Enteral Granulocyte-Colony Stimulating Factor and Erythropoietin Early in Life Improves Feeding Tolerance in Preterm Infants: A Randomized Controlled Trial Mona M. El-Ganzoury, MD1, Hesham A. Awad, MD1, Rania A. El-Farrash, MD1, Tarek M. El-Gammasy, MD1, Eman A. Ismail, MD2, Hanan E. Mohamed, MD3, and Sarah M. Suliman, MB BCh1 Objective To evaluate the efficacy and safety of enteral recombinant human granulocyte colony-stimulating factor (rhG-CSF) and recombinant human erythropoietin (rhEPO) in preventing feeding intolerance.
Study design An interventional randomized control trial was conducted in 90 preterm infants born at #33 weeks gestational age. The neonates were assigned to 4 groups; 20 received rhG-CSF, 20 received rhEPO, 20 received both, and 30 received distilled water (placebo control). The test solution was given at the beginning of enteral feeding and was discontinued when enteral intake reached 100 mL/kg/day or after a maximum of 7 days, whichever came first. Feeding tolerance and adverse effects of treatment were assessed. Serum granulocyte colonystimulating factor and erythropoietin levels were measured on days 0 and 7 of treatment. Results All neonates tolerated the treatment without side effects. Neonates who received rhG-CSF and/or rhEPO had better feeding tolerance, as reflected by earlier achievement of 75 mL/kg/day, 100 mL/kg/day, and full enteral feeding of 150 mL/kg/day with earlier weight gain and a shorter hospital stay (P < .05). The risk of necrotizing enterocolitis was reduced from 10% to 0% in all treatment groups (P < .05). There was a shorter duration of withholding of feeding secondary to feeding intolerance among neonates receiving both rhG-CSF and rhEPO compared with those receiving placebo (P < .05). Serum levels of granulocyte colony-stimulating factor and erythropoietin at 0 and 7 days did not differ across the treatment groups. Conclusions Enteral administration of rhG-CSF and/or rhEPO improves feeding outcome and decreases the risk of necrotizing enterocolitis in preterm neonates. The mechanism may involve the prevention of villous atrophy. (J Pediatr 2014;-:---). See editorial, p
T
he human fetus normally swallows more than 200 mL of amniotic fluid per kilogram body weight each day at term. Amniotic fluid contains multiple intestinal growth factors that have cognate receptors on the luminal surface of developing villous enterocytes and are critical to small bowel villous development.1 These growth factors, including granulocyte colony-stimulating factor (G-CSF) and erythropoietin (EPO), have important nonhematopoietic functions, including trophic actions on villous height and bowel length of the developing intestine.2 With preterm birth, the ingestion of amniotic fluid containing G-CSF, EPO, and other enterocyte trophic factors ceases abruptly, likely predisposing the gut to villous atrophy, feeding intolerance, and necrotizing enterocolitis (NEC) once feedings are instituted.3 Human milk feeding is the only currently accepted modality for NEC prevention. This finding may be related to the presence of growth factors in human milk.4 G-CSF and EPO have the practical advantage of being available as sterile human recombinant factors.5 The ingested G-CSF and EPO are highly protected from digestion in the neonatal intestine and remain biologically active. They bind to specific receptors expressed on the surface of fetal villous enterocytes.6 The effect of enteral intake of these factors compared with each other in preventing NEC in preterm infants has not been widely explored, however. The present study aimed to evaluate the efficacy and safety of enteral recombinant human G-CSF (rhG-CSF) and recombinant human EPO (rhEPO) in preventing feeding intolerance and/or NEC in preterm infants. ARR EPO G-CSF NEC rhEPO rhG-CSF TPN VLBW WBC
Absolute risk reduction Erythropoietin Granulocyte colony-stimulating factor Necrotizing enterocolitis Recombinant human erythropoietin Recombinant human granulocyte colony-stimulating factor Total parenteral nutrition Very low birth weight White blood cell
