EBNEO Commentary: Umbilical cord milking in nonvigorous infants: a cluster-randomized crossover trial

November 30, 2022


Vonita Chawla MD
Assistant Professor
Department of Pediatrics – Division of neonatology
University of Arkansas for Medical Sciences, Little Rock, AR.


Vonita Chawla
Department of pediatrics, ACH neonatology
#1 Children’s Way, Slot 512-5 Little Rock, AR 72202, USA
Office: (501) 364-1028
Fax: (501) 364-4264
Email: chawlavonita@uams.edu

Title of write-up     

EBNEO Commentary: Umbilical cord milking in nonvigorous infants: a cluster-randomized crossover trial

Manuscript citation

Katheria AC, Clark E, Yoder B, Schmölzer GM, Yan Law BH, El-Naggar W, Rittenberg D, Sheth S, Mohamed MA, Martin C, Vora F, Lakshminrusimha S, Underwood M, Mazela J, Kaempf J, Tomlinson M, Gollin Y, Fulford K, Goff Y, Wozniak P, Baker K, Rich W, Morales A, Varner M, Poeltler D, Vaucher Y, Mercer J, Finer N, El Ghormli L, Rice MM; of the Milking In Nonvigorous Infants group. Umbilical cord milking in nonvigorous infants: a cluster-randomized crossover trial. Am J Obstet Gynecol. 2022 Aug 13:S0002-9378(22)00649-4. doi: 10.1016/j.ajog.2022.08.015. PMID: 35970202. 

Type of investigation

Randomized controlled trial


 In (P) non-vigorous newborns born between 35 and 42 weeks’ gestation, does (I) umbilical cord milking (UMC) versus (C) early cord clamping (ECC) reduce (O) admission to the neonatal intensive care unit in the (T) first 24 hours of life.


  • Design: Cluster-randomized crossover trial that included ten medical centers in 3 countries between January 2019 and May 2021
  • Allocation: A computer-generated allocation sequence was used to randomize participating hospitals 1:1 to Umbilical Cord Milking (UCM) or Early Cord Clamping (ECC) in the first period until half of the required enrollment was reached. Hospitals were then crossed over to the other intervention during the second period for the remaining half of consented subjects.
  • Blinding: Due to the nature of the interventions, personnel providing UCM or ECC could not be blinded. All the NICU admissions were independently and blindly reviewed. The outcome assessors were blinded, and statisticians reported data as groups A and B until the data were locked.
  • Follow-up period: The primary outcome was admission to the NICU in the first 24 hours of life. For secondary outcomes, such as peak bilirubin, the data were collected for the length of the hospitalization.
  • Setting: 10 multinational sites (7 in the United States, 2 in Canada, and 1 in Poland)
  • Patients: Newborns delivered between January 2019 and May 2021 were screened for eligibility.

Inclusion criteria:

  • Newborns delivered at 35 – 42 weeks’ gestation who were non-vigorous at birth.

Exclusion criteria:

  • Newborns who were vigorous at birth
  • Newborns born with known major congenital or chromosomal anomalies
  • Newborns born with known cardiac defects other than small VSDs
  • Newborn births complicated by complete placental abruption or cutting through the placenta at the time of delivery.
  • Monochorionic multiple births
  • Cord anomalies such as avulsion or true knots, presence of a nonreducible nuchal cord.
  • Newborn births with incomplete delivery data
  • Neonates delivered by an obstetrical provider not trained in the study protocol.
  • Intervention: Non-vigorous newborns received UCM or ECC according to the randomization assignment. For UCM, during C-section, the obstetrical providers placed the newborn below the level of the incision while a second team member milked the cord 4 times. For vaginal delivery, the obstetrical provider held the infant or placed the infant on the mother’s abdomen, and the cord was milked 4 times by either the obstetrical provider or a second team member. For UCM, the provider milked 20 cm of cord over 2 seconds, repeating 3 additional times. For ECC, the provider clamped the cord within 60 seconds of birth.
  • Outcomes:

Primary outcome:  Admission to the NICU in the first 24 hours of life for:

      • respiratory distress
      • bradycardia or tachycardia
      • hypotonia/hypertonia or irritability
      • lethargy or difficulty arousing
      • poor feeding or emesis
      • hypoglycemia
      • oxygen desaturations or cyanosis
      • need for oxygen
      • apnea
      • seizures or seizure-like activity
      • hyperbilirubinemia
      • temperature instability

Secondary outcomes:

      • need for therapeutic hypothermia, volume expanders, phototherapy
      • hemoglobin at 24 hours of life
      • peak serum bilirubinThe principal safety outcome was HIE.

