Umbilical Cord Milking in Infants Born at <37 Weeks of Gestation: A Systematic Review and Meta-Analysis.

Umbilical cord milking (UCM) could be an alternative in cases where delayed umbilical cord clamping cannot be performed, therefore our objective was to evaluate the effects of UCM in newborns <37 weeks' gestation. In this systematic review and meta-analysis, we searched MEDLINE, EMBASE, CINAHL, the Cochrane Database of Clinical Trials, the clinicaltrails.gov database for randomized UCM clinical trials with no language restrictions, which we then compared with other strategies. The sample included 2083 preterm infants. The results of our meta-analysis suggest that UCM in premature infants can reduce the risk of transfusion (relative risk (RR)= 0.78 [95% confidence interval (CI),0.67-0.90]) and increase hemoglobin(pooled weighted mean difference (PWMD)= 0.89 g/L[95%CI 0.55-1.22]) and mean blood pressure (PWMD=1.92 mmHg [95% CI 0.55-3.25]). Conversely, UCM seems to increase the risk of respiratory distress syndrome (RR = 1.54 [95% CI 1.03-2.29]), compared to the control groups. In infants born at <33 weeks, UCM was associated with a reduced risk of transfusion (RR= 0.81 [95%CI 0.66-0.99]), as well as higher quantities of hemoglobin (PWMD= 0.91 g/L[95%CI 0.50-1.32]). UCM reduces the risk of transfusion in preterm infants, and increases initial hemoglobin, hematocrit, and mean blood pressure levels with respect to controls.


Introduction
Placental transfusion is the transfer of residual placental blood to the baby during birth and umbilical cord clamping. This transfer is part of the physiological transition from fetal to neonatal circulation [1].
The placenta can contain up to 40% of fetal blood volume [2]. In full-term infants, when delayed umbilical cord clamping (DCC) is performed, an additional 80-100 mL of blood is transferred and can contribute one third to one quarter of neonatal blood volume at birth [3]. In preterm infants, a randomized study of DCC versus immediate umbilical cord clamping (ICC) found an 18% increase in blood volume in the DCC group [4]. This is why the benefits and risks derived from the different ways of managing the umbilical cord in infants have been studied. With DCC, the observed effects include an increase in hemoglobin levels, reduced need for transfusion, an increase in iron deposits, and reduced rates of necrotizing enterocolitis [5][6][7].
Clinical practice guidelines (CPG) [8,9] and various scientific associations [10][11][12] recommend DCC in all births, whenever possible, due to its positive impact on neonatal health. Occasionally, it is not possible to perform DCC for varying reasons, such as immediate neonatal resuscitation or maternal hemodynamic instability. Umbilical cord milking (UCM) has been suggested as an alternative to DCC in these cases. This technique consists of milking the umbilical cord two to four times along a 10 cm or 20 cm length of cord, from the placenta toward the newborn [13,14].
In 2015, a meta-analysis was published on the use of UCM which included seven randomized controlled trials (RCTs) [13]. The control groups were made up of infants who had received DCC or ICC. One study included in this meta-analysis was on full-term infants. The authors concluded that UCM in preterm infants resulted in higher levels of hemoglobin and hematocrit than in other types of clamping, and also found a reduced risk of oxygen being needed and a reduced risk of intraventricular hemorrhage if UCM was performed. Furthermore, in 2018, another meta-analysis was published including only two RCTs of preterm infants comparing the practice of UCM with DCC. The authors concluded that UCM may reduce intraventricular hemorrhage compared to DCC [15].
Despite the demonstrated benefits of UCM for preterm infants, it is still not standard practice in delivery care and requires a new review due to the large number of RCTs published in recent years [16][17][18][19][20][21][22][23][24][25][26]. It would be especially interesting to determine the benefits and risks by gestational age and by type of clamping (DCC or ICC) in the most significant variables. Therefore, the main objective of this systematic review and meta-analysis was to evaluate the effects of UCM in infants born at less than 37 weeks' gestation. The secondary objective was to evaluate the effects of UCM stratified by gestational age (<33/≥33 weeks) and type of clamping (ICC/DCC).

Materials and Methods
This systematic review with meta-analysis was done in accordance with the preferred reporting items for systematic review and meta-analyses (PRISMA) declaration [27].

