Recombinant Human Granulocyte Colony-stimulating Factor in Preterm Neonates with Sepsis and Relative Neutropenia: a Randomized, Single-Blind, Non-placebo-controlled Trial

Abstract

We performed a prospective, randomized, single-blind, non-placebo-controlled trial on preterm (<37 weeks) neonates (birth weight <2000g) with sepsis and absolute neutrophil counts (ANC) <5000 cells mm⁻³ to study the effect of recombinant human granulocyte colony-stimulating factor (rhG-CSF) on all-cause-neonatal mortality and hematological parameters (total leucocyte (TLC, ANC, absolute monocyte and absolute platelet counts). The rhG-CSF group (n = 20) received 10 µg/kg/day of intravenous infusion of rhG-CSF once daily for 5 days along with conventional therapy, and the control group (n = 20) received conventional therapy alone. Hematological parameters on Days 0, 1, 3, 5, 7 and 14 of study entry and all-cause mortality rates at discharge were recorded. Baseline characteristics between the rhG-CSF and control group including mean birth weight (1395 ± 289 vs. 1500 ± 231g), mean gestational age (31.5 ± 2.68 vs. 32.6 ± 2.23 weeks), initial neonatal complaints and maternal characteristics were comparable. Mortality rates were significantly less among the rhG-CSF group (3/20 (15%) vs. 7/20 (35%), p < 0.05). By Day 5 (for TLC) and Day 3 (for ANC) of start of the intervention, rhG-CSF group had significantly higher TLC (8189 ± 1570 vs. 6936 ± 1128 cells mm⁻³, p < 0.05) and ANC (4756 ± 1089 vs. 4213 ± 354 cells mm⁻³, p < 0.05) compared to controls. ANC levels recovered to levels >5000 cells mm⁻³ faster in the rhG-CSF group, with 80% babies having ANC >5000 cells mm⁻³ by Day 7 of study entry compared with 35% in the control group (p < 0.05). Preterm neonates with sepsis and neutropenia treated with rhG-CSF adjunctive therapy have decreased all-cause mortality at discharge and a quicker recovery of their total leucocyte and ANC.

Introduction

Neonates, especially preterm, have an increased susceptibility for infection and sepsis is the major cause of neonatal mortality. Estimates suggest 20% of very low birthweight (<1500g) babies suffer from culture proven sepsis and 10–20% die from sepsis in spite of treatment [1,2]. Sepsis-related neonatal mortality increases to >50% when associated with severe neutropenia [3]. Neonatal host defense is characterized by developmental immaturities of humoral, cellular and phagocytic components. Neonates, especially those born before 32 weeks, have an increased susceptibility to neutropenia during sepsis which is often associated with adverse prognosis. This deficient-cellular immunity, the first line of host defense, predisposes newborns to life-threatening bacterial and fungal infections.

The increased susceptibility of neonates to sepsis-related neutropenia is primarily due to a smaller neutrophil-storage pool, reduced neutrophil mobilization from bone marrow reserves and slower regeneration of neutrophils in the bone marrow [4]. In addition, neutrophil functions like chemotaxis, adherence, deformability, polarization are also impaired in the preterm babies [5]. Proliferation and differentiation of myeloid-progenitor cells is controlled by a family of hematopoietic growth factors, of which granulocyte colony-stimulating factor (G-CSF) and granulocyte macrophage colony-stimulating factor (GM-CSF) are commercially available. The GM-CSF and G-CSF are naturally occurring cytokines that stimulate the production and antibacterial function of monocytes and neutrophils. The actions of G-CSF include clonal maturation of neutrophil progenitors and augmentation of neutrophil functions including chemotaxis, phagocytosis, superoxide production and bactericidal activity. The rationale for administration of G-CSF in neonatal sepsis is based on rodent model of experimental sepsis where these cytokines were found to reduce mortality, but were more effective when given before bacterial inoculation [6]. Thereafter, Phase I and II clinical trials demonstrated that the hematopoietic growth factor administration was safe and well tolerated by very low birthweight neonates [7]. Two potential roles of G-CSF in neonatal sepsis are currently being explored. First, their use as an adjunctive treatment in newborns with established sepsis and neutropenia. Second, prophylactic option in high-risk newborns to prevent sepsis by stimulating-neutrophil production and enhance their phagocytic function.

