Premature infants are at high risk for anemia development because of their unique physiological conditions and iatrogenic causes. During the initial stay in the neonatal intensive care unit (NICU), >60% of very preterm infants and 90% of infants with extremely low birth weight receive blood transfusions (1). Blood transfusions can lead to specific complications, such as transfusion-associated necrotizing enterocolitis (TA-NEC) (2). Additionally, blood transfusions for preterm infants may increase the risks of retinopathy of prematurity (ROP), intraventricular hemorrhage (IVH), and bronchopulmonary dysplasia (BPD) (3).

TA-NEC refers to NEC episodes that are temporally related to packed red blood cell (RBC) transfusion and typically occur within 48 h after transfusion (4, 5). According to several studies (6, 7), 25%−35% of infants with NEC received blood transfusions prior to being diagnosed. Packed RBC transfusions, enteral feedings, and gastrointestinal immaturity are all risk factors for TA-NEC (8). Anemia can lead to intestinal ischemia and hypoxia, and blood transfusions can lead to intestinal tissue reperfusion and oxidative stress damage, thus increasing the risk of TA-NEC (9). Using near-infrared spectroscopy (NIRS) to monitor the blood oxygen saturation of local intestinal tissues, one study showed that the blood oxygen saturation of the mesentery in infants with TA-NEC decreased more significantly than that in infants without NEC after blood transfusion (10). Infants, particularly premature infants, have immature autoregulation of intestinal blood flow. If the intestine cannot obtain sufficient oxygen, then the local tissues may experience hypoxia and ischemia, leading to TA-NEC. Feeding during blood transfusions for preterm infants with anemia leads to decreased intestinal oxygenation, which may be related to TA-NEC (11, 12).

Withholding enteral feeding during blood transfusions for preterm infants may prevent TA-NEC. One systematic review of non-randomized controlled trials (non-RCTs) that included seven studies and 7,492 infants found that when feeding was withheld throughout the transfusion, the odds ratio (OR) of NEC development within 48–72 h was significantly low [OR, 0.47; 95% confidence interval (CI), 0.28–0.80] (13). A recent Cochrane review (14) (including one RCT and 22 preterm infants) found evidence that could not sufficiently suggest that delaying enteral feeding for infants who needed blood transfusions had an impact on NEC development. An RCT with a small sample size that divided preterm infants into three groups according to the feeding protocol found no significant differences in the mean splanchnic cerebral oxygenation ratio (SCOR), mean splanchnic fractional oxygen extraction (FOE), and TA-NEC when enteral feeds were withheld, continued, and restricted during blood transfusions; however, the feeding intolerance (FI) incidences were 0%, 25%, and 5% (P = 0.02), respectively (15).

Studies of feeding and NEC during blood transfusions for premature infants have primarily focused on preterm infants and a feeding amount ≥120 ml/kg. However, clinically, many premature infants receive blood transfusions before their feeding amount reaches 120 ml/kg. Moreover, studies have not limited the frequency of blood transfusions for the included infants. Multiple blood transfusions are associated with the severity of intestinal mucosal damage (16). One case study revealed that the administration of two full-volume transfusions within a relatively short timeframe was associated with pneumoperitoneum and TA-NEC (17). Therefore, this study aimed to examine the changes in splanchnic tissue oxygenation during the implementation of a protocol comprising the withholding of enteral feeding compared with those associated with routine enteral feeding during the peri-transfusion period for preterm infants who received partial enteral nutrition. Furthermore, this study analyzed the effects of blood transfusion and various feeding strategies on splanchnic tissue oxygenation and transfusion-related complications among very premature infants to provide clinical evidence for a management protocol for premature infants who require RBC transfusions.

Figure 1 presents the CONSORT recruiting diagram. Sixty infants met the inclusion criteria; therefore, 30 infants were assigned to each group. All parents or legal guardians of the included infants provided informed consent. Because of inadequate data, the study excluded two and four events from the feeding and withholding feeding groups, respectively.

Table 1 shows the demographic and clinical characteristics of each group at baseline. Differences in baseline characteristics of the infants in the withholding feeding and feeding groups were not significant. Similar prenatal characteristics were observed in both groups, and effective blood transfusions were performed for each patient.

