Hidden acidosis: an explanation of acid–base and lactate changes occurring in umbilical cord blood after delayed sampling

Authors

  • P Mokarami,

    Corresponding author
    • Institution of Clinical Sciences, Department of Obstetrics and Gynaecology, Skåne University Hospital, Lund University, Malmö, Sweden
    Search for more papers by this author
  • N Wiberg,

    1. Institution of Clinical Sciences, Department of Obstetrics and Gynaecology, Skåne University Hospital, Lund University, Lund, Sweden
    Search for more papers by this author
  • P Olofsson

    1. Institution of Clinical Sciences, Department of Obstetrics and Gynaecology, Skåne University Hospital, Lund University, Malmö, Sweden
    Search for more papers by this author

Correspondence: Dr P Mokarami, Department of Neurology, Skåne University Hospital, S–20502 Malmö, Sweden. Email parisa.mokarami@med.lu.se

Abstract

Objective

To explore the ‘hidden acidosis’ phenomenon, in which there is a washout of acid metabolites from peripheral tissues in both vaginal and abdominal deliveries, by investigating temporal umbilical cord blood acid–base and lactate changes after delayed blood sampling.

Design

Prospective comparative study.

Setting

University hospital.

Sample

Umbilical cord blood from 124 newborns.

Methods

Arterial and venous cord blood was sampled immediately after birth (T0), and at 45 seconds (T45), from unclamped cords with intact pulsations taken from 66 neonates born vaginally and 58 neonates born via planned caesarean section at 36–42 weeks of gestation. Non-parametric tests were used for statistical comparisons, with < 0.05 considered significant.

Main outcome measures

Temporal changes (T0T45) in umbilical cord blood pH, the partial pressure of CO2 (math formula) and O2 (math formula), and in the concentrations of lactate, haematocrit (Hct), and haemoglobin (Hb).

Results

In both groups all arterial parameters, except for math formula in the group delivered by caesarean section, changed significantly (pH decreased and the other variables increased). There were corresponding changes in venous acid–base parameters. When temporal arterial changes were compared between the two groups, the decrease in pH and increase in math formula were more pronounced in the group delivered vaginally. Neonates born vaginally had significantly lower pH and higher lactate, Hct, and Hb concentrations at T0 and T45 in both the artery and the vein. At T45, arterial math formula and math formula levels in the group delivered vaginally were also significantly higher.

Conclusions

Delayed umbilical cord sampling affected the acid–base balance and haematological parameters after both vaginal and caesarean deliveries, although the effect was more marked in the group delivered vaginally. The hidden acidosis phenomenon explains this change towards acidaemia and lactaemia. Arterial haemoconcentration was not the explanation of the acid–base drift.

Introduction

Delayed umbilical cord clamping at vaginal delivery results in a decrease in pH and base excess (BE), and an increase in the partial pressure of O2 (math formula), the partial pressure of CO2 (math formula), and lactate concentration in the umbilical artery.[1-3] These changes towards acidaemia and lactaemia can be explained by the ‘hidden acidosis’ phenomenon. During uterine contractions, the fetal circulation is centralised at the expense of perfusion of low-priority organs and peripheral tissues,[4] with a build-up of acid metabolites peripherally. When the newborn starts to breathe sufficiently the peripheral perfusion is restored and the ‘trapped’ metabolites surge into the central circulation and, after some seconds, can be detected in umbilical cord blood.[3] The phenomenon has also been demonstrated in animal studies at the restoration of the peripheral circulation after provoked hypovolaemic shock.[5, 6] Soon after volume expansion has started, a rapid drop in pH and increase in lactate concentration are seen. In animal limb tourniquet ischaemia–reperfusion experiments, a similar phenomenon is seen during reperfusion.[7, 8]

Our hypothesis was that hidden acidosis occurs in the newborn (Figure 1). As newborns after planned caesarean delivery (caesarean section) seldom show acrocyanosis, we hypothesized that hidden acidosis would be most pronounced after vaginal delivery. The opening of peripheral vascular beds might result in changes in haemoconcentration in the cord blood, and therefore we investigated temporal changes not only in blood gases and lactate concentration, but also in haematocrit (Hct) and total haemoglobin (ctHb) concentration.

Figure 1.

