Neonatal sepsis with neutropenia: granulocyte-colony stimulating factor (G-CSF)

The Guidelines Project, an initiative of the Brazilian Medical Association, aims to combine information from the medical field in order to standardize producers to assist the reasoning and decision-making of doctors.

The information provided through this project must be assessed and criticized by the physician responsible for the conduct that will be adopted, depending on the conditions and the clinical status of each patient.

Sepsis is a serious complication in newborns and an important cause of mortality, especially in premature infants. One of the factors that contribute to the susceptibility of these patients to infections is the immaturity of their immune system. One of the possible genesis of this phenomenon seems to lie in the presence of an inhibitor of placental origin, which acts by decreasing the endogenous production of the granulocyte-colony stimulating factor (G-CSF). A systematic review was performed using the PICO search system. Newborns with neonatal sepsis and neutropenia (patient), Treatment with G-CSF plus antibiotics compared with conventional therapy (intervention), Death (any cause), and adverse effects (outcome). The search resulted in 85 articles, of which 12 studies were included. The details of the methodology and results of this guideline are set out in ^{Annex 1}.

INTRODUCTION

Sepsis is a serious complication in newborns and an important cause of mortality, especially in premature infants¹,². Although there has been a reduction in its incidence in recent decades³, it is estimated that 1.4 million newborns still die annually as a result of invasive infection⁴.

In addition to mortality, septic conditions are associated with a worse neurological prognosis and the occurrence of chronic pulmonary disease⁵, impairing the quality of life and raising the social costs of survivors.

One of the factors that contribute to the susceptibility of these patients to infections is the immaturity of their immune system⁴. Structural and functional abnormalities of the immune components have been described in this age group, and the changes probably more related to the occurrence of sepsis are those of polymorphonuclear cells⁷. Among them, neutrophils also presented a quantitative reduction in up to 58% of preterm infants, depending on the definition of neutropenia used⁸.

Neutrophils play a fundamental role in the defense against bacterial infections, and people with a chronic deficiency of these cells exhibit recurrent infections and early mortality. In neonates, risk factors such as gestational hypertension, intrauterine growth restriction, fetus-fetal transfusion syndrome, and hemolytic anemia due to Rh incompatibility are related to the absolute reduction in the count of these cells⁹.

One of the possible genesis of this phenomenon seems to lie in the presence of an inhibitor of placental origin, which acts by decreasing the endogenous production of the granulocyte-colony stimulating factor (G-CSF)¹⁰. In addition, it has been observed that newborns with a low number of neutrophils in the bloodstream, in fact, have a hidden a medullary reserve, which can quickly respond to the use of a recombinant stimulating factor (rG-CSF)¹¹.

For this reason, rG-CSF has been suggested as prevention and adjuvant treatment of infections in the neonatal period¹². Clinical trials and systematic reviews concluded that the drug is safe for this population, but were not sufficient to recommend its routine use¹³.

RESULTS

The population included 355 newborns with sepsis and undergoing treatment with G-CSF plus antibiotics (N=183), compared (N=172) with conventional therapy (N=179), and followed-up to measure the death outcomes due to all causes related (in time) to the septic episode during hospitalization and death at 14 or 28 days, in addition to adverse events (Table 1).

