Materials and Methods
After obtaining institutional review board approval, we implemented our QI process for MUC in September 2011 and prospectively monitored all eligible infants through August 2013. We compared the MUC group with a retrospective cohort of ELGANs who were delivered at the University of Utah between January 2010 and August 2011. Inborn singleton, twin, and triplet pregnancies that underwent delivery at <30 completed weeks of gestation (eligible up to 29 weeks 6 days), for whom full resuscitation was intended, were included. Data were not collected in either time period for deliveries in which resuscitation was not intended. There was no lower limit for inclusion based on gestational age. Mode of delivery (vaginal or cesarean) was not an exclusion criterion, nor were maternal complications such as placental abruption, placental implantation disorders (previa or accreta spectrum), and chorioamnionitis. Deliveries were excluded for the following fetal conditions: suspected major congenital anomaly or genetic disorder, fetal anemia, and hydrops fetalis.
The MUC procedure was adapted from that described by Hosono et al. Immediately after delivery, the neonate was held at or approximately 10 cm below the level of the placenta, and twisting or nuchal cord loops were released. The cord was pinched as close to the placenta as possible and milked towards the infant over a 2-3 second duration. The cord was then released, which allowed for a brief 2-3 second pause between each milking motion; this was repeated for a total of 3 times over a duration of <30 seconds. After completion, the cord was clamped per usual practice, and the neonate was handed to the resuscitation team. No resuscitation efforts were initiated during the milking process. Time of birth and Apgar timing were initiated when the infant was completely delivered, not when the cord was clamped.
Prospective data collection included maternal demographics, obstetric complications, antenatal steroid use, and features of labor and delivery. The number of times the cord was milked and the total duration of MUC was recorded by the delivering provider. Neonatal data variables included approach to resuscitation, Apgar scores, admission temperature, and time of initial intubation and surfactant therapy, as indicated. We also recorded the initial hematocrit value and any additional hematocrit values in the first 24 hours of life. Additionally, any supplemental volume infusions above calculated maintenance fluids and the use of inotropes or hydrocortisone within the first 72 hours of life were determined. We recorded the highest and lowest blood pressures (oscillometric or indwelling catheter) over 6-hour time epochs during the first 24 hours of life. We determined whether any red blood cell transfusions were provided, the age at the first transfusion, the total number of transfusions, and the total volume of blood transfused over the course of hospitalization. We recorded data on invasive and noninvasive respiratory support and a diagnosis of bronchopulmonary dysplasia at 36 weeks postmenstrual age. We monitored for other common neonatal morbidities that included NEC Bell’s stage ≥2, any operative intervention for retinopathy of prematurity, and the use of any medical or surgical therapy for patent ductus arteriosus. As standard of care in our unit, all infants born at <30 weeks of gestation receive at least 2 cranial ultrasound scans, typically from 5-10 of days age and approximately 28 days of age. IVH was graded based on the criteria developed by Papile et al. Posthemorrhagic ventriculomegaly, placement of a shunt, and evidence of periventricular leukomalacia were recorded.
Statistical analysis was performed with SPSS software (version 19; IBM, Armonk, NY). χ 2 or Fisher exact test was used for categoric comparisons. The Student t test was used for analysis of normally distributed continuous data. The Mann-Whitney U test was applied for ordinal data or continuous data that were not distributed normally. Given a baseline rate of 39% for the composite outcome in the historic cohort, with an α level of .05 and a β level of 0.80, and accepting a 2-sided probability value of < .05 by Fisher exact test as statistically significant for the primary outcome, we determined that at least 115 infants would be needed in the prospective group that underwent MUC to demonstrate a 30% reduction in the composite outcome (defined as severe IVH, NEC, and/or death before hospital discharge). Prespecified analyses also evaluated the individual outcomes of IVH, NEC, and death before hospital discharge. A higher risk subgroup of neonates who were born at <27 weeks’ gestation was also analyzed. Stepwise logistic regression analysis was performed to test independent risk factors for the composite outcome; variables that were significant at a probability value of < .05 by univariate analysis were selected for inclusion in the multivariate model. All analyses were performed by intention-to-treat.
