Introduction

Lower gastrointestinal haemorrhage (LGIH) is a challenging phenomenon in a comorbid, elderly population. CT mesenteric angiography (CTMA) allows localisation of the site of haemorrhage, and provides a target for interventional techniques, but the intermittent nature of LGIH makes it challenging to reliably demonstrate extravasation. This study aimed to identify objective factors that may predict scan outcomes.

Methods

In this retrospective cohort study, all patients undergoing CTMA for LGIH at Monash Health from January 2011 to December 2019 (n = 854) were included. Baseline patient characteristics included age, bowel resection/endoscopic intervention within the past 14 days, known bowel malignancy, anticoagulant/antiplatelet use, duration of symptoms, vital signs, transfusion requirements in the past 24 h and investigation results (recent haemoglobin levels, platelet count, international normalised ratio and creatinine levels). Univariate analysis was performed, and significant factors were entered into a multivariate model.

Results

The final multivariate model was statistically significant (P < 0.001) and consisted of bowel resection/endoscopic intervention within the past 14 days (OR = 2.15), use of antiplatelet agents (OR = 2.03), blood transfusion requirement greater than 3 units per 24 h (OR = 1.79), systolic blood pressure less than 100 mmHg (OR = 1.56) and heart rate greater than 100 beats per minute (OR = 1.52).

Conclusion

The factors identified above are objective, independently associated with positive scan outcomes, readily available to radiologists and are useful for more judicious patient selection.