From the Departments of 1Pediatrics, 2Clinical Pathology, and 3Radiodiagnosis, Faculty of Medicine, Ain Shams University, Cairo, Egypt The authors declare no conflicts of interest. Registered with ClinicalTrials.gov: NCT01441427. 0022-3476/$ - see front matter. Copyright ª 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2014.07.034
1
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Methods This double-blind, randomized trial was conducted in 90 preterm infants of gestational ages #33 weeks in the neonatal intensive care units of Ain-Shams University hospitals between March 2013 and March 2014. Exclusion criteria were congenital or acquired anomalies of the gastrointestinal tract and previous use of cytokine or intravenous immunoglobulin. The mothers of the newborns provided written consent for the protocol, which was approved by the Ethical Committee of Ain-Shams University hospitals. Reporting of the study conforms to Consolidated Standards of Reporting Trials 2010 statement.7 Of 120 neonates assessed for eligibility, 30 were excluded, including 3 who did not meet the inclusion criteria, 12 whose guardian declined participation, 13 who died before receiving the study drug or placebo, and 2 because of insufficient blood samples. The 90 included newborns were randomly assigned into 4 groups; 20 received enteral rhG-CSF, 20 received enteral rhEPO, 20 received both enteral rhG-CSF and rhEPO, and 30 received distilled water as a placebo (Figure 1; available at www.jpeds.com). Drug administration followed a predetermined schedule generated from random numbers in a 1:1 manner based on a computer-generated randomization sequence maintained within the investigational drug pharmacy, with allocation concealment in opaque sequentially numbered sealed envelopes. All of the newborns had a detailed perinatal and family history obtained, underwent a thorough clinical examination, and received standard neonatal care. Laboratory investigations included a complete blood count and a high-sensitivity Creactive protein measurement. Serum levels of G-CSF and EPO were measured on days 0 and 7. Serum G-CSF concentrations were determined using the RayBio Human G-CSF ELISA Kit (RayBiotech, Norcross, Georgia), and EPO concentrations were measured with the DRG EPO ELISA Kit (DRG International, Springfield, New Jersey). The newborns also underwent baseline evaluations in the first 72 hours of life that included abdominal and cranial ultrasounds as well as echocardiographic assessment. Abdominal ultrasound examination was done for evaluation of abnormally large bowel loops and intestinal wall edema with Doppler examination of superior mesenteric artery, as described previously.8 On the day that the attending neonatologist chose to begin feedings, the infants were assigned to receive enteral G-CSF and/or EPO or water placebo. A single daily dose of enteral rhG-CSF, 4.5 mg/kg (Geneleukin 300 mg/mL; Shandong Geneleuk Biopharmaceutical, Jinan, China), and/or enteral rhEPO, 88 IU/kg (Epoetin 2000 IU/mL; SEDICO Pharmaceuticals, 6th of October City, Egypt) was given with the start of enteral feeding. This dose was calculated to provide the same amount of growth factors ingested by a fetus swallowing 200 mL/kg/day of amniotic fluid, according to Christensen et al.1 The daily dose for each patient was diluted in sterile distilled water in accordance with the manufacturer’s instructions and as described by Canpolat et al,9 kept in a separate 2
Vol. -, No. opaque aliquot at 2-8 C, and allowed to warm to room temperature before administration through the orogastric/nasogastric tube with milk feedings for 7 days. The control group was given 1 mL of distilled water (placebo) once daily. Feeding Protocol The unit’s feeding policy was to commence early trophic feeding starting with 10-20 mL/kg/day, preferably breast milk if available, and progressing by 10-20 mL/kg/day for as long as tolerated (as judged by the attending neonatologist). This feeding policy was not altered during the study period. Feeding intolerance was defined as the presence of any of the following that interfered with the written enteral feeding plan: increased abdominal