Effect of umbilical cord milking versus delayed cord clamping on preterm neonates in Kenya: A randomized controlled trial

Background Delayed cord clamping (DCC) is a placental to new-born transfusion strategy recommended by obstetric and gynaecological societies. Though not widely adopted, umbilical cord milking (UCM) may achieve faster transfusion when DCC cannot be performed such as when a neonate requires resuscitation. Methods Pragmatic, two-arm, randomized clinical trial in which consenting women in spontaneous labour or provider-initiated delivery at 28 to less than 37 weeks at Kenyatta National Hospital in Nairobi, Kenya, were enrolled. At delivery, stable preterm infants were randomized to UCM (4 times) or DCC (60 seconds). Neonatal samples were collected for analysis at 24 hours after delivery. Maternal primary PPH (within 24 hours) and neonatal jaundice (within 1 week) were evaluated clinically. The primary outcome was the mean neonatal haemoglobin level at 24 hours after birth. Modified Intention to treat analysis was used for all outcomes. P-value was significant at p<0.05. Results Between March 2018 to March 2019, 344 pregnant women underwent screening, and 280 eligible participants were randomized when delivery was imminent. The intervention was not performed on 19 ineligible neonates. Of the remaining 260 neonates, 133 underwent UCM while 128 underwent DCC. Maternal and neonatal baseline characteristics were similar. The mean neonatal haemoglobin (17.1 vs 17.5 grams per decilitre, p = 0.191), haematocrit (49.6% vs 50.3%, p = 0.362), anaemia (9.8% vs 11.7%, p = 0.627), maternal PPH (2.3% vs 3.1%, p = 0.719) were similar between UCM and DCC respectfully. However, neonatal polycythaemia (2.3% vs 8.6%, p = 0.024) and neonatal jaundice (6.8% vs 15.6%, p = 0.024) were statistically significantly lower in UCM compared to DCC. Conclusion UCM compared to DCC for preterm neonates resulted in similar outcomes for neonatal haemoglobin, haematocrit, anaemia and maternal primary PPH and a lower proportion of neonatal polycythaemia and clinical jaundice. UCM offers a comparable method of placental transfusion compared to DCC and may be considered as an alternative to DCC in preterm neonates at 28 to <37 weeks’ gestation. This was a parallel, pragmatic, randomised controlled trial of intact umbilical cord milking (intervention) versus delayed cord clamping (control) in a ratio of 1:1. The trial protocol was developed by the authors and approved by the Kenyatta National Hospital/University of Nairobi Ethical Review Committee (KNH-UON-ERC), P677/11/2017. Ethical principles expressed in the Declaration of Helsinki were used to conduct the trial. Verbal and written informed consents were sought from each participant. Adequate explanation and counselling were done before attaining consent. Due to the nature of the study, Antenatal clinic (ANC) consent was not practical due to the unpredictability of PTB and generally erratic ANC attendance; majority of the consent was taken while the mother was already in labour. Although this is a vulnerable period for the mother, the nature of such a condition did not allow for earlier consent. The consent process was thorough and delicately handled so as not to overwhelm the mother. This process was free from coercion and any patient who declined consent was excluded from the study. Study participants were monitored by study assistants and clinicians in the respective wards. An independent data safety monitoring board (DSMB) was formed, whose members reviewed trial safety and progress. Safety and progress reports were submitted to the KNH-UON-ERC in case of occurrences of any adverse events. This study was carried out at the Kenyatta National Hospital (KNH) labour ward, maternity theatres, post-natal wards, new born unit (NBU) and the haematology laboratory. The KNH is the largest public referral and teaching hospital in Kenya receiving patients from Nairobi and its environs as well as referrals from all other hospitals in Kenya. It has an average of 1,000 deliveries per month. These deliveries occur amongst mothers of varying socioeconomic status. The standard of care after delivery for newborns in KNH is DCC and even though UCM is not contraindicated it is rarely practiced. Neonates fitting to the admission criteria of birth weight < 1900 and other medical conditions were managed in the NBU. There was availability of a neonatal ICU (NICU) at KNH in case a neonate required advanced care. The NICU is equipped with a limited number of ventilation machines. Surfactant is available although majority of the patients are unable to afford the cost. Entry criteria included mother-baby pairs between 28 to < 37 weeks gestational age. Gestational age was confirmed by use of last menstrual period dating or if available a first-trimester ultrasound scan. Further to validate the gestational age, a Ballard score was undertaken for each newborn in the study. We excluded pregnancies with multiple gestations, neonates with congenital abnormalities, red cell isoimmunisation, HIV and VDRL positive women, advanced stage of labour (>7cm dilated), antepartum haemorrhage including placenta praevia or placenta abruption. Pregnancies in which neonates required resuscitation as a priori during the antenatal and immediate postnatal period as well as patients who were unwilling to undergo randomization were also excluded. Randomization occurred in blocks to obtain 140 participants in each arm. Block randomization was used in equal blocks of 14 until the total sample size was achieved. For allocation concealment, the randomization instructions were given to the investigator and research assistants in sequentially numbered, opaque, sealed, identical envelopes with an unpredictable allocation code. Randomization was done by the study midwives when delivery was imminent, for mothers in labour or, at the induction of anesthesia for caesarean deliveries. Potential study participants were recruited by the attending midwife at the triage area in the labour ward and through the antenatal wards. The subjects were identified and chosen for the study if they met the eligibility criteria. Women were enrolled at the onset of spontaneous labour (cut off point >7 cm cervical dilation) if they were in labour, or when the decision had been made for caesarean delivery for those not in labour. All consenting participants were monitored