nutrients

Systematic Review
Relationship between Iron Deficiency and Thyroid Function:
A Systematic Review and Meta-Analysis
Vincenzo Garofalo 1 , Rosita A. Condorelli 1, * , Rossella Cannarella 1,2 , Antonio Aversa 3 ,
Aldo E. Calogero 1 and Sandro La Vignera 1

1 Department of Clinical and Experimental Medicine, University of Catania, 95124 Catania, Italy;
[email protected] (V.G.); [email protected] (R.C.); [email protected] (A.E.C.);
[email protected] (S.L.V.)
2 Cleveland Clinic Foundation, Glickman Urological & Kidney Institute, Cleveland, OH 44195, USA
3 Department of Experimental and Clinical Medicine, University Magna Graecia of Catanzaro,
88100 Catanzaro, Italy; [email protected]
* Correspondence: [email protected]

Abstract: Objective: Iron deficiency (ID) is the most prevalent nutritional deficiency worldwide.
Low levels of serum ferritin (SF) could affect the thyroid gland and its functioning. The purpose of
this systematic review and meta-analysis is to summarize the main currently available evidence and
analyze data on the relationship between ID and thyroid function. Methods: This study included all
articles evaluating the relationship between ID and thyroid function. Quality assessment was per-
formed using Cambridge Quality Checklists. The search strategy included the following combination
of Medical Subjects Headings terms and keywords: “iron deficiency”, “thyroid function”, “thyroid
disease”, “thyroid dysfunction”, and “hypothyroidism”. A meta-analysis was performed to evaluate
whether thyroid stimulating hormone (TSH), free thyroxine (FT4), and free triiodothyronine (FT3)
levels differed between patients with ID and healthy controls without ID. For statistical comparison
between cases and controls, the mean difference (MD) was calculated, and a subgroup analysis of
pregnant and non-pregnant women was performed. Cochran’s Q testing and heterogeneity indices
(I2 ) were used to assess statistical heterogeneity. Sensitivity analysis and publication bias analy-
ses were also performed, both qualitatively and quantitatively. Finally, a meta-regression analysis
Citation: Garofalo, V.; Condorelli,
was performed to evaluate the correlation between serum TSH or FT4 levels and SF in the study
R.A.; Cannarella, R.; Aversa, A.;
Calogero, A.E.; La Vignera, S.
population. Results: Ten cross-sectional studies were identified and reviewed. Patients with ID
Relationship between Iron Deficiency showed TSH (MD: −0.24 mIU/L; 95% CI −0.41, −0.07; I2 = 100%, p = 0.005), FT4 (MD: −1.18 pmol/L;
and Thyroid Function: A Systematic 95% CI −1.43, −0.94; I2 = 99%, p < 0.000001), and FT3 (MD: −0.22 pmol/L; 95% CI −0.32, −0.12;
Review and Meta-Analysis. Nutrients I2 = 99%, p < 0.00001) levels that were significantly lower. Subgroup analysis confirmed significantly
2023, 15, 4790. https://doi.org/ lower TSH, FT4, and FT3 levels in pregnant women. Non-pregnant women showed significantly
10.3390/nu15224790 lower serum FT4 and FT3 levels but no difference in TSH values. Meta-regression analysis showed
Academic Editor: Hongbing Sun
that serum TSH and FT4 levels were positively correlated with SF levels. Our systematic review
of the literature found that ID significantly increases the prevalence of thyroid autoantibody (anti-
Received: 4 October 2023 thyroglobulin antibodies and anti-thyroid peroxidase antibodies) positivity both individually and
Revised: 5 November 2023
collectively. Conclusion: Studies currently published in the literature indicate a possible relationship
Accepted: 13 November 2023
between ID, thyroid function, and autoimmunity, especially in some patient groups. Data analysis
Published: 15 November 2023
shows that thyroid hormone levels are lower in patients with ID and, in particular, in pregnant
women. Further studies are needed to understand the role played by iron in thyroid metabolism.

Copyright: © 2023 by the authors. Keywords: iron deficiency; thyroid function; thyroid dysfunction; thyroid disease; hypothyroidism
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons 1. Introduction
Attribution (CC BY) license (https://
Iron is an essential nutrient required for various physiological functions. Iron defi-
creativecommons.org/licenses/by/
ciency (ID) is a widespread nutritional disorder worldwide, affecting about two billion
4.0/).

Nutrients 2023, 15, 4790. https://doi.org/10.3390/nu15224790 https://www.mdpi.com/journal/nutrients

Nutrients 2023, 15, 4790 2 of 19

people, mainly pregnant women and women of childbearing age [1,2]. ID is a condition in
which the body lacks adequate amounts of iron. Iron is a key component of hemoglobin,
a protein present in red blood cells that carries oxygen from the lungs to different tissues
and organs in the body. It is also crucial for other enzymes and proteins involved in energy
production and cell function. This common deficiency can lead to adverse effects on the
thyroid gland, especially the function of the thyroid peroxidase enzyme [3,4]. Furthermore,
anemia due to ID is a common comorbidity in patients with thyroid dysfunction, affecting
nearly 5% of the population [5].
A growing body of evidence suggests that ID may play a significant role in the patho-
genesis of thyroid dysfunction. Several studies have reported a high prevalence of ID in pa-
tients with thyroid diseases, particularly hypothyroidism and thyroid autoimmunity (TAI),
which can impair the synthesis and function of thyroid hormones [6]. The production of
thyroid hormones is negatively affected by ID, and their deficiency reduces the proliferation
of erythrocyte precursors, both directly and through reduced secretion of erythropoietin
by the kidneys [7,8]. Additionally, ID can affect the hypothalamic–pituitary–thyroid axis,
leading to altered thyroid hormone levels and a decreased response to thyroid-stimulating
hormone. Iron is also essential for the activity of thyroid peroxidase, an enzyme that
catalyzes the iodination of tyrosine residues in thyroglobulin, a precursor protein for thy-
roid hormone synthesis [9]. Correlation between ID and hypothyroidism is likely due
to impaired thyroperoxidase (TPO) hemoprotein biosynthesis, as shown in a rat study
in which ID reduced TPO activity [4]. Furthermore, animal studies have shown that ID
can interfere with thyroxine deiodinase activity by reducing the conversion of thyroxine
(T4) to triiodothyronine (T3) and with the regulation of thyroid metabolism at the central
level [10,11]. Moreover, the interaction between thyroid hormones and iron is bidirectional
since, through the TRα receptor, TH directly stimulates erythropoiesis [12,13].
Despite growing interest in the relationship between ID and thyroid dysfunction,
scientific evidence is still inconclusive. Several studies have been conducted to evaluate the
association between iron status and thyroid function, but the results are conflicting. While
some studies suggest that ID is associated with an increased risk of thyroid dysfunction,
others have found no significant association.
A recent systematic review and meta-analysis show that the prevalence of overt and
subclinical hypothyroidism is higher in pregnant women and women of childbearing age
with ID who had higher thyroid-stimulating hormone (TSH) values and reduced free
thyroxine (FT4) values, with a possible increase in the risk of autoantibody positivity [14].
Therefore, a comprehensive and up-to-date systematic review of the literature and meta-
analysis is needed to provide a better understanding of the relationship between ID and
thyroid dysfunction in the general population. This study aims to summarize and critically
evaluate the available evidence on the association between iron status and thyroid function
in the general adult population, including the mechanisms underlying this association, the
prevalence of ID in patients with thyroid diseases, and the impact of iron supplementation
on thyroid function. In addition, this study aims to analyze currently available data on the
impact of serum ferritin values on thyroid hormones.

2. Methods
2.1. Search Strategy
The Meta-Analysis and Systematic Reviews of Observational Studies (MOOSE) guide-
lines [15] (Supplementary Table S1) and the Preferred Reporting Items for Systematic
Review and Meta-Analysis Protocols (PRISMA-P) [16] (Supplementary Table S2) were
employed to carry out our systematic review. A systematic search was performed from
April 2023 to July 2023, across the Pubmed and Scopus databases from the earliest avail-
able through July 2023, using Medical Subjects Headings (MeSH) indexes and keyword
searches. The string used included the entry term “iron deficiency*”, which was searched
in combination (AND) with the terms “thyroid function” OR “thyroid dysfunction” OR
“thyroid disease” OR “hypothyroidism”. The “LIMIT-TO (DOCTYPE, “ar”)” query was
Nutrients 2023, 15, 4790 3 of 19

used to limit the retrieval of original studies only. The same combination of terms was
used for all databases. Abstracts of the retrieved articles were independently screened by
two researchers in duplicate (V.G. and R.A.C.). Disagreements were resolved by a third
person (A.E.C.).