TABLE 1
SEPSIS IN NEWBORNS WITH NEUTROPENIA ≤5,000 CELLS/MM3

DESCRIPTION OF THE STUDIES INCLUDED
Study	Population	Intervention	Control	Outcome/ Time
Bedford-Russell, 2001 Multicenter	28 NB with clinical signs of sepsis and neutrophils <5 x 10(9)/L. IG >25s, PC 500 – 1,500g, Id ≤28 d	13 NB, G-CSF 10 mcg/kg/d, IV, maximum of 14 d	15 NB, placebo	Death during hospitalization Death at 14 days Death at 28 days
Drossou-Agakidou, 1998 Single Center	35 NB with clinical signs of sepsis and neutrophils <5 x 10(9)/L. IG 24 – 37s, PC 720 – 2,940 g, Id <28 d	19 NB, G-CSF 10 mcg/kg/d, SC, 3 d	16 NB, placebo	Death at 14 days
Schibler, 1998 Multicenter	20 NB with early-onset sepsis and neutropenia <1.7x109/L. IG 24 – 40s, PC 530 – 3,667 g, Id <3 d	10 NB, G-CSF 10 mcg/kg/d, IV, 3 d	10 NB, placebo	Death during hospitalization Death at 14 days Death at 28 days
Gathwala, 2011 Single Center	40 NB with clinical signs of sepsis and neutrophils <5 x 10(9)/L. IG <37 s, PC <2,000 g	20 NB, G-CSF 10 mcg/kg/d, IV, 5 d	20 NB, conventional therapy	Death during hospitalization
Chaudhuri, 2012 Single Center	78 NB with early-onset sepsis, neutropenia <1,500 cells/mm3 IG <34 s; PC 1,100 to 2,500 g; Id 3 d	39 NB, G-CSF 10 mcg/kg/d, 3 d	39 NB, placebo	Death during hospitalization
Borjianyazdi, 2013 Single Center	46 NB with clinical signs of sepsis; neutropenia (CAN ≤5,000/μL) IG 30 – 37 s, PC 530 – 3,667 g, Id <10 d	23 NB, G-CSF 10 mcg/ kg/d, SC, 5 d	23 NB, placebo	Death at 14 days
Aktaş, 2015 Single Center	56 NB with sepsis, neutropenic (<1.0 x 109/L) IG 32 s, PC 1,001 ± 240	33 NV, G-CSF 10 mcg/ kg/d, IV, (1– 4 d; median, 2 d) up to CAN ≥1.0 x 109/L	23 NB, conventional therapy	Death during hospitalization
Gupta, 2016 Single Center	52 NB with clinical signs of sepsis, neutropenia <1,800 cells/mm3 IG 32 s, PC 1500.38 ±306.35, Id <25 d	26 NB, G-CSF 10 mg/kg/d, 3 d, SC	26 NB, conventional therapy	Death during hospitalization

• NB = newborns, CAN = absolute neutrophil count, IG = gestational age, s = weeks, PC = body weight, Id = age, d = days, IV = intravenous, SC = subcutaneous

In order to achieve some comparability between studies, the data for the death outcome were recovered in such a way as to allow an analysis of mortality at the 14th or 28th day, from the beginning G-CSF therapy.

Among the studies selected, the mean gestational age of the patients ranged from 24 to 40 weeks (only one study included NB <40 weeks²⁴; seven studies¹⁷-²³ included NB <37 weeks). The body weight ranged from 500 to 3,667g, and age was ≤28 days. The median duration of G-CSF treatment was four days (range: 3-14 days) with a dose of 10 mcg/kg/day, intravenously, or subcutaneously. Four studies²⁰,²³,²⁴,¹⁷ included NB with a cutoff point for neutropenia <1,800 cells/mm³, and four others¹⁸,¹⁹,²¹,²² with neutropenia ≤5,000 cells/mm³.

In relation to the risk of bias, there was no description of the randomization in two studies¹⁷ and of the blinded allocation in three¹⁷,²¹,²⁴. No blinding of the team was described in one study¹⁷, and there was no double-blinding in another²¹. One study presented a loss of ≥20%, and no intention to treat analysis was conducted¹⁷. The overall risk of bias of the studies included is considered not severe (Table 2).

TABLE 2
DESCRIPTION OF THE BIASES OF THE STUDIES INCLUDED

Study/Year	Rand	Blinded allocation	Blinded participants and team	Blinded evaluator	Losses	Prog. characteristics	Appropriate outcomes	AITT	Sample size calculation
Aktas 2015
Bedford-Russel 2001
Borjianyazdi 2013
Chaudihuri 2012
Drossou-Agakidou 1998
Gathwala 2011
Gupta 2016
Schibler 1998

• ITTA = Intention to Treat Analysis; Boxes: green = absence of bias, red = presence of bias, yellow = uncertain risk of bias

META-ANALYSIS FOR THE DEATH OUTCOME

The meta-analyzed outcomes were death during hospitalization and death at 14 or 28 days from the beginning of the G-CSF therapy.

In the analysis of in-hospital mortality, there was a reduction in the risk by 19% (95% CI, 10 to 29; I²=26%) with the use of G-CSF in comparison to the conventional treatment with or without placebo; it was necessary to treat about 5 NB to prevent one death (NNT). The quality of evidence is high to sustain this outcome (Table 3).