Schematic illustration of the hidden acidosis phenomenon. The grey box represents the first few minutes after birth, when a steep decrease in pH and an increase in lactate concentration are first seen, according to the hypothesis.

Methods

Arterial and venous umbilical cord blood were sampled from 124 newborn singletons immediately after birth (T0), and again at 45 seconds (T45), from unclamped umbilical cords with intact pulsations. The women's length of gestation was determined at an early second trimester ultrasound, and all were found to be at 36–42 weeks of gestation. Of the 124 neonates, 66 were born vaginally in cephalic presentation and 58 were delivered by planned caesarean section. The newborns included in the study were expected to have no need of immediate rescue procedures that would interfere with the delayed cord clamping. The women who delivered vaginally were included in a previously published study.[3]

Women in the group delivering vaginally were recruited to the study at admission to the labour and delivery ward, and women in the group delivering by caesarean section were asked to participate a few hours before the operation. All caesarean sections were planned and the indications were breech presentation or maternal request. Women undergoing spinal anaesthesia were placed in supine position, tilted 15º to the left, and received prehydration. Bupivacaine and fentanyl were used for spinal anaesthesia. Simultaneously, an intravenous infusion of ephedrine (50 mg in 500 ml of sodium chloride solution) was started and adjusted with the aim to maintain a mean arterial pressure within 25% of its initial value. Women undergoing general anaesthesia also received prehydration. Drugs administered at general anaesthesia were thiopental, suxamethonium, and sevoflorane. After cord clamping, all women received oxytocin.

During cord blood sampling, babies delivered vaginally were placed on the abdomen of the mother, whereas babies born by caesarean section were placed between the mother's legs and kept warm under a towel. The procedure was meticulously prepared, and the samples were taken and analysed by one of the authors (N.W.), who was not involved in the obstetric care of the women. Blood was drawn first from the cord artery and then, within a few seconds, and at the same location on the cord, from the vein. The next pair of samples were taken 45 seconds later, and the needle punctures were made a few millimetres closer to the placenta. A 0.6- or 0.9-mm needle was used, and the samples were collected in 2–ml pre-heparinised plastic syringes. A minimum of 0.5 ml of blood from each vessel was used for analysis in the blood gas analyser (ABL735; Radiometer A/S, Copenhagen, Denmark). All samples were analysed within 15 min, in chronological order. The radiometer analyser works by measuring pH and math formula by potentiometry, PO2 and lactate by amperometry, and ctHb by spectrophotometry. ctHb includes deoxy-, oxy-, carboxy-, and methemoglobin. Hct is available as a derived parameter, calculated according to the formula: Hct = 0.0485 × ctHb + 8.3 × 10−3. The analyser was operated in an accredited laboratory (Laboratory Medicine Skåne, Clinical Chemistry, Lund and Malmö).

All women in labour were monitored with cardiotocography during the second stage of labour. Small for gestational age (SGA) was defined as a birthweight below – 2 SD from the gestational age-adjusted mean value, appropriate for gestational age (AGA) was defined as a birthweight within the mean ± 2 SD range, and large for gestational age (LGA) was defined as a birthweight above the mean + 2 SD.[9]

Statistical analyses

The Mann–Whitney U test was used for comparison of continuous parameters between groups, and the Wilcoxon signed-ranks matched-pairs test was used for longitudinal comparisons. Values are reported as median and range or mean with 95% confidence interval (95% CI), as appropriate. A two-tailed P < 0.05 was considered to be statistically significant. Statistical analyses were performed with the aid of StatView® (SAS Institute, Cary, NC, USA). As umbilical cord blood gas and lactate values are dependent on gestational age,[10-12] comparisons between groups delivered vaginally and by caesarean section were also performed using cord arterial pH adjusted to a gestational age of 280 days, according to the regression coefficient −0.00096 per day of gestational age.[10]

Results

The characteristics of the study population are shown in Table 1. Gestational age at delivery was significantly lower, and Apgar score (AS) at 1 minute was significantly higher, in the group delivered by caesarean section. One newborn had an AS of 4 at 1 minute, but otherwise all scores at 1 minute were ≥8 and at 5 and 10 minutes were ≥9.

Table 1. Characteristics of the study population (= 124)
 Vaginal delivery (= 66)Caesarean delivery (= 58)
  1. a

    The difference in gestational age and Apgar score at 1 minute was statistically significant (Mann–Whitney U test; ≤ 0.03) between the two groups.