girth, emesis, gastric residual of $25% of the previous feed volume, abdominal distension, or the presence of macroscopic blood in stools. No specific value for increase in abdominal girth or volume of emesis was required.10 Withholding feeding was indicated in cases of feeding intolerance with a gastric residual $25% of the previous feed volume, suspected or proven NEC, heavily bile-stained or large gastric residuals or vomiting, clear abdominal pathology, significant abdominal distension or discoloration, blood in stool, or an unstable condition causing clinical concern, including significant cardiorespiratory instability. Follow-Up Strategy and Study Endpoint All infants in the treatment and control groups were followed up clinically with at least 2 complete physical examinations (1 in the morning and the other in the evening) and radiologically by abdominal radiography every other day. The following outcome data were recorded: day of successful start of enteral feeding; times to establish one-half, two-thirds, and full enteral feedings after drug/placebo administration (at least 150 mL/kg/day); time to the end of total parenteral nutrition (TPN); weight gain, expressed as g/kg/day; the incidence of NEC of stage I or worse according to Bell modified criteria11; surgery for NEC; NEC-related death; length of hospital stay; hospital readmission; and adverse effects of treatment (if any), including emesis, increased gastric residual, increased abdominal girth, diarrhea, skin rash, or blood pressure instability. The test solution (or placebo treatment) was discontinued when the patient’s enteral intake reached 100 mL/kg or after a maximum of 7 days, whichever came first. Statistical Analyses The sample size was calculated using power and sample size calculation program EpiInfo version 6.0 (Centers of Disease Control and Prevention, Atlanta, Georgia). Our unit statistics had a consistently high prevalence of milk intolerance in preterm infants of very low birth weight (VLBW) during the 4 consecutive years before the study. The mean total time to attain full enteral feeding in our unit was 35 5 days. Accordingly, a sample size of at least 20 neonates in each group was sufficient to detect a 30% difference in the time needed to achieve full feedings (a reduction from 35 days to 25 days) with a = 0.05 and 80% power of the study. El-Ganzoury et al
- 2014
ORIGINAL ARTICLES
Data analysis was done using SPSS version 15 (SPSS Inc, Chicago, Illinois). The Student t test was used to compare quantitative parametric variables between 2 groups, and ANOVA with a post hoc Bonferroni test was used for comparisons of more than 2 groups. The c2 test was used to test the association variables for categorical data. The Spearman rank correlation coefficient was used to measure the strength and direction of the linear relationship between 2 variables. A P value .05) (Figure 2). Moreover, no significant correlations were found between hemoglobin level and serum EPO level
(r = 0.07, P > .05) or between WBC count and serum G-CSF level (r = 0.2; P > .05) measured on the seventh day of enteral intake of growth factors (P > .05). Assessment of the relationships among various clinical variables in the whole treatment group revealed a significant negative correlation between gestational age and length of hospital stay (r = 0.368; P = .004). Moreover, there were significant negative linear correlations between birth weight and the time needed to achieve one-half (r = 0.414; P = .001), two-thirds (r = 0.386; P = .002), and full enteral feeding (r = 0.487; P < .001). Birth weight was also negatively correlated with weight gain (r = 0.323; P = .12) and length of hospital stay (r = 0.722; P < .001). The duration of hospital stay was positively correlated with the time needed to achieve one-half (r = 0.378; P = .003), two-thirds (r = 0.355; P = .005), and full enteral feeding (r = 0.38; P = .003). The time needed to discontinue TPN also was positively correlated with the achievement of one-half (r = 0.601; P = .001), two-thirds (r = 0.89; P < .001), and full enteral feeding (r = 0.867; P < .001) (data not shown).
