Objective: The Apgar score should be an objective method to assess the state of newborns; however, its applicability in preterm infants is hampered by large variations among different observers. The study tested whether physicians that give low scores to written case descriptions also apply lower scores to preterm infants. Patients and Methods: Descriptions (BMJ 2004; 329: 143-4) were sent to 14 neonatal units. Physicians were asked to evaluate the Apgar (case score). From seven units Apgar scores of all very low birth weight infants (VLBW) born between January 2004 and December 2006 were obtained from charts (clinical score). Results: In total, 121 physicians from 14 institutions (median 9, range 3-15) replied: 24 residents with <6-month and 28 with >6-month neonatal experience, and 69 consultants. The assessment of the case scores was very heterogeneous with large variations in respiration, muscle tone and reflexes. Clinical scores were obtained from 1000 VLBW infants. The score depended on the gestational age, with a median of 4 at 24 and 7 at 27 weeks. With one exception, centres that assigned low case scores had also low clinical scores. Conclusion: There is considerable variation in assigning Apgar scores. Definitions are required to apply the Apgar score to infants under clinical conditions such as preterm delivery, resuscitation or artificial ventilation.