A comparative analysis of heme vs non-heme iron administration: a systematic review and meta-analysis of randomized controlled trials

Abstract

Background and purpose

Bioavailability studies and observational evidence suggest that heme iron (HI) may have greater impact on iron status indicators compared with non-heme iron (NHI). This systematic review and meta-analysis aimed to review the current evidence on the effect of the administration of HI compared with NHI for improving iron status in non-hospitalized population groups.

Methods

We searched Pubmed, CENTRAL, Scopus, Web of Science, and LILACS from inception to July 2024. There was no language restriction or exclusion based on age or iron status. Only randomized controlled trials comparing HI with NHI were considered. A random-effects meta-analysis was performed to compare the effect of treatments for iron status indicators and total side effects (including gastrointestinal side effects). We measured the certainty of the evidence (CoE) using GRADE assessment.

Results

After screening 3097 articles, 13 studies were included. Most of the interventions used HI in low doses combined with NHI. The meta-analysis showed higher hemoglobin increases in children with anemia or low iron stores receiving HI (MD 1.06 g/dL; 95% CI: 0.34; 1.78; CoE: very low). No statistically significant difference between interventions were found for any iron status indicator in the other population subgroups (CoE: very low). Participants receiving HI had a 38% relative risk reduction of total side effects compared to NHI (RR 0.62; 95% CI 0.40; 0.96; CoE: very low).

Conclusion

The current evidence comparing HI with NHI is very limited, preliminary findings suggest that interventions using HI may result in fewer side effects and may be superior in children with iron deficiency or anemia. However, given the very low certainty of the evidence, these results need further investigation through high-quality clinical trials.

Protocol registration

CRD42023483157.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00394-024-03564-y.

Introduction

Iron deficiency anemia (IDA) is a condition characterized by a reduction in hemoglobin levels, hindering the transport of oxygen in the tissues and the subsequent energy production of the body [1]. In low and middle-income countries (LMIC), it is considered a major public health problem that can have serious consequences for human health, especially during cognitive development [2, 3]. According to the Global Burden of Disease Study (2021), the global prevalence of anemia stands at an estimated 24.3% across all ages [4]. In LMIC, most affected subgroups are children between 6 and 59 months (59.8%) [5], pregnant women (45.2%) [6] and non-pregnant women of reproductive age (39.7%) [7].

The conceptual framework of IDA encompasses a multitude of underlying causes, including but not limited to factors such as low educational attainment, poverty, conflict, cultural norms, and health policies. These risk factors contribute to food insecurity and micronutrient deficiencies [8, 9]. In LMIC, inadequate iron intake is postulated as one of the leading causes of IDA [10], accounting for 66.2% of total anemia cases [4]. The current World Health Organization strategy is to act in multiple key outcomes, such as improving the consumption of specific micronutrients (iron, folate, vitamin B12) through dietary diversification, food fortification and supplementation [9].

Oral iron supplementation has proven to be an effective approach to addressing IDA [11]. However, an important barrier to treatment adherence is the occurrence of undesirable side effects [12, 13], which are more commonly associated with ferrous salts [14]. In this context, the use of micronutrients and food fortification has shown to be viable public health strategies to aid in the treatment and prevention of IDA [15–19]. However, problems with the lack of sustainable supply and manufacturing are found to be barriers to their proper implementation [20]. Exploring further interventions with lower production costs and comparable or superior efficacy to standard ferrous salt treatments while mitigating side effects becomes imperative. One of the possible options is heme iron (HI), a molecule found in the protoporphyrin IX ring in hemoglobin and myoglobin [21]; this compound has shown superior bioavailability compared to non-heme iron (NHI) [22–26] and positive results for hemoglobin regeneration both in intervention trials compared to placebo [27–29] and in large-scale interventions in Chile [30].

The current evidence about the comparative efficacy of HI with standard treatment (NHI orally or intravenous) in clinical settings has shown heterogeneous results [31–35]. However, in non-clinical settings, no systematic analysis has been performed to assess and compare the effect of HI with NHI on iron status. Hence, in this systematic review, we aim to evaluate the current evidence on the effect of HI administration compared with NHI for improving iron status indicators in non-hospitalized population groups, including individuals with and without anemia.

Methods

The study was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2020 guidelines [38] and the Cochrane Handbook for Systematic Reviews of Interventions [39]. A protocol was published in PROSPERO (International Prospective Register of Systematic Reviews database) under registration number CRD42023483157.

Search strategy

The following databases were searched: Pubmed, CENTRAL, SCOPUS, Web of Science (Core Collection) and LILACS (from inception to July 2024); we also included studies from other reviews using citation search strategy. The search strategy included terms related to heme iron supplementation, iron status and study design: (“Heme iron” OR “Iron”) AND (“Hemoglobin” OR “Ferritin” OR “Total iron binding capacity” OR “Serum iron” OR “Erythrocyte Protoporphyrin”) AND (“Randomized Controlled Trial” OR “Randomized Clinical Trial”). The detailed search strategy is presented in Online resource Table 1.

Eligibility criteria

The articles retrieved in the search strategy were uploaded to Rayyan, and three pairs of reviewers (AA, GAG, KCR, GDM, GRL and CDP) independently evaluated the eligibility of the studies by title and abstract screening. Afterwards, the texts of potentially suitable articles were obtained for additional assessment. Any disagreements were resolved by discussion with the seventh reviewer (MGR). Only randomized controlled trials that met the following criteria were included: (1) study participants of any population group (children, adolescents, pregnant women and women of reproductive age) that were allocated to an intervention that used HI in its composition (i.e. HI as a supplement or in fortified foods), (2) participants were allocated to a control group that included NHI (e.g. ferrous sulphate, fumarate, polymaltose iron).

The included studies needed at least one outcome that reflected iron status (hemoglobin, ferritin, total iron binding capacity, serum iron, transferrin receptor, transferrin saturation, erythrocyte or zinc protoporphyrin, erythrocyte indices, body iron). In addition, side effects, defined as the effect of a treatment that is in addition to or beyond its desired effect [40], were also included as a secondary outcome. The exclusion criteria were as follows: (1) Studies with hospitalized patients with some acute or chronic illness that are decompensated and/or uncontrolled, (2) studies whose objective was only to evaluate bioavailability by measuring iron absorption, (3) Comparator group using intravenous iron formulations. No study was excluded based on treatment duration.

