Abstract Background Thioguanine (TG) is a thiopurine which has been used for patients with inflammatory bowel disease (IBD), who have failed azathioprine (AZA) or mercaptopurine (MP) due to adverse events or suboptimal response. Its widespread use has been hampered due to concerns about nodular regenerative hyperplasia (NRH) of the liver. The aim of this study was to investigate the long-term efficacy and safety of low-dose TG therapy in IBD patients failing AZA and MP. Methods A retrospective multicentre study was performed in IBD patients who failed prior treatment with conventional thiopurines with or without following immunomodulation (thiopurine-allopurinol, biologicals, methotrexate, tacrolimus) and were subsequently treated with TG as rescue monotherapy between 2003 and 2019 at three hospitals in the United Kingdom. Clinical response, adverse events, laboratory results, imaging and liver biopsies were retrospectively collected. Results A total of 193 patients (57% female and 64% Crohn’s disease) were included, with a median daily TG dose of 20 mg (range: 20–40 mg), a median treatment duration of 23 months (IQR 10–47) and a median follow-up of 36 months (IQR 22–53). The clinical response rate at 12 months was 65 and 54% remained on TG until the end of follow-up. Adverse events consisted primarily of elevated liver tests (6%), myelotoxicity (7%) and rash (5%). NRH was histologically diagnosed in two patients and two other patients (1%) developed non-cirrhotic portal hypertension. The median 6-TGN and TPMT levels were 953 pmol/8 × 10⁸ RBC (IQR 145–1761) and 47 mu/L (IQR 34.5–96). Conclusions Long-term follow-up suggests that TG can be an effective and well-tolerated therapy in more than half of difficult-to-treat and multi-therapy failing IBD patients. Findings of this study indicate that TG can be used safely and the occurrence of hepatotoxicity was low. The incidence rate of NRH was within the background incidence.