Efficacy and safety of iron supplements for restless leg syndrome, a systematic review, meta-analysis, meta-regression, and trial sequential analysis of randomized controlled trials

Abstract

Background

Restless leg syndrome (RLS) is a common neurological condition that can affect the quality of life. Iron is one of the available options for the treatment of RLS.

Methods

This systematic review and meta-analysis assessed the efficacy and safety of different iron regimens in treating RLS. We searched EMBASE, SCOPUS, Web of Science, PubMed, and Cochrane databases for studies that assessed the effects of different iron supplements compared to placebo on the International RLS score (IRLS), RLS-Quality of Life (QOL) score, sleep visual analog score (VAS), and adverse events (AEs).

Results

Twelve randomized controlled trials with 511 patients were included. Iron improved the IRLS score (mean difference [MD] = −5.28; 95% confidence interval [CI] − 7.66, −2.90; P < 0.0001), the number of patients with improved IRLS score (risk ratio [RR] = 2.06; 95% CI 1.49, 2.84; P < 0.0001), RLS-QOL (MD = 7.42; 95% CI 1.32, 13.51; P = 0.02), and sleep VAS score (MD = −24.83; 95% CI −40.08, −9.58; P = 0.001), but yielded more overall AEs (RR = 2.04; 95% CI 1.46, 2.85; P < 0.0001), with no difference in serious AEs or those leading to drug discontinuation (RR = 2.04; 95% CI 0.39, 10.81; P = 0.40 and RR = 6.25; 95% CI 0.79, 49.54; P = 0.08), respectively.

Conclusion

Iron, especially ferric carboxymaltose, is a valuable option for RLS, which showed improvement in several indices.

Restless legs syndrome (RLS) is a prevalent sensorimotor disorder affecting 5% of the population worldwide. RLS is characterized by the urge to move the legs, particularly during the night, and is accompanied by abnormal leg sensation. These symptoms vary in intensity, ranging from mildly bothersome to disrupting sleep and diminishing the quality of life, which necessitates medical intervention.^1–3

Iron deficiency plays an essential role in the pathophysiology of RLS. As opposed to normal individuals, patients with RLS exhibit reduced ferritin levels in their cerebrospinal fluid and elevated transferrin levels. However, when comparing serum samples from healthy persons and RLS patients, there were no meaningful alterations in iron, ferritin, or transferrin between both groups.^4,5 The prevalence of RLS in individuals with iron deficiency anemia (IDA) is markedly elevated, being four to five times greater than in the general population.⁶ Also, the most noteworthy finding is the confirmation of a high occurrence of clinically significant RLS in IDA cases. Approximately one out of every four individuals with IDA suffers from clinically significant RLS, a prevalence rate nine times higher than that observed in the general population.⁶ Furthermore, patients with IDA who have RLS also suffer from notable sleep-related health issues, such as sleep duration reduction to 5 to 6 hours and disrupted sleep.⁶ Apart from changes in the brain induced by iron deficiency, exposure to various environmental, social, and occupational factors can play a role in the pathogenesis of RLS. Several interesting studies have revealed that such exposure can significantly impact sleep quality, leading to sleep disturbances such as increased body movement. Moreover, specific occupations, such as firefighting, have been identified with an elevated prevalence of sleep disorders, including RLS.^7,8

In their most recent guidelines, the International RLS Study Group recommends using oral iron therapy to address RLS in individuals with serum ferritin levels <75 mg/L. In comparison, intravenous (IV) iron treatment is recommended for managing moderate to severe RLS in those with serum ferritin levels <300 mg/L.⁹

Previous research has indicated that the focal point of interest in RLS is the brain, and disordered brain iron regulation is a fundamental pathological factor in RLS. Numerous investigations employing biochemical and imaging techniques have pointed to a connection between the iron within the brain and the functioning of dopamine neurotransmission in the brain.¹⁰ Furthermore, various imaging techniques have consistently demonstrated a decrease in brain iron levels among individuals with RLS.^11,12 A previously published meta-analysis concluded that iron supplementation is associated with improvement of the IRLS score, which supports the use of iron, oral or IV, as effective therapy for patients with RLS.¹³

This systematic review and meta-analysis with trial sequential analysis (TSA) examined the efficacy and safety of different iron supplement regimens in treating RLS. We conducted a meta-regression to explore the relationship between ferric carboxymaltose (FCM) dosage and IRLS score changes, and used TSA to evaluate the stability of current evidence on FCM. We also applied the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system to assess the quality of evidence. Additionally, we extracted more data, allowing us to report additional outcomes such as the Pittsburgh Sleep Quality Index (PSQI), visual analogue score (VAS), RLS-Quality of Life (QOL) score, and side effects. Finally, we employed the updated ROB2 tool and provided a detailed figure to illustrate our findings.¹³

METHODS

Our study was conducted based on the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement¹⁴ and the Cochrane Handbook of Systematic Reviews and Meta-analysis.¹⁵ The prespecified protocol was registered with PROSPERO with the ID CRD42023460631.

