Abstract
Background:
ADHD is associated with oxidative stress (OS), possibly stemming from deficiencies in essential nutrients. Previous randomized controlled trials (RCTs), including Micronutrients for ADHD in Youth (MADDY), demonstrated improved symptoms in response to treatment with multinutrients compared to placebo. It remains unknown whether multinutrient supplementation influences antioxidant status and OS, and if these factors contributed to improvements observed in children with ADHD in the MADDY RCT.
Objectives:
Utilizing samples from the MADDY RCT, (1) compare the change in antioxidant (AO) and OS biomarkers after 8 weeks of multinutrient supplementation vs placebo, and (2) evaluate these biomarkers at baseline, and their change after 8 weeks, as moderators/mediators of treatment response.
Methods:
Activity of glutathione peroxidase (GPx) and glutathione reductase (GR) enzymes, plus oxidative stress index (OSI) based on ratio of reactive oxygen metabolites (ROM) to biological antioxidant potential (BAP), were measured in plasma at baseline and week 8. Differences between groups were determined using two-sample t-test or Mann-Whitney U-test. Logistic regression models assessed AO/OS biomarkers for mediation/moderation of treatment response.
Results:
Plasma from 77 children (aged 9.9 ± 1.7 years; 71% male) treated with multinutrients (n = 45) or placebo (n = 32) was analyzed. After 8 weeks, ROM decreased with multinutrients and increased with placebo (−14.3 vs. +26.8 Carratelli units, p = .017); but no significant differences in OSI, BAP, GPx, and GR between groups. None of the baseline AO/OS biomarker levels were moderators of treatment response. Eight-week change in both OSI and ROM trended toward mediation of treatment response (OR = 0.00058, 95% CI [0.000, 2.30], p = .078 and OR = 0.985, 95% CI [0.968, 1.002], p = 0.086, respectively) but did not reach significance.
Conclusions:
Eight weeks of multinutrient supplementation in children with ADHD reduced ROM without significant change in antioxidant status, suggesting decrease in oxidative stress. Given the preliminary signals associating a decrease in OS with symptom improvement following multinutrient supplementation, future research is warranted to understand OS in ADHD pathogenesis.
Clinical Trial Registry:
Keywords: ADHD, children, oxidative stress, antioxidant, multinutrient, supplement, glutathione, RCT
Introduction
Nearly 10% of children in the United States (U.S.) are affected by ADHD, a common neurodevelopmental disorder that is associated with poor academic, social, economic, and health outcomes (Danielson et al., 2018; Zhao et al., 2019). The etiology of ADHD is multifactorial including genetic diathesis, differences in brain circuitry, environment, and psychosocial factors (Brown et al., 2018). Environmental risk factors for ADHD, such as air pollution, toxic metals (e.g., lead), pesticides, and cigarette smoke, are commonly associated with increased production of reactive oxygen species (ROS), which can overwhelm the body’s antioxidant defenses, leading to increased oxidative stress (OS; Al-Gubory, 2014; Lange et al., 2017).
There are three major sources of ROS in humans: (1) pathophysiological byproduct of oxidative phosphorylation and other metabolic processes, (2) inflammatory ROS produced for defense by innate immune cells, and (3) environmental (diet, pollution, radiation, and external sources of heavy metals; Figueira et al., 2013). The body’s main defense mechanism against ROS damage utilizes antioxidant enzymes including glutathione peroxidase (GPx), superoxide dismutase (SOD), and catalase (CAT). These enzymes require essential minerals manganese (Mn), copper (Cu), zinc (Zn), selenium (Se), and iron (Fe) as cofactors in their activities. The secondary defense layer includes non-enzymatic antioxidants, either dietary (i.e., vitamin C and vitamin E) or endogenous antioxidants, with reduced glutathione (GSH) being most prevalent antioxidant (Lee et al., 2020; Singh et al., 2019). GSH is readily oxidized to form glutathione disulfide (GSSG) by seleno-enzyme GPx during periods of oxidative stress and reduced to its antioxidant GSH form via glutathione reductase (GR). Thus, the activity of both enzymes (GPx and GR) is critical to antioxidant defense (Lee et al., 2020; Singh et al., 2019). In addition to selenium (Se), various other nutrients, including amino acids, riboflavin (B2), vitamin B12, magnesium (Mg), and selenium (Se), are required for glutathione synthesis and redox cycling (Lange et al., 2017). Given the large number of components that compose the antioxidant defense systems and contribute to oxidative stress, several global measures of overall oxidative balance have been developed. For example, an oxidative stress index (OSI) calculated as a ratio of derivatives of reactive oxygen metabolites (ROM) and biological antioxidant potential (BAP) has demonstrated elevated OS in children with ADHD and autism spectrum disorder compared to typically developing children (Kitaoka et al., 2020; Morimoto et al., 2020; Morimoto et al., 2023; Sanchez-Rodriguez & Mendoza-Nunez, 2019).
Accumulating evidence suggests that children with ADHD may have an imbalance between oxidants and antioxidants, leading to OS and oxidative damage (Joseph et al., 2015; Verlaet et al., 2018). Physiologically, the brain is especially vulnerable to the effects of OS due to its high level of oxygen utilization, high lipid content, and lower antioxidant concentration (Lee et al., 2020; Lopresti, 2015). Additionally, degradation of dopamine by auto-oxidation or metabolism via monoamine oxidase (MAO) enzyme generates ROS, which contributes to degeneration of dopaminergic neurons, suppression of striatal dopamine release, and impact on dopamine receptor function (Lopresti, 2015). This mechanism is particularly relevant in the context of ADHD, where dopamine is a key neurotransmitter in theorized pathology of the disorder (Polanczyk et al., 2010). Oxidative stress also leads to disruption of the blood brain barrier, which can increase permeability and lead to increased neuroinflammation and neuronal damage (Song et al., 2020). Despite the recent progress in understanding nervous system OS, the specific mechanisms regulating OS and antioxidant interaction in ADHD are not well understood.
A recent placebo-controlled double-blind RCT of broad-spectrum multinutrient supplementation for treatment of ADHD with emotional dysregulation in children ages 6 to 12 years (Micronutrients for ADHD in Youth [MADDY] Study) demonstrated that three times as many children in the multinutrient group had behavioral improvements according to clinician global rating (also described as “treatment response”) compared to the placebo group (Johnstone et al., 2022). Further, according to parents’ report about their child’s most significant problems, inattention and internalizing symptoms (anxiety and depression) improved more with multinutrients compared to placebo (Tost et al., 2024). The multinutrient formula contained all vitamins and essential minerals and some amino acids, encompassing many nutrients needed to support antioxidant defenses (i.e., vitamins E and C, riboflavin, Cu, Fe, Mn, Zn, Se, and N-acetyl cysteine) as illustrated in Figure 1. Therefore, the aim of this analysis is to scrutinize oxidative stress as a potential contributing mechanism to treatment response in the MADDY RCT. We hypothesized that OS would decrease after 8 weeks of multinutrient supplementation, and subsequently contribute to treatment efficacy as indicated by mediation or moderation of clinician-rated treatment response in the MADDY RCT.
Figure 1.

Select nutrients contained in multinutrient supplement involved in antioxidant synthesis or activity.
Source. Created in BioRender. Robinette (2025). https://BioRender.com/of8ij7s
Methods
Trial Design and Participants
These data were generated from biological samples collected in the MADDY Study, an 8-week double blind RCT that examined the efficacy of a multi-vitamin/mineral supplement (referred to hereafter as multinutrient) as treatment for ADHD and emotional dysregulation in children aged 6 to 12 years. The intervention contained all essential vitamins and minerals and certain amino acids and antioxidants. Details of intervention dosing, study design, and primary RCT outcomes of behavioral improvements and safety profile are published elsewhere (Johnstone et al., 2019, 2022; Leung et al., 2023). Pertinent details of the RCT, including inclusion/exclusion criteria, clinical trial registration, review board approvals, original RCT sample size considerations, and intervention formula and placebo contents are summarized in Supplemental Materials.
Sample Size
One-hundred twenty-three participants completed the RCT (Johnstone et al., 2022). Participants at the two U.S. sites were required by FDA to provide fasted blood and urine samples for contraindications and adverse event monitoring and additional quantities of blood were collected for future mechanistic research (Johnstone et al., 2022). Blood was separated into plasma using standard methods and aliquoted and stored at −80°C. Frozen plasma samples were available for baseline and week 8 visits for 77 participants (45 multinutrient and 32 placebo) from the two U.S. sites (Figure 2). Missing samples were primarily due to insufficient quantity of blood drawn during the visit to freeze for future research.
