Cerebral palsy is defined as a static neuromuscular disorder characterized by an abnormal control of movement or posture appearing early in life that is not the result of a recognized progressive disease.1 The prevalence of cerebral palsy is approximately 2 per 1,000 live births. Of these, a large proportion has no identifiable cause for brain impairment.2
Of those for whom asphyxia is likely to have played a role, differences in the timing of the asphyxial episode may alter the neonatal presentation and clinical course. For example, some neonates may have been injured early in gestation and yet survived in utero to be liveborn; others may have been injured acutely intrapartum and may have died if not for intervention. In fact, the contribution of intrapartum events to the overall incidence of cerebral palsy attributable to “asphyxia” is relatively small and has been estimated to be approximately 10% of all instances.3 Among neonates for whom asphyxia is the most likely cause of later cerebral palsy, the clinical presentation may be complicated by the timing, mechanism, and severity of an asphyxial episode.4,5
The physiological mechanisms leading to asphyxia remain poorly understood. Because an asphyxial episode occurring intrapartum may be preventable, a variety of approaches has been used to ascertain the timing of this event. These approaches have largely depended on fetal monitoring characteristics, supplemented by adjunctive fetal and neonatal clinical observations.6–9 In 1999, a multidisciplinary International Cerebral Palsy Task Force set out criteria for the retrospective definition of an intrapartum hypoxic event that would have been sufficient to cause cerebral palsy.7 In 2003, the Task Force on Neonatal Encephalopathy and Cerebral Palsy, a joint effort of the American College of Obstetricians and Gynecologists and the American Academy of Pediatrics, updated and refined the criteria (Box 1), principally by adding a fourth essential criterion that there be no other identifiable etiologies for the injury and modifying the nonspecific criteria to include nonreassuring fetal heart rate status and to decrease the relevant Apgar score to 3 or less at 5 minutes.8
Strijbis et al,10 using a study population of 46 children with cerebral palsy born at term, were unable to demonstrate the presence of these criteria after reviewing the neonatal medical records. They concluded that the criteria appeared useful for excluding intrapartum asphyxia as a cause of cerebral palsy and, rather, identified multiple potential causes other than acute injury. We performed a MEDLINE search for all articles in all languages using the key word “neonatal encephalopathy” from 2003 (the year of publication of the Task Force report “Neonatal Encephalopathy and Cerebral Palsy: Defining the Pathogenesis and Pathophysiology”) through June 2011 and found no original research articles related to testing the sensitivity or specificity of the Task Force criteria. Here, we attempt to construct a study population in which the acute nature of the asphyxial event is likely, and we describe the presence or absence of the consensus criteria.
MATERIALS AND METHODS
This is a descriptive study designed to estimate the prevalence of the consensus criteria in a study population of acutely asphyxiated neonates with permanent brain injury. To enable an improved understanding of the pathogenesis of fetal brain injury, a national registry of brain-damaged infants was established in 1991.11,12 Eligibility criteria for this registry required a diagnosis of hypoxic ischemic encephalopathy in the neonatal period as recorded in the medical record by a neonatologist or pediatric neurologist (or both), and evidence of permanent neurological injury on follow-up records. In no instance did we examine or establish a diagnosis in these infants. All infants were free of identifiable factors potentially associated with cerebral palsy other than the occurrence of an intrapartum event, and all were singleton gestations of at least 37 weeks, with sufficient medical records to evaluate the following: the prenatal course, the intrapartum course (including intrapartum fetal monitoring), and the neonatal course. All patients had U.S. births occurring between 1976 and 2001, and all data were retrieved from medical record review.
Because of the strict requirement that the cases included in the present study must be consistent with the occurrence of an intrapartum asphyxial event, the following registry cases were excluded: intrapartum fetal deaths; cases of no available fetal monitor strip; cases of a persistent nonreactive fetal heart rate pattern from admission until delivery; cases of a fetal heart rate bradycardia on admission to the hospital, cases of a reactive fetal heart rate pattern on admission to the hospital followed by a nonreactive fetal heart rate, a substantial increase in the baseline fetal heart rate to a level of tachycardia with repetitive fetal heart rate decelerations and usually a loss of variability; and cases of an unclassifiable fetal heart rate pattern.