Two independent reviewers (A.M. and B.M.) conducted a thorough search in EMBASE, SCOPUS, Web of Science, MEDLINE (PubMed), and Cochrane Central Register of Controlled Trials databases for eligible studies. No filters were applied. Supplemental Table 1 shows the complete search strategy.

Eligibility criteria

Randomized controlled trials (RCTs) that matched our criteria were included in this study. Our PICO criteria were as follows: Population, patients with RLS; Intervention, iron (oral or IV preparations including FCM, iron sucrose, and iron dextran); Control, placebo; and Outcomes, change in the IRLS, RLS-QOL, VAS, and PSQI scores, number of patients with improvement in the IRLS score, any adverse events (AEs), serious AEs, and AEs requiring intervention discontinuation. We excluded non-RCT studies, such as in vitro studies, conference abstracts, observational studies, animal studies, and single-arm clinical trials.

Study selection and data extraction

Covidence online software was used by two independent reviewers (A.M. and M.A.) for study selection. Duplicates were automatically eliminated by the software, and the remaining studies were screened by the reviewers using the title and abstract of each study. Studies that were included in the last step went through further screening using the full text of each study. Another reviewer (M.A.) was called to settle the conflicts at each step of the screening.

Four reviewers (A.A.S.A., A.M., H.A., and M.R.) independently used an Excel sheet that was generated after reading the full text of the included RCTs to extract the following data: summary characteristics, baseline characteristics, efficacy outcomes, and safety outcomes. More information about the extracted data can be found in Tables 1 and 2.

Table 1.

