Table 3.

Summary of relevant results. HVLT, Hopkins Verbal Learning Test; MCV, mean corpuscular volume.

1: Iron
Study	Dependent Variable	Domains	Statistical Analysis	Changes in Iron/Zinc Status	Findings
Bruner et al., 1996 [15]	Cognition	Attention: Assessed using Brief Test of Attention (BTA), Symbol Digit Modalities Test (SDMT) and Visual Search and Attention Test (VSAT). Memory: multicomponent test, Hopkins Verbal Learning Test (HVLT). Learning: multicomponent test, HVLT	Intention-to-treat analysis; multiple linear regression, assessment of post iron treatment cognitive scores; analysis of variance (ANOVA), assessed HVLT learning curve	Iron girls: significantly higher serum ferritin concentration (18.2 (SD 12.6) vs. 3.5 (6.6) µg/L, p < 0.001) and higher mean Hb concentration (135 (8.0) vs. 127 (7.0) g/L, p < 0.001)	No significant effect on attention were found. Significant improvement on total recall scores of HVLT for girls who took iron treatment, compared to control group (p < 0.02). No significant differences were found between delayed recall and recognition. Learning: All participants recalled more words after each trial (F = 273.7), p < 0.001), although no significant differences were found between groups. Despite this, participants who took iron treatment performed better compared to the control group (F = 6.3, p < 0.02).
Beard et al., 2005 [21]	Cognition	Intellectual abilities: Raven’s Coloured Progressive Matrices (RPM). Memory: Digit Symbol Test	Repeated measures ANOVA; Pearson’s correlation, strength of association between variables	IDA iron group: Significant improvement in Hb transferrin saturation (TSAT) and serum ferritin (Ft) values. IDA placebo: Also significant increase in Hb, reflecting natural state of iron status restoration postpartum. Iron sufficient: No change in haematology	Iron treatment effected significant improvement on RPMs and Digit Symbol scores. Iron-treated mother’s scores were almost identical to non-anaemic mothers at 9 months. An association was found between—RPM scores and Hb, 10 weeks—Digit symbol test and MCV at 9 months
	Emotions (postpartum depression)	Postpartum depression: Edinburgh Postnatal Depression Scale (EPDS)	Repeated Measures ANOVA; Pearson’s Correlation		Results of EPDS were unclear. Pearson’s correlation found that scores on the EPDS Hb and MCV.
Murray-Kolb and Beard, 2007 [2]	Cognition	IQ: Shipley Institute of Living Scale (20), Cognitive Abilities test (CAT) measuring 3 domains. Attention: Reaction Time Task. Memory: Reaction Time task (short-term memory (STM)), Probe Recall Task (recognition), Sternberg Memory Search Task (STM, retrieval and executive functioning). Learning: CAT	ANCOVA, assessed differences between groups with IQ as a covariate; repeated-measures ANOVA, assessed change in cognitive variables over time	Iron-treated participants were found to have significantly improved iron status. Significant increases occurred in ID iron treatment and IDA iron treatment groups for ferritin (p < 0.001 and p < 0.01, respectively) and body iron (p < 0.0001 and p < 0.01, respectively) The IDA iron-treated participants also had a significant increase in Hb (p < 0.0001), haematocrit (p < 0.001) and transferrin saturation (p = 0.018).	Cross-sectional baseline comparisons:
					Attention: Composite score significantly lower in IDA group, between iron sufficient and ID (p = 0.008; p = 0.003, respectively). Performance in attention significantly better than IDA group, in control group and iron-deficient groups (p = 0.047 and 0.008, respectively). Memory: Composite scores show that iron-sufficient and ID groups scored equally well and significantly better than IDA (p < 0.001). Performance in this domain showed the same pattern. Learning: Composite score on the learning domain showed that iron-sufficient group scored better than ID group, and there was a significant difference between iron-sufficient and IDA group (p = 0.013) and ID and IDA group (p = 0.042).
					Longitudinal analysis:
					Significant improvement in serum ferritin (n = 66) was related to improvements in the attention and learning and memory domains, 5-/7-times, respectively, in women that were non-responders (n = 66) (p < 0.001). No significant correlation between the size of ferritin change and the cognitive improvement. Women who had significant change in Hb concentration (n = 33) completed the attention and memory tasks significantly faster than Hb non-responders (n = 80) (p < 0.001).
McClung et al., 2009 [16]	Mood and physical performance	General mood: Profile of Mood State (POMS)	Two-way repeated measures ANOVA, treatment and time effects	Participation in 8-wk battle combat training had effects on iron status, as shown by elevation (p < 0.05) of Hb, red blood cell distribution width and soluble transferrin receptor (sTfR) in both placebo and iron-treated groups. Serum ferritin was diminished (p < 0.05) in placebo, but not iron-treated groups.	Group x Time interaction for vigour on the POMS: Iron supplementation had a significant effect on vigour, after Battle CombatTraining (BCT). After stratification by iron status at the beginning of BCT, the positive effects of time on mood stayed significant (p < 0.05) for all indicators, except anger, in the iron-deficient group
Verdon et al., 2003 [24]	Depression, anxiety and fatigue	Depressive and anxiety symptoms: Questionnaire bref d’auto-évaluation des dimensions dépressive, asthénique et anxieuse	Intention-to-treat protocol χ2 and linear regression	Unclear	The iron group showed the largest decrease in cumulative −4.6 (7.5) score for fatigue (−7.5 (8.0)), a difference of 3.0 points, 0.3 to 5.6, p = 0.03. Scores for depression were not significant between groups. Depression was not associated with any markers of iron status
Lambert et al., 2002 [23]	Cognition	Verbal Working Memory, free recall: Hopkins Verbal Learning Test. Attention: Stroop Task. Processing speed and attention: Visual Search. Working memory: Reading Span Task.	Multiple regression analysis	Ferritin level was increased for both placebo (t (58) = 3.73, p < 0.001) and iron group (t (56) = 9.16, p < 0.001). It was found that the extent of increase was larger in the iron group. A significant decrease in Hb concentration was found in the placebo group pre-/post-treatment (t (56) = 4.09, p < 0.001), and there was no change in Hb level for iron groups pre-/post-treatment (t (56) = 1.41, not significant).	Hopkins Verbal Learning Test: For the iron group, significant improvement in recall of words from baseline to post-treatment (t (56) = 2.40, p = 0.01) in the second half list, but no significant relationship between iron treatment and the first half of the list. A significant relationship between Hb change and post-treatment performance on recall for the iron treatment group, second half of the list only (p < 0.004). Reading Span Test: Multiple regression found a significant relationship between change in serum ferritin and reading span (p < 0.01). Stroop Task: No relationship found between performance on this task with changed iron status or supplementation. Visual Search Task. No treatment x testing interaction found.