  2. Values are median (range) or number of cases (%).

Maternal characteristics
Duration of second stage of labour (min)41 (5–234)
Duration of pushing (min)24 (4–90)
Induction of labour5 (7.6%)
Instrumental birth9 (13.6%)
Drugs administered
Pethidin6 (9.1%)
Oxytocin31 (47.0%)
Nitrous oxide50 (75.8%)
Anaesthesia
Epidural15 (22.7%)
Spinal52 (90.0%)
General6 (10.0%)
Newborn characteristics
Gestational age (weeks)a40+0 (36+0 – 42+0)38+4 (36+4 – 40+3)
Birthweight (g)3595 (2560–4405)3535 (2516–5320)
SGA3 (4.5%)0
AGA62 (93.9%)47 (81.0%)
LGA1 (1.5%)11 (19.0%)
Apgar score
1 minutea9 (4–10)9 (8–10)
5 minute10 (8–10)10 (7–10)
10 minute10 (9–10)10 (9–10)
Cardiotocography
Intermediate13 (19.7%)
Pathological3 (4.5%)

Serial blood samples were taken in all 124 cases, but four analyses at T0 (one vaginal delivery and three caesarean sections) and ten analyses at T45 (six vaginal deliveries and four caesarean sections) failed because of instrument failure or blood clotting. For each parameter, only cases with valid measurements obtained at both T0 and T45 were included in the statistical analyses. Data for arterial and venous acid–base and haematological measurements are shown in Tables 2 and 3.

Table 2. Arterial blood gas, lactate, haematocrit (Hct), and total haemoglobin (ctHb) concentration median (range) values obtained immediately after birth (time T0), and again 45 seconds later (T45), in unclamped umbilical cords with intact pulsations after vaginal delivery and caesarean delivery
 VaginalCaesarean T 0 T 45 Vaginal versus caesarean
VaginalCaesareanVaginalCaesareanSignificance of difference (P)
n n Median (range)Median (range)Median (range)Median (range) T 0 T 45
  1. The Mann–Whitney U test was used for group comparisons.

pH58397.235 (7.008–7.379)7.305 (7.162–7.397)7.207 (7.005–7.384)7.296 (7.116–7.424)<0.0001<0.0001
math formula (kPa)58397.55 (5.24–11.6)7.30 (5.86–9.56)7.87 (5.94–11.8)7.57 (5.56–10.4)0.30.03
math formula (kPa)57392.31 (0.62–7.93)1.99 (1.18–3.72)2.66 (1.09–4.94)2.28 (1.18–3.25)0.10.02
Lactate (mmol/l)56374.8 (2.0–13.3)1.8 (1.1–4.8)5.5 (2.3–13.3)2.2 (1.5–6.2)<0.0001<0.0001
Hct57380.507 (0.051–0.625)0.452 (0.409–0.585)0.514 (0.423–0.635)0.460 (0.372–0.583)<0.0001<0.0001
ctHb (g/l)5738167 (134–205)148 (133–191)168 (138–208)151 (121–191)<0.0001<0.0001
Table 3. Venous blood gas, lactate, haematocrit (Hct), and total haemoglobin (ctHb) concentration median (range) values obtained immediately after birth (time T0), and again 45 seconds later (T45), in unclamped umbilical cords with intact pulsations after vaginal delivery and caesarean delivery
 VaginalCaesarean T 0 T 45 Vaginal versus caesarean
VaginalCaesareanVaginalCaesareanSignificance of difference (P)
n n Median (range)Median (range)Median (range)Median (range) T 0 T 45
  1. The Mann–Whitney U test was used for group comparisons.

pH64417.331 (7.068–7.471)7.371 (7.320–7.479)7.329 (7.470–7.474)7.367 (7.318–7.469)<0.0001<0.0001
math formula (kPa)64415.49 (3.91–9.70)5.78 (4.37–7.46)5.42 (4.05–9.54)5.77 (4.69–7.54)0.20.1
math formula (kPa)63413.57 (1.46–15.70)3.46 (1.87–7.45)3.68 (1.52–7.38)3.46 (1.40–6.43)0.60.9
Lactate (mmol/l)60404.6 (1.9–10.9)1.5 (1.1–2.7)4.7 (2.1–10.8)1.6 (1.2–3.0)<0.0001<0.0001
Hct63380.515 (0.401–0.648)0.455 (0.410–0.585)0.513 (0.058–0.633)0.456 (0.389–0.590)<0.0001<0.0001
ctHb (g/l)6439168 (131–212)148 (133–191)168 (126–208)149 (127–193)<0.0001<0.0001

