Postoperative Intensive Care Management in Children

Neurological Care

Neurological events are an important cause of morbidity and mortality in the early postoperative period; a significant proportion of deaths occur in the first several postoperative days as a result of cerebral edema and herniation. In one study, all postoperative deaths were found to be associated with neuropathological findings on postmortem examination, thus suggesting that neurological alterations were present even in patients not dying of a neurological cause. Initial postoperative assessment after transplantation requires knowledge of both the patient’s preoperative neurological condition and the presence of any drugs that might alter the neurological status of the patient. Medications given during the surgical procedure will have been reported by the anesthesiologist, but it may be difficult to predict the duration of action of some of these agents in the altered physiological state of the patient. If the patient is not moving spontaneously, the presence of neuromuscular blocking agents must be assessed with the use of a nerve stimulator. The presence of all four responses to train-of-four stimulation indicates that less than 75% of the receptors are blocked and that neuromuscular blocking agents are not significantly affecting clinical assessment. The pharmacodynamics of analgesic agents is also altered in transplantation patients, and a more prolonged effect may be seen. Nevertheless, a patient without significant preoperative liver encephalopathy should be waking within a few hours of returning to the ICU.

Assessment of the patient’s neurological status must include the level of consciousness, cranial nerve function, motor function, sensation, and reflexes. New cranial nerve findings, abnormal posturing, or significant asymmetry in the examination mandates an urgent computed tomographic scan of the brain to evaluate the possibility of intracranial hemorrhage. The computed tomographic scan must also be evaluated for diffuse findings consistent with increased intracranial pressure (ICP), such as loss of the sulci cerebri and interhemispheric and sylvian fissures and diminution in the body of the lateral ventricle. A further rise in ICP results in complete loss of the sulci and fissures, as well as loss of the chiasmatic, quadrigeminal, and interpeduncular cisterns. Finally, with severe increases in ICP, there will be complete loss of all perimesencephalic cisterns.

If computed tomography reveals the presence of central nervous system hemorrhage, neurosurgical consultation and correction of any underlying coagulopathy are required. If increased ICP is present, as manifested by examination or computed tomographic scan, consideration should be given to ICP monitoring. Clinical findings of increased ICP include decreased mental status, increased muscle tone and deep tendon reflexes, hyperventilation, and dilation of pupils with a sluggish response to light. With a further increase the patient postures, initially decorticate and then decerebrate, and later the pupils become fixed and dilated from compression of the third cranial nerves.

Several options are available to monitor ICP. Blei et al performed a survey to examine complications of ICP monitoring in liver transplantation candidates (262 patients). The use of epidural monitors as opposed to subdural or parenchymal monitoring appears to have a lower complication rate, with an infection rate of less than 1% and a 3% incidence of hemorrhage. Intraparenchymal monitors have been associated with a 13% complication rate from hemorrhage and a 4% infection rate, whereas subdural monitors have had an 18% incidence of hemorrhage with 5% of monitored patients dying of intracranial bleeding. A more recent study involving subdural ICP monitors in similar patients reported only 3 episodes of subdural bleeding in 161 patients and only 1 death from intracranial hemorrhage. The recent use of recombinant activated factor VII immediately before insertion of intraparenchymal ICP monitors has demonstrated a further diminution in intracranial bleeding complications. Management of increased ICP should be directed at maintaining adequate cerebral perfusion pressure, which is the difference between mean arterial pressure and ICP.

