Zinc Supplementation Does Not Alter Indicators of Insulin Secretion and Sensitivity in Black and White Female Adolescents

Abstract

Background: Zinc is a micronutrient involved in the production of, and peripheral sensitivity to, pancreatic β cell–derived insulin. To our knowledge, the effect of zinc supplementation on insulin outcomes, and potential risk of diabetes, in otherwise healthy children in the United States has not been investigated.

Objective: The objective of this study was to determine the influence of zinc supplementation on insulin outcomes in black and white girls in the early stages of adolescence. A secondary objective was to determine relations between baseline zinc concentrations and insulin outcomes.

Methods: Healthy black and white girls aged 9–11 y were randomly assigned to daily supplementation of zinc (9 mg elemental Zn/d; n = 75; blacks: n = 35) or placebo (n = 72; blacks: n = 32) for 4 wk. Fasting serum insulin, glucose, and C-peptide were assessed at baseline and at 4 wk. C-peptide and glucose values were used to calculate the computer model–derived homeostatic model assessment of insulin resistance (HOMA2-IR). Changes in outcome measures were compared by using repeated-measures, mixed-model ANOVA.

Results: Baseline plasma zinc was not correlated with C-peptide (r = −0.07), insulin (r = −0.06), or HOMA2-IR (r = −0.09) (all P > 0.05) after controlling for race and age. Treatment × time interactions for C-peptide and HOMA2-IR were not significant (both P > 0.05). Although the treatment × race × time interactions for C-peptide and HOMA2-IR were not significant (both P = 0.08), black girls who received the placebo experienced slight increases in C-peptide (15.7%) and HOMA2-IR (17.7%) (P = 0.06).

Conclusions: Four weeks of zinc supplementation had no effect on insulin outcomes in healthy black and white early-adolescent girls, although C-peptide and HOMA2-IR tended to increase in black girls who received placebo. Additional trials that are appropriately powered should further explore the effect of zinc on markers of diabetes risk, and whether race affects this relation. This trial was registered at clinicaltrials.gov as NCT01892098.

Introduction

There has been a striking increase in type 2 diabetes and prediabetes in the pediatric population in recent years (1), mirroring the surge in childhood obesity. Of US adolescents, 1 in 4 has prediabetes (2), and this is concerning because early development of diabetes increases the risk of complications later in life (3). Given that insulin resistance predisposes children to type 2 diabetes in adulthood, it is critical to identify novel, safe, and inexpensive strategies for diabetes prevention.

Zinc is an essential micronutrient with multiple biological functions, but its potential link to diabetes has drawn attention in the past decade (4–7). Evidence suggests that zinc may exert an endocrine regulation of insulin production (4) and insulin signaling (7). Indeed, the process of insulin maturation, storage, and secretion from pancreatic β cells is zinc dependent (4). In addition, results from animal studies have shown that zinc stimulates the translocation of glucose transporter type 4 to the plasma membrane of insulin target tissues, ultimately increasing glucose uptake into cells and lowering blood glucose concentrations (7). In vitro studies that further explored this mechanism in preadipocytes and adipocytes found that zinc induces phosphorylation of the β-subunit of the insulin receptor in addition to other proteins involved in the signaling cascade leading to glucose transporter type 4 translocation to the cell surface (7). In adults, alterations in zinc status may play a role in the development of prediabetes and diabetes (5, 6). Serum zinc is shown to be negatively associated with insulin resistance (8, 9), and people with diabetes often exhibit increased urinary zinc excretion and decreased plasma zinc status than do healthy individuals (7, 10). Taken together, these findings suggest that zinc may be a therapeutic option in treating diabetes or insulin resistance.

When investigating the effect of zinc supplementation on circulating insulin concentrations and glycemic control in adults, researchers have produced mixed findings (9, 11–14), likely due to differences in the age and health status of participants, zinc doses, duration of supplementation, and outcome variables. In addition, the relation between zinc and both insulin secretion and sensitivity has not been well researched in children. Cross-sectional analyses have shown that children with lower serum zinc concentrations and low dietary zinc intakes have significantly higher serum insulin concentrations and insulin resistance indexes (15–18). To date, only one randomized controlled trial (RCT) has been conducted in children. After 4 wk of zinc supplementation in obese children (19), fasting plasma glucose, insulin, and HOMA-IR were shown to decrease significantly. Importantly, to our knowledge, no previous zinc trial in children has used C-peptide to assess insulin secretion, which is a more reliable indicator of β cell function than serum insulin (20, 21). In addition, no previous studies, to our knowledge, have assessed racial differences in response to zinc supplementation. This is important because black children are often more insulin resistant than white children and present with greater values of fasting insulin and C-peptide (22–25).

