By David Kiefer, MD, Editor

Clinical Assistant Professor, Department of Family Medicine, University of Wisconsin; Clinical Assistant Professor of Medicine, Arizona Center for Integrative Medicine, University of Arizona, Tucson

Dr. Kiefer reports no financial relationships relevant to this field of study.


  • Researchers in Bangladesh randomized a small group of adults with prediabetes to either 30 mg zinc sulphate or placebo daily for six months.
  • After six months, there were improvements in fasting glucose, lipids, C-reactive protein, and calculated measures of insulin and pancreas function (HOMA scale).
  • Generalizability of these results is unknown, especially given that the population studied historically has had low dietary zinc intake, making them possibly more amenable to zinc repletion.

SYNOPSIS: Thirty milligrams of zinc sulphate over six months improved a variety of metabolic parameters in a Bangladeshi population with prediabetes.

SOURCE: Islam MR, Attia J, Ali L, et al. Zinc supplementation for improving glucose handling in pre-diabetes: A double blind randomized placebo controlled pilot study. Diabetes Res Clin Pract 2016;115:39-46.

Zinc may, in fact, not just be for the treatment of colds anymore. This essential mineral, a cofactor in many biological reactions, also has been found to nudge a variety of clinical conditions.1 The researchers conducting this clinical trial were following up on previous studies that found a correlation between higher dietary zinc intake and a lower risk of type 2 diabetes. Furthermore, they cited work that shows a small, likely clinically insignificant, reduction in fasting glucose and hemoglobin A1c with zinc supplementation, although they posit that zinc supplementation might stave off the progression from insulin resistance or metabolic syndrome to overt diabetes. Islam et al attempted to add to some preliminary work on this hypothesis.

This study took place in Bangladesh, involving a collaboration of Bangladeshi and Australian researchers. Adults aged 30-65 years diagnosed with prediabetes within a “catchment area” of the Bangladesh Institute of Health Sciences hospital were screened (n = 224, n = 177 with phone numbers allowing contact) for randomization to either placebo (n = 27) or 30 milligrams of zinc sulphate (n = 28) daily for six months. Prediabetes was defined as fasting blood glucose concentration between 5.6-6.9 mmol/L.

Exclusion criteria were extensive and are detailed in Table 1. The zinc and placebo tablets were identical in size, taste, color, and smell, and study participants received a six-month supply of tablets upon recruitment for participation. Both groups received advice on diet and lifestyle activities (details not provided) and were called every two weeks to check on “compliance, adverse events, and morbid events.” See Table 2 for a timeline of interventions and data collection during the study. Note that socioeconomic data, 24-hour diet information, and blood tests were obtained. The primary outcome of the study was any change in fasting glucose levels. Secondary outcomes included calculation of beta-cell function, insulin sensitivity, and insulin resistance by way of the Homeostasis Model Assessment (HOMA), and serum zinc and lipid values at baseline and six months.

Table 1: Exclusion Criteria, Preempting Inclusion in This Trial

  • Diabetes diagnosis
  • Pregnancy
  • Morbid obesity
  • Renal insufficiency
  • Cardiovascular disease
  • Thyroid disease
  • Liver disease
  • “Any other known chronic diseases”
  • Multiple comorbidities
  • Physical inactivity (wheelchair bound)
  • Psychiatric disorders
  • Gastrointestinal disorders
  • Taking vitamins, diuretics, or “complementary medicines”

Table 2: Data Collection and Interventions During the Six-month Study


Every Two Weeks

Six Months

Home visit to confirm diagnosis

Advice on “healthy eating and lifestyle”

Face-to-face interview*

24-hour dietary recall

Blood tests**

Calculation of HOMA

Phone call: compliance, adverse effects, morbid events

Blood tests**

Calculation of HOMA

* Face-to-face interview collected “… socioeconomic and other relevant information.” (details not provided)

** Blood tests included fasting glucose, serum zinc, serum insulin, hemoglobin A1c, lipid profile, C-reactive protein

With respect to baseline characteristics, the placebo group was more affluent and had higher lipid values than the zinc group; other parameters were similar. Table 3 shows the results for primary and secondary variables for zinc vs. placebo over the course of the research trial. For fasting glucose and the three subsets of the HOMA calculation, the zinc group showed statistically significant improvements from baseline to six months; the placebo group improved slightly for fasting glucose and insulin sensitivity, although the beta-cell function and insulin resistance were unchanged (see P values in Table 3). Also interesting was the comparison between the zinc and the placebo group at the six-month mark. The zinc group showed statistically significant improvements compared to the placebo group for all four parameters (P values ranging from 0.01 to < 0.001). In addition, the zinc group showed a lower serum triglyceride, low-density lipoprotein, and C-reactive protein, and higher high-density lipoprotein (all P values < 0.001) than the control group after six months.