The objective of our study was to determine in preterm neonates, with sepsis and relative neutropenia, the adjunctive role of recombinant human G-CSF (rhG-CSF) administration. Our primary outcome variable was mortality rate and the secondary outcome variable was comparison of hematological parameters [total leucocyte count (TLC), absolute neutrophil count (ANC), absolute monocyte count (AMC) and absolute platelet counts (APC)] between the rhG-CSF and control group.

Materials and Methods

Study design and ethics

This prospective, randomized, single-blind, non-placebo-controlled trial was conducted from September 2003 to December 2004 at the NICU, PGIMS, Rohtak, Haryana, India. The study protocol was approved by the institutional review board and the institutional ethics committee and written-informed consent was taken from the parents of the babies before enrollment.

Patient eligibility

Preterm newborns (gestation <37 weeks), with birthweight <2000g, born at our hospital during the study period and admitted in the NICU with clinical suspicion of sepsis were assessed for eligibility. Inclusion criteria were clinical signs of sepsis plus at least one positive blood culture plus relative neutropenia (defined as ANC <5000 cells mm⁻³). Babies with congenital malformation, suspected intrauterine infection, deranged liver function (SGOT/SGPT >4× normal) and deranged renal function (serum creatinine >2 mg dl⁻¹) were excluded.

Clinical sepsis was defined as any one of the following in the absence of alternative explanation: (i) fever (axillary temperature ≥38°C) or hypothermia (axillary temperature <36°C); (ii) respiratory distress or apnea; (iii) prolonged capillary-refill time (≥3 s) or shock; (iv) abdominal signs, e.g. abdominal distension, vomiting, radiologically confirmed necrotizing enterocolitis; (v) neurological problems, e.g. lethargy, seizures, excessive cry.

All babies with suspected sepsis were started on initial antimicrobial therapy of cefotaxime 100 mg kg⁻¹ day⁻¹ and amikacin 15 mg kg⁻¹ day⁻¹, which were later revised according to blood-culture sensitivity report.

Randomization and intervention procedure

Babies were randomized using computer generated random numbers into rhG-CSF or control group. Babies in the rhG-CSF group in addition to anti-microbial therapy, received injection rhG-CSF 10 µg kg⁻¹ slow intravenous infusion over 2 h once daily for 5 consecutive days. The study medication was stopped prior to the scheduled 5 days of therapy, if ANC increased >20 000 cells mm⁻³. The allocation sequence was concealed in sealed non-transparent envelopes until the start of the intervention to the baby. The investigator collecting the study data and following the babies during the study period and the hematologist analyzing the hematological parameters were blinded.

Evaluations

Basic information about maternal data including obstetric details, maternal complaints and neonatal data, on gestational age, anthropometry, Apgar score, neonatal problems, examination, vital signs, serum biochemistry, C-reactive protein (CRP) and blood sugar were recorded prior to the administration of the study intervention. Babies were classified as appropriate for gestational age (AGA), if birth weight was between 10th and 90th percentile for gestational age and small for gestational age (SGA) when birthweight was <10th percentile for gestational age according to Lubchenco’s charts [8].

The gestational age was assessed from the first day of the last menstrual period of the mother and confirmed by New Ballard’s scoring system [9]. When the gestational age by dates differed from the Ballard’s score by >2 weeks, the latter was taken as the gestation of the baby.

Hematological evaluations including TLC, manual differential leucocyte cell count, ANC, AMC and APC were estimated on Days 0, 1, 3, 5, 7 and 14 of study entry, drawn in ethylenediaminetetra-acetic acid-anti-coagulated specimen collection bottles. These hematological indices were estimated using standard procedure [10]. Blood cultures were repeated every 48–72 h until reported sterile.

Statistical analysis

Data are expressed as numbers (%), median (range) and mean ± SD. Groups were compared using Student’s t-test, the Wilcoxon rank-sum test and χ²-test, as appropriate. Logistic regression analysis was done to identify risk factors for mortality. All data were analyzed based on intention to treat. A p = 0.05 was taken as the cut-off for significance. Statistical analysis was done using intercooled STATA 8.0 (STATA Corp., College Station, TX, USA).

