Umbilical Cord Management for Newborns <34 Weeks' Gestation: A Meta-analysis

To determine the effects of different umbilical cord management strategies (including timing of clamping and cord milking) at preterm birth <34 weeks’ gestational age.

This review was conducted by following the methodology outlined in the Cochrane Handbook for Systematic Reviews of Interventions and adheres to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guidelines.^19,20  The protocol was registered prospectively with the International Prospective Register of Systematic Reviews (PROSPERO, CRD42019155475). Full methods are detailed in Appendix 1 in Supplemental Information.

We considered all RCTs and cluster RCTs in which researchers compared alternative umbilical cord management strategies at preterm birth <34 weeks’ gestational age or with low birth weight <2500 g. Studies were included if the authors reported a mean gestational age of <34 weeks or if >80% of the births were <34 weeks’ gestation.

Studies in which researchers compare the following umbilical cord management interventions were included in this review:

1. ECC, defined as application of a clamp to the cord <30 seconds after birth, without cord milking;

2. DCC, defined as application of a clamp to the cord ≥30 seconds after birth or based on physiologic parameters (such as when cord pulsation has ceased or breathing has been initiated), without cord milking;

3. intact-cord milking (ICM) (also referred to as “stripping”), defined as repeated compression of the cord from the placental side toward the infant with the connection to the placenta intact at any time point within the first few minutes after birth; and

4. cut-cord milking (CCM) (also referred to as “stripping”), defined as drainage of the cord by compression from the cut end toward the infant after clamping and cutting a long segment.

Review outcomes were selected in consultation with representatives from the World Health Organization and ILCOR. They comprised infant and maternal outcomes that were seen as clinically relevant and therefore likely to change clinical practice.²¹  All outcomes and their definitions have been summarized in Table 1. Prespecified subgroup analyses, search strategy, study selection, data extraction, risk of bias evaluation, certainty of evidence assessment, and data synthesis are detailed in Appendix 1 in Supplemental Information.

Forty-two studies (reported in 102 articles) including 5772 infants met the inclusion criteria for the review, of which 41 studies (including 5676 infants) had data that could be included in the meta-analysis (Fig 1, Appendix 3 in Supplemental Information: full list of included studies per comparison).

Study characteristics and participant characteristics for the included studies are outlined for each comparison in Tables 1a–1d in the Supplemental Information and Tables 2–5, respectively.

All of the included studies were individual RCTs (unit of randomization was either the mother or the infant). Studies were undertaken in a range of countries (although most were high income by World Bank country classifications^22). Most studies excluded infants with complications such as major malformations or congenital anomalies.

Risk of bias is summarized in Fig 2. The majority of studies were at low risk of selection bias (62% low for random sequence generation, 71% low for allocation concealment). All included studies were at high risk of performance bias, because it is difficult, if not impossible, to blind the clinicians managing the infant’s care. Blinding of outcome assessment was rated separately for delivery room outcomes and outcomes assessed at a later stage. Although risk of bias was high across all studies for delivery room outcomes (because of the nature of the intervention), it was low for most studies (55%) for other outcomes. Most studies were at low risk of attrition bias. There were some concerns regarding selective outcome reporting bias. Evidence profile tables were collated for primary and key secondary outcomes applying the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework. These include details on risk of bias (Tables 2–5 in Supplemental Information).

We identified 23 studies including 3514 infants comparing DCC to ECC. Studies were undertaken in a range of countries, mostly high-income. Most studies included births before 32 to 34 weeks’ gestation and were conducted at a single center (78%), but the largest RCTs were multicenter (22%). The studies covered a variety of timings of cord clamping and positioning of the infant. Timing of DCC ranged between 30 and ≥120 seconds, with half the studies (52%) delayed by 30 to 45 seconds. Timing of early or immediate cord clamping ranged from within 5 seconds to within 30 seconds across studies; in most studies (69%), clamping was within 10 seconds.

Results for all primary and key outcomes are summarized in Table 6. Compared to ECC, DCC may improve neonatal survival (or reduce neonatal mortality) or may make no difference (survival: risk ratio [RR]: 1.02, 95% confidence interval [CI]: 1.00 to 1.04 (Fig 3); Number needed to benefit: 50, 95% CI: 25 to no benefit; 16 studies, 2988 infants; I² = 0%, certainty of evidence moderate). This translates into an RR of 0.80 (95% CI: 0.63 to 1.02) for the inverse outcome of mortality (post hoc analysis, Table 6 in Supplemental Information).