2.2. Selection Criteria

This systematic review included all published articles evaluating the relationship
between ID and thyroid function. All eligible studies were selected following the PECOS
(population, exposure, comparison/comparator, outcomes, study design) [17] model (Table 1).
Specifically, all the studies reporting information on thyroid function in patients exposed to
ID, consisting of a decrease in extracellular iron with a lower-than-normal serum ferritin
(SF), were included. Only original articles in English language reporting complete data of
clinical relevance to the present review were included in the analysis. Duplicates have been
carefully checked and removed.

Table 1. Inclusion and exclusion criteria of the current systematic review, according to the PECOS
model [17].

Inclusion Exclusion
Patients with comorbidities or taking
Population Adults, pregnant women drugs that affect thyroid function or
iron status, age < 18 years.
Exposure Iron deficiency -
Comparison Subjects without ID -
Outcome TSH, FT4, FT3, TPOAb, gAb -
Cross-sectional study, In vitro, animal studies, case reports,
Study Type randomized controlled study, editorials, communications, reviews,
case–control, cohort and meta-analyses
Abbreviations. FT3, free triiodothyronine; FT4, free thyroxine; ID, iron deficiency; SF, serum ferritin; TgAb,
antithyroglobulin antibodies; TPOAb, anti-thyroid peroxidase antibodies; TSH, thyroid-stimulating hormone.

2.3. Data Extraction

We collected information on the first author, year of publication, study design, age,
type of population (gender, pregnant or non-pregnant), diagnostic criteria for ID and SF
levels, and information on primary outcomes (TSH, FT4, FT3, TgAb, and TPOAb). For
analysis, TSH values were converted to mIU/L and FT4 and FT3 in pmol/L, if provided
in different units of measure, according to the conversion tables. For each parameter, the
number of patients/controls, mean value and standard deviation (SD), median value and
interquartile range (IQR) range, or median value and minimum and maximum values
were reported, depending on how the authors reported the data. For studies that express
data as median and IQR or as median and minimum and maximum, the formula by Wan
and colleagues [18] was used to estimate the mean and SD. The data were independently
extracted by V.G.

2.4. Quality Assessment

The quality of evidence (QoE) of the studies was evaluated by VG, by using the
Cambridge Quality Checklists [19], which consists of three checklists to identify high-
quality studies of (1) correlates, (2) risk factors, and (3) causal risk factors. The correlates
checklist evaluates the appropriateness of the sample size and the quality of the outcome
measurements. The risk factor checklist assigns high-quality scores to those studies with
appropriate time-ordered data. Finally, the causal risk factors checklist assesses the type of
study design. To draw confident conclusions about correlates, risk factors, and causal risk
factors, all three checklist scores must score high.
As no randomized control trials were included, no further evaluation was needed.
Nutrients 2023, 15, 4790 4 of 19

2.5. Statistical Analysis

Quantitative data analysis was performed using Comprehensive Meta-Analysis Software
(Version 3) (Biostat Inc., Englewood, NJ, USA) and RevMan software v. 5.4 (Cochrane Col-
laboration, Oxford, UK). The mean difference (MD) was calculated for statistical comparison
of cases and controls, including subgroup analysis between pregnant and non-pregnant
women. Cochran’s Q testing and heterogeneity indices (I2 ) were used to assess statistical
heterogeneity. In particular, if I2 was less than or equal to 50%, the variation of the studies
was considered homogenous, and the fixed effect model was adopted to calculate the
pooled effect size. However, if I2 was greater than 50%, there was significant heterogeneity
between studies, and the random effect model was used. Sensitivity analysis and analysis
of publication bias were also performed. Publication bias was qualitatively analyzed by
the asymmetry of the funnel plot, which suggested some missing studies on one side of
the graph. For the quantitative analysis of publication bias, we used Egger’s intercept
test, which evaluated the statistical significance of publication bias. We also performed a
meta-regression analysis to investigate the correlation between serum TSH of FT4 levels
and SF in the study population. Statistical significance was accepted for p-values less than
or equal to 0.05.

3. Results
The search strategy mentioned above identified a total of 418 records. After 191 duplicates
were excluded, the remaining 227 articles were considered potentially relevant for this
review. After reading the abstracts, 199 articles were excluded because they were not
pertinent. Ten studies were conducted on animals, and one was an in vitro study. The full
text of the remaining seventeen articles were downloaded and read carefully: seven of
these were excluded because they did not include a control group or because data were
insufficient. In conclusion, ten studies were considered for this systematic review (Figure 1).
The main characteristics of the included studies are reported in Table 2. Nine of them
evaluated the association between ID and thyroid function in pregnant women or women
of childbearing age [20–28], and only one study was conducted on the general popula-
tion [29]. Among the included studies, they all evaluated different parameters regarding
thyroid function. All studies [20–29] measured blood levels of TSH and FT4 to evalu-
ate thyroid function, and four of them also evaluated serum FT3 levels [20,21,25,29]. Six
studies [21–24,26,27] evaluated the positivity of TPOAb in the blood of patients, and four
of these [21–24] evaluated the TgAb to determine the possible presence of autoimmune thy-
roiditis. For the diagnosis of ID, four studies [20,22,25,26] used ferritin values < 20 ng/dL, four
studies [21,23,27,29] used ferritin values < 15 ng/dL, and two studies [24,28] values < 12 ng/dL.
All studies used electrochemiluminescence immunoassay to measure blood levels of TSH,
FT4, FT3, TPOAb, TgAb, and SF. The reference ranges of thyroid parameters are reported
in Table 2.

3.1. Quality of Evidence of Included Studies

The QoE of ten studies included revealed that, out of a total score of fifteen, five studies
scored an eleven, four studies scored a ten, and only one study scored a nine (Table 3),
supporting the high quality of the majority of the studies.
Nutrients 2023, 15, 4790 5 of 19

Table 2. Main characteristics of the studies included in the analysis.

Population (n)
Diagnostic Items and TSH FT4 FT3 TPOAb TgAb
Type of Type of
Age (y) Criteria Reference Range
Study Population ID Controls for ID (SF)
ID Controls ID Controls ID Controls ID Controls ID Controls
TSH
(0.24–4.2 mIU/L),
Wang Chinese
29 FT4 1.54 0.93 13.23 16.58 4.15 4.79
et al., CSS Pregnant 387 1831 <20 ng/dL NR NR NR NR
(25–33) (12–22 pmol/L), (0.92–2.35) (0.43–1.56) (11.26–16.08) (14.67–18.67) (3.51–5.04) (4.33–5.32)
2022 [20] women
FT3
(3.1–6.8 pmol/L)
TSH
(0.35–4.2 mIU/L),
FT4
Non-pregnant (0.58–1.64 ng/dL),
Okuroglu
women of FT3
et al., CSS 31.5 ± 8.4 203 155 <15 ng/dL 1.90 ± 0.97 1.74 ± 0.91 0.96 ± 0.11 0.97 ± 0.11 2.77 ± 0.42 2.81 ± 0.60 8.37% 3.87% 9.85% 7.10%
reproductive (1.71–3.71 pg/mL),
2020 [21]
ages AbTPO
(<5.6 IU/mL),
AbTg
(<10 IU/mL)
TSH
(0.27–4.2 mIU/L),
FT4
Zhang Chinese
(12–22 pmol/L), 1.85 1.69 13.94 14.63
et al., CSS 28.3 ± 0.1 pregnant 373 1219 <20 ng/dL NR NR 3.8% 5.2% 4.6% 3.1%
AbTPO (0.01–7.84) (0.01–10.2) (8.91–29.82) (8.22–47.24)
2020 [22] women
(<34 IU/mL),
AbTg
(<115 IU/mL)
TSH
(0.14–4.87 mIU/L),
Chinese
Zhang FT4 (NR),
pregnant and 2.06 1.99 8.9 8.21 12.33 11.35
et al., CSS 28.3 ± 3.6 638 8229 <15 ng/dL AbTPO 14.88 ± 2.32 16.05 ± 3.19 NR NR
non-pregnant (1.17–3.09) (1.14–3.05) (5–15.6) (5–13.82) (10–42.45) (10–24.34)
2019 [23] (<34 IUmL),
women
AbTg
(<115 IU/mL)
Maldonado-
Spanish adult TSH (µUI/mL),
Araque
CSS 50 ± 17.1 non-pregnant 523 3323 <15 ng/dL FT4 (pmol/L), 1.98 ± 0.07 2.04 ± 0.04 14.83 ± 0.10 15.16 ± 0.07 4.91 ± 0.03 5.01 ± 0.02 NR NR NR NR
et al.,
population FT3 (pmol/L)
2018 [29]
TSH
(0.27–4.2 mIU/L),
FT4
He et al., Pregnant (12–22 pmol/L), 9.7 10.4 10 10.02
CSS 27.2 ± 4.5 69 140 <12 ng/dL 3.25 ± 1.72 2.36 ± 1.48 14.26 ± 1.45 14.45 ± 1.54 NR NR
2018 [24] women AbTPO (9.3–14.3) (9.6–14.5) (10–15.95) (9.2–15.09)
(<34 IU/mL),
AbTg
(<115 IU/mL)
Nutrients 2023, 15, 4790 6 of 19

Table 2. Cont.