Results
We identified 342 inborn preterm infants at <30 weeks of gestation who were admitted to our neonatal intensive care unit between Jan. 1, 2010, and Aug. 31, 2013; 166 infants before and 176 infants after institution of the MUC protocol. Among historic control infants, 6 were excluded for fetal anomalies or hydrops compared with 18 among infants who were born during the MUC period. Thus, 160 infants from the historic control group and 158 infants from the MUC group were included in our analysis. Among the 158 infants who were eligible for MUC, 139 infants underwent MUC: 1 infant was milked 2 times; 113 infants (82%) were milked 3 times; 21 infants (15%) were milked 4 times, and 3 infants were milked 5 times. No adverse events were reported with the MUC procedure, and all MUCs were completed in <30 seconds.
There were small but significant differences between groups in reported maternal race, clinical chorioamnionitis, placental abruption, and duration of ruptured membranes >24 hours before birth ( Table 1 ). Infants who were born during the MUC QI period were older and heavier than the historic control group ( Table 2 ). Despite the brief delay in resuscitation that is associated with MUC, there were no differences in admission temperature, Apgar scores, or extent of resuscitation between groups ( Table 2 ). There were no differences between the 2 groups for respiratory support needs, which included intubation rates, surfactant therapy, duration of ventilator support, or diagnosis of bronchopulmonary dysplasia at 36 weeks of gestation (data not shown). There were no differences in rates and severity of retinopathy of prematurity, patent ductus arteriosus that required treatment, ventriculomegaly with or without need for surgical management, or periventricular leukomalacia (data not shown).
| Variable | Historic control infants (n = 160) | Milking of umbilical card era (n = 158) | P value |
|---|---|---|---|
| Maternal age, y a | 27.7 ± 6.2 | 27.6 ± 6.6 | .939 |
| Maternal race, n (%) | .037 | ||
| White | 106 (66) | 97 (61) | |
| Black | 10 (6) | 4 (3) | |
| Hispanic | 25 (16) | 43 (27) | |
| Other | 19 (12) | 14 (9) | |
| Multiple gestation, n (%) | 49 (31) | 48 (30) | .962 |
| Any antenatal steroids, n (%) | 155 (97) | 156 (99) | .421 |
| Steroid doses, n b | 2 (2–2) | 2 (1–2) | .848 |
| Cesarean delivery, n (%) | 110 (69) | 103 (65) | .500 |
| Chorioamnionitis, n (%) | 48 (30) | 30 (19) | .022 |
| Preeclampsia, n (%) | 21 (13) | 22 (14) | .835 |
| Intrauterine growth restriction, n (%) | 17 (11) | 23 (15) | .290 |
| Abruption, n (%) | 36 (22) | 21 (14) | .032 |
| Preterm rupture of membranes, n (%) | 77 (48) | 69 (44) | .425 |
| Rupture of membranes at >24 hours, n (%) | 30 (19) | 45 (29) | .041 |
| Variable | Historic control group (n = 160) | Milking of umbilical cord era (n = 158) | P value |
|---|---|---|---|
| Gestation, wk a | 27.1 (23.0, 29.9) | 27.4 (23.1, 29.9) | .010 |
| Birthweight, g a | 880 (375, 2050) | 960 (410, 1910) | .009 |
| Male, n (%) | 8 (54) | 91 (57) | .490 |
| Apgar score, n a | |||
| 1 minute | 5 (2, 7) | 4 (2, 6) | .570 |
| 5 minutes a | 7 (6, 8) | 7 (6, 8) | .720 |
| Temperature on admission, °C a | 36.3 (35.8, 36.7) | 36.5 (36.0, 36.8) | .277 |
| Maximum resuscitation, n (%) | 20 (13) | 34 (22) | .127 |
| Blow by oxygen or continuous positive airway pressure, n (%) | 23 (14) | 13 (8) | |
| Positive pressure breaths, n (%) | 109 (68) | 104 (64) | |
| Intubation | |||
| Chest compressions, n (%) | 8 (5) | 10 (6) | |
| Age at intubation, min a | 5 (4, 10) | 6 (3, 11) | .764 |
| Age at surfactant, min a | 29 (20, 60) | 38 (23, 60) | .012 |
Table 3 shows outcome data for the historic control group compared with infants who were born during the MUC QI period. The hematocrit value at birth was significantly higher during the MUC QI period, and the proportion of infants who received packed red blood cell transfusions was significantly lower. These effects were significantly different by subgroup analysis for infants 27 weeks of gestation to 29 weeks 6 days of gestation, but only the hematocrit value remained significantly different among infants who were born at <27 weeks of gestation ( Table 3 ). Four infants (2.5%) in the historic control period had a hematocrit value of >60%, but none of the levels were >65%. Among the MUC QI infants, 10 infants (6.3%) had hematocrit values of >60%, with hematocrit value >65% (66.9%). None of the infants with hematocrit values of >60% manifested signs of polycythemia-hyperviscosity syndrome. Bilirubin levels were no different at 24 hours age (mean values: MUC group, 4.7 mg/dL, vs control group, 4.4 mg/dL; P = .60), although peak bilirubin level in the first week of life was slightly, but significantly, higher in the MUC era compared with the historic control group (9.1 ± 2.2 mg/dL vs 7.9 ± 2.4 mg/dL, respectively; P < .001). Nearly all of the infants received phototherapy in the first week: 94% of those in the MUC group and 95% of those in the historic control group. No infants underwent exchange transfusion.