Summary characteristics of the included studies

Study	Total participants	Hematological inclusion	Type of studied RLS	Intervention dose and duration	Comparator dose and duration	Time of efficacy measurement	Othermedications for RLS	Included patients	Follow-up duration	Main conclusion
Allen et al 201119	43	Ferritin <300 ng/d, or TSAT <45%	Primary, as patients were excluded from the study if they had RLS secondary to CNS disease, CNS injury, or chronic kidney disease	IV FCM 500 mg, 2× over 2 wk	Placebo	4 wk	Not allowed	IRLS score ≥15	24 wk	IV FCM provided a safe and effective treatment for RLS that lasted for at least 24 weeks for some patients
Bae et al 202120	29	Hb <12 g/dL and either ferritin <20 ng/mL or ferritin <100 ng/mL with TSAT <18%	Primary, as patients were excluded from the study if they had severe medical diseases including liver cirrhosis and heart failure	1500 mg FCM	Placebo	6 wk	Not allowed	IRLS score ≥15	52 wk	The study showed significant efficacy and safety of FCM 1500 mg treatment both in the short-term (6 weeks) and long-term (52 weeks) in RLS patients with IDA
Bae et al 202321	18	Hb <12 g/dL and either ferritin <20 ng/mL or ferritin <100 ng/mL with TSAT <18%	Primary, as patients were excluded from the study if they had severe medical diseases including liver cirrhosis and heart failure	1500 mg FCM	Placebo	6 wk	Not allowed	IRLS score ≥15	6 wk	1500 mg FCM effectively treats RLS symptoms in IDA patients over 6 weeks, with MRI measurements of improved brain iron content serving as a potential biomarker for RLS patients
Cho et al 201622	64	Ferritin <300 ng/dL, Hb >12 mg/dL, or TSAT <45%	Primary, as patients were excluded if they had polyneuropathy, neurodegenerative disease, chronic kidney disease, pregnancy	IV FCM 1000 mg	Placebo	6 wk	Not allowed	IRLS score ≥15	30 wk	The value of FCM treatment was shown in the short term (6 weeks) and long term (30 weeks) for improving RLS symptoms
Cho et al 201823	64	Ferritin <300 ng/dL, Hb >12 mg/dL, or TSAT <45%	Primary, as patients were excluded if they had polyneuropathy, neurodegenerative disease, chronic kidney disease, pregnancy	IV FCM 500 mg, once	Placebo	6 wk	Not allowed	IRLS score ≥15	30 wk	This study did not show a significant benefit of a single 500 mg FCM treatment for RLS symptoms
Davis et al 200024	28	Hb < 12	Patients were included regardless of other potential causes of RLS, such as neuropathy and renal disease	Oral iron sulfate, 325 mg twice daily for 16 wk	Placebo	12 wk	Allowed	Symptomatic RLS	26 wk	Iron sulfate does not appear to be an effective empiric treatment for RLS
Deng et al 201725	32	Ferritin level of <200 ng/mL, and TSAT of <20%	Secondary, as all the participants were on dialysis	Three times per week for a total dose of 1000 mg	Placebo	2 wk	Not allowed	Symptomatic RLS	2 wk	IV iron sucrose is a safe and effective treatment for reducing RLS symptoms in HD patients over the short term
Early et al 200926	18	Hb <12 g/dL	Primary, as patients were excluded if they had possible secondary forms of RLS	1000 mg iron IV 2 wk	Placebo	2 wk	Not allowed	PLMS >15/h	2 yr	High-dose IV iron failed to demonstrate the robust changes reported in three prior open-label studies
Grote et al 200927	60	Ferritin level <30 ng/L and normal folic acid/ B12 vitamin serum values	Primary, as exclusion criteria included clinical or laboratory findings suggestive of secondary RLS	FCM 200 mg, 5× times over 5 wk	Placebo	3 wk	Not allowed	IRLS score ≥ 10	1 yr	This study showed a lack of superiority of iron sucrose at 11 weeks but found evidence that iron sucrose reduced RLS symptoms both in the acute phase (7 weeks) and during long-term follow-up
Sloand et al 200428	25	NA	Secondary, as all the participants were on dialysis	1000 mg of iron dextran	Placebo	4 wk	Not allowed	Symptomatic RLS	4 wk	High-dose iron dextran infusion is associated with a significant but transient reduction in symptoms of RLS in patients with ESRD
Trenkwalder et al 201729	110	Ferritin < 75 [patients were also included if ferritin (75–300) ng/L and TSAT <20%], Hb < 11.5/12.5	No sufficient data on whether secondary causes of RLS were excluded or not	IV FCM 1000	Placebo	4 wk	Not allowed	RLS score ≥ 15	12 wk	In patients who responded to treatment, FCM may require more time to stabilize RLS than previously assumed
Wang et al 200930	18	Ferritin level of 15–75 ng/mL	Primary, as patients were excluded if they had a possible secondary cause of RLS	Ferrous sulfate 325 mg twice daily	Placebo	12 wk	Not allowed	IRLS score >11	12 wk	There was significant improvement in RLS symptoms using oral iron therapy in patients with low to normal ferritin

CNS indicates central nervous system; ESRD, end-stage renal disease; FCM, ferric carboxymaltose; Hb, hemoglobulin; IDA, iron deficiency anemia; IRLS, International RLS score; IV, intravenous; MRI, magnetic resonance imaging; PLMS, periodic limb movements of sleep; RLS, restless leg syndrome; TSAT, transferrin saturation.

Table 2.