There was no clear difference in the number of infants with severe IVH (RR: 0.98, 95% CI: 0.67 to 1.42) and necrotizing enterocolitis (NEC) (RR: 0.83, 95% CI: 0.61 to 1.13). There was little to no difference for chronic lung disease (RR: 1.03, 95% CI: 0.94 to 1.13) and hyperbilirubinemia treated by phototherapy (RR: 0.99, 95% CI: 0.95 to 1.03).

DCC probably improves hematologic measures. Peak hemoglobin and hematocrit (%) were probably higher for DCC compared to ECC within 24 hours after birth (peak hemoglobin: mean difference [MD]: 1.24 g/dL, 95% CI: 0.01 to 2.47; peak hematocrit: MD: 2.63%, 95% CI: 1.85 to 3.42), and peak hematocrit was higher within 7 days after birth (MD: 2.70%, 95% CI: 1.88 to 3.52).

The evidence was unclear for survival without moderate or severe neurodevelopmental impairment in early childhood (RR: 0.96, 95% CI: 0.78 to 1.17). None of the included studies assessed other early childhood outcomes. Compared to ECC, DCC may make little or no difference to maternal complications, including any postpartum hemorrhage ≥500 mL (RR: 0.93, 95% CI: 0.54 to 1.62), severe postpartum hemorrhage ≥1000 mL, use of therapeutic uterotonic agents, blood transfusion, manual removal of the placenta, or postpartum infection (Table 6). No researchers reported on maternal deaths, severe morbidity, or additional treatment of postpartum hemorrhage. Authors of 1 study reported on mothers’ views and experiences.^23,24 

Other outcomes are detailed in Table 7a in Supplemental Information. Few differences were found except for hematologic outcomes. Compared with infants in the ECC group, infants in the DCC group had less inotropic support for hypotension during the first 24 hours of life (RR: 0.36, 95% CI: 0.17 to 0.75), a higher measurement of lowest mean arterial blood pressure in the first 12 hours of life (MD: 1.79 mm Hg, 95% CI: 0.53 to 3.05), lower incidence of any blood transfusion (RR: 0.83, 95% CI: 0.77 to 0.90), and a lower total number of blood transfusions per infant (MD: −0.63, 95% CI: −1.08 to −0.17) during hospital course.

We identified 13 studies comparing ICM to ECC (Table 3). Studies in comparison 2 included 1170 infants, and all were single center. Two studies (18%) included only preterm births <30 weeks. Timing of ECC ranged between clamping immediately and within 20 seconds of birth, and in most studies (69%), clamping was immediately. For ICM, the cord was milked between 2 and 4 times, with most studies (54%) reporting milking 3 times.

Compared to ECC, ICM may make no difference, slightly decrease, or slightly improve survival to discharge (RR: 1.02, 95% CI: 0.98 to 1.06; I² = 24%, 10 studies, 945 infants; certainty of evidence moderate) (Fig 4). This translates into an RR of 0.77 (95% CI: 0.49 to 1.23) for the inverse outcome of mortality (post hoc analysis, Table 6 in Supplemental Information).

We found no clear difference for severe IVH (RR: 0.72, 95% CI: 0.44 to 1.19), chronic lung disease (RR: 1.02, 95% CI: 0.63 to 1.65), and NEC (RR: 0.80, 95% CI: 0.55 to 1.18), and there was little or no difference for hyperbilirubinemia treated by phototherapy (RR: 1.04, 95% CI: 0.94 to 1.16).

ICM probably improves hematologic measures within 24 hours after birth. Peak hemoglobin and hematocrit (%) were higher for ICM compared to ECC within 24 hours after birth (peak hemoglobin: MD: 1.18 g/dL, 95% CI: 0.65 to 1.71; peak hematocrit: MD: 3.04%, 95% CI: 1.28 to 4.80). Evidence was uncertain for peak hematocrit and hemoglobin within 7 days after birth.

Limited data are available regarding outcomes in later infancy. Certainty of evidence was very low for moderate to severe neurodevelopmental impairment in early childhood (RR: 0.75, 95% CI: 0.21 to 2.71) and cerebral palsy in early childhood (RR: 2.65, 95% CI: 0.88 to 7.97). There were no researchers assessing sensory outcomes in later infancy.

The evidence is uncertain about maternal complications, including severe postpartum hemorrhage ≥1000 mL or blood transfusion, and there were no researchers assessing other maternal complications such as postpartum hemorrhage ≥500 mL (Table 7).

