Non-opioid analgesics and the risk of restless leg syndrome—a spurious association?

Abstract

Objectives: The aim of this study was to evaluate the prevalence and risk factors of restless leg syndrome in a secondary care patient population maintained on tricyclic and serotonin re-uptake inhibiting antidepressants. Study design: Prospective observational study. Study population: Two hundred and forty-three patients of a single practice specializing in the treatment of affective and anxiety disorders. Patients younger than 30 years and those with neurologic diseases such as Parkinson's disease, multiples sclerosis and dementia were excluded, as well as those with schizophrenia or other illusional disturbances, eating disorders, histories of opioid or codeine-containing analgesic abuse or of alcohol or illicit drug abuse. Patients remained in the study population only if they had gained complete or partial recovery of their anxiety and/or depressive symptoms and been treated for at least 6 months in the practice's ambulatory care (85.0% of the patients were finally included). Methods: Histories of leg discomfort, akathisia, medication use and socio-demographic factors were assessed at the time of enrolment. RLS was assumed if all minimal criteria of the international RLS study group were fulfilled. RLS patients also had a neurologic examination focusing on signs and symptoms of neuropathy. It is unclear whether those without RLS were also examined neurologically. RLS case status was assessed at 6-month intervals for patients in the study. It is unclear if the reported RLS prevalence reflects case status at enrolment and/or during the treatment course. The prevalence of various pain syndromes, as well as the intake of analgesics and average coffee intake per day was compared between those with and without RLS. Lifetime and current treatment exposure to antidepressants and dopamine antagonists was evaluated according to RLS status. It is unclear what the mean treatment time was. Various univariant, descriptive statistics were used. The statistical analysis and logistic regression was applied using RLS case status as the dependent and co-morbidities, and medications/treatments as independent variables. Results: The prevalence of RLS among the patients included in the analysis was 27.0%, with female patients more often affected than men (31.0% vs. 19.0%). RLS symptom severity during the course of treatment was low for 43 of 65 cases. Somatic co-morbidities were more prevalent in RLS affected women and less prevalent in RLS affected men than in gender specific non-RLS patients. No significant differences were observed for lifetime exposure to dopamine antagonist between those with and without RLS (38.0% vs. 35.0%). Chronic pain syndromes were observed more often in RLS cases (women: 86.0% vs. 32.0%, men: 62.5% vs. 29.0%). Twelve percent of those without RLS and 75.0% of RLS cases had a history of regular intake of analgesics prior to study enrolment. In a multivariate logistic regression model the authors found RLS case status to be associated with the intake of analgesics (OR: 25.5, P<0.001), psychiatric co-morbidity (OR: 2.97, P=0.01) and an interaction between age and the intake of analgesics (OR: 2.6, P=0.03). Conclusions: A 27.0% prevalence of RLS was found in this highly selected population treated with tricyclic and serotonin re-uptake inhibiting antidepressants. The authors infer from their results that regular intake of analgesics is a risk factor for RLS and that the observed association of gender with RLS might be explained by a higher intake of analgesics in women compared to men. They further conclude that the intake of analgesics, rather than the pain co-morbidity itself, enhances the risk of RLS.