Longitudinal changes between T0 and T45

Longitudinal changes in arterial blood gases, and in lactate, Hct, and ctHb concentrations are illustrated in Figure 2. With the exception of math formula in the group delivered by caesarean section (P = 0.4), all blood gas and lactate parameters changed significantly. Acid–base changes in venous blood were in the same directions as in arterial blood, although in the group delivered vaginally only the increase in lactate was significant (= 0.001), and in the group delivered by caesarean section only the decrease in pH (P = 0.03) and increase in lactate (P < 0.0001) were significant (not shown in Figure 2). Hct and ctHb increased significantly in the artery in both groups, whereas venous values decreased significantly in the group delivered vaginally (P ≤ 0.04), and remained unchanged in the group delivered by caesarean section (P ≥ 0.2).

Figure 2.

Measurements of arterial umbilical cord blood gases, and concentrations of lactate, haematocrit (Hct), and total haemoglobin (ctHb) obtained immediately after birth (T0), and then again 45 seconds later (T45), in unclamped umbilical cords with intact pulsations after vaginal and caesarean deliveries. The figure shows mean values and 95% confidence intervals. The Wilcoxon signed-ranks test was used to compare values at T0 and T45: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; NS, not significant.

Vaginal versus caesarean delivery

When longitudinal arterial pH, lactate, and math formula changes were compared between groups, the decrease in pH and increase in math formula were found to be significantly greater in the group delivered vaginally (P ≤ 0.04), but there was no statistically significant difference between the groups regarding the increase in lactate concentration from T0 to T45 (P = 0.9). Adjusting pH for the difference in gestational age between the groups did not change the results.

Neonates born by vaginal delivery had significantly lower pH values and higher lactate, Hct, and ctHb concentrations at T0 and T45, in both the artery and the vein, compared with neonates delivered by caesarean section (Tables 1 and 2). At T45, math formula and math formula in the artery in the group delivered vaginally were also significantly higher.

Spinal versus general anaesthesia

Neonates in the group delivered by caesarean section with spinal anaesthesia (= 52) had lower pH values, and higher math formula and lactate concentration at T0, compared with neonates in the general anaesthesia group (= 6), but only the difference in lactate concentration was statistically significant (P = 0.03).

Discussion

This study showed significant changes in acid–base and haematological parameters in umbilical cord blood when sampling was delayed by 45 seconds, with these changes being more marked for pH and math formula in the group delivered vaginally. The similar increases in lactate concentration in the two groups indicate that considerable hidden acidosis was also present in the group delivered by caesarean section.

The lack of change in venous math formula indicates that placental perfusion and gas exchange were maintained during the first 45 seconds, after both vaginal and abdominal deliveries. Thus, the temporal increase in arterial math formula must be a result of CO2 inflow from the newborn, and not from the placenta, or of an accumulation of CO2 in the blood circuit. Moreover, the significant increase in math formula indicates the rapid establishment of functional pulmonary ventilation, which would result in the escape of CO2 and in a lowering of math formula unless there was a considerable continuing fetal contribution. As it is unlikely that the CO2 contribution was a result of a sudden rise in neonatal metabolism, a washout of CO2 from peripheral tissues is the most plausible explanation for this finding.

After 45 seconds, arterial blood showed a small but significant haemoconcentration and venous blood showed a haemodilution in the group delivered vaginally. A relevant question is, then, whether these concentration changes could have influenced the temporal acid–base and lactate changes. According to Stewart's physicochemical concept, a change towards alkalosis should occur during haemoconcentration, as dehydration results in a higher [OH].[13] In the present study, the changes in haemoconcentration paralleled changes towards acidosis in the artery, indicating that the temporal acetous change was not a result of the haemoconcentration.

