Nosocomial infections (NIs) are a public health problem. Close monitor
ing is essential to the prevention of NIs. A prospective study was con
ducted in a neonatal unit to evaluate the number of neonates treated v
ia intravenous (IV) lines over a one-year period, to determine the nat
ure and incidence of IV line-related infections, and to provide data o
f use for improving the quality of care. Population and methods. all n
eonates who had an IV line started in the unit between February 15, 19
92, and February 14, 1993, were included in the study. The IV line inf
usion protocol had been written several years earlier and included uni
versal precautions. A monitoring sheet was completed for each infant a
nd each IV line. Results: An IV line was started in 313 infants, i.e.,
56.2% of the total population. Mean birth weight was 2405g (820-4560)
, mean gestational age was 35.5 weeks (25-42), and 57.8% of infants we
re preterm. One thousand IV lines were set up and monitored; mean numb
er of IV lines per infant was 3.12. A short catheter was used in 841 c
ases and an epicranial needle in 146. Mean number of interventions per
line was 3.95. Mean running lime was 25.59 h/line, for a mean of 2.91
days of infusion/day for one year. Most patients (96.9%) remained fre
e of local complications; there were 26 cases of local inflammation an
d two of lymphangitis. No IV line-related systemic infections were rec
orded. Conclusion: these data are objective evidence that the IV line
protocol used for this study is effective in preventing systemic infec
tions and local incidents. Stringent adherence to this protocol is ess
ential.