The dietary recommendations for mineral intakes by the Institute of Medicine were revised in the late 1990s. Intake recommendations for Zn in children aged 1–4 years changed dramatically from an RDA of 10 mg/d in 19891 to an estimated average requirement (EAR) of 2·5 mg/d, and an RDA of 3 mg/d2. These were based on a factorial approach assuming obligatory urinary losses of 7·5 μg/kg per d, integumental losses of 6·5 μg/kg per d, endogenous faecal losses of 34 μg/kg per d and a requirement for growth of 120 μg/d2. Most of these assumptions were based on data in older subjects, due to the lack of suitable data from young children2.
The principal reason for the absence of data in this age group is the impracticality of prolonged dietary regulation and complete urine and faecal collections that are required for traditional balance studies, especially in active children who are often not toilet-trained. Studies in this age group are now more feasible with modern stable-isotope methods in which Zn absorption can be directly assessed with a single timed urine sampleReference Abrams3, Reference Griffin4. Thus, our goal in the present study was to utilize stable isotopes to evaluate the relationships among Zn intake, absorption, endogenous faecal excretion, and retention in healthy small children on diets common in the USA.
Zn intakes in children 1 to 4 years in the USA (median intake 5·81 mg/d) are well above the current EAR (2·5 mg/d). In fact, according to the Continuing Survey of Food Intakes in Individuals, the 75th percentile of Zn intake in the USA in this age range is 7·74 mg/d, a value higher than the tolerable upper limit for Zn2. As the potential hazard identified to determine the tolerable upper limit for Zn intake was a possible adverse effect of high Zn intakes on Cu absorption or Cu status, we were also interested in collecting preliminary data on Cu absorption in children on Zn intakes typical of the 1–4-year-old US population.
We hypothesized that fractional Zn absorption would be significantly negatively correlated with Zn intake, and that urinary Zn excretion, endogenous faecal Zn excretion, absolute Zn absorption and Zn balance would be significantly positively correlated with Zn intake. We also wished to use the relationship between Zn retention and Zn intake to estimate the Zn intake needed to meet the estimated requirement for normal growth (120 μg/d) for 1–4-year-old children, as this would be a suitable basis for determining the EAR.
Subjects and methods
Healthy children aged 1–4 years in the greater Houston area were recruited through public advertising. Subjects were selected to reflect the approximate racial and ethnic distribution of the greater Houston population. The Institutional Review Board of Baylor College of Medicine and Affiliated Hospitals approved the protocol, and informed written consent was obtained from the subjects' parents for all studies.
Screening visit
Children were eligible for enrolment if they were healthy, not taking any medications (except multivitamins), were born at term ( ≥ 37 weeks gestation) and had a birth weight ≥ 2500 g. Children were excluded from participating if they had chronic health problems, were below the 3rd or above the 97th percentile of weight- or height-for-age, or were below the 5th or above the 95th percentile of weight-for-height. Those subjects taking multivitamins were required to discontinue them for 2 weeks before participating in the mineral absorption study. Families were offered the option of participating in a 2 d in-patient Zn absorption study or a 5 d in-patient study in which faecal samples would also be collected and endogenous excretion would be measured.
The research dietitian met with the parent and obtained a complete dietary history to evaluate usual daily micronutrient and energy intake. Dietary intake data were collected using Nutrition Data System for Research software (Nutrition Coordinating Center, University of Minnesota, Minneapolis, MN, USA).
Home dietary adaptation
After screening, a dietary plan was developed for each child that would be consumed at home for the 7 d before the in-patient mineral absorption study. This was to ensure that children did not alter their eating habits immediately before the mineral absorption study, and was designed to be reflective of the child's usual dietary intake. All foods and beverages to be consumed during these 7 d were provided by the research centre and were pre-weighed before delivery to the family. Parents were instructed to return all the uneaten food and beverage items for the first 3 d of the pack-out so items could be post-weighed.
Isotope preparation and mineral absorption study
67Zn (88 % enrichment) and 70Zn (96 % enrichment) were purchased from Trace Sciences Incorporated (Toronto, ON, Canada) and prepared for human use as the chloride salt by the Investigational Pharmacy Service of Texas Children's Hospital, Houston, Texas. All isotopes were tested for sterility and pyrogenicity before use.
