Carbohydrate solution intake during labour just before the start of the second stage: a double-blind study on metabolic effects and clinical outcome


Dr H. C. J. Scheepers, Department of Obstetrics, Leiden University Medical Centre (LUMC), PO Box 9600, 2300 RC Leiden, The Netherlands.


Objective  To study the effects of oral carbohydrate ingestion on clinical outcome and on maternal and fetal metabolism.

Design  Prospective, double-blind, randomised study.

Setting  Leyenburg Hospital, The Hague, The Netherlands.

Population  Two hundred and two nulliparous women.

Methods  In labour, at 8 to 10 cm of cervical dilatation, the women were asked to drink a solution containing either 25 g carbohydrates or placebo. In a subgroup of 28 women, metabolic parameters were measured.

Main outcome measures  Number of instrumental deliveries, fetal and maternal glucose, free fatty acids, lactate, pH, Pco2, base excess/deficit and β-hydroxybutyrate.

Results  Drinking a carbohydrate-enriched solution just before starting the second stage of labour did not reduce instrumental delivery rate (RR 1.1, 95% CI 0.9–1.3). Caesarean section rate was lower in the carbohydrate group, but the difference did not reach statistical significance (1%vs 7%, RR 0.2, 95% CI 0.02–1.2). In the carbohydrate group, maternal free fatty acids decreased and the lactate increased. In the umbilical cord there was a positive venous–arterial lactate difference in the carbohydrate group and a negative one in the placebo group, but the differences in pH and base deficit were comparable.

Conclusion  Intake of carbohydrates just before the second stage does not reduce instrumental delivery rate. The venous–arterial difference in the umbilical cord suggested lactate transport to the fetal circulation but did not result in fetal acidaemia.


For decades, women have been advised against eating and drinking during labour, due to fear of gastric aspiration. In recent years, however, there has been a tendency towards a more liberal policy.1,2 In The Netherlands, women are allowed to eat and drink during normal labour. Despite this advice, the incidence of mortality due to gastric aspiration is comparable to countries with a restrictive policy.3,4

In sports medicine, the intake of carbohydrates during exercise has been proven not only to enhance performance, but also to delay fatigue.5–7 The effects of caloric intake on the process of parturition are unclear.

Our first study on actual intake during labour in women with free choice showed that women without caloric intake during labour had a higher incidence of instrumental deliveries due to an obstructed labour in the second stage.8 The study design made it impossible to draw conclusion regarding cause and effect. A following double-blind study on carbohydrate intake during early stages of labour showed a higher incidence of caesarean sections in the carbohydrate group.9 Our suggestion was that, like in athletes, not all carbohydrate intake regimes are beneficial.

Caloric requirements are highest in the second stage of labour and instrumental deliveries due to an obstructed labour in the second stage are common in nulliparous women. It was our hypothesis that intake of easily absorbed and digestible carbohydrates, just before starting the second stage, might reduce the incidence of instrumental deliveries. A second objective was to measure the metabolic consequences of carbohydrate intake. Hardly any data are available on the effects of eating and drinking on maternal and fetal metabolism during labour. In the past, intravenous glucose was administered in high concentrations, sometimes resulting in a fetal metabolic acidosis due to an increased lactate formation.10–13 During normal labour, most lactate formation occurs during the second stage and the concentration depends on the duration of this stage.14,15 Therefore, our hypothesis was that glucose ingestion would most likely cause acidosis due to an increased lactate level in nulliparous women, who drank carbohydrates just before starting the second stage.


In the period October 1999 to February 2001, this randomised, double-blind, placebo-controlled study was conducted at the Leyenburg Hospital, The Hague, The Netherlands. We included nulliparous women with a single fetus in cephalic presentation. Women with diabetes, a preterm delivery and women who where considered to have a direct risk for a caesarean section were excluded. Women with non-progressing labour were not excluded.