Data extraction

Six authors (AA, GAG, KCR, GDM, GRL and CDP) independently extracted data from selected articles, protocols, and associated published appendices from each study. If a study presented values only in figures, WebPlotDigitizer v4 software was used to extract numeric information, this tool was used due to the high levels of reliability and validity [41]. When required, authors were contacted for additional details [42, 43]. In cases where studies included more than two experimental groups, consensus was reached to select the two most appropriate groups. A standardized form for data storage was developed, containing the author, year of publication, type of population, type of intervention, number of participants, follow-up period, eligibility criteria, and primary and secondary outcomes. The primary outcome sought in the studies was the final concentration of iron status indicators or its difference from baseline; the secondary outcome was the incidence of side effects related to iron administration.

Risk of bias assessment

Four authors (AA, GAG, GDM, MGR) independently assessed the quality of the studies according to the Revised Cochrane Risk-of-bias tool for randomized trials version 2.0. (RoB-2) [44]. The five domains considered were randomization process (1), deviations from the intended interventions (2), missing outcome data (3), measurement of the outcome (4) and selection of the reported result (5). The risk of bias was assessed at the level of each outcome, using the RoB-2 Excel tool algorithm [45], study outcomes were considered as low risk of bias if all domains scored as “low risk”; some concerns when at least one domain was labelled as “some concerns” and high risk when at least one domain scored as “high risk of bias”. Based on previous literature, to assess the comparative response to iron administration, an effect size of 1 g/dL for hemoglobin was considered clinically significant [46, 47].

GRADE assessment of certainty of the evidence

The Grading of Recommendations Assessment, Development and Evaluation (GRADE) evaluation was performed by two trained methodologists (AA and GDM) and a diriment (MGR) to assess the general CoE for each outcome presented in this systematic review. CoE was assessed based on the following criteria: (1) risk of bias, (2) inconsistency, (3) indirectness, (4) imprecision and (5) publication bias. The certainty of the evidence was categorized as very low, low, moderate, or high [48, 49].

Statistical analysis

We performed a meta-analysis of continuous outcomes for iron status indicators using mean differences (MD) with 95% confidence intervals (CI). We considered change from baseline for comparison between groups; for studies that did not report change from baseline, we followed Cochrane recommendations and analyzed only the endpoint values for comparison [50]. Study weights were assigned using the inverse of the variance method, and calculations were based on a random-effects model using the restricted maximum likelihood method.

We calculated the diversity adjusted required information size using the TSA software (Copenhagen Trial Unit, Centre for Clinical Intervention Research, Copenhagen) [51]. The analysis incorporated the effect sizes and the standard deviation (SD) from the pooled analysis of each population subgroup outcome, using 5% type I error rate and an 80% statistical power. The results from the analysis were considered to grade the imprecision of the evidence [52].

If an outcome was summarized as the median and interquartile range (IQR), the skewness was evaluated using Shi et al. (2023) method [46] and then converted to mean and SD using McGrath et al. (2020) method [47]. If only the IQR was presented as a difference of Q3-Q1 value, we obtained the SD using the Cochrane method [28]. When values were given as confidence intervals, we used the Review Manager calculator to obtain the SD [53]. We used relative risk (RR) with 95% confidence intervals to compare total side effects between groups. The significance level (α) adopted was 5%. For data transformation and statistical analysis, we used R version 4.2.2 through R Studio^®.

Heterogeneity was evaluated among the studies using the point estimates from the forest plot and secondarily using the Cochran Q test [50], a p-value below 0.05 was considered statistically significant for this analysis. Given the diversity of studies, we conducted a subgroup analysis to evaluate the effect of the type of population and their iron status on the outcome; a p-value below 0.05 was considered statistically significant for subgroup differences.

We conducted sensitivity analyses to assess the robustness of the findings. First, we used the leave-one-out method to evaluate potential sources of bias from individual studies. Second, we excluded studies that initially presented their values as medians and IQR, studies that included blood donors and studies with high risk of bias. Third, considering that two studies had two NHI groups (Palomino Quispe b and Hoppe b), we made separate meta-analyses with each control arm to assess whether the outcome changed within the population subgroup. Fourth, we performed a meta-analysis of medians using a Quantile Matching Estimation to account for the possible effect of the studies skewness on the overall result [54]. Publication bias was assessed primarily through funnel plot visual inspection and Egger’s test (statistically significant p-value < 0.1). Trim and fill method was applied as a sensitivity analysis to evaluate how the effect estimate changed after theoretical imputation of missing studies.

Results

A total of 3097 articles were retrieved from the databases and registers; two articles were identified using citation search. After duplicate removal, 2171 articles were screened, and 24 were considered for full-text review, from which 10 articles were excluded (Online resource Table 2), from the 14 articles left, 13 articles were included for the systematic review, and one was an ongoing trial (Fig. 1).

Fig. 1.

PRISMA flow diagram of trial selection

Included trials

The 13 trials included [42, 43, 55–65] accounted for a total of 910 participants. Six studies were conducted in adult population, including adult women of reproductive age (n = 273) [42, 57, 62–65] and men (n = 49) [56, 63], three studies in children (n = 154) [43, 55, 60], two in female adolescents (n = 200) [58, 59] and two in pregnant women (n = 232) [61, 62]. The trials were conducted in Norway [56, 61, 62, 64, 65], Sweden [57, 63], Mexico [55, 58, 59], United States [42], Guatemala [60] and Peru [43]. Most of the evidence was indirect as only one study administered HI as a single supplement (n = 1) [42], while the others combined HI with NHI formulations through supplements (n = 7) [43, 56, 61–65], or use dfortified foods with HI  (n = 5) [55, 57–60]. Baseline iron stores were different among studies, where participants initiate either with variable iron stores (low to moderate or normal levels) (n = 9) [42, 55, 57–63] or with iron deficiency (n = 4) [43, 56, 64, 65]. The follow-up intervention period ranged from 7 to 36 weeks. One ongoing trial was found, using HI polypeptide compared with ferrous salts in Gambian children aged 6 to 12 months with IDA [66]. Details about the characteristics of the included trials are provided in Table 1 and the summary of the outcomes are available in Tables 2 and 3.