In this study, transfusions were associated with an upward trend in splanchnic tissue oxygenation in both groups (Table 2). A comparison of values observed before and after transfusions showed that the feeding group a greater increase in the intestinal rSO₂ (T = 2.381; P = 0.031), whereas the withholding feeding group experienced a greater increase in the renal rSO₂ (T = 5.019; P < 0.001). This difference was significant.

No significant differences in the rSO₂ and FOE (Table 3, Figures 2, 3) were observed in the brain, intestines, liver, and SCOR of the groups at various times before and after transfusions. No interactions between the various time points before and after transfusions and feeding strategies were observed.

Significant changes in the renal rSO₂ and FOE (Table 3 and Figure 3) were observed between groups at 48 h after transfusion. An interaction between renal oxygenation and FOE with the feeding protocol techniques was noted. At 48 h after transfusion, the renal rSO₂ and FOE of the withholding feeding group were significantly different from those of the feeding group; however, no significant difference was observed at other time points.

Differences in clinical outcomes were not observed (Table 4). One infant in the withholding feeding group developed gastrointestinal symptoms 2 h after transfusion and was diagnosed with NEC. No cases of death occurred. No difference in the FI incidence between groups was observed (21.4 vs. 7.7%; P = 0.331).

No increase in IVH incidence was observed. During this study, the BPD incidence rate was 50% in both groups; however, the ROP incidence rates were 42.9% in the withholding feeding group and 57.7% in the feeding group, with no significant difference between groups. A subsequent investigation revealed no association between BPD and the total volume of the RBC transfusions (cc/kg) [R² = 0.228; P = 0.218 (>0.05)]; however, a strong positive correlation was observed between ROP and the number of transfusions (R² = 0.42; P = 0.021(<0.05)].

Recently, NIRS monitoring of rSO₂ has been widely applied to manage critically ill infants in the NICU. Human cerebral circulation has a certain degree of autonomous regulation, and the cerebral rSO₂ value is relatively stable. However, gastrointestinal rSO₂ is easily affected by the infant's position, intestinal peristalsis, and feeding factors, and high variability has been observed among different individuals (29). Therefore, this study not only directly compared rSO₂ but also measured intestinal and cerebral rSO₂ simultaneously. Subsequently, the SCOR was calculated to eliminate factors that influence intestinal rSO₂. Additionally, we recorded the SpO₂ of the patients and calculated the local tissue oxygen uptake fraction (FTOE) to reflect oxygen use in tissues. By these measurements, more accurate comparisons between groups and individuals within a group were possible.

A blood transfusion is a common therapy administered in the NICU. Blood transfusions for premature infants can lead to TA-NEC, which is closely related to IVH, BPD, and ROP (30–32). Additionally, the pathogenesis of TA-NEC may be closely related to intestinal mucosal hypoxic-ischemic reperfusion injury, inflammatory cascade reactions, and intestinal microbiota imbalance caused by anemia (8, 10, 16, 33). Withholding feeding decreased the risk of TA-NEC (13). When feeding is continued during blood transfusions, gut oxygenation in premature infants shows a declining trend, suggesting hypoxia-reperfusion injury and a probable association with TA-NEC development (34). This study found no significant differences between the peri-transfusion rates of cerebral oxygenation, intestinal oxygenation, liver oxygenation, FTOE, and SCOR. Additionally, no significant variations in gastrointestinal and preterm complications The findings of this study are consistent with those of the study by Schindler et al. (15), who randomly divided preterm infants who received blood transfusion into three groups (withholding enteral feeds for 12 h group, continuing enteral feeds, and restricting the enteral feed volume to 120 cc/kg/day group). Schindler et al. (15) compared the intestinal to-brain oxygen ratio (SCOR) and FTOE values of those three groups transfusion 1 h before transfusion, immediately after transfusion, and 12 and 24 h transfusion. No significant differences in the average SCOR and FTOE were observed among those three groups at any of the aforementioned times, and no difference in the TA-NEC incidence was observed (15). The gestational age, birth weight, postnatal age at the time of transfusion, and hemoglobin levels of the infants before transfusion observed during the present study are consistent with those observed by Schindler et al. (15). However, the average enteral feed volume of the feeding group was lower than that observed during this study (66.31 ± 47.50 ml/kg). This may have occurred because our hospital does not have a breast milk bank. Most infants in our NICU lived far from the hospital. During hospitalization, most infants received mixed feedings and a conservative protocol for adding milk. The feed withholding time during this study was shorter than that; however, no differences in the incidence of TA-NEC and other symptoms of the feeding and withholding enteral feeding groups were observed. The intestinal transit times of premature infants are four-times greater than those of full-term infants; therefore, a 4-h fast may not have the desired effect of lowering NEC (35). In contrast to the studies by Schindler et al. and others (15, 36), we adopted a protocol involving an 8-h feeding withholding period for the withholding feeding group, from the start of the transfusion. This study indicated that withholding enteral feeding for shorter periods (<12 h) during and after the transfusion may be safe.