Population (n)
Diagnostic Items and TSH FT4 FT3 TPOAb TgAb
Type of Type of
Age (y) Criteria Reference Range
Study Population ID Controls for ID (SF)
ID Controls ID Controls ID Controls ID Controls ID Controls
TSH
(0.98–2.26 mIU/L),
Chinese
Fu et al., FT4 1.4 10.36 10.88 4.6
CSS 27 ± 3 pregnant 932 832 <20 ng/dL 1.6 (1–2.3) 4.4 (3.9–5) NR NR NR NR
2017 [25] (8.85–12.25 pmol/L), (0.7–2.2) (8.8–12.2) (8.9–13.44) (4.2–5.2)
women
FT3
(4.06–5.06 pmol/L)
TSH
(0.1–2.5 mIU/L),
Li et al., Pregnant FT4 1.35 1.07 13.22 13.99 3 (23.6%) 3 (11.2%)
CSS 28.5 ± 3.7 431 2150 <20 ng/dL NR NR NR NR
2016 [26] women (11.5–18.8 pmol/L), (0.71–2.4) (0.6–1.67) (11.95–14.84) (13.12–15.07) (3–4.28) (3–3)
AbTPO
(<5.6 mIU/mL)
TSH
(0.3–4 mIU/L),
Veltri
30 Pregnant FT4 1.5 1.3 1.0 1.1 29 (10%) 28 (6%)
et al., CSS 674 1226 <15 ng/dL NR NR NR NR
(15–47) women (0.8–2 ng/dL), (0.0–9.6) (0.0–30.5) (0.7–2.2) (0.6–3.1) (15–9704) (25–13,000)
2016 [27]
AbTPO
(<60 kIU/L)
Chinese TSH
Yu et al., pregnant and (0.14–4.87 mIU/L), 1.94 1.92
CSS 28.9 ± 3.6 143 4249 <12 ng/dL 15.22 ± 2.61 16.06 ± 2.47 NR NR NR NR NR NR
2015 [28] non-pregnant FT4 (1.2–2.73) (1.12–2.85)
women (12.4–20.7 pmol/L)

Abbreviations: AbTg, antithyroglobulin antibodies; AbTPO, anti-thyroid peroxidase antibodies; CSS, cross-sectional study; FT4, free thyroxine; FT3, free triiodothyronine; ID, iron
deficiency; SF, serum ferritin; TSH: thyroid-stimulating hormone.
Nutrients 2023,15,
Nutrients2023, 15,4790
x FOR PEER REVIEW 7 7ofof20
19

Figure1.1.The
Figure The PRISMA
PRISMA diagram
diagram for
for Databases
Databases and
and Registers
Registers [30].
[30].

3.1. Quality
Table of Evidence
3. Quality of Included
of evidence Studies
assessment of the included studies (results of the Cambridge Quality
Checklist [19]).
The QoE of ten studies included revealed that, out of a total score of fifteen, five stud-
ies scored an eleven, four studies scored a ten, and only one study scored a nine (Table 3),
Cambridge Quality Checklists
supporting the high quality of the majority of the studies.
Study Name Type of Study Checklist for
Checklist for Checklist for
Causal Risk
Correlates
Table 3. Quality of evidence assessment of the included Risk Factors
studies (results of the Cambridge Quality
Factors
Checklist [19]).
Wang et al., 2022 [20] CSS 4 1 5
Okuroglu et al., 2020 [21] CSS Cambridge
3 Quality
1 Checklists 5
Type of
Study
Zhang et al.,Name
2020 [22] CSS Checklist for
4 Checklist 1for Checklist5 for
Study
Zhang et al., 2019 [23] CSS Correlates5 Risk Factors1 Causal Risk5 Factors
Wang et al., et
Maldonado-Araque 2022 [20][29]
al., 2018 CSSCSS 4 5 1 1 55
Okuroglu et2018
He et al., al., 2020
[24] [21] CSSCSS 3 4 1 1 55
Zhang
Fu et al., 2017 [25] [22]
et al., 2020 CSSCSS 4 5 1 1 55
ZhangLi et et
al.,al.,
20162019
[26] [23] CSSCSS 5 5 1 1 55
Maldonado-Araque
Veltri et al., 2016 [27]et al., CSSCSS 5 5 1 1 55
2018 [29]
Yu et al., 2015 [28] CSS 4 1 5
He et al.,
Abbreviations. CSS:2018 [24]
cross-sectional CSS
study. 4 1 5
Fu et al., 2017 [25] CSS 5 1 5
 Li et al., 2016 [26] CSS 5 1 5
Veltri et al., 2016 [27] CSS 5 1 5
Yu et al., 2015 [28] CSS 4 1 5
Nutrients 2023, 15, 4790 8 of 19
Abbreviations. CSS: cross-sectional study.

3.2. Results of the Quantitative Synthesis

3.2. Results of the Quantitative Synthesis
3.2.1. TSH
3.2.1. TSH
Ten studies, including 4395 patients and 23,354 controls, assessed serum TSH levels.
Ten studies, including 4395 patients and 23,354 controls, assessed serum TSH levels.
Our analysis showed that, overall, TSH was significantly lower in patients than in controls,
Our analysis showed that, overall, TSH was significantly lower in patients than in controls,
although
although interstudy heterogeneitywas
interstudy heterogeneity wasfound
found(MD:
(MD:−−0.24 mIU/L; 95%
0.24 mIU/L; 95%CI CI−−0.41, −0.07;
0.41, − 0.07; I =
2

100%, p = 0.005).
I2 = 100%, Subgroup
p = 0.005). analysis
Subgroup revealed
analysis significantly
revealed lower
significantly TSH
lower TSHvalues in in
values thethe
preg-
nant population (MD: −0.41 mIU/L; 95% CI −0.73, −0.09; I2 = 2100%, p = 0.01), while they
pregnant population (MD: −0.41 mIU/L; 95% CI −0.73, −0.09; I = 100%, p = 0.01), while
were non-significantly
they were non-significantly different in non-pregnant
different in non-pregnantsubjects (MD:
subjects (MD:0.01 mIU/L;
0.01 95%
mIU/L; 95%CI CI
−0.09,
0.11; I2 =0.11;
−0.09, 96%,I2 p= =96%,
0.86)p (Figure 2).
= 0.86) (Figure 2).

Figure 2.
Figure Forest plot of
2. Forest of mean
meandifference
differenceininserumserumTSH TSH levels between
levels between patients andand
patients controls and and
controls
subgroup analysis
subgroup analysis in pregnant
pregnantand andnon-pregnant
non-pregnantwomen. women. Green
Green boxes estimate
boxes estimatethe the
study results,
study results,
and
andgive
give aa representation
representation of ofthe
thesize
sizeofofthe
thestudies.
studies. Horizontal
Horizontal lines
lines represent
represent the the
95%95% confidence
confidence
intervals
intervals of
of the
the study results,with
study results, witheach
eachend endofofthe
theline
line representing
representing thethe boundaries
boundaries of the
of the confidence
confidence
interval. The black diamond represents the point estimate and confidence
interval. The black diamond represents the point estimate and confidence intervals of the combined intervals of the combined
studies. The left arrow indicates a study whose estimate point and confidence
studies. The left arrow indicates a study whose estimate point and confidence interval are lower interval are lower
than -2. Fu et al., 2017 [25]; He et al., 2018 [24]; Li et al., 2016 [26]; Veltri et al., 2016 [27];
than −2. Fu et al., 2017 [25]; He et al., 2018 [24]; Li et al., 2016 [26]; Veltri et al., 2016 [27]; Wang et al., Wang et al.,
2022
2022 [20]; Yu et al., 2015 [28]; Zhang et al., 2019 [23]; Zhang et al., 2020 [22]; Maldonado-Araque et al., et
[20]; Yu et al., 2015 [28]; Zhang et al., 2019 [23]; Zhang et al., 2020 [22]; Maldonado-Araque
al.,
20152015 [29];
[29]; Okuroglu
Okuroglu et2020
et al., al., 2020
[21]. [21].