Thus, only those liveborn neonates who had a sudden, rapid, and sustained deterioration of the fetal heart rate from a previously reactive fetal monitor strip and a normal baseline heart rate on admission to the hospital were included.5,11 In all patients, once the sudden, rapid, and sustained deterioration of the fetal heart rate began, the fetal heart rate did not return to its original baseline, was unresponsive to remedial measures, and lasted until the delivery of the fetus. The length of the prolonged deceleration was recorded from its onset until the delivery of the fetus. This specific case definition, although limiting the number of patients for study inclusion through the elimination of potentially misclassified neonates, should provide an estimate of the prevalence of the consensus criteria in the population of interest, ie, neonates with acute intrapartum injury. This approach is in contrast to that of Stribjis et al,10 in which the etiology of cerebral palsy occurred through a variety of mechanisms, and an attempt was made to determine what proportion may have had an intrapartum cause. Here, we attempt to establish a study population that is likely to have an intrapartum cause and to determine the sensitivity of the consensus criteria for identifying the timing of the injury.
Data regarding the criteria (Box 1) set forth by the Task Force on Neonatal Encephalopathy and Cerebral Palsy8 were then abstracted from the medical record. Because the umbilical artery pH is one of the four essential criteria, patients in the case group without an umbilical arterial cord blood gas obtained at the time of delivery were excluded. To determine the presence of moderate or severe neonatal encephalopathy, Sarnat criteria were applied by J.P.P. to each patient in the case group based on the neonatal records to specify the degree of neonatal encephalopathy.13 The diagnosis of spastic quadriplegia or dyskinetic cerebral palsy was established through the medical records. Using the examples described by the Task Force on Neonatal Encephalopathy and Cerebral Palsy,8 hypoxic sentinel events were recorded, if present, for each patient in the case group. Whether a neonate had an acute nonfocal cerebral abnormality was based solely on the neuroimaging report. At no time did we read or interpret the neuroimaging performed on these neonates.
Documentation of multisystem involvement was fundamentally consistent with the definitions set forth by Hankins et al,14 with the additional specificity provided here. First, hepatic dysfunction was defined by the presence of elevated liver transaminases (serum glutamic-oxaloacetic transaminase [SGOT] and serum glutamic-pyruvic transaminase [SGPT]) of more than 100 units/L or L-lactate dehydrogenase (LDH) levels of 600 units/L or more within the first 5 days of life. Second, gastrointestinal dysfunction was defined by the reporting of the radiographic appearance of a gasless abdomen within the first 24 hours of life. At no time did we review the radiographs in these patients. The presence or absence of a gasless abdomen was based solely on the radiograph report within the medical records. If the radiographic report did not document the intestinal findings, then the status of the intestines was considered unknown. Third, pulmonary dysfunction was defined as the use of mechanical ventilation. It is acknowledged that cerebral injury itself may be associated with apnea and the need for mechanical ventilation.15 The total duration of mechanical ventilation in hours for each neonate was recorded. Fourth, cardiac dysfunction was defined as the use of vasopressors, the presence of an abnormal echocardiogram, or an elevation of creatinine phosphokinase (CPK) of 400 units/L or more within the first 5 days of life. Fifth, renal dysfunction was defined as the presence of oliguria (less than 1.00 mL/kg/h) during the first 24 hours of life or the first serum creatinine level of 1.0 mg/dL or more within the first 48 hours of life. Sixth, hematological dysfunction was not defined. However, we recorded the number of nucleated red blood cells per 100 white blood cells (expressed as the percent nucleated red blood cells), the absolute nucleated red blood cells count, the nucleated red blood cells clearance time, which was defined as the number of hours from birth to the time the nucleated red blood cells count first reached 0% or 1%, the initial platelet count, and whether the neonatal platelet count declined to less than 100,000 per mm3 within the first 5 days of life. Seventh, because all neonates had brain injury diagnosed, multisystem involvement was defined as dysfunction in any one of the following organ systems: hepatic, gastrointestinal, pulmonary, cardiac, or renal.
