A 23-day-old female neonate presented to the emergency department lethargic and mottled. The neonate had been in her usual state of health until 18 hours prior to presentation when she had an episode of nonbloody, nonbilious vomiting. There were no additional episodes of vomiting the following morning and afternoon. However, the neonate took only 4 to 5 ounces of formula during this time. As the day progressed, the neonate grew increasingly lethargic and had a weak cry and respiratory distress. The mother denied any increased irritability, fever, diarrhea, or bloody stools.
The neonate had been delivered via cesarean section (secondary to failure of the pregnancy to progress) by a 20-year-old gravida 1 para 1 woman with no history of prenatal care. The neonate weighed 7 pounds, 12 ounces at birth. She had an unremarkable course in the newborn nursery. At her 2-week follow-up, she weighed 9 pounds, 14 ounces.
Maternal history included chlamydia that was medically treated shortly before the pregnancy, but no test of the cure was documented. The mother was negative for group B streptococci at delivery. She had a postpartum fever and was treated for 5 days with intravenous antibiotics.
On presentation to the emergency department, the neonate was immediately taken to the resuscitation room. Her vital signs included a core temperature of 35.8°C, a heart rate of 255 beats per minute, a respiratory rate of 60 breams per minute, an unobtainable blood pressure, and a glucose level of 16 mg/dL. She weighed 7 pounds, 12 ounces. Supplementary oxygen was administered via a bag valve mask and intravenous access was obtained in the right upper extremity. A 16-cc 10% bolus of dextrose and three 20-cc/kg boluses of normal saline solution were administered. An ice bag over the patient's face failed to alter her heart rate. Central access was obtained via the left femoral vein to facilitate fluid resuscitation and administration of medicine. Intravenous tobramycin and ampicillin were administered empirically. Maintenance fluids were started with 10% dextrose, V* normal saline solution at 17 cc/h. Eight milliequivalents of half-strength sodium bicarbonate was also adininistered.
A physical examination revealed a lethargic female neonate with a sunken anterior fontanelle, dry mucous membranes, and a supple neck. The respiratory effort was shallow and rapid with intercostal and subcostal retractions. Coarse breath sounds with monchi and rales were noted on pulmonary examination. The heart rate was extremely rapid and difficult to auscultate. The abdomen was nontender and nondistended with a palpable liver edge. Femoral pulses were bilaterally thready and almost imperceptible. The distal extremities were cool and mottled. There were no rashes or petechiae noted.
Once respiratory, hemodynamic, and metabolic supports were in place, the etiology was entertained. Initial thoughts revolved largely around severe dehydration and hypoglycemia secondary to bacterial or viral sepsis.
Before the full differential diagnosis is considered for the neonate who appears septic, it is important to address immediate life-threatening issues. Addressing the airway, breathing, and circulatory issues and the critically low dextrose stick reading was imperative for this patient despite an incomplete understanding of the etiology of her condition. Signs of early sepsis in the neonate can be exceedingly subtle and require low thresholds for detection. Lrritability, vomiting, and excessive somnolence (especially failure to wake for feedings) may all be harbingers of invasive bacterial disease. Fever, for all the reliance placed on it as an indicator of disease, is often absent. In fact, the neonate is more likely to be hypothermic in the face of sepsis.
Although sepsis is by far the most likely etiology in a neonate presenting with this degree of illness, the physician must be able to entertain and systematically rule out a wide variety of other life-threatening conditions that cross several different organ systems. They may be divided into the following categories: (1) neurologic or child abuse; (2) cardiac; (3) gastrointestinal; (4) renal; (5) hematologic; and (6) metabolic, endocrine, or toxic.
Neurologic conditions mimicking sepsis are intracranial processes, frequently the result of child abuse or a malfunctioning ventriculoperitoneal shunt. Child abuse remains disturbingly common in our society and must be constantly entertained as an etiology for septic appearance. Evidence of cutaneous injury or old fractures on chest x-rays are the simplest clues one can use early in the evaluation. Noncontrast computed tomography remains the gold standard for evaluation of acute intracranial injury and mass effect.