The objective of this study was to conduct a secondary analysis of a previously completed RCT to determine the influence of zinc supplementation on insulin resistance and insulin secretion (via C-peptide) in healthy black and white early-adolescent girls. A secondary objective was to determine the association between zinc status and insulin outcomes. To our knowledge, this is the first study in the United States conducted in otherwise healthy children, and the first to assess racial differences, in response to zinc supplementation.

Methods

Study design and participants.

This study is an ancillary analysis to a previously completed, double-blind, zinc-supplementation RCT. This trial was registered at clinicaltrials.gov as NCT01892098. The aim of the original parent study was to determine the effects of 4 wk of zinc supplementation on biochemical markers of bone turnover in healthy female adolescents. Details on recruitment, randomization, enrollment, and compliance following the CONSORT (Consolidated Standards of Reporting Trials) guidelines have been published previously (26). Participants were healthy non-Hispanic white and black girls aged 9–11 y. Exclusion criteria included onset of menses, as measured by self-report, and unwillingness to provide a blood sample. All of the participants were required to have been at the early stages of breast development (27). All of the participants and legal guardians provided written consent and permission, respectively, and the Institutional Review Board on Human Subjects at the University of Georgia approved all study protocols and procedures.

Zinc supplementation.

Tablets containing 9 mg elemental Zn (zinc sulfate; 23 mg) and identical placebo pills (i.e., in color, size, and odor) were provided by Vesta Pharmaceuticals, Inc. Covance Laboratories verified the zinc content of the supplements via atomic absorption. Enrolled participants were randomly assigned to receive either the supplemental zinc (zinc group: n = 75; blacks: n = 35) or placebo (placebo group: n = 72; blacks: n = 32). All of the investigators, research personnel, and participants were blinded to the treatment conditions. Participants were made aware of the potential adverse events with the consumption of 9 mg Zn/d, although this level is considered safe. At the baseline visit, participants were provided a 4-wk supply of zinc or placebo tablets and were instructed to consume 1 tablet/d. Empty pill bottles and unused tablets were returned to the laboratory staff at the completion of the 4-wk trial. Compliance was measured by pill count and was calculated according to the following formula: compliance (percentage) = (number of tablets taken/number of tablets that should have been taken) × 100.

Biochemical analyses.

Fasting blood samples were collected in zinc-free tubes by a trained phlebotomist after an overnight fast at the baseline and 4-wk time points. Samples were stored at −80°C until assayed. Plasma zinc was determined by atomic absorption spectrophotometry by using a Perkin Elmer Analyst 400, and accuracy was verified on the basis of standards from the US Institute of Standards and Technology. Serum glucose concentrations were determined in triplicate by using a microtiter modification of the enzymatic Autokit Glucose method (Wako Chemicals USA). The measurement range for this assay is 0–500 mg/dL, and the mean intra- and interassay CVs were 1.8% and 2.2%, respectively. Serum insulin was analyzed in duplicate by using a Human Insulin Specific RIA (HI-14K; Millipore). The measurement range is 3.125–100 μU/mL. The mean intra- and interassay CVs were 3.5% and 5.3%, respectively. C-peptide was analyzed on an automated enzyme immunoassay analyzer (AIA-600II; Tosoh Bioscience, Inc.) in triplicate by using immunofluorescence technology. The minimum sensitivity is 0.2 ng/mL, and the mean intra- and interassay CVs were 1.67% and 1.20%, respectively. As a function of fasting C-peptide and glucose concentrations, computer model–derived HOMA-IR (HOMA2-IR) was calculated by using the HOMA2 calculator version 2.2.3 (28).

Dietary assessment.

All of the participants were instructed to avoid altering their typical dietary intake patterns during the study period. Energy, protein, and zinc intakes were assessed at baseline and at 4 wk by using 3-d diet records; and each participant and her parent or guardian received instructions on how to complete these records at home (which included 2 weekdays and 1 weekend day). Diet records were analyzed by using the Food Processor SQL version 9.7.3 (ESHA Research) and the average over 3 d was reported. In our laboratory, 1-way random-effects model, average-measure (i.e., 3-d) intraclass correlations were calculated in girls aged 6–10 y (n = 10), whose 3-d diet records were completed twice over a 2-wk period and calculated for vitamin D (r = 0.94), calcium (r = 0.71), and energy (r = 0.47).

Anthropometric, body-composition, and sexual maturation measures.