The analysisshowed
The analysis showed thethe absence
absence of publication
of publication bias asbias as inferred
inferred by Egger’sby test
Egger’s test (in-
(intercept
tercept 8.12,
8.12, 95% −37.7;
CI95% CI −37.7;
53.9; p 53.9; p =and
= 0.69) 0.69)symmetry
and symmetry of the funnel
of the funnel plots (Scheme
plots (Scheme 1, panel1,A).
panel
However, all studies were found to be sensitive enough to change the conclusion
A). However, all studies were found to be sensitive enough to change the conclusion that that serum
TSH levels
serum are lower
TSH levels in patients
are lower thanthan
in patients in controls (Scheme
in controls (Scheme1, panel B). B).
1, panel Meta-regression
Meta-regression
analysis showed
analysis showed that
that serum
serumTSH TSHlevels
levelscorrelated positively
correlated positively with SF.SF.
with In this analysis,
In this the the
analysis,
magnitude of mean serum levels in patients and controls increased as a function ofSFtheir
magnitude of mean serum levels in patients and controls increased as a function of their
values (Scheme 2).
SF values (Scheme 2).
EER REVIEW 24 of 20
Nutrients 2023, 15, 4790 9 of 19

Scheme 1. Funnel plot Scheme

(A) and sensitivity
1. Funnel plot (A) analysis (B)analysis
and sensitivity of the (B)meanof thedifference
mean difference in in
serum
serum TSH values
TSH values
in patients and controls. In Panel
in patients A, circles
and controls. indicate
In Panel theindicate
(A), circles standard error error
the standard plotted bybythe
plotted standard
the standard
difference in means of each study. Rhombus indicates a standard difference in means ofzero.
difference in means of each study. Rhombus indicates a standard difference in means of zero.In In
Panel (B), black boxes estimate the study results, and give a representation of the size of the studies.
Panel B, black boxes estimate the study results, and give a representation of the size of the studies.
horizontal lines represent the 95% confidence intervals of the study results, with each end of the
horizontal lines representline the 95% confidence
representing the boundaries intervals of the interval.
of the confidence study results, with each
The blue diamond end of
represents thethe
pointline
representing the boundaries
estimate andof confidence
the confidenceintervals ofinterval.
the combinedThestudies.
blue Fu diamond
et al., 2017 represents the [24];
[25]; He et al., 2018 point
estimate and confidence intervals of the combined studies. Fu et al., 2017 [25]; He et al., 2018 [24];
Li et al., 2016 [26]; Veltri et al., 2016 [27]; Wang et al., 2022 [20]; Yu et al., 2015 [28]; Zhang et al., Li
2019 [23]; Zhang et al., 2020 [22]; Maldonado-Araque et al., 2015 [29]; Okuroglu et al., 2020 [21].
et al., 2016 [26]; Veltri et al., 2016 [27]; Wang et al., 2022 [20]; Yu et al., 2015 [28]; Zhang et al., 2019
[23]; Zhang et al., 2020 [22]; Maldonado-Araque et al., 2015 [29]; Okuroglu et al., 2020 [21].
Nutrients 2023, 15, x FOR PEER REVIEW 25 of 20
Nutrients 2023, 15, 4790 10 of 19

Scheme
Scheme2.2.Meta-regression showing
Meta-regression showing thethe influence
influence of serum
of serum ferritin
ferritin (SF) (moderator)
(SF) levels levels (moderator)
on the on the
mean
meandifference
difference of serum
serumTSHTSHlevels
levelsin in patients
patients from from the study
the study population.
population. The sizeThe sizecircles
of the of the circles
indicates
indicatesthe
the size ofthe
size of thestudy.
study.TheThe
redred
lineline represents
represents linear linear predictions
predictions for meanfor mean difference
difference of TSH as of TSH
asaafunction
function of mean
of the the mean absolute
absolute increaseincrease
in SF. Theincurved
SF. Thelinescurved
indicatelines indicate
the 95% the 95%
confidence confidence
interval
interval lines around
lines around the regression
the regression line. line.

3.2.2. Free Thyroxin

3.2.2. Free Thyroxin
Ten studies, including 4395 patients and 23,354 controls, evaluated serum FT4 levels.
FT4Tenwasstudies, including
significantly 4395
lower in patients
patients thanand 23,354 despite
in controls, controls, evaluated
interstudy serum FT4 levels.
heterogeneity
FT4
(MD:was significantly
−1.18 pmol/L; 95% lower
CI −in1.43,
patients
−0.94;than
2 in controls,
I = 99%, despite
p < 0.000001). By interstudy heterogeneity
subgroup analysis,
2 = 98%,
(MD: −1.18 pmol/L; 95% CI −1.43, −0.94; I = 99%, p < 0.000001). By subgroup analysis, FT4
FT4 was lower in both pregnant women (MD: 2 − 1.43 pmol/L; 95% CI − 1.81, − 1.05; I
p < lower
was 0.000001) and non-pregnant
in both pregnant women (MD: −
women(MD: 0.72 pmol/L;
−1.43 pmol/L;95%95%CICI−−1.81,
1.39, 0.06; I2 = I96%,
−1.05; 2 = 98%, p <
p = 0.03) (Figure
0.000001) 3).
and non-pregnant women (MD: −0.72 pmol/L; 95% CI −1.39, 0.06; I2 = 96%, p =
The analysis showed the absence of publication bias as inferred by Egger’s test (in-
0.03) (Figure 3).
tercept −5.35, 95% CI −52.8; 42.1; p = 0.80) and symmetry of the funnel plots (Scheme 3,
panel A). No study was sensitive enough to alter the results (Scheme 3, panel B). In meta-
regression analysis, we found a positive correlation between serum FT4 levels and SF levels
in the study population. Indeed, the magnitude of mean serum FT4 levels in patients and
controls increased as a function of their SF levels (Scheme 4).
Nutrients 2023, 15, 4790 Nutrients 2023, 15, x FOR PEER REVIEW 11 of 19 11 of

Figure 3. Forest plot of mean difference in serum free thyroxin levels between patients and controls
Figure 3. Forest plot of mean difference in serum free thyroxin levels between patients and contr
and subgroup analysis andinsubgroup
pregnantanalysis
and non-pregnant
in pregnant women. Green boxes
and non-pregnant estimate
women. theboxes
Green studyestimate
results,the study
and give a representation of the
sults, and givesize of the studies.
a representation Horizontal
of the size of thelines represent
studies. the 95%
Horizontal linesconfidence
represent the 95% con
intervals of the studydence intervals
results, of theend
with each study results,
of the with each endthe
line representing of boundaries
the line representing the boundaries of
of the confidence
interval. The black diamond represents the point estimate and confidence intervals of the combined intervals
confidence interval. The black diamond represents the point estimate and confidence
the combined studies. Fu et al., 2017 [25]; He et al., 2018 [24]; Li et al., 2016 [26]; Veltri et al., 20
studies. Fu et al., 2017 [25]; He et al., 2018 [24]; Li et al., 2016 [26]; Veltri et al., 2016 [27]; Wang et al.,
[27]; Wang et al., 2022 [20]; Yu et al., 2015 [28]; Zhang et al., 2019 [23]; Zhang et al., 2020 [2
2022 [20]; Yu et al., 2015
Nutrients 2023, 15, x FOR PEER REVIEW [28]; Zhang et al.,
Maldonado-Araque 2019
et al., [23];
2015 Zhang
[29]; et al.,et2020
Okuroglu [22];[21].
al., 2020 Maldonado-Araque et 20
27 of al.,
2015 [29]; Okuroglu et al., 2020 [21].
The analysis showed the absence of publication bias as inferred by Egger’s test (
tercept −5.35, 95% CI −52.8; 42.1; p = 0.80) and symmetry of the funnel plots (Scheme
panel A). No study was sensitive enough to alter the results (Scheme 3, panel B). In me
regression analysis, we found a positive correlation between serum FT4 levels and SF le
els in the study population. Indeed, the magnitude of mean serum FT4 levels in patien
and controls increased as a function of their SF levels (Scheme 4).

Scheme
Scheme 3. Funnel
3. Funnel plotplot (A) and
(A) and sensitivity
sensitivity analysis
analysis (B) of(B) of mean
mean difference
difference in serum
in serum free thyroxin
free thyroxin (FT4)
(FT4) levels in patients and controls. In Panel A, circles indicate the standard error plotted by the
standard difference in means of each study. Rhombus indicates a standard difference in means of
zero. In Panel B, black boxes estimate the study results, and give a representation of the size of the
studies. horizontal lines represent the 95% confidence intervals of the study results, with each end
of the line representing the boundaries of the confidence interval. The blue diamond represents the
point estimate and confidence intervals of the combined studies. Fu et al., 2017 [25]; He et al., 2018
 Nutrients 2023, 15, 4790 12 of 19

levels in patients and controls. In Panel (A), circles indicate the standard error plotted by the standard
difference in means of each study. Rhombus indicates a standard difference in means of zero. In
Panel (B), black boxes estimate the study results, and give a representation of the size of the studies.
horizontal lines represent the 95% confidence intervals of the study results, with each end of the
line representing the boundaries of the confidence interval. The blue diamond represents the point
estimate and confidence intervals of the combined studies. Fu et al., 2017 [25]; He et al., 2018 [24];
Nutrients 2023, 15, x FOR PEER REVIEW 28 of 20
Li et al., 2016 [26]; Veltri et al., 2016 [27]; Wang et al., 2022 [20]; Yu et al., 2015 [28]; Zhang et al.,
2019 [23]; Zhang et al., 2020 [22]; Maldonado-Araque et al., 2015 [29]; Okuroglu et al., 2020 [21].