Figure 2.

CONSORT flow diagram for MADDY RCT updated to include biological samples used in these analyses. Adapted from Johnstone et al. (2022).
Measures
Sociodemographic and Anthropometric Measures
Sociodemographic information was collected including race, ethnicity, parent/guardians’ level of education, and family income at the baseline visit (Johnstone et al., 2022). At each visit, body mass index (BMI) was also calculated based on height, measured using a stadiometer with adjustable headpiece, and weight, measured using a calibrated digital scale (Johnstone et al., 2022).
Behavioral Measures
An a priori defined primary outcome, Clinical Global Impression-Improvement (CGI-I), assessed improvement in global functioning and behaviors (i.e., treatment response) compared to baseline (rated from 1 to 7, with 1 = very much improved to 7 = very much worse, with 4 = no change; Guy, 1976; Johnstone et al., 2022). CGI-I rating at week 8 was dichotomized into “treatment responders” (rating of 1 or 2 [“very much improved” or “much improved”]) or “treatment non-responders” (rating of 3 [somewhat improved] to 7 [very much worse]; Johnstone et al., 2022).
Oxidative Stress (OS) Biomarkers
OS was assessed using five biomarkers of antioxidants and oxidative stress (GR and GPx enzyme activities, biological antioxidant potential [BAP], reactive oxygen metabolites [ROM], and oxidative stress index [OSI], referred to collectively hereafter as “OS biomarkers” for brevity). These biomarkers were chosen based on the expected effects of specific ingredients in the multinutrient formulation, which were well documented in other pathophysiological contexts. GPx enzyme activity was selected since Se was included in the formula and is a required co-factor for the GPx enzyme (Lee et al., 2020; Minich & Brown, 2019). Similarly, thiamine’s, riboflavin’s and niacin’s role in NADPH serving as a cofactor for GR was a basis for measurement of GR enzyme activity for the assessment of these components of supplement (Lee et al., 2020; Minich & Brown, 2019; Pinto & Zempleni, 2016). BAP was chosen as a measure of overall antioxidant capacity since the multinutrient formula contains multiple antioxidants such as vitamins C and E plus N-acetyl cysteine which may support increased glutathione synthesis (Minich & Brown, 2019). Finally, ROM was selected as an overall measure of ROS activity due to the inclusion of essential minerals Fe, Cu, Mn, and Zn in the formulation, which play multiple roles serving as co-factors for antioxidant enzyme or participating in ROS production (Lee et al., 2020).
All assays were performed on baseline and week 8 plasma samples and measurements of absorbance at appropriate wavelengths were conducted using Biotek Synergy H1 Microplate Reader (Agilent, Santa Clara, CA). All standards and samples were run in duplicate.
Biological Antioxidant Potential (BAP) was measured using Diacron BAP kit (Diacron International, Grosseto, Italy) using manufacturer’s instructions for use with microplates (except for incubation temperature, which was increased from the recommend 37°C to 40°C to allow sufficient heating of all wells on microplate for the reaction to complete; Diacron International, 2023a, 2023c). BAP is a measure of the total non-enzymatic antioxidant capacity of blood plasma. It works on the basis that a complex of ferric ions (Fe3+, reagent R2) and a chromogen (reagent R1) create a colored working solution that decolors when the ferric ions are reduced to ferrous ions (Fe2+). First, a working solution of R1 and R2 was made according to a ratio of 1:20 and allowed to equilibrate to room temperature. Each sample well was prepared by adding 2 μL of sample (or standard solutions) and 200 μL of working solution (R1 + R2) then incubated at 40°C for 5 min. Absorbance was then measured at 505 nm. The intensity of the decolorization was directly proportional to the ability of the plasma to reduce the ferric ions. BAP is reported in units of μmol/L.
Reactive oxygen metabolites (ROM) were measured using Diacron ROM (sometimes referred to as d-ROM) kit (Diacron International, Grosseto, Italy) according to manufacturer’s instructions for use with microplates (Diacron International, 2023b, 2023d). It is a measure of the amount of ROM (hydroperoxides, a stable product of the lipid peroxidation process) in a sample using an iron-mediated reaction. In a biological sample, hydroperoxides, in iron’s presence (released from plasma proteins by an acidic buffer—the R2 reagent), can generate alkoxyl and peroxyl radicals according to the Fenton reaction. These radicals can oxidize an alkyl substituted aromatic amine (N,N-diethyl-p-phenylenediamine solubilized in a chromogenic mixture—the R1 reagent) to form a pink colored derivative, which can be photometrically quantified. First, a mastermix of R1 and R2 was made according to a ratio of 1:100 and allowed to equilibrate to room temperature. Each sample well was prepared by adding 1 μL of sample (or standard solutions) and 200 μL of mastermix (R1 + R2) then incubated at 37°C for 90 min. Absorbance was then measured at 505 nm. The color’s intensity was directly proportional to the concentration of hydroperoxides (ROMs) in the sample. ROM is reported in units of Carratelli Units (UCARR), with 1 UCARR equivalent to 0.08 mg H2O2/dL.
Oxidative stress index (OSI) was calculated as the ratio of ROM/BAP.
Glutathione reductase (GR) enzyme activity was measured using Cayman Chemicals Glutathione Reductase Assay Kit (Cayman Chemical Company, Ann Arbor, MI, Cat. #: 703202) according to manufacturer’s instructions (Cayman Chemical Company, 2016). Briefly, this kit measures the rate of nicotinamide adenine dinucleotide phosphate (NADPH) oxidation, accompanied by a decrease in absorbance at 340 nm that is directly proportional to the GR activity in the sample. Before the assay, the following reagents were prepared: Assay Buffer, Sample Buffer, GR control, GSSG, and NADPH. Plasma was diluted 1:1 with sample buffer before assaying. The sample wells were prepared with 100 μL of Assay Buffer, 20 μL of GSSG, and 20 μL of sample. Reactions were initiated by adding 50 μL of NADPH to all the wells. Absorbance was monitored at 340 nm once every minute to obtain at least five time points. Activity was quantified as nmol/min/mL.
Glutathione Peroxidase (GPx) enzyme activity was measured using Cayman Chemicals Glutathione Reductase Assay Kit (Cayman Chemical Company, Ann Arbor, MI, Cat. #: 703102) according to manufacturer’s instructions (Cayman Chemical Company, 2019). This kit measured GPx activity indirectly by a coupled reaction with GR, in which GSSG that is produced upon reduction of hydroperoxide by GPx is recycled back to its reduced state by GR and NADPH. The oxidation of NADPH to NADP+ is accompanied by a decrease in absorbance at 340 nm. Before the assay, the following reagents were prepared: assay buffer, sample buffer, GPx control, GPx co-substrate mixture (containing lyophilized glutathione and GR), NADPH, and cumene hydroperoxide. Plasma was diluted 1:1 with sample buffer before assaying. The sample wells were prepared with 50 μL of Assay Buffer, 50 μL of co-substrate mixture, 50 μL NADPH, and 20 μL of sample. The reaction was initiated by adding 20 μL of Cumene Hydroperoxide to all the wells. Absorbance was measured at 340 nm once every minute to obtain at least five timepoints. Activity was quantified as nmol/min/mL.
Hemoglobin was measured using a Hemoglobin Assay Kit (Cayman Chemical Company, Ann Arbor, MI, Cat. #: 700540; Cayman Chemical Company, 2020). In the presence of heme, the detection reagent selectively reacts with the heme molecule to create a product which absorbs between 560 and 590 nm. The intensity of the color is directly proportional to the total concentration of hemoglobin in the sample, based on the 4:1 molar ratio of heme to hemoglobin. Before the assay, the following reagents were prepared: hemoglobin sample buffer, hemoglobin detector, and hemoglobin standard. Eight concentrations of hemoglobin standard were prepared, and 200 μL of each standard was added to each standard well. The sample wells were prepared with 20 μL of undiluted plasma and 180 μL of hemoglobin detector. The plate was covered and incubated at room temperature for 15 min, then the plate cover was removed, and absorbance read at 560 nm. The absorbance values of the hemoglobin standards were used to create a standard curve and the linear regression equation was used to calculate the hemoglobin concentrations in each sample. Hemoglobin was reported in g/dL (Kristal et al., 2014).