Table 1.

Baseline characteristics of included randomized controlled trial (RCT) studies

References	Design	Population	Intervention	Control	Outcomes assessed	Main findings	Number of participants at baseline HI/NHI	Mean age, y	Follow-up
								HI/NHI
Palomino Quispe [43]	Parallel open	Children with mild or moderate anemia (previously dewormed)	Supplement: A total of 25,9 mg of elemental iron as heme iron + non-heme iron as microencapsulated ferric pyrophosphate in 5 g of food complement for home fortification	Supplement: (a) 3 mg/kg of elemental iron daily as Ferrous sulphatec,e (b) 12.5 mg of elemental iron as Ferrous fumarate in sprinkles for home fortification	MD: Hb RR: Total side effects (diarrhea, constipation) Baseline Hb – Mean ± SD g/dL) HI: 10.2 ± 0.8 NHI: (a) 10.1 ± 0.8 (b) 10.4 ± 0.5	The use of home food fortification with a HI supplement showed ↑ increases in hemoglobin compared with NHI in both ferrous sulphate and ferrous fumarate in the form of sprinkles	24/23	1.8/2.2	12 weeks
Quintero-Gutiérrez [55]	Parallel double-blind	Children with Hb between 11.5 and 13.5 g/dL (previously dewormed)	Fortified food: 4.2 mg of elemental iron as heme iron in biscuits daily	Fortified food: 4.2 mg of elemental iron as Ferrous sulphate in biscuits daily	MD: Hb, ferritin, MCH, HCT, MCV Baseline Hb—Mean ± SD g/dL HI: 12.8 ± 0.7 NHI: 12.9 ± 0.4	The use of food fortification based on HI and NHI showed equal effectiveness in increasing iron status indicators	15/19	5.6/5.6	10 weeks
Ulvik [56]	Parallel double -blind	Mostly adult women of reproductive age registering as blood donors for the first time with IDA (18%) or iron deficiency (82%), ferritin < 15 µg/La	Supplement: 27.6 of elemental iron as heme iron (3.6 mg) + ferrous fumarate (24 mg) + 1 control-like placebo tablet daily	Supplement: 100 mg of elemental iron as Ferrous sulphate + 3 intervention-like placebo tablets daily	MD: Hemoglobin, ferritin, serum iron, TIBC RR: Total side effects (gastrointestinal side effects) Baseline Hb—Median (95% CI) g/dL HI: 12.8 (12.3–13) NHI: 12.5 (12.1–12.8)	The HI supplement was as effective as NHI supplement for hemoglobin regeneration but not for ferritin Participants using HI supplements showed ↑ tolerance and ↓ side effects compared with NHI group	40/36	29/33	12 weeks
Hoppe [57]	Parallel open	Adult women of reproductive age	Fortified food: 27 mg of elemental iron as Heme iron in 75 g of crisp bread (based on rye flour) daily	Supplement: (a) 35 mg of elemental iron as Ferrous fumarate dailye (b) 60 mg of elemental iron as Ferrous fumarate daily	MD: Hemoglobin, ferritin, transferrin receptor, body iron RR: Total side effects (constipation, nausea, gastric pain, flatulence) Baseline Hb -Median (IQR) g/dL HI: 13.6 (0.8) NHI: (a) 13.6 (1.5) (b) 12.9 (1.2)	No statistically significant difference in iron status indicators between HI and NHI was observed	34/12 (a) and 15 (b)	24/22	12 weeks
Quintero-Gutiérrez [58]	Parallel double-blind	Female adolescents	Fortified food: -First 4 weeks: 9.6 mg of elemental iron as heme iron in biscuits -Next 9 weeks: 10.8 mg of elemental iron as heme iron in biscuits	Fortified food: -First 4 weeks: 9.6 mg of elemental iron as ferrous sulphate in biscuits -Next 9 weeks: 10.8 mg of elemental iron as ferrous sulphate in biscuits	MD: Hemoglobin, ferritin, serum iron, MCH, HCT, MCV Baseline Hb—Mean ± SD g/dL HI: 13.7 ± 0.6 NHI: 13.7 ± 0.6	Food fortification based on HI showed statistically significant ↑ values in hemoglobin compared with NHI group	59/64	14/13.7	13 weeks
González-Rosendo [59]	Parallel double-blind	Female adolescents	Fortified food: Biscuits with heme iron -First 16 days: 9.6 mg of elemental iron as heme iron in biscuits -Next 21 days: 10.8 mg of elemental iron as heme iron in biscuits	Fortified food: Biscuits with heme iron -First 16 days: 9.6 mg of elemental iron as ferrous sulphate in biscuits -Next 21 days: 10.8 mg of elemental iron as ferrous sulphate in biscuits	MD: Hemoglobin, ferritin, serum iron, MCH, HCT, MCV Baseline Hb—Mean ± SD g/dL HI: 13.6 ± 0.8 NHI: 13.5 ± 0.9	Food fortification based on HI and NHI showed similar effects at increasing hemoglobin, while HI group showed a better preservation of ferritin levels	40/37	14.6/14.6	7 weeks
Swaind [42]	Parallel double-blind	Adult women of reproductive age with ferritin < 53 µg/L after blood phlebotomy	Supplement: 5 mg of elemental iron as heme iron in gelatin capsules daily	Supplement: 50 mg of elemental iron as Ferrous sulphate monohydrate in wheat rolls daily	MD: Hemoglobin, ferritin, TIBC, ZPP, transferrin receptor and saturation, body iron Baseline Hb (NA)	The use of NHI as ferrous sulphate improved iron status indicators, but a very low dose of HI showed no effect on iron status	9/12	40b	12 weeks
Schuman [60]	Parallel double-blind	Children with Hb between 10 and 11.5 g/dL (previously dewormed)	Fortified food: 35 mg of elemental iron as heme iron (31.2 mg) in 156 g cans of black beans (P.Vulgaris) + non-heme iron from beans (3.7 mg) 5 times per week	Fortified food: 35 mg of elemental iron as Ferrous sulphate hydrated (31.2 mg elemental iron) in 156 g cans of black beans (P.Vulgaris) + non-heme iron from beans (3.7 mg) 5 times per week	MD: Hemoglobin, ferritin Baseline Hb—Mean ± SD g/dL HI: 10.9 ± 0.5 NHI: 10.9 ± 0.5	The use of food fortification based on HI and NHI showed an equal effect to increase iron status indicators; pos hoc analysis showed a statistically significant higher increase in hemoglobin for the HI group in children with low ferritin levels	36/37	1.7/1.7	10 weeks
Sandstad [61]	Parallel open	Pregnant women	Supplement: 27.6 mg of elemental iron as Heme 3.6 mg of iron + ferrous fumarate (24 mg) daily	Supplement: 60 mg of elemental iron as Ferrous sulphate daily	MD: Hemoglobin, ferritin Baseline Hb—Mean ± SD g/dL HI: 12.6 ± 0.8 NHI: 12.5 ± 0.8	An iron supplementation regime based on baseline iron status resulted in adequate iron status for both HI and NHI in pregnant women	94/77	29.1b	From early pregnancy (10–14 weeks) to 6 weeks postpartum, the treatment period assumed to be between 32 to 36 weeks
Eskeland [62]	Parallel double-blind	Pregnant women	Supplement: 27.6 mg of elemental iron as Heme iron (3.6 mg) + ferrous fumarate (24 mg) daily	Supplement: 27 mg of elemental iron as Ferrous fumarate + 100 mg of ascorbic acid daily	MD: Hemoglobin, ferritin, serum iron, transferrin saturation, erythrocyte protoporphyrin, MCV Baseline Hb (Median (IQR) g/dL) HI: 12.5 (12.1–13.4) NHI: 12.5 (12.2–13.2)	A daily dose of a supplement containing HI prevents the depletion of iron stores with similar efficacy as NHI in pregnant women	31/30	28/26	From early pregnancy (< 13 weeks) to 24 weeks post-partum, only 18 weeks corresponding to the treatment period were considered
Frykman [63]	Parallel double-blind	Adult women of reproductive age and men (blood donors)a	Supplement: 18.4 mg of elemental iron as Heme iron (2.4 mg) + ferrous fumarate (16 mg) daily	Supplement: 60 mg of elemental iron as Ferrous fumarate daily	MD: Hemoglobin, ferritin RR: Total side effects (diarrhea, constipation, gastric pain, nausea) Baseline Hb—Median (95% CI) g/dL HI: 12.6 (12.4–13.6) NHI: 13.3 (12.8–13.7)	No changes in iron status indicators were observed in participants receiving HI or NHI Side effects were statistically significant lower in the group receiving the supplement containing HI	49/48	44.5/45	12 weeks
Borch-Iohnsen [64]	Parallel open	Adult women of reproductive age (blood donors and non-blood donors with low ferritin < 20 µg/L)	Supplement: 18 mg of elemental iron as Heme iron (2 mg) + ferrous fumarate (16 mg) daily	Supplement: 20 mg of elemental iron as ferrous fumarate + 120 mg of ascorbic acid daily	MD: Hemoglobin, ferritin Baseline Hb—Mean (group) g/dL, SD not reported HI: 13.3 (BD) NHI: 13.0 (BD) HI: 13.4 (NBD) NHI: 13.4 (NBD)	Both HI and NHI supplements showed similar effects at increasing hemoglobin in both BD and NBD; ferritin increases were ↑ in BD receiving HI	(a) BD 16/18 (b) NBD 11/11	30–50	20 weeks
Borch-Iohnsen [65]	Parallel open	Women of reproductive age (blood donors and non-blood donors with low ferritin < 20 µg/L)	Supplement: 18 mg of elemental iron as Heme iron (2 mg) + ferrous fumarate (16 mg) daily	Supplement: 20 mg of elemental iron as ferrous fumarate + 120 mg of ascorbic acid daily	MD: Hemoglobin, ferritin, TIBC Baseline Hb – Mean ± SD g/dL HI: 13.4 ± 0.8 NHI: 13.4 ± 0.8	Both HI and NHI supplements showed similar effects at increasing ferritin values and decreasing TIBC; no changes in hemoglobin were observed because that population was not with IDA, only with iron deficiency	14/14	30b	24 weeks

