Incidence of infectious complications after an ultrasound-guided intervention

OBJECTIVE. The objective of our study was to determine the incidence of infectious complications of common ultrasound-guided procedures including fine-needle aspiration (FNA), drain placement, biopsy, pseudoaneurysm thrombin injection, thoracentesis, and paracentesis. SUBJECTS AND METHODS. The infection prevention and control (IPAC) committee at the Mayo Clinic, Rochester, MN, conducts surveillance of selected infections including radiology procedures. When a positive culture, hospital admission, or operating room visit for infection is identified, the patient's electronic records are thoroughly reviewed by an infection control practitioner looking for information about prior interventions. Similarly, the department of radiology prospectively follows all patients who have undergone ultrasound-guided hepatic, renal, and pancreatic biopsies for complications 24 hours, 3 months, and 12 months after biopsy. We reviewed 2 years of these data to determine the incidence of infections after common ultrasound-guided procedures. RESULTS. We performed 13,534 ultrasound-guided procedures from January 2006 to December 2007. There were 11 likely and three possible procedure-related infections for an overall incidence of 0.1% (14/13,534). The infections consisted of five abscesses, four bloodstream infections, four cases of peritonitis, and one urinary tract infection. The highest incidence of infections occurred after ultrasound-guided biopsy (0.2%, 10/5,487), with biopsy of a hepatic transplant having the highest incidence (1.0%, 2/192). No infections occurred after thoracentesis and FNA despite the large number of procedures performed (2,489 and 2,340, respectively). Nearly all patients improved on antibiotics. One patient died 5 days after paracentesis; however, death was likely due to multiorgan failure in the setting of fulminant liver failure with hepatorenal syndrome. CONCLUSION. The incidence of a serious infectious complication after ultrasoundguided intervention is low. Radiologists can use these data to provide more accurate information to patients when asking for consent before procedures and to reassure their patients.