Statistical Methods
Continuous baseline characteristics of age, BMI, and OS biomarker levels were reported as mean and standard deviation (SD) or median and interquartile range (IQR) for normally distributed data and non-parametric data, respectively. For OS biomarkers, any measured value that was below detection limits was coded as ½ of the detection limit for data analysis (Handelsman & Ly, 2019). Categorical baseline characteristics of sex, race, ethnicity, and family income level were reported as frequency (percent). Characteristics were calculated overall and compared between intervention arms using two-sample t-tests with equal variances (or unequal variance, if ratio of standard deviations between intervention groups was >2) for continuous variables and the Pearson chi squared or Fisher’s exact test (if expected counts <5) for categorical variables. To evaluate for potential confounding factors, site, age, sex, and BMI were evaluated for associations with each OS biomarker using Pearson correlation coefficients and two-sample t-test for normal data and Spearman correlation coefficients and Mann-Whitney U-test for non-normal data.
For each OS biomarker, change after the 8-week RCT was calculated for each participant using the value at week 8 minus the value at baseline (referred to as “8-week change”). Two-sample t-test (for BAP, ROM, OSI, and GPx) or Mann Whitney U-test (for GR) were used to determine if the 8-week change for each OS biomarker was different between multinutrient and placebo group. To assess for potential effects of hemolyzed plasma samples, a sensitivity analysis was conducted by excluding samples that had a value of hemoglobin above the detection limit of 0.005 g/dL (Thomas, 2002).
Based on analytical framework for mediation/moderation in an RCT recommended by Kraemer et al. (2002), logistic regression models were used to estimate the odds ratio (OR) and 95% confidence interval (CI). Independent variables included intervention arm (multinutrient or placebo group, I), the potential mediator/moderator (M), and the multiplicative interaction variable (IxM; Kraemer et al., 2002) . The outcome measure was behavioral improvements represented by the CGI-I at week 8 as a dichotomous variable (responder vs. non-responder). Covariates included site, sex, age, BMI, and baseline ROM (for ROM change models only).
Potential moderators were baseline OS values and potential mediators were any OS biomarker with significant difference in the 8-week change between multinutrient and placebo group. To assess for potential effects of hemolyzed plasma samples, sensitivity analyses included baseline and/or week 8 hemoglobin concentration as covariate(s) in the regression models. Significant interaction terms were probed to determine the direction of moderation/mediation.
Note on power calculations:
Power calculations are commonly used before a study begins to determine the probability of detecting an effect, if one is present. Determining the probability of an event after that event has occurred is not logical, and for this reason, conducting post hoc power analyses is discouraged by statisticians (Jiroutek & Turner, 2018; Lakens, 2014). A previous 3-month RCT of a different multinutrient supplement in children with ASD detected a medium effect size (Cohen’s d = 0.52) difference between treatment and placebo groups in changes in an oxidative stress measure (ratio of oxidized to reduced glutathione) in 56 children after 3 months (Adams et al., 2011). In that study, the difference in means was 0.026 and standard deviation within a group was 0.05 (Adams et al., 2011). Compared with our sample size of 77 children, we expected to be able to detect similar effect sizes. The available sample size is not expected to yield sufficient statistical power to examine interaction effects of moderator or mediator analyses, therefore we regard these analyses as exploratory.
Given the hypothesis-generating nature of mediation and moderation analyses, we did not correct for multiple testing. Statistical significance was defined as a two-sided p-value <.05. All analyses were conducted using Stata version 18 software (College Station, TX: StataCorp LLC; StataCorp, 2023).
Results
Study Population Characteristics
Table 1 reports the biometric, sociodemographic, and baseline OS characteristics of this subset of MADDY study participants with plasma samples at baseline and week 8 (n = 77). This sample had a median age of 10.2 years (IQR = 8.6–11.1), was 70% male, and predominantly White (83%) and non-Hispanic (67%). There were no significant differences in sociodemographic and biometric characteristics between the multinutrient and placebo groups in this subset of participants, except sex, in which the placebo group had a higher proportion of females compared to the multinutrient group (43.7% vs. 20.0%, p = .025; Table 1).
Table 1.
Characteristics of the Study Population Comparing Multinutrient and Placebo Groups.
| Characteristics | Total (n = 77) | Multinutrient (n = 45) | Placebo (n = 32) | p-valueb |
|---|---|---|---|---|
|
| ||||
| Child’s age (years), median (IQR) | 10.3 (8.8–11.1) | 10.4 (8.6–11.3) | 10.1 (8.9–11.0) | .400 |
| BMI (kg/m2), median (IQR) | 16.6 (15.2–19.1) | 16.8 (15.5–18.9) | 16.1 (14.5–19.3) | .306 |
| Child’s sex, n (%) | .025 | |||
| Male | 54 (70.1) | 36 (80.0) | 18 (56.2) | |
| Female | 23 (29.9) | 9 (20.0) | 14 (43.7) | |
| Family income, annually, n (%) | .607 | |||
| <$30,000 | 6 (7.8) | 4 (8.9) | 2 (6.2) | |
| $30,001–60,000 | 12 (15.6) | 6 (13.3) | 6 (18.7) | |
| $60,001–80,000 | 9(11.7) | 7 (15.6) | 2 (6.2) | |
| >$80,001 | 50 (65.0) | 28 (62.2) | 22 (68.8) | |
| Parent marital status, n (%) | .251 | |||
| Married | 58 (75.3) | 32 (71.1) | 26 (81.2) | |
| Divorced | 15 (19.5) | 9 (20.0) | 6 (18.7) | |
| Single | 4 (5.2) | 4 (8.9) | 0 (0.0) | |
| Parent educational level, n (%) | .932 | |||
| High school | 7 (9.1) | 4 (8.9) | 3 (9.4) | |
| Technical college/trade school | 16 (20.8) | 10 (22.2) | 6 (18.7) | |
| University or higher | 54 (70.1) | 31 (68.9) | 23 (71.9) | |
| Ethnicitya, n (%) | .093 | |||
| Not Hispanic or Latino | 52 (67.5) | 27 (60.0) | 25 (78.1) | |
| Hispanic or Latino | 6 (7.8) | 6 (13.3) | 0 (0.0) | |
| Otherd | 2 (2.6) | 1 (2.2) | 1 (3.1) | |
| Racea, n (%) | .343 | |||
| Asian | 3 (3.9) | 3 (6.7) | 0 (0.0) | |
| Black | 7 (9.01) | 3 (6.7) | 4 (12.5) | |
| White | 64 (83.1) | 36 (80.0) | 28 (87.5) | |
| Othere | 1 (1.3) | 1 (2.2) | 0 (0) | |
|
| ||||
| Baseline oxidative stress | Total | Multinutrient | Placebo | p-valuec |
|
| ||||
| BAP (μmol/L), mean (SD) | 1,476 (414) | 1,425 (422) | 1,548 (399) | .198 |
| ROM (UCarr), median (IQR) | 375.5 (337.7–425.4) | 401.9 (345.5–426.2) | 364.8 (326.6–398.2) | .070 |
| OSI, median (IQR) | 0.264 (0.216–0.323) | 0.264 (0.218–0.358) | 0.256 (0.201–0.256) | .162 |
| GR (nmol/min/mL), mean (SD) | 17.7 (6.7) | 18.1 (7.0) | 17.2 (6.3) | .530 |
| GPx (nmol/min/mL), median (IQR) | 147.8 (134.3–158.0) | 145.6 (131.9–156.9) | 147.9 (138.9–159.6) | .457 |
Note. Values are presented as medians (IQR) or means (SD) where noted for continuous variables or as frequencies (percentages) for categorical variables. IQR = interquartile range; SD = standard deviation; BMI = Body Mass Index; BAP = biological antioxidant potential; ROM = reactive oxygen metabolites; OSI = oxidative stress index; GR = glutathione reductase; GPx = glutathione peroxidase; UCarr = Carratelli units
17 participants did not report ethnicity and 2 participants did not report race.
p-values comparing multinutrient to placebo groups using two-sample t-tests with equal variances (or unequal variance, if ratio of standard deviations >2) for continuous variables and the Pearson chi squared or Fisher’s exact test (if expected counts <5) for categorical variables.
p-values comparing multinutrient to placebo groups for baseline OS calculated using t-test for parametric and Mann-Whitney U-test for non-parametric variables.
Includes Jewish, Japanese, or other for ethnicity.
Includes American Indian/Native American or Alaska Native, Metis, Native Hawaiian/Pacific Islander, or other for Race.