The absolute value of adequate cerebral perfusion pressure varies from patient to patient but appears to be the value at which ICP spikes are minimized and below which ICP increases. Attainment of this goal often requires maintenance of mean arterial pressure above normal for age, which many postoperative patients maintain spontaneously. Intravascular volume must be adequately preserved, and hypovolemia should not be allowed to occur because it will adversely affect cerebral perfusion pressure. Thus a hypovolemic or euvolemic patient should not be fluid restricted. The use of hyperventilation in patients with increased ICP is controversial. Although hyperventilation does initially lower ICP, its effect is relatively transitory and may make segments of the brain ischemic and thus result in further brain injury and swelling. Most centers have discontinued the use of hyperventilation for raised ICP and maintain the partial pressure of arterial carbon dioxide between 35 and 40 mm Hg. Mannitol may be useful for treating increased ICP by decreasing blood viscosity and the water content of the brain. Care must be exercised, however, because posttransplantation patients are often hyperosmolar initially, and mannitol should not be used if osmolality is greater than 315 mOsm/kg. An alternative means of raising serum osmolality is to use hypertonic saline to raise the serum sodium level, and this technique is increasingly being used to treat raised ICP. A recent randomized controlled study in adults with acute liver failure and grade 3 or 4 encephalopathy demonstrated decreased intracranial hypertension in the group in whom hypertonic saline was given to maintain serum sodium at 145 to 155 mmol/L. Moderate hypothermia with temperatures as low as 32° C (89.6° F) has been used with some evidence of efficacy in patients with acute liver failure and intracranial hypertension unresponsive to other medical therapies.

For a patient not under the effect of neuromuscular blockade, analgesic and sedative medications should be withheld until the patient begins to wake because the duration of intraoperative analgesic and sedative agents may be quite prolonged. When the patient wakes, narcotics should be administered to provide analgesia; intermittent dosing of morphine sulfate is usually adequate, although the dosing will be highly variable and titration will be necessary. If hypotension is a concern, fentanyl may be advantageous. Orthotopic liver transplant recipients typically require less analgesia than comparable hepatic resection patients do. Although benzodiazepines are frequently used as sedatives in the pediatric ICU, caution must be exercised in their use in postoperative transplantation patients because those with liver encephalopathy have an endogenous substance with benzodiazepine activity and use of the benzodiazepine antagonist flumazenil has been shown to improve liver encephalopathy transiently. For this reason, benzodiazepines should be avoided in an encephalopathic patient. Dexmedetomidine is a selective α- _2a adrenergic agonist that offers potent sedative effects and some analgesia. It has the advantage of not being associated with respiratory depression and may facilitate earlier extubation.

Seizures occur in approximately 8% of children after liver transplantation. Although hypoglycemia is an unusual cause of early posttransplantation seizures because patients are nearly always hyperglycemic, the glucose level must be assessed immediately and treated if low. Electrolytes, including magnesium, should be measured and abnormalities corrected. Hypomagnesemia was previously a significant cause of seizures. Seizures often correlate with high tacrolimus or cyclosporine levels, and thus levels should be monitored and the dosage reduced if possible. Any seizure without a discernible cause and all focal seizures necessitate a computed tomographic scan to evaluate the possibility of central nervous system bleeding. The seizure episode itself may be treated with lorazepam initially if the patient is not encephalopathic. The optimal anticonvulsant for maintenance therapy in this patient population is unclear.

Levetiracetam has the advantage of not needing dose adjustment in hepatic failure and not interfering with the metabolism of commonly used immunosuppressive agents. In the absence of structural lesions on the computed tomographic scan and with no subsequent seizures, anticonvulsant therapy may be discontinued 2 weeks after the seizure. Structural lesions require a longer duration of treatment.

Respiratory Care

Nearly all liver transplantation patients return to the ICU endotracheally intubated and maintained on positive pressure ventilation. Occasionally an older child who was previously in excellent health other than the underlying liver disease may be extubated in the operating room, but such is the exception rather than the rule. All intubated patients are continuously observed via end-tidal carbon dioxide monitoring, pulse oximetry, and intermittent arterial blood gas analysis, in addition to the standard ventilator monitoring and alarm systems.

No evidence supports the superiority of any specific mode of ventilation in children. If there is a significant leak around the endotracheal tube, which is common with the uncuffed tubes often used in children, a pressure-limited mode of ventilation often provides more consistent tidal volumes. Either an assist control or a synchronized intermittent mandatory ventilation mode may be used. Compliance should be monitored closely because it may be significantly altered by changes in abdominal distention.