BD Blood donor, dL Deciliter, g Gram, Hb Hemoglobin, HI Heme iron, kg Kilogram, MCH Mean corpuscular haemoglobin, MCV Mean corpuscular volume, MD Mean difference, mg Milligram, NA Not available, NBD Non-blood donor, NHI Non-heme iron, RR Risk ratio, TIBC Total iron binding capacity, ↑ Higher, ↓ Lower

^aFor Frykman study only the women cohort was considered for the main iron status indicators meta-analysis, for Ulvik trial no separated data of only women was available. However, as only three men were in the analysis, the subgroup of women of reproductive age was maintained as a pragmatic approach

^bMean age of both groups

^cSide effects were not measured in this group

^dHemoglobin outcome not available

^eIntervention groups considered for the main analysis

Table 2.

Summary of other iron status indicators

Iron status indicators	Overall analysis (Random-effects model)
	n	Estimate MD (95% CI)	I2	p1	p2
Total iron binding capacity (µmol/L)	3	2.72 (− 0.46; 5.90)	0%	0.09	0.69
Serum iron (µg/dL)	3	1.18 (− 5.98; 8.33)	0%	0.75	0.82
Transferrin receptor (nmol/L)	2	− 0.59 (− 5.88; 4.70)	74%	0.83	0.05
Mean corpuscular hemoglobin (pg/cell)	3	0.63 (− 0.06; 1.32)	31%	0.07	0.23
Mean corpuscular volume (fL)	4	0.51 (− 0.67; 1.69)	43%	0.40	0.16
Hematocrit (%)	3	− 0.47 (− 1.06; 0.11)	0%	0.11	0.51
Body iron (mg/kg)	2	− 0.66 (− 1.79; 0.48)	0%	0.26	0.74

p¹ p-value for effect, p² value for heterogeneity, n Number of studies, MD Mean difference, CI Confidence interval, I² Statistic assessment of heterogeneity, % Percentage, g Gram, µg Micrograms, L Liter, nmol Nanomole, fL Femtoliter, mg Milligram, kg Kilogram, “- “ No data available or not measured

Table 3.