TABLE 3
G-CSF COMPARED TO THE CONVENTIONAL TREATMENT IN NEONATAL SEPSIS WITH NEUTROPENIA (CAN <5,000 CELLS/MM3)

Evaluation of certainty							Summary of Results
No. of Participants (studies) Follow-up	Risk of bias	Inconsistency	Indirect evidence	Imprecision	Publication bias	Overall certainty of the evidence	Study event rates (%)		Relative Effect (95% CI)	Potential absolute effects
							With control	With G-CSF		Risk with control	Risk difference with G-CSF
DEATH (HOSPITALIZATION)
274 (6 RCTs)	not severe	not severe	not severe	not severe	None	⊕ ⊕ ⊕ ⊕ HIGH	46/133 (34.6%)	22/141 (15.6%)	RR 0.43 (0.27 to 0.69)	346 per 1,000	minus 197 per 1,000 (from minus 252 to minus 107)
DEATH (14 DAYS)
129 (4 RCTs)	not severe	not severe	not severe	severe a	none	⊕ ⊕ ⊕ Օ MODERATE	10/64 (15.6%)	8/65 (12.3%)	RR 0.79 (0.34 to 1.81)	156 per 1,000	33 less per 1,000 (from 103 less to 127 more)
DEATH (28 DAYS)
48 (2 RCTs)	not severe	not severe	not severe	severe a	none	⊕ ⊕ ⊕ Օ MODERATE	7/25 (28.0%)	3/23 (13.0%)	RR 0.46 (0.13 to 1.57)	280 per 1,000	151 less per 1,000 (from 244 less to 160 more)

• CI: Confidence interval; RR: Risk ratio. Explanations = ^a. Non-significant difference.

In the analyses of mortality at 14 or 28 days, there was no difference in the use of G-CSF in comparison to the conventional treatment, with or without a placebo. The quality of evidence is moderate to sustain this outcome (Table 3- Annexes).

Heterogeneity

There was no evidence of significant heterogeneity between the studies.

Adverse effects of G-CSF administered in newborn sepsis

No serious adverse effects were reported with the use of G-CSF in the studies evaluated in this review. There were mild adverse effects, such as cutaneous reactions, irritability, electrolytic alterations, cardiac alterations, peripheral edema, and thrombocytopenia, but with no significant difference between G-CSF and the placebo. Four of the eight studies reported no adverse effects¹⁹-²².

SYNTHESIS OF EVIDENCE

In newborns with sepsis and neutropenia ≤5,000 cells/mm³, the use of G-CSF reduces the risk of in-hospital death by 19% (NNT = 5). G-CSF has proved to be a safe drug.

DISCUSSION

In a previous meta-analysis conducted by Bernstein (2001), five studies with a total of 155 patients were evaluated, and the mortality rate was lower among patients who received G-CSF than among those who received the placebo. However, when the non-randomized studies were excluded, the beneficial effects of G-CSF therapy were found to be less consistent. Thus, the routine use of G-CSF could not be recommended for all newborns with sepsis²⁵.

FIGURE 2
DEATH DURING HOSPITALIZATION. COMPARISON CHART: G-CSF VERSUS CONVENTIONAL THERAPY, OUTCOME: ALL-CAUSE MORTALITY DURING HOSPITALIZATION.

FIGURE 3
DEATH AT 14 OR 28 DAYS FROM THE BEGINNING OF THE G-CSF THERAPY. COMPARISON CHART: RHG-CSF VERSUS CONVENTIONAL THERAPY, OUTCOME: MORTALITY DUE TO ALL CAUSES: 14 DAYS

FIGURE 4
COMPARISON CHART: RHG-CSF VERSUS CONVENTIONAL THERAPY, OUTCOME: MORTALITY DUE TO ALL CAUSES AT 28 DAYS.