At the end of the home adaptation period, patients were admitted to the General Clinical Research Center at Texas Children's Hospital for the mineral absorption study. On the morning of the in-patient study, subjects had a heparin-lock intravenous catheter placed, using topical 4 % lidocaine cream (L-M-X-4; Ferndale Laboratories, Ferndale MI, USA) as an analgesic. Subsequently, 0·5 mg 67Zn was given intravenously over 1 min. The subgroup of subjects in which endogenous faecal Zn was being measured had a higher intravenous dose of 1 mg 67Zn given intravenously. Subjects were then given a breakfast that included 30 ml apple juice to which 0·12 mg 70Zn and 30 μg 65Cu had been added. After the subject consumed the isotope-containing juice, the subject consumed another 30 ml apple juice without isotope from the same cup as a rinse to ensure that none of the isotope was left in the cup. This process was repeated with lunch using the same amount of isotopes as breakfast.
Each meal provided approximately one-third of the daily mineral intake of the subject. Menus for the in-patient study visit were based on the subject's usual mineral intake that they had received at home for the previous 7 d. All foods and beverages during the in-patient visits were pre- and post-weighed to accurately determine intake. Dietary intakes used in the results section were based on these intakes.
Subjects remained in our in-patient unit for 48 h and their urine was collected in 24 h pools for the duration of their hospitalization. If the subject was not well toilet-trained urine bags were used for the sample collection. Subjects returned to the out-patient unit 96 h after isotope administration for collection of a spot urine sample. The subset of subjects in whom endogenous faecal excretion was measured remained in the in-patient unit for 120 h during which time their urine and stools were collected in 24 h pools.
Isotope ratio measurement
Urine samples were prepared for MS analysis as previously describedReference Herman, Griffin, Suwarti, Ernawati, Permaesih, Pambudi and Abrams5. Faecal samples were homogenized with an approximately equal volume of water, the final volume recorded and a sample frozen. A 2 ml portion of the sample was digested in a microwave (MARS 4; CEM Inc., Matthews, NC, USA) until the sample was clear and all organic matter digested. The sample was evaporated on a hotplate at sub-boiling temperature and samples used for Zn and Cu concentration measurement and Zn and Cu isotope ratio measurement.
Samples were analysed for Zn isotopic enrichment with a Finnigan MAT 261 (Thermo Finnigan MAT GmbH, Bremen, Germany) magnetic sector thermal ionization mass spectrometer. Each sample was analysed for the 67Zn:66Zn and 70Zn:66Zn ratios with correction for fractionation to the reference 64Zn:66Zn ratio. Accuracy and precision of this technique for natural abundance samples compared with standard data are 0·15 % or better, depending on the ratio being measured. Cu isotope ratios were measured by a high-resolution inductively coupled plasma MS (Element 2; Thermo Electron, Bremen, Germany) equipped with a 100 μl/min Micromist concentric nebulizer and a Scott type spray chamber. Diluted digested solution containing about 25 μg Cu/l was analysed at a mass resolution of about 4000 for 2 min. Analytical precision for the 65Cu:63Cu ratio was < 0·3 %.
Zn concentrations were measured by flame atomic absorption spectroscopy, and Cu concentrations by inductively coupled plasma MS.
Measurement of zinc absorption and zinc excretion
Zn absorption was measured from the ratio of the urinary excretion of the orally administered stable isotope to the intravenous isotope in the 96 h urine sample as described elsewhereReference Griffin4, Reference Herman, Griffin, Suwarti, Ernawati, Permaesih, Pambudi and Abrams5.
Urinary Zn excretion was calculated from the urine collection (either 48 or 120 h) and expressed as mg/d.
Endogenous faecal Zn excretion was measured from the urinary and faecal excretion of the intravenously administered tracer and the urinary excretion of tracee Zn in the 5 d urine and faecal collection using the equationReference Griffin4:
Zn balance was estimated from the equation:
Endogenous Zn excretion was measured in nine subjects. These data were used to extrapolate Zn losses in those subjects (n 21) in whom it was not directly measured (see later). Integumental losses were estimated to be 6·5 μg/kg2.
Measurement of copper absorption
Cu absorption was estimated from the difference between the amount of oral tracer given, and the cumulative excretion of the oral Cu isotope during the 5 d faecal collection.
Statistical analysis and sample size determination
Assuming that the smallest clinically significant correlation between mineral intake and absorption is when changes in intake explain 25 % of the variability in mineral absorption (r 2 0·25; r 0·5) a sample size of thirty would have a power of greater than 80 % to detect such a difference at P < 0·05. This degree of correlation is similar to previous studies that have shown a correlation of r 0·5 between intake and net Ca balance in this age rangeReference Matkovic and Heaney6, and is therefore biologically plausible.
Relationships between Zn intake and different parameters of Zn homeostasis (Zn absorption etc) were assessed by simple regression analysis. χ2 Tests, unpaired t tests and one-sample t tests were used as appropriate. Statistical analysis was carried out using StatsView 5·0·1 for Macintosh (SAS Institute Inc., Cary, NC, USA). Data are presented as mean values and standard deviations or as n and percentage, as appropriate. Statistical significance was assumed at P < 0·05.