Characterization of Baseline Oxidative Stress
The baseline values of the OS biomarkers (GR, GPx, BAP, ROM, and OSI) were similar among males and females and participants from the two recruitment sites (Table 2). Baseline ROM was negatively correlated with child’s age (r = −0.24, p = .04; Table 2). The remaining OS biomarkers were not associated with either age or BMI. To explore the potential effect of this different proportion of females between treatment groups, we checked the baseline values of the OS biomarkers and found they were similar among males and females (Table 2).
Table 2.
Relationship Between Baseline Oxidative Stress Measures and Sociodemographic and Biometric Characteristics.
| Baseline oxidative stress values | Sex |
Site |
Child age (years) |
BMI (kg/m2) |
||||
|---|---|---|---|---|---|---|---|---|
| Male (n = 54) | Female (n = 23) | p-value | Oregon (n = 38) | Ohio (n = 39) | p-value | r (p-value) | r (p-value) | |
|
| ||||||||
| BAP (μmol/L), mean (SD) | 1,463 (430.3) | 1,505 (381.7) | .691a | 1,475 (406.3) | 1,476 (427.3) | .985a | −.11 (.35)b | −.12 (.32)b |
| ROM (UCARR), median (IQR) | 393.0 (344.2–431.6) | 363.7 (332.6–410.2) | .142 | 368.7 (330.6–422.4) | 391.4 (337.7–425.4) | .691 | −.24 (.04) | .08 (.48) |
| OSI, median (IQR) | 0.259 (0.210–0.348) | 0.272 (0.214–0.311) | .703 | 0.256 (0.216–0.334) | 0.264(0.214–0.298) | .972 | −.06 (.61) | .13 (.26) |
| GR (nmol/min/mL), mean (SD) | 18.08 (7.12) | 16.92 (5.53) | .489a | 17.28 (5.97) | 18.17 (7.35) | .561a | .11 (.32)b | −.01 (.91)b |
| GPx (nmol/min/mL), median (IQR) | 147.4 (131.9–156.1) | 147.8 (134.6–161.3) | .470 | 152.4 (139.9–162.6) | 141.8 (131.6–152.4) | .064 | −.18 (.13) | −.17 (.14) |
Note. IQ R = interquartile range; SD = standard deviation; BMI = Body Mass Index; BAP = biological antioxidant potential; ROM = reactive oxygen metabolites; OSI = oxidative stress index; GR = glutathione reductase; GPx = glutathione peroxidase; UCarr = Carratelli units.
Two-sample t-test for parametric and Mann-Whitney U-test for non-parametric.
Pearson correlation coefficient; all others used Spearman correlation.
8-Week Change in Oxidative Stress Biomarkers
Between Group Change.
The multinutrient group had a significant decrease in ROM after 8 weeks compared to placebo (between group difference = −41.1 UCARR, SE = 16.8, p = .017). A similar trend was detected in OSI (between group difference = −0.07, SE = 0.04, p = .055) though it did not reach significance (Table 3 and Figure 3). The between group differences in the 8-week change for BAP, GR, and GPx were not significant. To explore the potential effect of this different proportion of females between treatment groups, we analyzed the difference in mean 8-week changes on males and females separately for ROM. Eight-week change between treatment groups continued to be significant among males only (n = 54; between group difference = −43.8 UCARR, SE = 19.7, p = .03) and was insignificant among females only (n = 23) but followed the same general trend of decrease in multinutrient compared to placebo (between group difference = −25.8 UCARR, SE = 36.1, p = 0.48).
Table 3.
Change in Oxidative Stress Measures After 8 Weeks.
| Oxidative stress value | Multinutrient |
Placebo |
Between group difference |
|||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Baseline |
Week 8 |
8-week change |
Baseline |
Week 8 |
8-week change |
|||||
| Mean (SD) | Mean (SD) | Mean (SD) | p-value | Mean (SD) | Mean (SD) | Mean (SD) | p-value | Mean (SE) | p-value | |
|
| ||||||||||
| BAP (μmol/L) | 1,424.6 (422.0) | 1,479.5 (452.5) | 54.9 (299.4) | .225 | 1,548.3 (398.5) | 1,547.1 (532.3) | −1.2 (356.1) | .985 | 56.1 (74.9) | .456 |
| ROM (UCARR) | 399.6 (67.4) | 385.3 (72.4) | −14.3 (68.1) | .165 | 374.8 (69.9) | 401.6 (66.8) | 26.8 (78.5) | .063 | −41.1 (16.8) | .017 |
| OSI | 0.32 (0.17) | 0.30 (0.17) | −0.02 (0.14) | .273 | 0.26 (0.09) | 0.31 (0.19) | 0.05 (0.16) | .089 | −0.07 (0.04) | .055a |
| GR (nmol/min/mL) | 18.14 (6.98) | 16.08 (6.20) | −2.06 (8.31) | .103 | 17.16 (6.28) | 18.05 (6.50) | 0.89 (6.34) | .434 | −2.95 (1.75) | .297a |
| GPx (nmol/min/mL) | 145.48 (24.27) | 145.09 (24.08) | −0.39 (16.49) | .875 | 149.44 (21.86) | 150.42 (21.44) | 0.98 (17.24) | .751 | −1.37 (3.89) | .726 |
Note. SD = standard deviation; BAP = biological antioxidant potential; ROM = reactive oxygen metabolites; OSI = oxidative stress index; GR = glutathione reductase; GPx = glutathione peroxidase; UCarr = Carratelli units.
Mann Whitney U-test, all others t-test.
Figure 3.

Within and between group changes in oxidative stress and antioxidant biomarkers after 8-weeks of multinutrient supplementation. p-values denote significance of between group change as measured by two sample t-test or Mann-Whitney U-test: (a) ROM, BAP, and OSI 8-week changes and (b) GPx and GR 8-week changes.
Within Group Change.
The multinutrient group showed a non-significant increase in BAP and non-significant decreases in ROM, OSI, GR, and GPx (Table 3 and Figure 3). Meanwhile, the placebo group showed non-significant increases in ROM, OSI, GR, and GPx, and non-significant decrease in BAP.
Sensitivity Analysis.
Reanalyzing the data while excluding samples with Hb >0.005 g/dL to account for potential influence of hemolysis did not change the significance or direction of the primary results, except the OSI between group change became significant (between group difference = −0.099, SE = 0.045, p = .048; Supplemental Table 1).
Baseline Oxidative Stress as Moderators and Predictors of Treatment Response
The baseline OS biomarkers were not significant moderators or independent predictors of treatment response (Table 4). Both BAP and OSI trended as independent predictors of treatment response (BAP: OR = 0.998, 95% CI [0.996, 1.000], z = −1.8, p = .072; OSI: OR = 320.1, 95% CI [0.7, 140,000], z = 1.86, p = .063). In the sensitivity analysis, including baseline hemoglobin concentration as a covariate did not change the significance or direction of the findings (Supplemental Table 2).
Table 4.
Baseline OS Measurements as Moderators and Predictors of Treatment Response.
| Primary analysisa | ||||||
|---|---|---|---|---|---|---|
| Independent predictor |
Moderator |
|||||
| Baseline OS value |
Baseline OS × treatment group interaction |
|||||
| OR [95% CI] | z | p-value | OR [95% CI] | z | p-value | |
|
| ||||||
| BAP (μmol/L) | 0.998 [0.996, 1.000] | −1.8 | .072 | 1.003 [0.999, 1.008] | 1.58 | .114 |
| ROM (UCARR) | 1.002 [0.993, 1.011] | 0.43 | .667 | 1.007 [0.989, 1.025] | 0.75 | .453 |
| OSI | 320.1 [0.7, 139,925.7] | 1.86 | .063 | 0.00017 [0.000, 25.629] | −1.43 | .153 |
| GR (nmol/min/mL) | 1.059 [0.962, 1.167] | 1.17 | .24 | 1.033 [0.848, 1.260] | 0.33 | .745 |
| GPx (nmol/min/mL) | 1.010 [0.984, 1.037] | 0.77 | .439 | 0.982 [0.933, 1.033] | −0.72 | .473 |
Note. SD = standard deviation; BAP = biological antioxidant potential; ROM = reactive oxygen metabolites; OSI = oxidative stress index; GR = glutathione reductase; GPx = glutathione peroxidase; UCarr = Carratelli units.
Logistic regression model with child’s age, sex, BMI, and study site as covariates.