Data Sources and Searches
The search strategy was: (stripping OR milking OR squeezing) AND (umbilicus OR umbilical cord OR cord). A systematic search was performed of main database: Cochrane Library Plus, EMBASE, Scopus, PubMed, and ClinicalTrials.gov. The specific search strategy adapted to each database is provided in detail in Appendix A (Table A1).
The inclusion criteria were: (I) the type of study: RCT; (II) the population; including infants born at <37 weeks gestational age (GA).We made an initial decision to study three populations (<37, <33, and ≥33 weeks GA) separately because the effects and the results of interest would be different for these two groups; and (III) the type of procedure, where we included studies that compared UCM with a control procedure (ICC or DCC). The exclusion criteria were RCTs that included both preterm and full-term infants without the possibility of obtaining separate information for each group.
RCTs were selected with no time or language restrictions. Two reviewers (IOE and JRA) independently evaluated the articles obtained from a literature search done using titles and summaries, in an initial stage. They then evaluated the full texts that had been selected. Any dispute was resolved by reaching a consensus. If this was not possible, a third reviewer (AHM) evaluated the articles.
The main outcome of our study was neonatal mortality before discharge from hospital and the secondary results were adaptation at birth variables (cord arterial pH, Apgar score at 1 and 5 min) and hematological variables (first hematocrit and hemoglobin levels measured within the first 24 h after birth, the need for red blood cell transfusion before being discharged, peak serum bilirubin, and hyperbilirubinemia requiring phototherapy). We also included mean blood pressure within the first 6h after birth and short-term morbidities such as respiratory distress syndrome, hypotension in the first 24 h after birth requiring volume or inotropic support, intraventricular hemorrhage (any grade), need for oxygen at 28 days, necrotizing enterocolitis, sepsis, retinopathy of prematurity, patent ductus arteriosus, and duration of hospital stay.

Data Extraction and Quality Assessment
The three reviewers (IOE, AAA, and AHM) compiled the data and evaluated the quality independently. For the continuous outcomes, the averages and standard deviations (SD) were compiled whenever possible. When the averages and the SD were not available and originally appeared as the median and range or interquartile range, we attempted to contact the authors and ask for the results. When this was not possible, the results were converted to the mean and SD, using the methodology recommended in the Cochrane Handbook [28]. For the categorical outcomes, the counts of the study events were compiled.
The risk of bias in each study included was assessed using the criteria described in the Cochrane Handbook for Systematic Reviews of Interventions [29]. Seven domains were evaluated related with the risk of bias in each included study because there is evidence that these problems are associated with biased estimates of the treatment effect: (1) random sequence generation, (2) allocation concealment, (3) blinding of participants and personnel, (4) blinding of outcome assessment, (5) incomplete outcome data, (6) selective reporting, and (7) other bias. The opinions of the review authors were classified as "low risk", "high risk", or "uncertain risk" of bias.

Data Synthesis
For the categorical outcomes, relative risk (RR) was used together with confidence intervals of 95% (95% CI). Mantel-Haenszel fixed-effects models and Der Simonian-Laird random-effects models were used depending on the absence or presence of heterogeneity, respectively. The heterogeneity of the studies was estimated using I 2 tests and Cochran's Q. I 2 values of <25%, 25%-50%, and >50% normally correspond to small, moderate, and large amounts of heterogeneity, respectively [30,31].
For the quantitative outcomes, the pooled weighted mean difference (PWMD) was used with a 95% confidence interval (CI). The publication bias was also evaluated using Egger's asymmetry test and Funnel plots ( Figure A1) [32]. The statistical significance level was defined as 0.05.

Role of the Funding Source
The funders of the study had no role in study design, data collection, data analysis, data interpretation, writing of the report, or the decision to submit the paper for publication. All authors had full access to all data in the study and had final responsibility for the decision to submit for publication.

Study Selection
A total of 1579 studies were identified in the literature search. After eliminating duplicate articles, the 477 remaining documents were screened by title and summary. After applying the inclusion/exclusion criteria, 17 articles were selected for qualitative and quantitative analysis (meta-analysis; Figure 1)
The description of the UCM technique varied by study, including the number of times the cord was milked toward the baby (between two and four times) and the milking speed (between 5cm within 1s and 20 cm within 2s).
The number of infants in each study, the description of the UCM method, how the cord was managed in the control group, and the exclusion criteria are shown in Table 1.

Study and Data Quality
The risk of bias for the seven domains of each study is shown in Figure A2. Ten of the 17studies were assessed as being low risk for random sequence generation; in four studies, the risk was not clear, and three were assessed as being high risk because the details of the methods used for randomization were not described. All of the studies stated that the health professionals involved could not be blinded due to the nature of the intervention. Table A2.
In six studies [18,21,24,32,33,35], it was observed that the mean blood pressure of the UCM group was greater than that in the control group (PWMD =2.47 mmHg [95% CI 0.39-4.55]). We did not find significant differences in the analysis by type of control and GA ( Figure 2g, Table 2, and Table A4).

Respiratory Distress Syndrome
Four studies were used to assess respiratory distress syndrome [18,23,32,34]. We observed that the risk of respiratory distress syndrome was higher in the UCM group than in the control group (RR = 1.54 [95% CI 1.03-2.29]). However, when we studied 338 infants born at >33 weeks, we observed that the risk of respiratory distress syndrome was higher in the intervention group. No significant differences were found in the sub-analysis of the type of control ( Figure 2f, Table 2, and Table A3).

Other Variables
No inter-group differences were found for length of hospital stay, cord arterial pH, Apgar scores 1 min, Apgar scores 5 min, oxygen at birth, oxygen at 28 days, retinopathy of prematurity, using hypotensive expanders, using hypotensive drugs, necrotizing enterocolitis, patent ductus arteriosus, sepsis, and intraventricular hemorrhage (Tables 2, A3 and A4).