Ventriculoperitoneal shunts are commonly placed for several conditions causing hydrocephalus in early life. The mere existence of the shunt requires its evaluation and élimination as the cause of septic appearance. Shunt malfunctions may present f\ilnvinantly or, more likely, subtly. It is vital to ascertain the existence of a shunt, the date of placement, and the last revision, if any, during the history. Shunt infections are highly unlikely after a shunt has been in place for 6 months. Malfunctions, however, may occur at any time and must be evaluated by computed tomography scan of the head to detennine ventricular size and shunt series and to examine the entire length of the tubing for kinks or fractures. On examination it is possible to palpate the shunt beneath the skin and pump the bubble beneath the scalp. A bubble that is difficult to pump is a clue to a malfunctioning shunt.
Infection by Clostridium botulinum represents the final neurologic entity commonly able to masquerade as sepsis in the neonate. Infantile botulism often has an insidious onset with poor suck (and therefore poor feeding), a weak cry, and sometimes constipation. These patients are usually afebrile. Risk factors for the disease include living in a rural area, ingesting honey, breastfeeding, and recently changing feeding practices. Clostridium spores reside in the soil and may be released into the air at construction sites, thus increasing the risk for the disease among neonates üving in homes near new development in an endemic area.
There are many cardiac causes of septic appearance in the infant and the neonate. They include disturbances in blood flow (eg, coarctation of the aorta, hypoplastic left heart syndrome, atrial septal defects, and ventricular septal defects) and electrophysiologic disturbances in cardiac rhythm (eg, supraventricular tachycardia and congenital heart block). In the neonate, disturbances of flow may be divided into ductaldependent and nonductal-dependent lesions. Congenital lesions necessitating a patent ductus arteriosus include aortic arch interruption, hypoplastic left heart syndrome, critical pulmonary stenosis, pulmonary atresia (without ventricular septal defects), and tricuspid atresia. These lesions depend on ductal flow to ensure adequate pulmonary or systemic circulation. The ductus tends to close on day 4 of life and may lead to the rapid onset of congestive heart failure. Therefore, the differential diagnosis oí any neonate who appears septic with an acute change in appearance and hemodynamics on approximately the fourth day of life must include a ductal-dependent congenital lesion.
Finally, infants with ductal-dependent lesions whose ducts have recently closed may not present with overt signs of congestive heart failure such as loud murmurs, organomegaly, or pulmonary congestion manifested as rales. Hypoplastic left heart syndrome is notorious for this; the diagnosis must be considered based on the timing of symptoms and the general clinical appearance. Treatment for these patients includes standard measures for sepsis, but also mandates the administration of prostaglandin as a temporizing measure to maintain the patency of the patent ductus arteriosus.
Beyond the first week of life, congenital lesions may also become significant up to 6 months of age or older and are either cyanotic or acyanotic. These abnormalities can include ventricular and atrial septal defects, coarctations of the aorta, atrioventricular canal malformations, coronary artery anomalies, and other complex lesions.1,2
Disturbances of rhythm include congenital heart block and supraventricular tachycardia. Supraventricular tachycardia is more common and may present insidiously as episodes of poor feeding and lethargy as the patient enters and exits the rhythm. Fever may be a precipitant of the rhythm and may further complicate the clinical picture. Supraventricular tachycardia encompasses several dysrhythmias ranging from reentrant rhythms to increased automaticity within certain atrial tissues. Wolff-Parkinson-White syndrome involves a reentrant mechanism by a concealed accessory pathway distinct from the atrioventricular node, which predisposes the patient to supraventricular tachycardia.1
Gastrointestinal causes of septic appearance involve both surgical conditions and medical disease. Volvulus secondary to malrotation represents one of the single most life-threatening surgical conditions in the first month of life. Any neonate presenting with a history of bilious vomiting must initially be considered to have maiiotation until proved otherwise by a fluoroscopic study of the upper gastrointestinal tract extending beyond the ligament of Treitz. Although the "double bubble" sign on an upright plain abdominal radiograph is helpful for diagnosing malrotation, clinical suspicion is the most important tool. Occasionally, when a complete volvulus has occurred, there will be only rninimal air remaining in the abdominal cavity, giving a nonspecific, white, ground-glass appearance to the film.