8-Week Change in Oxidative Stress as Mediators of Treatment Response
ROM and OSI had a significant 8-week change in the primary or sensitivity analysis, respectively; therefore, both were evaluated as potential mediators. While neither were significant mediators, both trended toward mediation (ROM: OR = 0.985, 95% CI [0.968, 1.002], z = −1.72, p = .086; OSI: OR = 0.00058, 95% CI [0.000, 2.30], z = −1.76, p = .078; Table 5). The direction of this mediation trend in the multinutrient group was the ROM change negatively associated with the probability of being a treatment responder (i.e., children whose ROM decreased after 8 weeks were more likely to be treatment responders), with an opposite trend in the placebo group (Figure 4(a)). The mediation trends for OSI were similar (Figure 4(b)). In the sensitivity analysis, including hemoglobin concentrations as covariates did not change the findings appreciably (Supplemental Table 3).
Table 5.
Eight-week Change in OS Biomarkers as Mediators of Treatment Response.
| Primary analysisa | |||
|---|---|---|---|
| 8-week change x treatment group interaction |
|||
| OR [95% CI] | z | p-value | |
|
| |||
| ROM (UCARR) | 0.985 [0.968, 1.002] | −1.72 | .086 |
| OSI | 0.00058 [0.000, 2.30] | −1.76 | .078 |
Note. ROM = reactive oxygen metabolites; OSI = oxidative stress index; UCarr = Carratelli units; OR = odds ratio; CI = confidence interval.
Logistic regression model with child’s age, sex, BMI, study site, and baseline ROM as covariates.
Figure 4.

Direction of the mediation trend for being a treatment responder: (a) The relationship of 8-week change in ROM and probability of treatment response by treatment group and (b) The relationship of 8-week change in OSI and probability of treatment response by treatment group.
Discussion
In this study, five biomarkers of oxidative stress were evaluated in an 8-week RCT of multinutrient supplementation for children with ADHD. It was hypothesized that the multinutrient formulation which included many nutrients that support antioxidant defenses (i.e., vitamins E and C, riboflavin, Cu, Fe, Mn, Zn, Se, and N-acetyl cysteine) would lead to improved oxidative stress after 8 weeks. The primary finding was that 8 weeks of multinutrients significantly decreased ROM and OSI compared to placebo. To probe the robustness of this finding in view of the unequal proportion of females between treatment groups, the 8-week change in ROM was further analyzed by each sex. In both subsets of male and female participants, ROM decreased more in the multinutrient group compared to placebo, although it only reached statistical significance for males. The lack of statistical significance in females may be due to lack of power attributed to the small sample size of females (n = 23). However, there was no significant change in antioxidant activity or capacity as measured by GPx, GR, and BAP between the two treatment groups. These findings suggest that the multinutrient supplement reduced the ROS burden possibly through decreased production or increased metabolism of ROS metabolites.
Counterintuitive to the assumption that antioxidant measures (GPx, GR, and BAP) would increase in the supplemented group, our study did not identify any significant group differences between supplement and placebo. Previous studies in adults showed that Se supplementation was linked to higher levels of antioxidants such as GSH and total antioxidant capacity (TAC), and lower levels of malondialdehyde (MDA; Zakeri et al., 2021). Additionally, Se supplementation has been shown to increase GPx activity (Alfthan et al., 2000; Neve, 1995; Wilke et al., 1992). A meta-analysis showed that Zn supplementation can elevate GSH, TAC, and SOD levels, but did not affect GPx (Faghfouri et al., 2021). Dose-dependent increases in erythrocyte GPx and GR activities, but no change in serum BAP and ROM, were observed in a study of 60 days of a low-dose (1–2× RDI) multinutrient supplement in older adults (E. Jansen, Beekhof, Tamosiunas, et al., 2015), though the difference in nutrient dosages and age group make these findings difficult to compare to our study. There is also some evidence that basal erythrocyte GR activity reflects changes in riboflavin intake in both lactating women and children/adolescents (Hoey et al., 2009).
This study adds to four previous studies that have examined changes in oxidative stress from other supplemental interventions for children with ADHD—three studies examining the use of pycnogenol (pine bark extract, an antioxidant), and one study examining omega-3 fatty acid supplementation (Dvoráková et al., 2006; Hariri et al., 2012; Hsu et al., 2021; Weyns et al., 2022). While all four studies reported improvement in some measures of ADHD symptoms for the treatment group compared to placebo, the secondary analyses of changes in oxidative stress measures were less consistent. Two of the three PBE studies reported evidence of PBE effect on oxidative stress through increased GSH, increased GSH/GSSG, and lower levels of oxidative damage (8-oxoguanine and thiobarbituric acid reactive substances after 1 month of pycnogenol compared to placebo; Chovanová et al., 2006; Dvoráková et al., 2006; Hsu et al., 2021). This contrasts with a third study that was a 3-way comparison of 10 weeks of pycnogenol, methylphenidate, and placebo, which found no treatment effect on 11 biomarkers of antioxidants or oxidative damage between the three groups (Weyns et al., 2022). Meanwhile, supplementation with omega-3 showed increased activity in GR and SOD antioxidant enzyme activity, but not CAT, in the omega-3 group compared to placebo (Hariri et al., 2012).
This is the first analysis exploring oxidative stress biomarkers as potential moderators and mediators of treatment response to multinutrient supplementation for ADHD, therefore comparison of these results with previously published moderator/mediator analyses is not possible (Rucklidge et al., 2014, 2019, 2021). These potential mechanisms of multivitamin/mineral action need to be studied in future research and were beyond the scope of this clinical intervention.
This study adds to a growing body of literature indicating the potential value of ROM as a biomarker in children with neurodevelopmental disorders. One study of children with autism spectrum disorder (ASD) and co-occurring ADHD found that ROM significantly decreased after drug therapy for ASD (e.g., guanfacine hydrochloride; Kitaoka et al., 2020), plus ROM was found to be significantly higher in children with ASD (Morimoto et al., 2020) and children with Down Syndrome (Carratelli et al., 2001) compared to neurotypical children. ROM has also been shown to be correlated with ASD severity (Morimoto et al., 2020), in addition to being correlated with inflammatory markers in children with ASD (Morimoto et al., 2023). Elevated levels of a similar measure, which also assesses iron-mediated reactions of hydroperoxides (total oxidant status [TOS]; RelAssay, Gaziantep, Turkey), has consistently been reported in children with ADHD compared to healthy controls (Guney et al., 2015; Kul et al., 2015; Miniksar et al., 2023; Sezen et al., 2016). Additionally, the difference in TOS was neutralized after 12 weeks of treatment with methylphenidate (Guney et al., 2015). Our study demonstrated a negative correlation between ROM and age in children with ADHD. This is consistent with a published study of ROM and BAP in healthy children aged 2 to 15 years, which found that ROM levels decreased as age increased, while BAP was not related to age (Morimoto et al., 2019). This finding suggests that it is critical to include age as a factor when measuring ROM as a biomarker in children and perhaps developing age-specific reference values. Further study is needed to understand why ROM would decrease as a child matures.
Strengths and Limitations
A strength of this study is the novel evaluation of oxidative stress in relation to multinutrient supplementation and antioxidant status in children recruited at different clinical and geographical sites to offset the variability related to sample collection. Second is the use of four independent oxidative stress biomarkers, and a calculated oxidative stress index, to capture the balance between pro-oxidants and antioxidants that influence oxidative stress. Finally, the use of the placebo control condition in the mediation and moderation analyses directly attributes changes to the intervention.
There are several limitations to this analysis. One limiatation is the age of the biological samples and stability of oxidative stress biomarkers. This was mitigated through storing the samples frozen at −80°C, and careful selection of biomarkers that have evidence of stability over longer storage times at −80°C (Abiaka et al., 2000; E. Jansen, Beekhof, Viezeliene, et al., 2015; E. H. J. M. Jansen et al., 2013, 2017). The different proportion of females in the treatment groups, in this subset of participants with biological samples, could potentially introduce bias; the impact has been minimized by confirming that baseline value of the OS biomarkers did not differ by gender, and gender was added as a covariate in regression models. It is unknown whether plasma levels of these oxidative stress biomarkers are representative of processes in the brain. The measurement of GPx activity in plasma is susceptible to influence from hemolysis due to the much larger concentration of GPx in erythrocytes compared to plasma. This was accounted for by measuring hemoglobin and performing sensitivity analyses, though assessment of GPx in erythrocytes may be appropriate for future studies. The relatively small sample size of participants with biological samples available may not have yielded sufficient statistical power to examine moderator- or mediator-defined subgroups, therefore we regard these analyses as exploratory. Although it did not reach significance, the difference in baseline ROM between treatment groups could contribute to the between group change; to account for this where possible, baseline ROM was added as a covariate in all regression models contained ROM 8-week change. Finally, since moderation/mediation analysis of an RCT is a hypothesis generating activity, this analysis did not correct for multiple testing, meaning that some findings could be false positives because of Type 1 error. Larger number of RCTs accounting for these limitations could provide further insight for the efficacy of multivitamins in combating OS in ADHD. Additional studies including a larger cohort of participants and testing dose dependencies will be required to establish ROM as a valid biomarker of treatment response.