Baseline characteristics of the included studies

Study	Patients in each group		Age (year): Mean (SD)		Gender (female): N (%)		IRLS score: mean (SD)		Mean baseline ferritin (mg/L): Mean (SD)		Mean baseline Hb (g/dL): Mean (SD)
	Iron	Placebo	Iron	Placebo	Iron	Placebo	Iron	Placebo	Iron	Placebo	Iron	Placebo
Allen et al 201119	24	19	49.5 (11.4)	54.8 (13.6)	17 (70.8)	10 (52.6)	25.0 (5.8)	24.2 (5.5)	28.1 (22.9)	24.8 (20.2)	NA	NA
Bae et al 202120	15	14	48.00 (13.50)	47.71 (6.04)	13 (86.67)	13 (92.86)	25.60 (7.35)	25.21 (5.58)	5.82 (5.05)	7.9 (9.84)	10.27 (22.41)	10.46 (0.94)
Bae et al 202321	10	8	42.10 (7.11)	46.25 (4.53)	9 (90.0)	8 (100.0)	23.70 (8.04)	24.75 (7.21)	3.73 (2.56)	3.05 (0.91)	10.16 (1.55)	10.08 (1.01)
Cho et al 201622	32	32	49.7 (13.7)	52.3 (10.7)	26 (81.3)	24 (75)	27.4 (4.03)	28.0 (5.16)	53.5 (41.8)	69.3 (55.4)	13.3 (1.42)	13.5 (1.11)
Cho et al 201823	32	32	47.3 (13.3)	51.5 (12.0)	26 (81.3)	26 (81.3)	28.1 (5.7)	27.3 (5.30)	50.7 (40.8)	70.2 (59.5)	13.5 (1.2)	13.3 (0.9)
Davis et al 200024	14	14	58.6 (33–80)	59.9 (33–76)	9 (62)	10 (71.4)	NA	NA	134.8 (9–680)	100.6 (8–335)	14.3 (12.7–16.9)	13.7 (11.6–15.6)
Deng et al 201725	16	16	63.63 (4.83)	64.19 (7.93)	20 (62.5) for both groups		26.06 (6.84)	26.31 (7.14)	154.75 (38.34)	157.13 (35.33)	10.5 (0.9)	10.8 (1.006)
Early et al 200926	11	7	66.4 (11.4)	61.4 (10.0)	11 (55)	7 (71)	30.8 (9.2)	29.7 (2.9)	78.3 (41.7)	70.3 (21.5)	15.0 (1.2)	14.0 (0.84)
Grote et al 200927	29	31	47 (10)	46 (8)	25 (86.2)	28 (90)	23.75 (7.8)	26 (5.75)	20.1 (12)	20.4 (11)	12.9 (1.8)	13.1 (1.2)
Sloand et al 200428	11	14	57.25 (4.9)	53.75 (7.79)	5 (45)	10 (71)	7 (1.7)	9 (0.61)	148 (93.5)	186.5 (79.7)	11.05 (0.45)	11.2 (0.33)
Trenkwalder et al 201729	59	51	53.0 (15.7)	55.5 (15.9)	48 (81.4)	42 (82.4)	25.9 (5.65)	26 (5.78)	41.93 (34.55)	48.85 (45.95)	NA	NA
Wang et al 200930	11	7	60 (36–82)	58 (33–72)	6 (54.5)	5 (71.1)	24.8 (5.72)	23 (5.03)	40.6 (15.3)	36.7 (20.8)	14.5 (1.30)	13.7 (1.50)

Hb indicates hemoglobulin; IRLS, International Restless Legs Syndrome Score; NA, not available; SD, standard deviation.

Risk of bias assessment

Four reviewers (A.A.S.A., A.M., H.A., and M.R.) assessed the quality of the studies included in the research independently using the Cochrane ROB2 tool.¹⁶ The ROB2 tool examined the risk of bias in five domains: the randomization process, deviation from the intended intervention, missing outcome data, measurement of the outcome, and selecting the reported results. The reviewers discussed and settled any disagreements. To estimate the quality of evidence, two reviewers (M.A. and B.A.) utilized the GRADE instructions.^17,18

Statistical analysis

The analysis was done with Revman version 5.4. For dichotomous variables, we estimated the pooled risk ratio (RR) and 95% confidence interval (CI), and for continuous variables, we calculated mean differences (MD) with a 95% CI. The chi-squared test and I² were used to evaluate statistical heterogeneity. When heterogeneity was deemed significant (P < 0.1 or I² > 60%), we employed a random effects model; otherwise, we used a fixed effects model. We performed a subgroup analysis to examine the impact of different iron preparations. A meta-regression was performed using Comprehensive Meta-Analysis software to examine the relationship between the effect of iron and the baseline IRLS score and iron doses. A funnel plot was created to check for possible publication bias via Comprehensive Meta-analysis Software. Finally, we performed a TSA with a 5% risk of type I error and a 20% risk of type II error (80% power). The TSA was conducted in chronological order by year of publication. TSA was performed using the TSA Viewer, version 0.9 beta (Copenhagen Trial Unit, Copenhagen, Denmark). We used the funnel plot and the Egger test to measure publication bias when there were more than 10 papers in the analysis. A P value < 0.05 indicates a significant publication bias.

RESULTS

We retrieved 4017 records from searching the data sources and excluded 1763 identical records via the Covidence tool, with 2254 records eligible for title and abstract screening. Thirty records entered full-text screening, and 12 RCTs were included in the study^19–30 (Figure 1). These 12 RCTs involved a total of 511 patients. Complete summary and baseline characteristics of the included RCTs are outlined in Tables 1 and 2.

Figure 1.

PRISMA flow chart of the screening process, which included searches of databases, registers, and other sources.