Scheme
Scheme4.4.Meta-regression showingthe
Meta-regression showing theinfluence
influence of serum
of serum ferritin
ferritin (SF) (SF)
levelslevels (moderator)
(moderator) on the on the
mean
meandifference in serum
difference in serumfree
free thyroxin
thyroxin (FT4)
(FT4) levelslevels
among among patients
patients in the
in the study study population.
population. The size The
size of the
of the circles
circles indicates
indicates the sample
the sample sizestudy.
size of the of theThestudy. The
red line red line linear
represents represents linearforpredictions
predictions the
formean
the mean difference
difference inaFT4
in FT4 as as a of
function function
the mean of absolute
the mean absolute
increase increase
in SF. in SF.
The curved The
lines curved lines
indicate
indicate the 95% confidence interval lines around the
the 95% confidence interval lines around the regression line. regression line.

3.2.3.Free
3.2.3. FreeTriiodothyronine
Triiodothyronine
Four studies, including 2047 patients and 6143 controls, evaluated serum FT3 levels.
Four studies, including 2047 patients and 6143 controls, evaluated serum FT3 levels.
FT3 was significantly lower in patients than in controls, despite the presence of interstudy
FT3 was significantly
heterogeneity (MD: −lower in patients
0.22 pmol/L; than
95% CI in controls,
−0.32, −0.12; I2despite
= 99%, pthe presenceItsoflevels
< 0.00001). interstudy
heterogeneity (MD: −0.22 pmol/L; 95% CI −0.32, −0.12; I
were lower in both pregnant (MD: −0.40 pmol/L; 95% CI −0.75, −0.04; I = 97%, p = 0.03) were
2 = 99%, p < 0.00001).
2 Its levels
lower in both pregnant
and non-pregnant women (MD: −0.40
(MD: −0.10pmol/L;
pmol/L;95%95%CI
CI−0.75,
−0.12,−0.04;
−0.07;I2I2= =97%,
11%,pp=<0.03) and non-
0.00001)
pregnant women (MD: −0.10 pmol/L; 95% CI −0.12, −0.07; I2 = 11%, p < 0.00001) (Figure 4).
(Figure 4).
The analysis showed the absence of publication bias as inferred by Egger’s test (in-
tercept −6.9, 95% CI −312.2; 298.3; p = 0.93) and symmetry of the funnel plots (Scheme 5,
panel A). Three studies were found to be sensitive enough to alter the conclusion that FT3
is lower in patients than in controls [20,21,25] (Scheme 5, panel B).
 3.2.3. Free Triiodothyronine
Four studies, including 2047 patients and 6143 controls, evaluated serum FT3 levels.
FT3 was significantly lower in patients than in controls, despite the presence of interstudy
heterogeneity (MD: −0.22 pmol/L; 95% CI −0.32, −0.12; I2 = 99%, p < 0.00001). Its levels were
Nutrients 2023, 15, 4790 13 of 19
lower in both pregnant (MD: −0.40 pmol/L; 95% CI −0.75, −0.04; I2 = 97%, p = 0.03) and non-
pregnant women (MD: −0.10 pmol/L; 95% CI −0.12, −0.07; I2 = 11%, p < 0.00001) (Figure 4).

Nutrients 2023, 15, x FOR PEER REVIEW 29 of 20

the study results, and give a representation of the size of the studies. Horizontal lines represent the
95% confidence intervals of the study results, with each end of the line representing the boundaries
of the confidence
Figure 4. Forest plotinterval.
of theThe
mean black diamond
difference in represents
serum freethe point estimatelevels
triiodothyronine and confidence intervals
between patients
Figure 4. Forest plot of the mean difference in serum free triiodothyronine levels between patients
of the
and combined
controls and studies.
subgroup Fuanalysis
et al., 2017 [25]; Wang
in pregnant andetnon-pregnant
al., 2022 [20]; women.
Maldonado-Araque
Green boxeset al., 2015
estimate
and
[29];controls
Okurogluandetsubgroup analysis in pregnant and non-pregnant women. Green boxes estimate the
al., 2020 [21].
study results, and give a representation of the size of the studies. Horizontal lines represent the 95%
confidence
The intervals
analysisofshowed the studytheresults, with each
absence end of the line
of publication representing
bias the boundaries
as inferred by Egger’sof testthe
confidence interval. The black diamond represents the point estimate and
(intercept −6.9, 95% CI −312.2; 298.3; p = 0.93) and symmetry of the funnel plots (Scheme confidence intervals of
the combined studies. Fu et al., 2017 [25]; Wang et al., 2022 [20]; Maldonado-Araque
5, panel A). Three studies were found to be sensitive enough to alter the conclusion that et al., 2015 [29];
Okuroglu et al.,in2020
FT3 is lower [21]. than in controls [20,21,25] (Scheme 5, panel B).
patients

Scheme5.5.Funnel
Scheme Funnel plot
plot (A) (A)
and and sensitivity
sensitivity analysis
analysis (B)mean
(B) of the of the mean in
difference difference
serum freeintriiodothyro-
serum free
nine (FT3) levels in patients and controls. In Panel (A), circles indicate the standard error plotted byerror
triiodothyronine (FT3) levels in patients and controls. In Panel A, circles indicate the standard the
plotted by the standard difference in means of each study. Rhombus indicates a standard difference
standard difference in means of each study. Rhombus indicates a standard difference in means of zero.
in means of zero. In Panel B, black boxes estimate the study results, and give a representation of the
size of the studies. horizontal lines represent the 95% confidence intervals of the study results, with
each end of the line representing the boundaries of the confidence interval. The blue diamond
represents the point estimate and confidence intervals of the combined studies. Fu et al., 2017 [25];
Wang et al., 2022 [20]; Maldonado-Araque et al., 2015 [29]; Okuroglu et al., 2020 [21].
Nutrients 2023, 15, 4790 14 of 19

In Panel (B), black boxes estimate the study results, and give a representation of the size of the studies.
horizontal lines represent the 95% confidence intervals of the study results, with each end of the
line representing the boundaries of the confidence interval. The blue diamond represents the point
estimate and confidence intervals of the combined studies. Fu et al., 2017 [25]; Wang et al., 2022 [20];
Maldonado-Araque et al., 2015 [29]; Okuroglu et al., 2020 [21].

3.3. Results of the Qualitative Synthesis

All studies showed a significant reduction in blood FT4 levels in patients with ID, and
only two studies [20,29] showed a significant reduction in plasma FT3 levels. Regarding
TSH levels, seven studies [20–22,24–27] reported a direct relationship with the severity
of ID; ID patients in these studies had higher values than the control group without ID.
However, three studies did not show significant differences in TSH levels between patients
with and without ID [23,28,29]. This may be due to the different populations included
across the studies. Indeed, some included non-pregnant women [23,28] and others included
the general population [29].
Regarding autoimmunity, in patients with ID, an increase in TPOAb positivity preva-
lence was reported in four studies [21,23,26,27], while TgAb positivity prevalence was
reported in two studies [21,22]. Okuroglu and colleagues reported an increase in the posi-
tivity prevalence of both autoantibodies in non-pregnant women of childbearing age [21].
Zhang Y. and collaborators are the only ones to report that ID is a risk factor for TgAb
positivity and not for TPOAb during the second trimester of pregnancy [22]. However,
in another study [23], the prevalence of isolated TPOAb positivity in the ID group was
significantly higher than the control group in non-pregnant women, while the prevalence
of isolated TgAb positivity was similar between the ID group and the control group, both
in pregnant and non-pregnant women. The difference could be due to the presence of
non-pregnant women in the second study. Li and colleagues, instead, showed greater
positivity prevalence for TPOAb in women with ID in early pregnancy than in the control
group [26]. Only one study [24] found no differences in TPOAb and TgAb levels in relation
to SF levels.