Conclusions
Eight weeks of multinutrient supplementation in children with ADHD decreased reactive oxygen metabolites suggesting that the intervention reduced production of ROS or more promptly metabolized them than in the placebo group. These findings suggest that reduced oxidative stress is one potential mechanism of action for treatment response to multinutrients that should be further explored.
Supplementary Material
Supplemental material for this article is available online.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was supported by the Virginia Vivian Research Award from the Ohio State University Department of Human Sciences and the Alumni Grants for Graduate Research and Scholarship from The Ohio State University Graduate School. The MADDY study was funded through private donations to the Nutrition and Mental Health Research Fund, managed by the Foundation for Excellence in Mental Health Care (FEMHC), plus a direct grant from FEMHC, and from the Gratis Foundation. None of the private donors provided any input on the study or its reporting. Authors also received support from the National Institutes of Health (NIH) National Center for Complementary and Integrative Health (NCCIH) K23 AT012068 (Johnstone) R90AT008924 to the National University for Natural Medicine (Johnstone), NIH-NCCIH T32 AT002688 to OHSU (Ast, Bruton, Johnstone); the National Center for Advancing Translational Sciences of the NIH, UL1TR002369; UL1TR002733 at OHSU and Ohio State University; OHSU’s Department of Child and Adolescent Psychiatry; the Department of Psychiatry and Behavioral Health as well as the Department of Human Sciences at Ohio State University. In Canada, funding was received through the Nutrition and Mental Health Fund, administered by the Calgary Foundation. The study funders had no role in study design; collection, analysis, and interpretation of data; writing of the report; or restrictions regarding the submission of the report for publication.
Biographies
Lisa M. Robinette, PhD, is a Postdoctoral Scientist at the Abigail Wexner Research Institute at Nationwide Children’s Hospital (Columbus, Ohio, USA). She earned her PhD in Nutrition from The Ohio State University (Columbus, Ohio, USA). Her research focuses on nutrition and neurodevelopment.
Irene E. Hatsu, PhD, RD is an Associate Professor of Human Nutrition and an Extension State Specialist at the Ohio State University (Columbus, Ohio, USA). Her research focuses on the nexus between food and nutrition security and chronic physical and mental diseases in vulnerable and underserved populations.
Olorunfemi Adetona is an Associate Professor of Environmental Health Science in the College of Public Health at The Ohio State University (Columbus, Ohio).
Chieh-Ming Wu is an epidemiologist in occupational health and tobacco research. His overall interest is in investigating potential health effects of biomass smoke exposure using different epidemiological approaches.
Jeanette M. Johnstone, PhD, MA, MFT is an Associate Professor in the Center for Mental Health Innovation at Oregon Health & Science University, Portland, OR. She earned her PhD at the University of Otago School of Medicine in Christchurch, New Zealand.
Alisha M. Bruton ND, MS, MS is a research scientist and biostatistician at the Center for Mental Health Innovation at Oregon Health & Science University in Portland, OR. Her research focuses on nutrition and emotional well-being in people with ADHD.
Hayleigh K. Ast, ND is a Postdoctoral Scientist at the Center for Mental Health and Innovation in the Department of Psychiatry at Oregon Health & Science University (Portland, OR, USA). She earned her naturopathic doctoral degree from the National University of Natural Medicine (Portland, OR, USA) and completed her residency training at the University of California, Irvine, Susan Samueli Integrative Health Institute (Irvine, CA, USA).
James B. Odei, PhD is a Clinical Associate Professor of Biostatistics at the Ohio State University (Columbus, Ohio, USA). His research interests include spatial, temporal, spatio-temporal modeling using Bayesian methods, and hierarchical modeling and applications to ecology and health.
Ouliana Ziouzenkova, PhD, attended University of Kyiv, Ukraine, and her PhD in Biochemistry from the University of Graz, Austria, followed by postdoctoral training at the University of Southern California and molecular biology training at Brigham and Women’s Hospital, Harvard University. Dr. Ziouzenkova’s research focuses on nutrigenomics and vitamin A metabolism, particularly the neuroendocrine role of retinaldehyde in regulating metabolic pathways. Her work has led to the discovery of mechanisms by which vitamin A metabolites reprogram fat storage and promote thermogenic adipocytes, resulting in improved glucose metabolism and neuroprotection. She has developed a novel nanotechnology-based therapy that exceeds the therapeutic potential of insulin in the brain, targeting diabetes and neurodegenerative complications. She serves as Associate Editor for BMC Endocrine Disorders and PLOS One. Dr. Ziouzenkova has authored 61 publications, including book chapters, and holds 5 patents. Her Google Scholar h-index is 31 and i10-index is 53.
Brenda M.Y. Leung, PhD, is an Associate Professor, Emmy Droog Chair in Complementary and Alternative Healthcare, Faculty of Health Sciences, University of Lethbridge, Lethbridge, AB Canada.
L. Eugene Arnold, MD, MEd (deceased), was an emeritus professor at Ohio State University where he previously served as the director of the division of child psychiatry and vice-chair of the department of psychiatry. His extensive research legacy includes involvement in the NIMH Multimodal Treatment Study of Children with ADHD, and in studies of alternative and complementary treatments for neurodevelopmental disorders.
Footnotes
Declaration of Conflicting Interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: The multinutrient and placebo capsules were provided by Hardy Nutritionals at no charge, but had no role in study design, data analysis, or interpretation of results. Dr. Arnold has received research funding from Axial, Forest, Lilly, Noven, Otsuka, Shire, Supernus, Roche, Yamo, Maplight, and YoungLiving (as well as NIH and Autism Speaks); has consulted with Pfizer, Tris Pharma, and Waypoint; and been on advisory boards for Arbor, Ironshore, Otsuka, Pfizer, Roche, Seaside Therapeutics, and Shire. The other authors declare that the research was conducted without any conflict of interest.
Data Availability Statement
Data described in the manuscript will be made available upon request pending study team review and approval.
References
- Abiaka C, Al-Awadi F, & Olusi S (2000). Effect of prolonged storage on the activities of superoxide dismutase, glutathione reductase, and glutathione peroxidase. Clinical Chemistry, 46(4), 566–567. [PubMed] [Google Scholar]
- Adams JB, Audhya T, McDonough-Means S, Rubin RA, Quig D, Geis E, Gehn E, Loresto M, Mitchell J, Atwood S, & Barnhouse S (2011). Effect of a vitamin/mineral supplement on children and adults with autism. BMC Pediatrics, 11, 1–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Alfthan G, Xu GL, Tan WH, Aro A, Wu J, Yang YX, Liang WS, Xue WL, & Kong LH (2000). Selenium supplementation of children in a selenium-deficient area in China: Blood selenium levels and glutathione peroxidase activities. Biological Trace Element Research, 73, 113–125. [DOI] [PubMed] [Google Scholar]
- Al-Gubory KH (2014). Environmental pollutants and lifestyle factors induce oxidative stress and poor prenatal development. Reproductive BioMedicine, 29(1), 17–31. 10.1016/j.rbmo.2014.03.002 [DOI] [PubMed] [Google Scholar]
- Brown KA, Samuel S, & Patel DR (2018). Pharmacologic management of attention deficit hyperactivity disorder in children and adolescents: A review for practitioners. Translational Pediatrics, 7(1), 36–47. 10.21037/tp.2017.08.02 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Carratelli M, Porcaro L, Ruscica M, De Simone E, Bertelli AA, & Corsi MM (2001). Reactive oxygen metabolites and prooxidant status in children with Down’s syndrome. International Journal of Clinical Pharmacology Research, 21(2), 79–84. [PubMed] [Google Scholar]
- Cayman Chemical Company. (2016). Glutathione reductase assay kit no. 703202 booklet.
- Cayman Chemical Company. (2020). Hemoglobin assay kit no. 700540 booklet.