Summary of findings (SoF) of GRADE evaluation

Population	Outcome	Relative effect (95% CI)	Anticipated absolute effects (95% CI)		No of participants (studies)	Certainty of the evidence (GRADE)
			Risk difference with heme iron	95% CI
Overall children population	Hemoglobin	–	MD: 0.53 g/dL	(− 0.40 to 1.47)	135 (3 RCT’s)	⨁ ⊝  ⊝  ⊝  Very low a,b
	Ferritin	–	MD: 3.47 µg/L	(− 2.25 to 9.19)	88 (2 RCT’s)	⨁ ⊝  ⊝  ⊝  Very low c,d
Female adolescents	Hemoglobin	–	MD: 0.36 g/dL	(− 0.32 to 1.05)	200 (2 RCT’s)	⨁ ⊝  ⊝  ⊝  Very low e,f
	Ferritin	–	MD: -1.08 µg/L	(− 8.92 to 6.76)	200 (2 RCT’s)	⨁ ⊝  ⊝  ⊝  Very low g,h
Pregnant women	Hemoglobin	–	MD: 0.07 g/dL	(− 0.41 to 0.54)	225 (2 RCT’s)	⨁ ⊝  ⊝  ⊝  Very low i,j
	Ferritin	–	MD: -1.62 µg/L	(− 15.4 to 12.1)	227 (2 RCT’s)	⨁ ⊝  ⊝  ⊝  Very low e,k
Overall population of adult women of reproductive age	Hemoglobin	–	MD: 0.04 g/dL	(− 0.19 to 0.27)	240 (5 RCT’s)	⨁ ⊝  ⊝  ⊝  Very low i,l
	Ferritin	–	MD: 0.84 µg/L	(− 3.03 to 4.72)	257 (6 RCT’s)	⨁ ⊝  ⊝  ⊝  Very low m,n
Adult women of reproductive age, men and children	Total side effects	RR 0.62 (0.40 to 0.96)	14 fewer events per 100 participants	(from 22 to 2 fewer)	251 (4 RCT’s)	⨁ ⊝  ⊝  ⊝  Very low a,o

CI confidence interval, RCT Randomized controlled trial, GRADE Grading of Recommendations Assessment, Development, and Evaluation

^aRisk of bias: High risk of bias due to missing outcome data domain in ≥ 25% studies (33% for hemoglobin in children studies and 25% for total side effects). Downgraded one level

^bImprecision: Number of participants < 30% of the Diversity-adjusted required information size of 867 participants. Downgraded two levels

^cRisk of bias: High risk of bias due to missing outcome data domain in ≥ 25% studies (50% in children studies). Downgraded one level

^dImprecision: Number of participants < 30% of the Diversity-adjusted required information size of 489 participants. Downgraded two levels

^eInconsistency: Heterogeneity in the point estimates of the forest plot could not be explained by subgroup analysis of iron status due to low number of studies. Downgraded one level

^fImprecision: Number of participants < 30% of the Diversity-adjusted required information size of 1490 participants. Downgraded two levels

^gIndirectness: Uncertainty of the generalizability of the findings due to the chronic inflammation status of the population assessed that yields high baseline ferritin levels. Downgraded one level

^hImprecision: Number of participants < 30% of the Diversity-adjusted required information size of 21,946 participants. Downgraded two levels

ⁱIndirectness: HI used in low proportion and mixed with ferrous salts. Downgraded one level

^jImprecision: Number of participants < 30% of the Diversity-adjusted required information size of 6620 participants. Downgraded two levels

^kImprecision: Number of participants < 30% of the Diversity-adjusted required information size of 10,756 participants. Downgraded two levels

^lImprecision: Number of participants < 30% of the Diversity-adjusted required information size of 12,069 participants. Downgraded two levels

^mInconsistency: Heterogeneity in the point estimates was not resolved after only considering data from women with iron deficiency. Downgraded one level

ⁿImprecision: Number of participants < 30% of the Diversity-adjusted required information size of 4682 participants. Downgraded two levels

^oImprecision: Number of participants < 30% of the Diversity-adjusted required information size of 517 participants. Downgraded two levels

Risk of bias assessment

The overall risk of bias assessment showed that most study outcomes presented some concerns due to a lack of available study protocols to assess selection of reported results [43, 55, 56, 58–64, 64]. Other studies presented a high risk of bias because outcomes reported to be measured in methodology could not be found in the manuscript or with data requests to the authors [42, 62]. Some studies presented some concerns in the randomization process since the precise randomization methods were not clearly described [42, 43, 57, 61, 62, 64]. A high risk of bias due to missing outcome data was also present; two studies excluded or lost participants due to infections or intolerance, and presented unequal proportions of participants in their groups at endpoint [55, 57]. Some studies showed some concerns in deviations from intended interventions arising from manufacturing problems or using an open-label design [43, 57, 61, 65] (Online resource Table 3).

Primary outcomes

Hemoglobin

All included studies reported this outcome [42, 43, 55–65]. However, the study by Swain et al. [42] did not provide quantitative data, instead stating that no difference was observed between the groups [42]. The meta-analysis showed no statistically significant difference comparing HI with NHI (MD 0.22 g/dL; 95% CI: − 0.05; 0.48; I² = 80%; Fig. 2). The subgroup analysis of the overall population showed no difference between treatments for children (MD 0.54 g/dL; 95% CI: − 0.39; 1.46; I² = 90%), female adolescents (MD 0.36 g/dL; 95% CI: − 0.32; 1.05; I² = 89%), pregnant women (MD 0.07 g/dL; 95% CI: − 0.41; 0.54; I² = 70%) and women of reproductive age (MD 0.04 g/dL; 95% CI: − 0.19; 0.27; I² = 0%). However, point estimates were heterogeneous for children, female adolescents, and pregnant women population subgroups. The test for subgroup differences was not statistically significant for this analysis (p = 0.64; Fig. 2).

Fig. 2.