According to previous studies of TA-NEC, most patients received multiple RBC transfusions (37). For example, Mohankumar et al. (16) developed a neonatal murine model of TA-NEC and demonstrated that multiple RBC transfusions increased the risk of bowel injury. Therefore, we limited our study to infants who received one or more blood transfusions before the study. Most studies did not specify the frequency of blood transfusions received by the enrolled infants. Therefore, this study is more in line with TA-NEC high-risk factors.

During this study, we compared the changes in splanchnic oxygenation in infants in the same group at different time points before and after transfusion. Splanchnic oxygenation in infants in both groups showed an increasing trend immediately after the transfusion. However, the increase in intestinal oxygenation in the feeding group was significant. In contrast, no significant difference was observed in the withholding enteral feeding group. The same feeding protocol group experienced no significant changes in intestinal oxygenation at 8, 24, and 48 h after the transfusion. These results indicated that feeding premature infants during blood transfusions may temporarily increase intestinal oxygenation. However, no significant differences in the SCOR and FOE were observed at different time points before and after transfusions within the same group, and no significant difference between the different feeding strategy groups was observed. Under normal circumstances, enteral feeding promotes the release of substance P by the central nervous system, thus leading to intestinal mucosal vasodilation, increased intestinal blood flow, and oxygen delivery (38). Braski et al. reported that enteral feeding led to a decreased SCOR for infants with anemia and a hematocrit level <28%, which may lead to decreased intestinal perfusion and increased risks of FI and NEC (32). Blood transfusions can also lead to decreased postprandial intestinal oxygen utilization efficiency and splanchnic tissue oxygen levels (34, 39, 40). The difference between the results of this study and those of the study by Balegar et al. may be related to different feeding strategies and the storage time of infused RBCs (39). Balegar et al. reported the use of a feeding protocol with a syringe pump at a constant speed (120 ml/h); however, we adopted a feeding protocol involving gravity. Furthermore, in the study by Balegar et al., the median age of the transfused RBCs was 9 days; however, in our study, RBCs were used within 7 days of blood collection. The storage time of banked blood may affect the ability of RBCs to release oxygen (41). The hematocrit (HCT) level before transfusion during our study was 25% which was lower than that during Braski et al.'s (26.3%), Marin et al.'s study (26%) and Balegar et al.'s study (Hb 9.3 g/dl). Therefore, the level of anemia could not have led to the difference between our research results and that of previous studies. Marin et al. (34) compared the intestinal oxygenation of preterm infants who were fed and those of preterm infants who were not fed during RBC transfusion (all had a gestational age <33 weeks); however, they did not provide information regarding cerebral oxygenation and oxygen saturation. In the study by Marin et al. (34), decreased postprandial intestinal oxygenation mainly occurred in preterm infants with a postmenstrual age <33 weeks, and especially in those with a postmenstrual age <30 weeks. Conversely, in the present study, most preterm infants had a postmenstrual age ≥34 weeks. However, the median postmenstrual age of the infants in our study was very similar with that of in the study of Braski et al. and Balegar et al. Therefore, it was not yet certain that infants with lower postmenstrual age have greater risk for altered gut rSO₂ during reperfusion phase following transfusion.

