We conducted a randomised controlled trial (parallel group study with 1:1 randomisation) comparing delayed and early cord clamping. The study was conducted between April 2008 and September 2009 at the Hospital of Halland, Halmstad, Sweden.
Pregnant women were eligible if they met the following criteria: non-smoking; healthy (no haemolytic disease, no treatment with any of the following drugs (anticonvulsants, antidepressants, thyroid hormone, insulin, chemotherapy, or cortisone)); normal pregnancy (no pre-eclampsia, no diabetes, no prolonged rupture of membranes or signs of infection); singleton; term pregnancy (gestational age 37+0 to 41+6 weeks); expected vaginal delivery with cephalic presentation. The mother also had to understand Swedish well enough to participate in the study and live close enough to the hospital to be willing to return for follow-up after four months. Exclusion criteria were serious congenital malformations, syndromes, or other congenital diseases that could affect the outcome measures.
Pregnant women were first given information about the trial at the antenatal healthcare centre. Those fulfilling the inclusion criteria at the time of admission to the delivery ward were again informed about the study by the attending midwife. Written informed consent was obtained, from both parents when possible. The study was approved by the regional ethical review board at Lund University (41/2008).
When delivery was imminent (expected within 10 minutes), the midwife opened a sealed, numbered, opaque envelope containing the treatment allocation. The interventions consisted of delayed clamping of the umbilical cord (≥180 seconds after delivery) or early clamping of the umbilical cord (≤10 s). In both intervention arms, the midwife was instructed to hold the newborn infant at a level about 20 cm below the vulva for 30 seconds and then place the baby on the mother’s abdomen (20 cm was chosen as a practically obtainable level because of the facility to fold down the lower part of the delivery bed). Babies born by caesarean section were placed in the mother’s lap before clamping in accordance with clinical routines. Oxytocin (10 IU) was administered intravenously immediately after cord clamping. The time from complete delivery of the baby to the first clamp on the umbilical cord was measured with a stopwatch by the midwife’s assistant.
Venous and arterial cord blood samples for assessment of acid-base status were taken within 30 seconds from the unclamped cord in the delayed cord clamping group and within 10 minutes from the double clamped segment of the umbilical cord in the early clamping group. Study samples were taken from the clamped umbilical cord in both groups. The remaining fetal blood in the placenta was measured by placing the free end of the cut umbilical cord in a measuring glass and elevating the placenta until all blood had been drained. All other aspects of obstetric care were managed according to standard practice at the hospital. All staff in the delivery unit were trained in the study procedures before the trial started. Early cord clamping was the clinical standard procedure in the hospital before the study.
After the delivery, the babies were cared for according to clinical routines, and early breast feeding was encouraged. As part of the study, the infant was assessed at 1 and 6 hours by the midwife, who recorded if the baby had been breast fed and the presence of respiratory symptoms (that is, respiratory rate >60 breaths/minute, presence of nostril flaring, grunting, or intercostal retractions).
Infants stayed at the postnatal ward with their mothers for two or three days, except for well babies whose mothers preferred to leave the hospital earlier and infants who were admitted to the neonatal unit. All infants were examined by a physician during the first 72 hours, in accordance with clinical routines. All neonatal diagnoses were reported in the study protocol. At 48–72 hours after birth, study samples were taken by a midwife or a neonatal nurse in conjunction with routine venous blood sampling for metabolic screening. Results from any additional blood samples that were taken on clinical indications were also recorded in the study protocol. The results from study samples were reviewed once a week by a physician, and appropriate action was taken if necessary. Six infants were considered to have anaemia at 2 days of age and were referred for further diagnostics, but none needed treatment.
At 4 months of age, infants were scheduled for a follow-up visit including blood sampling and weight and length measurements. Venous blood sampling was performed after application of a local anaesthetic (EMLA, AstraZeneca). Before the visit, parents were asked to complete a three day food diary to assess whether the infant was exclusively or partially breast fed and whether formula or solids were given.
Additional data collection
The following information was collected from maternal healthcare records: reported illness, medication, parity, weight, height, body mass index, smoking habits, blood group Rhesus factor status, and haemoglobin concentration at the time of admission to antenatal care. The median duration of pregnancy at this visit was 76 days (interquartile range 68 to 84). If the objective of the randomised intervention (that is, ≥180 s between delivery and umbilical cord clamping for delayed clamping or ≤10 s for early clamping) was not achieved, the reason for this was noted.
Umbilical cord blood was analysed for complete blood count (haemoglobin, packed cell volume, mean cell volume, mean cell haemoglobin concentration, reticulocyte count, and reticulocyte haemoglobin); iron status (serum iron, transferrin, serum ferritin, transferrin saturation, and soluble transferrin receptors); and C reactive protein. At 2 days of age, venous blood was analysed for complete blood count, iron status, C reactive protein, and bilirubin. At 4 months, venous blood samples were analysed for complete blood count, iron status, and C reactive protein.
Blood was collected in EDTA tubes (BD Vacutainer, Plymouth, UK) for complete blood count, and in serum separator tubes (BD Vacutainer) for iron status, bilirubin, and C reactive protein.