4. Discussion
ID is the most prevalent nutritional deficiency worldwide. The most affected popula-
tions are pregnant women and women of childbearing age. The World Health Organization
reports that ID is the leading cause of anemia in pregnant women (38.2%) and women of
reproductive age (29.4%) [31]. Several studies have shown that ID has negative effects on
thyroid function. Iron is an essential trace element for the biosynthesis and function of
thyroid hormones. In the body, iron is mainly contained in hemoglobin and myoglobin, but
a small part is present in various cytochromes and other hemoproteins, including thyroper-
oxidase (TPO), where it represents the central atom of the prosthetic groups in their active
site. Expressed inside thyrocytes, TPO is the key enzyme in thyroid hormone synthesis,
catalyzing the H2 O2 -dependent oxidative reaction of iodide ions to elemental iodine which,
in turn, will be incorporated into thyroglobulin to produce thyroid hormones [32].
The most important clinical indicator of ID is SF, which reflects the body’s iron stores.
Restoration of adequate SF concentrations promotes the recovery of normal thyroid function
in ID girls [33]. Furthermore, in patients with ID and subclinical hypothyroidism, combined
treatment with levothyroxine and iron supplements demonstrated better effects than either
factor alone [34,35].
Most studies in the literature related to ID and thyroid function have been conducted
on pregnant women and women of childbearing age, while only one study has been carried
out on large samples of the general adult population [29]. Subclinical hypothyroidism
ranges from 4% to 17% during pregnancy, and thyroid hormones play a crucial role in brain
development, especially during the first months of intrauterine life [36]. Impaired thyroid
function and TAI during pregnancy are associated with gestational diabetes mellitus,
blood hypertension, miscarriage, fetal dysplasia, and neurological deficits in children [37].
Nutrients 2023, 15, 4790 15 of 19

Pregnant women are more vulnerable to ID as their needs increase due to the expansion
of the erythrocyte mass and growth of the fetus and placenta [38]. Zimmermann and
colleagues reported that an insufficient maternal iron reserve predicts lower blood total
thyroxin levels and higher TSH levels during pregnancy [39].
In this systematic review and meta-analysis, we collected studies regarding the corre-
lation between ID and thyroid function. A careful literature review collected cross-sectional
studies comparing SF levels and thyroid function markers, specifically, TSH, FT4, FT3,
TPOAb, and TgAb, in the general population of adults, pregnant women, or women of
childbearing age.
Wang and colleagues, in their study of Chinese pregnant women with ID or ID anemia,
observed that serum FT3 and FT4 values had a significant downward trend in the ID and ID
anemia groups, with lower levels in the second group compared with the control group. In
addition, significantly higher TSH values have been reported in pregnant women with ID or
ID anemia. Finally, compared with the control group, the rate of hypothyroidism was higher
in both these groups of patients. Using canonical correlation analysis, SF and hemoglobin
values were positively correlated with FT3 and FT4 and negatively correlated with TSH,
thus providing further evidence of the correlation between iron status and thyroid function.
This was further validated by ROC curve analysis using selected variables from Lasso’s
model to predict the level of FT4, where the accuracy was good [20].
Another study of Chinese women in the first trimester of pregnancy shows that
the serum TSH level in women with SF < 20 ng/dL was significantly higher than in
the SF 20–100 ng/dL and SF > 100 ng/dL groups. Conversely, the blood FT4 level in
women with SF < 20 ng/dL is significantly lower than in groups with SF > 20 ng/dL.
Furthermore, the relationship between urinary iodine concentration (UIC) with iron status
and thyroid function was evaluated. Observations showed a positive and significant
association between UIC and SF and between UIC and FT3 when UIC is less than 249 µg/L.
In addition, observations showed a negative and significant association between UIC and
SF and between UIC and FT3 levels when UIC is greater than 250 µg/L. These data indicate
that more than adequate iodine nutritional status is associated with reduced SF and FT3
concentrations [25]. These results are in line with other studies where UIC correlates
positively with serum TSH during pregnancy [40,41].
In contrast to the two previous studies, Yu and colleagues reported that both pregnant
and non-pregnant women with ID had an increased risk of isolated hypothyroxinemia,
with no significant differences in serum TSH levels [28]. The difference in results may be
explained by the different iodine status, the presence of non-pregnant women in the study
by Yu and colleagues, and the use of different cut-offs to define ID. Another explanation
could be the lack of TAI assessment in these studies [20,25,28].
A study on a large sample of subjects without thyroid diseases, representative of the
Spanish adult population, shows that subjects with SF levels < 30 mL/dL were more likely
to have low FT4 and FT3 levels than subjects with SF ≥ 30 mg/dL, confirming a possible
association between ID and hypothyroxinemia. However, mean TSH concentrations did
not show significant changes [29].
Okuroglu and coworkers, in their study, support the association between ID and TAI.
In the ID group, positivity for TPOAb plus TgAb was significantly higher than in the
control group [21]. Zhang and colleagues found lower FT4 levels in pregnant and non-ID
Chinese women than in the control group, but the median TSH levels were similar in both
the ID and control groups. Regarding TAI, the prevalence of isolated TPOAb positivity
was higher in the ID group than in women without ID. After adjusting for confounding
factors in multivariable logistic regression, ID remained associated with isolated TPOAb
positivity in both pregnant and non-pregnant women. However, ID was not associated
with isolated TgAb positivity [23]. These results are in line with two previous studies in
which the prevalence of positivity for TPOAb was higher in women with ID during early
pregnancy and in which SF levels were directly proportional to FT4 levels and inversely
proportional to TSH levels [26,27].
Nutrients 2023, 15, 4790 16 of 19

The mechanism by which ID is associated with TPOAb but not TgAb positivity
remains unclear. It has been hypothesized that ID may produce post-translational changes
in TPO, leading to the generation of new epitopes or the exposure of previously hidden
epitopes and thus increasing the immunogenicity of TPO. Another hypothesis is that the
decrease in TPO activity resulting from ID causes an increase in TAI [27].
Zhang et al. focused on the association of TAI and ID during the second trimester of
pregnancy in Chinese women. Unlike other studies, they were the first to report that ID
was a risk factor only for increased TgAb levels and not for TPOAb [22]. As previously
mentioned, ID reduces the activity of heme-containing enzymes including myeloperoxidase
(MPO). Women with ID may exhibit increased MPOAb which, in turn, may cross-react
with AbTPO and, consequently, lead to increased thyroid autoantibody positivity [42]. At
present, the mechanisms by which SF deficiency increases TgAb levels is not known. It
is hypothesized that following SF deficiency, the decrease in FT4 results in a reduction of
feedback inhibition on TSH, thereby increasing its levels. Since TSH promotes TG mRNA
and TG antigen expression, this may lead to increased TgAb production [43].
Only one study [24] has excluded a correlation between TAI and ID in second-trimester
pregnant women with normal urinary iodine levels, thus emphasizing an important role of
this micronutrient in the pathogenesis of TAI [44].
Analysis of data on TSH and free thyroid hormones showed that their serum levels
were significantly lower in ID patients than in controls, with lower values in pregnant
women than in non-pregnant women. The behavior of TSH is curious: it goes in the
same direction as FT4. This could be explained by the heterogeneity of the examined
population, as serum TSH levels were lower in pregnant women than in the general
population. However, the FT4 values should be more reliable than the TSH levels since the
significance of the latter is lost in the sensitivity analysis. Meta-regression analysis shows
that mean serum TSH and FT4 levels in patients and controls increase as a function of SF
values, indicating a possible relationship between iron status and thyroid function.
ID may also be associated with an increased risk of developing TAI, particularly
in pregnant women, and should be included in the screening of women planning to
have a child. However, most of the studies considered do not take into account iodine
status, and some of them are conducted in areas with mild iodine deficiency. Also, other
parameters should be considered. First, the role of inflammation on SF, particularly during
pregnancy, as well as the role of other hormones that might affect thyroid function and
iron status. Inflammation hinders the interpretation of iron biomarkers, particularly serum
ferritin and hepcidin. In fact, several pro-inflammatory cytokines increase the synthesis
of ferritin and hepcidin with a consequent increase in iron trapping within cells. In case
of acute inflammation, serum ferritin levels of 50 ng/mL or greater may indicate ID.
Therefore, inflammation reduces the predictive values of ferritin and hepcidin in the case
of ID [45]. Recent studies have shown a correlation between hepcidin levels and the risk of
developing subacute and chronic autoimmune thyroiditis. A reduction in hepcidin levels
has been observed in patients with Hashimoto’s thyroiditis in whom euthyroidism has
been reverted [46]. Thus, it is possible that hypothyroidism results in increased hepcidin
with a consequent reduction in iron absorption. Testosterone is negatively associated with
serum ferritin. In obese hypogonadal elderly men receiving testosterone replacement
therapy, serum ferritin levels significantly decreased, suggesting a regulatory function of
testosterone on ferritin synthesis. Indeed, in obese patients, decreased testosterone levels
can lead to an increase in serum ferritin [47,48]. Insulin resistance and metabolic syndrome
also appear to positively correlate with serum ferritin levels [49]. Studies conducted in
pregnant women did not consider the role of human chorionic gonadotropin (hCG) on TSH
levels. In fact, hCG has structural similarities with TSH, allowing it to bind to TSH receptors
at the thyroid level, stimulating the secretory activity of the gland. This explains the slight
increase in serum free thyroxine levels during the first weeks of pregnancy, accompanied
by a reduction in serum TSH concentration [50]. Finally, the different studies considered
different SF cut-offs for the diagnosis of ID. However, as indicated in the guidelines, all
Nutrients 2023, 15, 4790 17 of 19

studies report SF values <20 ng/dL as the cut-off for the diagnosis of ID, and the lower
limit of normal for most laboratories is between 15 and 30 ng/dL [51].
This systematic review and meta-analysis has some limitations. First, the lack of
studies conducted on the general population. Most of the studies are performed on
pregnant women and this could affect the results. Second, the reviewed studies considered
different SF cut-offs for the diagnosis of ID. Furthermore, most studies in the literature that
include a control group are cross-sectional, so we do not know the results of prospective
or longitudinal studies. Finally, many of the studies analyzed do not consider the iodine
status of the patients examined, which as is known influences thyroid function. In addition
to iodine, other parameters such as inflammation and the action of other hormones affect
iron status and thyroid function and could therefore influence the results.
In conclusion, ID may adversely affect thyroid function and autoimmunity, especially
in some groups, such as pregnant women and women of childbearing age. The articles
analyzed in this systematic review and meta-analysis are cross-sectional studies, and it was
not possible to distinguish randomness between ID and thyroid function. The relationship
between trace elements and thyroid disorders is still unclear, and more research is needed
to clarify this issue and improve our understanding of how trace elements mediate thyroid
function and metabolism. Prospective randomized controlled trials are needed to clarify
the importance of iron’s nutritional status on thyroid health, including in the general
population or in other patient groups.