Results
Subject demographics
Zn absorption was successfully measured in thirty subjects (sixteen girls and fourteen boys). The ethnic distribution of the study population was 47 % Caucasian, 27 % Hispanic, 17 % African-American, and 10 % multi-ethnic (Table 1). Ten subjects were taking multivitamins; all discontinued them at least 2 weeks before the start of the study. Results are little changed if these subjects are omitted from the analysis.
Table 1 Demographics of the study subjects (numbers and percentages)
* Percentages do not equal 100 % due to rounding.
Twenty-one subjects opted for the 48 h admission and only had measurements of Zn absorption and urinary Zn excretion made, nine were admitted for 120 h and also had endogenous faecal Zn excretion and Cu absorption measured. There were no significant differences between the subjects who were admitted for 48 h compared with those admitted for 120 h in terms of sex (χ2P = 0·52), ethnicity (χ2P = 0·15), weight (P = 0·93), height (P = 0·57), age (P = 0·18) or Zn intake (P = 0·51).
Nutrient intakes
Mean Zn intake was 5·04 (sd 2·08, range 1·09–10·2) mg/d and was significantly greater than the current EAR of 2·5 mg/d (P < 0·0001), with four subjects (13 %) exceeding the current upper limit for Zn. Mean Cu intake was 408 (sd 116) μg/d, significantly above the EAR of 260 μg/d (P < 0·0001), but no subjects exceeded the upper limit for Cu (1000 μg/d). Mean Fe intake (6·77 (sd 2·42) mg/d) was also significantly above the EAR (P < 0·0001).
Zinc homeostasis
Neither fractional Zn absorption (31·1 (sd 7·8) %; y = 0·324 − 0·0003x; r 2 0·005; P = 0·71; Fig. 1) nor urinary Zn excretion was related to Zn intake (0·116 (sd 0·067) mg/d; 9·7 (sd 6·1) μg/kg per d; y = 0·14 − 0·005x; r 2 0·022; P = 0·44).
Fig. 1 Fractional Zn absorption in 1–4-year-old children is not related to Zn intake (y = 0·324 − 0·0003x; r 2 0·005; P = 0·71).
Endogenous faecal Zn excretion (0·995 (sd 0·401) mg/d; 82·8 (sd 69·0) μg/kg per d; n 9) was unaffected by Zn intake (y = 0·385+0·131x; r 2 0·060; P = 0·53), nor was it correlated with the subjects' age (r 2 0·049; P = 0·57), weight (r 2 0·035; P = 0·63) height (r 2 0·028; P = 0·67) or the total amount of Zn absorbed (the product of fractional Zn absorption and Zn intake) (r 2 0·076; P = 0·47). In the subsequent analysis, therefore, endogenous faecal Zn excretion was estimated to be 0·955 mg/d in those subjects in whom it was not measured (equal to the mean value for the nine subjects in which it was measured).
Absolute Zn absorption (Fig. 2) was significantly positively correlated with Zn intake (1·56 (sd 0·71) mg/d; y = 0·111+0·286x; r 2 0·696; P < 0·0001).
Fig. 2 Absolute Zn absorption in 1–4-year-old children is significantly positively correlated with Zn intake (y = 0·111+0·286x; r 2 0·696; P < 0·0001).
Zn balance, or Zn retention, was estimated by extrapolating the data on endogenous faecal Zn excretion, and by assuming integumental losses of 6·5 μg/d.2 For the entire population it was 0·361 (sd 0·794) mg/d which was not significantly different from the 0·120 mg/d estimated to be required for normal growth in 1–4-year-old children2 (P = 0·11). Zn retention was significantly correlated with Zn intake (Fig. 3; y = − 1·008+0·272x; r 2 0·506; P < 0·0001).
Fig. 3 Zn retention in 1–4-year-old children is significantly positively related to Zn intake (y = − 1·008+0·272x; r 2 0·506; P < 0·0001).
Zinc intake required for normal growth in 1–4 year-olds
Two different methods were used to estimate the dietary Zn intake needed for normal growth.
About 0·12 mg/d is required for normal growth of 1–4-year-old infants2. Based on the relationship shown between Zn intake and Zn retention (Fig. 3) this would be achieved at a Zn intake of 4·15 mg/d.