Forest plot comparing hemoglobin mean difference between HI vs NHI across population subgroups

Conversely, when comparing subgroups based on the iron status of children and women of reproductive age, the analysis revealed a statistically significant difference among children (p < 0.01; Fig. 3), showing a clinically and statistically significant higher increase in hemoglobin levels favoring HI in children with anemia or low iron stores (MD 1.06 g/dL; 95% CI: 0.34; 1.78; I² = 78%, p = 0.0001; Fig. 3) compared with children with moderate to normal iron stores or without anemia (MD − 0.19 g/dL; 95% CI: − 0.72; 0.34; I² = 0%, p = 0.48; Fig. 3). However, no statistically significant subgroup differences were found comparing women with low iron stores with women with low to moderate iron stores (p = 0.37; Fig. 3). The CoE was very low for this outcome in all population subgroups (Table 3 Summary of findings).

Fig. 3.

Forest plot comparing hemoglobin mean difference between HI vs NHI across population subgroups according to iron status

Ferritin

This outcome was assessed in 12 studies [42, 55–65], the overall analysis showed no statistically significant difference between HI and NHI in ferritin levels (MD 0.71 µg/L; 95% CI: − 2.18; 3.60; I² = 46%; Fig. 4). The subgroup population analysis showed no difference between treatments for children (MD 3.47 µg/L; 95% CI: − 2.25; 9.19; I² = 0%, female adolescents (MD − 1.08 µg/L; 95% CI: − 8.32; 6.76; I² = 26%), pregnant women (MD − 1.62 µg/L; 95% CI: − 15.38; 12.14; I² = 86%) and women of reproductive age (MD 0.84 µg/L; 95% CI: − 3.03; 4.72; I² = 52%). However, heterogeneity in the point estimates was found for pregnant women subgroup. No statistically significant subgroup differences were found among population subgroups (p = 0.77; Fig. 4). Similarly, no statistically significant difference was observed in the analysis according to iron stores for neither children (p = 0.95; Fig. 5) nor women (p = 0.38; Fig. 5). The CoE was very low for this outcome in all population subgroups (Table 3 Summary of findings).

Fig. 4.

Forest plot comparing ferritin mean the difference between HI vs NHI in the general population (A) and in the population subgroup with iron deficiency (B)

Fig. 5.

Forest plot comparing ferritin mean difference between HI vs NHI across population subgroups according to iron status

Other iron status indicators

Erythrocyte indices

Data for mean corpuscular hemoglobin (MCH) and hematocrit (HCT) were available in three studies [55, 58, 59], one in children [55] and two in female adolescents [58, 59]. The pooled analysis showed no difference between HI and NHI for MCH (MD 0.63 pg/cell; 95% CI: − 0.06; 1.32; I² = 31%; Table 2) nor HCT (MD − 0.47%; 95% CI: − 1.06; 0.11; I² = 0%; Table 2). Likewise, the subgroup analysis (Online resource Table 4) showed no difference in children for MCH (MD 0.50 pg/cell; 95% CI − 1.18; 2.18) and HCT (MD − 1.30%; 95% CI − 2.83; 0.23) nor in female adolescents for MCH (MD 0.65 pg/cell; 95% CI − 0.23; 1.53; I² = 65%) and HCT (MD − 0.33%; 95% CI − 0.97; 0.30; I² = 0%). The CoE for MCH and HCT was very low for both outcomes in the two populations (Online resource Tables 7 and 8).

The outcome of MCV was assessed in 4 studies [55, 58, 59, 62], two in female adolescents [58, 59], one in children [55] and one in pregnant women [62]. The pooled analysis showed no difference between HI and NHI for MCV (MD 0.51 fL; 95% CI: − 0.67; 1.69; I² = 43%). Similarly, the subgroup analysis (Online resource Table 4) showed no difference in female adolescents (MD 1.38 fL; 95% CI: − 0.43; 3.19; I² = 25%), children (MD − 0.30 fL; 95% CI: − 1.21; 0.61) and pregnant women (MD 0.60 fL; 95% CI: − 1.55; 2.75). The CoE was found to be very low for all populations subgroups (Online resource Tables 7–9).

Total iron binding capacity (TIBC)

Data from this outcome were reported in three studies involving women of reproductive age [42, 56, 65]. The pooled analysis showed no difference between HI and NHI (MD 2.72 µmol/L; 95% CI: − 0.46; 5.90; I² = 0%) (Table 2; more details in Online resource Table 4). The CoE for this outcome was rated as very low (Online resource Table 10).

Serum iron

This outcome was reported in four studies [56, 58, 59, 62], one in women of reproductive age [56], two in female adolescents [58, 59] and one in pregnant women [62], but no quantitative data was available for the latter (See narrative description in Online resource Table 9). The pooled analysis showed no difference between HI and NHI regarding serum iron (MD 1.18 µg/dL; 95% CI: − 5.98; 8.33; I² = 0%; Table 2). Similarly, there was no difference neither for women of reproductive age (MD 7.60 µg/dL; 95% CI: − 13.68; 28.88; Online resource Table 4) nor female adolescents (MD 0.36 µg/dL; 95% CI: − 7.24; 7.96; I² = 0%; Online resource Table 4). The CoE was very low for this outcome in women of reproductive age (Online resource Table 10) and female adolescents (Online resource Table 8).

Transferrin outcomes

Two studies in women of reproductive age analyzed the transferrin receptor outcome [42, 57]; the pooled analysis showed no difference between HI and NHI regarding transferrin receptor reduction (MD − 0.59 nmol/L; 95% CI: − 5.88 to 4.70; I² = 74%; Table 2, more details in Online resource Table 4). Two studies reported transferrin saturation for women of reproductive age [42] and pregnant women [62], but no quantitative data was available (See narrative description in Online resource Tables 9 and 10). The CoE was very low for all the outcomes (Online resource Tables 9 and 10).

Body iron

This outcome was available in two studies in women of reproductive age [42, 57]; the pooled analysis showed no difference between HI and NHI regarding body iron increment (MD − 0.66 mg/kg; 95% CI: − 1.79; 0.48; I² = 0%) (Table 2). The CoE for this outcome was very low (Online resource Table 10).

Erythrocyte and zinc protoporphyrin (ZPP and EPP)

One study in the pregnant women population assessed EPP [58], and the authors reported no difference between treatments (MD 0.08 µmol/L; CI 95% − 0.17;0.32; Online resource Table 4). ZPP outcome was available in one women of reproductive age study [42]; the authors reported that NHI showed a statistically significant reduction from the baseline of ZPP in comparison to HI in (MD 28 µg/L; CI 95% 0.73; 57.27), no p-value was reported (Online resource Table 4). The CoE was very low for EPP (Online resource Table 9) and ZPP (Online resource Table 10).