In the Cochrane Systematic Review, Carr and Modi (2003) published a meta-analysis of seven studies to determine the safety and efficacy of rhG-CSF to reduce mortality in the treatment of suspected or proven systemic infections. They concluded there was no evidence to support the addition of G-CSF or granulocyte-macrophage colony-stimulating factor (GM-CSF) to antibiotic therapy in preterm infants with suspected systemic infection to reduce mortality. A subgroup analysis of 97 infants of three treatment studies which, in addition to systemic infection, presented clinically significant neutropenia (<1.7, 109/L) at the time they were enrolled in the study shows a significant reduction in mortality until the 14th day [RR, 0.34 (95% CI, 0.12, 0.92); RD, 0.18 (95% CI, 0.33, 0.03); number needed to treat, 6 (95% CI, 3,33)]²⁶.

In this systematic review with meta-analysis, we included eight randomized clinical trials on G-CSF therapies in newborns with sepsis and neutropenia ≤5,000 cells/mm³ comparing them with the conventional treatment, and that reported all-cause mortality. The review showed evidence, based on data from six studies included in the meta-analysis, that the addition of G-CSF to antibiotic therapy in newborns with sepsis and neutropenia <5,000 cells/mm³reduces mortality in hospitalization due to all causes [RD -0.19 (95% CI -0.29, 0.10); number needed to treat, 5 (95% CI, 3, 11)].

No significant advantage in survival was observed at 14 and 28 days from the beginning of therapy [RD -0.03 (95% CI -0.15, 0.08) and RD -0.15 (95% CI -0.37, 0.07), respectively]. However, all eight treatment studies were small, with the largest including only 78 newborns.

Our study has some limitations. The studies included in our meta-analysis were performed in a wide range of time, during which the definition of neonatal sepsis and the diagnostic and treatment methods changed. Four different definitions of neutropenia were used in different studies, and the definitions of sepsis varied between the clinic one and proof through positive culture. Patients of multiple birth weights and gestational ages were reported in all eight studies, and it is possible that this is a significant confounding factor.

