Volume 196, Issue 5 , Pages 422-423, May 2007
The outcome of infants with an Apgar score of zero at 10 minutes: past and future
Article Outline
The study by Harrington et al,1 which appears in this issue of the Journal, describes the outcomes of a series of patients with Apgar scores of zero at 10 minutes and systematically reviews published reports of similar infants. From a total of 94 neonates with Apgar scores of zero at 10 minutes, 78 infants (83%) died in the hospital or after discharge home; 10 infants (11%) had severe disability; 2 infants (2%) had moderate disability; 1 infant (1%) had mild disability; and 3 other infants had unknown neurodevelopmental outcomes. Although there are some limitations with the systematic review that affect the validity of the results, these findings provide evidence that resuscitation beyond 10 minutes in neonates without cardiac activity or other signs of life may not be justified at the present time.
See related article, page 463
To assess the validity of the results, a well-defined and representative population that is assembled at a common point is needed. The authors used a large prospective database from babies who were born in a single hospital from 1991-2004. Neonates were included if they had an Apgar score of zero at 10 minutes and were given continued resuscitation. Neonates with extreme prematurity (< 24 weeks of gestation or not fully resuscitated) or with major congenital abnormalities and the stillbirths were excluded. Thus, the babies who were included from their cohort were those in whom full resuscitation was provided and are a well-defined group in a hospital-based population. Important data on gestational age, birthweight, peripartum events, cord blood gases, and specific outcomes that included the timing of death and follow-up evaluation are included in the series of new patients, which helps in the extrapolation of the results; but the sample size is very small. In contrast to the well-described characteristics of the series of patients from their hospital database, the selection criteria of the neonates included from the review of the literature cannot be described well because of incomplete data in many of the original reports; thus, selection bias cannot be excluded.
A sufficiently long and complete follow-up period and an objective assessment of the main outcomes are also necessary to improve validity. Although the 2 survivors in the current study were followed for a relatively long period (2 and 5 years), many of the survivors reported in the reviewed literature had only short-term outcomes. A follow-up period, at least to school age, would be ideal, because earlier evaluations may not be predictive. Another limitation of most reports is the lack of a formal neurodevelopmental assessment to better describe the type and degree of functional disability. Further studies are needed to follow infants longer and perform formal neurodevelopmental evaluation to reduce assessment bias.
Subgroup analyses are ideal for studies on prognosis. The small number of patients and variables that are reported in the series and systematic review limits subgroup analyses that are necessary to adjust for prognostic or risk factors. For example, advanced statistical approaches have been used in neonates with hypoxic ischemic encephalopathy to predict death or disability.2 Specific aspects of the neurologic examination after birth (eg, posture and spontaneous activity), cord/initial blood gases, and certain maternal risks were found to be most predictive of outcome. Further studies should collect data on these and other prognostic/risk factors to allow better subgroup analyses.
Reporting and publication bias affects validity. It is more likely that clinicians will compile and attempt to publish what they consider interesting or unusual findings. Likewise, the reviewers and editors may be more willing to accept manuscripts that report these types of findings. High or low mortality or morbidity rates in this population may be considered of interest or unusual, so it is difficult to predict the effect of bias. Comparable results were reported in the small and large series that were reviewed. Therefore, a funnel-type analysis would suggest that publication/reporting biases are not a major problem with the current systematic review.3 Notably, the new patients reported by Harrington et al1 have mortality and morbidity rates similar to those of the reviewed patients. This strengthens the argument that the previously reported patients may be generalizable.