Secondary outcomes

Total side effects

Six studies provided results on side effects [43, 55–57, 62, 63], Two of these studies reported their results qualitatively [55, 62]. One study reported no side effects from using HI and NHI-fortified foods in children [55], while another reported similar side effects in both groups of pregnant women [62]. In contrast, a total of four studies [43, 56, 57, 63] reported quantitative data on the incidence of side effects, three in adults (men and women) [56, 57, 63] and one in children [43]. The reported outcomes were mainly gastrointestinal side effects, such as constipation and diarrhea (see more detail in Online resource Table 5). The pooled risk ratio indicated that participants receiving HI had a 38% relative risk reduction of total side effects compared to those receiving NHI (ferrous sulphate and ferrous fumarate in the form of sprinkles) (RR 0.62; CI 95% 0.40; 0.96; I²:14%; p = 0.031) (Fig. 7). The CoE for this outcome was very low (Table 3 Summary of findings).

Sensitivity analysis

The leave-one-out analysis indicated that excluding any individual study did not affect the pooled results for hemoglobin (Online resource Fig. 2a) and ferritin (Online resource Fig. 2b). However, the omission of the Palomino Quispe (2024) (a) study led to a noticeable reduction in heterogeneity, dropping from 80 to 60% in the MD of hemoglobin (Online resource Fig. 2a).

The leave-one-out analysis for total side effects RR showed that all studies exclusion except Hoppe (2013) changed the pooled outcome to a null effect (Online resource Fig. 3).

Following the exclusion of potentially skewed studies, studies with a high risk of bias, and studies with blood donors, there were no changes in the pooled outcome for hemoglobin and ferritin. Similarly, performing a meta-analysis of medians showed no changes in the overall analysis of both outcomes (Online resource Table 6). The comparator exchanges showed no difference for the overall outcome in children (Online resource Fig. 4a) and women of reproductive age (Online resource Fig. 5a and b). However, when using Palomino Quispe (b) sprinkles comparator, the subgroup outcome of hemoglobin in children with iron deficiency became not clinically significant (MD 0.94 g/dL; 95% CI: 0.49; 1.39; I²:46%; p < 0.01; Online resource Fig. 4b). Finally, no changes in the direction of the result were found after including men data from Frykman et al. study in the hemoglobin and ferritin analysis (Online resource Fig. 6a and b).

Fig. 6.

Risk ratio of side effects comparing HI vs NHI iron administration

Publication bias

After visual inspection, we found evidence of funnel plot asymmetry for hemoglobin (Online resource Fig. 7a) and ferritin outcomes (Online resource Fig. 7b). However, Egger’s test showed no statistical funnel plot asymmetry for hemoglobin (p = 0.40) and ferritin (p = 0.48). The Trim and Fill method showed that after the theoretical imputation of three studies, there was a statistically significant difference for hemoglobin in favor of heme iron (MD 0.46 g/dL 95% CI: 0.19; 0.72; I²: 84.6%; p = 0.0008; Online resource Fig. 7c), but not clinically significant (< 1 g/dL). For ferritin, the imputation of one study showed no statistically significant difference in the overall result (MD: 1.37 µg/L; 95% CI: − 1.64; 4.38; I²: 53%; p = 0.37; Online resource Fig. 7d).

Discussion

This systematic review found that the pooled analysis of the current evidence suggests that interventions using HI may have comparable effects to NHI for improving iron status. However, the overall effect may be underpowered and influenced by the iron status and physiological variability from the assessed population subgroups. Our preliminary findings indicate that children with iron deficiency or anemia may show a higher hemoglobin increase after HI administration compared with NHI. The evidence assessing the effects of HI on pregnant women and female adolescents was highly inconsistent to make any preliminary claim. Considering the very low CoE, our confidence in these effects is limited.

Our overall effect estimates are partially aligned with previous research on other population groups not considered for this analysis. In patients with chronic kidney disease [31, 32, 35], similar responses to treatment between HI and NHI (either orally or intravenous) have been reported. However, as explained by Dull et al. [67], the methodological concerns related to the low statistical power in those trials limits the establishment of equivalence between interventions [67]. Similarly, in pregnant women HI administration compared with intravenous NHI presented similar effects on hemoglobin and ferritin improvement, but the authors did not properly report if their sample size was sufficient to establish equivalence [33]. In contrast, Mischler et al. (2017) reported no efficacy of HI in treating iron deficiency in women following roux-en-y gastric bypass surgery, the authors hypothesized that the low solubility of HI polypeptide might be an explanation for their findings, but no further trials to confirm those results have been conducted [34].

Nevertheless, our population subgroup analysis revealed that this equivalence pattern was not homogeneous in all the population subgroups. In children with iron deficiency, part of the higher hemoglobin increases observed after  receiving HI might be explained by the differing absorption profiles of HI and NHI and the role of iron stores on the regulatory mechanisms of iron absorption. While the precise molecular pathway of HI absorption is currently not fully understood [68], clinical evidence shows higher HI absorption (15–35%) compared with NHI (2–20%) [69–71], especially in iron deficiency states [72, 73]. This can be explained considering the more extended mechanisms of NHI absorption, which require a reduction process by duodenal cytochrome b (Dcytb) or other reductant agents (ascorbic acid, citric acid), followed by the uptake through the divalent metal transporter 1 (DMT-1). After absorption, free iron obtained from both NIH and HI is transported into plasma by ferroportin; this latter process is upregulated in iron deficiency and downregulated in inflammatory conditions by hepcidin [74]. However, this finding was not observed in the subgroup of women of reproductive age with iron deficiency, the main factors that could explain these differences could be related to the low dose of HI employed in those studies, and the insufficient iron status contrast in the subgroups, as only a low number of women presented anemia.