Data are expressed mean±standard deviation, and median and range are included for descriptive purposes. Exact 95% confidence intervals of the proportion of neonates with the essential criteria and nonspecific criteria (Table 1) were calculated using PASS statistical software (PASS 2005). This protocol was approved by the Institutional Review Board of Citrus Valley Medical Center.
A total of 524 singleton term neonates with permanent central nervous system (CNS) impairment or intrapartum fetal death was present in the registry. Forty-eight (9%) died intrapartum, fetal heart rate monitor strips were not present in 21 (4%), leaving 455 patients appropriate for evaluation. Of these, 114 (25%) satisfied the fetal heart rate requirements. Because of the requirement to have documentation regarding umbilical artery cord blood at the time of birth, this yielded a final study population of 39 (34%) patients. Maternal characteristics are described in Table 1.
The umbilical cord arterial blood obtained at delivery demonstrated a pH less than 7.00 in 38 of 39 (97.4%) and a base deficit of 12 mmol/L or more in 30 of 30 (100%) patients. Nine patients without a base deficit had only an umbilical cord blood arterial pH performed. The single patient with an umbilical blood pH more than 7.00 had a neonatal heart rate in excess of 100 beats per minute at birth and an umbilical arterial blood pH of 7.08, with a base deficit of 17 mmol/L.
The presence of moderate or severe neonatal encephalopathy was found in 34 (87%) of the patients. Of these, 33 (97%) demonstrated either moderate (26 [79%]) or severe (7 [21%]) neonatal encephalopathy. A single fetus with mild neonatal encephalopathy had an umbilical artery cord blood pH of 6.86, and although there was no evidence of a sentinel hypoxic event, the fetus did have meconium-stained amniotic fluid and bradycardia that lasted 53 minutes from the time of onset until delivery.
Of the 39 patients in the case group available for study, neonatal outcomes were known in 36 (92%) of the patients and were as follows. Death occurred before a diagnosis could be established clinically in 12 (33%) of the patients and ranged from 1 day to 54 months from the time of birth. Of the 27 (67%) survivors, the long-term neurologic outcome was known in 24 (89%) of the patients. Spastic quadriplegia or dyskinetic cerebral palsy was present in 22 (92%) of the 24 patients. Of the remaining two patients in the case group, one had hemiplegia and one had ataxia. None of the neonates had any other identifiable etiologies found other than hypoxic ischemic encephalopathy for their cerebral palsy.
The proportion of each of the Task Force on Neonatal Encephalopathy and Cerebral Palsy's nonspecific criteria satisfied are illustrated in Table 2. The presence of a sentinel hypoxic event was known in 31 (80%) and occurred for the following reasons: uterine rupture with abruption in 14 (45%); uterine rupture alone in 12 (39%); shoulder dystocia in one (3%), cord prolapse in one (3%); fetal exsanguination in one (3%); amniotic fluid embolus in one (3%); and placental abruption alone in one (3%). In the eight patients without an identifiable sentinel hypoxic event (20%), five (13%) patients had evidence of meconium 2+ or thicker. In the remaining three patients (8%), no factor could be identified to account for the sudden, rapid, and sustained bradycardia.
Given the definition of the study population, all the patients satisfied the fetal heart rate requirements. Each patient had a reactive fetal monitor strip on admission with a normal baseline rate followed by a sudden, rapid, and sustained deterioration that was unresponsive to remedial measures and lasted until delivery. For the 36 fetuses for which the duration of the deceleration could be calculated, the bradycardia lasted a mean of 32±9.6 minutes, with a median of 32 minutes and a range of 16 to 57 minutes (Fig. 1).
