Abstract Background Video capsule endoscopy (VCE) is a non-invasive and accurate diagnostic modality that allows for direct visualization of the small intestinal mucosa in Crohn’s disease (CD). In rare cases, lesions not detected by other diagnostic modalities in patients suspected to have CD are solely identified through VCE. We aim to evaluate the clinical characteristics and long-term outcomes of these rare cases. Methods A retrospective multicentre matched case-control analysis was conducted in patients with CD across three hospitals in Korea, from January 2007 to April 2023. To perform a matched case-control study, we defined cases (VCE group) as patients with normal findings at colonoscopy and cross-sectional imaging, ultimately diagnosed with CD by VCE; the controls were patients with CD diagnosed with specific findings at endoscopy or cross-sectional imaging (non-VCE group). Controls were matched to cases with a ratio of 3:1 for sex, calendar year of diagnosis, and age at diagnosis. Clinical characteristics and outcomes between the two groups were compared. Results Of the 3,137 patients diagnosed with CD during the study period, 24 were diagnosed solely by VCE. When compared to 72 matched patients with CD, the mean duration from symptom onset to diagnosis did not significantly differ (27.2 months [SD 10.2] vs. 25.3 months [SD 8.9], p=0.394). There was a significant difference in the proportion of Montreal behaviour (p<0.001). In the VCE group, non-stricturing and non-penetrating (B1) type and stricturing (B2) type accounted for 91.7 and 8.3%, respectively. However, there was no significant difference in the perianal fistula modifier (25% vs. 33.3%, p=0.446). Within the VCE group, 45.8% underwent the examination due to abdominal pain with diarrhoea, followed by 25.0% with perianal disease, and 20.8% with unexplained iron deficiency anaemia. Regarding capsule findings, the ileum was mostly involved (87.5%), while the jejunum and duodenum were affected in 70.8 and 16.7% of cases, respectively. Fifteen patients had two or three segments affected. The Median Lewis score was 838 [IQR 393–1803]. Concerning the cumulative incidence of clinical outcomes over 10 years, complicated behaviour, need for biologics, and CD-related hospitalization and surgery were all significantly lower in the VCE group than in the non-VCE group (p<0.005). Conclusion Patients diagnosed with CD solely by VCE are rare. In this study, they exhibited different clinical characteristics and a more favourable long-term disease course compared to the general population of patients with CD. This included a lower incidence of complicated behaviour, reduced need for biologics, and CD-related hospitalization and surgery.