Background: There is no consensus definition to distinguish infectious from non-infectious phlebitis associated to peripheral venous catheter. Aims: To evaluate the impact of an intervention program on the rate of infectious (those associated to bacteremia or local purulent discharge) and non-infectious phlebitis (the rest) and identify differential features. Methods: Interventional study developed in three stages: basal measurement, intervention, and evaluation. Results: Ten infectious and 186 non-infectious phlebitis were registered. Infectious phlebitis diminished after intervention (0.2 to 0.04 events per 1,000 bed-days; p = 0.02) but not the rest (2.3 per 1,000 bed-days before and after). Five of 10 patients with infectious phlebitis had bacteremia, one with infectious endocarditis and valve replacement, and other with septic shock and a fatal outcome. None of the patients in the non-infectious group presented complications. Infectious phlebitis appeared later (mean 4.1 versus 2.4 days; p = 0.007) and were associated to fever (40% vs 5.9%, p = 0.004). Non-infectious phlebitis was associated to irritating compounds (OR 6.1; IC95 1.3-29, p < 0.05). Conclusions: The intervention program demonstrated a favorable impact only on the rate of infectious phlebitis. Distinction appears to be relevant because those of infectious origin are associated with fever, complications or death, respond to an intervention program, and emerge lately.