Other outcomes are detailed in Table 7b in Supplemental Information. In infants, few differences were found, with the exception of less inotropic support for hypotension during the first 24 hours of life (RR: 0.61, 0.44 to 0.84) and fewer infants receiving ≥1 blood transfusion (RR: 0.73, 95% CI: 0.56 to 0.94) in the ICM group.

We identified 1 single-center study of 60 infants evaluating CCM compared to ECC. The evidence was uncertain for the incidence of survival or its inverse mortality to hospital discharge, with no deaths in either group (Table 8). Evidence was also uncertain for severe IVH (RR: 0.33, 95% CI: 0.01 to 7.87), chronic lung disease (RR: 1.00, 95% CI: 0.07 to 15.26), and NEC (RR: 0.50, 95% CI: 0.05 to 5.22). CCM may increase peak hematocrit concentrations within 24 hours after birth (MD: 3.34%, 95% CI: 0.60 to 6.08). The authors of the study did not report other hematologic measures and did not assess any of the included early childhood or maternal outcomes. Other outcomes are detailed in Table 7c in Supplemental Information.

We identified 7 studies including 1073 infants comparing DCC to ICM. The studies were published between 2011 and 2019, and most were single center (71%). Timing of DCC ranged between 30 and 180 seconds, and most studies (71%) reported delay of 30 to 60 seconds. For ICM, the cord was milked between 3 and 4 times, with most studies (71%) reporting milking 4 times.

Compared to ICM, DCC may make no difference, slightly decrease, or slightly improve survival to discharge (RR: 0.99, 95% CI: 0.95 to 1.02; I² = 14%; 5 studies, 1000 infants, certainty of evidence moderate) (Fig 5, Table 9). This translates into an RR of 1.21 (95% CI: 0.76 to 1.94) for the inverse outcome of mortality (post hoc analysis, Table 6 in Supplemental Information).

There were no clear differences for key neonatal morbidities of severe IVH (RR: 0.60, 95% CI: 0.32 to 1.12), chronic lung disease (RR: 0.91, 95% CI: 0.67 to 1.25), NEC (RR: 1.57, 95% CI: 0.83 to 2.97), and hyperbilirubinemia treated phototherapy (RR: 1.05, 95% CI: 0.90 to 1.24).

There were also no clear differences between DCC and ICM for hematologic measures within 24 hours (peak hemoglobin concentrations [g/dL]: MD: −0.02, 95% CI: −0.56 to 0.53, peak hematocrit concentrations [%] MD: −0.18, 95% CI: −1.90 to 1.54). No study authors reported data on peak hemoglobin or peak hematocrit concentration within 7 days after birth.

Limited data were available regarding outcomes in later infancy. Certainty of evidence was low for moderate to severe neurodevelopmental impairment (RR: 0.22, 95% CI: 0.01 to 4.40), cerebral palsy in early childhood (RR: 0.36, 95% CI: 0.01 to 8.65), and significant developmental delay in early childhood (RR: 14.06, 95% CI: 0.83 to 237.84). Researchers of 1 study assessed legal blindness and reported no events, and no researchers assessed hearing deficits.

No researchers reported the included maternal outcomes. Other outcomes are detailed in Table 7d in Supplemental Information. Few differences were found between ICM and DCC.

No studies were identified for any of these comparisons (DCC versus CCM, ICM versus CCM, DCC <60 seconds versus DCC ≥60 seconds, time-based DCC versus physiologic DCC).

No patterns were identified in the subgroup analyses (Table 8 in Supplemental Information). The number of prespecified subgroup analyses was large, and P values were not adjusted for multiple comparisons. Researchers of only 2 studies reported data by subgroup, limiting the ability to perform subgroup analyses.

In this systematic review and meta-analysis, we identified 42 eligible studies with 5722 infants comparing cord management interventions. Compared to early clamping, delayed clamping may slightly improve infant survival but may make no difference (moderate quality evidence). We found moderate- to high-quality evidence that delayed clamping does not reduce or increase major neonatal morbidities, but it probably improves hematologic measures and may reduce the use of inotropes and blood transfusions in infants.

Compared to early clamping, intact milking may result in increased survival, slightly reduced survival, or make no difference. We found low to moderate quality evidence indicating no clear difference in major neonatal morbidities such as chronic lung disease, IVH, and NEC. Intact milking probably improves hematologic measures.

For the 1 study in which researchers compared ECC to CCM, the evidence was uncertain for infant survival and major morbidities. CCM may increase peak hematocrit within 24 hours after birth.