The Institute of Medicine2 calculated the EAR using a factorial approach. They estimated that the requirement for absorbed Zn was 744 μg/d (Table 2)2. If we substitute our measurements of endogenous faecal Zn losses (83 μg/kg) and urinary Zn excretion (9·7 μg/d) for the estimates used by the Institute of Medicine, the requirement for absorbed Zn becomes 1410 μg/d (Table 2). If we assume fractional Zn absorption to be 30 % (consistent with the data used by the Institute of Medicine, and our own data from the present study) this equates to a dietary requirement of 4·7 mg/d.
Table 2 Estimates for total requirement for absorbed zinc for 1–4-year-old children using estimates made by the Institute of Medicine2 and made by modifying those estimates from data from the present study
Copper homeostasis
Cu absorption averaged 75·1 (sd 10·8, range 52·5–88·5) %, and was unrelated to Cu intake (n 9; y = 0·866 − 0·0003x; r 2 0·051; P = 0·56).
Discussion
We studied 1–4-year-old children on Zn intakes typical of the US population and found no significant relationship between Zn intake and fractional Zn absorption. This was contrary to our hypothesis and to the general expectation that such a relationship should existReference King, Cousins, Shike, Ross, Caballero and Cousins7. For example, in a study of stunted 3–4-year-old Peruvian children, de Romana et al. Reference de Romana, Salazar, Hambidge, Penny, Peerson, Krebs and Brown8 showed a significant fall in fractional Zn absorption in children receiving different Zn intakes, from 34·2 % in those consuming 2·14 mg Zn/d, to 23·7 % in those on 4·72 mg/d to 13·3 % for those on 10·04 mg/d.Reference de Romana, Salazar, Hambidge, Penny, Peerson, Krebs and Brown8 Our subjects consumed a similar range of intakes, but fractional Zn absorption of our population was very close to the Zn absorption of the lowest Zn intake group of de Romana'sReference de Romana, Salazar, Hambidge, Penny, Peerson, Krebs and Brown8 with no fall in fractional absorption as Zn intake increased. There are a variety of possible explanations for these differing findings. The diets used in the present study differed very greatly from those of de RomanaReference de Romana, Salazar, Hambidge, Penny, Peerson, Krebs and Brown8, and the greater bioavailability from our diets may have obscured the effect of Zn intake on fractional Zn absorption. Alternatively, the length of dietary adaptation may be important. We have previously shown a significant inverse relationship between fractional Zn absorption and Zn intake from a single meal with different amounts of haem Fe and inorganic Zn, given without any preceding adaptation period, in 4–8-year-old childrenReference Etcheverry, Hawthorne, Liang, Abrams and Griffin9. However, in a study of 9–14-year-old girls adapted to intakes of 4 and 12 mg/d for 3 weeks, no relationship between fractional Zn absorption and Zn intake was seenReference Griffin, Hicks, Liang and Abrams10. In the present study we did not modify subjects' usual dietary intake, so they had a prolonged period to adapt to their usual Zn intake. de Romana studied children for 2–3 d and 51–52 d after a dietary modification was madeReference de Romana, Salazar, Hambidge, Penny, Peerson, Krebs and Brown8 and differences in Zn absorption in relation to intake were seen at both times. However, when adults are placed on marginal Zn intakes, an initial intake in fractional Zn absorption is seen, but this is not maintained and decreases to baseline by 6 monthsReference Lee, Prasad, Hydrick-Adair, Brewer and Johnson11.
A final explanation relates to how the dietary changes were made. In those studies where a significant relationship between Zn intake and Zn absorption were seen, either oursReference Etcheverry, Hawthorne, Liang, Abrams and Griffin9 or de Romana'sReference de Romana, Salazar, Hambidge, Penny, Peerson, Krebs and Brown8, the dietary change was affected by the addition of a Zn salt to the test mealReference Etcheverry, Hawthorne, Liang, Abrams and Griffin9 or to fortified wheat flour provided to the familiesReference de Romana, Salazar, Hambidge, Penny, Peerson, Krebs and Brown8. In those studies where no relationship between Zn absorption and Zn intake were seen, the differences in Zn intake were due to differences in foods consumed in the diet – either self-selected by the subject's parents (the present study) or under the guidance of a dietitian to effect the desired changes in Zn intakeReference Griffin, Hicks, Liang and Abrams10. The Zn present in whole foods is probably more heterogeneous in its solubility and bioavailability than single salts added to the diet. Furthermore, changes in whole foods (rather than changes in single nutrients) may also lead to changes in the intake of inhibitors and enhancers of mineral absorption. This suggests that modifications in single nutrients in metabolic studies are not likely to produce results generalizable to the ‘real world’ where subjects consume complex foods rather than single nutrients in isolation. The present results are also consistent with data comparing two populations of young Chinese women whose self-selected diets contained either low or marginal intakes of ZnReference Sian, Mingyan, Miller, Tong, Krebs and Hambidge12 and where no significant relationship between Zn absorption and Zn intake could be demonstrated. Conversely in a study of 7–9-year-old females where Zn intakes between 5·61 and 14·61 mg/d were produced by adding aqueous zinc sulfate to meals, a significant inverse relationship between Zn intake and fractional Zn absorption was seenReference Ritchey, Korslund, Gilbert, Fay and Robinson13.
