Restless legs syndrome: clinical experience with long-term treatment

Background and purpose: There are limited data on long-term treatment efficacy, and almost none on predictors of treatment response in patients with restless legs syndrome (RLS). To assess: (1) long-term efficacy of RLS treatment in a clinical setting, (2) predictors of a good treatment response, and (3) the value of the RLS-severity score according to the criteria of the International Restless Legs Syndrome Study Group (IRLSSG). Patients and methods: Over three years 70 patients (36 men, 34 women; mean age: 59 years; range: 29-79) with RLS were prospectively assessed. Diagnosis of RLS was made according to international criteria Severity of RLS symptoms was were assessed at the outset by the IRLSSG rating scale. Treatment was chosen individually according to clinical judgement. After a mean follow-up time of 16 months (range: 1-106 months) evolution of symptoms was assessed by both overall clinical impression and IRLSSG rating scale. Clinical characteristics and treatment effect were compared between patients never treated for RLS before this study ('naive' = N-pts) and those with previous treatment ('treated'= T-pts). Predictors of treatment response were sought for comparing patients with good treatment response (good, better or much better on follow-up) and those with bad (B-pts) treatment response. Results: There were 40 N-pts and 30 T-pts. The mean IRLSSG score (hereinafter, IRLSSG) at baseline was 26 (range 12-38). No significant differences were found between N-and T-pts in age, gender, etiology and duration of RLS, positive family history, presenting sleep complaint, IRLSSG, or percentage of patients with periodic limb movements in sleep (PLMS) on polysomnography (PSG). At final follow-up 30 (76%) of 40 N-pts and 23 (77%) of 30 T-pts had a good (G-pts) treatment response. The mean IRLSSG at follow-up was 19 (range: 1-36). There was a significant correlation between improvement of overall clinical impression (better or much better on final follow-up) and reduction of IRLSSG (P<0.0001). PLMS were more common in B- than G-pts (100 vs 58% of patients, P=0.02). In all other variables considered the two groups were similar. Conclusion: (1) A good long-term treatment response can be obtained and maintained in a clinical setting in about 80% of RLS patients. (2) Patients with RLS and without PLMS may have a better long-term treatment response, and (3) the IRLSSG is a useful tool for assessment of evolution of RLS symptoms over time in individual patients. (C) 2004 Published by Elsevier B.V.