Study design
This multicenter randomized controlled trial was conducted at two clinics and one birth center in Kanagawa, Japan from December 2015 to November 2016. The study protocol was approved by the Institutional Review Board of St. Luke’s International University and registered with UMIN-CTR in Japan (UMIN000022573; dated June 01, 2016 - retrospectively registered, https://upload.umin.ac.jp/cgi-open-bin/ctr/ctr_view.cgi?recptno=R000023056). This clinical trial adhered to the clinical research ethical guidelines for human subjects established on April 27, 2015 by the Ministry of Education, Culture, Sports, Science and Technology, and the Ministry of Health, Labor and Welfare, Japan.
Participants
Participants were non-smoking pregnant Japanese women planning a vaginal birth and exclusive breastfeeding and neonates who were term, singleton, and cephalic presentations.
Participants were excluded if they had any maternal complications, fetal complications, or emergency cesarean section, were transferred to another hospital during pregnancy or delivery, were not literate in Japanese, or were unable to return in 4 months.
E.S chose study participants using medical records along with the study criteria. E.S or research assistants not in charge of the prenatal check-ups then recruited those participants who met the inclusion criteria. All potential participants who met the inclusion criteria were provided with both verbal and written research information, as well as a consent form to initial if they decided to partake in the study. After agreeing to participate in the study, they signed the consent form. For the inclusion of neonates in the study, their guardians (mother or father) were provided with the same explanation as that provided to pregnant women; they then signed the consent form on the infants’ behalf.
Randomization and masking
Randomization was performed centrally using the Mujinwari system (Iruka System Corporation [17], Tokyo, Japan) with a block size of four via access to the internet. Because of different characteristics of participants in each facility, randomization was stratified by institution. E.S or research assistants performed allocations when normal progress leading to vaginal delivery was predicted (i.e. full dilation of the cervix in primiparas or 6–8 cm dilation of the cervix in multiparas). Allocations were provided to the midwives who were to perform the intervention, and assistant midwives, who measured the time from neonatal delivery to cord clamping, were informed on the group assignment of the participants at the time of delivery. Masking was not applied in this trial. Because of the characteristics of the intervention, participants, and midwives in charge of the intervention were aware of treatment allocations. Research assistants and E.S, who evaluated the outcomes, were also aware of allocations.
Intervention
Midwives clamped the cord using a Kocher clamp after more than a minute following neonatal delivery or when the cord pulsation stopped in the intervention group and within 15 seconds after neonatal delivery in the control group. To assure similarity in settings, all neonates were placed on the chest or abdomen of their mothers just after delivery. Moreover, the position of the mothers was set to about 30-degree semi-Fowler position after neonatal delivery.
Midwives who were involved in the deliveries were instructed on the intervention procedure using the study protocol and had been trained before the study commenced.
Outcomes
The primary outcome was the spectrophotometric hemoglobin (SpHb) level at 4 months. Secondary outcomes were as follows: incidence of anemia (SpHb < 11.0 g/dL), TsB on days 3–5, incidence of “over the excess value of TsB,” incidence of phototherapy, transcutaneous bilirubin level on days 1–4, hematocrit on days 3–5, incidence of polycythemia, birth weight, infant vital signs (heart rate, respiratory rate, and temperature) after birth, infant growth (weight, height, head circumference, and chest circumference) at 1 and 4 months old, adverse effects (seizures, admission to neonatal intensive care units, neonatal death), and maternal hemorrhage (third stage, within 2 hours). Data were collected at the following time points: during hospital or clinic stay and at 1 and 4 months old.
Measurements
SpHb monitoring
Non-invasive and continuous Hb measurements using pulse oximeter (Radical-7®; Mashimo, Irvine, CA, USA) technology was conducted to determine the SpHb levels at 4 months. The non-invasive estimation of the Hb level by pulse-CO-oximetry in infants and neonates undergoing surgery showed a significant correlation with the invasive standard laboratory measurement of total Hb (r = 0.73, p < .00) and demonstrated clinically acceptable agreement with standard laboratory Hb measurements [18]. This continuous monitoring of SpHb in the stable state was required during measurement. E.S measured the value 2 minutes after a stable SpHb was detected.
Anemia
If the SpHb level was less than 11.0 g/dL at 4 months, infants were noted to be anemic.
