Development and validation of a nomogram for spontaneous closure of duodenal fistulas after infection control in septic patients.

BACKGROUND

In septic patients, the rate of spontaneous closure is lower than in non-infected patients. Few studies have focused on predicting spontaneous closure after sepsis control, which is crucial for guiding treatment decisions such as surgery or observation. This study developed and validated a nomogram to predict the likelihood of spontaneous closure in external duodenal fistulas after infection control in septic patients.

METHODS

Between January 2004 and October 2024, patients from centers A and B were included in the development cohort, while those from centers C and D constituted the external validation cohort. Key risk factors for spontaneous closure were identified and incorporated into the nomogram. Its performance was evaluated using calibration and decision curve.

RESULTS

A total of 449 patients were included in the study: 342 in the development cohort and 107 in the external validation cohort. The rates of spontaneous closure were 53.2% and 57%, respectively. Six significant factors influencing closure were identified: the interval between fistula occurrence and infection control, the extent of infection invading the retroperitoneum, the occurrence of emergency laparotomy, fistula diameter greater than 2 cm, duodenal decompression, and the need for intermittent albumin transfusions after infection control. The nomogram demonstrated strong predictive accuracy, with C-indices of 0.82 for the development cohort and 0.76 for the validation cohort. An optimal cut-off score of 155 exhibited high sensitivity and specificity across both cohorts.

CONCLUSION

This study introduces a practical nomogram to assess the risk of spontaneous closure in external duodenal fistulas after infection control in septic patients.