Risk of bias and quality of evidence

Six studies—Allen et al 2011, Bae et al 2021, Bae et al 2023, Cho et al 2016, Cho et al 2018, and Trenkwalder et al 2017^19–24—showed an overall low risk of bias, while four studies showed some concerns (Deng et al 2017, Early et al 2009, Grote et al 2009, and Wang et al 2009^25–28). Two studies, Davis et al 2000 and Sloand et al 2004,^29,30 showed a high risk of bias. More detailed information is shown in Figure 2. The quality of evidence is illustrated via GRADE instructions in Table 3.

Figure 2.

Summary of risk of bias: (a) authors’ judgments about each risk of bias item for each included study; (b) authors’ judgments about each risk of bias item presented as percentages across all included studies.

Table 3.

Quality assessment of the outcomes using GRADE criteria

Certainty assessment							No. of patients		Effect		Certainty	Importance
No. of studies	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Iron therapy	Placebo	Relative (95% CI)	Absolute (95% CI)
IRLS score
11	Randomized trials	Not serious	Very seriousa	Not serious	Not serious	Dose response gradient	250	231	–	MD 5.28 points lower (7.66 lower to 2.9 lower)	⨁⨁⨁◯ Moderate	CRITICAL
RLS-QOL
4	Randomized trials	Not serious	Not serious	Not serious	Seriousb	None	103	97	–	MD 7.42 points higher (1.32 higher to 13.51 higher)	⨁⨁⨁◯ Moderate	IMPORTANT
PSQI score
4	Randomized trials	Not serious	Very seriousa	Not serious	Seriousb	None	89	86	–	MD 3.15 points lower (7.03 lower to 0.73 higher)	⨁◯◯◯ Very low	IMPORTANT
Sleep VAS score
5	Randomized trials	Not serious	Very seriousa	Not serious	Seriousb	None	97	99	–	MD 24.83 points lower (40.08 lower to 9.58 lower)	⨁◯◯◯ Very low	IMPORTANT
Number of patients with improvement in IRLS score
6	Randomized trials	Not serious	Not serious	Not serious	Not serious	None	80/166 (48.2%)	35/147 (23.8%)	RR 2.06 (1.49 to 2.84)	252 more per 1000 (from 117 more to 438 more)	⨁⨁⨁⨁ High	CRITICAL
Any adverse events
7	Randomized trials	Not serious	Not serious	Not serious	Not serious	None	73/180 (40.6%)	34/170 (20.0%)	RR 2.04 (1.46 to 2.85)	208 more per 1000 (from 92 more to 370 more)	⨁⨁⨁⨁ High	CRITICAL

CI indicates confidence interval; IRLS, International Restless Leg Syndrome Score; MD, mean difference; PSQI, Pittsburgh Sleep Quality Index; RLS-QOL, Restless Leg Syndrome–Quality of Life score; RR, risk ratio; VAS, sleep visual analog score.

^aUnexplained high heterogeneity; b. Wide confidence interval.

Efficacy outcomes

Change in IRLS score

The analysis included 11 studies,^19–27,30 with a total of 250 patients in the iron therapy arm and 231 patients in the placebo arm. Moderate-certainty evidence showed that iron therapy significantly decreased the IRLS score compared to placebo (MD = −5.28; 95% CI −7.66, −2.90; P < 0.0001) (Table 3, Figure 3a). The pooled analysis was heterogenous (I² = 77%, P < 0.00001), which sensitivity analysis could not resolve.

Figure 3.

Forest plots showing the mean difference in International Restless Leg Syndrome Score for (a) overall iron therapy and for IV and oral iron subgroups, and (b) different forms of IV iron.

The subgroup analysis revealed that IV iron significantly outperformed the placebo group (MD = −4.98; 95% CI −7.48, −2.48; P < 0.0001), and the pooled analysis was heterogenous (I² = 79%, P < 0.00001) (Figure 3a). Furthermore, subgroup analysis showed that FCM significantly decreased the IRLS score (MD = −6.37; 95% CI −10.19, −2.55; P = 0.001), while iron sucrose did not reach statistical significance (MD = −3.32; 95% CI −8.81, 2.18; P = 0.24) (Figure 3b). The pooled analysis was heterogeneous for the FCM and iron sucrose subgroups (I² = 81%, P = 0.005; I² = 81%, P < 0.00001, respectively).

Change in RLS-QOL score

The analysis included four studies^19,20,22,23 with 103 patients in the iron therapy arm and 97 patients in the placebo arm. All of the studies included in the analysis administrated IV FCM. Moderate-certainty evidence showed that iron therapy significantly increased the RLS-QOL score compared to placebo (MD = 7.42; 95% CI 1.32, 13.51; P = 0.02) (Table 3, Figure 4a). The pooled analysis was homogenous (I² = 27%, P = 0.25).