It is important to note that almost uniform mortality or severe morbidity rates in neonates with a zero Apgar score at 10 minutes occurred in the context of current clinical practice. It is possible that future interventions, such as neuroprotective ones, may improve outcomes. For example, therapeutic hypothermia in infants with neonatal encephalopathy that is associated with acute peripartum events has been tested in several randomized controlled trials.4, 5, 6, 7 The 3 most recent multicenter randomized trials by Gluckman et al,4 Shankaran et al,5 and Eichner et al6 generally enrolled neonates with Apgar scores of ≤ 5 at 10 minutes (among other criteria) and a previous trial4 included neonates with Apgar scores of ≤6 at 10 minutes. Infants with Apgar scores of zero at 10 minutes were not excluded. There was some heterogeneity of results and not universal consensus in the interpretation of the results.8 We analyzed outcomes of infants who were treated with hypothermia for neuroprotection from the existing randomized, controlled trials. There were 518 neonates included, and there was a reduction in death/disability (disability was defined as severe neurodevelopmental, moderate or severe neurodevelopmental cerebral palsy, respectively) from 62.9% in the control group to 46.2% in the hypothermia group (P < .001) with a number of 6 needed to treat (Figure). Mortality rates were decreased from 36.7% in the control group to 27.3% in the treated group (relative risk, 0.77; 95% CI, 0.60, 0.98; P < .05; number needed to treat, 11). Furthermore, there was no evidence of increased hospital or long-term adverse effects. Subgroup analyses by lower Apgar scores or Apgar score of zero at 10 minutes has not been performed, so it is impossible to determine the effect of therapeutic hypothermia on the type of patients who are addressed in the review by Harrington et al.1 Nonetheless, with further data on the subjects being collected in ongoing trials, the effect of hypothermia could be ascertained in different prognostic or risk subgroups in the future.
The findings of Harrington et al1 support the recent guidelines that were developed by the International Liaison Committee on Resuscitation (ILCOR),9 which state that “if there are no signs of life after 10 minutes of continuous and adequate resuscitation effort, it may be justifiable to stop resuscitation.” It should be emphasized that the ILCOR guidelines require adequate resuscitation efforts. It is known that resuscitation efforts frequently are not adequate, but the adequacy of resuscitation could not be assessed in the current review. Of note, the ILCOR recommendations were based on the evaluation of outcomes of fewer than one-half of the total number of patients who were reported by Harrington et al.1
The current review more than doubles the previous reviews on the subject and supports the new ILCOR guidelines on withdrawal of support in neonates with an Apgar score of zero at 10 minutes. However, the strength of the inference is limited by the various limitations and sources of biases that are addressed. Furthermore, in view of emerging neuroprotective therapies such as hypothermia, further research and data analyses are needed to determine whether these very high-risk neonates could benefit from newer interventions.
References
- . The long-term outcome in surviving infants with Apgar zero at ten minutes: a systemic review of the literature and hospital based cohort. Am J Obstet Gynecol. 2007;196:463.e1–463.e5
- . Predicting outcomes of neonates diagnosed with hypoxemic-ischemic encephalopathy. Pediatrics. 2006;119:2084–2093
- . Bias in meta-analysis detected by a simple, graphical test: asymmetry detected in funnel plot was probably due to true heterogeneity [letter]. BJM. 1998;316:469
- Selective head cooling with mild systemic hypothermia after neonatal encephalopathy: multicentre randomized trial. Lancet. 2005;365:663–670
- Whole-body hypothermia for neonates with hypoxic-ischemic encephalopathy. N Eng J Med. 2005;353:1574–1584
- Moderate hypothermia in neonatal encephalopathy: efficacy outcomes. Pediatr Neurol. 2005;32:11–17
- Neurodevelopmental outcome of infants treated with head cooling and mold hypothermia after perinatal asphyxia. Pediatrics. 2001;107:480–484
- Hypothermia and perinatal asphyxia: executive summary of the National Institute of Child Health and Human Development workshop. J Pediatr. 2006;148:170–175
- . The International Liaison Committee on Resuscitation (ILCOR) consensus on science with treatment recommendations for pediatric and neonatal patients: neonatal resuscitation. Pediatrics. 2006;117:e978–e988
PII: S0002-9378(06)02143-0
doi:10.1016/j.ajog.2006.10.878
© 2007 Mosby, Inc. All rights reserved.
Refers to article:
- The long-term outcome in surviving infants with Apgar zero at 10 minutes: a systematic review of the literature and hospital-based cohort
Volume 196, Issue 5 , Pages 422-423, May 2007