as labour progressed until delivery was imminent, when at this point they were immediately randomized into either intervention group or control group. Once delivery was imminent, the recruited midwife immediately opened a sealed, numbered, opaque envelope containing the treatment allocation. Neonates underwent either intact UCM or DCC. For neonates that underwent intact UCM, with one hand, the cord was pinched and held closest to the placental end, the other hand was used to milk blood towards the infant where at the umbilical end point the cord was held. This served as one milking motion. The cord was then released at the placental end and allowed to refill over 1 to 2 seconds between each milking motion. This was repeated a total of 4 times. After milking, the cord was clamped and cut, and the neonate handed over to the neonatal receiving team. For neonates that underwent DCC, the cord was clamped and cut after waiting for 60 seconds. The time was recorded using a stop watch timer in the labour ward delivery room for those who had spontaneous vaginal delivery and a stop watch timer in the maternity theatre for those who had a caesarean section. During the intervention the new-born was positioned on the maternal abdomen after vaginal delivery and on the maternal anterior thigh after cesarean delivery. The participants, research assistants collecting data and the laboratory staff performing analysis of blood samples were blinded but, the study staff were not due to the nature of the intervention. After the intervention, routine new-born care was initiated. Immediate cord clamping was implemented in the event that neonatal resuscitation was required. After delivery, the estimated maternal postpartum blood loss was recorded by the midwife by visual estimation, where a quantitative or semi quantitative estimate was made within the immediate post-partum period. This is the method routinely used at KNH to estimate PPH. If the mother was stable, she was then discharged to the post-natal ward. During post-natal care, mothers were interviewed within 24 hours of delivery to complete the questionnaire. Clinicians and midwives involved in the delivery were trained on the process of intact UCM via clinical tutorials and videos. With assistance from the paediatrician, who was part of the study team, clinical examination of the neonate was carried out observing for yellow discolouration in the periphery of the ocular conjunctiva and in the oral mucous membranes. Mothers were trained by the paediatrician before discharge on how to detect jaundice and were given the contact information of the principal investigator. This was guided by pictures and videos on how to detect jaundice [9]. Neonatal follow up was until 1 week after delivery, Mothers who were discharged earlier received telephone calls on day three as well as day seven postpartum for neonatal follow-up and data collection. Neonatal blood samples were collected using a 23-gauge needle from a peripheral vein amounting to 0.5mls. Blood was collected in EDTA tubes (BD Vacutainer®) 24 hours after delivery. The blood samples were transported in a blood transport box by the research assistant to the Kenyatta National Hospital laboratory where they were analysed for a complete blood count using the Sysmex XN-550 automated haematology analyser. The analyser equipment was calibrated with standard calibrators. The internal quality control was set daily every morning, done by running a known quality control sample along with tests, to confirm the validity of the values of the tests. There was no time lag between the withdrawal of blood samples and analysis. The primary outcome was the mean neonatal venous haemoglobin measured at 24 hours of life. The secondary outcomes were the mean neonatal haematocrit at 24 hours of life, the prevalence of neonatal anaemia and neonatal polycythemia at 24 hours of life, the prevalence of neonatal clinical jaundice within the first week of life and the incidence of maternal primary postpartum haemorrhage. Neonatal anaemia was defined as hemoglobin concentration <2SD below the normal value for gestational age [10]. Neonatal polycythemia was defined as hematocrit concentration >2 SD above the normal value for gestational age [10]. Primary PPH was defined as within 24 hours a blood loss of >500 ml after vaginal or >1000 ml after cesarean delivery. Data were collected using a closed ended pre-tested data extraction form checked for completeness and correctness and entered into Microsoft Access with in-built consistency and validation checks. Obstetric and medical information were collected from maternal health care records from the time of admission to discharge. The sample size was calculated on the basis of a previous trial done by Katheria et al, and using the formula by Chow S (Shao J, Wang H. 2008) for a 2 tailed α value of 0.05 and a power of 80%, 128 participants were required in each group to detect a mean difference in haemoglobin of 0.7g/dl with a standard deviation of 2. To account for attrition, an additional 10 percent were added to the total sample size. Therefore, 280 participants (140 in each arm) were required for the study. The Biostatistician participated in study design and performed all the analysis and was blinded to allocation groups. Modified intention to treat analysis were used where, some patients who were deemed ineligible after randomization or certain patients who never received the intervention were excluded from analysis. Data were locked until completion of the study thereafter exported to STATA version 13.0 for analysis. Continuous variables were compared using two-sample t-tests and categorical variables compared using chi-squared tests or Fischer’s exact tests. Outcomes were compared between the two arms using t-test for continuous variables and chi-square test of independence for categorical variables. Risk estimates and corresponding 95% confidence intervals were obtained for primary and secondary outcomes. P< 0.05 was considered statistically significant. All statistical assumptions were met, and validation of final analysis was done. This trial was registered at the Pan African Trial Registry (PACTR201906770028735). Although the trial was registered on 3rd March 2018, before enrollment of any patient, the approval was obtained later, therefore, the trial has been registered as a retrospective trial. http://www.pactr.org.