The women were asked to participate either before or during labour. Those who agreed to participate were randomised when the cervix was 8–10 cm dilated. Before randomisation, women were allowed to eat or drink at will. Allocation was undertaken by selecting consecutively numbered envelopes with the numbers of corresponding bottles containing 200 mL of either a carbohydrate solution (carbohydrates 12.6%: polysach 9.8%/Na 50 mg, Osm: 280 mOsm/L, Numico, Wageningen) or placebo (artificial aroma and sweeteners: aspartame). Randomisation occurred by a computer generated list and data collection was blinded. The bottles had an identical appearance and contained no information other than the randomisation number. The solutions had an identical appearance and taste. Intake of other solutions was not allowed. When intravenous infusion was required (e.g. augmentation), 0.9% saline was used, but in none of the women did this exceed 500 cc.

The following parameters were recorded: the time until reaching full dilatation, the duration of the second stage, the mode of delivery, the need for additional augmentation, Apgar scores and arterial umbilical cord pH. Furthermore, women were asked to record what caloric intake they had had before starting the study.

In our previous study on actual intake during labour, the incidence of an instrumental delivery due to obstructed second stage was 13% in nulliparous women with caloric intake versus 26% in those without.8 In this present study, women were included just before the second stage. This would mean that the major part of the assisted deliveries due to a non-progressing first stage would already have been performed earlier. Therefore, a higher incidence of vaginal birth and a higher incidence of vaginal instrumental deliveries were expected than in an unselected group. A need for 105 patients in each group was calculated using a prior estimation of 20%versus 40% incidence of obstructed second stage with an α= 0.05 and β= 0.10. Statistical analysis included descriptive statistics, χ2 tests and non-parametric testing (Mann–Whitney U test) were appropriate.

We studied the metabolic effects in a subgroup of 30 women. A venous blood sample was taken at randomisation and after delivery. The umbilical cord was clamped within 15 seconds and both arterial and venous umbilical cord blood samples were obtained. Within 5 minutes after delivery, a second venous maternal blood sample was taken. In all samples, the pH, Pco2, Po2, HCO3 and base excess were measured within 10 minutes at the delivery unit itself. In order to assess glucose, free fatty acids, lactate and β-hydroxybutyrate at a later time, the samples were kept on ice, centrifuged immediately in a pre-cooled centrifuge (4°C, 6 minutes 3390 n/minute) and stored at −70°C until analysis. Venous plasma glucose measurements were performed on a Synchron LX 20 from Beckman Coulter, using a hexokinase method. Lactate and β-hydroxybutyrate were assessed by an enzymatic/spectrophotometric method as described by Bergmeyer.16 Free fatty acids were assessed as described by Shimizu et al.17

Because no data on possible effects of oral carbohydrate intake on maternal and fetal metabolites were available at the start of the study, we choose a sample size of two groups of 15 women. We calculated the difference between the venous and arterial umbilical cord parameters. In the umbilical cord, the difference between the venous and arterial concentrations (ΔV − A) shows in which direction metabolites are transported. Positive values suggest fetal uptake and negative ones show transport to the maternal circulation. Because the maternal metabolic parameters were not completely comparable at randomisation, we calculated the differences in the changes in maternal blood samples before and after randomisation. We used non-parametric testing (Mann–Whitney). Both studies were approved by the Medical Ethical Committee and informed consent was obtained from all patients.


The clinical characteristics at the time of randomisation are outlined in Table 1. In both groups, the intake before randomisation had been approximately 10 kJ/hour and in both groups, almost half of the women had no intake after the start of contraction but before randomisation. The median total intake of the study solution was 200 mL (range 15–200) in the placebo group and 200 mL (range 10–200) in the carbohydrate group (P= 0.42).