Discussion Efforts to reduce mucosal atrophy in preterm neonates and attempts to prevent feeding intolerance among VLBW neonates have met with varying success. Provision of minimal enteral nutrition has been used in VLBW neonates as a means of avoiding disuse atrophy of the small bowel mucosa. Using an experimental model, Owens et al13 observed that minimal enteral nutrition did indeed result in maturation of intestinal motor function, but had little beneficial effect on the small bowel mucosa. Erythromycin administration has been tested but deemed generally not useful in reducing feeding intolerance in this population.14 The main objective in the present study was to improve feeding tolerance in preterm neonates by enterally administering intestinal growth factors, along with other benefits, such as decreasing nothing-by-mouth days and reducing the need for TPN to avoid its complications and cost. We postulated that VLBW infants could be provided with physiological quantities of rhG-CSF and/or rhEPO via intermittent orogastric/nasogastric administration in concentrations comparable with what they would have ingested from amniotic fluid had they remained in utero. Our results showed that the test solutions were well tolerated for up to 7 days of administration without related side effects. Barney et al15 used a sterile isotonic solution patterned after human amniotic fluid to treat feeding intolerance in infants and reported that none of the study infants had a recognized adverse reaction to the test solution. Juul et al16 also safely used enteral rhEPO in neonatal rats. We observed a significant improvement in feeding tolerance among all treatment groups, as reflected by the reduced time to achieve one-half, two-thirds, and full enteral feeding. Furthermore, neonates in the treatment groups had earlier weight gain, faster hospital discharge, and shorter duration of withholding of feeding. An important finding was the
Enteral Granulocyte-Colony Stimulating Factor and Erythropoietin Early in Life Improves Feeding Tolerance in Preterm Infants: A Randomized Controlled Trial
3
THE JOURNAL OF PEDIATRICS
Vol. -, No. -
www.jpeds.com
Table. Clinical and laboratory characteristics of the treatment groups and placebo control group Variables Male sex, n (%) Gestational age, wk, mean SD Birth weight, kg, mean SD Type of milk, n (%) Human milk Formula milk Mixed Age at the start of enteral feeding, d, mean SD Time to achieve one-half enteral feeding, d, mean SD Time to achieve two-thirds enteral feeding, d, mean SD Time to achieve full enteral feeding, d, mean SD Withholding feeding, d, mean SD Day of weight gain, d, mean SD Day to discontinue TPN, d, mean SD Hospital stay, d, mean SD Duration of treatment/placebo, d mean SD Gastric residuals >25%, n (%) NEC, n (%) Gaseous distention (radiography), n (%) Intestinal wall edema (ultrasound), n (%) Deaths, n (%) WBC count, 109/L, mean SD Hemoglobin, g/dL, mean SD Platelet count, 109/L, mean SD
Placebo group G-CSF group (n = 30) (n = 20)
EPO group (n = 20)
P value, treatment groups vs control group G-CSF + EPO group (n = 20) G-CSF group EPO group G-CSF + EPO group
10 (33.3) 30.5 1.5 1.36 0.29
8 (40) 30.8 1.8 1.26 0.27
5 (25) 30.2 1.8 1.31 0.31
8 (40) 30.4 1.9 1.19 0.28
.65 .46 .36 .23
.65 .56 .91 .23
.65 .91 .06 .23
0 (0) 14 (47) 16 (53) 1.7 1.4
1 (5) 9 (45) 10 (50) 1.8 1.4
0 (0) 5 (25) 15 (75) 2.3 1.75
0 (0) 5 (25) 15 (75) 2.5 3.3
.92
.16
.55
8.2 4.2
6.2 4.4
7.2 4.1
5.9 2.4
.027*
.04*
.013*
11.7 4.4
9.7 3.0
10.5 4.3
9.5 4.9
.11
.23
.033*
16.3 5.3
12.6 5.4
13.4 4.9
12.4 3.1
.005*
.032*
.006*
8.4 7.3 14.9 6.5 15 5.43 57.9 10.8 6.5 2.7 11 (36.7) 3 (10) 11 (36.7) 6 (20) 3 (10) 15.5 7.3 15.4 2.9 215 94
3.1 1.3 5 2.1 10.3 3.7 10.8 4.0 13.5 6.0 13.75 5.0 44.6 11.9 43.5 11.1 6.5 1.3 6.4 1.4 4 (20) 5 (25) 0 (0) 0 (0) 3 (15) 3 (15) 2 (10) 3 (15) 2 (10) 2 (10) 13.7 5.3 17.8 6.6 16.8 4.3 17.7 5.5 220 96 260 133
.05 .002* .17