Pyloric stenosis represents another surgical condition in which patients may present with lethargy and septic appearance. It is most common in first-born males between the ages of 5 and 8 weeks. These patients present with a history of projectile vomiting following reeding, after which they tend to remain hungry. As the disease progresses, they will have hypochloremic metabolic alkalosis. The characteristic "olive" (the hypertrophied pylorus muscle) palpated on physical examination is present in fewer than 25% to 50% of patients and should not be heavily relied on for clinical decision making. Rarely, a peristaltic wave can be seen passing over the epigastric region. When visualized, this can be virtually pathognomonic. It can be challenging to distinguish pyloric stenosis from either reflux or gastroenteritis. The clinical history and a willingness to check for the condition using abdominal ultrasound are sometimes the only tools the physician has for detecting this serious but ultimately repairable problem.
Intussusception is the final surgical entity that should be considered in the differential diagnosis oí die septic infant. This is a rare condition in the infant younger than 5 months and is more common in the healthy toddler. These patients may be febrile and lethargic, but have a wide array of clinical presentations. Patients with histories of colicky behavior followed by periods of apparent listlessness should be strongly considered to have intussusception. Evidence of bloody stool (ie, currant jelly stools) is sensitive for intussusception. but not specific. A high degree of suspicion is vital for detection. Patients presenting with currant jelly stools have a well-advanced disease process. The goal is to detect intussusception far in advance to prevent this extent of intra-abdominal injury.
Seemingly simple gastroenteritis may cause severe dehydration and hypoglycemia in a neonate with limited reserve and should be respected as a potent neonatal disease process. Specific gastrointestinal organisms likely to lead to sepsis include Salmonella and Shigella; bloody diarrhea is the clue in many of these patients. Seizures may be an additional clue for patients with Shigella infection.
Renal etiologies for septic appearance are generally detected within the first week of life. Posterior urethral valves, more common in male infants, may cause bladder outlet obstruction and consequent renal failure, urosepsis, or both. Approximately half of these patients escape early detection and present as late as 2 to 3 months of age. Histories may include vomiting, abdominal distention, and poor feeding and growth. Laboratory information is helpful, especially with elevated levels of blood urea nitrogen and creatinine. Abdominal ultrasound is the most useful diagnostic test.
Hematologic causes of septic appearance include anemia in its many forms, ranging from nutritional and hemolytic to disorders of hemoglobin structure. Methemoglobinemia is discussed in depth here because basic laboratory tests will not reveal its presence unless a methemoglobin level is ordered. Methemoglobinemia is a disorder of hemoglobin structure related to the ionic state of iron within the heme molecule. The natural balance between the ferric state and the ferrous state of iron is altered by enzymatic defects, overall hemoglobin structural issues, or exposure to oxidant drugs or chemicals, such as nitrates.
With the iron in the ferric state, the ability of hemoglobin to bind oxygen is limited. In many respects, methemoglobinemia should be considered in the same realm as gastrointestinal causes because it may result from dehydration caused by excessive vomiting and diarrhea in infants leading to acidosis and a consequent oxidant stress. Patients may present with cyanosis and lethargy with normal oxygen saturation levels by pulse oximetry.
