Can the Apgar score be used for international comparisons of newborn health?

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 10s, as opposed to 9s, 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 Apgar an unsuitable indicator for benchmarking newborn health across countries. However, country-level trends over time in Apgar may reflect real changes and merit further investigation.