Study design An interventional randomized control trial was conducted in 90 preterm infants born at #33 weeks gestational age. The neonates were assigned to 4 groups; 20 received rhG-CSF, 20 received rhEPO, 20 received both, and 30 received distilled water (placebo control). The test solution was given at the beginning of enteral feeding and was discontinued when enteral intake reached 100 mL/kg/day or after a maximum of 7 days, whichever came first. Feeding tolerance and adverse effects of treatment were assessed. Serum granulocyte colonystimulating factor and erythropoietin levels were measured on days 0 and 7 of treatment. Results All neonates tolerated the treatment without side effects. Neonates who received rhG-CSF and/or rhEPO had better feeding tolerance, as reflected by earlier achievement of 75 mL/kg/day, 100 mL/kg/day, and full enteral feeding of 150 mL/kg/day with earlier weight gain and a shorter hospital stay (P < .05). The risk of necrotizing enterocolitis was reduced from 10% to 0% in all treatment groups (P < .05). There was a shorter duration of withholding of feeding secondary to feeding intolerance among neonates receiving both rhG-CSF and rhEPO compared with those receiving placebo (P < .05). Serum levels of granulocyte colony-stimulating factor and erythropoietin at 0 and 7 days did not differ across the treatment groups. Conclusions Enteral administration of rhG-CSF and/or rhEPO improves feeding outcome and decreases the risk of necrotizing enterocolitis in preterm neonates. The mechanism may involve the prevention of villous atrophy. (J Pediatr 2014;-:---). See editorial, p
T
he human fetus normally swallows more than 200 mL of amniotic fluid per kilogram body weight each day at term. Amniotic fluid contains multiple intestinal growth factors that have cognate receptors on the luminal surface of developing villous enterocytes and are critical to small bowel villous development.1 These growth factors, including granulocyte colony-stimulating factor (G-CSF) and erythropoietin (EPO), have important nonhematopoietic functions, including trophic actions on villous height and bowel length of the developing intestine.2 With preterm birth, the ingestion of amniotic fluid containing G-CSF, EPO, and other enterocyte trophic factors ceases abruptly, likely predisposing the gut to villous atrophy, feeding intolerance, and necrotizing enterocolitis (NEC) once feedings are instituted.3 Human milk feeding is the only currently accepted modality for NEC prevention. This finding may be related to the presence of growth factors in human milk.4 G-CSF and EPO have the practical advantage of being available as sterile human recombinant factors.5 The ingested G-CSF and EPO are highly protected from digestion in the neonatal intestine and remain biologically active. They bind to specific receptors expressed on the surface of fetal villous enterocytes.6 The effect of enteral intake of these factors compared with each other in preventing NEC in preterm infants has not been widely explored, however. The present study aimed to evaluate the efficacy and safety of enteral recombinant human G-CSF (rhG-CSF) and recombinant human EPO (rhEPO) in preventing feeding intolerance and/or NEC in preterm infants. ARR EPO G-CSF NEC rhEPO rhG-CSF TPN VLBW WBC
Absolute risk reduction Erythropoietin Granulocyte colony-stimulating factor Necrotizing enterocolitis Recombinant human erythropoietin Recombinant human granulocyte colony-stimulating factor Total parenteral nutrition Very low birth weight White blood cell
From the Departments of 1Pediatrics, 2Clinical Pathology, and 3Radiodiagnosis, Faculty of Medicine, Ain Shams University, Cairo, Egypt The authors declare no conflicts of interest. Registered with ClinicalTrials.gov: NCT01441427. 0022-3476/$ - see front matter. Copyright ª 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2014.07.034
1