Results

Baseline characteristics and study flow

A total of 1533 preterm babies weighing <2000g were born at the institute during the study period, of which 425 were admitted in the NICU with clinical sepsis and were assessed for eligibility. Figure 1 shows the flow of the subjects through the phases of the study. A total of 40 babies were enrolled and randomized to receive either rhG-CSF (n = 20) plus conventional care or conventional care alone (n = 20). All babies completed the study. Table 1 shows the baseline characteristics of the babies in the two treatment groups.

Blood-culture isolates

Microorganisms isolated at study entry in the blood culture of rhG-CSF group were Pseudomonas aeruginosa (n = 15), Enterobacter spp (n = 3), one case each with Acinetobacter spp and Enterococcus spp. Microorganisms at study entry in the control group were P. aeruginosa (n = 12), Acinetobacter spp (n = 3), Staphylococcus aureus (n = 2); one case each with Enterobacter spp, Citrobacter spp and Candida albicans. All cases had late onset sepsis. In 70% babies of rhG-CSF group initial blood culture became sterile by Day 6 of the therapy. In the control group, a second organism was isolated in 40% of the babies by Day 6 of therapy and Staphyloccus aureus and Citrobacter were the most common isolates. In 3 (15%) babies in rhG-CSF group and 6 (30%) babies in the control group, the blood culture remained positive at death.

Hematological indices

At study entry (before first dose of the intervention), the TLC and ANC were lower in the rhG-CSF group. The baseline AMC and APC were comparable between the two groups (Table 2).

By Day 5 of start of the intervention, rhG-CSF group had significantly higher TLC compared to controls (8189 ± 1570 vs. 6936 ± 1128 cells mm⁻³, p < 0.05). TLC remained significantly more in the rhG-CSF on Day 7 of the therapy. By Day 14, the mean TLC was comparable between the rhG-CSF and control group.

The mean ANC increased above the critical level of 5000 cells mm⁻³ by Day 5 of the start of rhG-CSF infusion, whereas in the controls this critical level was attained by Day 14 of the enrollment. A total of 60% babies had ANC >5000 cells mm⁻³ by Day 5 of the start of therapy in the rhG-CSF group compared with one baby (5%) in the control group. ANC remained <5000 cells mm⁻³ in all babies who expired, in both the groups (Table 2).

AMC were significantly higher in the rhG-CSF group on Days 5 and 7 compared to the control group. By Day 14, AMC was comparable between the two groups. There was no significant difference in the APC on Days 0, 1, 3, 5, 7 and 14 in the two groups (Table 2).

Mortality

There were 7 (35%) deaths in the control group compared with 3 (15%) in the treatment group (p < 0.05) (Table 3). The cause of deaths among the rhG-CSF group was necrotizing enterocolitis (NEC) Stage III (n = 1), septic shock (n = 1), intraventricular hemorrhage, Grade III and IV (n = 1). In the control group, causes of death were pulmonary hemorrhage (n = 3), septic shock (n = 2), one each due to intraventricular hemorrhage and respiratory failure. The relative risk of all cause mortality among the control group and rhG-CSF group was 2.33. Multiple regression analysis to determine the relation between mortality (dependent variable) and independent variables including birth weight, rhG-CSF administration, meningitis, ANC not recovering to >5000 cells mm⁻³ by Day 7, blood culture becoming sterile by Day 6, revealed that ANC not recovering to >5000 cells mm⁻³ by Day 7 was an independent risk factor for mortality (p = 0.004).