REFERENCES

• ¹
  Stoll BJ, Hansen N. Infections in VLBW infants: studies from the NICHD Neonatal Research Network. Semin Perinatol 2003;27:293-301.
• ²
  Patel RM et al. Causes and timing of death in extremely premature infants from 2000 through 2011. NEJM 2015;372:331-40.
• ³
  Adams M, Bassler D. Practice variations and rates of late onset sepsis and necrotizing enterocolitis in very preterm born infants, a review. Transl Pediatr 2019;8:212-26.
• ⁴
  Shane AL, Stoll BJ. Neonatal sepsis: progress towards improved outcomes. J Infect 2014;68(Suppl 1):S24-32.
• ⁵
  Stoll BJ et al. Neurodevelopmental and growth impairment among extremely low-birth-weight infants with neonatal infection. JAMA 2004;292:2357-65.
• ⁶
  Adams-Chapman I. Long-term impact of infection on the preterm neonate. Semin Perinatol 212;36:462-70.
• ⁷
  Hill HR. Biochemical, structural, and functional abnormalities of polymorphonuclear leukocytes in the neonate. Pediatr Res 1987;22(4):375-82.
• ⁸
  Maheshwari A. Neutropenia in the newborn. Curr Opin Hematol 2014;21:43-49.
• ⁹
  Christensen RD. The role of recombinant leukocyte colony-stimulating factors in the neonatal intensive care unit. In: Ohls RK, Maheshwari A. Hematology, immunology and infectious disease: neonatology questions and controversies. 2.ed. Philadelphia: Elsevier, 2012.
• ¹⁰
  Koenig JM, Christensen RD. The mechanism responsible for diminished neotrophil production in neonates delivered of women with pregnancy-induced hypertension. Am J Obstet Gynecol 1991;165:467-73.
• ¹¹
  Juul SE, Christensen RD. Effect of recombinant granulocyte colony-stimulating fator on blood neutrophil concentrations among patients with “idiopathic neonatal neutropenia”: a randomized, placebo-controleed trial. J Perinatol 2003;23:493-7.
• ¹²
  Parravicini E et al. Myeloid hematopoietic growth factors and their role in prevetion and/or treatment of neonatal sepsis. Transfus Med Rev 2002;16:11-24.
• ¹³
  Carr R, Modi N, Doré CJ. G-CSF and GM-CSF for treating or preventing neonatal infections (Review). Cochrane Database Sys Rev 2003;(3):CD003066.
• ¹⁴
  Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ. 2009;339:b2535.
• ¹⁵
  Review Manager (RevMan) [Computer program]. Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014.
• ¹⁶
  GRADEpro GDT: GRADEpro Guideline Development Tool [Software]. McMaster University, 2015 (developed by Evidence Prime, Inc.). Available from gradepro.org.
• ¹⁷
  Aktaş D, Demirel B, Gürsoy T, Ovalı F. A randomized case-controlled study of recombinant human granulocyte colony stimulating factor for the treatment of sepsis in preterm neutropenic infants. Pediatr Neonatol 2015;56:171-5. PMID: 25458637
• ¹⁸
  Bedford Russell AR, Emmerson AJ, Wilkinson N, Chant T, Sweet DG, Halliday HL, et al. A trial of recombinant human granulocyte colony stimulating factor for the treatment of very low birthweight infants with presumed sepsis and neutropenia. Arch Dis Child Fetal Neonatal Ed 2001;84:F172-6. PMID: 11320043
• ¹⁹
  Borjianyazdi L, Froomandi M, Noori Shadkam M, Hashemi A, Fallah R. The effect of granulocyte colony stimulating factor administration on preterm infant with neutropenia and clinical sepsis: a randomized clinical trial. Iran J Ped Hematol Oncol 2013;3:64-8. PMID: 24575272
• ²⁰
  Chaudhuri J, Mitra S, Mukhopadhyay D, Chakraborty S, Chatterjee S. Granulocyte Colony-stimulating Factor for Preterms with Sepsis and Neutropenia: A Randomized Controlled Trial. J Clin Neonatol 2012;1(4):202-6. PMID: 24027727
• ²¹
  Drossou-Agakidou V, Kanakoudi-Tsakalidou F, Sarafidis K, Taparkou A, Tzimouli V, Tsandali H, et al. Administration of recombinant human granulocyte-colony stimulating factor to septic neonates induces neutrophilia and enhances the neutrophil respiratory burst and beta2 integrin expression. Results of a randomized controlled trial. Eur J Pediatr 1998;157:583-8. PMID: 9686822
• ²²
  Gathwala G, Walia M, Bala H, Singh S. Recombinant human granulocyte colony-stimulating factor in preterm neonates with sepsis and relative neutropenia: a randomized, single-blind, non-placebo-controlled trial. J Trop Pediatr 2012;58:12-8. PMID: 21296862
• ²³
  Gupta P, Sethi RS, Chaurasia OS, Sethi A. Role of Granulocyte Colony Stimulating Factor (G-Csf) in Neonatal Sepsis with Neutropenia. People’s Journal of Scientific Research July 2016 ; V o lume 9, Issue 2
• ²⁴
  Schibler KR, Osborne KA, Leung LY, Le TV, Baker SI, Thompson DD. A randomized, placebo-controlled trial of granulocyte colony-stimulating factor administration to newborn infants with neutropenia and clinical signs of early onset sepsis. Pediatrics 1998;102:6-13. PMID: 9651406
• ²⁵
  Bernstein HM, Pollock BH, Calhoun DA, Christensen RD. Administration of recombinant granulocyte colony-stimulating factor to neonates with septicemia: a meta-analysis. J Pediatr. 2001;138(6):917–20.
• ²⁶
  Carr R, Modi N, Dore´ C. G-CSF and GM-CSF for treating or preventing neonatal infections. Cochrane Database Syst Rev 2003;3:CD003066.

