The association between restless legs syndrome, cardiovascular and metabolic diseases: hypotheses and evidence from the literature

The association between RLS and both CVRFs, such as hypertension and diabetes, and CVDs still remains elusive. Although several shared physiopathological causes could explain these possible relationships, the emerging body of literature focusing on these disorders remains controversial. The reasons for these inconsistent findings are mainly due to the different methodologies applied. First, considering that RLS, CVRFs and CVDs are influenced by age and sex, many clinical and population-based studies performed a selection bias by restricting the sample collection to these covariates. Second, assessments of covariates are often incomparable and the methods applied for diseases assessment are often affected by low sensitivity and specificity. Only few population-based studies collected data by means of face-to-face interview or physical examination in order to limit the false positive rate compared to questionnaires administered by mail or telephone. The assessment of RLS was not always performed according to IRLSSG criteria and anyway the four diagnostic criteria did not allow the exclusion of other disorders that may act as mimics (Hening et al., 2009; Allen et al., 2014). Disease assessment ranged from a self-reported diagnosis, information on the use of specific medications, or a direct measurement of BP and blood glucose levels. Moreover, some antihypertensive medications, such as beta-blockers and certain calcium channel blockers, could both ameliorate and aggravate RLS symptoms (Innes et al., 2012) and therefore it would be important to consider medications as confounding factors. In addition, the co-occurrence of several CVRFs is frequent and they may influence each other. Therefore, the cross-sectional nature of most studies cannot assess the causal relationship between them and the variables of interest (i.e., RLS and/ or CVDs). Finally, only few studies adjusted their analyses for other cardiovascular risk factors, such as diabetes mellitus, history of myocardial infarction, BMI, dyslipidemia, and smoking status, that might act as confounders or mediators. In summary, longitudinal population-based studies and meta-analyses will be necessary in order to build a sufficiently robust body of evidence on this topic.