Table 1.  Clinical data at randomisation mean [SD] or median (range) where appropriate.
 Carbohydrates (n= 100)Placebo (n= 102)
Maternal age (years)26.9 [4.1]26.4 [5]
Duration of pregnancy (days)279 [12]281 [11]
Body Mass Index
Before pregnancy23.7 [3.8]23.7 [3.6]
At the end of pregnancy29.6 [4.1]30.0 [4.4]
No. of women with
High risk pregnancies7371
Spontaneous start of contractions7476
Oxytocin augmentation1924
Opiate treatment1813
Epidural anaesthesia1613
Duration of regular contractions (hour)8 (1–30)8 (1–45)
Cervical dilatation (cm)8.9 [0.9]9.0 [0.8]
Caloric intake (kJ/hour)10 (0–413)10 (0–817)
No. of women without caloric intake4546
Fetal presentation
Occiput anterior9289

Table 2 shows the clinical parameters. The median duration from randomisation until the birth of the child was 81 minutes (range 9–540) in the placebo group versus 100 minutes (range 13–379) in the carbohydrate group (P= 0.25). No differences were seen in the total incidence of instrumental deliveries, but the number of caesarean sections however was higher in the placebo group The reasons for the caesarean sections were in the placebo group: obstructed labour in the first stage (n= 4), obstructed labour in the second stage (n= 1), fetal distress (n= 2). The only caesarean section that was performed in the carbohydrate group was due to obstructed labour in the second stage. No differences were observed in neonatal outcome.

Table 2.  Characteristics of labour after randomisation. Values are presented as n or median (range).
 Carbohydrates (n= 100)Placebo (n= 102)P or RR (95% CI)
Duration until start of the second stage (minutes)45 (0–325)30 (0–524)0.64
Duration second stage (minutes)47 (6–195)40 (5–118)0.09
Additional augmentation14150.89
Spontaneous delivery68651.07 (0.88–1.30)
Vaginal instrumental delivery31301.05 (0.69–1.60)
Caesarean section170.15 (0.02–1.16)
Reasons for instrumental deliveries
Non-progressing first stage040.05
Non-progressing second stage22200.68
Fetal distress990.97
Combination of fetal distress and non-progressing second stage140.18
Apgar 1 minute9 (2–10)9 (4–10)0.22
Apgar 5 minutes10 (7–10)10 (6–10)0.32
Arterial umbilical cord pH7.22 (6.97–7.35)7.21 (6.94–7.38)0.80

In the subgroup of 30 women in whom metabolic parameters were measured, two women were excluded since no blood samples were taken after the baby was born. Both women had a caesarean section and had been randomised in the placebo group. At randomisation, maternal age (28 vs 26 years), body mass index (both 24), duration of active labour (6.75 vs 8 hours), caloric intake before randomisation (55 vs 71 kJ/hr), the number of women using pain medication (5 vs 6) or augmentation (4 vs 2) were comparable. In the carbohydrate group, the median cervical dilation was lower (8 vs 9 in the placebo group, P= 0.05) and more women delivered spontaneously (67%vs 54%).

At randomisation, maternal lactate: 2.85 (1.46; 6.05) versus 1.87 (1.25; 4.69) and maternal base excess: −6.4 (−8.7; −2.3) versus−4.7 (−6.6; −0.6) were higher in the carbohydrate group All other values were comparable: pH: 7.45 (7.36; 7.61) versus 7.43 (7.41; 7.50); Pco2: 24.2 (14.9; 33.8) versus 27.3 (20.8; 33.7); glucose: 6.1 (5.2; 10.6) versus 6.1 (4.6; 11.9); free fatty acids: 0.9 (0.36; 1.54) versus 0.88 (0.41; 1.67); and β-hydroxybutyrate: 0.6 (0.11; 1.57) versus 0.68 (0.09; 2.52).

The maternal serum metabolic parameters in both groups after the birth of the child are outlined in Table 3. Because the groups were not completely comparable at randomisation, the changes during delivery are important.

