Dear Editor

We read with great enthusiasm the article on use of rhG-CSF for treatment of sepsis in VLBW by Bedford et al.[1] It emphasises the previous studies on the subject of potential use of rhG-CSF in management of neonatal sepsis, especially given its fewer side effects and its significant effect on neutrophil counts.

While it is true that VLBW neonates have low neutrophil counts and storage pool, accounting for neutropoenia; we feel that other causes such as pregnancy induced hypertension (PIH)[2] and birth asphyxia,[3] would have needed exclusion from the study group, as they are likely to confound the results. Carr and Modi[2] have reported in their article that maternal PIH was associated with neutropoenia in 50% of neonates. In the study by Engle et al, birth asphyxia had associated neutropoenia in 35% of cases.

Although the authors have taken sensitivity to products of E. Coli as an exclusion criteria, we would like to know the parameters for its basis so that it could be incorporated in future studies. Necrotising enterocolitis (NEC) was taken as one of the major criteria for inclusion in the "confirmed clinical sepsis" group. At our centre , we have had cases of NEC with out other markers of sepsis. Besides, NEC is not always associated with infection, unless it is of epidemic variety. Sporadic cases of NEC may be related to other risk factors besides infection. In the study mentioned, the authors have found positive bacterial cultures in only two out of five cases of NEC. We feel that the subpopulation of NEC should have been included in the "Probable sepsis" group.

It need to be highlighted that the postmenstrual age is the sum of gestational age and postnatal age. Authors have mentioned that" -- mortality at 6 and 12 months of postmenstrual age was significantly lower in the treatment group". We feel that the authors have misinterpreted postnatal age for postmenstrual age. We would appreciate if due cognisance is given to this fact by them.

G-CSF has definitely proven itself as a potential treatment modality in the management of neonatal sepsis. The study of Bedfordet al, reconfirms its role while underlining minimum side effects of such a therapy.

References

(1) Bedford AR , Emmersonb A J B , Wilkinsona N, Chantb T, Sweetc DG,Hallidayc HL, Hollandd B, Daviesa EG. 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-F176 (May)

(2) Robert C, Modi N. Hematopoietic colony stimulating factors for preterm neonates. Arch Dis Child 1997;76: F128-F133.

(3) Engle RD, Rosenfeld CR. Neutropoenia in high risk neonates. J Pediatr 1984;105(6):982-86

(4) Newell SJ. Necrotising Enterocolitis. In Textbook of Neonatology, 3rd ed. Rennie JM, Roberton NRC (Eds). London: Churchill Livingstone. 1999:747-55

Dear Editor,

We thank Drs Jog and Patole for their extremely valid comments regarding our paper, which provide us with the opportunity to share the information requested. The complete dataset was omitted in the interests of brevity and summarised in the statement that there were no significant differences in maternal and peripartum characteristics between the two groups.

Antenatal steroids were administered to the mothers of 12/13 rhG-CSF treated vs 10/15 placebo group babies; surfactant to 6/13 vs 4/15; indomethacin to 3/13 vs 2/15; and postnatal glucocorticoids to 5/13 vs 4/15. Ventilation mode at the time of randomisation was by IPPV in 7/13 rhG-CSF vs 9/15 placebo group babies, HFOV in 2/13 vs 4/15, CPAP in 3/13 vs 1/15 and one baby in each group was spontaneously ventilating in air.

Weaning strategies from ventilation and fluid and electrolyte management was according to participating unit policy, but with protocol guidelines. As time to removal from mechanical ventilation was an exploratory efficacy endpoint, and it was acknowledged that management between units may vary, the details of ventilatory management were documented in the case report forms. No significant differences were identified when the study was unblinded.

Early deaths in the placebo group were as a result of confirmed sepsis (n=2) and culture-negative sepsis with respiratory distress syndrome (n=2). One baby receiving rhG-CSF died from candidaemia and staphylococcus epidermidis septicaemia. As stated in the paper all three late deaths were in the placebo group and secondary to chronic lung disease in 2 babies who had suffered recurrent episodes of suspected pneumonia. The third baby died from hepatorenal failure following necrotising enterocolitis. Although death could not be directly attributed to acute infection in these infants, infection is a recognised risk factor for late mortality in such a group [1].

We hope that this clarifies the reasons for our conclusions. Yours sincerely,

AR Bedford Russell, EG Davies.

Reference

(1) Stoll BJ et al,J Pediatr 1996;129:72-80

Dear Editor,

We read with interest the article by Bedford Russell et al on the use of rhG-CSF in very low birth weight neonates with presumed sepsis in a prospective randomised, placebo controlled trial. [1] Considering that the safety (primary objective) and efficacy of rhG-CSF (secondary outcome) was evaluated by factors like worsening of respiratory distress syndrome or chronic lung disease (CLD), thrombocytopenia, rise in absolute neutrophil counts, duration of ventilation, and antibiotic use, it would have been useful to provide the data on variables which would have influenced such outcomes. These include the following: Use of antenatal steroids, surfactant therapy, mode of ventilation, strategy for weaning ventilation, approach to management of fluid-electrolyte balance, use of indomethacin for patent ductus arteriosus [2], frequency of red cell transfusions for anemia of prematurity [3] and postnatal exposure to glucocorticoids.4 Additionally data on sepsis related mortality is not provided. Considering the multifactorial nature of CLD and thrombocytopenia (markers of the safety of rhG-CSF) such data may be helpful before concluding that rhG-CSF therapy is safe.

JOG SM, PATOLE SK
Department of Neonatology, Kirwan Hospital for Women
Townsville, Queensland 4817, Australia

References:

(1) Bedford Russell AR, Emmerson AJB, Wilkinson N, 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; 84:172-76.
(2) Herson VC, Krause PJ, Eisenfeld LI, Pontius L, Maderazo EG. Indomethacin -associated sepsis in very –low- birth-weight infants. Am J Dis Child 1988; 5:555-8.
(3) Hirano K, Morinobu T, Kim H, et al. Blood transfusion increases radical promoting non-transferrin bound iron in preterm infants. Arch Dis Child Fetal Neonatal Ed 2001; 84: F188-93M
(4) Wright IM, Skinner AM. Post –transfusion white cell count in the sick preterm neonate. J Paediatr Child Health 2001; 37(1): 44-6.