- Chovanová Z, Muchová J, Sivonová M, Dvoráková M, Zitnanová I, Waczulíková I, Trebatická J, Skodácek I, & Duracková Z (2006). Effect of polyphenolic extract, Pycnogenol, on the level of 8-oxoguanine in children suffering from attention deficit/hyperactivity disorder. Free Radical Research, 40(9), 1003–1010. 10.1080/10715760600824902 [DOI] [PubMed] [Google Scholar]
- Company, C. C. (2019). Glutathione peroxidase assay kit no. 703102 booklet. https://cdn.caymanchem.com/cdn/insert/703102.pdf
- Danielson ML, Bitsko RH, Ghandour RM, Holbrook JR, Kogan MD, & Blumberg SJ (2018). Prevalence of parent-reported ADHD diagnosis and associated treatment among US children and adolescents, 2016. Journal of Clinical Child & Adolescent Psychology, 47(2), 199–212. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Diacron International. (2023a). BAP test: Colorimetric determination of biological antioxidant potential in plasma and serum (rev 5.03). [Google Scholar]
- Diacron International. (2023b). d-ROMs test: Colorimetric determination of reactive oxygen metabolites (free radical compounds) in plasma and serum (rev 7.02). [Google Scholar]
- Diacron International. (2023c). Microplates procedure: BAP test.
- Diacron International. (2023d). Microplates procedure: d-ROMs test.
- Dvoráková M, Sivonová M, Trebatická J, Skodácek I, Waczuliková I, Muchová J, & Duracková Z (2006). The effect of polyphenolic extract from pine bark, Pycnogenol on the level of glutathione in children suffering from attention deficit hyperactivity disorder (ADHD). Redox Report, 11(4), 163–172. 10.1179/135100006X116664 [DOI] [PubMed] [Google Scholar]
- Faghfouri AH, Zarezadeh M, Aghapour B, Izadi A, Rostamkhani H, Majnouni A, Abu-Zaid A, Kord Varkaneh H, Ghoreishi Z, & Ostadrahimi A (2021). Clinical efficacy of zinc supplementation in improving antioxidant defense system: A comprehensive systematic review and time-response meta-analysis of controlled clinical trials. European Journal of Pharmacology, 907, Article 174243. 10.1016/j.ejphar.2021.174243 [DOI] [PubMed] [Google Scholar]
- Figueira TR, Barros MH, Camargo AA, Castilho RF, Ferreira JC, Kowaltowski AJ, Sluse FE, Souza-Pinto NC, & Vercesi AE (2013). Mitochondria as a source of reactive oxygen and nitrogen species: From molecular mechanisms to human health. Antioxidants & Redox Signaling, 18(16), 2029–2074. 10.1089/ars.2012.4729 [DOI] [PubMed] [Google Scholar]
- Guney E, Cetin FH, Alisik M, Tunca H, Tas Torun Y, Iseri E, Isik Taner Y, Cayci B, & Erel O (2015). Attention Deficit Hyperactivity Disorder and oxidative stress: A short term follow up study. Psychiatry Research, 229(1–2), 310–317. 10.1016/j.psychres.2015.07.003 [DOI] [PubMed] [Google Scholar]
- Guy W (1976). ECDEU assessment manual for psychopharmacology In US Department of Health, Education, and Welfare Publication (ADM) (pp. 76–338). National Institute of Mental Health. [Google Scholar]
- Handelsman DJ, & Ly LP (2019). An accurate substitution method to minimize left censoring bias in serum steroid measurements. Endocrinology, 160(10), 2395–2400. 10.1210/en.2019-00340 [DOI] [PubMed] [Google Scholar]
- Hariri M, Djazayery A, Djalali M, Saedisomeolia A, Rahimi A, & Abdolahian E (2012). Effect of n-3 supplementation on hyperactivity, oxidative stress and inflammatory mediators in children with attention-deficit-hyperactivity disorder. Malaysian Journal of Nutrition, 18(3), 329–335. [PubMed] [Google Scholar]
- Hoey L, McNulty H, & Strain JJ (2009). Studies of biomarker responses to intervention with riboflavin: A systematic review. The American Journal of Clinical Nutrition, 89(6), 1960S–1980S. 10.3945/ajcn.2009.27230B [DOI] [PubMed] [Google Scholar]
- Hsu CD, Hsieh LH, Chen YL, Lin IC, Chen YR, Chen CC, Shirakawa H, & 7 Yang SC. (2021). Complementary effects of pine bark extract supplementation on inattention, impulsivity, and antioxidative status in children with attention-deficit hyperactivity disorder: A double-blinded randomized placebo-controlled cross-over study. Phytotherapy Research, 35(6), 3226–3235. 10.1002/ptr.7036 [DOI] [PubMed] [Google Scholar]
- Jansen E, Beekhof P, Tamosiunas A, Luksiene D, & Baceviciene M (2015). Biomarkers of oxidative stress and redox status in a short-term low-dosed multivitamin and mineral supplementation study in two human age groups. Biogerontology, 16(5), 645–653. 10.1007/s10522-015-9568-x [DOI] [PubMed] [Google Scholar]
- Jansen E, Beekhof P, Viezeliene D, Muzakova V, & Skalicky J (2015). Long-term stability of cancer biomarkers in human serum: Biomarkers of oxidative stress and redox status, homocysteine, CRP and the enzymes ALT and GGT. Biomarkers in Medicine, 9(5), 425–432. 10.2217/bmm.15.14 [DOI] [PubMed] [Google Scholar]
- Jansen EHJM, Beekhof PK, Cremers JW, Viezeliene D, Muzakova V, & Skalicky J (2013). Long-term stability of parameters of antioxidant status in human serum. Free Radical Research, 47(6–7), 535–540. 10.3109/10715762.2013.797969 [DOI] [PubMed] [Google Scholar]
- Jansen EHJM, Beekhof PK, Viezeliene D, Muzakova V, & Skalicky J (2017). Long-term stability of oxidative stress biomarkers in human serum. Free Radical Research, 51(11–12), 970–977. 10.1080/10715762.2017.1398403 [DOI] [PubMed] [Google Scholar]
- Jiroutek MR, & Turner JR (2018). Why it is nonsensical to use retrospective power analyses to conduct a postmortem on your study. The Journal of Clinical Hypertension, 20(2), 408–410. 10.1111/jch.13173 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Johnstone JM, Hatsu I, Tost G, Srikanth P, Eiterman LP, Bruton AM, Ast HK, Robinette LM, Stern MM, Millington EG, Gracious BL, Hughes AJ, Leung BMY, & Arnold LE (2022). Micronutrients for attention-deficit/hyperactivity disorder in youths: A placebo-controlled randomized clinical trial. Journal of the American Academy of Child and Adolescent Psychiatry, 61(5), 647–661. 10.1016/j.jaac.2021.07.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Johnstone JM, Leung B, Gracious B, Perez L, Tost G, Savoy A, Hatsu I, Hughes A, Bruton A, & Arnold LE (2019). Rationale and design of an international randomized placebo-controlled trial of a 36-ingredient micronutrient supplement for children with ADHD and irritable mood: The Micronutrients for ADHD in Youth (MADDY) study. Contemporary Clinical Trials Communications, 16, Article 100478. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Joseph N, Zhang-James Y, Perl A, & Faraone SV (2015). Oxidative stress and ADHD: A meta-analysis. Journal of Attention Disorders, 19(11), 915–924. 10.1177/1087054713510354 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kitaoka T, Morimoto M, Hashimoto T, Tsuda Y, Nakatsu T, & Kyotani S (2020). Evaluation of the efficacy of drug treatment based on measurement of the oxidative stress, using reactive oxygen metabolites and biological antioxidant potential, in children with autism spectrum disorder and attention deficit hyperactivity disorder. Journal of Pharmaceutical Health Care and Sciences, 6, 8. 10.1186/s40780-020-00164-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kraemer HC, Wilson GT, Fairburn CG, & Agras WS (2002). Mediators and moderators of treatment effects in randomized clinical trials. Archives of General Psychiatry, 59(10), 877–883. 10.1001/archpsyc.59.10.877 [DOI] [PubMed] [Google Scholar]
- Kristal AR, Kolar AS, Fisher JL, Plascak JJ, Stumbo PJ, Weiss R, & Paskett ED (2014). Evaluation of webbased, self-administered, graphical food frequency questionnaire. Journal of the Academy of Nutrition and Dietetics, 114(4), 613–621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kul M, Unal F, Kandemir H, Sarkarati B, Kilinc K, & Kandemir SB (2015). Evaluation of oxidative metabolism in child and adolescent patients with attention deficit hyperactivity disorder. Psychiatry Investigation, 12(3), 361–366. 10.4306/pi.2015.12.3.361 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lakens D (2014). The 20% statistician: Observed power, and what to do if your editor asks for post-hoc power analyses. The 20% Statistician. https://daniellakens.blogspot.com/2014/12/observed-power-and-what-to-do-if-your.html [Google Scholar]