Table 3.  Maternal metabolic parameters at the birth of the child. Values are presented as median (range).
 Carbohydrates (n= 15)Placebo (n= 13)P
pH7.31 (7.25; 7.42)7.33 (7.23; 7.39)0.59
Pco2 (mmHg)28.4 (21.4; 35.4)28.8 (17.8; 38.4)0.74
Base excess−9.75 (−14.4; −3.9)−9.15 (−17.4; −6.2)0.90
Glucose (mmol/L)7.2 (5.1; 10.8)6.4 (5.9; 14.3)0.56
Free fatty acids (mmol/L)0.65 (0.14; 1.24)0.87 (0.58; 1.91)0.02
Lactate (mmol/L)7.47 (2.94; 10.29)5.22 (2.54; 9.45)0.009
β-hydroxybutyrate (mmol/L)0.43 (0.06; 3.02)0.6 (0.2; 2.90)0.20

Table 4 shows the changes in both groups. The changes in pH, base excess, glucose or β-hydroxybutyrate were comparable. Free fatty acid concentrations hardly changed in the placebo group, but decreased in the carbohydrate group. Lactate levels increased in both groups, but tended to increase to a higher degree in the carbohydrate group.

Table 4.  Changes in maternal metabolic parameters in both the carbohydrate and placebo groups. P value shows difference in changes. Values are presented as median (range).
 Carbohydrates (n= 15)Placebo (n= 13)P
ΔpH−0.12 (−0.28; −0.02)−0.11 (−0.20; −0.03)0.62
ΔBase excess−4 (−7.8; 4.3)−5.9 (−11.9; −1.5)0.15
ΔGlucose1.0 (−4.3; 4.5)0.9 (−0.6; 2.4)1.00
ΔFree fatty acids−0.32 (−1.19; 0.23)0.05 (−0.54; 0.77)0.02
ΔLactate4.41 (−0.55; 8.21)2.37 (0.65; 7.95)0.07
Δβ-hydroxybutyrate−0.03 (−1.09; 2.20)0.11 (−0.18; 0.76)0.21

Table 5 shows the metabolic parameters in venous and umbilical cord blood in both the placebo and carbohydrate groups. No differences were observed in the arterial and venous concentrations of any of the parameters. The difference between the venous and arterial concentrations is shown in Table 6. Positive values are found in both groups for glucose, free fatty acids and β-hydroxybutyrate. In the placebo group, the difference between venous and arterial lactate concentrations was negative (−0.39), suggesting lactate transfer from the fetus to the maternal circulation. In the carbohydrate group, this difference was positive (0.56), suggesting that lactate is transferred from the mother to the fetus. This difference was not significant (P= 0.12).

Table 5.  Venous and arterial umbilical cord metabolic parameters. Values are presented as median (range).
 Carbohydrates (n= 15)Placebo (n= 13)P
Venous umbilical cord
pH7.31 (7.19; 7.36)7.29 (7.15; 7.36)0.28
Pco2 (mmHg)37.3 (25; 56)39.1 (34; 57)0.19
Base excess−6.8 (−13.4; −3)−7.1 (−9.7; −2.9)0.83
Glucose (mmol/L)6.3 (3.7; 7.4)5.9 (5.4; 12.8)0.56
Free fatty acids (mmol/L)0.12 (0.03; 0.23)0.14 (0.08; 0.37)0.09
Lactate (mmol/L)5.6 (2.94; 10.4)5.1 (2.7; 8.9)0.45
β-Hydroxybutyrate (mmol/L)0.18 (0.01; 2.05)0.4 (0; 3.0)0.26
Arterial umbilical cord
pH7.22 (7.06; 7.34)7.21 (7.03; 7.30)0.39
Pco2 (mmHg)49.9 (30.7; 74.8)54.3 (41.2; 76.5)0.24
Base excess−6.3 (−15.7; −2.4)−6.3 (−11.4; −1.9)0.60
Glucose (mmol/L)5.9 (3.4; 8.5)5.3 (3.9; 6.4)0.17
Free fatty acids (mmol/L)0.07 (0.01; 0.22)0.09 (0.02; 1.06)0.85
Lactate (mmol/L)5.43 (2.78; 11.48)5.77 (2.23; 7.24)0.86
β-hydroxybutyrate (mmol/L)0.12 (0.02; 1.80)0.25 (0.09; 2.76)0.15
Table 6.  Differences between venous and arterial umbilical cord values (ΔV − A). Values are presented as median (range).
 Carbohydrates (n= 15)Placebo (n= 13)P
Glucose (mmol/L)
Free fatty acids (mmol/L)
Lactate (mmol/L)0.56−0.390.12
β-hydroxybutyrate (mmol/L)


In this study, we observed no differences in the total number of instrumental deliveries. A lower number of caesarean sections were observed in the women who drank carbohydrate-enriched liquids compared with the women who drank placebo, but this was not significant.