Table 3: Study Results for the Zinc and Placebo Groups Over the Six-month Study Period





Six Months


Six Months

Fasting glucose (mmol/L)


5.69 (P = 0.05)


5.37 (P < 0.001)

Beta cell function (HOMA mean)


78.7 (P = 0.14)


87.7 (P < 0.001)

Insulin sensitivity (HOMA mean)


88.6 (P = 0.001)


90.4 (P = 0.012)

Insulin resistance (HOMA mean)


1.13 (P = 0.10)


1.09 (P = 0.002)

The P values displayed compare the placebo and zinc baseline values with their own values at six months.

Of note, there were no dropouts during the six months, and no adverse effects nor side effects occurred. Furthermore, the authors reported that none of the participants progressed from prediabetes to type 2 diabetes over the six-month period.


This study is an interesting example of the effects of a single nutrient on a variety of metabolic parameters. These results should be somewhat predictable, given the range of physiological systems touched by zinc. For example, the authors provided a comprehensive review of how zinc might be involved with insulin and glucose regulation. However, it is important to remember the specific demographic studied here. The researchers mentioned the fact that Bangladesh has low dietary zinc intake and a high dietary phylate intake, further compromising zinc absorption. It is likely that the study participants were zinc-deficient, and there are examples of how zinc repletion in the context of zinc deficiency might lead to positive clinical outcomes;1 it’s less clear whether such results would be achieved if a person has adequate amounts of zinc. Hence, the results from this study may or may not apply to the typical patient in the United States.

On a further note relevant to applicability, the exclusions seemed extreme enough that it’s hard to believe there were any study participants left meeting the criteria. The authors justified the list, referencing prior work, as with exclusion criteria in general, that they wanted to minimize confounders and drug interactions. Nonetheless, it seems overly limiting and leaves clinicians wondering how to apply these results to their own patient population. Many patients with prediabetes probably are taking some vitamin or have even the most minor of gastrointestinal disorders (occasional gastroesophageal reflux disease or indigestion). It would be interesting to see a comparable study on “more normal” people with a mixture of conditions and medications or dietary supplements.

The lack of adverse effects or side effects almost seems too good to be true. Most placebo-controlled trials report side effects even for the placebo group, so the lack of this in the current study casts some doubt on the methodological rigor and data collection. The phrasing of the phone call query into adverse effects is not listed; perhaps leading questions dissuaded study participants from reporting their symptoms. We don’t know.

And how clinically relevant are these results? The authors mentioned a number needed to treat (NNT) of 12 for metformin in cases of prediabetes over three years, based on an approximate improvement in hemoglobin A1c of 1%. Meta-analyses estimate that zinc improves hemoglobin A1c by approximately 0.5%, so the authors postulated a NNT of 24 (this study didn’t measure hemoglobin A1c). This is not a negligible effect, and given its safety profile (a few minor safety concerns, though, at higher doses, possible copper deficiency), perhaps clinicians might consider turning to zinc early in the prediabetes treatment algorithm. In some cases, a serum zinc level may also be helpful in guiding therapy, although there are nuances to interpreting the results.2 A quick glance at the data showed marginal improvements in fasting glucose, but (not included above, but provided in the research article) more robust changes in lipids and C-reactive protein, both between groups and over the six-month study period. The mechanism is plausible, and these findings are believable. Now, the next step is to broaden the demographic and study size and see if the results translate to a wider audience, even if zinc deficiency is not present.


  1. Saper RB, Rash R. Zinc: An essential micronutrient. Am Fam Physician 2009;79:768-772.
  2. Roohani N. Zinc and its importance for human health: An integrative review. J Res Med Sci 2013;18:144-157.