Management depends on disease severity. Levels of 30% result in cyanosis; levels of 30% to 50% lead to tachycardia, dyspnea, dizziness, headache, and fatigue; and levels of 50% to 70% lead to lethargy and stupor. Levels greater than 70% are generally fatal. The primary goal of treatment is to identify an oxidant stress. In patients with methemoglobin levels of less than 30%, no specific treatment beyond addressing the oxidant stress itself should be required. For levels of more of than 30%, methylene blue (1 to 2 mg /kg in a 1% solution in saline adrninistered intravenously for 5 minutes) should be given. Failure to respond to the first treatment within 1 hour should prompt a second treatment. Failure of a second treatment may be a clue to a concurrent case of glucose-6-phosphate dehydrogenase deficiency. Ascorbic acid (500 mg orally) or an exchange transfusion may be required to gain a response.2
Metabolic, Endocrine, or Toxic
Several metabolic or endocrine disorders and toxin exposures may present as sepsis. Congenital adrenal hypoplasia may present with a history of vomiting, diarrhea, and lethargy. The physical examination usually reveals evidence of ambiguous genitalia. Laboratory evaluation is also helpful when there is evidence of hyperkalemia with concurrent hyponatremia. Hypoglycemia is another clue, although it is somewhat nonspecific. Dehydration taken to extremes may cause electrolyte disturbances, which cause a septic appearance (ie, hypernatremic dehydration). At the other extreme, hyponatremia caused by water intoxication (from improperly mixed formula concentrates) may explain the septic appearance of an infant presenting with a history of seizures. Some infants with hyponatremia should be evaluated for cystic fibrosis. A history of poor feeding and growth, prolonged neonatal jaundice, meconium plug syndrome, repeated pulmonary infections, and hypoelectrolytemia in the absence of gastroenteritis should prompt a sweat test.
Toxins may also cause metabolic disturbances that mimic sepsis. Infants may have inappropriate medicines administered by well-meaning parents. These can include salicylates or cough preparations containing alcohol. Carbon monoxide is a common but underrecognized household toxin that may be responsible for septic appearance in the face of hypoxia. Generally, one would expect other members of the household (including pets) to be affected, but this is not always the case. This cause must be sought by obtaining a carboxyhemoglobin level. Treatment will vary depending on the level detected and may include hyperbaric oxygen therapy.
An electrocardiogram was obtained for our patient. With little improvement in her hemodynamic status, adenosine was administered in a dose of 0.1 mg /kg by rapid intravenous perfusion without effect. This was followed by several subsequent doses ranging from 0.1 to 0.2 mg /kg, all given via the femoral line. Modest short-lived conversions to sinus rhythm were noted following a dose of 0.8 mg /kg. Not until 0.3 mg /kg of adenosine was given did the patient convert to sinus rhythm for a prolonged period of time.
The patient was endotracheally intubated and transferred to the neonatal intensive care unit. Digoxin was administered and a full cardiology evaluation ensued, which revealed WolffParkinson-White syndrome. The patient was extubated on hospital day 3. Evidence of cardiac failure, including hepatomegaly and pulmonary congestion on chest radiography, resolved spontaneously. Following digitalization there were no recurrences of supraventricular tachycardia. On hospital day 7, the patient was switched to propranolol. She was discharged on hospital day 11. All cultures were negative.
This case highlights the need to maintain an open mind to a wide variety of disease processes when faced with a neonate with septic appearance. Initially, given the history of poor feeding (for more than 12 hours prior to admission) and the neonate's general appearance, the diagnosis of severe dehydration and sepsis was ascertained. In reality, the poor perfusion and dehydration were due to factors other than simply poor oral intake or invasive bacterial disease. The supraventricular tachycardia in all likelihood contributed to the poor feeding, which led to gradually worsening dehydration and hypoglycemia. All infants with septic appearance have various forms of shock as the final common pathway. Therefore, it is the shock that we are treating initially and from there we work backward to elucidate its cause.
In this case, the patient was in cardiogenic shock from supraventricular tachycardia secondary to Wolff-Parkinson-White syndrome. There are several varieties of supraventricular tachycardia, including atrial tachycardia, atrioventricular nodal tachycardia, and atrioventricular reentrant tachycardia. Atrioventricular reentrant mechanisms are the most common form of supraventricular tachycardia presenting in the pediatric age group.1 The reentry may occur via either an accessory pathway within the atrioventricular node or a separate pathway anatomically distinct from the atrioventricular node. Patients with Wolfi-Parkinson-White syndrome possess an accessory pathway between the atrium and the atrioventricular node through which antegrade conduction occurs, bypassing the usual conduction delay of the atrioventricular node. This results in the ventricular pre-excitation seen on the electrocardiograms of patients with WolffParkinson-White syndrome with the characteristic delta wave. Patients with Wolff-ParkinsonWhite syndrome comprise approximately 10% to 20% of patients with supraventricular tachycardia. However, it is impossible to discern whether the patient has pre-excitation while supraventricular tachycardia is occurring. Only after a sinus rhythm has been reestablished can electrocardiogram criteria be applied.1
Many patients with supraventricular tachycardia have histories of repaired congenital heart disease lesions (eg, Ebstein's anomaly or corrected transposition) or infection, fever, and drug exposure (most likely to sympathomimetic amines from cold preparations). Half of all infant presentations are classified as idiopathic.1-2
Figure 1 . Schematic drawing of orthodromic conduction via the accessory pathway. RV = right ventricle; AV = atrioventricular; LV = left ventricle.