Figure 4.

Forest plots showing (a) the mean difference in Restless Leg Syndrome–Quality of Life score, (b) the mean difference in Pittsburgh Sleep Quality Index score, (c) the mean difference in sleep visual analog score, (d) the risk ratio of the number of patients with improvement in the International Restless Leg Syndrome score for overall iron therapy and for IV and oral iron subgroups, and (e) the risk ratio for different forms of IV iron.

Change in PSQI score

The analysis included four studies^20–23 with a total of 89 patients in the iron therapy arm and 86 patients in the placebo arm. All of these studies administrated IV FCM. Very low-certainty evidence showed that iron therapy decreased the PSQI score; however, the result was statistically insignificant compared to placebo (MD = −3.15; 95% CI −7.03, 0.73; P = 0.11) (Table 3, Figure 4b). The pooled analysis was heterogenous (I² = 84%, P = 0.0002), which sensitivity analysis could not resolve.

Change in sleep VAS score

The analysis included five studies^20–23 with a total of 97 patients in the iron therapy arm and 99 patients in the placebo arm. Very low-certainty evidence showed that iron therapy significantly decreased the sleep VAS score compared to placebo (MD = −24.83; 95% CI −40.08, −9.58; P = 0.001) (Table 3, Figure 4c). The pooled analysis was heterogenous (I² = 83%, P < 0.0001), which sensitivity analysis could not resolve.

A subgroup analysis revealed that IV iron (FCM) reduced the sleep VAS score significantly more than placebo (MD = −30.09; 95% CI −45.78, −14.40; P = 0.0002) (Figure 4c). The pooled analysis was heterogeneous (I² = 79%, P = 0.002).

Number of patients with improvement in IRLS score

The analysis included six studies^{19,22,24,26–28} with a total of 166 patients in the iron therapy arm and 147 patients in the placebo arm. High-certainty evidence showed that the incidence of improvement in IRLS score was significantly higher in the iron therapy group compared to the placebo group (RR = 2.06; 95% CI 1.49, 2.84; P < 0.0001) (Table 3, Figure 4d). The pooled analysis was homogenous (I² = 8%, P = 0.36).

The subgroup analysis showed that IV iron significantly increased the incidence of improvement in IRLS score (RR = 1.97; 95% CI 1.42, 2.74; P < 0.0001). The pooled analysis was homogenous (I² = 12%, P = 0.34) (Figure 4d). Moreover, FCM accounted for this significant effect (RR = 2.33; 95% CI 1.48, 3.67; P = 0.0002), while iron sucrose did not show a significant effect (RR = 1.52; 95% CI 0.96, 2.40; P = 0.08) (Figure 4e). The pooled analysis was homogenous for the FCM (I² = 0%, P = 0.89) and iron sucrose (I² = 59%, P = 0.12) subgroups.

Safety outcomes

Any AEs

The analysis included seven studies^{19,22,24,26,27,29,30} with a total of 180 patients in the iron therapy arm and 170 patients in the placebo arm. High-certainty evidence showed that iron therapy had a significantly higher number of any AEs compared to placebo (RR = 2.04; 95% CI 1.46, 2.85; P < 0.0001) (Table 3, Figure 5a). The pooled analysis was homogenous (I² = 24%, P = 0.24).

Figure 5.

Forest plots showing the risk ratio for (a) any adverse events for overall iron therapy and for IV and oral iron subgroups; (b) any adverse events for different forms of IV iron; (c) severe adverse events; and (d) adverse events requiring discontinuation.

The subgroup analysis showed that IV iron had a significantly higher rate of any AEs compared to placebo (RR = 1.80; 95% CI 1.28, 2.51; P = 0.0006) (Figure 5a). The pooled analysis was homogenous (I² = 0%, P = 0.53). Similarly, FCM accounted for this significantly higher any AEs result (RR = 2.19; 95% CI 1.22, 3.90; P = 0.008), while iron sucrose did not show a significant incidence of any AEs compared to placebo (RR = 1.51; 95% CI 0.88, 2.60; P = 0.14) (Figure 5b). The pooled analysis was homogenous for FCM (I² = 0%, P = 0.94) but heterogeneous for iron sucrose (I² = 65%, P = 0.09).

