Background Despite many efforts, to date there has been no focused attempt to derive a robust estimate of the prevalence of rheumatic diseases (RDs) to quantify how this is influenced by other factors than them examined in every local study, however, the problems magnitude is rising and due the demographic transition and the increase in the life expectancy
Objectives To determine, through a systematic review and meta-analysis, the prevalence of RDs in the adult general population and explore its heterogeneity
Methods MEDLINE, EMBASE, BIREME, LILLACS and Web of Science were searched using a search strategy combining key words and related database-specific subject terms to identify relevant cross-sectional based on COPCORD methodology studies. Also was developed a manual search. Included articles were assessed for risk of bias and quality based on the STROBE statement. Prevalence figures for RDs were analyzed according to female percentage of sampled individuals, mean age and sample size. A mixed effect model was used to obtain the combined prevalence and a meta-regression to estimate the effects of other variables
Results 44 out from 127 papers were included in English, Spanish or Portuguese. Estimates for any RDs prevalence ranged from 7.2% to 62.3% (26.9%; 95% CI 18.3%>25.6%). For rheumatoid arthritis (RA), the prevalence varied between 0.2% and 6.2% (1.04%; 95% CI 0.4%>1.6%); fibromyalgia (FM) had a mean prevalence of 2.1% (95% CI 1.0%>3.2%) and osteoarthritis: 13.5% (95% CI: 10.6%>16.4%). SLE (systematic lupus erythematosus) was the less frequent condition with average prevalence of 0.14% (95% CI: 0.005%>0.28%). The random-effects pooled prevalence for any RDs was 25% (95% CI: 18.0%, 31.1%). Prevalence was higher in studies with bigger sample size (random effect coefficient: 0.0014, p=002). There was evidence of relevant heterogeneity in the analysis (p<0,001) and for RDs, RA and FM the sample size was positively associated to the perceived heterogeneity. No effects were found for SLE
Conclusions It was found significant variation among the prevalence across this review, in particular, related to the sample size used in each study. Two facts must be accounted for, first the statistical difficulties associated to the estimation of small prevalence and the consequent heterogeneity of the estimates, and, second, the limited number of studies included in this meta-analysis. Nonetheless, there is evidence about big heterogeneity what can correspond to non-observed variables, in particular, life-styles, and environmental or genetic traits.
Disclosure of Interest None declared