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Methods This double-blind, randomized trial was conducted in 90 preterm infants of gestational ages #33 weeks in the neonatal intensive care units of Ain-Shams University hospitals between March 2013 and March 2014. Exclusion criteria were congenital or acquired anomalies of the gastrointestinal tract and previous use of cytokine or intravenous immunoglobulin. The mothers of the newborns provided written consent for the protocol, which was approved by the Ethical Committee of Ain-Shams University hospitals. Reporting of the study conforms to Consolidated Standards of Reporting Trials 2010 statement.7 Of 120 neonates assessed for eligibility, 30 were excluded, including 3 who did not meet the inclusion criteria, 12 whose guardian declined participation, 13 who died before receiving the study drug or placebo, and 2 because of insufficient blood samples. The 90 included newborns were randomly assigned into 4 groups; 20 received enteral rhG-CSF, 20 received enteral rhEPO, 20 received both enteral rhG-CSF and rhEPO, and 30 received distilled water as a placebo (Figure 1; available at www.jpeds.com). Drug administration followed a predetermined schedule generated from random numbers in a 1:1 manner based on a computer-generated randomization sequence maintained within the investigational drug pharmacy, with allocation concealment in opaque sequentially numbered sealed envelopes. All of the newborns had a detailed perinatal and family history obtained, underwent a thorough clinical examination, and received standard neonatal care. Laboratory investigations included a complete blood count and a high-sensitivity Creactive protein measurement. Serum levels of G-CSF and EPO were measured on days 0 and 7. Serum G-CSF concentrations were determined using the RayBio Human G-CSF ELISA Kit (RayBiotech, Norcross, Georgia), and EPO concentrations were measured with the DRG EPO ELISA Kit (DRG International, Springfield, New Jersey). The newborns also underwent baseline evaluations in the first 72 hours of life that included abdominal and cranial ultrasounds as well as echocardiographic assessment. Abdominal ultrasound examination was done for evaluation of abnormally large bowel loops and intestinal wall edema with Doppler examination of superior mesenteric artery, as described previously.8 On the day that the attending neonatologist chose to begin feedings, the infants were assigned to receive enteral G-CSF and/or EPO or water placebo. A single daily dose of enteral rhG-CSF, 4.5 mg/kg (Geneleukin 300 mg/mL; Shandong Geneleuk Biopharmaceutical, Jinan, China), and/or enteral rhEPO, 88 IU/kg (Epoetin 2000 IU/mL; SEDICO Pharmaceuticals, 6th of October City, Egypt) was given with the start of enteral feeding. This dose was calculated to provide the same amount of growth factors ingested by a fetus swallowing 200 mL/kg/day of amniotic fluid, according to Christensen et al.1 The daily dose for each patient was diluted in sterile distilled water in accordance with the manufacturer’s instructions and as described by Canpolat et al,9 kept in a separate 2
Vol. -, No. opaque aliquot at 2-8 C, and allowed to warm to room temperature before administration through the orogastric/nasogastric tube with milk feedings for 7 days. The control group was given 1 mL of distilled water (placebo) once daily. Feeding Protocol The unit’s feeding policy was to commence early trophic feeding starting with 10-20 mL/kg/day, preferably breast milk if available, and progressing by 10-20 mL/kg/day for as long as tolerated (as judged by the attending neonatologist). This feeding policy was not altered during the study period. Feeding intolerance was defined as the presence of any of the following that interfered with the written enteral feeding plan: increased abdominal girth, emesis, gastric residual of $25% of the previous feed volume, abdominal distension, or the presence of macroscopic blood in stools. No specific value for increase in abdominal girth or volume