The ventilatory rate is chosen to provide adequate minute ventilation as manifested by the partial pressure of carbon dioxide, with the initial set rate approximating the patient’s physiological respiratory rate. To increase the small functional residual capacity commonly seen with these patients’ abdominal distention, positive end-expiratory pressure (PEEP) of 4 cm H ₂ O is initially used. PEEP is titrated upward if adequate oxygenation cannot be achieved at a fractional inspired oxygen concentration (FIO ₂ ) of 0.60 or less. The FIO ₂ is initially chosen to provide an oxygen saturation as measured by pulse oximetry (SpO ₂ ) of greater than 95%, with the FIO ₂ being weaned as tolerated.

In a relatively stable postoperative patient it is typically possible to extubate the patient within the first 12 hours. Several factors may make such extubation impossible in selected patients ( Table 70-1 ). Persistent paralysis or a depressed respiratory drive secondary to encephalopathy or excessive sedation precludes rapid extubation. Reversal of paralytic agents and narcotics is not usually required, and these agents are allowed to wear off as they are metabolized by the body. Fluid overload may delay extubation as a result of both pulmonary edema and pleural effusions, which are typically on the right. The effusion is generally a transudate through the traumatized diaphragm from ascitic fluid in the abdomen. Aggressive diuresis is usually efficacious for both pulmonary edema and pleural effusions. Placement of chest tubes is avoided in view of the often dilated chest wall vasculature and the coagulopathy generally present in the immediate postoperative period.

The metabolic alkalosis commonly seen in the immediate postoperative period leads to a compensatory respiratory acidosis. Although the resultant high partial pressure of carbon dioxide is not in itself a contraindication to extubation, the hypoventilation often causes or exacerbates atelectasis and results in an increasing oxygen requirement, which makes successful extubation less likely. For this reason, a base excess greater than +7 is often treated with acetazolamide. Occasionally, with severe metabolic alkalosis, 0.2 N hydrochloric acid is initiated at 0.5 mL/kg/hr and titrated as necessary on the basis of serial serum bicarbonate levels. Other metabolic abnormalities such as hypophosphatemia, hypomagnesemia, hypocalcemia, and hypokalemia may lead to respiratory muscle dysfunction and an inability to wean from the ventilator.

A small infant often presents a challenge to wean from the ventilator. These patients have variable and fluctuating degrees of abdominal distention from ascites, edematous bowel, and large liver grafts. This combination results in a decrease in effective pulmonary compliance and highly variable delivered tidal volumes with the pressure-limited ventilation typically used. Peak inspiratory pressure may need to be frequently altered, and successful extubation may need to await improvement of the abdominal distention. In an analysis of infants transplanted at less than 90 days of age from the Studies of Pediatric Liver Transplantation (SPLIT) database, the mean duration of mechanical ventilation was over 16 days.

An infant who is otherwise doing well but fails to wean from the ventilator probably has either abnormal mechanics of breathing or an infection. In infants with chronic ascites and abdominal distention, a short, broad chest develops, with decreased diaphragmatic motion often leading to atelectasis. Aggressive chest physiotherapy may be helpful to open collapsed areas of lung, but bronchoscopy may be necessary to reexpand the atelectatic lung if the patient is unable to wean from the ventilator. Diaphragmatic paralysis from phrenic nerve injury or direct injury to the diaphragm itself will delay extubation in a small infant. Diaphragmatic motion under conditions of spontaneous respiratory effort needs to be examined with either ultrasonography or fluoroscopy to look for paradoxical diaphragmatic movement. Rarely, plication of the diaphragm to prevent paradoxical motion is necessary to facilitate successful extubation. Another cause of failure to wean from the ventilator is an inappropriately small endotracheal tube, which significantly increases resistance to airflow and thus increases the work of breathing. This problem requires either changing to a larger endotracheal tube or extubating from higher levels of ventilatory support.