TsB
Blood samples were collected for measuring the concentration of TsB on days 3–5. To minimize the invasiveness of the procedure for neonates, blood samples for measuring TsB were collected during routine blood sampling for congenital metabolic disorder mass screening. Blood was collected using hematocrit capillary tubes by a midwife or nurse and immediately centrifuged; then, the TsB level was measured using a BL-300 jaundice meter® (TOITU, Tokyo, Japan) or similar device.
Excess value of total serum bilirubin
The bilirubin nomogram by Imura [19] for prediction of hyperbilirubinemia was used because it is clinically most commonly used in Japan.
Phototherapy
The actual number of neonates who received phototherapy was counted.
Transcutaneous bilirubin
A non-invasive bilirubinometer (JM-103® or JM-105®; Konica Minolta, Tokyo, Japan) was used to measure the transcutaneous bilirubin concentration by putting the device to the chest and forehead of the neonate on days 1–4. The value used was determined as the average of one measurement each from the forehead and chest. Transcutaneous bilirubin shows a high correlation with TsB [20] and is routinely used for screening for hyperbilirubinemia in clinical settings.
Hematocrit
The same blood samples collected for measuring TsB were also used for determining the hematocrit. The hematocrit was measured after centrifuging capillary blood.
Polycythemia on days 3–5
If the hematocrit level was ≥ 65%, infants were noted to be polycythemic.
Data collection
Mothers and neonates were checked by midwives or nurses immediately after birth, at 1 hour, and then at 2 hours and cared for according to routine practice. The data on maternal hemorrhage and neonatal vital signs were collected at those times. In two facilities, mothers stayed 5 days, except for those who were willing to be discharged earlier or needed to be sent to another hospital for abnormalities. In one clinic, the routine stay was 3 days. In all facilities, they had rooming-in, and breastfeeding was encouraged from soon after birth. All neonates’ transcutaneous bilirubin levels were checked using a bilirubinometer every morning. At day 4, they usually had blood collection for mass screening. At that time, additional blood was collected for measuring TsB and hematocrits. However, the neonates whose transcutaneous bilirubin levels were high during daily checks had blood tests performed on other days as well. For the neonates who had blood tested several times, the highest values of the TsB or hematocrit were used as outcome data. After discharge, they usually had a check-up at 2 weeks to evaluate weight and jaundice, and the mothers also had support for breastfeeding from midwives. At the 1-month check-up, the doctor or midwives evaluated the neonates’ health including growth, jaundice, and nutrition (exclusive breastfeeding, mixed, or formula milk). One month’s data were collected at that point. At 4 months, the researcher collected the data of neonates’ growth, nutrition, and SpHb. At a time convenient for the mothers, E.S or research assistants collected demographic characteristics of mothers and neonates and delivery outcomes.
Sample size
This study explored the effects of cord clamping timing on Hb status at 4 months old by comparing the DCC and ECC groups. From a previous cohort study [15], the estimated mean SpHb value in the ECC group was about 11.5 g/dL. Previous studies [21, 22] indicated that the Hb level was 11.5 g/dL in the ECC group and 12.0 g/dL in the DCC group, and an effect size of 0.5 g/dL was predicted. For the primary outcome measure of SpHb, based on 80% power to detect a significant difference of 0.5 g/dL with a 0.8 g/dL standard deviation (α = 0.05, two-sided), 40 participants were required for each study group. Assuming a dropout rate of 20%, a minimum total of 100 patients was required.
Statistical analysis
Descriptive statistics were used to summarize the participant’s backgrounds. To compare the DCC and ECC groups, a t-test was used for the primary outcome of mean difference (MD) in SpHb. For secondary outcomes, t-testing was used for MDs, the Mann-Whitney U test was used for non-parametric variables, and the chi-squared test was used for bivariate analysis. Risk ratios (RR) and their confidence intervals (CI) were calculated, as well as confidence intervals for difference in means.
The primary outcome was assessed according to the intention-to-treat principle. One participant in the DCC group was not included in the analysis because their data was missing due to technical difficulties. However, one subject in the DCC group who did not complete the protocol was included. Secondary outcomes were assessed using intention-to-treat principles for continuous variables and intention-to-treat for nominal variables. Per-protocol analysis and as-treated analysis were performed as adjunct methods. In per-protocol analysis, participants who deviated from the protocol were excluded from the analysis. In as-treated analysis, analysis was conducted as if participants were treated, regardless of allocation. All data were analyzed using IBM SPSS Statistics for Windows version 24.0 (IBM Corp., Armonk, NY).