In some parts of the world, policy recommends cord blood samples are routinely collected at all births for analysis of either gases or lactate. For example, in my part of the world the Royal Australian and New Zealand College of Obstetricians and Gynaecologists policy says the following:
Paired umbilical cord blood gas or lactate analysis should be taken at delivery either routinely or when any of the following are present:
RANZCOG, 2019, p. 25
- Apgar score < 4 at 1 minute
- Apgar score < 7 at 5 minutes
- Fetal scalp sampling performed in labour
- Operative delivery undertaken for fetal compromise
They evidence they cited to support this recommendation for routine blood gas analysis came from a paper by White, et al. published in 2014, and another by Allanson, et al., published in 2018. Regular readers of my blog will know this particular guideline is far from a shining beacon of evidence based practice, so let’s see how they did this time.
Cord gas analysis in Perth, Western Australia in the 2000s
The RANZCOG guideline describes research conducted in a tertiary hospital with data collection occurring between January 2003 and December 2006. They then cite the 2014 paper by White, et al. so I presumed this was where this research was published. Ah no. The 2014 paper describes research about cord blood gas analysis conducted in a primary and two secondary hospitals around 2010. So RANZCOG is off to a flying start here with attention to detail.
I was able to identify the paper they were describing – White, et al., 2010. This research was undertaken at the King Edward Memorial Hospital in the period described in the RANZCOG guideline. During this period the hospital had a policy to collect paired cord blood gas samples at all live births. The policy was clearly enacted, with at least one cord blood gas sample collected for 97% of all births during this period.
Umbilical artery pH levels showed a statistically significant rise over time. But it is important to note here that the absolute difference was tiny and not clinically relevant – rising from 7.253 in 2003 to 7.259 in 2006. A similar small, not clinically relevant, but statistically significant fall in lactate levels also occurred over time, from 4.29 in 2003 to 3.77 in 2006. Other research linking cord gas results with neonatal outcomes shows only weak correlations between the two (Johnson, et al., 2021; Leinonen, et al., 2019). So it is important to see what changed in relation to neonatal clinical outcomes.
The rate of admissions to the special care nursery fell slightly over time, from 18% in 2003 to 15% in 2006. Intensive care nursery admission rates rose marginally, from 6% to 7% but this difference didn’t reach statistical significance. There were no differences in rates of neonatal mortality, hypoxic ischaemic encephalopathy, or the use of resuscitation at birth. The rate of Apgar scores under 4 at 5 minutes showed a statistically significant fall over time, from 1.8% in 2003 to 0.5% in 2006.
In summary, there were small improvements over time in surrogate endpoints (Apgar scores, pH levels) but little change in important clinical outcomes (death, hypoxic ischaemic encephalopathy). Only short term data were collected, so this study was unable to provide evidence about whether there were benefits, or not, after hospital discharge. It also struck me as odd that they chose not to compare a period prior to the enactment of the new policy, when cord blood gases were collected less often, with the period after the new policy was introduced. This would have provided more conclusive evidence about whether introducing the policy was the thing that made the difference, or whether there were other things going on between 2003 and 2006 that might have produced the findings reported here.
And that’s the real story here. What wasn’t mention in the paper, or in the RANZCOG guideline, was the context in which this research was conducted. There had been significant issues with clinical care at King Edward Memorial Hospital, culminating in the Douglas enquiry. In December 2001, the enquiry group tabled a list of 237 recommendations, and there was ongoing oversight to ensure these were enacted until April 2005. So it is safe to say there were a LOT of changes, other than just a new policy about cord blood gases, that might have modified the outcomes of maternity care during the period 2003 to 2006. Context is everything.
What about the papers that were cited?
Most of the authors from the White, et al., 2010 paper also contributed to their 2014 paper. A similar approach was used, but this time three hospitals, one providing primary care and the other two secondary care, all in Western Australia adopted a policy of universal cord blood collection at birth. Two units tested only lactate, and the other the full panel of blood gases. I found it difficult to get their time line straight. The authors speak of four time epochs, but the dates that mark the start and end of each were never provided. The only information about timing was that Unit A had 1,073 births between December 2009 and February 2011, Unit B had 1,891 births between July 2009 and September 2011, and Unit C had 3,340 births between July 2009 and June 2011. I am unable to tell whether any of the time epochs represent a period before the introduction of the new policy.
The findings don’t make a compelling argument in favour of routine testing. At Unit A (regional primary centre) there was no change in lactate levels but the incidence of abnormal levels fell over time. At Unit B (regional secondary centre) lactate levels rose over time but there was no change in the incidence of abnormal levels. And at Unit C (metropolitan secondary centre) there no change in pH, base excess or lactate levels over time, and no change in abnormal levels. The only clinical outcome assessed was Apgars scores, and there was no change over time. Despite these findings, the authors concluded that routine assessment of cord blood showed benefit.
The other paper cited came from South Africa. Allenson et al. (2018) set out to assess the feasibility of introducing umbilical cord lactate assessment in a tertiary centre. Despite the policy of routine collection, cord lactate results were available for only 20% of births. The authors examined the sensitivity (when the test is abnormal how often do poor outcomes occur?) and specificity (when the test is normal how often do good outcomes occur?) of the test in relation to the use of resuscitation, nursery admission, and Apgar score of less than 7 at 5 minutes of age. This study did not make any attempt to assess whether the use of cord lactate improved perinatal outcomes and therefore doesn’t assist in supporting the recommendation in the RANZCOG policy.
So routine cord blood gas testing joins our list of practices for which RANZCOG fails to generate an argument based in sound evidence.
