NEONATAL SOCIETY ABSTRACTS
Neutropenia, systemic infection and mortality in preterm, small for gestational age neonates
Presented at the Neonatal Society 2009 Summer Meeting (programme).
Carr R1, Doré C2, Brocklehurst P3, Modi N4
1 Guy’s Hospital, King’s College London, UK
2 MRC Clinical Trials Unit, London, UK
3 National Perinatal Epidemiology Unit, University of Oxford, UK
4 Imperial College, Chelsea & Westminster campus, London, UK
Background and Aims: Systemic infection (SI) is frequent in immature preterm babies and neutropenia is believed to increase the risk. The PROGRAMS trial (1) found that prophylactic Granulocyte-Macrophage Colony Stimulating Factor (GM-CSF) significantly increased neutrophil counts but there was no associated reduction in SI. In PROGRAMS, haematological, clinical and microbiological information were collected daily for 28 days from trial entry. We aimed to utilise these data to examine the impact of neutropenia and SI on mortality in this high risk population.
Methods: Infants <32 completed weeks gestation and birthweight <10th centile, who did not have strong indications of early onset SI, were recruited to the PROGRAMS trial within 72 h of birth. PROGRAMS was designed to test the primary hypothesis that 5 days prophylactic GM-CSF would lead to a reduction in SI and pre-discharge all cause mortality. As the trial results demonstrated no difference in SI or mortality (no rejection of the null hypothesis), data from the two randomisation arms were combined for the present analysis.
During 28 days from trial entry, absolute neutrophil count including band forms (ANC), platelet count, C-reactive protein (CRP), results of microbiological cultures and the presence of 12 predefined clinical signs associated with acute onset SI (2) were recorded daily.
Neutropenia was defined as ANC ≤1.0 x109/l. Culture positive SI defined as (i) a positive culture of blood, CSF, or suprapubic or catheter urine specimen, together with (ii) the acute onset of 3 or more predefined clinical signs of SI, plus (iii) a simultaneous rise in CRP, or fall in platelet count. Probable SI (culture-negative) was defined similarly, but without the presence of a positive culture. SI and concurrent neutropenia defined as SI plus ANC ≤1.0 x109/l occurring from 24 hours before to 48 hours after SI onset. For this analysis SI includes both culture positive and probable sepsis. Cox proportional hazards models were used to predict all cause mortality to discharge. Logistic regression analysis was used to predict SI during days 1-28. All models included testing for interaction with GM-CSF randomisation. The interaction term was never significant, therefore results reported are for the randomisation groups combined.
Results: 280 infants were recruited from 26 UK neonatal units. All are included in this analysis. During the 28 day study period 127 (45%) infants experienced one or more days of neutropenia, 114 (41%) experienced one or more episodes of SI, 102 (36%) had neither neutropenia nor SI. Infants experiencing neutropenia had an increased incidence of SI compared to those with no days of neutropenia (63/127, 50% v 51/153, 33%; odds ratio 2.0, 95%CI 1.2 to 3.2, p=0.006). Infants neutropenic at any time had an increased risk of pre-discharge mortality (neutropenia 40/127, 32% v no neutropenia 22/153, 14%; hazard ratio 2.4, 95%CI 1.4 to 4.1, p=0.001). Infants who had an episode of SI had increased risk of pre-discharge mortality (SI 46/114, 40% v no SI 16/166, 10%; hazard ratio 4.9, 95%CI 2.8 to 8.6, p<0.001). Mortality was even higher in infants who developed SI and concurrent neutropenia (15/25, 60% v no SI 16/166, 10%; hazard ratio 8.8, 95%CI 4.4 to 17.9, p<0.001). Mortality in the 102 infants who had neither neutropenia or SI during the study period was 6%. Neutropenia and SI were independent predictors of pre-discharge mortality.
Conclusions and Discussion: In the high risk infants recruited to PROGRAMS, the occurrence of both neutropenia and SI independently increased the risk of pre-discharge all cause mortality. This analysis confirms the premise underpinning the PROGRAMS trial, that neutropenia increases the risk of SI in the newborn. However, the finding that neutropenia increases mortality risk independent of SI, and that this is not ameliorated by strategies which boost neutrophil numbers, suggests neutropenia may be a marker for other as yet uncharacterised immunological deficits in these infants which will require a different therapeutic approach.
Acknowledgements: The assistance of Mrs Anne Smith, PROGRAMS trial administrator, is gratefully acknowledged.
1. Carr R, Brocklehurst P, Doré CJ, Modi N. Lancet 2009, 373:26-233
2. Modi N, Doré C, Saraswatula A et al. Arch Dis Child 2009, 94:F8-12