of emesis was required.10 Withholding feeding was indicated in cases of feeding intolerance with a gastric residual $25% of the previous feed volume, suspected or proven NEC, heavily bile-stained or large gastric residuals or vomiting, clear abdominal pathology, significant abdominal distension or discoloration, blood in stool, or an unstable condition causing clinical concern, including significant cardiorespiratory instability. Follow-Up Strategy and Study Endpoint All infants in the treatment and control groups were followed up clinically with at least 2 complete physical examinations (1 in the morning and the other in the evening) and radiologically by abdominal radiography every other day. The following outcome data were recorded: day of successful start of enteral feeding; times to establish one-half, two-thirds, and full enteral feedings after drug/placebo administration (at least 150 mL/kg/day); time to the end of total parenteral nutrition (TPN); weight gain, expressed as g/kg/day; the incidence of NEC of stage I or worse according to Bell modified criteria11; surgery for NEC; NEC-related death; length of hospital stay; hospital readmission; and adverse effects of treatment (if any), including emesis, increased gastric residual, increased abdominal girth, diarrhea, skin rash, or blood pressure instability. The test solution (or placebo treatment) was discontinued when the patient’s enteral intake reached 100 mL/kg or after a maximum of 7 days, whichever came first. Statistical Analyses The sample size was calculated using power and sample size calculation program EpiInfo version 6.0 (Centers of Disease Control and Prevention, Atlanta, Georgia). Our unit statistics had a consistently high prevalence of milk intolerance in preterm infants of very low birth weight (VLBW) during the 4 consecutive years before the study. The mean total time to attain full enteral feeding in our unit was 35 5 days. Accordingly, a sample size of at least 20 neonates in each group was sufficient to detect a 30% difference in the time needed to achieve full feedings (a reduction from 35 days to 25 days) with a = 0.05 and 80% power of the study. El-Ganzoury et al
- 2014
ORIGINAL ARTICLES
Data analysis was done using SPSS version 15 (SPSS Inc, Chicago, Illinois). The Student t test was used to compare quantitative parametric variables between 2 groups, and ANOVA with a post hoc Bonferroni test was used for comparisons of more than 2 groups. The c2 test was used to test the association variables for categorical data. The Spearman rank correlation coefficient was used to measure the strength and direction of the linear relationship between 2 variables. A P value .05) (Figure 2). Moreover, no significant correlations were found between hemoglobin level and serum EPO level
(r = 0.07, P > .05) or between WBC count and serum G-CSF level (r = 0.2; P > .05) measured on the seventh day of enteral intake of growth factors (P > .05). Assessment of the relationships among various clinical variables in the whole treatment group revealed a significant negative correlation between gestational age and length of hospital stay (r = 0.368; P = .004). Moreover, there were significant negative linear correlations between birth weight and the time needed to achieve one-half (r = 0.414; P = .001), two-thirds (r = 0.386; P = .002), and full enteral feeding (r = 0.487; P < .001). Birth weight was also negatively correlated with weight gain (r = 0.323; P = .12) and length of hospital stay (r = 0.722; P < .001). The duration of hospital stay was positively correlated with the time needed to achieve one-half (r = 0.378; P = .003), two-thirds (r = 0.355; P = .005), and full enteral feeding (r = 0.38; P = .003). The time needed to discontinue TPN also was positively correlated with the achievement of one-half (r = 0.601; P = .001), two-thirds (r = 0.89; P < .001), and full enteral feeding (r = 0.867; P < .001) (data not shown).