The exploratory outcomes included resuscitation interventions, blood pressure, and length of hospitalization

Analyses and Sample Size:

Sample size

A sample size of 1200 was estimated as necessary based on the following:

  • 35% relative reduction in NICU admissions (16% for UCM vs 25% for ECC)
  • Type I error alpha=0.05, 85% power, 0.02 rho (within cluster within period correlation)
  • 02 eta (within cluster between period correlation)
  • a correction factor (4 x cluster size) for the small number of clusters


  • Analyses were performed using the intention-to-treat concept.
  • Wilcoxon rank-sum test was used for continuous variables and the chi-square or Fisher exact tests were used for categorical variables.
  • For the primary outcome and secondary outcomes with adequate numbers (i.e., cardiorespiratory support, Apgar score at 1 minute ≤3, and Apgar score at 5 minutes ≤6), hierarchical generalized linear mixed models were used.
  • The odds ratio (OR) and 95% confidence intervals (CI) were estimated.
  • A mixed model using least squared means was employed for continuous outcomes to estimate the least square means difference and 95% CI.
  • For multinomial outcomes, multinomial logistic regression was used with fixed treatment group effect.
  • For all rare secondary binomial outcomes, logistic regression was used with fixed treatment group effect.
  • No interim analysis for efficacy was performed, a P value <.05 was used to define statistical significance, and all tests were 2-tailed. The P values for secondary outcomes were not adjusted for multiple testing for safety reasons.
  • Analyses were performed using SAS, version 9.4; SAS Institute Inc, Cary, NC.

Subgroup analyses                            

  • An evaluation of the consistency of the primary outcome across racial-ethnic and neonatal sex subgroups was performed.
  • An evaluation of treatment by site interaction was included.
  • Patient follow-up:
    • 16,234 newborns were screened at 10 hospitals and 1780 were eligible (905 UCM, 875 ECC).
    • 1730 had primary outcome data for analysis (97% of eligible; 872 UCM, 858 ECC).

Main results

  • The frequency of NICU admission (primary outcome) between the UCM (199/872 = 23%) and ECC (239/858 = 28%) groups was different, however, after accounting for study design by site, it was not statistically significant (OR, 0.69; 95% CI, 0.41-1.14).
  • The most common cause for NICU admission was respiratory distress (UCM 19% vs ECC 22%).
  • The infants in the UCM group were less likely to have delivered in the first period i.e. the first phase of the study during which hospitals carried out UCM or ECC as per initial randomization, prior to crossover (42% vs 58%; P<.001) and more likely to have maternal hypertension (20% vs 16%; P=.04).
  • There was 1 death in the delivery room in the ECC group with a postmortem diagnosis of HIE.

Secondary outcomes

Secondary outcome UCM




Crude odds ratio


Odds ratio (95% CI) accounting for study design
Cardiorespiratory support 460/758 (60.7) 515/730 (70.5) 0.64 (0.52-0.80) 0.57 (0.33-0.99)
Therapeutic hypothermia 21/829 (2.5) 35/806 (4.3) 0.57 (0.33-0.99)
Volume bolus 58/823 (7.1) 56/790 (7.1) 0.99 (0.68-1.45)
Phototherapy 88/606 (14.5) 77/613 (12.6) 1.18 (0.85-1.64)
HIE 27/828 (3.3) 38/806 (4.7) 0.68 (0.41-1.13)
Mild HIE 15 (1.8) 14 (1.7)
Moderate HIE 8 (1.0) 19 (2.4)
Severe HIE 4 (0.5) 5 (0.6)
Moderate-to-severe HIE 12/828 (1.4) 24/806 (3.0) 0.48 (0.24-0.96)
Apgar score at 1 min ≤3 227/757 (30) 248/730 (34) 0.83 (0.67-1.04) 0.79 (0.53-1.18)
Apgar score at 5 min ≤6 119/756 (15.7) 137/730 (18.8) 0.81 (0.62-1.06) 0.77 (0.58-1.03)
Death before discharge 0 4 (0.5)
Severe IVH 1 (0.1) 2 (0.2)


UCM in the near-term/term non-vigorous newborn was not associated with a reduction in NICU admissions compared with ECC. UCM was associated with a reduction in cardiorespiratory support in the delivery room, fewer cases of moderate-to-severe HIE, lower use of therapeutic hypothermia, and higher hemoglobin. There was no evidence of harm associated with UCM compared with ECC.


Earliest reports of umbilical cord milking are from the 1950s when higher hemoglobin levels were reported in babies that underwent umbilical cord milking (UCM) compared to babies that underwent immediate cord clamping (1). A recent survey conducted by Chiruvolu et al showed large practice variation across the United States regarding umbilical cord management (2). In this study, 57.3% of neonatologists reported not performing UCM, and it was less likely to be performed in teaching hospitals and in level 4 NICUs (2).

Neonatal Resuscitation Program (NRP) recommends delayed cord clamping (DCC) for most vigorous term and preterm infants (3) and does not recommend UCM in infants <28 weeks GA (gestational age) due to increased risk of IVH in this population (4, 5). However, there is debate surrounding UCM in non-vigorous or asphyxiated infants >28 weeks GA. The European Resuscitation Council 2021 guidelines recommend considering UCM for infants >28 weeks GA when DCC is not possible (6).