In this study, no significant changes were observed in the liver and brain oxygen levels before and after transfusion. However, the renal rSO₂ levels significantly increased and the renal FTOE levels decreased in the withholding feeding group, but no significant change in renal oxygenation was observed in the feeding group. No significant differences in the renal function, renal tissue oxygenation, baseline hemoglobin and hematocrit levels, caffeine use, or diuretic use were observed between the two groups before blood transfusion. However, in the feeding group, a significant increase in intestinal oxygen content was observed immediately after blood transfusion. The increased renal tissue oxygenation in the withholding feeding group may have been caused by the differences in feeding strategies, leading to different blood flow distribution ratios between organs after blood transfusion. We did not perform routine ultrasound examinations of renal hemodynamics during this study; therefore, we cannot rule out the possibility that differences may exist in renal hemodynamics before blood transfusion in patients with different feeding strategies, leading to differences in renal oxygenation after blood transfusion.

Currently, the effects of blood transfusions on the kidneys are unclear, and conflicting results of the effects of different transfusion thresholds on renal blood flow before and after transfusion have been reported (42–44). Limited research of the effects of renal oxygen before and after transfusion has been performed; therefore, further research is required.

Prior studies have shown that infants who were fed formula are at higher risk for TA-NEC (37), but that breastfeeding can reduce the risk of NEC (45). However, during this study, few infants were exclusively breastfed and fortified with breast milk; the majority received breast milk and formula. However, the sample size was small, and group comparisons were insufficient.

The primary strength of this study is that we controlled for corresponding confounding factors. This study implemented a blind method for NIRS measurements, thus eliminating subjective bias and allowing more reliable results. Thirty minutes before and after the corresponding time points of the measurement, nurses checked the position of the probe and did not perform nursing procedures to ensure that the child was quiet and did not experience any changes in body position, thus avoiding the effects of the probe or body position changes on the child. The feeding group received a standardized feeding method after transfusion that involved gravity feeding for 5–10 min.

This study had some limitations. First, we examined the effects of various feeding methods on splanchnic tissue oxygenation during the peri-transfusion period, but we did not monitor the splanchnic tissue oxygenation changes during transfusion and splanchnic hemodynamics. Second, we did not continuously monitor rSO₂. The NIR monitoring instrument used during our study was the EGOS-600B near-infrared blood oxygen monitoring instrument (Aiqin Medical Biomedical Electronics), which collected tissue rSO₂ data every 2 s. However previous studies have indicated that measuring rSO₂ within 10 min could provide accurate values when those values were stabilized (with fluctuations not exceeding 2%). Most of the research conducted with the use of this device adopted intermittent monitoring rather than continuous monitoring. Because of the lack of cameras and other software in our NICU, we were unable to ensure accurate recordings of infant movements and manage movement artifacts. Therefore, we chose several rSO₂ measurement time points. We used clear monitoring time points for each patient in this study to ensure the scientific validity of our research data. Additionally, because of the small sample size, we could not completely rule out the possibility of a second type of error that may have affected findings and reduced the importance of some differences. Infants with hemodynamic instability are at high risk for TA-NEC; therefore, were excluded from this study. Another reason for the lack of difference in the TA-NEC incidence between groups may have been the relatively high hemoglobin levels of the included infants. Patel et al. (46) reported that hemoglobin levels <8 g/dl significantly increased the incidence of TA-NEC. An in-depth study with a larger sample size is required.

The results of this study showed that splanchnic oxygenation in very preterm infants during the transfusion phase was unaffected by their feeding status. Feeding during RBC transfusion did not increase the incidences of intestinal ischemia after transfusion and TA-NEC in preterm infants compared with those of infants whose feedings were withheld during blood transfusions. Future studies should include multicenter trials with larger sample sizes to assess the long-term effects of different feeding strategies during RBC transfusions for preterm infants. Additionally, an investigation of the effects of the feeding volume, composition, RBC storage time, and hemodynamic monitoring on tissue oxygenation and transfusion-related complications would provide valuable insights regarding the optimization of clinical management protocols for this vulnerable population.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

The studies involving humans were approved by the medical Ethics Committee of West China Second University Hospital, Sichuan University. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation in this study was provided by the participants' legal guardians/next of kin.

JH: Writing – original draft, Conceptualization, Formal analysis, Investigation. XS: Supervision, Conceptualization, Writing – original draft. XX: Data curation, Software, Writing – original draft. HL: Formal analysis, Methodology, Writing – original draft. JT: Validation, Project administration, Writing – review & editing. TX: Resources, Visualization, Writing – review & editing. JZ: Supervision, Writing – review & editing. JS: Resources, Writing – review & editing.