Figure 2. Schematic drawing of antidromic conduction via the accessory pathway. RV = right ventricle; AV = atrioventricular; LV = left ventricle.
Many patients will spontaneously convert in and out of supraventricular tachycardia, making the diagnosis difficult. Among patients who persistently remain in supraventricular tachycardia, many tolerate it well for up to 24 to 48 hours, after which time approximately half will have congestive heart failure. Characteristically, patients present with histories of poor feeding, irritability, and, eventually, lethargy.
Treatment of supraventricular tachycardia beyond the supportive measures ensuring adequate airway and breathing involves the use of adenosine in doses ranging from 0.1 to 0.2 mg/ kg by rapid intravenous perfusion. Adenosine is an endogenous nucleoside with extremely shortlived (t Vz < 6 seconds) negative chronotropic, dromotropic, and inotropic effects. Its hemodynamic effects are rmmmal, making it the drug of choice for termination of supraventricular tachycardia in which the atrioventricular node plays a role in the reentry circuit. Adenosine transiently blocks the pacemaking ability of the atrioventricular node and the sinoatrial node. Multiple doses may be used in succession. However, one should not hesitate to consider direct current cardioversion (0.5 joules synchronized cardioversion) in the unstable patient resistant to medical therapy.
There is some concern regarding the use of adenosine for patients with supraventricular tachycardia and known Wolff-Parkinson-White syndrome. Several case reports and case series point to the induction of atrial fibrillation and even ventricular tachycardia in certain patient subsets.3,4 Although there are several case reports alluding to this phenomenon, there are no truly illustrative prospective studies demonstrating a trend within acute care settings such as the emergency department.
Pathophysiologically, it appears that patients with narrow complex tachycardias are likely conducting through their accessory pathway in the more common orthodromic fashion and may safely receive adenosine in most cases (Fig. 1). Conversely, patients with wider complex tachycardias may be conducting antidromically through their accessory pathway (Fig. 2). The addition of adenosine in this instance may encourage a preponderance of atrioventricular conduction through the accessory pathway because the atrioventricular node will be blocked due to adenosine effects and may potentially induce atrial fibrillation, ventricular tachycardia, or vocal fremitus. The safe conclusion to draw from this is to have a defibrillator ready when admimstering adenosine to patients in supraventricular tachycardia with known Wolff-Parkinson-White syndrome. These complications are as serious as they are rare, but are ultimately treatable as long as they are recognized.
In conclusion, the differential diagnosis for the neonate who appears septic is wide and potentially complex. It is vital not to categorize all illappearing neonates as septic and simply requiring timely acbministration of empiric antibiotics and aggressive fluid resuscitation for successful treatment. Although these approaches are appropriate in many cases, it is important to consider a wide array of explanations for the patient's condition. The patient's history, vital signs, laboratory values, and general condition must be closely examined and frequently reevaluated to eliminate other viable explanations for a septic appearance.
1. Park MK. Pediatric Cardiology for Practitioners, 3rd ed. Baltimore: Mosby; 1996:338-341.
2. Fleisher GR, Ludwig S, eds. Textbook of Pediatric Emergency Medicine, 4th ed. Baltimore: Williams and Wilkins; 2000:565-572.
3. Mulla N, Karpawich PP. Ventricular fibrillation following therapy for SVT in a neonate with concealed WPW syndrome treated with digoxin. Pediatr Emerg Care. 1995; 11:238-239.
4. Ros SP, Fisher EA, Bell TJ. In the emergency management of SVT. Pediatr Emerg Care. 1991;7:222-223.