In 2021, the International Liaison Committee on Resuscitation (ILCOR) published two meta-analyses pertaining to umbilical cord management in preterm and term infants (7,8). The current ILCOR recommendations are summarized here:

For infants <34 weeks GA:

      1. Intact cord milking (ICM) is a reasonable alternative to deferring cord clamping (weak recommendation, moderate-certainty evidence), in infants born between 28-34 weeks’ GA, not requiring immediate resuscitation.
      2. ICM for infants born at <28 weeks’ GA is advised against (weak recommendation, very low–certainty evidence).
      3. In infants born at <34 weeks’ GA requiring immediate resuscitation, there is insufficient evidence to make a recommendation.

For infants =/>34 weeks GA:

       In the systematic review comparing DCC vs ICM, only a single study was identified. As a result, no recommendations are made.

In this study, the authors have studied the effects of UCM vs ECM on non-vigorous, near-term and term infants. Specific postnatal criteria were used to define “non-vigorous” infants (presence of poor tone, pallor, lack of breathing despite initial resuscitation efforts). While there was no statistically significant difference in NICU admissions, infants in the UCM group required less cardiorespiratory support in the delivery room and had a decreased incidence of moderate-to-severe HIE compared to infants that underwent ECC. The risk of death was not increased with UCM.

The authors mention that the infants in this study are being followed for two-year neurodevelopmental outcomes. It will be interesting to see the impact of UCM on the neurodevelopment of these infants that were non-vigorous at birth but had a lower incidence of moderate-to-severe HIE compared to the ECC group.

Based on this study’s outcome, UCM is recommended as a safe alternative to DCC in non-vigorous infants =/>35 weeks GA, compared to ECC. This is especially relevant in settings where resuscitation skills and resources needed for caring for asphyxiated neonates are limited. Additionally, per ILCOR recommendations, the use of UCM in vigorous infants 28-34 weeks may be considered, while keeping in mind that this a weak recommendation. Furthermore, a standardized approach to cord milking is needed, while also accounting for patient-related factors.


        1. Siddall RS, Crissey RR, Knapp WL. Effect on cesarean section babies of stripping or milking of the umbilical cords. Am J Obstet Gynecol. 1952 May;63(5):1059-64. doi: 10.1016/0002-9378(52)90546-2. PMID: 14923706.
        2. Chiruvolu A, Mallett LH, Govande VP, Raju VN, Hammonds K, Katheria AC. Variations in umbilical cord clamping practices in the United States: a national survey of neonatologists. J Matern Fetal Neonatal Med. 2022 Oct;35(19):3646-3652. doi: 10.1080/14767058.2020.1836150. Epub 2020 Oct 20. PMID: 33081557.
        3. Weiner GM, Zaichkin J. Initial steps of newborn care. Textbook of neonatal resuscitation – 8th edition. 2021 Jun.
        4. Aziz K, Lee HC, Escobedo MB, Hoover AV, Kamath-Rayne BD, Kapadia VS, et al. Part 5: Neonatal Resuscitation: 2020 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Circulation. 2020 Oct 20;142(16_suppl_2):S524-S550. doi: 10.1161/CIR.0000000000000902. Epub 2020 Oct 21. PMID: 33081528.
        5. Katheria A, Reister F, Essers J, Mendler M, Hummler H, Subramaniam A, et al. Association of Umbilical Cord Milking vs Delayed Umbilical Cord Clamping With Death or Severe Intraventricular Hemorrhage Among Preterm Infants. JAMA. 2019 Nov 19;322(19):1877-1886. doi: 10.1001/jama.2019.16004. PMID: 31742630; PMCID: PMC6865839.
        6. Perkins GD, Graesnar J, Semeraro F, Olasveengen T, Soar J, Lott C et al., European Resuscitation Council Guidelines 2021: Executive summary, Resuscitation (2021), https://doi.org/10.1016/j.resuscitation.2021.02.003
        7. Seidler AL, Gyte GML, Rabe H, Díaz-Rossello JL, Duley L, Aziz K, et al. International Liaison Committee On Resuscitation Neonatal Life Support Task Force. Umbilical Cord Management for Newborns <34 Weeks’ Gestation: A Meta-analysis. Pediatrics. 2021 Mar;147(3):e20200576. doi: 10.1542/peds.2020-0576. PMID: 33632931; PMCID: PMC7924139.
        8. Gomersall J, Berber S, Middleton P, McDonald SJ, Niermeyer S, El-Naggar W, et al. International Liaison Committee On Resuscitation Neonatal Life Support Task Force. Umbilical Cord Management at Term and Late Preterm Birth: A Meta-analysis. Pediatrics. 2021 Mar;147(3):e2020015404. doi: 10.1542/peds.2020-015404. PMID: 33632933.

Leave a comment