Objective To identify the distribution of carbon dioxide tension (pCO2) relative to pH in validated umbilical cord acid–base data. Design Observational study. Setting European hospital labour wards. Population Data for 36 432 term newborns were obtained from three sources: two trials of fetal monitoring with electrocardiography (ECG; the Swedish randomised controlled trial and the European Union Fetal ECG trial) and data from Mölndal Hospital. Methods From the total study population, cases with missing values or obvious typing errors were excluded. The remaining data were validated based on specified criteria. Percentiles of arterial pCO2 by pH were calculated using multilevel regression modelling. Main outcome measures Umbilical cord pH, pCO2 and base deficit. Results Acid–base values were considered invalid in one out of seven cases. Percentiles for arterial pCO2 corresponding to specified values of arterial pH were developed from the validated data of 26 690 cases. Conclusions Percentiles for arterial pCO2 for a specified arterial pH can be used as a tool to identify cases with erroneously low pCO2 values, and thus avoid an incorrect interpretation of the newborn’s acid–base status.