Background This study aimed to validate and modify the recently released Chiang Mai University Intussusception (CMUI) scoring system in predicting failure of non-surgical management of infantile intussusception.

Methods A retrospective review was conducted in 151 cases of infantile intussusception who were primarily treated with pneumatic reduction in our institute during 2008–2018. The analysis focused on the validation of the CMUI scoring system. Following this analysis, the scoring system was modified to be more suitable for our clinical practice, in which we perform pneumatic reduction in all cases.

Results Pneumatic reduction was successfully performed in 120/151 cases (79.5%). A high CMUI score was significantly associated with increased likelihood of failure at the positive likelihood ratio of 1.49 (sensitivity 25.8 and specificity 82.7). The area under the receiver operating characteristic curve (AUC) was 0.73. We modified the CMUI System in two ways, first by removing the item ‘method of reduction’ and replacing it with ‘hyponatremia’, and by changing the definition of low body weight to less than 9 kg. The modified CMUI had an AUC of 0.76. A high score (>9 points) on the modified version gave a positive likelihood ratio of 4.77 (sensitivity 53.0 and specificity 80.9).

Conclusion In infantile intussusception primarily treated with pneumatic reduction, the modified CMUI scoring system gave a better prediction reliability than the original.