Survival of extremely premature infants has been significantly higher in th
e last decade than previously, and may well have improved. The majority of
infants greater than or equal to 25 weeks' gestation survive today. Surviva
l of infants 23 and 24 weeks' gestation is significantly lower, but is by n
o means negligible. Reports of survival of infants less than 22 weeks or 50
0-g birth weight are not unique. Moreover, the maximum survival of infants
less than or equal to 25 weeks with current state-of-the-art care is not kn
own. Currently available data do not allow survival of the individual extre
mely low-birth weight or extremely premature infant to be predicted with cl
inically acceptable accuracy. The concept of a limit of viability is vague
and clinically and ethically simplistic. The provision of neonatal intensiv
e care is not necessarily beneficial or justified merely because it affords
some chance of survival.