How good is cord blood gas assessment at predicting poor outcomes?
The thing that sparked my interest in what RANZCOG had recommended and how they built the “truth” of their argument was new research from Israel (Gonen, et al., 2023). This study retrospectively examined cord blood gas results collected at births during the period 2014 – 2022, at a hospital with a policy of routine sampling. Only those from “low risk deliveries” were included, defined as singleton non-instrumental vaginal births between 37 and 41 weeks, without congenital abnormality, chorioamnionitis, preeclampsia, low birth weight, stillbirth, or meconium stained liquor. The ability of a pH of less than 7.1 alone or in combination with a base excess of less than -12 mmol/L to predict a composite outcome including neonatal sepsis, necrotising enterocolitis, phototherapy, transfusion, neurological morbidity, or respiratory morbidity was assessed.
Data were available for 14,338 births. A pH level of under 7.1 occurred at a rate of 11 per 10,000 births, and this low pH level in combination with a base excess of less than -12 mmol/L occurred at a rate of 3 per 10,000 births. The primary composite outcome was more common in infants with a low pH at 7% vs 1% with a pH above 7.1, and in combination with base excess the rates were 25% and 1%, with both reaching statistical significance. While the negative predictive values (the ability of a normal test to predict a normal outcome) were high, this is hardly surprising given the low rate of the composite outcome at 1.2% for the total population. The positive predictive values were low, at 7.1% for pH alone and 25% for pH plus base excess. In other words, most infants with one of the poor outcomes had normal test results. This was abundantly clear even in the raw data – there were 178 infants with a poor outcome and normal test results, and only one with a poor outcome and abnormal test results.
The authors argue that routine use of cord blood gas testing in low risk vaginal births is neither clinically nor financially effective. Their research, like the studies cited in the RANZCOG policy, didn’t set out to examine whether the introduction of routine cord blood gases as a policy modified outcomes in a clinically meaningful way, or not. While we therefore can’t say it doesn’t work, this new research highlights that routine cord blood gas assessment is highly unlikely to generate clinical benefit.
Why not do it anyway? It can’t hurt, right?
It could be argued that there remains the possibility that routine assessment of cod blood might still be beneficial, and we simply haven’t done the right research to show it is. Some might say it’s just a quick blood test on “left over” blood so there’s no possible harm to the baby in terms of testing – so why not do it on the off chance that it helps in some way.
While technically speaking, it is possible to sample cord blood with the cord intact, it is far more common for early cord clamping to take place to facilitate the collection of a cord blood sample. And early cord clamping isn’t harmless nor is it “left over” blood. As Andersson and Mercer (2021) remind us, this un-physiological practice has consequences –
- increased need for resuscitation at birth and longer time to establish regular respiration during resuscitation
- lower cerebral oxygenation at 12 hours of life
- lower iron stores, with higher rates of iron deficiency and anaemia in the first year of life
- reduced myelination in the brain at 4 and 12 months of age
- neurodevelopment consequences at 4 years of age, particularly in boys
Therefore, there is more potential for harm than benefit from the practice of routine cord blood collection. It is time our guideline writers reconsidered this practice.
References
Allanson, E. R., Pattinson, R. C., Nathan, E. A., & Dickinson, J. E. (2018, May). The introduction of umbilical cord lactate measurement and associated neonatal outcomes in a South African tertiary hospital labor ward. Journal of Maternal-Fetal & Neonatal Medicine, 31(10), 1272-1278. https://doi.org/10.1080/14767058.2017.1315094
Andersson, O., & Mercer, J. S. (2021). Cord management of the term newborn. Clinics in Perinatology, 48(3), 447-470. https://doi.org/10.1016/j.clp.2021.05.002
Gonen, N., Cohen, I., Gluck, O., Jhucha, D., Shmueli, A., Barda, G., Weiner, E., & Barber, E. (2023, Feb 19). Umbilical cord blood gases sampling in low-risk vaginal deliveries as a predictor of adverse neonatal outcome. Archives of Gynecology & Obstetrics, in press. https://doi.org/10.1007/s00404-023-06965-2
Johnson, G. J., Salmanian, B., Denning, S. G., Belfort, M. A., Sundgren, N. C., & Clark, S. L. (2021, Sep 1). Relationship between umbilical cord gas values and neonatal outcomes: Implications for electronic fetal heart rate monitoring. Obstetrics & Gynecology, 138(3), 366-373. https://doi.org/10.1097/AOG.0000000000004515
Leinonen, E., Gissler, M., Haataja, L., Andersson, S., Rahkonen, P., Rahkonen, L., & Metsäranta, M. (2019, Apr 07). Umbilical artery pH and base excess at birth are poor predictors of neurodevelopmental morbidity in early childhood. Acta Paediatrica, 108(10), 1801-1810. https://doi.org/10.1111/apa.14812
Royal Australian and New Zealand College of Obstetricians and Gynaecologists. (2019). Intrapartum Fetal Surveillance Clinical Guideline. 4th Edn. https://ranzcog.edu.au/statements-guidelines https://ranzcog.edu.au/statements-guidelines
White, C. R., Doherty, D. A., Henderson, J. J., Kohan, R., Newnham, J. P., & Pennell, C. E. (2010). Benefits of introducing universal umbilical cord blood gas and lactate analysis into an obstetric unit. Australian & New Zealand Journal of Obstetrics & Gynaecology, 50(4), 318-328. https://doi.org/10.1111/j.1479-828X.2010.01192.x
Categories: New research, Obstetrics, Perinatal brain injury
Tags: cord clamping, Evidence, guidelines, lactate, pH
Birth Small Talk 