A prospective, monocentric, unblinded, randomized clinical study was performed. This randomized, controlled study was conducted between January 2018 and December 2021 at the NICU, Dubai Hospital. The study was approved by the institutional review boards of the Dubai Scientific Research Ethics Committee, and the trial protocol and the statistical analysis plan are provided in Supplement 1. The primary outcome variable was the composite of death or severe IVH, and the secondary outcome variables were neonatal sepsis, NEC, feeding intolerance, and central line associated complications. Other outcome variables included: retinopathy of prematurity (ROP), bronchopulmonary dysplasia (BPD), and hemodynamically significant patent ductus arteriosus (Hs PDA). All preterm infants between 23- and 29.6 weeks gestation admitted to the NICU were candidates for inclusion in the study. The gestational age assessment by the last menstrual period and/or early pregnancy ultrasound was used to establish eligibility for the study. Exclusion criteria included infants with major congenital malformations, infants born outside hospital or referred from other hospitals, placental abruption, cord prolapse, and hydrops fetalis.
Recruitment and randomization procedures
The infants were screened for eligibility immediately after birth. If eligibility criteria were met, infants were randomized using SNOSE (sequentially numbered opaque sealed envelopes), created by independent support staff. Separate sets of randomization envelopes were used for neonates according to gestational age (23 weeks 0 days through 26 weeks 6 days or 27 weeks 0 days through 29 weeks 6 days). Two sets of cards labelled for randomization were enclosed in sequenced, opaque envelopes containing group assignment (UVC, or PICC) and kept in the NICU. After the birth of eligible preterm infant, the principal investigator or the on-call registrar opened the next randomization card, informed the staff of the group assignment, and inserted the central line according to the unit protocol. For the UVC group, the attending registrar inserted the UVC according to the formula: wt. x 3+9 divided by 2 then add 1, or by measuring the umbilical shoulder length [8,9]. For the PICC group, double lumen Vygon catheters were used, and the length of insertion was determined by a measuring tape from the site of insertion to the left or right second intercostal space. The position of central line insertion was verified by the use of X-ray or echocardiography and adjusted accordingly. UVCs found to be malpositioned below the diaphragm or twisted in the liver were removed and replaced by a PICC and the infant was assigned to PICC group. Similar to PICC malposition or insertion failure, the line was replaced by UVC, and the infant was assigned to the UVC group.
Demographic data, including gestational age (GA), weight, sex, antenatal steroids, mode of delivery, Apgar score, resuscitation, mode of ventilation support, hemodynamic stability, and surfactant administration were considered equally on assignment of infants to either group.
Blinding/Bias
Randomization was concealed to prevent allocation bias. Care-givers were not blinded to the infants’ assignments. In addition, ascertainment of the severity of IVH was performed by radiologists who were blinded to the study groups.
The primary outcome was the frequency of death and severity of IVH in the two studied groups between randomization and the study endpoint when the infant reached 28 days post menstrual age. IVH was graded according to the Papille and Volpe classification where grade 3 and grade 4 were considered as severe IVH. For each infant, the highest IVH grade out of the three brain scans was recorded and considered.
Secondary outcomes included any grade of IVH, central line-associated complications, feeding intolerance and/or necrotizing enterocolitis (NEC), and late-onset neonatal sepsis.
Definition of the Outcome Measures
-
Severe intraventricular hemorrhage was defined as grade III: intraventricular hemorrhage with ventricular dilation or grade IV: parenchymal hemorrhage in any one of the 3 cranial scans performed [10].
-
Necrotizing enterocolitis was defined according to Bell’s staging criteria considering only surgical NEC [11].
-
Feeding intolerance was defined as gastric residual volume >50% of the preceding feed, abdominal distension, or emesis and disruption of the infant’s feeding plan [12].
-
Late onset neonatal sepsis (LOS) was defined as sepsis occurring at or after 72 hours of life [13].
Cranial ultrasound:
A total of 669 cranial ultrasound scans were performed and recorded during the study period. The unit protocol is to scan all preterm infants < 32 weeks gestation for IVH within 48 hours of birth. A second scan occurred on Days 3-7, and a third scan occurred on Days 10-14 of life. Additional scans are performed for follow-up of IVH or post-hemorrhagic sequalae or periventricular leukomalacia. All ultrasound scans were performed by radiologists blinded to the study. Furthermore, one hundred and fifty scan electronic images belonging to a50 randomly selected infants from both groups were reviewed by an independent radiologist to evaluate interrater variability between the readers.
Sample Size Calculation
we calculated that a sample size of 258 infants is required to detect a difference of 5% in the incidence of severe IVH between the UVC group and the PICC groups, with power of 0.90 and p-value of 0.05. According to our local data, total admissions in the study period average of 1300 infants. Assuming a 10% dropout rate, we decided to enrol 280 infants. We chose 5% difference based on previous 2 trials adopted measures to reduce IVH by controlling fluctuations of cerebral blood flow and resulted in decrease incidence of severe IVH by 3% to 7% [14,15].
Statistical Analysis
Statistical analyses were performed using SPSS, version 25.0 (SPSS, Chicago, IL, USA). T tests were used for comparing continuous data. Chi-square or Fisher’s exact tests were used to compare categorical variables. Mann–Whitney U test was used for ordinal or non-normally distributed continuous variables. A P value <0.05 was considered as statistically significant. The impact of risk factors on the odds ratio of the observed IVH was estimated by logistic regression analysis using standard statistical techniques. Cohen's kappa coefficient (κ) was used to measure interrater reliability between the two cranial ultrasound readers, where 0.41–0.60 is a moderate agreement, 0.61–0.80 is substantial, and 0.81–1 is almost perfect agreement.