Using characteristic changes in laboratory values to assist in the diagnosis of hemorrhagic shock and encephalopathy syndrome.

OBJECTIVES: (1) To determine whether the laboratory abnormalities in hemorrhagic shock encephalopathy syndrome (HSES) follow a characteristic pattern of evolution; and (2) to describe the clinical characteristics and outcome of this group of children. DESIGN: Retrospective review of hospital records. SETTING: Pediatric intensive care unit in an academic children's hospital. PATIENTS: Seventeen children who received the diagnosis of HSES. MEASUREMENTS: (1) To support the diagnosis of HSES, the following laboratory tests were obtained: creatine phosphokinase, alanine aminotransferase, aspartate aminotransferase, blood urea nitrogen, creatinine, anemia, thrombocytopenia, coagulation abnormalities, and a metabolic acidosis; and (2) patient characteristics at the time of admission and during the illness and patient outcome were recorded. RESULTS: The median time for laboratory abnormalities in HSES to reach the most aberrant levels was 1.2 to 1.4 days. After reaching maximal deviation, the laboratory values gradually returned to normal. The severity of laboratory abnormalities was not predictive of mortality. All patients were comatose, had seizures, and required fluid resuscitation. Respiratory failure occurred in 12 (71%) of 17 patients. Vasopressor infusions were used for 8 (47%). Viral cultures were performed in 15 patients and were positive in 8 (53%). Eleven (64%) patients survived. CONCLUSIONS: The laboratory abnormalities that are characteristic of HSES follow a distinctive pattern of evolution during the course of the illness. Although the sensitivity, specificity, and predictive value of this pattern of laboratory abnormalities are not known, it might provide a useful tool in diagnosing HSES and excluding other illnesses. Viral illness is common in HSES. The severity of illness and mortality in this group of patients is similar to other descriptions of HSES.

Relationship of benzyl alcohol to kernicterus, intraventricular hemorrhage, and mortality in preterm infants.

Intraventricular hemorrhage and death in preterm neonates has been associated with the use of fluid containing benzyl alcohol, a bacteriostatic agent, to flush intravascular catheters. The hospital and autopsy records of infants admitted to a nursery during the last 18 months that benzyl alcohol was in use (218 patients) were reviewed and compared with those of infants admitted in the first 18 months after benzyl alcohol was withdrawn (232 patients). The volume of flush solution administered to each patient was estimated. Exposure to benzyl alcohol was significantly associated with the development of kernicterus (P less than .005), and intraventricular hemorrhage (P less than .000,000,5). Kernicterus did not develop in any patient after benzyl alcohol was withdrawn. Many patients with kernicterus or intraventricular hemorrhages received small daily volumes of fluid containing benzyl alcohol. Withdrawal of benzyl alcohol from clinical use had no demonstrable effect on mortality. Medications intended for neonatal use should not contain benzyl alcohol. Our data indicate that patients not exposed to benzyl alcohol have a greatly reduced risk of kernicterus. If this finding is confirmed by other investigators, present indications for exchange transfusions in preterm infants with moderate elevations of serum bilirubin should be reconsidered.

Legionellosis in children with leukemia in relapse.

Two children with legionellosis complicating a relapse of acute lymphoblastic leukemia are reported. A 5-year-old boy with pneumonia had Legionella pneumophila cultured from a tracheal aspirate following a rapid deterioration in his respiratory status and intubation. This child had severe and irreversible granulocytopenia and died in spite of therapy with erythromycin and rifampin added five days later. Combination antimicrobial therapy is suggested for immunosuppressed children with legionellosis if resolution of neutropenia is not readily anticipated. Culture of Legionella sp from respiratory tract secretions or sputum, as reported for the first time in the pediatric literature, should be attempted in all children in whom this infection is suspected. A 13-year-old boy with pneumonia recovered in spite of therapy with antimicrobial agents not proven to be effective against the legionellae. Clinical improvement coincided with increase in absolute granulocyte count. A retrospective diagnosis was made when seroconversion to Legionella micdadei (less than 1:16 to 1:1,024) was determined during a survey of unselected sera from 255 hospitalized children. This is the first documented case of Pittsburgh pneumonia described in a child.

Toxoplasmosis: a treatable neurologic disease in the immunologically compromised patient.

A 10-year-old girl with aplastic anemia developed seizures and a mild hemiparesis following a bone marrow transplant. Based on serologic evidence and a computed tomography scan, which showed a left parietal lucency with ring enhancement, a diagnosis of toxoplasmosis was considered. A brain biopsy of the lucent area demonstrated the inflammation and necrosis but no organisms were seen. During a six-week course of pyrimethamine, sulfadiazine, and folinic acid therapy there was clinical and neuroradiologic resolution. The short course of therapy as well as the inadvertent substitution of folic acid for folinic acid and trimethoprim-sulfamethoxazole for sulfadiazine resulted in the reappearance of neurologic deficits. Reinstitution of appropriate therapy produced gradual improvement over a nine-month period. Serial computer tomography scans correlated with the clinical course. In the immunologically compromised host CNS toxoplasmosis should be considered in the differential diagnosis of an evolving CNS syndrome. Early detection and prolonged therapy with appropriate drugs can result in a favorable outcome. Computed tomography scanning may be helpful in diagnosis and follow-up.

