Clinically significant bleeding complications occur in >30% of patients undergoing percutaneous renal biopsy (PRB) of native kidneys and can be severe in up to 10% of patients. A noninvasive measure that would reliably predict which patients will do well with an uncomplicated post-biopsy course or which patients may be at risk of developing a clinically significant complication is in great demand.

PRB of native kidneys was performed in 162 adult patients from February 2002 through February 2007 using real-time ultrasound and automated needle. Renal ultrasound (US) was performed at 1-h post-PRB to assess biopsy-related bleeding. Patients were observed for 24 h post-PRB to monitor clinically apparent biopsy-related complications. The value of the post-biopsy ultrasound in predicting complications was assessed.

A clinically apparent complication was observed in 26 (16%) patients post-PRB (13 minor not requiring any intervention and 13 major requiring intervention). In patients with complicated courses, a haematoma at 1 h was seen in 77% (69% with minor and 87% with major complications). However, only 27 (20%) of 136 patients without complications (P < 0.0001) had a haematoma at 1 h. The presence of a haematoma 1-h post-PRB had a sensitivity of 77%, specificity of 80%, positive predictive value of 43% but a negative predictive value of 95% for predicting clinical complications.

We find that with the use of renal ultrasound 1-h post-PRB, the absence of perinephric bleeding is predictive of an uncomplicated course while the presence of a perinephric haematoma is not reliably predictive of a clinically significant complication post-renal biopsy.