Discussion

Neutropenia in neonates has been variably described as ANC <1500 cells mm⁻³, 1800 cells mm⁻³, 1100 cells mm⁻³, by different authors [11,12]. When defined as ANC <1500 cells mm⁻³, currently the most accepted definition, it is estimated to affect between 6% and 60% preterm infants [13,14]. Estimates suggest that, when sepsis is associated with severe neutropenia, mortality exceeds 50% [3]. We chose to include in this study neonates with relative neutropenia, defined as ANCs <5000 cells mm⁻³, instead of the conventional definition of 1500 cells mm⁻³, as ANC <1500 cells mm⁻³ is associated with a very high risk of mortality. Relative neutropenia is considered to be a low-risk group for sepsis-related morbidity and mortality in the developed world. However, in developing country settings as India, with infrastructural inadequacies and less advanced NICU facilities, sepsis-related relative neutropenia is also a significant cause of neonatal morbidity and mortality. In resource limited settings, it is compelling, therefore, to identify at-risk babies before their ANC falls too low and explore interventions that can prevent the downhill cascade of events related to sepsis and neutropenia.

Our study demonstrated that preterm babies with sepsis and neutropenia who were treated with rhG-CSF for 5 days along with conventional care had a significantly lower mortality compared with those treated with conventional care alone (15% vs. 35 %, p < 0.05). There was × 2.3 the probability of death among the control group compared with rhG-CSF group. ANC not recovering to level >5000 cells mm⁻³ by Day 7 was an independent predictor of mortality. The rhG-CSF administration by itself did not independently correlate with mortality, but since rhG-CSF administration resulted in higher ANC and faster recovery of ANC to levels >5000 cells mm⁻³, we conclude that rhG-CSF contributed to reduced mortality in the treatment group. For the overall favorable prognosis in the management of newborn sepsis and its related morbidity, it is important to have normal hematological indices, and thereby, effective host defense mechanism early in the course of the disease.

Since the first reported use of G-CSF to a 654g SGA neonate, with persistent neutropenia and sepsis published more than a decade ago [15], several studies have explored the potential use of hematopoietic growth factor administration in neonatal sepsis. The two main approaches to augment cellular immunity in neonates with the use of CSF have been either the prophylactic use in high-risk newborns or as treatment adjunct in established sepsis. Published studies on this issue, however, are remarkably heterogeneous with regard to patient selection, duration of intervention and outcome criteria.

A recently published multicenter randomized-controlled trial involving 280 neonates explored the prophylactic use of GM-CSF for the reduction of sepsis in preterm- (≤31 weeks) SGA neonates (PROGRAMS trial). Although the intervention significantly raised the neutrophil counts in the treatment group, there was no significant difference in sepsis-free survival, mortality or short-term clinical morbidity [16]. Another study utilizing GM-CSF prophylactically, also demonstrated a significant increase in peripheral neutrophil counts by the intervention, but incidence of confirmed nosocomial infections was not found to be significantly different in the two groups [17]).

The therapeutic application of hematopoietic CSFs in the treatment of neonatal sepsis has been addressed by several studies [18–20]. A Cochrane review [21] to this effect has revealed that in seven randomized-controlled trials involving 257 neonates, there was no evidence that addition of G-CSF or GM-CSF to the antibiotic therapy reduced immediate all-cause mortality. No significant survival advantage was seen at 14 days from the start of the therapy [typical relative risk (RR) 0.71 (95% CI 0.38, 1.33); typical risk difference (RD) −0.05 (95% CI −0.14, −0.04)]. The subgroup analysis of 97 infants from three treatment studies who, in addition to systemic infection, had clinically significant neutropenia (<1.7 × 109/l) at trial entry, did show a significant reduction in mortality by Day 14 [RR 0.34 (95% CI 0.12, −0.92); RD −0.18 (95% CI −0.33, −0.03); number needed to treat (NNT) 6 (95% CI 3,−33)]. The authors conclude that at the present there is not sufficient evidence to support the regular use of either G-CSF or GM-CSF into neonatal practice, either as treatment of established systemic infection to reduce resulting mortality, or as prophylaxis to prevent systemic infection in high-risk neonates.

The debate regarding the potential role of rhG-CSF and its place in the management of preterm neonates with sepsis and neutropenia is yet unsettled. Further studies with a larger sample size are warranted to establish whether rhG-CSF as an adjunctive therapy has a significant role in reducing neonatal morbidity and mortality. The potential role of combination of G-CSF and immunoglobulin in neonatal sepsis to augment both cellular and humoral immunity also warrants further research.

References