Randomised controlled trials (RCTs) or quasi-randomised trials that assessed in-line filters and reported sufficient data to calculate the risk of phlebitis in both treatment and control groups were eligible for inclusion. Studies needed to define phlebitis.
Study populations included surgical, medical, general hospital, paediatric oncology and cardiac patients. Patients were assumed to have received short peripheral catheters (<7.5cm) except in two trials that reported using long peripheral catheters (7.5cm to 20cm). Intravenous solutions included hyperalimentation, isotonic glucose, cephalothin, clear fluids, buffered solutions and "various kinds of fluids and additives". In-line filters ranged from 0.22μm to 0.5μm. Control groups received dummy filters or no filter. Phlebitis was diagnosed based on various clinical findings.
Two reviewers independently assessed studies for inclusion. Disagreements were resolved through consensus.