Zinc status in attention-deficit/hyperactivity disorder: a systematic review and meta-analysis of observational studies

Previous studies regarding the zinc status in attention-deficit/hyperactivity disorder (ADHD) yielded inconsistent results. Thus, the present meta-analysis was aimed to estimate the association between hair and serum/plasma zinc levels and ADHD. Online databases of Medline, EMBASE, and Scopus were searched up to October 2020 with no limitation in time and language. Weighted mean differences (WMDs) of hair and serum/plasma zinc levels were calculated using a random-effects model. Overall, 22 articles with 1280 subjects with ADHD and 1200 controls were included. The pooled effect size indicated that serum/plasma zinc levels in subjects with ADHD were not statistically different than their controls (WMD = − 1.26 µmol/L; 95% CI − 3.72, 1.20). Interestingly, the exclusion of one study from the analysis showed that people with ADHD significantly have lower circulating levels of zinc compared to their controls (WMD: − 2.49 µmol/L; 95% CI − 4.29, − 0.69). Also, the pooled effect size indicated that hair zinc levels in cases with ADHD were not statistically different than their controls (WMD = − 24.19 μg/g; 95% CI − 61.80, 13.42). Present meta-analysis raises the possibility that subjects with ADHD are prone to have declined levels of zinc levels. Based on current findings, screening the zinc levels in subjects with ADHD could be reasonable. Further well-designed studies are needed to clarify the role of zinc in the etiology of ADHD.

Quality assessment. Newcastle-Ottawa Scale (NOS) was used to evaluate the quality of the included studies. Articles with a total score of 0-4, 5-7, and 8-10 were considered as low, moderate and high quality, respectively.

Statistical analysis.
To merge data, we used the random-effects model. In order to the calculation of effect size, the concentrations of zinc were converted to µmol/L. Heterogeneity of studies is determined using I 2 and Chi-square test, high heterogeneity is determined by I 2 above 50% and P value < 0.1. In order to determine the origin of heterogeneity, we used subgroup analysis and sensitivity analysis. Subgroup analysis was performed according to the type of samples (serum, plasma, blood), year of publication (≤ 2010, > 2010), method of zinc assessment (atomic absorption spectrophotometer, others), study design (case-control, cross-sectional), and sample size (< 100, > 100). Begg's and Egger's tests were used to examine the publication bias. STATA software (version 14) 18 was used for statistical analysis. Statistically significance is confirmed with P value less than 0.05.

Results
Study selection. The study selection process is illustrated in Fig. 1. Among the 1800 studies found based on an electronic and manual search for all trace elements, 513 studies were duplicate. Screening the title and abstracts excluded 1185 documents. Among the remaining 102 articles 75 studies were related to other trace elements. Three studies were excluded due to unavailability of information regarding zinc levels [19][20][21] and two studies were excluded due to lack of reporting information in case and control groups separately and reporting the level of zinc in hemoglobin 22,23 . Finally, 22 articles were included in this meta-analysis, 14 studies assessed circulating levels of zinc 13,24-36 and 8 studies reported hair zinc levels 12,37-43 . Characteristics of included studies. Characteristics of the included studies were summarized in Table 1.
Included studies were conducted between 1990 and 2020, examining 1280 people with ADHD and 1200 controls. Except for one study 24 in which both case and control groups had diabetes, other studies used healthy individuals and people with ADHD as control and case groups, respectively. Of the available studies, four were conducted in Egypt 24,28,29,33  Twenty studies had case-control design and only two studies 13,28 were conducted by cross-sectional design. According to the Newcastle-Ottawa scale, nine studies were assigned to moderate quality 12,25,27,29,30,32,36,40,41 , and the rest of them had high quality (Table 2).

Meta-analysis of mean blood zinc levels.
Fourteen studies assessed the association between blood zinc level and ADHD 13,24-36 , involving 902 cases and 818 controls. The pooled effect size indicated that serum/plasma zinc levels in subjects with ADHD were not statistically different than their controls (WMD = − 1.26 µmol/L; 95% CI − 3.72, 1.20, P = 0.31, Fig. 2). Interestingly, sensitivity analysis showed that exclusion of Abdelnaby's study 28 from the analysis changed the overall effect size (WMD: − 2.49 µmol/L; 95% CI − 4.29, − 0.69). A significant heterogeneity was detected among studies (I 2 = 98.2%, P < 0.001). Despite classification of the studies, no possible source of heterogeneity was found and the result remained non-significant in all categories (Table 3). No evidence of publication bias was observed among included studies (P = 0.87, Begg's test and P = 0.45, Egger's test).

