Association between diabetes and levels of micro-nutrients in a Middle Eastern population - A case-control study

Design, population, and setting

A case control study, using data collected by Qatar Biobank (QBB), was conducted on Qatari citizens and long-term residents. The QBB is a population-based research initiative that creates a depository of biological samples and information on the health and lifestyle of Qatari citizens and long-term residents, aiming to prospectively recruit 60,000 volunteers (13, 14). Participants above 18 years of age with data on micronutrients were eligible for inclusion in the present study. During this study, participants with diabetes were randomly selected from the QBB database and matched, using age, gender and nationality, to those without diabetes. Participants without data on micronutrients, those aged below 18 years and pregnant women were excluded.

Data collection

Each participant completed a questionnaire, had anthropometry measured, and had blood drawn for measuring serum micronutrients. The data collected for each participant were: demographics, data on self-reported doctor-diagnosed diabetes, duration of diabetes, nutritional supplement usage, income, smoking status, blood pressure, anthropometry, HbA1c level, serum levels of lipids, random glucose, and micronutrients. The micronutrients measured were copper (Cu), zinc (Zn), magnesium (Mg), iron (Fe), folate, and vitamin B12.

Diabetes measurements

Participants were classified into the diabetes group if they self-reported being diagnosed by a doctor or if they had an HbA1c ≥6.5% or a random blood glucose level ≥11.1 mmol/L, according to the American Diabetes Association (ADA) criteria (14). Controlled diabetes was defined based on ADA guidelines as HbA1c ≤ 7% (13).

Micro, macro elements and vitamins measurements

For each participant, micronutrients were assessed from serum. Levels of micronutrients were classified as ‘low’ according to the normal laboratory ranges for adults on UpToDate (14), using the following cut-offs: Mg (≤0.66mmol/L), Fe (≤9 μmol/L), Zn (≤11.5μmol/L), Cu (≤15.7μmol/L), folate (≤4.1nmol/L), and vitamin B12 (≤147.5pmol/L).

Ethics and informed consent

At the QBB, participants gave written informed consent, before data were collected. This study received ethical approval from the Qatar University’s Institutional Review Board (QU-IRB) (reference number: QU-IRB1228-E/20), and the Qatar Biobank’s IRB (reference number: QF-QBB-RES-ACC-0186).

Data analysis

Histograms were used to check for the normality of data distribution for measured variables. Descriptive data were presented as mean and standard deviation (SD) if normally distributed, or as median and interquartile range (IQR) if not normally distributed. Categorical variables were presented using frequencies and percentages. Boxplots were used to compare the levels of the micronutrients between participants without diabetes, those with controlled and participants with uncontrolled diabetes. The micronutrient levels were compared between participants with and without diabetes using the t-test for independent groups if normally distributed or Wilcoxon rank-sum test if not normally distributed. Comparisons of micronutrient levels between the three groups (controlled, uncontrolled, and no-diabetes) were tested using one-way ANOVA if data were normally distributed, or Dunntest if data were not normally distributed with post hoc test done using Bonferroni adjustment.

Multivariable logistic regression was carried out for the association between diabetes and the micronutrients, with separate models for each micronutrient. To investigate the association between micronutrients and uncontrolled diabetes, multivariable multinomial logistic regression was used, with “no diabetes” as the base-outcome. In all models, we adjusted for age, gender, BMI, smoking, and income, as they are known to be associated with both micronutrients’ intake and diabetes (15-17). For each multivariable regression model, we tested model specification using linktest. Exact p-values were reported and interpreted as evidence against null hypothesis and 95% confidence intervals (95% CI) were reported for odds ratios (OR). All data analyses were performed using Stata 16.0 statistical software.

Characteristics of participants by diabetes status

A total of 1118 participants, of which 33.5% had diabetes, were included in the study. Participants with diabetes were slightly older than those without (mean age 44.4 years old (9.4) vs 40.3 years old (9.9), P<0.001). Participants with diabetes compared to those without, had significantly higher mean BMI (31.5 kg/m² (SD 5.4) vs 29.8 kg/m² (SD 5.6), respectively, P<0.001) (Table 1).

Comparison of characteristics by controlled diabetes status

As expected, median HbA1C was higher in the participants with uncontrolled diabetes, compared to those with controlled diabetes (7.8% (IQR 7.4-8.9) vs 6.0% (IQR 5.5-6.4), respectively, P<0.001) (Table 1). The median duration of diabetes for participants with uncontrolled diabetes was longer compared to the participants with controlled diabetes (7 years (IQR 4-13) vs 4 years (IQR 2-9), respectively, P<0.001). A greater percentage of participants with controlled diabetes reported taking multi-vitamins/minerals supplements compared to those with uncontrolled diabetes (51.1% vs 45.0%, respectively, P=0.011)

Comparison of micronutrients levels between participants with and without diabetes

Median serum levels of iron and folate were significantly higher in participants with diabetes compared to those without (Table 2). Median Mg serum levels were significantly lower in participants with diabetes, compared to those without diabetes (0.79 mmol/L (SD 0.07) vs. 0.84 mmol/L (SD 0.06), P<0.001). For categorized micronutrients, compared to participants without diabetes, a significantly higher proportion of participants with diabetes had low magnesium (48.9% vs 22.3%, p<0.001) (Table 2).

The boxplots in Figure 1 and Table 2 show a comparison of the levels of micronutrients between participants without diabetes, those with controlled diabetes and participants with uncontrolled diabetes. The mean Mg level was significantly higher in participants with controlled diabetes compared to those with uncontrolled diabetes (0.81 mmol/L (SD 0.07) vs 0.76 mmol/L (SD 0.07), respectively, P<0.001)) (Table 2 & Fig. 1). The median vitamin B12 levels were significantly higher in participants with uncontrolled diabetes compared to participants with either controlled or without diabetes (Table 2). There were no significant differences between the three groups in the remaining micronutrients (Table 2).

• Download figure
• Open in new tab

Fig. 1: Boxplots comparing selected micronutrent levels across the three groups

Association between diabetes and micronutrients– multivariable logistic regression

After multivariable logistic regression, diabetes, compared to no diabetes, was significantly associated with a 3-fold odds of having low magnesium (OR 3.32, 95% CI 2.47-4.47, p<0.001), a 49% increase in the odds of having low iron (OR 1.49, 95% CI 1.03 – 2.16, p = 0.034). Diabetes was also associated with a 49% increase in the odds of having low Zn, although with weak evidence against the null hypothesis (p=0.334) (Table 3).

Association between micronutrients and controlled diabetes – multivariable multinomial logistic regression

In further stratified analyses, uncontrolled diabetes was associated with 5.6-fold increased odds of having low magnesium (OR 5.57, 95% CI 3.65-8.57, p<0.001) and two-fold increased odds of having low iron (OR 1.75, 95% CI 1.01-3.03, p=0.047). These associations were still observed in participants with controlled diabetes, but they were noticeably weaker (Table 3). Lastly, in participants with controlled diabetes, diabetes was associated with two-and-a-half-fold increase in the odds of having low Zn, although there was weak evidence against the null hypothesis at the study’s sample size (Table 3).