Serious AEs

The analysis included three studies^24,27,29 with a total of 94 patients in the iron therapy arm and 96 patients in the placebo arm. Iron therapy did not cause any significant serious AEs compared to placebo (RR = 2.04; 95% CI 0.39, 10.81; P = 0.40) (Figure 5c). The pooled analysis was homogenous (I² = 0%, P = 0.82).

Subgroup analysis showed that IV iron was comparable to placebo in terms of serious AEs (RR = 1.72; 95% CI 0.23, 12.60; P = 0.59) (Figure 5c). The pooled analysis was homogenous (I² = 0%, P = 0.59).

AEs requiring discontinuation

The analysis included two studies^27,29 with a total of 43 patients in the iron therapy arm and 45 patients in the placebo arm. There was no significant difference in discontinuation due to AEs in the iron therapy arm compared to placebo (RR = 6.25; 95% CI 0.79, 49.54; P = 0.08) (Figure 5d). The pooled analysis was homogenous (I² = 0%, P = 0.59).

Trial sequential analysis

The TSA shows that the cumulative z-curve had passed the trial sequential monitoring edge for the favorable effect of iron therapy on the IRLS score, and the actual cumulative sample size exceeded the diversity-adjusted required information size (Supplemental Figure 1). TSA of the IV iron and FCM subgroups showed similar findings (Supplemental Figures 2 and 3). We can now confidently confirm the beneficial effect of iron therapy, IV iron, and FCM on the IRLS score, and further studies are unlikely to alter our findings. However, the TSA of iron sucrose revealed that there was insufficient evidence at the time to draw firm conclusions about the effect of iron sucrose on IRLS score, and further research into the role of iron sucrose is required (Supplemental Figure 4).

Meta-regression

There was a significant correlation between the dose of FCM and its effect on the reduction of the IRLS score (Supplemental Figure 5). However, the meta-regression analysis revealed no significant interaction (P = 0.63) between the effect of iron therapy on the IRLS score and the IRLS baseline score (Supplemental Figure 6).

Publication bias

The funnel plot did not rule out publication bias, as shown in Supplemental Figure 7. However, Egger’s (P = 0.69) and Begg’s (P = 0.31) tests indicated no evidence of publication bias among the included studies for change in IRLS score outcome. We could not assess publication bias for the remaining outcomes due to the presence of <10 studies in the meta-analysis.

DISCUSSION

This is an updated systematic review and meta-analysis with a TSA that studied the effectiveness and safety of iron therapy in treating RLS. Our results show that iron therapy led to improvement in IRLS score, RLS-QOL, and sleep VAS score, with no significant improvement in PSIQ score compared to placebo. Iron therapy led to a significantly higher overall incidence of AEs compared to placebo. However, the increase in serious AEs and those requiring treatment discontinuation was not statistically significant.

According to the latest guidelines, FCM (at a dosage of 1000 mg) demonstrates effectiveness in managing moderate to severe RLS among individuals with serum ferritin levels <300 μg/L. It can be considered a primary treatment option for RLS in adults. For those with serum ferritin levels ≤75 μg/L, oral iron may also be a potentially effective treatment for RLS. However, there is currently insufficient evidence to draw conclusions regarding the efficacy of oral or IV iron for RLS in pediatric populations.⁹

IV iron treatment changes brain iron levels, which correlate with reduced RLS severity. This indicates that symptom relief is connected to higher iron content in certain areas of the sensorimotor network. A study using quantitative susceptibility mapping (QSM) found significant changes in brain iron levels in individuals with RLS. Six weeks after receiving 1000 mg of FCM (t1), QSM values increased in the caudate nucleus and decreased in the putamen by the 24-week mark (t2). These changes in brain regions were linked to the severity of RLS symptoms.³¹ Moreover, Bae et al identified a statistically significant correlation between the difference in IRLS scores and the difference in brain iron levels measured by QSM in the thalamus.²¹

In our study, the improvement in IRLS score was significant in overall iron, and the point estimate favored IV formulations, with our GRADE assessment indicating moderate quality of evidence. The significant improvement in RLS symptoms was primarily driven by IV FCM, while IV iron sucrose and dextran had minimal impact. This difference may be due to the limited research and small patient populations studied for iron sucrose and dextran formulations. Furthermore, it is worth noting that the treatment effectiveness of iron sucrose gradually diminished over time, leading to the reappearance of RLS symptoms. This decline could potentially be linked to a decrease in body iron levels due to insufficient absorption of iron in the gastrointestinal tract or an increase in iron loss through renal or gastrointestinal means.²⁷ Allen et al and Bae et al conducted long-term evaluations of IV FCM, both revealing significant improvements in the IRLS score at 24 weeks and 52 weeks, respectively.^19,20 Notably, Cho et al used a treatment dose of 500 mg FCM, which did not yield a substantial improvement in outcomes compared to the placebo after 6 weeks.²³ However, in studies employing doses of 1000 mg and 1500 mg of FCM, a notable improvement in outcomes was observed. This suggests that there might be a correlation between higher iron doses and more favorable treatment outcomes.^20–22