Supplementary Materials: The following supporting information can be downloaded at: https://
www.mdpi.com/article/10.3390/nu15224790/s1, Table S1: MOOSE Checklist for Meta-analyses of
Observational Studies; Table S2: PRISMA 2009 Checklist. References [15,52] are cited in supplemen-
tary materials.
Author Contributions: Conceptualization, V.G. and S.L.V.; methodology, V.G. and R.C.; software,
R.C.; validation, R.A.C. and A.A.; formal analysis, R.C.; investigation, V.G.; data curation, V.G.;
writing—original draft preparation, V.G.; writing—review and editing, A.E.C., R.C., and S.L.V.;
visualization, R.A.C. and A.A..; supervision, A.E.C.; project administration, S.L.V. All authors have
read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Data are contained within the article.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Kumar, S.B.; Arnipalli, S.R.; Mehta, P.; Carrau, S.; Ziouzenkova, O. Iron Deficiency Anemia: Efficacy and Limitations of Nutritional
and Comprehensive Mitigation Strategies. Nutrients 2022, 14, 2976. [CrossRef] [PubMed]
2. Zimmermann, M.B.; Hurrell, R.F. Nutritional iron deficiency. Lancet 2007, 370, 511–520. [CrossRef] [PubMed]
3. Pasricha, S.R.; Tye-Din, J.; Muckenthaler, M.U.; Swinkels, D.W. Iron deficiency. Lancet 2021, 397, 233–248. [CrossRef] [PubMed]
4. Hess, S.Y.; Zimmermann, M.B.; Arnold, M.; Langhans, W.; Hurrell, R.F. Iron deficiency anemia reduces thyroid peroxidase activity
in rats. J. Nutr. 2002, 132, 1951–1955. [CrossRef]
5. Szczepanek-Parulska, E.; Hernik, A.; Ruchała, M. Anemia in thyroid diseases. Pol. Arch. Intern. Med. 2017, 127, 352–360.
[CrossRef]
6. Soliman, A.T.; De Sanctis, V.; Yassin, M.; Wagdy, M.; Soliman, N. Chronic anemia and thyroid function. Acta Biomed. 2017, 88,
119–127. [CrossRef]
7. Golde, D.W.; Bersch, N.; Chopra, I.J.; Cline, M.J. Thyroid hormones stimulate erythropoiesis in vitro. Br. J. Haematol. 1977, 37,
173–177. [CrossRef]
8. Maggio, M.; De Vita, F.; Fisichella, A.; Lauretani, F.; Ticinesi, A.; Ceresini, G.; Cappola, A.; Ferrucci, L.; Ceda, G.P. The Role of the
Multiple Hormonal Dysregulation in the Onset of “Anemia of Aging”: Focus on Testosterone, IGF-1, and Thyroid Hormones. Int.
J. Endocrinol. 2015, 2015, 292574. [CrossRef]
9. Zimmermann, M.B. The impact of iron and selenium deficiencies on iodine and thyroid metabolism: Biochemistry and relevance
to public health. Thyroid 2002, 12, 867–878. [CrossRef]
Nutrients 2023, 15, 4790 18 of 19

10. Brigham, D.E.; Beard, J.L. Effect of thyroid hormone replacement in iron-deficient rats. Am. J. Physiol. 1995, 269 Pt. 2, R1140–R1147.
[CrossRef]
11. Beard, J.L.; Brigham, D.E.; Kelley, S.K.; Green, M.H. Plasma thyroid hormone kinetics are altered in iron-deficient rats. J. Nutr.
1998, 128, 1401–1408. [CrossRef] [PubMed]
12. van Gucht, A.L.M.; Meima, M.E.; Moran, C.; Agostini, M.; Tylki-Szymanska, A.; Krajewska, M.W.; Chrzanowska, K.; Efthymi-
adou, A.; Chrysis, D.; Demir, K.; et al. Anemia in Patients with Resistance to Thyroid Hormone α: A Role for Thyroid Hormone
Receptor α in Human Erythropoiesis. J. Clin. Endocrinol. Metab. 2017, 102, 3517–3525. [CrossRef] [PubMed]
13. Pastori, V.; Pozzi, S.; Labedz, A.; Ahmed, S.; Ronchi, A.E. Role of Nuclear Receptors in Controlling Erythropoiesis. Int. J. Mol. Sci.
2022, 23, 2800. [CrossRef] [PubMed]
14. Luo, J.; Wang, X.; Yuan, L.; Guo, L. Iron Deficiency, a Risk Factor of Thyroid Disorders in Reproductive-Age and Pregnant Women:
A Systematic Review and Meta-Analysis. Front Endocrinol. 2021, 12, 629831. [CrossRef] [PubMed]
15. Stroup, D.F.; Berlin, J.A.; Morton, S.C.; Olkin, I.; Williamson, G.D.; Rennie, D.; Moher, D.; Becker, B.J.; Sipe, T.A.; Thacker, S.B.
Meta-analysis of observational studies in epidemiology: A proposal for reporting. Meta-analysis of 14 Observational Studies in
Epidemiology (MOOSE) group. J. Am. Med. Assoc. 2000, 283, 2008–2012. [CrossRef]
16. Shamseer, L.; Moher, D.; Clarke, M.; Ghersi, D.; Liberati, A.; Petticrewet, M.; Shekelle, P.; Stewart, L.A. Preferred reporting items
for systematic review and meta-analysis protocols (prisma-p): Elaboration and explanation. BMJ 2015, 349, q7647. [CrossRef]
17. Methley, A.M.; Campbell, S.; Chew-Graham, C.; McNally, R.; Cheraghi-Sohi, S. PICO, PICOS and SPIDER: A comparison study of
specificity and sensitivity in three search tools for qualitative systematic reviews. BMC Health Serv. Res. 2014, 14, 579. [CrossRef]
18. Wan, X.; Wang, W.; Liu, J.; Tong, T. Estimating the sample mean and standard deviation from the sample size, median, range
and/or interquartile range. BMC Med. Res. Methodol. 2014, 14, 135. [CrossRef]
19. Murray, J.; Farrington, D.P.; Eisner, M.P. Drawing conclusions about causes from systematic reviews of risk factors: The Cambridge
Quality Checklists. J. Exp. Criminol. 2009, 5, 1–23. [CrossRef]
20. Wang, F.; Zhang, Y.; Yuan, Z.; Li, Y.; Liu, S.; Zeng, X.; Qiu, X.; Ye, L.; Huang, D. The association between iron status and thyroid
hormone levels during pregnancy. J. Trace Elem. Med. Biol. 2022, 74, 127047. [CrossRef]
21. Okuroglu, N.; Ozturk, A.; Özdemir, A. Is iron deficiency a risk factor for the development of thyroid autoantibodies in euthyroid
women with reproductive ages? Acta Endocrinol. 2020, 16, 49–52. [CrossRef]
22. Zhang, Y.; Huang, X.; Chen, Z.; Yang, Q.; Li, X.; Zhang, R.; Yu, Z.; Li, Q.; Liu, J.; Zha, B. Iron deficiency, a risk factor for thyroid
autoimmunity during second trimester of pregnancy in china. Endocr. Pract. 2020, 26, 595–603. [CrossRef] [PubMed]
23. Zhang, H.Y.; Teng, X.C.; Shan, Z.Y.; Wang, Z.J.; Li, C.Y.; Yu, X.H.; Mao, J.Y.; Wang, W.W.; Xie, X.C.; Teng, W.P. Association
between iron deficiency and prevalence of thyroid autoimmunity in pregnant and non-pregnant women of childbearing age:
A cross-sectional study. Chin. Med. J. 2019, 132, 2143–2149. [CrossRef] [PubMed]
24. He, L.; Shen, C.; Zhang, Y.; Chen, Z.; Ding, H.; Liu, J.; Zha, B. Evaluation of serum ferritin and thyroid function in the second
trimester of pregnancy. Endocr. J. 2018, 65, 75–82. [CrossRef]
25. Fu, J.; Yang, A.; Zhao, J.; Zhu, Y.; Gu, Y.; Xu, Y.; Chen, D. The relationship between iron level and thyroid function during the first
trimester of pregnancy: A cross-sectional study in Wuxi, China. J. Trace Elem. Med. Biol. 2017, 43, 148–152. [CrossRef]
26. Li, S.; Gao, X.; Wei, Y.; Zhu, G.; Yang, C. The Relationship between Iron Deficiency and Thyroid Function in Chinese Women
during Early Pregnancy. J. Nutr. Sci. Vitaminol. 2016, 62, 397–401. [CrossRef]
27. Veltri, F.; Decaillet, S.; Kleynen, P.; Grabczan, L.; Belhomme, J.; Rozenberg, S.; Pepersack, T.; Poppe, K. Prevalence of thyroid
autoimmunity and dysfunction in women with iron deficiency during early pregnancy: Is it altered? Eur. J. Endocrinol. 2016, 175,
191–199. [CrossRef]
28. Yu, X.; Shan, Z.; Li, C.; Mao, J.; Wang, W.; Xie, X.; Liu, A.; Teng, X.; Zhou, W.; Li, C.; et al. Iron deficiency, an independent risk
factor for isolated hypothyroxinemia in pregnant and nonpregnant women of childbearing age in China. J. Clin. Endocrinol.
Metab. 2015, 100, 1594–1601. [CrossRef]
29. Maldonado-Araque, C.; Valdés, S.; Lago-Sampedro, A.; Lillo-Muñoz, J.A.; Garcia-Fuentes, E.; Perez-Valero, V.; Gutierrez-
Repiso, C.; Goday, A.; Urrutia, I.; Peláez, L.; et al. Iron deficiency is associated with Hypothyroxinemia and Hypotriiodothy-
roninemia in the Spanish general adult population: [email protected] study. Sci. Rep. 2018, 8, 6571. [CrossRef]
30. Page, M.J.; McKenzie, J.E.; Bossuyt, P.M.; Boutron, I.; Hoffmann, T.C.; Mulrow, C.D.; Shamseer, L.; Tetzlaff, J.M.; Akl, E.A.;
Brennan, S.E.; et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ 2021, 372, n71.
[CrossRef]
31. Grabowski, M.K.; Gray, R.H.; Makumbi, F.; Kagaayi, J.; Redd, A.D.; Kigozi, G.; Reynolds, S.J.; Nalugoda, F.; Lutalo, T.; Wawer, M.J.; et al.
The Global Prevalence of Anaemia in 2011; WHO: Geneva, Switzerland, 2015; Volume 126, pp. 5409–5418. [CrossRef]
32. Köhrle, J. Selenium, Iodine and Iron-Essential Trace Elements for Thyroid Hormone Synthesis and Metabolism. Int. J. Mol. Sci.
2023, 24, 3393. [CrossRef] [PubMed]
33. Eftekhari, M.H.; Simondon, K.B.; Jalali, M.; Keshavarz, S.A.; Elguero, E.; Eshraghian, M.R.; Saadat, N. Effects of administration of
iron, iodine and simultaneous iron-plus-iodine on the thyroid hormone profile in iron-deficient adolescent Iranian girls. Eur.
J. Clin. Nutr. 2006, 60, 545–552. [CrossRef] [PubMed]
34. Ravanbod, M.; Asadipooya, K.; Kalantarhormozi, M.; Nabipour, I.; Omrani, G.R. Treatment of iron-deficiency anemia in patients
with subclinical hypothyroidism. Am. J. Med. 2013, 126, 420–424. [CrossRef] [PubMed]
Nutrients 2023, 15, 4790 19 of 19