Standing height and body weight were measured at baseline and at 4 wk. Height was measured by using a wall-mounted stadiometer to the nearest 0.10 cm (Novel Products, Inc.). Weight was measured to the nearest 0.1 kg by using an electric scale (Seca Bella 840). Height and weight were used to calculate BMI-for-age percentiles for each participant (29). Percentage body fat was determined by DXA (Discovery A; Hologic, Inc.). Sexual maturation rating stage was determined by self-assessment with the use of the stages of breast development method as described by Tanner (27).

Statistical analyses.

Power analyses were computed on the basis of the primary outcome of C-peptide by using the Simple Interactive Statistical Analysis Program (30); α was set at 0.05 with the use of 80% power and a 2-tailed approach for both outcomes. By using a zinc-supplementation study in diabetic and nondiabetic adults by Oh and Yoon (12), a sample size of 127 was estimated for observing significant changes in C-peptide with zinc supplementation. Therefore, our sample size of 147 was sufficient to observe significant results for C-peptide.

Data were examined for normality, and logarithmic transformations were applied if necessary for analyses, but back-transformed for presentation of results. Baseline differences between all 4 treatment × race groups were determined by 1-factor ANOVA. Changes in outcome measures were compared by using repeated-measures, mixed-model ANOVA in an intention-to-treat analysis strategy with the use of all available data. Time was used as the within-subjects factor and treatment and race as between-subjects factors. Post hoc tests with the use of least significant difference comparisons were conducted to compare mean values at baseline and at 4 wk for C-peptide and HOMA2-IR for all 4 groups. Further analyses used Pearson bivariate correlations to quantify relations between body composition and the insulin outcomes of interest and partial correlations to quantify associations between baseline plasma zinc status and biochemical measures of insulin secretion controlling for race and age. All statistical analyses were performed by using SPSS software (version 22; IBM SPSS Statistics), and significance was set at P < 0.05.

Results

Of the 147 total participants, 4 did not complete the study (2 from the zinc group, 2 from the placebo group). Baseline descriptive participant characteristics for the total cohort as well as our treatment × race groups are provided in Table 1. At baseline, black participants had significantly higher plasma zinc than did white participants (P < 0.001). There were no other group differences observed in baseline characteristics. At baseline, percentage of body fat, fat mass, and lean mass were positively correlated with HOMA2-IR (r = 0.36, 0.43, and 0.36, respectively; all P < 0.001) and C-peptide (r = 0.377, 0.44, and 0.36, respectively; all P < 0.001). Forty-five participants (30.6%) had a BMI-for-age percentile ≥95th percentile for age. The mean dietary zinc intake for both treatment groups was less than the RDA of 8 mg/d for children in this age group (31), and 53.7% of the participants (n = 79) consumed less than two-thirds of the RDA. In addition, mean fasting serum glucose concentrations were normal in both groups, although 4 participants, 2 black and 2 white, had fasting glucose concentrations within the prediabetic range (100–110 mg/dL).

TABLE 1.

Baseline characteristics of black and white girls aged 9–11 y randomly assigned to zinc or placebo groups¹

Variable	Overall (n = 147)	Zn-White (n = 40)	Zn-Black (n = 35)	PL-White (n = 40)	PL-Black (n = 32)	P2
Age, y	10.5 ± 0.7	10.6 ± 0.7	10.5 ± 0.1	10.7 ± 0.7	10.3 ± 0.7	0.08
Sexual maturation rating stage (1–3)	2.3 ± 0.5	2.4 ± 0.5	2.3 ± 0.5	2.3 ± 0.44	2.3 ± 0.5	0.77
Weight, kg	47.0 ± 11.3	45.6 ± 10.6	48.2 ± 12.7	46.8 ± 9.5	48.0 ± 12.6	0.74
Height, cm	148 ± 6.7	147 ± 6.2	149 ± 7.5	148 ± 5.5	149 ± 7.7	0.50
BMI-for-age percentile	75.1 ± 26.1	75.7 ± 21.4	73.0 ± 31.3	77.4 ± 23.5	73.7 ± 29.0	0.88
Body fat, %	30.7 ± 8.1	31.2 ± 7.7	30.6 ± 8.7	31.5 ± 7.6	29.3 ± 8.7	0.71
Fat mass, kg	15.4 ± 7.3	15.0 ± 7.1	15.9 ± 8.2	15.4 ± 6.3	15.3 ± 7.8	0.08
Lean mass, kg	31.1 ± 5.1	30.0 ± 4.7	31.7 ± 5.5	30.7 ± 4.4	32.4 ± 5.8	0.23
Energy intake, kcal/d	1902 ± 553	1996 ± 522	1901 ± 652	1894 ± 513	1786 ± 532	0.49
Zinc intake, mg/d	4.6 ± 1.7	4.6 ± 1.7	4.8 ± 1.7	4.8 ± 1.7	4.0 ± 1.8	0.44
Plasma zinc, μmol/L	11.7 ± 1.3	10.0 ± 1.3a	13.9 ± 1.2b	10.0 ± 1.3a	14.4 ± 1.2b	<0.001
Insulin, μU/mL	24.0 ± 1.6	21.2 ± 1.5	27.2 ± 1.6	22.9 ± 1.4	26.0 ± 1.6	0.06
Glucose, mg/dL	86.1 ± 7.6	86.6 ± 7.3	84.9 ± 8.8	87.1 ± 7.2	85.3 ± 7.1	0.55
C-peptide, ng/mL	1.9 ± 1.5	1.8 ± 1.5	2.0 ± 1.6	2.0 ± 1.6	1.7 ± 1.6	0.39
HOMA2-IR	1.4 ± 1.6	1.3 ± 1.5	1.4 ± 1.6	2.0 ± 1.6	1.5 ± 1.6	0.40