• REFERENCES OF THE STUDIES EXCLUDED FROM THE REVIEW AND REASONS
  Khan TH, Shahidullah M, Mannan MA, Nahar N. Effect of recombinant human granulocyte colony stimulating factor (rhG-CSF) for the treatment of neonates in presumed sepsis with neutropenia. Mymensingh Med J 2012;21:469-74. PMID: 22828545.
  COULD NOT ACCESS FULL TEXT. ABSTRACT DID NOT REPORT RANDOMIZATION
  El-Ganzoury MM, El-Farrash RA, Saad AA, Mohamed AG, El-Sherbini IG. In vivo effect of recombinant human granulocyte colony-stimulating factor on neutrophilic expression of CD11b in septic neonates: a randomized controlled trial. Pediatr Hematol Oncol 2012;29:272-84. PMID: 22475305.
  EVALUATES AN INTERMEDIATE OUTCOME (production of neutrophils and expression of CD11b)
  Nayeri F, Soheili H, Kaveh M, Oloomi Yazdi Z, Shariat M, Dalili H. Comparison of two regimens of RhG-CSF in neutropenic neonatal septicemia: a randomized clinical trial. Acta Med Iran 2011;49:575-8. PMID: 22052139
  EVALUATES DIFFERENT DOSES OF THE DRUG
  Kücüködük S, Sezer T, Yildiran A, Albayrak D. Randomized, double-blinded, placebo-controlled trial of early administration of recombinant human granulocyte colony stimulating factor to non-neutropenic preterm newborns between 33 and 36 weeks with presumed sepsis. Scand J Infect Dis 2002;34:893-7. PMID: 12587621.
  NOT NEUTROPENIC
  Kocherlakota P, La Gamma EF. Human granulocyte colony-stimulating factor may improve outcome attributable to neonatal sepsis complicated by neutropenia. Pediatrics 1997;100:E6. PMID: 9200380.
  NOT RANDOMIZED
  Rodwell RL, Gray PH, Taylor KM, Minchinton R. Granulocyte colony stimulating factor treatment for alloimmune neonatal neutropenia. Arch Dis Child Fetal Neonatal Ed 1996;75:F57-8. PMID: 8795359
  NOT PATIENTS WITH SEPSIS (autoimmune neonatal neutropenia)
  Barak Y, Leibovitz E, Mogilner B, Juster-Reicher A, Amitay M, Ballin A, et al. The in vivo effect of recombinant human granulocyte colony-stimulating factor in neutropenic neonates with sepsis. Eur J Pediatr 1997;156:643-6.
  NOT RANDOMIZED
  Ahmad A, Laborada G, Bussel J, Nesin M. Comparison of recombinant granulocyte colony-stimulating factor, recombinant human granulocyte-macrophage colony-stimulating factor and placebo for treatment of septic preterm infants. Pediatr Infect Dis J 2002;21:1061-5. PMID: 12442030
  NEUTROPENIA WAS NOT A COMPULSORY CRITERION AT ENROLLMENT
  Gillan ER, Christensen RD, Suen Y, Ellis R, van de Ven C, Cairo MS. A randomized, placebo-controlled trial of recombinant human granulocyte colony-stimulating factor administration in newborn infants with presumed sepsis: significant induction of peripheral and bone marrow neutrophilia. Blood 1994 1;84:1427-33. PMID: 7520770
  NEUTROPENIA WAS NOT A COMPULSORY CRITERION AT ENROLLMENT
  Miura E, Procianoy RS, Bittar C, Miura CS, Miura MS, Mello C, Christensen RD. A randomized, double-masked, placebo-controlled trial of recombinant granulocyte colony-stimulating factor administration to preterm infants with the clinical diagnosis of early-onset sepsis. Pediatrics 2001;107:30-5. PMID: 11134430
  NEUTROPENIA WAS NOT A COMPULSORY CRITERION AT ENROLLMENT
  Saeidi R, Akhavan H, Tavakkolafshari J , Shapouri-Moghaddam A. Clinical and Biochemical Effects of Recombinant Human Granulocyte Colony-Stimulating Factor on the Prognosis of Preterm Infants with Early Onset Neonatal Sepsis, Arch Pediatr Infect Dis. 2019 ; 7:e67807. doi: 10.5812/pedinfect.67807.
  MIGHT HAVE INCLUDED NON NEUTROPENIC

ANNEX I

Clinical question

What is the impact on overall mortality outcomes (death from any cause) and adverse events of G-CSF in the treatment of newborns with neonatal sepsis and neutropenia [Absolute Neutrophil Count (ANC)<5,000 mm³] compared to the conventional therapy?

Objective

The objective of this assessment is to identify the efficacy and safety of granulocyte-colony stimulating factor (G-CSF) in the treatment of neonatal sepsis with neutropenia less than 5,000 cells/mm³, in comparison with conventional therapy.

Structured question

P	Newborns with neonatal sepsis and neutropenia (CAN ≤5,000 cells/mm3)
I	Treatment with G-CSF plus antibiotics compared with the conventional therapy
C	-
O	Death (due to any cause) and adverse effects

Eligibility criteria

We followed the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (Prism) to conduct this systematic review¹⁴.