The 5-minute Apgar score was documented for all 39 neonates, with 22 (56%) neonates demonstrating a score of 0 to 3. The 10-minute Apgar score was known in 34 (87%) of the 39 neonates; 10 (29%) had Apgar scores of 3 or less.
Multisystem involvement was assessed in each patient, and the results are demonstrated in Table 3. Renal dysfunction as evidenced by oliguria within 24 hours of life or an elevated first serum creatinine level obtained within the first 48 hours of life was identified in 21 of 27 (78%) and 16 of 29 (55%), respectively. Of these, seven patients resumed normal renal function during the first 48 hours of life. Of those with two or more urine outputs available during the first 3 days of life, the urine output returned to normal (1.00 mL/kg/h or more) on days of life 1–6 (29%), 2–13 (62%), and 3–2 (10%). The serum creatinine levels were known in 21 patients during the first 3 days of life, and these increased in four (19%), remained the same in one (5%), and decreased in 16 (76%).
Hepatic dysfunction as measured by SGOT, SGPT, and LDH is demonstrated in Table 3. Of the 14 neonates in whom hepatic status could be evaluated, 10 (71%) had hepatic dysfunction. Overall, SGOT was found to be abnormal in 7 of 14 (50%), and the values ranged from 44 units/L to 426 units/L. The SGPT level was abnormal in only 1 of 12 (8%), and the values ranged from 12 units/L to 132 units/L. The LDH level was considered abnormal in seven of nine (78%), with a range from 553 to 2,550 units/L.
Cardiac dysfunction was evaluated in 33 neonates, of whom 18 (55%) had evidence of cardiac dysfunction. Of the 12 survivors who received dopamine, the mean duration of use was 78.0±35.6 hours, with a median of 72.6 and a range of 32.8–153 hours. The CPK level was obtained in seven (18%) patients. The CPK values ranged from 344 units/L to 10,923 units/L.
Mechanical ventilation was medically necessary in 32 of the 32 neonates for whom information was available (100%). The duration of mechanical ventilation was known in 24 (75%). Of these, the mean duration of ventilator use was 91.8±105.1 hours, with a median of 81.1 hours and a range of 0.3–252 hours. It is acknowledged that the use of mechanical ventilation was necessary because of the underlying CNS injury and not necessarily as a result of pulmonary dysfunction alone.
Hematologic dysfunction as measured by an nucleated red blood cells count of 26% or more or a platelet count of 100,000 per mm3 or less was encountered in 3 of 32 (9%) and 1 of 27 (4%), respectively. The mean nucleated red blood cells count was 10.6%±14.5%, with a median of 6.5% and a range of 0%–76%. Of 25 patients (64%) with sufficient data to determine the nucleated red blood cells clearance time, none demonstrated a clearance time more than 80 hours. The mean nucleated red blood cells clearance time was 23.6±20.0 hours, with a median of 15.7 hours and a range of 0.8–60.4 hours.
For the 28 neonates with available data, the initial mean platelet count was 201,000±64,000 per mm3, with a median of 201,000 per mm3 and a range of 86,000 per mm3–407,000 per mm3. The proportion of neonates whose platelet count declined to less than 100,000 per mm3 within 5 days of birth was 1 of 27 (4%). This single neonate had a deceleration to delivery interval of 23 minutes, a sentinel hypoxic event attributable to a shoulder dystocia, and a zero assisted Apgar score at 10 minutes. This neonate died at 15 days of life. None of the other neonates demonstrated a decline in platelets to less than 100,000 per mm3 during the first 5 days of life. Intestinal complications as exemplified by the radiographic observation of a gasless abdomen were encountered in 7 of 15 patients (47%).
Overall, organ dysfunction was identified in 34 of 34 (100%) patients with available neonatal data. Of these, the number of organs affected was 1–4 (12%), 2–9 (26%), 3–14 (41%), and 4–7 (21%). No fetus had five organ systems affected. The most common organ systems involved were the lungs, followed by the kidneys.