A very ill, malnourished infant may not have adequate strength to wean from the ventilator, and a period of ventilatory support followed by a gradual defined weaning program will thus be required. To constantly try to wean such a patient before improvement in nutritional status only exhausts the patient, uses precious calories, and delays extubation. Increasing periods of “sprinting” are often used to wean such patients, although there is no documentation of the superiority of such a weaning technique over a gradual reduction in the intermittent mandatory ventilation rate. Nutritional support should provide sufficient, but not excessive, calories with adequate noncarbohydrate calories to avoid an increased respiratory quotient with the resulting high carbon dioxide load. A small child who continues to require ventilatory support despite correction of these factors may have a pulmonary infection. Infants with an increased alveolar-arterial oxygen gradient, fever, diffuse interstitial infiltrates on chest roentgenograms, and inflammatory cells on sputum Gram stain in the absence of a bacterial process often have a viral pneumonitis. Viral cultures and antigen detection studies should be performed on an endotracheal sputum sample. Management is generally limited to good supportive care, minimization of immunosuppressive agents as tolerated, and time. Symptomatic adenoviral pneumonia has a significant mortality rate in this population.

Reliable guidelines for the timing of extubation in pediatric patients do not exist. Application of the rapid, shallow breathing index, which is useful in the adult population, would incorrectly predict failure of a normal infant to tolerate extubation. A maximum negative inspiratory force greater than 45 cm H ₂ O and a crying vital capacity greater than 15 mL/kg were predictive of successful discontinuation of positive pressure ventilation in one group of infants undergoing ventilation postoperatively. Randolph et al compared a pressure support method of ventilatory weaning with a volume support method and found no difference. Of note is the readiness-to-extubate criteria used for that comparison trial. This extubation readiness test consisted of a maximal FIO ₂ of 0.50, maximal PEEP of 5 cm H ₂ O, and if the patient then maintained an Sp o ₂ of 95% or higher, changing the ventilatory mode to pressure support ventilation at a minimal level based on endotracheal tube size. If the patient tolerated a 2-hour trial with acceptable SpO ₂ , tidal volume, and respiratory rate, extubation was performed. In a recent review, Newth et al advocated using an extubation readiness test using 2 hours of spontaneous breathing on continuous positive airway pressure of less than 5 cm H ₂ O or a T-piece.

Cardiovascular Care

All patients are monitored with an electrocardiogram, invasive and noninvasive blood pressure devices, and central venous pressure recordings. Occasionally a patient requires placement of a pulmonary artery catheter for an ongoing shock state, often sepsis, but this is unusual. No catecholamines are systematically used. A minimum systolic blood pressure of 90 to 100 mm Hg is maintained in all patients in the immediate postoperative period for adequate graft perfusion; fluid support and occasionally inotropic agents are used to maintain these perfusion pressures. Vasopressor agents are avoided and used only if the patient remains in shock despite vigorous fluid support and is inappropriately vasodilated. These patients are invariably septic and are treated in the same manner as other septic patients. The typical postoperative patient is not hypotensive but is actually hypertensive; more than 70% of pediatric patients require therapeutic intervention for their hypertension. The cause of the hypertension appears to be multifactorial, with fluid overload, cyclosporine, tacrolimus, steroids, and elevated renin levels implicated. Attention to adequate analgesia and appropriate diuresis for hypervolemia is sufficient therapy for many patients who are hypertensive. In the absence of clinically significant bleeding and without severe coagulopathy, some degree of moderate hypertension is tolerated. Systolic pressures up to 140 mm Hg and diastolic pressures up to 90 mm Hg are not treated in all but the smallest patients for the first several days postoperatively. If other pharmacological intervention is deemed necessary, nifedipine in a low dose may be used. In an unstable patient a nicardipine infusion may be titrated to optimize blood pressure control. One should err on the side of underdosing initially and increase the dosage as necessary to minimize the chance of acutely making the patient hypotensive and perhaps increasing the risk for hepatic artery thrombosis.