Answers are grounded in this paper and cite their sources.
Dear Editor, We would like to respond to the article ‘Prevalence and characteristics of extra-intestinal manifestations in a large cohort of Greek patients with inflammatory bowel disease’ by Karmiris et al.,1 and indirectly to epidemiologic inflammatory bowel disease [IBD] studies derived from other hospital-based [HB] settings. Albeit Karmiris et al.’s study is performed elegantly and their findings are clearly of importance, we would like to emphasize that HB settings may be associated with an overrepresentation of IBD patients with a more severe disease course, due to less frequent outpatient visits of mildly diseased IBD patients. To illustrate this, we performed a small systematic review with meta-analysis to compare population-based [PB] settings with HB settings with respect to phenotype characteristics at diagnosis that are associated with severe disease course, namely, extensive colitis for ulcerative colitis [UC]2,3 and perianal disease for Crohn’s disease [CD].4,5 Pubmed and Embase were searched for studies meeting the following conditions: PB or HB studies on IBD [A study was defined as PB, when the authors strived to capture all patients within a well-defined region and timeframe. HB studies included the total hospital registration or a large and unselected sample hereof with a cut-off of at least 250 patients]; IBD diagnosis by Lennard-Jones criteria; phenotype assessment at diagnosis by Montreal classification; studies on adult patients [including combined cohorts on adult and paediatric patients]; studies published between 2000 and 2015 and written in English language. The proportions of patients with extensive colitis or perianal disease at diagnosis were extracted, then pooled for PB and HB settings separately, and finally compared by random effects metaregression analysis [α = 0.05]. Sensitivity analyses comprised cohorts from Europe, from high incident countries, with prospective patient inclusion, and adult cohorts. The forest plot clearly showed that the pooled proportion of UC patients with extensive colitis at diagnosis, as well as of CD patients with perianal disease at diagnosis, was significantly higher in HB than PB settings [Figure 1]. As these phenotype characteristics are associated with severe disease course, these results indicate that HB settings are prone to have relatively more patients with a severe IBD disease course compared with PB settings. The sensitivity analyses confirmed our findings for CD, but not for UC [data not shown]. It should, however, be noted that heterogeneity remained high in all analyses, and that some HB cohorts do resemble PB cohorts in terms of patients with phenotype characteristics at diagnosis that are associated with severe disease course. Forest plots of population-based [PB] and hospital-based [HB] IBD studies in the metaregression analysis for ulcerative colitis and Crohn’s disease. Results remained significant when the outlying data of Jiang were excluded. In conclusion, in this letter we emphasized that epidemiological data on IBD from HB settings may suffer from external validity, as they may comprise an overrepresentation of patients with phenotype characteristics at diagnosis that are associated with severe disease course. This should be taken into account when initiating studies and interpreting epidemiologic findings. This work was supported by funding from the European Union Seventh Framework Program [grant agreement number 305564]. The authors have no conflicts of interest to declare. TH was involved in designing the study; collecting, analysing and interpreting data; and drafting the manuscript. MP and DJ were involved in designing the study; critically revising the manuscript; and giving final approval for submission.
No takes yet. Share an insight, caveat, or question.
Heuvel et al. (2016) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: