The germinal matrix is particularly prone to hemorrhage. This fragility is attributed, at least in part, to its extensive vascularity, which is naturally weak due to a lack of pericytes to support the arterial structural integrity [10].

In neonatal intensive care medicine, the transfusion of PRBCs is vital. In some cases, it can be lifesaving. Each transfusion, however, comes with its own set of risk and benefits. Some transfusion complications have been well identified, while others have been less so, and yet others may not be recognized as transfusion-related events at all, but rather as a clinical deterioration in the complex intensive care course [2].

The aim of this prospective observational study was to evaluate the impact of packed RBC transfusion on the incidence and severity of the germinal matrix hemorrhage/intra-ventricular hemorrhage. One hundred preterm neonates delivered between 27 and 35 weeks gestational age hospitalized at neonatal intensive care unit and presented with GM/IVH were examined over a period of 9 months.

Regarding demographic data and birth weight, we found that the mean gestational age of our patients was 30.91 + 2.23 weeks with the mean body weight 1473.20 + 436.15 g. Sixty-seven percent were females and 33% were males. This goes in line with the study conducted by Bordbar and Farjadnia, where 115 infants with GM/IVH and 120 infants without GM/IVH were evaluated. The neonates with and without GM/IVH had a mean age of 29.50 ± 2.9 weeks and 30.74 ± 1.9 weeks, respectively (P = 0.1). Furthermore, neonates with GM/IVH had a lower mean weight than those without: 1978 ± 644 g GM/IVH vs 2188 ± 864 g, respectively (P = 0.2) [11].

Regarding the peri-natal history, cesarean section was performed in the majority of our cases (71%) versus (29%) had a normal vaginal delivery); these findings are contradictory to the studies conducted by Riskin et al. and Humberg et al. who found that GM/IVH was more prevalent in the preterm neonates born by vaginal delivery, than those with planned CS [12, 13].

In our study, 54% of the mothers had no history of illness, 18% had a history of PIH, 10% had a history of PROM, 5% had a history of antepartum hemorrhage, 3% had a history of DM, and 10% had a history of other diseases. This is in accordance with the study results conducted by Bordbar and Farjadnia, and Linder et al. who found that the rate of GM/IVH is not related to maternal and perinatal factors such as preeclampsia, premature rupture of membranes, and chorioamnionitis [11, 14].

Regarding the clinical presentation, most of our cases (78%) were presenting with RDS, 11% of the cases were admitted by apnea, and two cases (2%) had sepsis. This was supported by Khanafer-Larocque et al.’s study which revealed that neonates in the severe GM/IVH group had more RDS and were more likely to be ventilated and to receive inhaled nitric oxide during the first 72 h of age [15].

Fifty-seven percent of our cases had a history of anemia, 1% was mild, 30% were moderate, and 26% were severe. In agreement, Rocha et al. reported a significant association between the presence of anemia on admission among their low-birth-weight neonates and high incidence of severe intra-periventricular hemorrhage [16].

On clinical examination, 75% of the cases had seizures, 39% suffered from apnea, and 71% had a bulging AF. The presence of seizures and pallor was reported to be significantly associated with GM/IVH according to a study by Egwu et al. [17].

Regarding the packed RBC transfusion, in our study, the percentage of the cases that needed packed RBC transfusion was 58% and was indicated for anemia in all these patients. Forty percent required one transfusion, 14% required twice, and 2% required either 3 or 4 transfusions. Forty-one percent of cases required transfusion during the 1st week, 12% during the 2nd week, 4% during the 3rd week, and 1% after 3rd week. The total amount of PRBC transfusion was 24.05 cc ± 7.946 cc, with a median of 20 cc. The higher percentage was noted in the first week and may be attributed to frequent blood sampling done early in life.

This is rather similar to the study done by Lee et al. who reported that infants with birth weight < 1000 g received 3.1 ± 2.2 transfusions, with a first transfusion at 22.1 ± 16.6 days. However, infants with birth weight 1000 to 1500 g received only 1.5 ± 0.8 transfusions, with a first transfusion at 32.6 ± 15.2 days. PRBC transfusions were administered significantly earlier (P = 0.001) and more frequently (P < 0.001) with smaller birth weights. Preterm infants in the NICU tend to receive more PRBC transfusions [18].

The hemoglobin level in our study group ranged from 3.00 to 20.50 g/dL with a mean hemoglobin of 10.83 + 3.58, while on discharge, it ranged from 9.90 to 17.60 with a mean of 13.08 + 1.77. For the hematocrit value on admission, it ranged from 10 to 61.20 g/dL with a mean hematocrit value of 33.08 + 11.77 while on discharge, it ranged from 30 to 55 g/dL with a mean 37.88 + 4.77. This was in accordance with the study conducted by Verhagen et al., who reported that most cases of severe GM/IVH occurred on the first day of life, corresponding with a sharp drop in hematocrit value between days 1 and 2. The low hematocrit value on the first day of life in infants with severe GM/IVH can be due to bleeding; however, a low hematocrit value can also accelerate cerebral blood flow and contribute to further bleeding. It remains uncertain whether low hematocrit value is the cause of hemorrhage or occurs secondary to it [19].

When we correlated the onset of GM/IVH with the date of the first PRBC transfusion received, 30% of the patients developed GM/IVH within 4 days of receiving their first PRBC transfusion while 13% developed GM/IVH within 5–20 days of receiving their first PRBC transfusion. Fifteen percent of the patients developed GM/IVH before PRBC transfusion. Forty-seven percent of the GM/IVH patients understudy were grade II while 36% were grade III and only 17% were grade I (Fig. 4).

Similarly, in the study by Sarkar et al., 10.5% of the very low birth weight infants surviving beyond 3 days of age had severe GH/IVH during the first 7 to 10 days of life [20].

Regarding the correlation of the GM/IVH grade to the patient’s outcome, 27% of our preterm neonates with GM/IVH died, while 53% survived and 20% survived but suffered from complications. The mortalities were high among grade II GM/IVH patients (52%) followed by grade III GM/IVH patients (30%), and lastly grade I GM/IVH patients (19%) with no statistically significant association between the grade of GM/IVH and the poor outcome of the patients.

This goes in line with the study conducted by Brouwer et al., who found that grade I GM/IVH had the most minor neuro-developmental complications [21].

Our results are also in concordance with Radic et al., who found that all preterm (≤ 30 completed weeks) patients included in the study by grades 2, 3, and 4 GM/IVH were significantly associated with an increased overall mortality, primarily in the neonatal period, and the risk increased with increasing grade of GM/IVH. Grade 4 GM/IVH was significantly associated with an increased risk of disability (P < 0.001) [22].

In our study, no statistically significant correlation was found between the grade of GM/IVH and any of the transfusion data among the study group except for a statistically significant correlation found between the GM/IVH grade and the age at first PRBC transfusion.

This was supported by the results of the study conducted by Christensen that reported an association between “early” RBC transfusions and the subsequent occurrence of intraventricular hemorrhage, although the underlying pathophysiological mechanism of this association was not clarified [2].

Similarly, Christensen et al. observed that eliminating or reducing RBC transfusions by drawing all baseline laboratory blood tests from fetal blood in the placenta, and drawing none from the neonate, resulted in less GM/IVH [2].

However, in our study, no similar statistically significant association was observed between the status of PRBC transfusion therapy and the grade of GM/IVH (P > 0.05).

Similarly, Lee et al. retrospectively investigated the relationship between PRBC transfusion and short-term outcomes including GM/IVH. Forty patients had GM/IVH. In 30 patients diagnosed with GM/IVH in the transfusion group, 14 patients had been diagnosed after PRBC transfusions, and 2 patients had received PRBC transfusions within 1 week of the preceding diagnosis [18].

In our study, a statistically significant positive association was found between increased grading of GM/IVH and increased total amounts of PRBC transfused among the study group (r = 0.269, P = 0.04). A similar finding was reported by Lien et al. and Baer et al. who reported that cases who received more RBCT were at higher risk for developing a severe GH/IVH and suggested other relevant variables associated with increased GM/IVH extension including transfusion amount of RBCs [6, 23].

In disagreement with our results, a more extensive multicenter study by Kirpalani et al. showed no association between GM/IVH grade and transfusion guidelines used [24].

When correlating the GM/IVH with the mean Hb and HCT values on admission, we did not find a significant difference among the different GM/IVH grades among the cases that received packed RBCs and those who did not receive PRBC transfusion (P > 0.05). This is in agreement with Baer et al., who compared the blood hemoglobin (Hb) concentrations during the first 72 h that “triggered” the order for the first PRBC transfusion in each of the patients and controls. The controls had their first RBC transfusion after a Hb level of 11.1 ± 1.3 g/dL, while those who developed severe GM/IVH received their first RBC transfusion after a Hb level of 10.7 ± 1.6 g/dL (P = 0.22) with these results being very close to our results [6].

Also, in harmony with our study findings, Valieva et al. reported that there was no association between both HCT level at birth and HCT level immediately before the first transfusion and the severity of GM/IVH [25].

In our study, no statistically significant association was found between GM/IVH grade and gestational age among the study group (P > 0.05). In contrast, Swai et al. found that babies delivered before 28 weeks of pregnancy are more likely to suffer GM/IVH since the germinal matrix is entirely unsupported at this time. Furthermore, spontaneous involution of the germinal matrix usually occurs after 28 weeks of pregnancy, when the risk of hemorrhage is significantly reduced [26]. This controversy could be explained that the mean gestational age for our patients is 30.91 weeks ± 2.23, while in the other studies, the gestational age of the patients was less than 28 weeks.

In our study, an increased grade of GM/IVH was associated with statistically significant higher birth weight among the study group (P = 0.034). Egwu et al. did not find any association between the birth weight of neonates and grade of GM/IVH. This observation may be related to the few number of neonates with very low birth weights in both our and their study [17].

In our study, at 1, 5, and 10 min, lower Apgar scores were significantly associated with increased grades of GM/IVH (P = 0.004, P = 0.033, and P = 0.009, respectively).

This is in accordance with the study of Koksal et al. who reported that the prevalence of GM/IVH in very low birth weight neonates was highly associated with a low 5 min Apgar score at birth. Infants with lower Apgar scores were more likely to be subjected to procedures such as positive pressure breathing, endotracheal intubation, and chest compression, which could have resulted in a wide range of cerebral pressure fluctuations, increasing their risk of GM/IVH [27].

In our study, among GM/IVH grade 1 cases, the mean Apgar score at 5 min was lowest among cases who did not receive PRBC therapy (P = 0.03) while among GM/IVH grade III cases, the mean Apgar score at 1 min was significantly lower among cases who received PRBC transfusion (P = 0.01).

In our study, there were no statistically significant associations found between the transfusion data and any of the maternal data among the study group (P > 0.05). Similarly, Shanmugha Priya reported that maternal medical complications, obstetric complications, mode of delivery, multiple births, and antenatal steroid administration in mothers were not found to influence the need for transfusions statistically in their babies [28].

In our study, GM/IVH patients who received PRBC transfusion had significantly lower mean gestational age (P < 0.001), lower mean birth weight (P = 0.01), and significantly longer duration of NICU stay associated with more sampling or invasive interventions necessitating PRBC transfusion than those who never received a PRBC transfusion (P = 0.001). Similarly, Guzman et al. reported that statistically significant factors associated with intracranial hemorrhage were gestational age less than 32 weeks and blood transfusion [29].

In contrast, Shanmugha Priya reported that number of blood transfusions received by the preterm neonates did not make any significant difference in the length of hospital stay [28].

Limitations of our study included the absence of follow-up cranial ultrasound done at defined intervals for proper documentation of GM/IVH extension. Also, the small sample size was not enough for an accurate assessment of the problem among the targeted patient group.