- Lange KW, Hauser J, Lange KM, Makulska-Gertruda E, Nakamura Y, Reissmann A, Sakaue Y, Takano T, & Takeuchi Y (2017). The role of nutritional supplements in the treatment of ADHD: What the evidence says. Current Psychiatry Reports, 19(2), 8. 10.1007/s11920-017-0762-1 [DOI] [PubMed] [Google Scholar]
- Lee KH, Cha M, & Lee BH (2020). Neuroprotective effect of antioxidants in the brain. International Journal of Molecular Sciences, 21(19), Article 7152 10.3390/ijms21197152 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Leung BMY, Srikanth P, Robinette L, Bruton AM, Tost G, Hatsu I, Arnold LE, & Johnstone JM (2023). Micronutrients for ADHD in youth (MADDY) study: Comparison of results from RCT and open label extension. European Child & Adolescent Psychiatry, 33(5), 1355–1367. 10.1007/s00787-023-02236-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lopresti AL (2015). Oxidative and nitrosative stress in ADHD: Possible causes and the potential of antioxidant-targeted therapies. Attention Deficit Hyperactivity Disorder, 7(4), 237–247. 10.1007/s12402-015-0170-5 [DOI] [PubMed] [Google Scholar]
- Minich DM, & Brown BI (2019). A review of dietary (phyto) nutrients for glutathione support. Nutrients, 11(9), 2073. 10.3390/nu11092073 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Miniksar DY, Cansız MA, Göçmen AY, Kılıç M, & Miniksar Ö (2023). The effect of drug use, body mass index and blood pressure on oxidative stress levels in children and adolescents with attention deficit and hyperactivity disorder. Clinical Psychopharmacology and Neuroscience, 21(1), 88–98. 10.9758/cpn.2023.21.1.88 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Morimoto M, Hashimoto T, Tsuda Y, Kitaoka T, & Kyotani S (2019). Evaluation of oxidative stress and antioxidant capacity in healthy children. Journal of the Chinese Medical Association, 82(8), 651–654. 10.1097/JCMA.0000000000000045 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Morimoto M, Hashimoto T, Tsuda Y, Nakatsu T, Kitaoka T, & Kyotani S (2020). Assessment of oxidative stress in autism spectrum disorder using reactive oxygen metabolites and biological antioxidant potential. PLoS ONE, 15(5), e0233550. 10.1371/journal.pone.0233550 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Morimoto M, Hashimoto T, Tsuda Y, Suenaga M, Nakamura T, & Katoh S (2023). Study on oxidative stress and inflammatory/antioxidant substance levels in autism spectrum disorder. Journal of the Chinese Medical Association, 86(5), 489–493. 10.1097/JCMA.0000000000000917 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Neve J (1995). Human selenium supplementation as assessed by changes in blood selenium concentration and glutathione peroxidase activity. Journal of Trace Elements in Medicine and Biology, 9(2), 65–73. 10.1016/S0946-672X(11)80013-1 [DOI] [PubMed] [Google Scholar]
- Pinto JT, & Zempleni J (2016). Riboflavin. Advances in Nutrition, 7(5), 973–975. 10.3945/an.116.012716 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Polanczyk G, Bigarella M, Hutz M, & Rohde L (2010). Pharmacogenetic approach for a better drug treatment in children. Current Pharmaceutical Design, 16(22), 2462–2473. 10.2174/138161210791959872 [DOI] [PubMed] [Google Scholar]
- Rucklidge JJ, Eggleston MJF, Boggis A, Darling K, Gorman B, & Frampton CM (2021). Do changes in blood nutrient levels mediate treatment response in children and adults with ADHD consuming a vitamin-mineral supplement? Journal of Attention Disorders, 25(8), 1107–1119. 10.1177/1087054719886363 [DOI] [PubMed] [Google Scholar]
- Rucklidge JJ, Eggleston MJF, Darling KA, Stevens AJ, Kennedy MA, & Frampton CM (2019). Can we predict treatment response in children with ADHD to a vitamin-mineral supplement? An investigation into pre-treatment nutrient serum levels, MTHFR status, clinical correlates and demographic variables. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 89, 181–192. 10.1016/j.pnpbp.2018.09.007 [DOI] [PubMed] [Google Scholar]
- Rucklidge JJ, Johnstone J, Gorman B, Boggis A, & Frampton CM (2014). Moderators of treatment response in adults with ADHD treated with a vitamin-mineral supplement. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 50, 163–171. 10.1016/j.pnpbp.2013.12.014 [DOI] [PubMed] [Google Scholar]
- Sanchez-Rodriguez MA, & Mendoza-Nunez VM (2019). Oxidative stress indexes for diagnosis of health or disease in humans. Oxidative Medicine and Cellular Longevity, 2019, Article 4128152. 10.1155/2019/4128152 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sezen H, Kandemir H, Savik E, Basmacı Kandemir S, Kilicaslan F, Bilinc H, & Aksoy N (2016). Increased oxidative stress in children with attention deficit hyperactivity disorder. Redox Report, 21(6), 248–253. 10.1080/13510002.2015.1116729 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Singh A, Kukreti R, Saso L, & Kukreti S (2019). Oxidative stress: A key modulator in neurodegenerative diseases. Molecules, 24(8), 1583. 10.3390/molecules24081583 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Song K, Li Y, Zhang H, An N, Wei Y, Wang L, Tian C, Yuan M, Sun Y, Xing Y, & Gao Y (2020). Oxidative stress-mediated blood-brain barrier (BBB) disruption in neurological diseases. Oxidative Medicine and Cellular Longevity, 2020, 1–27. 10.1155/2020/4356386 [DOI] [Google Scholar]
- StataCorp. (2023). Stata statistical software: Release 18. StataCorp, LLC. [Google Scholar]
- Thomas L (2002). Haemolysis as influence & interference factor. The Electronic Journal of the International Federation of Clinical Chemistry and Laboratory Medicine, 13(4), 95–98. [PMC free article] [PubMed] [Google Scholar]
- Tost G, Srikanth P, Bruton A, Hatsu IE, Leung BM, Ast HK, Eiterman LP, Robinette LM, Williams C, Gracious B, Eugene Arnold L, & Johnstone JM (2024). Problems most concerning to parents of children with ADHD and emotional dysregulation in a randomized controlled trial of multinutrients: MADDY secondary analysis. European Child & Adolescent Psychiatry, 33(12), 4335–4347. 10.1007/s00787-024-02463-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Verlaet A, Robberecht H, De Bruyne T, & Hina N (2018). Evaluation of biomarkers of oxidative stress in attention-deficit/hyperactivity disorder (ADHD). Journal of Molecular Biomarkers & Diagnosis, 9(390), 10–4172. [Google Scholar]
- Weyns AS, Verlaet AA, Van Herreweghe M, Breynaert A, Fransen E, De Meester I, Logie E, Vanden Berghe W, Verhelst H, Van West D, Van Ingelghem T, Jonckheere AI, Beysen D, Sandra K, Moens E, van Roest APJ, Savelkoul HFJ, De Bruyne T, Pieters L, . . . Van Ingelghem I (2022). Clinical investigation of French maritime pine bark extract on attention-deficit hyperactivity disorder as compared to methylphenidate and placebo: Part 2: Oxidative stress and immunological modulation. Journal of Functional Foods, 97, Article 105247. [Google Scholar]
- Wilke BC, Vidailhet M, Favier A, Guillemin C, Ducros V, Arnaud J, & Richard MJ (1992). Selenium, glutathione peroxidase (GSH-Px) and lipid peroxidation products before and after selenium supplementation. Clinica Chimica Acta, 207, 137–142. [DOI] [PubMed] [Google Scholar]
- Zakeri N, Asbaghi O, Naeini F, Afsharfar M, Mirzadeh E, & kasra Naserizadeh S (2021). Selenium supplementation and oxidative stress: A review. PharmaNutrition, 17, Article 100263. [Google Scholar]
- Zhao X, Page TF, Altszuler AR, Pelham WE, Kipp H, Gnagy EM, Coxe S, Schatz NK, Merrill BM, Macphee FL, & Macphee FL (2019). Family burden of raising a child with ADHD. Journal of Abnormal Child Psychology, 47(8), 1327–1338. 10.1007/s10802-019-00518-5 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data described in the manuscript will be made available upon request pending study team review and approval.