The selection of the studies and the evaluation of the titles and abstracts obtained from the search strategy in the databases consulted were independently and blindly conducted in total accordance with the inclusion and exclusion criteria. Finally, studies with potential relevance were separated.

When the title and the summary were not enlightening, we sought for the full article.

Only studies with texts available in its entirety were considered for critical evaluation.

Excluded outcome - failure in the correction of preexisting neutropenia during treatment.

Randomized clinical trial studies were selected.

Without time or language restrictions.

Full text available for access.

Search for papers

Databases and research strategy

The search for evidence will be conducted on virtual scientific information databases, using the following search strategy:

• -Medline/PubMed - (Granulocyte Colony-Stimulating Factor OR rG-CSF OR G-CSF OR GCSF OR Recombinant Proteins) AND (Infant, Newborn OR Infant, Premature OR neonates OR preterm infants) AND (Sepsis OR Neutropenia) AND Random*.

• - Central/Cochrane Library - (Granulocyte Colony-Stimulating Factor OR rG-CSF OR G-CSF OR GCSF OR Recombinant Proteins) AND (Infant, Newborn OR Infant, Premature OR neonates OR preterm infants) AND (Sepsis OR Neutropenia).

• - Lilacs via BVS - same search as in Central/Cochrane Library.

We also searched sources of data still unpublished, in progress, in gray literature and manual, performed by checking the list of “References” of the studies included in this review or in previous reviews.

All searches were performed by October 2019.

CRITICAL EVALUATION

Relevance - clinical importance

This guideline was prepared by means of a clinically relevant question in order to gather information in medicine to standardize approaches and assist in decision-making.

Results application - External validity

The process of retrieving the studies, as well as the evaluation of the titles and abstracts obtained, will be conducted by two researchers with expertise in the development of systematic reviews (A. S. and W. M. B.), independently and blinded, in accordance with the eligibility criteria listed above. When there is any disagreement regarding the selection of studies between the researchers, a third reviewer will be consulted (I. F.).

RESULTS

The data extraction will be performed independently by three reviewers, the results compared, and disagreements resolved by discussing them.

We will extract the following data from the studies: name of the author and year of publication, study population, intervention and comparison methods, the absolute number of deaths and adverse events, time of follow-up.

The methodological quality of the eligible trials will be evaluated independently by individual reviewers, not blinded, using the full text of the paper. Randomized clinical trials will have their risk of biases analyzed according to the following criteria: randomization, blinded allocation, double-blinding, losses, prognostic characteristics, presence of relevant outcome, time for the outcome, the method for outcome measurement, sample size calculation, early interruption, presence of other biases.

The results of the studies included will be meta-analyzed using the RevMan 5.3 software ¹⁵. The heterogeneity of treatment effects between the tests was evaluated graphically and tested using the standard chi-square test. A weighted estimate of the relative risk (RR) between the studies was calculated using a fixed-effects model. The risk difference (RD) and the number needed to treat (NNT) were also calculated, and the final measure was used to sustain the synthesis of evidence, which will answer the clinical question of this review. The 95% CI was calculated for all outcomes.

During the search for evidence, we retrieved 85 papers, of which 12 were then selected based on their title and abstract; they all evaluate the use of G-CSF plus antibiotics for the treatment of newborns with sepsis and neutropenia, in comparison with conventional therapy. The 12 studies that met the eligibility criteria were then were accessed for analysis of their full text. Of the 12 studies, five¹⁷-²⁴ were selected to support this review; the grounds for exclusion and the list of studies excluded are available in the references and in Figure 1.

FIGURE 1
THE SELECTION OF RETRIEVED FROM THE VIRTUAL DATABASES OF SCIENTIFIC INFORMATION

Application of evidence - Recommendation

The quality of evidence will be graded as high, moderate, low, or very low using the Grade instrument¹⁶ and taking into account the risk of bias, the presence of inconsistency, vagueness or indirect evidence in the meta-analysis of the outcomes of death and adverse events, and the presence of publication bias.

Publication Dates

• Publication in this collection
  27 Feb 2020
• Date of issue
  2020

History

• Accepted
  1 Dec 2019