Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

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.

Miura E, Procianoy RS, Bittar C et al.

Pediatrics 2001;107:30-35. [abstract]

Reviewed by Waleska Arias MD, Section of Critical Care at Texas Children's Hospital, Baylor College of Medicine, Houston, TX.

Review posted July 15, 2002

---

I. What is being studied?:

The study objective:

The study objective was to determine if the administration of three doses of recombinant granulocyte colony stimulating factor (rG-CSF) (10 mcg/kg/day) to premature infants with early-onset bacterial sepsis would: a) reduce mortality rate by 50% and b) reduce subsequent acquisition of a nosocomial infection for a period of two weeks after dosing.

The study design:

The study design was a randomized, double blinded, placebo controlled trial of recombinant granulocyte colony stimulating factor administration to preterm infants with the clinical diagnosis of early-onset sepsis.

The patients included:

The study included all premature infants born from July 1996 to July 1997 at the Neonatal Intensive Care Unit of Hospital Porto Alegre with a birth weight between 500-2000gm. Additional inclusion criteria were patients with a gestational age of < 37 weeks, age < 5 days, and a clinical diagnosis of early-onset bacterial sepsis.

Other clinical criteria required were (three or more required): 1) abnormal body temperature not explained by environmental influences, 2) respiratory distress, 3) neurological dysfunction, 4) abdominal distension, 5) hyperbilirubinemia, and 6) diffuse hemorrhage. Laboratory criteria included were (one or more required): 1) neutropenia, 2) an immature to total neutrophil ratio > 0.2, 3) serum interleukin-6 levels of > 32 pg/ml and 4) a serum tumor necrosis factor > 12 pg/ml.

The patients excluded:

The patients were excluded from the trial if their mother received two or more of doses of antibiotics before delivery, if the neonate had severe congenital malformations, or if the neonate had severe asphyxia at birth without response to resuscitation measures.

The interventions compared:

Neonates in the treatment group received intravenous rG-CSF at a dose of 10 mcg/kg/day in 10 ml of 5% dextrose and 0.2% albumin infused by a pump over 30 minutes once daily for three consecutives days. The placebo group received an identical volume of the vehicle without rG-CSF.

The outcomes evaluated:

Two primary outcome variables were tested: a) mortality over a 30 day period and b) the development of nosocomial infections over a 2 week period after dosing. Secondary outcomes measured were length of time on a mechanical ventilator and length of time on supplemental oxygen.

II. Are the results of the study valid?

Primary questions:

1. Was the assignment of patients to treatments randomized?

Yes. All eligible neonates were randomized to receive either rG-CSF or placebo using a predetermined schedule generated from random numbers.

2. Were all patients who entered the trial properly accounted for and attributed at its conclusion?

Was followup complete?

Yes, all 44 preterm neonates enrolled were followed for a period of 30 days.

Were patients analyzed in the groups to which they were randomized?

Yes; on an "intention to treat basis", the patients were analyzed in each of the groups to which they were randomized. There were no crossovers.

Secondary questions:

3. Were patients, health workers, and study personnel "blind" to treatment?

Yes, the participants, investigators, and health care providers were blinded to the study assignment. Leukocyte and absolute neutrophil counts done at 24 and 48 hours after the first dose of rG-CSF were available to physicians. The fact that these values were significantly different in the two groups and also were available to the care providers, may potentially interfered with the blinding.

4. Were the groups similar at the start of the trial?

Yes. The randomization was done in order to assure that the baseline characteristics (birth weight, gestational age, etc.) were similar between treatment and control groups. The treatment and placebo groups were of similar gestational age (29 ± 3 vs. 31 ± 3 weeks) and birth weight (1376 ± 491 vs. 1404 ± 508 g). Both groups had similar APGARS (8; range 1-10 vs. 7; range 3-10) and 24 hour Neonatal Acute Physiology scores (9; range 0-23 vs. 8; range 3-28).

5. Aside from the experimental intervention, were the groups treated equally?

Maybe. All patients were "treated with conventional therapeutic interventions including antibiotics, supplemental oxygen, mechanical ventilation, intravenous fluids, and vasoactive drugs as deemed necessary by the attending neonatologist." The investigators did not report the number of patients that received each of the interventions listed. Since the study was blinded, it is likely that the groups were treated equally although the data to confirm this is not provided.

III. What were the results?

1. How large was the treatment effect?

The treatment effect was so small that the study was halted by the Data Safety Monitoring Board after only 44 patients of the anticipated 80 patients were enrolled.

A) G-CSF effect on mortality:

Table 1: Mortality

Measures of the Effect of Recombinant G-CSF Treatment on Mortality		95% CI
Risk without rG-CSF Treatment (Baseline risk): X	6/22 = 0.27 or 27%.	0.1315 to 0.4815
Risk with rG-CSF Treatment	5/22 = 0.23 or 23%	0.1012 to 0.4344
Absolute Risk Reduction (ARR) of Mortality with rG-CSF Treatment: X-Y	0.27 - 0.23 = 0.0455	-0.21 to 0.30
Relative Risk (RR) of Mortality with rG-CSF Treatment: Y/X	0.23/0.27 = 0.833	0.297 to 2.33
Relative Risk Reduction (RRR) of Mortality with rG-CSF Treatment: [1-Y/X] x 100 or [(X-Y) / X] x 100	[0.04/0.27] x 100 = 0.167 or 16.7%	-1.332 to 0.072
Number needed to treat (NNT): (1/ARR)	1/0.045=22	N/A

The relative risk reduction (RRR) of mortality after treatment was 16%; the treatment reduced the risk of mortality by 16%, but this effect was not significant. The number needed to treat (NNT) is 22, meaning that 22 patients would have to be treated in order to have one possible survival.