Compared to ICM, delayed clamping probably results in little to no difference in survival, major neonatal morbidities, and hematologic measures.

Across all comparisons, many of the infants could not be classified into the correct subgroup categories, and thus, meaningful subgroup differences are not possible to detect with the current data.

The latest comprehensive review in this area was a Cochrane review with searches conducted in November 2017.⁸  Authors of that review found a reduction in infant death for delayed compared to early clamping, a slight reduction in any IVH, but no reduction in severe IVH. There was insufficient evidence to derive conclusions for cord milking. With our review, we add new information, because we identified and included 11 additional recently published trials.^25–35 

Although previous reviews included preterm infants born at less than 37 weeks’ gestational age,^4,8  our review is limited to infants born at less than 34 weeks’. Although late preterm infants have increased risk for admission to neonatal intensive care and poor developmental outcome compared with term infants, they do not have the same serious morbidities experienced by less mature preterm infants.³⁶  Therefore, 18 studies included in the Cochrane review were excluded from the current review, leading to a slightly smaller total number of infants (188 less), despite the 11 additional trials.

Previous reviews included infant mortality as a primary outcome, whereas in this review, we assess the inverse of mortality, survival, because this is the standard ILCOR approach. This changes the relative effect measures, as shown in our post hoc sensitivity analysis comparing RRs for survival and mortality using the same data (Table 6 in Supplemental Information). The reason for this is that relative risk depends on the incidence of an event, which is higher for survival than mortality. Thus, the same absolute number of deaths can translate into different relative risk estimates for survival or mortality. For instance, in comparison 1, in the delayed clamping group, 1383 (93%) infants survived and 107 (7%) died. In the early clamping group, 1364 (91%) infants survived and 134 (9%) died. This equals a 2% absolute difference for both survival (93% to 91% = 2%) and mortality (9% to 7% = 2%). However, because survival was more common than mortality, the relative risk indicates a small 2% increase in survival (RR: 0.93/0.91 = 1.02) but a much larger 20% relative risk reduction for mortality (RR: 0.07/0.09 = 0.80).

For comparison 1 (early versus delayed clamping), the relative risk for mortality (indicating a 20% reduction) is similar to that reported in previous reviews (eg, 27% relative risk reduction in the Cochrane review).⁸  Although for previous reviews, this finding was statistically significant, in the current review, the CI touches the line of no effect. This may be due to different eligibility criteria for gestational age (as outlined above) or to the more-recent studies included in the current review. We did not find a difference in survival between ICM and delayed clamping (comparison 4). Point estimates for survival with intact milking compared to early clamping (comparison 2) are similar to point estimates for delayed compared to early clamping (comparison 1), but CIs are wider in comparison 2 because of fewer included studies. This suggests that intact milking may be comparable to delayed clamping for the outcome of survival, but more evidence is needed to confirm this.

In this review, we find improved hematologic measures and reduced use of inotropes for delayed clamping, and intact and cut milking compared to early clamping, in accordance with previous reviews.^4,8,13  This supports the proposed mechanism of placental transfusion (ie, increased net transfer of blood from the placenta to the infant) through delayed clamping or milking.^10,37  Our findings did not suggest a difference between delayed clamping and milking with respect to hematologic measures.

Although authors of previous reviews report differences in IVH rates for different cord management strategies,⁸  we did not find evidence for this in the current review. Animal models have been used to demonstrate that during umbilical cord milking, there was an increase in carotid blood flow and pressure.²⁶  In addition, a recent trial comparing delayed clamping to milking was stopped early in the subgroup of very preterm infants (<28 weeks’ gestation), because of a higher incidence of severe IVH in the milking group.²⁶  Thus, there may be different IVH risks related to cord management strategies depending on gestational age. Further evidence is required to resolve this question. In addition, not all studies in the current review were blinded for assessment of IVH, which is problematic because ultrasound diagnosis of IVH can be rater-dependent.³⁸  Consequently, we downgraded certainty of evidence for this outcome.

Few researchers reported developmental outcomes in early childhood, and the evidence was uncertain for all comparisons. One study published outcomes in early childhood for early clamping compared to delayed clamping (comparison 1) shortly after our search date and was therefore not included in the analysis.³⁹  Authors of this study found that delayed clamping may reduce the risk of death or adverse neurodevelopmental outcome at 2 years of age for children born <32 weeks, but confirmation in larger studies is needed.