Publication Bias
Publication bias was observed for the hematocrit study (Egger's test for asymmetry; p =0.037), mean blood pressure (Egger's test for asymmetry; p = 0.007), and for length of hospital stay (Egger's test for asymmetry; p = 0.039; Table 2 and Figure A1).

Main Findings
The results of our meta-analysis suggest that UCM in preterm infants may reduce the risk of transfusion and increase hemoglobin and mean arterial pressure values. The only adverse effect of UCM appears to be that it increases the risk of respiratory distress syndrome compared to control groups.
With regard to meta-analysis by gestational age, in infants born with <33 weeks of GA, UCM was associated with a reduced risk of transfusion and with higher hemoglobin levels compared to the control group. In infants born with >33 weeks, higher hematocrit levels were observed in the intervention group versus the control group.
Moreover, upon conducting the meta-analysis according to the type of controls, the only statistical differences observed was the increase in hemoglobin levels in the UCM group versus ICC and DCC.

Interpretation
In our review we found an increase in hemoglobin levels, which reduces the risk of anemia, as well as the need for transfusion and the complications associated with this practice [38,39]. Specifically, the improvement in hematology is due to the placenta containing approximately 15-20 mL of blood per kilogram of body weight, regardless of birth weight [2], and when performing the UCM there is an increase in systemic blood volume and, therefore, in fetal hemoglobin. These improvements in hematological values were also observed in previous studies by Al-Wassia et al. and Dang et al. [13,40].
In our study, no significant differences were found regarding intraventricular hemorrhage between the intervention group and the controls, nor when performing the sub-analysis by GA or type of clamping. However, in 2019, one study [26] was discontinued due to concerns of increased severe IVH (Intraventricular Hemorrhage) in the overall cohort and a significant difference in the premature subgroup, <28 weeks. In our study, the cut-off point of gestational age was set at <33 weeks, and previous trials included in the meta-analysis were underpowered to find such effects. Therefore, the result of this variable must be taken with caution as new trials are required to confirm these data. Therefore, until this safety is verified with other studies, this practice should not be recommended in fetuses <28 weeks.
With regard to the adaptive capacity of the infant to life, no significant differences were observed in the Apgar scores at one or five minutes, cord arterial pH, or need for oxygen at birth between the UCM group and the control groups. An increased risk of respiratory distress was observed in infants with UCM, although only four studies were included for this meta-analysis.
When we carried out sub-analysis by GA, we found a reduced need for transfusion, and in increase in hemoglobin levels in preterm infants born at <33 weeks GA. However, these improvements were not observed in the >33 weeks GA group, which is attributable to the fact that only 2-3 RCTs were included, and also to the lack of information on several of the variables included. In the sub-analysis by type of control, where we compared UCM to ICC or DCC, we can only confirm an improvement in hemoglobin levels in the UCM group compared to the other two types of controls. In this regard, new trials are needed to confirm the benefits of UCM in infants at >33 weeks GA and, especially, to compare it with DCC.
Studies are currently being conducted on animals, that have raised major concerns about the safety of UCM in lambs at 126 ± 1 days gestation (equivalent to approximately 26 weeks gestational age in humans), with spikes in blood pressure and cerebral blood flow during each milking, which are detrimental to the newborn. [41]. In this respect, more studies covering a wider range of gestational ages would be necessary to raise these concerns and thus be able to compare the different studies.

Strengths and Limitations
The main strength of this systematic review is the inclusion of 11 new RCTs compared with the last review published. It also analyzes the effect of UCM in different sub-populations by gestational age and type of control, as well as assessing publication bias.
One of the limitations of our systematic review is that the inclusion and exclusion criteria were very variable, therefore there is a large amount of heterogeneity in the study populations. We also observed that the variables measured varied widely between each study. Another limitation is the lack of standardization of the practice of UCM in the different studies, although the method is described in detail in the majority of them. We observed a publication bias in the result "mean blood pressure", which could mean that the UCM group is overestimated, so the results should be interpreted with caution. On the other hand, we have not assessed the effects of UCM on severe intra-ventricular hemorrhage, and we do not have these data on long-term effects. The low number of studies that compare UCM with DCC means that is not possible to establish conclusive results.

Conclusions
The main conclusion of our systematic review is that UCM increases initial hemoglobin and mean blood pressure levels and reduces the risk of transfusion in preterm infants. There are no complications associated with this practice regarding the studied variables, except for an increased risk of respiratory distress syndrome. The UCM does not increase neonatal mortality compared to DCC and ICC procedures. In this sense, UCM could be considered as an alternative to DCC in situations where this cannot be carried out in fetuses>28 weeks, although the preferred technique should still be DCC.

High Risk Low Risk Low Risk Unclear Low Risk Low Risk Low Risk
Were randomized before delivery to one of two groups using random permuted blocks of 10; an independent statistician provided the randomization sequence

Serially numbered opaque envelopes contained arm bands
Neonatologists and pediatric support staff were not blinded to treatment assignment They were not alerted for study participation or treatment assignment and no notation of study participation