Background
A repeat biopsy is recommended, but often omitted in coeliac disease patients on a gluten‐free diet. The effect of performing or not performing repeat biopsies is currently unknown.
Aim
To identify factors associated with and the significance of lacking biopsy for long‐term outcome. Predictors and the importance of incomplete histological recovery after 1 year was investigated in re‐biopsied patients.
Methods
A total of 760 patients participated in a nationwide follow‐up study. Medical data were gathered via interviews and patient records, and blood samples were drawn for serology. Current symptoms and well‐being were assessed by validated PGWB, SF‐36 and GSRS questionnaires.
Results
Malabsorption was more common among those with a repeat biopsy (46%) than those without repeat biopsy (33%), P < 0.001, as were severe symptoms at diagnosis (24% vs. 16%, P = 0.05) and concomitant gastrointestinal (40% vs. 32%, P = 0.049) or musculoskeletal (35% vs. 27%, P = 0.023) diseases such as arthritis, osteoporosis and back pain. Repeat biopsy was more rare in subjects diagnosed in private care (11% vs. 23%, P < 0.001) or by screening (10% vs. 16%, P = 0.010). The groups were comparable as to current symptoms and dietary adherence, but those without re‐biopsy were less confident of their diet (89% vs. 94%, P = 0.002) and more often seropositive on diet (14% vs. 9%, P = 0.012). They reported better SF‐36 physical functioning (P = 0.043) and less pain and indigestion (P = 0.013 and P = 0.046 respectively) and total GSRS (P = 0.052) score. Incomplete mucosal recovery was predicted by more advanced histological (P < 0.001) and serological (P = 0.001) disease at diagnosis, whereas the groups did not differ in long‐term adherence, symptoms, seropositivity, questionnaire scores, frequency of fractures or malignancies.
Conclusions
Severe disease at diagnosis predicted the record of a repeat biopsy and incomplete mucosal recovery. Neither lacking biopsy nor incomplete recovery in a relative short time span of 1 year was associated with poorer long‐term outcome, although survival bias cannot be excluded.