Arterio‐venous blood gas Δvalues for validation of umbilical cord blood samples at birth are not only biased by sample mix ups but also affected by clinical factors

Abstract

Introduction Traditional validation of umbilical cord blood samples with positive veno‐arterial ΔpH and arterio‐venous ΔpCO2 values confirms the source of samples, whereas negative Δvalues represent mix‐up of samples. To investigate whether this is true, the distributions of V‐A ΔpO2 and A‐V Δlactate were also explored and related to clinical characteristics. In addition, different cord blood sampling techniques were evaluated. Material and methods Register study with cord blood acid‐base and clinical data from 27 233 newborns. Clinical characteristics were related to positive, zero and negative Δvalues. Blood samplings from unclamped and double‐clamped cords were compared. A two‐sided P < 0.05 was considered significant. Results ΔpH and ΔpCO2 values distributed into positive, around zero, and negative sub‐populations, with significant differences in pH and clinical characteristics between sub‐populations. No such sub‐populations were distinguished for ΔpO2 and Δlactate. The 2.5th and 5th ΔpH percentiles were 0.013 and 0.022, respectively, and for ΔpCO2 0.30 and 0.53 kPa. Applying 5th percentile criteria resulted in 3.5% of “approved” cases showing a ΔpO2 ≤ 0. Puncture and sampling of the unclamped cord resulted in significantly better sample quality. Conclusions Unphysiological negative ΔpO2 values occurred despite correct validation with traditional criteria. Δlactate cannot be used for validation because both positive and negative values are physiological. Positive/around zero/negative ΔpH and ΔpCO2 sub‐populations were associated with significant differences in pH and clinical characteristics, indicating that defective sampling and sample handling are not the sole explanations for negative Δvalues. Prompt puncture and sampling of the unclamped cord resulted in best sample quality.

Proposed cutoff for fetal scalp blood lactate in intrapartum fetal surveillance based on neonatal outcomes: a large prospective observational study

“…As venous pH values are usually less acidotic than arterial, venous levels meeting the cutoff criteria for arterial samples are evidently pathological. 22 Secondary outcome measures included alternative combinations of the blood gas parameters, Apgar scores and neonatal outcomes: neonatal intensive care unit admission (NICU), continuous positive airway pressure therapy, bilevel positive airway pressure or ventilation. There is no international consensus on the definition of MA, so we chose to present cutoffs for several predefined combinations of blood gas parameters used in previous studies and defined by scientific societies.…”

“…If arterial cord samples were missing, the venous samples were used with the same cutoffs. As venous pH values are usually less acidotic than arterial, venous levels meeting the cutoff criteria for arterial samples are evidently pathological 22 …”

Proposed cut-off for fetal scalp blood lactate in intrapartum fetal surveillance based on neonatal outcomes: a prospective observational study

“…If arterial samples were missing, venous samples were used with the same cut-offs as arterial. Since venous pH values are usually less acidotic than arterial, venous levels meeting cut-off criteria for arterial samples are evidently pathological (19).…”

Gestational age‐related reference values for Apgar score and umbilical cord arterial and venous pH in preterm and term newborns

“…Overall, 207 684 sample analyses were retrieved from the blood gas analyzers from 2001 to 2010. Samples not identified as originating from umbilical cord blood, not being paired as arterial‐venous samples, not fulfilling the validation requirement that venous pH (VpH) should be at least 0.020 units higher than ApH, and where corresponding clinical data could not be retrieved, were excluded. By using maternal personal identification numbers for matching of clinical data from the regional Perinatal Revision South Register, we could identify 27 175 cases for the study.…”