B) G-CSF effect on the Occurrence of Nosocomial Infections:

Table 2: Nosocomial Infection

Risk for the Ocurrence of Nosocomial Infections after G-CSF Treatment		95% CI
Risk of Infection without rG-CSF Treatment: X	9/19 = 0.473 or 47.3%	0.273 to 0.682
Risk of Infection with rG-CSF Treatment: Y	2/22 = 0.091 or 9.1%	0.0253 to 0.2781
Absolute Risk Reduction (ARR) of Infection with rG-CSF Treatment: X-Y	0.473 - 0.091 = 0.382	0.128 to 0.637
Relative Risk (RR) of Infection with rG-CSF Treatment: Y/X	0.091/0.473 = 0.19	0.047 to 0.781
Relative Risk Reduction (RRR) of Infection with rG-CSF Treatment: [1-Y/X] x 100 or [(X-Y) / X] x 100	[1 - 0.19] x 100 = 81% or [0.38/0.473] x 100 = 81%	0.219 to 0.953
Number needed to treat (NNT): (1/ARR)	1/0.382 = 2.6	2 to 8

As the article presented, the relative risk of acquiring a nosocomial infection was only 0.19. The calculated relative risk reduction was 81%, so there is apparent decrease in the number of subsequent nosocomial infections in neonates after rG-CFS treatment.

2. How precise was the estimate of the treatment effect?

The 95% CI for mortality and the occurrence of nosocomial infections after treatment are shown in above in Tables 1 and 2. The confidence intervals (CI) reported above for mortality showed that we can be 95% confident that the true relative risk of mortality with r G-CSF is between 1/3 rd that of the controls or 2.3 times worse. The 95% confidence interval for the relative risk of nosocomial infections reported in Table 2 above allows us to be confident the true risk for the occurrence of nosocomial infection with r G-CSF treatment can be as little as 5% of the risk of controls, or as much as 0.78 times as likely as the controls. This relative risk for the acquisition of nosocomial infections do not cross one, so it is significant as presented in the article

IV. Will the results help me in caring for my patients?

1. Can the results be applied to my patient care?

The findings of this article are not strong enough to support the use of rG-CSF in a regular basis in my patient population. The article only studied neonates and did not included patients above the neonatal period. We have to remember that the neonates' immune system is less mature than that of the pediatric patients, so we should expect different outcomes in the latter group. I think that further studies should be done with r G-CSF in order to be able to apply this treatment to the entire pediatric population with sepsis in our intensive care units.

2. Were all clinically important outcomes considered?

The neonate immature immune system makes them more prone to the development of sepsis and nosocomial infections. Morbidity (nosocomial infections) and mortality secondary to neutropenia are the most important outcomes to be evaluated in the patient population being studied in this article.

3. Are the likely treatment benefits worth the potential harms and costs?

Although the treatment of infections during neutropenia with rG-CSF transfusions appeared to be safe and no adverse effects were reported in the article, this is an expensive therapy. The price of G-CSF in the US ranged from $395 to $ 569 per cycle (3-5 days similar length of time used in the trial being evaluated)(3). In this article the benefits do not worth the cost because there were no observed reduction in mortality. I think that the use of this treatment should be reserved for patients with complicated febrile neutropenia where the expected risk of infection is high and the duration of neutropenia is prolonged, or those with documented infections that are refractory to antibiotic treatment.

4. What is in the literature about treatment with rG-CSF?

Several trials in the literature have demonstrated the ability of endogenous myeloid colony-stimulating factors (CSFs) to enhance the clinical management of immunosuppressed patients, especially those with cancer (3). As summarized by Buschel et al (1), these agents are associated with significant decreases in chemotherapy-associated infections, antibiotic use, length of hospital stays and mortality. Two major endogenous recombinant myeloid CSFs currently are being manufactured. Granulocyte macrophage CSF (GM-CSF) has broad activity in the proliferation and differentiation of myeloid lineage progenitor cells, whereas granulocyte CSF, acts selectively on cells of the granulocyte lineage. These agents have clinical benefits beyond enhancing neutrophil recovery, including shortening the duration of mucositis and diarrhea, stimulating dendritic cells, preventing infection and facilitating antitumor activity.

The use of these agents as prophylaxis in a regular basis for the average neutropenic patient in the PICU, (for example neutropenia secondary to septic shock ) need to overcome important side effects such as migration of stimulated neutrophils to previously injured tissues and potentiation of proinflammatory cytokines. Researchers suggested that additional studies should be done to see if a role for G-CSF can be demonstrated in non-neutropenic states, (such as burns or diabetes where there are adequate numbers of neutrophils but they are functionally defective), or in the management of infections in patients with normal neutrophil counts commonly seen in our intensive care units (2).

References

1. Buchsel PC, Forgey A, Grape FB, Hamann SS. Granulocyte macrophage colony-stimulating factor: current practice and novel approaches. Clin J Oncol Nurs. 2002 Jul-Aug;6(4):198-205. [abstract]
2. Siddiqui T, Burney IA, Kakepoto GN, Khurshid M, Salam A, Smego RA Jr. Lack of benefit of granulocyte macrophage or granulocyte colony stimulating factor in patients with febrile neutropenia. J Pak Med Assoc. 2002 May;52(5):206-10. [abstract]
3. Messori A, Trippoli S, Tendi E. G-CSF for the prophylaxis of neutropenic fever in patients with small cell lung cancer receiving myelosuppressive antineoplastic chemotherapy: meta-analysis and pharmacoeconomic evaluation. J Clin Pharm Ther. 1996 Apr;21(2):57-63. [abstract]
4. CEBM Stats calculator, University Health Network, Mount Sinai Hospital.
5. Hyper Stat on Line Textbook by David M. Lane from Rice University, Houston, Texas.

Comments

Back to the EB Journal Club Index