Of the 39 patients, 21 (54%) had neuroimaging or autopsy findings consistent with an acute nonfocal CNS lesion or cerebral edema. Cerebral edema was documented in 7 of 25 (28%) patients. The timing of the diagnosis of cerebral edema ranged from 10.4 hours to 264 hours and reflected the retrospective nature of this investigation, the technology available, and the reliance on reports rather than the neuroimaging itself to establish the diagnosis.
Applying the template set forth by the Task Force on Neonatal Encephalopathy and Cerebral Palsy, 24 (62%) had sufficient information on all the essential criteria. Of these 24 patients, 21 (88%) satisfied all the essential criteria. The three patients (12%) that did not satisfy these requirements were as follows: a neonate with ataxia; a neonate with hemiplegia; and a neonate with umbilical artery pH of 7.08. Of the 21 patients who met all the essential criteria, six (29%) satisfied all the essential and nonspecific criteria. The reasons why patients did not satisfy the nonspecific criteria were as follows: Apgar score more than 3 beyond 5 minutes (n=7); no hypoxic sentinel event (n=2); and insufficient neuroimaging data (n=9). Some patients had more than one reason.
Because all the patients had no other identifiable etiology of hypoxic ischemic encephalopathy, 36 (95%) met three or more of the essential criteria. Of these 36, all (100%) exhibited one or more of the nonspecific criteria.
This study evaluated the consistency of the Task Force on Neonatal Encephalopathy and Cerebral Palsy criteria for acute intrapartum asphyxia in a homogeneous population of neonates with permanent neurologic impairment. The results indicate that the Task Force's criteria are consistent with the clinical findings in such patients. This level of consistency may, in large part, be attributable to limiting the study population to those with a specific acute fetal heart rate pattern and the presence of an umbilical arterial cord gas. As demonstrated herein, significant metabolic acidemia was observed in 38 of 39 (97.5%) patients. The sole exception was a neonate whose neonatal heart rate was in excess of 100 beats per minute at birth. Previously, Phelan et al16 demonstrated that the presence of metabolic acidemia (umbilical artery cord pH less than 7.00 and a base deficit of 12 mmol/L or more) was significantly linked to the presence of a slow heart rate (less than 100 beat per minute) at the time of birth.
The cerebral palsy type noted in the current study population, with its acute intrapartum fetal heart rate pattern, was consistent with the cerebral palsy type anticipated by the Task Force. It is interesting to note that of the 12 neonates who died before the establishment of a diagnosis, the neuroimaging in two (17%) of them demonstrated neurological insults in areas of the neonatal brain that would have, had the neonates survived, resulted in spastic quadriplegia or dyskinetic cerebral palsy.
In our series, 34 of 34 (100%) had evidence of one or more organs involved and 14 (41%) had three organ systems affected. The two most common organs affected were the lungs and kidneys. These findings are in keeping with our previous observations in acutely asphyxiated neonates with permanent brain injury.17
The strength of this study is that we were able to ascertain the prevalence of the Task Force criteria in a group of neonates who appeared to undergo acute asphyxial injury and to confirm that these neonates have characteristics that are consistent with the criteria. This study is limited by the rarity of this presentation of hypoxic ischemic encephalopathy, the need for case-finding through the use of a registry, and the lack of uniformity of documentation associated with such case-finding methods. Furthermore, because this is a descriptive study of affected neonates only, we cannot estimate the specificity of these criteria. Such issues also limit the degree of generalization of our results to other populations of acutely asphyxiated neonates.
In summary, our results suggest that the criteria developed by the Task Force on Neonatal Encephalopathy and Cerebral Palsy appear to provide a template that may be useful for the analysis of cases involving intrapartum neonatal asphyxial injury. Whether this approach will also be valuable in populations of neonates with differing intrapartum fetal heart rate patterns remains to be investigated. Nevertheless, fetuses with a sudden, rapid, and sustained deterioration of the fetal heart rate during labor from a previously reactive fetal heart rate pattern who were found to have cerebral palsy appeared to demonstrate clinical characteristics that were consistent with the criteria recommended by the Task Force on Neonatal Encephalopathy and Cerebral Palsy.