In our previous randomised placebo-controlled study, women were allowed to drink carbohydrate-enriched fluids at will, during the entire first stage of labour. In the carbohydrate group, a significant increase in caesarean sections was observed.9 It is difficult to find a logical explanation for these differences in results.

In athletes, not all carbohydrate regimens have proven to be beneficial. In athletes with a low activity level, a fasting period of several hours before intake and a relatively low carbohydrate intake, transient hypoglycaemia has been described.18 The women in our first study might have experienced a similar situation. They had low actual intake (±350 mL/8 hours) and, in many cases, most intake occurred just after randomisation and at the beginning of active labour. It has been hypothesised that hypoglycaemia could be caused by a an insulin response following carbohydrate intake, which in more strenuous exercise or high anxiety levels is counter-regulated by the sympathetic nervous system.18 When the level of sympathetic activation is lower, this counter-regulation is less active. Just before and during the second stage, a higher activity level, and probably also a higher anxiety level, is present. It has been described that the sympathetic activation rises gradually during labour.19

No data were available on the rate in which carbohydrate-enriched solutions are emptied from the stomach in women in labour and the availability of carbohydrates. The drinks we used in this study were previously studied in non-labouring women.20 They were completely emptied from the stomach within 60–90 minutes and the majority within the first half hour. The maximal blood glucose concentrations were observed at 40 minutes. In the present study, the intake of the solution was approximately 45 minutes before starting the second stage. The median time until delivery was 90 minutes. It was therefore likely that a substantial amount of glucose was absorbed. This would mean that, compared with the ‘at will’ intake in the beginning of the first stage, a much higher glucose load was present.

Glucose is the most important energy source during labour for both the mother and the fetus.21–23 During labour, in most cases, maternal plasma glucose levels rise,19,24,25 which can be explained by the increased gluconeogenesis due to cortisol and adrenaline, produced due to a stress reaction.26 The simultaneous rise in free fatty acids27,28 and ketobodies29,30 suggests, however, that a relative lack of glucose is present. Indeed, in some cases, during labour, maternal glucose concentrations decline.19

In the present study, maternal oral glucose ingestion at the end of labour did not lead to significant changes in maternal glucose levels when compared with the placebo group. However, the lower concentrations of free fatty acids suggest improved glucose availability. Lactate levels tended to increase to a higher extent in women who drank carbohydrates but did not result in a maternal metabolic acidosis. Schneider et al.31 have described that lactate is transferred by facilitated diffusion and therefore depends on the difference in concentrations. In almost all cases, fetal lactate concentrations are found to be higher than the maternal levels, suggesting that in most deliveries lactate is produced by the fetus and transferred to the maternal circulation.31,32 Piquard et al.33 described that in the presence of high maternal lactate levels, lactate may be transferred from the fetus to the mother. The present study suggests indeed that in the carbohydrate group lactate is transferred to the fetal circulation. No differences, however, were observed in fetal pH or base excess.

Theoretically, like in intravenous administration, the ingestion of high concentrations of oral carbohydrates might affect maternal and fetal lactate concentrations and therefore the fetal and maternal acid–base balance. Intravenous infusion of 100 g of glucose in one hour has been proven to cause fetal acidaemia.34 Dosages of 9 and 30 g of glucose per hour did not result in lactoacidosis or neonatal hypoglycaemia.32,35


At present, it is difficult to establish whether carbohydrate supplementation during labour is beneficial or not. Different intake strategies might have different metabolic and clinical consequences. In many countries, women in labour are nowadays allowed to have clear liquid intake, but studies regarding effects of this intake are lacking.

Accepted 21 April 2004