Discussion Efforts to reduce mucosal atrophy in preterm neonates and attempts to prevent feeding intolerance among VLBW neonates have met with varying success. Provision of minimal enteral nutrition has been used in VLBW neonates as a means of avoiding disuse atrophy of the small bowel mucosa. Using an experimental model, Owens et al13 observed that minimal enteral nutrition did indeed result in maturation of intestinal motor function, but had little beneficial effect on the small bowel mucosa. Erythromycin administration has been tested but deemed generally not useful in reducing feeding intolerance in this population.14 The main objective in the present study was to improve feeding tolerance in preterm neonates by enterally administering intestinal growth factors, along with other benefits, such as decreasing nothing-by-mouth days and reducing the need for TPN to avoid its complications and cost. We postulated that VLBW infants could be provided with physiological quantities of rhG-CSF and/or rhEPO via intermittent orogastric/nasogastric administration in concentrations comparable with what they would have ingested from amniotic fluid had they remained in utero. Our results showed that the test solutions were well tolerated for up to 7 days of administration without related side effects. Barney et al15 used a sterile isotonic solution patterned after human amniotic fluid to treat feeding intolerance in infants and reported that none of the study infants had a recognized adverse reaction to the test solution. Juul et al16 also safely used enteral rhEPO in neonatal rats. We observed a significant improvement in feeding tolerance among all treatment groups, as reflected by the reduced time to achieve one-half, two-thirds, and full enteral feeding. Furthermore, neonates in the treatment groups had earlier weight gain, faster hospital discharge, and shorter duration of withholding of feeding. An important finding was the
Enteral Granulocyte-Colony Stimulating Factor and Erythropoietin Early in Life Improves Feeding Tolerance in Preterm Infants: A Randomized Controlled Trial
3
THE JOURNAL OF PEDIATRICS
Vol. -, No. -
www.jpeds.com
Table. Clinical and laboratory characteristics of the treatment groups and placebo control group Variables Male sex, n (%) Gestational age, wk, mean SD Birth weight, kg, mean SD Type of milk, n (%) Human milk Formula milk Mixed Age at the start of enteral feeding, d, mean SD Time to achieve one-half enteral feeding, d, mean SD Time to achieve two-thirds enteral feeding, d, mean SD Time to achieve full enteral feeding, d, mean SD Withholding feeding, d, mean SD Day of weight gain, d, mean SD Day to discontinue TPN, d, mean SD Hospital stay, d, mean SD Duration of treatment/placebo, d mean SD Gastric residuals >25%, n (%) NEC, n (%) Gaseous distention (radiography), n (%) Intestinal wall edema (ultrasound), n (%) Deaths, n (%) WBC count, 109/L, mean SD Hemoglobin, g/dL, mean SD Platelet count, 109/L, mean SD
Placebo group G-CSF group (n = 30) (n = 20)
EPO group (n = 20)
P value, treatment groups vs control group G-CSF + EPO group (n = 20) G-CSF group EPO group G-CSF + EPO group
10 (33.3) 30.5 1.5 1.36 0.29
8 (40) 30.8 1.8 1.26 0.27
5 (25) 30.2 1.8 1.31 0.31
8 (40) 30.4 1.9 1.19 0.28
.65 .46 .36 .23
.65 .56 .91 .23
.65 .91 .06 .23
0 (0) 14 (47) 16 (53) 1.7 1.4
1 (5) 9 (45) 10 (50) 1.8 1.4
0 (0) 5 (25) 15 (75) 2.3 1.75
0 (0) 5 (25) 15 (75) 2.5 3.3
.92
.16
.55
8.2 4.2
6.2 4.4
7.2 4.1
5.9 2.4
.027*
.04*
.013*
11.7 4.4
9.7 3.0
10.5 4.3
9.5 4.9
.11
.23
.033*
16.3 5.3
12.6 5.4
13.4 4.9
12.4 3.1
.005*
.032*
.006*
8.4 7.3 14.9 6.5 15 5.43 57.9 10.8 6.5 2.7 11 (36.7) 3 (10) 11 (36.7) 6 (20) 3 (10) 15.5 7.3 15.4 2.9 215 94
3.1 1.3 5 2.1 10.3 3.7 10.8 4.0 13.5 6.0 13.75 5.0 44.6 11.9 43.5 11.1 6.5 1.3 6.4 1.4 4 (20) 5 (25) 0 (0) 0 (0) 3 (15) 3 (15) 2 (10) 3 (15) 2 (10) 2 (10) 13.7 5.3 17.8 6.6 16.8 4.3 17.7 5.5 220 96 260 133
.05 .002* .17
