MRI measurement of brain iron in patients with restless legs syndrome

Methods.

Multi-slice measurements of the relaxation rates R₂* (from the gradient-echo MRI) and R₂ (From the spin-echo MRI) were obtained from a single scan using the GESFIDE (gradient-echo sampling of FID and echo) sequence8 on a GE 1.5T Signa System (General Electric, Milwaukee, WI) following procedures previously described.5

R₂ and R₂* images were reconstructed as a single, multi-slice “stack” in the public-domain image-analysis program NIH Image 1.61. The slices showing the iron-containing structures on the R₂* images (which generally have the best contrast-to-noise ratio) were displayed. The structures were manually traced independently by two trained investigators (both blinded to the control/patient status of the subject, and to each other’s results) using standard anatomic guidelines for the slice with the best presentation of the area. No asymmetries were expected or observed; therefore, the relaxation rates were averaged for both left and right hemispheres. R₂′ was then calculated from the difference of R₂* and R₂.

RLS severity was assessed in two ways. A trained clinician, blinded to the patients’ iron status and age, reviewed the charts to rate severity using the Johns Hopkins RLS Severity Scale (JHRLSS). This scale is scored based on the usual time-of-day for onset of RLS symptoms: 0 for none, 1 for mild (bedtime only), 2 for moderate (after 6:00 pm), or 3 for severe (Before 6:00 pm). This standard measurement from prior studies has an inter-rater agreement of 92% for trained clinical raters.9 Severity was determined from the description of the patients’ symptoms prior to treatment for RLS. In addition, shortly prior to the MRI, a board certified sleep specialist experienced with RLS interviewed the patient using the same scale to assess the patient’s report of RLS severity when not taking the current RLS medications. There was 100% agreement between these two severity ratings.

All control subjects were interviewed by a diplomate of the American Board of Sleep Medicine experienced in diagnosing and treating RLS. They were included only if they showed no symptoms suggesting RLS, periodic leg movements in sleep, or any significant sleep disturbance.

Results.

The groups were well matched for age (average ± SD: patients with RLS = 66.2 ± 10.5; controls = 66.4 ± 16.8.). The MRI measurements made independently by the two well-trained investigators showed excellent inter-reader reliability with no systematic bias. The inter-reader reliability correlation for the R₂* and R₂ was 0.99, and intra-reader reliability correlation was 0.98.

Figure 1 gives an example of the R₂* MRI image used for identification of areas showing the clear visualization of the red nucleus and substantia nigra, despite the marked differences between a 70-year-old patient with severe RLS and a 71-year-old control subject. The Spearman correlations between R₂′ and RLS severity were significant for both the putamen (ρ = −0.64) and the substantia nigra (ρ = −0.69) ( figure 2). The lower values of R₂′ for patients with RLS compared to controls were marginally significant for the putamen and significant for the substantia nigra. There were no significant results from analyses of the caudate or the areas of secondary interest ( table).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. R₂* images in (A) a 70-year-old patient with restless legs syndrome (RLS) and (B) a 71-year-old control subject. Much lower R₂* relaxation rates are apparent in the RLS case in both red nucleus and substantia nigra.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. R₂` (1/s) versus restless legs syndrome (RLS) severity on the Johns Hopkins RLS Severity Scale (JHRLSS) for substantia nigra and putamen.

Discussion.

Although this is a limited initial study, the results showed two major features. First, the results were consistent with the primary hypothesis that decreases in iron concentration in the nigrostriatal areas correlate with RLS severity. This was found for the substantia nigra and the putamen, but not the caudate. These results are consistent with prior studies of patients with RLS showing iron insufficiency in CSF2 and striatal dopamine abnormalities in brain imaging studies.6,7⇓ This, of course, only establishes correlation and not causation. Iron insufficiency could contribute to the dopamine abnormalities or vice versa, or both could be caused by a third factor. All patients with RLS also had been treated with dopaminergic medications, which could possibly have affected brain iron concentrations. Further studies with larger sample sizes and, ideally, with patients off RLS-related medications are needed to confirm these results.

Second, iron decreases were most clearly evident for patients with severe RLS. This may characterize this aspect of RLS or may indicate limitations of the MRI measurements. Future MRI studies with larger samples should carefully evaluate RLS severity and ensure adequate inclusion of patients with severe RLS occurring day and night. Larger samples may also uncover smaller differences than could be detected in this initial study.

The pathology for RLS may also involve other CNS areas either too small or too iron-poor for quantitative iron measurement by MRI. Overall, these results do not prove a causal relation between low brain iron and RLS. There could be, for example, another factor independently causing decreases in both iron and dopamine, leading to RLS. Nonetheless, the results demonstrate potential utility and possible limitations (involving RLS severity) of this MRI method, and are consistent with the hypothesis that the pathophysiology of RLS involves iron-insufficiency in at least some iron-rich regions of the brain.