| Variable | Control infants at <30 wks of gestation (n = 160) | Milking of umbilical cord at <30 wks of gestation (n = 158) | Control infants at <27 wks of gestation (n = 76) | Milking of umbilical cord at <27 wks of gestation (n = 63) |
|---|---|---|---|---|
| Hematocrit at birth, % a | 45.1 ± 7.4 | 49.9 ± 7.6 b | 42.2 ± 5.5 | 46.2 ± 6.8 b |
| Packed red blood cell transfusion, n (%) | 127 (79) | 90 (57) b | 73 (96) | 56 (89) |
| Dopamine first 72 hr, n (%) | 51 (32) | 28 (18) b | 42 (55) | 19 (30) b |
| Volume bolus first 24 hr, n (%) | 61 (38) | 37 (24) | 44 (58) | 23 (37) |
| Necrotizing enterocolitis, n (%) | ||||
| Any | 32 (20) | 18 (11) c | 21 (28) | 9 (14) |
| Surgical intervention | 15 (9) | 8 (5) | 11 (15) | 5 (8) |
| Intraventricular hemorrhage, n (%) | ||||
| Any | 61 (38) | 49 (31) | 45 (59) | 26 (41) c |
| Severe d | 27 (17) | 15 (10) c | 21 (28) | 11 (18) |
| Survived, n (%) | 135 (84) | 148 (94) b | 58 (76) | 57 (91) c |
| Death or severe intraventricular hemorrhage, n (%) | 40 (25) | 22 (14) c | 29 (38) | 15 (24) |
| Composite outcome, n (%) | 63 (39) | 34 (22) b | 44 (58) | 21 (33) b |
| Age at discharge, d a | 97 ± 35 | 93 ± 38 | 123 ± 30 | 124 ± 37 |
b P < .01 vs historic control group
c P < .05 vs historic control group
d Severe intraventricular hemorrhage, intraventricular hemorrhage grade 3 or 4, or periventricular leukomalacia.
Lower rates were found in the MUC QI period for early dopamine use and supplemental volume administration in the first 24 hours of life, although only less dopamine use reached significance ( Table 3 ). The lowered rates of dopamine and volume support were primarily attributable to infants who were born at <27 weeks of gestation. Despite less inotrope and volume use, mean blood pressures were consistently higher over the initial 24 hours of life in the MUC QI period ( Figure ).
Mortality and major morbidity rates are shown in Table 3 . Infants who were born during the MUC era had a significantly decreased risk of the composite outcome of severe IVH, NEC, and death before hospital discharge. The individual outcomes of severe IVH, NEC, and death before hospital discharge were also significantly different between groups. The greatest effect appeared to be in infants at <27 weeks of gestation ( Table 3 ).
Regression modeling demonstrated that the most important risk factor for the composite outcome was gestational age at birth ( Table 4 ). Other independent risk factors included a diagnosis of clinical chorioamnionitis and placental abruption. Regression analysis also demonstrated a significant association between the composite outcome and birth during the MUC era compared with the historic control group.