Cord management at preterm birth is an active research field, evidenced by the number of additional studies included in this review compared to previous reviews. The searches for the latest Cochrane update were conducted in November 2017.⁸  In <2 years (search to July 2019), we identified 11 new studies. Still, more evidence is being generated; a search in February 2019 identified an additional 62 ongoing trials evaluating cord management strategies in preterm infants.⁴⁰ 

Ultimately, we want to answer the question: “which cord management strategy is the best and for whom?” With the current study, we take a step toward answering this question by looking at different comparisons analyzed in pairwise meta-analyses. Yet, there is insufficient evidence, when using aggregate data, to derive a definite answer, particularly when assessing differences for key infant subgroups. Once ongoing trials are completed, a network meta-analysis will be possible, which allows comparing and ranking of multiple interventions simultaneously.⁴¹  For assessing differential treatment effects across subgroups, the use of individual participant data can increase statistical power and reduce the risk of ecological bias.⁴²  The individual participant data on Cord Management at Preterm Birth (iCOMP) Collaboration is collating individual participant data from ongoing and completed trials to perform network meta-analysis and subgroup analyses to resolve remaining questions.⁴⁰  Investigators planning future trials in this area should follow a prospective meta-analysis framework in collaboration with the iCOMP Collaboration to target evidence gaps and avoid research waste.⁴³ 

Strengths of this review include its rigorous methods, including a prospectively registered protocol, a comprehensive search strategy, two reviewers independently completing each step of the review process, and the use of GRADE to determine certainty of evidence.⁴⁴  The author team constitutes a collaboration of world experts in systematic reviews, neonatology, and obstetrics, including the ILCOR taskforce, the Cochrane Neonatal and Pregnancy and Childbirth groups, and independent experts in cord management.

Yet, there are several limitations. For many reported comparisons and outcomes, certainty of evidence was low or very low, or no studies were available. This was mainly due to imprecision and, in some cases, due to inconsistency and risk of bias. For four of the prespecified comparisons, no studies were identified. In this review, only pairwise comparisons are presented; we did not conduct analyses comparing all available comparisons simultaneously (network meta-analysis). Our subgroup analyses were limited by authors of most studies not reporting outcomes separately by subgroup, highlighting the need for individual participant data to resolve these questions. Definitions for early and delayed clamping and milking varied across studies. Delayed clamping ranged from 30 seconds to >2 minutes, and early clamping ranged from within 5 seconds to within 30 seconds. Thus, in some instances, early and delayed clamping groups may have received similar interventions.

DCC at preterm birth may be beneficial compared to early clamping, and these benefits appear to be hemodynamic, but additional evidence is required to confirm this. There is some evidence that ICM may be similarly beneficial, but this needs further study. Additional evidence from ongoing trials and individual participant data network meta-analysis is required to determine which cord management strategies are the most advantageous and for whom.

We thank Carol Friesen (Cochrane Neonatal) for developing and conducting literature searches. We thank the following members of the ILCOR Scientific Advisory Committee and Neonatal Life Support Task Force for their input and assistance in developing the protocol and offering feedback on the review: Myra H. Wyckoff, MD, chair; Jonathan Wyllie MBChB, BSC, vice chair; Maria Fernanda de Almeida, MD; Jorge W. Fabres, MD; Joe Fawke, MD; Ruth Guinsburg, MD, PhD; Shigeharu Hosono, MD, PhD; Tetsuya Isayama, MD, MSc, PhD; Vishal S. Kapadia, MD, MSCS; Han-Suk Kim, MD, PhD; Helen G. Liley, MBChB; Chris JD McKinlay, MBChB, PhD; Lindsay Mildenhall, MBChB; Jeffrey M. Perlman, MBChB; Yacov Rabi, MD; Charles C. Roehr, MD, PhD; Edgardo Szyld, MD, MSc; Daniele Trevisanuto, MD; Sithembiso Velaphi, MBChB, FC Ped, PhD; and Gary Weiner, MD. We also thank Slavica Berber, PhD, Sol Libesman, BSc, Kylie Hunter, MSc, Angie Barba, MSc, Mason Aberoumand, MSc, and Hannah Ahern, MSc (National Health and Medical Research Council Clinical Trials Centre, University of Sydney) for assistance in the review process.

CCM

cut-cord milking

CI

confidence interval

DCC

delayed cord clamping

ECC

early cord clamping

GRADE

Grading of Recommendations Assessment, Development and Evaluation

ICM

intact-cord milking

ILCOR

International Liaison Committee on Resuscitation

IVH

intraventricular hemorrhage

MD

mean difference

NEC

necrotizing enterocolitis

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

RCT

randomized controlled trial

RR

risk ratio