Routine chest radiographs following repositioning of endotracheal tubes are necessary to assess correct position in pediatric patients.

Attempts to correctly reposition endotracheal tubes (ETTs) are not always successful in pediatric patients, even when chest radiographs (CXRs) are measured to determine the distance that the ETT deviates from the correct position. We determined the frequency of continued ETT malposition after repositioning in a pediatric intensive care unit (PICU). Forty children with malpositioned ETTs were identified during a 4-month period. After repositioning, ten (25 percent) continued to be malpositioned on the next CXR. Of 47 children with correctly positioned ETTs, only one ETT (2 percent) was found to be incorrectly positioned on the next routine CXR obtained 24 h later. The difference in frequency of ETT malposition between these two groups of children is significant (p < 0.0001). The children were similar in weight and age. Despite repositioning based on measurements taken from a CXR, a large percentage of pediatric patients had continued ETT malposition. However, after radiographic documentation of correct position, we demonstrated that significant movement was uncommon. Routine confirmation of ETT position by CXRs should be considered after repositioning ETTs in pediatric patients.

Serial neurologic examinations after near drowning and outcome.

OBJECTIVE: We evaluated serial neurologic examinations after warm water near drowning to determine how rapidly survivors with poor neurologic outcome could be identified. RESEARCH DESIGN: Retrospective chart review. SETTING: University-affiliated pediatric hospital. PATIENTS: Forty-four children admitted to the pediatric intensive care unit with an abnormal mental status after near drowning during a 5-year period. Follow-up was a minimum of 6 months. INTERVENTIONS: None. MEASUREMENTS: A 14-point coma scale was used to evaluate both cortical and brain-stem function at the time of hospital admission and then daily afterward. The Mann-Whitney U Test was used to compare patients grouped as having satisfactory outcomes (those who returned to their presubmersion baseline or had very mild deficits) and unsatisfactory outcomes (total custodial care or death). Significance was defined as P < .05. CONCLUSION: All satisfactory survivors were sufficiently awake to have spontaneous, purposeful movements 24 hours after near drowning and had normal brain-stem function. All children without spontaneous, purposeful movements and normal brain-stem function 24 hours after near drowning suffered severe neurologic deficits or death. In this retrospective investigation of 44 children, the cortical examination 24 hours after warm water near drowning distinguished satisfactory survivors from children who required total custodial care or died.

A mathematical model of life-threatening hyperthermia during infancy.

A mathematical model was created to test the hypothesis that a partially covered febrile infant may develop potentially lethal temperature elevation. Infants may be at special risk to develop hyperthermia because, unlike older children, infants may not be able to remove blankets in response to temperature elevation. The model compared heat production (MTsk) with heat loss (Qtot). The difference between these terms is the excess energy (E): MTsk - Qtot = E. In most situations the simulated infant transfers heat to the environment as rapidly as it is produced (E less than 0), so hyperthermia does not result. In some situations, heat production exceeds heat loss (E greater than 0), causing progressive warming. The time was calculated for the simulated infant to progress from 41 to 43.4 degrees C (defined as a lethal end point). In certain circumstances, this may occur in less than 90 min. An infant at high risk of hyperthermia may not appear to be covered by a conspicuous excess of insulation (less than or equal to 3.5 cm may be sufficient). In many situations, heat loss is more closely determined by exposed body surface area than by blanket thickness. These findings have important implications for understanding the antecedents of hyperthermia in infants and may help in understanding the role of hyperthermia in certain pediatric illnesses.

An animal model of life-threatening hyperthermia during infancy.

A mathematical model of heat balance in human infants suggests that it may be possible for severe hyperthermia to develop if an infant is unable to remove his blankets in response to overheating (thermal entrapment). This hypothesis was tested in an animal model of weanling piglets. Ten piglets were warmed in a radiant heater to rectal temperature of 41 degrees C to simulate a fever. Animals in the experimental and control groups were removed from the heater and covered with ordinary infant blankets (to a thickness of approximately 3 cm). Endogenously produced heat caused the animals to warm to 42 degrees C. At this point, the control animals were uncovered. They rapidly cooled to normal body temperature. Animals in the experimental group remained covered until they expired from hyperthermia at 43.9 +/- 0.7 degrees C (SD) after 96 +/- 43 (SD) min. These data show that lethal hyperthermia may result from thermal entrapment. This finding may help clarify the role that hyperthermia may play in illnesses such as hemorrhagic shock and encephalopathy syndrome and some cases of sudden infant death syndrome.

Nonpulmonary organ failure and outcome in children treated with high-frequency oscillatory ventilation.