Another factor to consider for the findings is treatment adherence. As evidenced in our sensitivity analysis, using iron supplements at lower doses (sprinkles as NHI comparator) in children showed similar results as the result using higher doses of ferrous sulphate as comparator, suggesting that the effect might depend more on other factors than solely iron content in this population. This is confirmed by many trials where sprinkles formulations at low iron doses showed fewer side effects and similar efficacy for improving iron status compared with ferrous salts [75–77]. However, the success of those strategies is highly dependent on age, educational level, socioeconomic status [78, 79] and the use of previous educational strategies to ensure adherence [80, 81]. In that line, the findings by Palomino Quispe [43] and Schuman [60] may be influenced by the socio-economic context of the population studied, and the effect is plausible to be different in other social contexts [62].

The inconsistency observed for hemoglobin in children was resolved by our subgroup analysis after only considering children with iron deficiency, stabilizing the point estimates. However, due to the low number of studies per population group, other sources of heterogeneity for the outcomes of hemoglobin in female adolescents, and ferritin in women of reproductive age and pregnant women could not be properly resolved. Possible explanations of heterogeneity in adolescent studies could be related to different treatment duration of iron supplementation; the additional six weeks from Quintero-Gutierrez [58] study could have affected hemoglobin values and altered the final mean differences on the basis that hemoglobin increase rate is estimated to be 0.1 g/dL per day [82]. This heterogeneity was not observed for ferritin, possibly because this adolescent population presented chronic elevation of this indicator due to parasite infection and inflammation [83].

For women of reproductive age and pregnant women, the high heterogeneity observed in ferritin levels could be explained by the different doses of NHI comparator used between studies. Serum ferritin increases have shown to be highly dependent on the amount of iron dose applied [84–86]. The studies by Ulvik [56] and Sandstad [45] used a higher dose of NHI compared to Borch-Iohnsen [64, 65] and Eskeland respectively [62]. In both scenarios, the NHI groups with larger doses presented higher ferritin increases.

Regarding side effects, as most reported disturbances were gastrointestinal, there is coherence with the proposed mechanisms for NHI supplementation-induced generation of free radicals in the colon through Haber–Weiss and Fenton-type reactions reported in clinical studies [87, 88], as well as alterations in the microbiota [89–91]. For HI, clinical evidence shows generation of metabolites related to inflammation and oxidative stress, which include N-nitroso compounds and end products of lipid peroxidation [92–96]. Preclinical evidence suggests that HI induces gut dysbiosis [97] and increase free radicals from heme degradation [98, 99].

As both interventions have biological plausibility to induce side effects, the lower incidence observed for HI interventions may be attributed to the dosing of elemental iron. Due to the expected higher bioavailability, most studies administered low doses of HI, potentially decreasing the amount of iron reaching the colon. Similarly, the dose of NHI was low, ranging from 35 to 60 mg, except for Ulvik [56] study, which used 100 mg. As shown in the forest plot analysis, the only study that individually presented statistically significant findings with the CI assessment was Ulvik study. This suggests that HI might exhibit a safer profile when compared to higher NHI dosage, this hypothesis should be formally tested in further trials.

Nonetheless, some safety concerns have been raised by the European Food Safety Authority (EFSA) when HI supplements based on blood peptonates are directed to the general population at high doses (relative to context of supplemental intake) [100]. These concerns are based on evidence from cohort studies [101] and controlled feeding trials [92, 93, 96, 102], suggesting a contributing role of HI to the relationship between red meat consumption and colorectal carcinogenesis [103]. However, it is important to highlight that these safety concerns  have not been specifically addressed in the context of IDA. There is uncertainty on how much of the long-term outcomes are entirely attributable to only HI or to the  synergistic effect of the food matrix where HI is delivered and the time of exposure. As HI is offered as a dietary supplement [104, 105] and not a drug, no long-term safety studies have been conducted. Consequently, the administration of this intervention should be addressed using a risk–benefit analysis that considers the burden of IDA and the socioeconomic and food availability context of the target population.

Implications for practice and policy

Our findings show that the current evidence of the use of HI, while promising in some age groups, remains limited. Future ongoing trials will be crucial in advancing understanding in this area. Moreover, IDA is a multifactorial problem that involves more determining factors that vary according to the population and context assessed [106–108]. Aspects related to the food environment, such as food availability, socio-economic status and educational level, are also important key aspects to address. Policymakers should consider an evidence-based framework that emphasizes changes to all determinants rather than solely focusing on iron supplementation. In addition, further research is necessary to address the concerns about the proper dosing of HI supplements according to the population group and their condition. Safety studies evaluating the health effects of long-term consumption of HI and comprehensive risk–benefit analyses using causal inference methods with the totality of the evidence are also required for future guidelines. On the other hand, if the use of HI supplements and food fortifiers is considered, food technology aspects and the development of adequate food systems will need to be addressed to ensure optimal and sustainable production, considering aspects such as sensory quality, shelf-life, food safety and the use of proper food vehicles [109].

Strengths and limitations

This systematic review has several strengths, to the best of our knowledge, this is the first meta-analysis that addresses the comparative effect between HI and NHI in the context of randomized controlled trials. We ensured a proper comparison by including studies where both interventions were administered orally. The risk of bias was systematically measured at the level of each outcome, enabling us to distinguish the bias of the effect according to each iron status indicator. The GRADE assessment offered valuable insights into the current CoE. Additionally, subgroup analyses based on population and iron status conditions helped to elucidate treatment response variations among different populations.

This review has limitations that are worth mentioning; first, the indirectness of the evidence, particularly regarding research on adult women of reproductive age and pregnant women. Many of the interventions in these trials did not use HI as the main compound, leading to uncertainty about the actual effect of heme iron-only therapies in this population. Second, none of the included studies provided a predefined analysis plan, which could have decreased bias due to selection of the reported results. Third, conflicts of interest could not be objectively measured; thus, results from trials funded by manufacturers of commercially available HI products may be susceptible to unmeasured bias.

Conclusion

The current evidence comparing HI with NHI is very limited, preliminary findings suggest that interventions using HI may result in fewer side effects and may be superior in children with iron deficiency. However, given the very low certainty of the evidence, these results need further investigation through high-quality clinical trials in multiple socioeconomic contexts and population groups.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

Instituto de Investigación Nutricional.

Data availability

The datasets supporting the main analyses are presented within the manuscript and figures. Additional details are available upon request from the authors.