Meta-analysis of mean hair zinc levels.
Eight studies reported sufficient data regarding hair zinc levels in ADHD and control subjects 12,37-43 , involving 375 cases and 382 controls. The pooled effect size indicated that hair zinc levels in cases with ADHD were not statistically different than their controls (WMD = − 24.19 μg/g; 95% CI − 61.80, 13.42, P = 0.20, Fig. 3). However, significant heterogeneity was detected across the studies (I 2 = 98.1%, P < 0.001). Despite the different subgroup analysis, we could not detect the potential source of observed hetero- www.nature.com/scientificreports/ geneity, as shown in Table 2. There was no evidence of publication bias among included studies (P = 0.62, Begg's test and P = 0.16, Egger's test).

Discussion
The present meta-analysis, including 22 studies and a total of 2428 people, showed that there was no statistically significant difference in serum/plasma and hair zinc levels between patients with ADHD and their controls. There was substantial heterogeneity among included studies. However, sensitivity analysis in studies examining the circulating zinc levels showed that excluding one study 28 changed the overall effect. Circulating levels of zinc were significantly lower in subjects with ADHD compared to healthy controls after excluding Abdelnaby's study 28 . Zinc deficiency is involved in a variety of neurological disorders including autism, seizures, depression, and anxiety disorders 44 . However, the exact mechanism of zinc in ADHD is still unclear. Dopamine is a neurotransmitter plays a crucial role in the pathophysiology of ADHD 9 . Previous studies reported that zinc is involved in the production of melatonin which could regulate dopamine levels and homeostasis 45,46 . Dysfunction in the dopamine transporter is another pathway that contributed to the etiology of ADHD 9 . Zinc binding to the dopamine receptors inhibits the dopamine re-uptake and increases the carrier-mediated dopamine efflux 9,46 . Also, zinc is an important cofactor for several enzymes in the brain involved in the neurotransmitters and prostaglandins production 9 .
Several studies have suggested the role of inflammation and oxidative stress in the pathogenesis of ADHD 47,48 . Although subjects with ADHD have normal levels of antioxidant capacity, their reaction to oxidative stress is impaired 47 . Elevated levels of pro-inflammatory cytokines could decrease the levels of zinc in patients with ADHD through the sequestration of zinc in the liver and spleen 49 . Zinc could exert anti-oxidative and antiinflammatory properties through the protection of sulfhydryl groups of proteins from oxidation 50 . Zinc takes part in antioxidant enzyme production and acts as a cofactor of several enzymes 50 . Also, zinc modulates the chronic inflammatory status by reducing pro-inflammatory cytokines 51 . On the other hand, zinc supplementation showed beneficial effects in the alleviation of hyperactivity symptoms in zinc-deficient ADHD subjects 52 . Moreover,   www.nature.com/scientificreports/ 150 mg/day zinc supplementation for 12 weeks led to a significant reduction in symptoms of hyperactivity, impulsivity, and impaired socialization in patients with ADHD 53 . Although, 30 mg/day zinc supplementation showed no significant effects on primary outcomes compared to the placebo, which might be due the low dosage of zinc 54 . Lower levels of zinc in subjects with ADHD may be attributed to the dietary zinc intake or zinc absorption 49 . Also, zinc-wasting in the urine is another possible cause of low levels of zinc in children with ADHD 55 . It has been suggested that hyperactive children have increased levels of urinary zinc and reduce levels of plasma 49 .
Sensitivity analysis showed that the exclusion of Abdelnaby's study 28 from the analysis changed the overall effect size. The pooled analysis without mentioned study showed significant lower levels of serum/plasma zinc in subjects with ADHD compared to their controls. Indeed, the mentioned study showed a significant higher levels of serum/plasma zinc in subjects with ADHD compared to the controls. This contradictory finding could be related to several factors e.g. different study design, small sample size, the different method in zinc measurement, and high risk of bias (NOS = 6).
The present study has some limitations that should be acknowledged. We observed a significant heterogeneity among included studies that could affect the generalizability of results. However, our attempts to detect the potential source of heterogeneity through different subgroup analysis were unsuccessful. The observed heterogeneity in the present meta-analysis could be related to several factors including demographic and clinical differences, BMI, study design, adjusted models for statistical analysis, risk of bias, and methods for assessing zinc levels. Small sample sizes of individual studies are another limitation of the present study. Almost all of the included studies except one 26 were performed on less than 200 participants. Moreover, included studies did not evaluate the dietary intake of zinc in study participants which could affect the results because that amount of zinc intake is related to the serum zinc concentration 56 . Also, many factors could affect hair zinc levels 57 , which should be taken into the interpretation of results. www.nature.com/scientificreports/

Conclusion
Present meta-analysis raises the possibility that subjects with ADHD are prone to have declined levels of zinc levels. Based on current findings screening the zinc levels at the beginning of the diagnosis in subjects with ADHD could be reasonable. Further well-designed studies are needed to clarify the role of zinc in the etiology of ADHD. Table 3. Subgroup analysis to assess the serum and hair zinc levels in subjects with ADHD.