yang Sheng, K Michae)