This view is also supported by our meta-regression analysis, which demonstrated a substantial correlation between the dosage of FCM and its impact on reducing the IRLS score. Specifically, higher doses were associated with a more pronounced reduction in IRLS score. This should be investigated in future studies. The meta-regression analysis also indicated that the effectiveness of iron in reducing IRLS scores was not significantly different between severe and mild cases.

There was a significant improvement in RLS-QOL in the analysis of four included RCTs that used IV FCM. The RLS-QOL demonstrates reliability and validity, effectively capturing improvements in patients with RLS, and addresses the aspects of life most affected by RLS symptoms.³² A statistically significant reduction in the sleep VAS score was found when analyzing five RCTs. The subgroup analysis showed that only IV FCM caused a significant reduction, with no significant effect when using oral iron sulfate. Brain iron deficiency can lead to a hypoadenosinergic state, which activates the arousal system and increases hyperarousal in RLS patients. These findings imply that addressing brain iron deficiency may alleviate the hyperarousal state and potentially enhance sleep quality.³³ However, the reduction in the PSQI score was statistically insignificant in the analysis of four RCTs that used FCM. The GRADE quality assessment showed moderate certainty for improvements in RLS-QOL. However, the evidence for enhancements in PSQI and sleep VAS scores was of very low certainty.

When examining safety, the administration of iron treatment resulted in a notably increased occurrence of any AEs, as supported by our GRADE assessment, which attests to the high quality of the evidence. These AEs primarily manifest as mild symptoms like headaches and nausea. Conversely, there was a lack of statistically significant occurrences of severe AEs or significant instances of treatment discontinuation when compared to the placebo group.

TSA was employed to manage the potential for type I error, evaluate whether the cumulative sample size reached a sufficient level, and add strength to the meta-analysis findings. The results, as determined by TSA, indicate that overall iron, IV iron, and FCM had a substantial positive influence on the IRLS score. Moreover, the sample size was sufficient to support a robust conclusion regarding the effectiveness of iron in the treatment of RLS.

Limitations and future research

Our study is an up-to-date systematic review, based on PRISMA guidelines, with performance of TSA, meta-regression, and evaluation of the certainty of the evidence using the GRADE system. Nonetheless, it has a few limitations, as the heterogeneity was high and was not resolved with subgroup and sensitivity analysis. This might be due to the variation in sample size, ferritin levels at inclusion, dosage, use of different regimens of iron, dosage interval, and the time of efficacy measurement among the included studies. A limited number of studies were available to support or refute the use of oral iron, iron sucrose, and iron dextran and allow us to postulate a definite conclusion concerning these supplements. Also, the number of studies in the meta-regression was limited, suggesting that the findings should be verified in the future.

Additional research is warranted to assess the cost-effectiveness of iron supplementation as a treatment for RLS and to investigate the effectiveness of alternative IV and oral iron supplementation compared to FCM, while also considering distinct subgroups of patients who exhibit varying characteristics and triggers, such as children and nonanemic patients. More studies with extended durations are needed to establish more robust evidence regarding the drug’s effectiveness and safety. Furthermore, additional research should be conducted to evaluate the iron content in the striatal structures of the brain as a means of assessing the effectiveness of IV iron in alleviating RLS symptoms.

CONCLUSION

Iron supplements, specifically FCM, are a valuable option for patients with low serum ferritin and RSL, improving IRLS scores, RLS-QOL scores, sleep VAS scores, and the number of patients with improvement in IRLS scores. However, overall, iron is associated with more side effects, but not serious ones or those leading to drug discontinuation. Meta-regression showed a positive correlation between the dose of FCM and reduction in the IRLS score. The TSA confidently confirmed the overall beneficial effect of iron therapy, specifically FCM, on the IRLS score. More trials are needed to study the effect of iron in patients with RLS and a normal iron profile, and to study different forms of iron compared to FCM.

Disclosure statement/Funding

The authors report no funding or conflicts of interest.