Endogenous faecal Zn excretion may be an important site of Zn homeostasisReference Lee, Prasad, Hydrick-Adair, Brewer and Johnson11, Reference King, Shames and Woodhouse14 and was measured in a subset of subjects. Ideally, we would have measured it in all subjects but the burden that would have placed on subjects and their families was felt to be too great, and would have greatly increased the difficulty in recruiting subjects. In the small number of subjects we studied we could not identify the expected inverse relationship with Zn intakeReference Lee, Prasad, Hydrick-Adair, Brewer and Johnson11, Reference Sian, Mingyan, Miller, Tong, Krebs and Hambidge12, Reference King, Shames and Woodhouse14 or the expected positive relationship with total absorbed Zn2, Reference Sian, Mingyan, Miller, Tong, Krebs and Hambidge12, although our statistical power was very limited and the study was not designed to detect such a relationship. Weight-specific endogenous faecal Zn excretion (89 (sd 69) μg/kg) did, however, tend to exceed the value of 34 μg/kg (P = 0·067) that the Institute of Medicine assumed for this age group2. Although we cannot say whether this higher than expected level of endogenous faecal Zn excretion was due to the relatively high Zn intake or to developmental differences in young children compared with older age groups (from whom the 34 μg/kg per d value was extrapolated from) it clearly has implications for accurate assessment of Zn requirements using a factorial approach.
We could detect no relationship between Zn intake and urinary Zn excretion, and this would be expected to be a late finding in Zn deficiency, or a response to very low Zn intakesReference King, Shames and Woodhouse14. The weight-specific urinary Zn excretion (9·2 μg/kg) was similar to that expected by the Institute of Medicine2, and somewhat lower than we have previously described in adolescents (averages between 13·4 and 15·5 μg/kg)Reference Griffin, Hicks, Liang and Abrams10.
Zn retention was linearly related to Zn intake. A Zn intake of 4·15 mg/d would be required to ensure the absorption of the 0·12 mg Zn/d required for normal growth in 1–4-year-old children2. An alternative approach to calculating the EAR for this age group, substituting our estimates of endogenous faecal Zn excretion and urinary Zn excretion into the Institute of Medicine's factorial calculation, leads to a slightly higher estimate of 4·7 mg/d. In either case it would appear that the current EAR (2·5 mg/d) is somewhat low and a level nearer 4·5–5·0 mg/d is more appropriate.
The Institute of Medicine also defines a tolerable upper limit of 7 mg/d, based on the theoretical risk of Zn interfering with Cu absorption and hence Cu status2. This level has been controversial as it is less than the previous RDA for this age group1, and is exceeded by many 1–4 year olds based on dietary surveys2. The present study design did not allow us to examine the effect of Zn intake on Cu absorption (or vice versa) as the intakes of the two minerals were strongly correlated. However, the demonstration that diets of young children that are high in Zn are also higher in Cu would appear to negate the potential effect of Zn on Cu status. However, in the subgroup of subjects where Cu absorption was measured it was relatively high, making a substantial effect of such Zn intakes on Cu absorption unlikely.
In summary, 1–4 year olds typically consume diets that exceed the current EAR and RDA for Zn; Zn retention increases as Zn intake increases; and a Zn intake of 4·2–4·7 mg/d appears to meet the needs for absorbed Zn in this age group.
Acknowledgements
The present study is a publication of the US Department of Agriculture (USDA)/Agricultural Research Service (ARS) Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine and Texas Children's Hospital, Houston, TX. This project has been funded in part with federal funds from the USDA/ARS under Cooperative Agreement number 58-6250-6-001 and the National Institutes of Health, National Center for Research Resources (NCRR) General Clinical Research for Children Grant number RR00188. Contents of this publication do not necessarily reflect the views or policies of the USDA, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.
The authors acknowledge the assistance of the nursing staff of the General Clinical Research Center of Texas Children's Hospital for caring for the study subjects and the investigational pharmacy of Texas Children's Hospital for preparation of the isotopes. We to thank Leslie Cruz and Dana McDonald for their help with patient enrolment and study visits, and Penni Hicks for her support and advice.
None of the authors has any conflict of interest to disclose.
