Background The Apgar score has been shown to be predictive of neonatal mortality in clinical and population studies, but has not been used for international comparisons. We examined population‐level distributions in Apgar scores and associations with neonatal mortality in Europe. Methods Aggregate data on the 5 minute Apgar score for live births and neonatal mortality rates from countries participating in the Euro‐Peristat project in 2004 and 2010 were analysed. Country level associations between the Apgar score and neonatal mortality were assessed using the Spearman rank correlation coefficient. Results Twenty‐three countries or regions provided data on Apgar at 5 minutes, covering 2 183 472 live births. Scores <7 ranged from 0.3% to 2.4% across countries in 2004 and 2010 and were correlated over time (ρ = 0.88, P < 0.01). There were large differences in healthy baby scores: scores of 10 ranged from 8.8% to 92.7% whereas scores of 9 or 10 ranged from 72.9% to 96.8%. Countries more likely to score 10 s, as opposed to 9 s, for healthy babies had lower proportions of Apgar <7 (ρ = −0.43, P = 0.04). Neonatal mortality rates were weakly correlated with Apgar score <7 (ρ = −0.06, P = 0.61), but differences over time in these two indicators were correlated (ρ =0.56, P = 0.02). Conclusions Large variations in the distribution of Apgar scores likely due to national scoring practices make the Apgar score an unsuitable indicator for benchmarking newborn health across countries. However, country‐level trends over time in the Apgar score may reflect real changes and merit further investigation.