Study design
We conducted a randomized controlled trial (parallel group study with 1:1 randomization) of pregnancies, finalizing inclusion only after birth of the baby. Ethical approval was obtained from the Nepal Health Research Council, reg. no. 92/2015. Trial registration: Clinicaltrials.gov NCT02727517.
Setting
The study was performed at the low risk-Maternal and Neonatal Service Centre (MNSC) at Paropakar Maternity and Women’s Hospital in Kathmandu, Nepal. Part of the research team had performed an intervention project introducing a simplified algorithm for neonatal resuscitation, Helping Babies Breathe (HBB) at the hospital [23]. The HBB algorithm is routine at the hospital and emphasizes that within the first minute after birth, “The Golden Minute”, a baby should breathe or receive positive-pressure ventilation. All babies who do not cry after drying should have their airways cleared and receive stimulation to breathe, followed by positive-pressure ventilation if necessary [4]. Before the study all staff underwent a recapitulation of the HBB algorithm including training sessions with the modified algorithm performing intact cord resuscitation. The research team members consisted of trained nurse-midwives with experience in clinical research, five on duty during all hours: one at the antenatal care unit, three at the delivery unit and one at the postnatal care unit. A 10-day training program was provided on selection criteria of women, taking consent, Apgar (including breathing) assessment and placement of the neonatal Masimo© pulse oximeter probe (Masimo Corporation, Switzerland).
Inclusion and consent
The hospital criteria for admission to MNSC were: uncomplicated pregnancies, no complication at admission, healthy mothers (no clinical history of hypertension, infection, diabetes, or chronic medical condition), expected vaginal delivery, gestational age between 34 and 41 weeks, and singleton pregnancy.
Women were eligible to participate in the study if they were assigned to MNSC. Mothers admitted to MNSC were informed and asked for consent. Four research team members were present at admission and at the delivery department at all hours and days of the week. Newborns were eligible and included if they had reached a gestational age ≥ 33 weeks, were in need of resuscitation according to the HBB algorithm, that is: no breathing despite thorough drying and additional stimulation within 30 s after birth [24]. Newborns not in need of resuscitation were included in a parallel study, Nepcord II [25].
Exclusion criteria: Monochorionic twins (from an ultrasound scan) or clinical evidence of twin-twin transfusion syndrome, triplets or higher order multiple pregnancy, and fetuses with known congenital malformation.
Randomization
Sequentially numbered, opaque envelopes containing the allocation were prepared by one of the authors (AKC) using a random digit generated list. AKC had no further clinical involvement in the trial.
Randomization was performed as follows: resuscitation with an intact umbilical cord close to the mother in her bed (Fig. 1), or standard care (resuscitation according to the HBB algorithm including early cord clamping) at a designated area, on a resuscitation table in a room next to the delivery room (Fig. 2). Delay in cord clamping was instructed to be at least 180 s in the intact cord group. Allocation was decided when birth was imminent, the research team member opened the next opaque numbered envelope and informed the staff managing the delivery on handling strategy of the allocated group.
Blinding
Due to the nature of the intervention and the close proximity in time to outcome measures, blinding was not possible.
Data collection
When the child was born, 2–3 research team members collaborated to start a timer, apply a pulse oximetry sensor on the right hand/arm of the newborn infant, and record time of first breath, time for regular breathing, time for cord clamping as well as Apgar score.
Primary outcome
Saturation by pulse oximetry (SpO2) at 10 min after birth.
Secondary outcome
-
SpO2 at five minutes after birth
-
Time to reach > 90% in SpO2
-
Newborn heart rate at one, five and ten minutes after birth measured by pulse oximeter.
-
Apgar score at one, five and ten minutes after birth
-
Timing of recorded first breathing effort and established regular breathing
-
Rectal temperature at 10–20 min
-
Bilirubin at discharge from hospital
-
Health status the day after birth (12–36 h after birth) – Alive or deceased
Background information regarding the women and pregnancies were collected from the hospital records by a research team member. All data were transferred to an individual datasheet. Data sheets were checked for quality and stored in a locked room at the hospital until the end of the study. Data were then entered into a database and used for statistical analyses.
Sample size
A power analysis shows that a group size of 99 would allow us to find a difference of 2% in SpO2 between groups at 10 minutes after birth with a power of 80% and a significance level of 0.05, assuming a mean saturation (SD) of 96% (5) in the delayed cord clamping group [26]. Allowing for an attrition of 20%, we planned to include 120 participants in each group.
Statistical analysis
For group comparisons of continuous variables, we used unpaired Student’s t-test for variables with normal distribution and Mann-Whitney U test for variables with skewed distribution. Categorical variables were compared between groups using Fisher’s exact test.
We calculated the numbers needed to treat, relative risk reduction, and their confidence intervals using the web based JavaStat calculator. A p-value < 0.05 was considered significant.
All analyses were made on an intention to treat basis as prespecified in the study protocol.
We performed all analyses for the primary and secondary outcomes, including only cases handled according to allocation (per protocol). Results are reported in Table 3.
Sensitivity analysis
To examine if adjusted analysis would yield results different from unadjusted analysis, we analyzed the correlation coefficients between baseline data that was not randomly distributed between treatment groups with primary and secondary outcomes with Spearman’s rank correlation.
Auxiliary analysis
Due to uneven group size after randomization and a large protocol violation rate in the intact cord group, we also chose to do alternate analysis including only infants where resuscitation was performed with bag and mask, reported in Table 4. An ANOVA analysis including post-hoc multiple comparisons comparing the infants divided into three groups, intact cord resuscitation per protocol (C ≥ 180 s), intact cord resuscitation protocol violation (CC < 180 s) and early CC was also performed [see Additional file 1].