¹Values are means ± SDs. Labeled means without a common superscript letter differ, P < 0.05. HOMA2-IR, computer model–derived homeostatic model assessment of insulin resistance; PL-Black, black girls, placebo group; PL-White, white girls, placebo group; Zn-Black, black girls, zinc group; Zn-White, white girls, zinc group.

²Tests of significance among all 4 treatment × race groups were based on 1-factor ANOVA.

Baseline plasma zinc was not significantly correlated with C-peptide (r = −0.07, P = 0.38), HOMA2-IR (r = −0.09, P = 0.30), or insulin (r = −0.06, P = 0.44) after statistically controlling for age and race. Results from the same correlations controlling for race and sexual maturation stage were not different. Figure 1 shows the mean changes in C-peptide and HOMA2-IR for each group after 4 wk of treatment. Black girls who received zinc experienced a slight decrease in C-peptide (−10.6%) and HOMA2-IR (−10.4%), whereas the black girls who received the placebo showed increases in C-peptide (15.7%) and HOMA2-IR (17.7%). However, the divergent pattern in these outcomes was not observed in white girls. There were no significant treatment × time interactions with C-peptide (P = 0.25) and HOMA2-IR (P = 0.24). Furthermore, the treatment × race × time interactions for C-peptide and HOMA2-IR did not reach significance (both P = 0.08; Figure 1). Post hoc comparisons showed that C-peptide and HOMA2-IR increased slightly from baseline to 4 wk in the black girls who received placebo, although the differences did not reach significance (both P = 0.06). Our post hoc analyses did not show significant changes in C-peptide or HOMA2-IR from baseline to 4 wk in the black girls who received zinc or in the white girls who received either treatment.

FIGURE 1.

Changes from baseline in C-peptide (A) and HOMA2-IR (B) in response to 4 wk of daily supplementation of either placebo or zinc randomly assigned to black and white girls aged 9–11 y. Values are means ± SDs. HOMA2-IR, computer model–derived homeostatic model assessment of insulin resistance; PL-Black, black girls, placebo group; PL-White, white girls, placebo group; Zn-Black, black girls, zinc group; Zn-White, white girls, zinc group.

Discussion

Much of the previous work related to zinc and insulin outcomes has focused on its potential clinical benefits for type 2 diabetes (5–7, 10). To our knowledge, this is the first study to examine the effects of zinc supplementation on insulin secretion and sensitivity in otherwise healthy adolescent girls through a randomized, double-blind, placebo-controlled study design. Four-week supplementation with 9 mg elemental Zn/d did not significantly influence insulin secretion (i.e., C-peptide) or sensitivity (i.e., HOMA2-IR). The black girls who received zinc experienced a slight decrease in C-peptide (10.6%) and HOMA2-IR (10.4%) after 4 wk, whereas the black girls who received placebo experienced an increase in these outcomes (15.7% for C-peptide and 17.7% for HOMA2-IR). However, this divergent pattern was not observed in the white girls, and the treatment × race × time interaction for both C-peptide and HOMA2-IR did not reach significance (both P = 0.08).

Two previous studies reported reductions in insulin outcomes with zinc supplementation (13, 19). For example, the provision of 20 mg elemental Zn given over 8 wk to nondiabetic obese children significantly decreased fasting insulin and HOMA-IR (19). A similar study in adults showed that fasting insulin decreased significantly after supplementation with 30 mg Zn daily for 4 wk (13). Findings from a recent cell culture study validate conclusions that zinc supplementation decreases insulin production. The study by Slepchenko et al. (32) showed that zinc ions co-secreted from pancreatic β cells with insulin may have an inhibitory effect on glucose-stimulated insulin secretion, thus acting through a negative feedback loop. The marginal reductions in C-peptide and HOMA2-IR that were observed in the current study, although not significant, are consistent with results from previous trials.