The study was performed in cases in which minimal neonatal assistance was expected to be required, and only two newborns in the group delivered vaginally and none in the group delivered by caesarean section had an umbilical artery pH <7.10 in the first samples. Both these newborns had a pathological cardiotocogram. One newborn was vigorous immediately, with 1-, 5-, and 10-minute AS scores of 8, 9, and 10, respectively, whereas the other was initially moderately depressed, and had corresponding AS scores of 4, 8, and 10. Interestingly, in the newborn with a 1-minute AS score of 8, the blood gas and lactate values deteriorated further by 45 seconds of age: pH changed from 7.06 to 7.02; math formula changed from 10.0 to 10.5 kPa; BE changed from –12.7 to –15.3 mmol/l; and lactate changed from 12.2 to 12.9 mmol/l. In the depressed newborn, the values remained mainly unchanged: pH was 7.01 at both time points, math formula changed from 11.2 to 11.8 kPa, BE changed from –14.9 to –14.4 mmol/l, and lactate concentration was 13.3 mmol/l at both time points. These observations further support the hypothesis that hidden acidosis is a physiological phenomenon, occurring in newborns with a rapidly established circulation.

It was not expected that the hidden acidosis phenomenon would occur so clearly in neonates born by caesarean section, as these neonates were not exposed to hypoxic stress by uterine labour contractions; however, it is well known that fetal/neonatal effects occur during regional anaesthesia for planned caesarean section. Despite precautions in terms of prehydration and vasopressor administration, spinal anaesthesia in particular is frequently associated with maternal hypotension and lower umbilical cord arterial pH.[14-18] Vasopressor substances can cross the placenta,[14, 19-22] and the maternal supine wedged position during caesarean section frequently results in fetal heart rate changes as a result of occult aortocaval compression.[23] Doppler ultrasound has shown uteroplacental circulation to be affected after spinal blockade.[16, 19, 24, 25] In concordance with these findings, the present study showed higher lactate values in the spinal anaesthesia group than in the general anaesthesia group. It seems that, even with the most modern techniques for spinal anaesthesia, this side effect is difficult to avoid.[26]

An interesting finding was that at T0, math formula was similar in the groups delivered vaginally and by caesarean sections, but at T45 it was significantly higher in the group delivered vaginally, as a result of a steeper increase. This demonstrates the protective role of vaginal delivery, with the more effective release of lung surfactant and alveolar expansion, absorption of pulmonary fluid, and rapid circulatory transition to extra-uterine life. At 45 seconds, alveolar clearance of fluid and alveolar expansion are the most important processes.[27]

Strengths and weaknesses

Repeated blood sampling performed by an experienced obstetrician and analyses within 15 minutes in chronological order minimised the sampling and measurement errors. The inclusion of only newborns presumed to be vigorous makes extrapolation to asphyxiated newborns problematic.

Interpretation

Even small blood gas changes can affect the interpretation of a newborn's status and lead to a false diagnosis of acidosis, as we have previously demonstrated.[3] Hypoxic neonates are expected to have a more pronounced circulatory centralisation and hidden acidosis, and, as they already have lower pH levels, an additional decrease is more likely to tip them below the lower limit of the reference interval. It would be difficult to create reliable normal reference intervals taking late cord blood sampling into account, because, as discussed above, vigorous newborns would show changes towards acidaemia, lactaemia, and hypercapnia, whereas depressed newborns would show small changes.

Conclusion

Delayed cord blood sampling with intact pulsations affected umbilical acid–base values and haematological parameters following both vaginal and caesarean deliveries. A change towards acidaemia and lactaemia can be explained by the hidden acidosis phenomenon. A small degree of haemoconcentration occurred in arterial blood, and haemodilution occurred in venous blood, but these changes could not explain the change in acid–base status.

Disclosure of interests

The authors state explicitly that there are no conflicts of interest in connection with this article.

Contribution to authorship

PM was involved in the conception and planning of the study, analysis of the data, and writing of the article; NW was involved in the conception, planning, and carrying out of the study, analysis of the data, and writing of the article. PO was involved in the conception and planning of the study, analysis of the data, and writing of the article.

Details of ethics approval

The study was approved on 24 February 2006 by the Central Ethical Review Board, Stockholm, Sweden (reference number Ö 50–200), and all the women gave their informed oral and written consent to participate in the study.

Funding

This study was supported by grants from Region Skåne and the Medical Faculty at Lund University (ALF). The funding sources had no role in the writing of the article or in the decision to submit it for publication.

Acknowledgement

None.

Ancillary