T Nakhla, S Imaizumi, J Saslow, Z Aghai, N Razi, G Stahl
extremely low birth weight, outcome, preterm, survival
T Nakhla, S Imaizumi, J Saslow, Z Aghai, N Razi, G Stahl. The Time To Death For Extremely Low Birth Weight Infants In The Neonatal Intensive Care Unit. The Internet Journal of Pediatrics and Neonatology. 2006 Volume 6 Number 2.
Conflict of interest
While there is a continued debate on what is considered the limit of viability, there is strong evidence that the survival of extremely low birth weight (ELBW) infants has significantly improved during the 1990s. 1,2,3,4,5,6,7 In our neonatal intensive care unit (NICU) we observed the deaths of some ELBW infants after a long hospital stay. The death of a baby in the NICU is a loss for the family and the health care team. 8 Others have raised the concern that although most NICU non-survivors expire early, they are not dying quite as early as they used to.4We sought to evaluate the length of stay (LOS) [time to death] among non-survivors over a 10 years period in our NICU. We projected that our data will assist in providing the parents with the most up-to-date knowledge on survival probabilities for ELBW infants who have extended hospital stays.
Material And Methods
All babies of birth weight 500-1000g, born at Cooper University Hospital between January 1, 1995 and December 31, 2004 were identified from a relational database, Neonatal Information System (NIS-Medical Data System, Wayne, PA). The NICU at Cooper University Hospital is a level III regional perinatal center in southern New Jersey providing specialized care to infants; the majority of them are inborn. Occasionally, an infant is transferred to another institution if a specialized care cannot be provided at our hospital or to be close to the parent's residence. Each infant was qualified for this study if; he/she was admitted to the NICU (i.e., survived the delivery room resuscitation) and was not transferred to another institution within 45 days of life. The Institutional Review Board at Cooper University Hospital exempted this retrospective study from review and waived HIPPA regulations since it included pre-existing medical records. The focus of this study was to evaluate the length of stay (LOS) among survivors [to discharge] and non-survivors [time to death]. Basic clinical data was collected on all ELBW infants. Clinical data for infants who survived to discharge was compared to those for infants who died using 2- tailed
There were 396 infants of birth weight 500-1000 grams born at Cooper Hospital between 1/1/1995 and 12/31/2004. Sixty infants died in the delivery room. Of the 336 infants admitted to the NICU, 14 infants were transferred to another institution prior to completing 45 days in the hospital and 2 were transferred to another institution after 45 days but were known to have eventually died prior to discharge. The remaining 320 infants are included in this report. Of the 320 infants included in this report, 105 expired in the NICU and 215 survived to discharge. Of the 215 infants that survived to discharge, 179 infants were discharged home, 10 infants were discharged to a rehabilitation facility and 26 infants were transferred to another institution for continuation of care. The characteristics of the 320 ELBW infants are shown in Table I. Over the 10-year period, the overall survival rate was 67% (49% in the birth weight subgroup 500-749g compared to 85% in the subgroup 750-1000g). The LOS among non-survivors [time to death] and survivors is shown in Table II broken down to percentage of infants in each group who represent a range of a corresponding length of stay. Forty-nine percent of the infants who died, died within the first week of life and 9% died after two months of life.
Figure 1: Shows the distribution of infants according to outcome (Died/Discharged/In hospital) in relation to time from birth plotted as a percentage of the whole group of 320 infants. Five percent (9/185) of all the infants who were still in the hospital after 60 days eventually died. Figure 2. Shows that survival improved by time but only reached 100% on day of life 180. percent survival on a given day.
This report puts to a glance the pattern of LOS [time to death] in non-survivors of ELBW infants. It has been well established in the literature that the probability of survival of an individual ELBW infant improves by time while the infant is in the NICU. 4,9,10 Furthermore, multi-center and single-center reports documented improvement in survival of ELBW infants in the 1990s and attributed this to major advances in management that included prenatal steroids and exogenous surfactant. 1,2,3,4,5,6,7,11We report our outcome data from 1995-2004 on infants with birth weight 500-1000 g. The overall survival rate for the 10 years was 67% compared to survival ranges from 57% to 71% in the most recent outcome reports. 1,2,4,7,12,13,14 Some recent reports compared the time to early death between the 1980s and early 1990s to the late 1990s and showed a significant decrease in early deaths in the late 1990s. 1,4,7,15Predictors of early death are discussed in the literature. 14,17,18,19,20 Previously published and utilized mortality risk prediction models are helpful tools in clinical trials differentiating treatment effects and possibly comparing outcomes across neonatal intensive care units. 14,17,18,19,20,21 They are, however, used early on in the clinical course and are not intended to limit the care of an individual infant since they are not sufficiently accurate to identify those patients who cannot be saved. 17,18,20
In this report, we had less deaths by DOL 4 (36% of all deaths) compared to others who reported ~ 80%. 2,4,10Forty-nine percent of the deaths occurred by one week in our cohort compared to 63% reported by Gould et al. 15Eighty percent of the deaths in our report occurred by DOL 30 which is comparable to 83% and 84% reported by others. 7,15It is known that early survival rates can be affected by the aggressiveness of delivery room resuscitation and initial management for which differences are reported between centers and even within the same center. 13,14,22,23 Our low rate of early death may reflect that we excluded all infants that died in the delivery room (60/396) in addition to that our report goes into the 2000s for which reports are not yet as abundant as the reports from 1980s and 1990s.
Our observation of some late deaths in the NICU prompted the question that was already raised by others concerning a possible effect of the advances in the management on time to death in ELBW infants. 4,15,24Data on late neonatal death is, however, not usually the focus of outcome reports. In this cohort 4% of the deaths occurred after DOL 90 compared to 5.7% of deaths after DOL 84 reported by Cooper et al in 1998. 9 Although inconsistent in the literature, we had significantly more survivors than non-survivors delivered by cesarean section. 2,16,25
The emotional impact of death in the NICU does not only affect the infant's family but it also affects the whole health care team. 8,26,27 Late death carries even a stronger impact since a long relationship has already developed between the health care team and the family in addition to the family's higher expectation for survival compared to earlier on in the hospitalization period.
Walsh et al reported that infants with prolonged mechanical ventilation 60 days had ‘less' chance of survival and ‘less' chance of survival free of neurodevelopmental impairment. 12
Our data along with the data reported by Walsh et al, suggest the possibility of developing mortality prediction models based on the clinical condition at a later time in the NICU course. 12Until this is reached, clinicians should find our data useful in overall evaluating the risk of not surviving till discharge based only on the length of stay but not tailored to the individual patient's condition.
We report that 95% (176/185) of the infants who were still in the hospital after DOL 60 survived till discharge. We suggest that; First, an agreement is needed among outcome reports on a time to death to be considered ‘late death'; Second, a larger cohort will be needed to evaluate the replication of our mortality rate after a specified LOS in the NICU and then look closer at this population for other possible predictors of late mortality.
Figure 1: Outcome (died, discharged or in-hospital) from birth to 12 month of 320 extremely low birth weight infants.
We appreciate the help of Nicole Kemble, RHIT in making the data available for our review. We also appreciate the help of B. Milcarek, PhD for his help with the statistical evaluation of the data.
Tarek Nakhla, MD Address: Division of Neonatology, Department of Pediatrics Children's Regional Center at Cooper University Hospital One Cooper Plaza, 755 Dorrance Bldg., Camden, NJ 08103 Telephone: (856) 342-2265 Fax: (856) 342-8007 E-mail: firstname.lastname@example.org