Although most studies conducted to date have used insulin resistance and fasting insulin as the primary outcome measures (9, 11, 14, 19), a strength of the current study is the additional use of C-peptide, which is a more reliable indicator of insulin secretion (20, 21). One other zinc supplementation study that used C-peptide as an outcome measure showed that C-peptide was significantly increased from baseline after zinc supplementation in diabetic adults with >4 y duration of diabetes and in those with marginal zinc status as determined by plasma zinc (12). The differences between these results and those of the current study are likely due to the differences in age and health status of the populations studied. The participants in the current study were considered healthy when we recruited them and only 4 participants were defined as prediabetic. Whether or not these 4 participants were included in the statistical analyses did not affect our findings.

A secondary aim of this study was to determine whether the effect of zinc supplementation on C-peptide and insulin resistance differed by race. Black children are often more insulin resistant than white children and present with greater values of fasting insulin and C-peptide (22–25). In addition, it has been shown that African-American adolescents secrete more insulin during a hyperglycemic clamp study than their white peers with comparable insulin sensitivity and body composition (23). These racial differences in insulin secretion may explain why the increase in C-peptide after 4 wk was greater in the black children than in the white children who were randomly assigned to receive the placebo. Although the population in the current study was more heterogeneous than those in other studies, our findings may only be applicable to apparently healthy black and white children during the early stages of maturation.

It is important to note that exploring the relation between zinc and insulin secretion and sensitivity in adolescents may be complicated by normal fluctuations in insulin secretion and insulin sensitivity throughout pubertal development. During the early stages of sexual maturation, a decline in insulin sensitivity is observed in both boys and girls (33–35). These changes have been observed in lean and obese children alike, although the insulin resistance may worsen with a longer duration of obesity (35). Children seem to experience a progressive decrease in insulin sensitivity from sexual maturation rating stage I to sexual maturation rating stages III and IV (36). Insulin secretion increases in response to the decreased insulin sensitivity, but whether or not it is able to fully compensate is still unclear (33, 36, 37). Moreover, it has been shown that insulin secretion and sensitivity differ by race, with black adolescents having greater insulin secretion and lower insulin sensitivity than white adolescents (22–24).

Strengths of the current study include the use of C-peptide as an outcome measure, which is a more reliable indicator of insulin secretion than fasting insulin (20, 21), and the inclusion of both black and white participants. Previous studies have assessed outcomes in a single-race population, and to our knowledge no previous study has assessed outcomes in black children.

We also acknowledge potential limitations with the current study. As discussed earlier, participants in our study received a lower amount of zinc, 9 mg/d, than in previous studies (9, 11–14). However, it was previously shown by our laboratory that 9 mg Zn/d given over 4 wk was sufficient to significantly increase plasma zinc concentrations (26), suggesting that 9 mg/d should also be sufficient to elicit changes in the intended biochemical outcomes in the current study. We were also limited by the 4-wk study duration, which may have been insufficient to detect changes in our outcomes of interest. However, other zinc-supplementation studies have reported significant changes in fasting insulin and HOMA-IR after 4 wk (12, 13, 19), suggesting that 4 wk should be sufficient to produce significant changes in C-peptide and HOMA2-IR. A final limitation of the current study is the use of fasting measures of insulin secretion and β cell function rather than dynamic measures of insulin response, such as with a mixed-meal challenge or an oral-glucose-tolerance test. To our knowledge, no study of zinc supplementation and insulin outcomes has used these dynamic methods to assess insulin response. Future studies in this area should use these dynamic methods in order to more accurately determine how zinc supplementation affects the insulin response to a glucose load.

In conclusion, zinc supplementation over 4 wk did not alter insulin secretion or insulin resistance in healthy female adolescents. Supplementation with 9 mg elemental Zn/d may have marginally attenuated increases in C-peptide and HOMA2-IR in black children, but future studies are needed to confirm this. If zinc supplementation promoted improved insulin outcomes in black adolescents, a high-risk group for type 2 diabetes, the clinical implications could be profound. However, based on the findings from the current study, it is premature to consider supplemental zinc at or above the RDA for improved clinical outcomes related to insulin. Future zinc RCTs among adolescents in the early stages of maturation should be considered, especially those that explore the racial differences in insulin outcomes with the use of dynamic measures of insulin sensitivity and glucose handling.