Blood samples were stored for a maximum of one hour in room temperature and then transported for analysis to the hospital clinical chemistry laboratory, where analyses were performed within one hour. Complete blood counts were analysed with an automated haematology analyser (Sysmex XE 2100, Sysmex, Kobe, Japan). Iron status indicators, bilirubin, and C reactive protein were analysed with Cobas 6000 (Roche Diagnostics, Basel, Switzerland).
We used the following definitions:
- At 2 days
- Anaemia—haemoglobin <145 g/L29
- Polycythaemia—packed cell volume >0.6530
- Hyperbilirubinaemia—bilirubin >257 μmol/L (corresponding to 15 mg/dL)
- At 4 months
- Anaemia—haemoglobin <105 g/L31
- Iron deficiency—≥2 indicators of iron status outside reference range (ferritin <20 μg/L,31 mean cell volume <73 fL,31 transferrin saturation <10%,32 soluble transferrin receptor >7 mg/L). (The soluble transferrin receptor cut-off value for the Roche immunoturbidimetric assay was calculated from the established cut-off value of 11 mg/L31 for the Ramco enzyme linked immunosorbent assay (ELISA) using the regression equation described by Pfeiffer et al33)
The log ratio of soluble transferrin receptor to ferritin (logTfR/Fer-ratio) was calculated.34
Soluble transferrin receptor was converted to values corresponding to the Ramco ELISA33
and then transformed to μg/L. The result was divided with the ferritin, and the logarithm (base 10) of the ratio was used. Total body iron (mg/kg) was approximated by the equation −(logTfR/Fer-ratio−2.8229)/0.1207.35
The primary outcome was infant haemoglobin and iron status (measured as serum ferritin, transferrin saturation, soluble transferrin receptors, reticulocyte haemoglobin, mean cell volume, mean cell haemoglobin concentration) at 4 months of age. To further assess iron status, we calculated iron deficiency, logTfR/Fer-ratio, and total body iron as described.
In this paper we report neonatal morbidity (including anaemia, polycythaemia, need for phototherapy, and respiratory symptoms). Other secondary outcomes are beyond the scope of the current paper, and will be reported separately: maternal postpartum haemorrhage and rates of successful umbilical arterial blood samples in relation to allocation group; neonatal serum immunoglobulin G level and infections during the first four months of life; effects of delayed and early cord clamping on iron status and haemoglobin at 12 months of age; psychomotor development at 4 and 12 months of age assessed by the Ages and Stages Questionnaire; reticulocyte haemoglobin as a marker of neonatal and infant iron stores.
As ferritin is the most sensitive indicator of iron status and has been shown to be the most efficient indicator of iron interventions,36
it was chosen for estimating the necessary group size. Evidence regarding the effects of early versus late cord clamping comes from a few trials in low or middle income countries, and we designed our study to be comparable with them. Beyond the neonatal period, the main effect observed in those studies is that delayed cord clamping leads to an increase in ferritin concentration at 3–6 months of age, which has been considered of importance to public health.16 20 22
We therefore chose iron status at 4 months as our main outcome and powered the study to find a difference in ferritin concentration.
A prestudy power analysis showed that a group size of 150 would allow us to find a difference of 29% in geometric mean serum ferritin concentration between groups at 4 months of age with a power of 80% and a significance level of 0.05, assuming a mean serum ferritin concentration of 110 μg/L in the delayed cord clamping group.37
The study was thus powered to find a slightly smaller difference in ferritin concentration than in a previous study in Mexico,20
where the delayed cord clamping group had 34% higher serum ferritin concentration at 6 months. Allowing for an attrition of 25%, we included 200 participants in each group.
Randomisation was performed by one of the investigators (MD) in advance by computer in blocks of 20 using the random number generator in MS Excel (Microsoft, Seattle, WA, USA).
The study design precluded either the mother giving birth or the midwife performing the intervention being blinded. Physicians performing neonatal examinations, staff members responsible for collection of blood samples and background data, and laboratory staff performing analyses of blood samples were blinded to each infant’s allocation group.
For group comparisons of continuous variables, we used Student’s t
test for variables with normal distribution. For variables with skewed distribution, we used Mann-Whitney U test for group comparisons and used Hodges and Lehmann estimator38
for confidence intervals across groups. Ferritin concentration was log10
transformed for analysis. Categorical variables were compared between groups by using Fisher’s exact test. We used SPSS for Windows, version 18.0 (SPSS, Chicago, IL, USA). We calculated the numbers needed to treat, relative risk reduction, and their confidence intervals using the web based JavaStat calculator.39
A P value <0.05 was considered significant.
All analyses were made on an intention to treat basis, except for 12 cases (four allocated delayed cord clamping, eight allocated early clamping) that were erroneously included in the study despite not fulfilling inclusion criteria. Of the 382 children finally analysed, 334 received the allocated intervention (168 for delayed clamping, 166 for early clamping) (figure).
Participant flow through study
We redid all analyses for the main and secondary outcomes, including cases of protocol breach at inclusion as well as per protocol (n=334), and this did not alter the results. We also did a sensitivity analysis for outcomes at 4 months of age by calculating imputed results for dropouts after 2 days of age (n=25), and the results did not change.