Devaki et al., 2009 [20]	Cognition and Emotion (mood)	STM: recall a list of 6 digit numbers, after interpolated activity, asked to recall (not as in Wechsler Intelligence Scale (WAIS)). Long-term memory (LTM): recited the stimulus number, presented four times and then asked to recall one h of interpolated activity (not as in WAIS). IQ: Ravens Progressive Matrices (RPM). WAIS. Affective behaviour: Emotional Quotient.	2 × 2 ANOVA with duration as one factor and study group as another	Iron-sufficient, supplement group, ID and IDA groups showed significant increases in Hb, transferrin saturation (TS) and serum ferritin 8 months post-treatment (p < 0.01). Iron-sufficient, placebo group showed no increase in iron status.	Iron supplementation groups showed improvements in mean scores of STM, LTM WAIS and RPM, at four- and eight-month follow up, which also reflected changes in haematological parameters. Changes in cognitive scores were much higher in the iron-deficient and iron-deficient anaemic group, compared to iron-sufficient, supplement group. No significant change in Emotion Quotient (EQ) score was seen for any group.
Vaucher et al., 2012 [22]	Primary Fatigue: Secondary: Mood	Fatigue and depressive symptoms: Current and Past Psychological Scale. Fatigue: Multidimensional Assessment of Fatigue score; self-reported health questionnaire, Global Fatigue Index and Severity.	Intention-to-treat analysis	After 6 weeks of iron treatment, significant effects on markers of iron status were found for iron-supplemented groups. Hb: (3 g/L; p = 0.001). Ferritin: (6.8 μg/L; p = 0.01). MCV: (1.2 fL; p = 0.01). sTfR: (−0.4 mg/L; p < 0.001) transferrin saturation (6.6%; p < 0.001). Similar effects were seen after Week 12.	Patients receiving iron supplement had a 3.5 point improvement (95% CI) in their fatigue score in current and past psychological scale compared to those in placebo group. Iron treatment had a significant effect on the global fatigue index from the Multidimensional Assessment of Fatigue Scale (p = 0.03) and its severity index (p = 0.03). Iron was not found to have a significant effect on anxiety or depression scores in this study.
	2: Zinc
Study	Dependent Variable	Domains	Statistical Analysis	Changes in Iron/Zinc Status	Findings
Siwek et al., 2009 [17]	Depression	Depressive Index scores on Clinical Global Impression (CGI), Montgomery-Åsberg Depression Rating Scale (MADRS), Hamilton Depression Rating Scale (HDRS), Beck Depression Inventory (BDI). Remission of depression was defined by a score of “very much” on CGI, plus scores of ≤7 on HDRS, ≤10 on MADRS or ≤9 on BDI.	Linear model, mixed-design ANOVA, with repeated factor being test number, and between factors being treatment, antidepressant and treatment resistance. A non-treatment-resistant group and a treatment-resistant group formed.	Zinc-supplemented groups had a significant increase in serum zinc levels at Weeks 6 and 12. Non-zinc-supplemented groups saw an increase of zinc at Week 12. Treatment-resistant groups demonstrated lower levels of zinc than treatment- and non-treatment-resistant groups.	A significant negative correlation between zinc levels and MADRS scores at Week 12 was found, when either all (p < 0.0001) or zinc supplemented patients were taken into the analysis
Nowak et al., 2003 [18]	Depression	Depression Index: HDRS and BDI.	Group differences were assessed using t-test and multiple ANOVA with two between-subjects factors (placebo vs. zinc treatment)	Unclear	Scores of the HDRS were significantly reduced over time: F (3, 28) = 5.091; p < 0.001, placebo, and F (3, 20) = 29.578 p < 0.001, zinc group. The group effect (F (1, 48) = 4.275, p = 0.049) and time effect (F (1, 48) = 21.683; p < 0.001) were statistically significant. There was no significant interaction effect. Zinc improved this reduction at Weeks 6 and 12, compared to treatment (ca. 55%). Zinc supplementation significantly improved the reduction in BDI scores at Week 12, compared to placebo (by 40%).
Sawada and Yokoi, 2010 [19]	Depression	Somatic symptoms and mood feelings, like anxiety, sensitivity, anger and tension: Profile of Mood State (POMS) and Cornell Medical Index (CMI).	Data were analysed with Wilcoxon’s signed-rank test	Neither intervention showed a significant change in serum ferritin or Hb concentration. Multivitamin (MV) + zinc significantly increased serum zinc concentration, whereas MV alone did not.	Women who took multivitamins with zinc showed a significant decrease in anger-hostility and depression-dejection scores on the POMS.