PURPOSE: The purpose of this study was to quantitate the contribution of nonpulmonary organ failure to mortality of patients treated with high-frequency oscillatory ventilation (HFOV) and to determine which gas-exchange differences are associated with improvement on HFOV. MATERIALS AND METHODS: Charts of all patients treated with HFOV in our pediatric intensive care unit from January 1992 until January 1997 were retrospectively reviewed. RESULTS: Sixty-six patients were treated and 21 patients improved during HFOV (group 1); 45 patients did not improve (group 2). Seventeen patients (26%) had isolated respiratory failure and their mortality was 12%. Percentages of patients with 2, and 3 or more organ failure were 45%, 29%, and their mortality was significantly higher, 67% and 95%, respectively. Patients with primary respiratory failure demonstrated a significantly greater risk of improvement on HFOV (RR ratio of 2.5, 95% CI 1.5 to 4.2). There was a significantly greater proportion of patients with primary cardiac failure who did not improve on HFOV compared with all other patients. Oxygenation index significantly improved over the first 72 hours for both groups, but then significantly worsened over the next 48 hours in group 2 but not in group 1. CONCLUSION: Patients with nonpulmonary organ failure were significantly less likely to improve on HFOV and had a significantly higher mortality than patients with isolated respiratory failure. Children who do not improve on HFOV appear to reach a plateau in oxygenation indices after 3 days of HFOV.

CT scan abnormalities in a series of patients with hemorrhagic shock and encephalopathy syndrome.

BACKGROUND: Hemorrhagic shock and encephalopathy syndrome (HSES) affects children under 1 year of age and is characterized by seizures, shock and certain laboratory abnormalities, including coagulation abnormalities. It has a high mortality and many of the survivors are neurologically abnormal. OBJECTIVE: To describe abnormalities observed on initial and follow-up CT scans in a group of patients suffering from HSES. MATERIALS AND METHODS: Retrospective review of records and CT scans of ten patients with HSES who were admitted to the intensive care unit of the Children's Hospital and Medical Center, Seattle. RESULTS: Cerebral edema was seen in all cases when the CT scan was obtained between 1 and 7 days after onset of HSES. The basal ganglia and cerebellum were relatively spared, and no hemorrhage was seen. Patients with moderate or marked cerebral edema usually had a poor prognosis. All survivors had significant neurologic sequelae. CT scans obtained after 7 days often showed encephalomalacia with ex vacuo ventricular enlargement. CT scans obtained between 24 h and 4 days after onset will show the acute changes of HSES. CT scans during the initial and convalescent stages of HSES can provide useful information about cerebral edema and encephalomalacia, which occur frequently with this illness.

Constitutive synthesis of heat shock protein (72 kD) in human peripheral blood mononuclear cells: implications for use as a clinical test of recent thermal stress.

There is no clinical laboratory marker to enhance the diagnosis of recent thermal stress in humans. The 72 kD heat shock protein, HSP 72, which is rapidly synthesized after heat stress could be useful in the diagnosis of illnesses associated with heat stress. In humans HSP, 72 is rapidly synthesized after thermal stress; however, conflicting data suggest it may also undergo low level constitutive synthesis. If HSP 72 is constitutively synthesized, a semi-quantitative test will be necessary to detect recent heat stress; if not, a qualitative test would be sufficient, peripheral blood mononuclear cells were chosen for this investigation because they can be isolated from a small sample (clinically acceptable) of blood. Following heat stress Western analysis and autoradiography of one- and two-dimensional electrophoresis samples demonstrated low levels of HSP 72 in unstressed cells. HSP 72 increased with heat stress, and remained elevated for up to 48 h. HSP 72 mRNA was detectable in small amounts in nonheat stressed cells. Heat stress increased HSP 72 mRNA 1 and 2 h after stress and remained elevated for 6 h. HSP 72 persists long enough to be potentially useful as a diagnostic probe of recent heat injury; however, a semi-quantitative assay will be necessary.

Pediatric intracranial pressure monitoring in hypoxic and nonhypoxic brain injury.

We reviewed the results of all pediatric patients undergoing intracranial pressure (ICP) monitoring in a 2-year period at our institution. The outcome of patients suffering hypoxia or ischemic injuries (HII) is compared to those suffering non-hypoxic or non-ischemic injuries (NHII). Thirty-four patients had ICP monitors placed during the study period. Incomplete patient information led to the exclusion of 5 patients. An additional 5 patients were excluded because no measures to control ICP were taken after the monitor was placed. Twenty-four patients required treatment for raised ICP (hyperventilation, 24; mannitol, 19; barbiturate coma, 6). Admission Glasgow Coma Score in patients suffering HII (median score 5) and NHII (median score 6) were not significantly different (Mann-Whitney U Test). Only 2 of 8 patients with HII were near-drowning victims. The remaining 6 had HII from other causes (5 survivors of various forms of asphyxia and 1 of cardiac arrest). All 8 patients had poor outcomes (1 severely disabled; 7 died). The 16 patients with NHII had a variety of diagnoses (6 trauma, 5 encephalitis, 4 bacterial meningitis, 1 diabetic ketoacidosis). Among these, 6 had good outcomes and 10 poor outcomes (2 severely disabled, 2 vegetative, and 6 died). The difference in outcome between patients with NHII and HII is significant at P = 0.059 (Fischer Exact test). Patients with NHII may benefit from ICP monitoring. Patients with HII from near-drowning and other causes did not appear to benefit from ICP monitoring and interventions directed at controlling ICP.