35. Cinemre, H.; Bilir, C.; Gokosmanoglu, F.; Bahcebasi, T. Hematologic effects of levothyroxine in iron-deficient subclinical
hypothyroid patients: A randomized, double-blind, controlled study. J. Clin. Endocrinol. Metab. 2009, 94, 151–156. [CrossRef]
[PubMed]
36. Prezioso, G.; Giannini, C.; Chiarelli, F. Effect of Thyroid Hormones on Neurons and Neurodevelopment. Horm. Res. Paediatr.
2018, 90, 73–81. [CrossRef]
37. Allan, W.C.; Haddow, J.E.; Palomaki, G.E.; Williams, J.R.; Mitchell, M.L.; Hermos, R.J.; Faix, J.D.; Klein, R.Z. Maternal thyroid
deficiency and pregnancy complications: Implications for population screening. J. Med. Screen. 2000, 7, 127–130. [CrossRef]
38. Bothwell, T.H. Iron requirements in pregnancy and strategies to meet them. Am. J. Clin. Nutr. 2000, 72 (Suppl. S1), 257S–264S.
[CrossRef]
39. Zimmermann, M.B.; Burgi, H.; Hurrell, R.F. Iron deficiency predicts poor maternal thyroid status during pregnancy. J. Clin.
Endocrinol. Metab. 2007, 92, 3436–3440. [CrossRef]
40. Fuse, Y.; Ohashi, T.; Yamaguchi, S.; Yamaguchi, M.; Shishiba, Y.; Irie, M. Iodine status of pregnant and postpartum japanese
women: Effect of iodine intake on maternal and neonatal thyroid function in an iodine-sufficient area. J. Clin. Endocrinol. Metab.
2011, 96, 3846–3854. [CrossRef]
41. Orito, Y.; Oku, H.; Kubota, S.; Amino, N.; Shimogaki, S.; Hata, M.; Manki, K.; Tanaka, Y.; Sugino, S.; Ueta, M.; et al. Thyroid
function in early pregnancy in japanese healthy women: Relation to urinary iodine excretion, emesis, and fetal and child
development. J. Clin. Endocrinol. Metab. 2009, 94, 1683–1688. [CrossRef]
42. Haapala, A.M.; Hyoty, H.; Parkkonen, P.; Mustonen, J.; Soppi, E. Antibody reactivity against thyroid peroxidase and myeloperoxi-
dase in autoimmune thyroiditis and systemic vasculitis. Scand. J. Immunol. 1997, 46, 78–85. [CrossRef] [PubMed]
43. Weetman, A.P. The immunopathogenesis of chronic autoimmune thyroiditis one century after hashimoto. Eur. Thyroid. J. 2013, 1,
243–250. [CrossRef]
44. Guastamacchia, E.; Giagulli, V.A.; Licchelli, B.; Triggiani, V. Selenium and Iodine in Autoimmune Thyroiditis. Endocr. Metab.
Immune Disord. Drug Targets 2015, 15, 288–292. [CrossRef]
45. Ueda, N.; Takasawa, K. Impact of Inflammation on Ferritin, Hepcidin and the Management of Iron Deficiency Anemia in Chronic
Kidney Disease. Nutrients 2018, 10, 1173. [CrossRef]
46. Hernik, A.; Szczepanek-Parulska, E.; Filipowicz, D.; Abdolall, A.; Borowczyk, M.; Wrotkowska, E.; Czarnywojtek, A.; Krasiński, Z.;
Ruchała, M. The hepcidin concentration decreases in hypothyroid patients with Hashimoto’s thyroiditis following restoration of
euthyroidism. Sci. Rep. 2019, 9, 16222. [CrossRef]
47. Chao, K.C.; Chang, C.C.; Chiou, H.Y.; Chang, J.S. Serum Ferritin Is Inversely Correlated with Testosterone in Boys and Young
Male Adolescents: A Cross-Sectional Study in Taiwan. PLoS ONE 2015, 10, e0144238. [CrossRef] [PubMed]
48. Liu, Z.; Ye, F.; Zhang, H.; Gao, Y.; Tan, A.; Zhang, S.; Xiao, Q.; Zhang, B.; Huang, L.; Ye, B.; et al. The association between the
levels of serum ferritin and sex hormones in a large scale of Chinese male population. PLoS ONE 2013, 8, e75908. [CrossRef]
49. Zafar, U.; Ali, Z.; Khaliq, S.; Lone, K.P. Correlation of insulin resistance with neutrophil to lymphocyte ratio and serum ferritin in
male patients of metabolic syndrome. J. Pak. Med. Assoc. 2022, 72, 696–701. [CrossRef]
50. Nwabuobi, C.; Arlier, S.; Schatz, F.; Guzeloglu-Kayisli, O.; Lockwood, C.J.; Kayisli, U.A. hCG: Biological Functions and Clinical
Applications. Int. J. Mol. Sci. 2017, 18, 2037. [CrossRef]
51. World Health Organization. Serum Ferritin Concentrations for the Assessment of Iron Status and Iron Deficiency in Populations; World
Health Organization: Geneva, Switzerland, 2011. Available online: https://iris.who.int/handle/10665/85843 (accessed on
12 November 2023).
52. Moher, D.; Liberati, A.; Tetzlaff, J.; Altman, D.G.; The PRISMA Group. Preferred Reporting Items for Systematic Reviews and
Meta-Analyses: The PRISMA Statement. PLoS Med. 2009, 6, e1000097. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.