Background The prevalence of systemic lupus erythematosus (SLE) reported in European studies has varied widely. It will be useful to have good regional estimates of the prevalence of SLE to inform public health plans and resource allocation.
Objectives To determine if a single estimate of the prevalence of SLE in Europe can be obtained from the critical review of European based studies. If not, the sources of variation will be described.
Methods A broad literature search of EMBASE and MEDLINE content up until 28/09/2011 was conducted using the terms “systemic lupus erythematosus” and “prevalence”. To be included each study must (1)be a primary research study (2) assess point or period prevalence of SLE after 1982 (when American College of Rheumatology classification criteria for SLE1were published), (3)be set within a European country or region, (4) not restrict by criteria other than age. All identified titles were screened by hand and relevant manuscripts were reviewed; citations from these were reviewed to identify further papers. Studies meeting inclusion criteria were rated for quality (“high”, “intermediate” or “low”) based on case ascertainment methods, diagnostic criteria and representativeness. Random-effects meta-analysis to assess heterogeneity and to produce a pooled prevalence estimate was performed on “high” and “intermediate” quality studies using the software StatsDirect.
Results The search yielded 2,949 citations; 58 manuscripts were reviewed by hand. A final 21 studies met the inclusion criteria, from which 11 of “high” and 2 of “intermediate” quality were selected for the meta-analysis. Prevalence estimates ranged from 18 per 100,000 to 71 per 100,000. One high quality study was dropped from the meta-analysis due to lack of a denominator figure in the text. The pooled random-effects prevalence estimate from the remaining 12 studies was 40 per 100,000 (95% CI 34, 46). Heterogeneity statistics (I2)indicated a very high degree of variability among the study estimates (Cochran Q=180.1, df =12, p<0.0001; I2 =93.3%). Limiting the analysis to high quality studies only (pooled estimate 39 per 100,000; Q=160.3, df=9, p<0.0001; I2=94.4%), to studies published after 1999 (46 per 100,000; Q=76.1, df=5, p<0.0001; I2=93.4%), to region (Northern Europe: 37 per 100,000; Q=131.5, df=8, p<0.0001; I2=93.9%; Southern Europe: 47 per 100,000; Q=40.5, df=3, p<0.0001; I2=92.6%) or to studies with populations aged >15 only (44 per 100,000; Q=140.4, df=7, p<0.0001; I2=95.0%) did not reduce the variability among estimates. The estimate of prevalence was higher in Southern than in Northern Europe. However, this may be explained by date of publication; Southern European studies have been published later and the prevalence may have increased over time.
Conclusions The best estimate of the European SLE prevalence is 40 per 100,000 based on the current studies, although this estimate is still associated with substantial variability. Subsetting the studies into potentially more homogeneous groups (e.g. on time period, region or age) did not decrease the variability.
Tan EM, Cohen AS, Fries JF. The 1982 revised criteria for the classification of systemic lupus erythrematosus. Arthritis Rheum 1982;25(11):1271-1277.
Disclosure of Interest J. Davidson Shareholder of: GlaxoSmithKline plc, Employee of: GlaxoSmithKline R&D, N. Galwey Shareholder of: Glaxosmithkline plc, Employee of: Glaxosmithkline R&D, P. Egger Shareholder of: GlaxoSmithKline plc, Employee of: Glaxosmithkline R&D