Accurate assessment of energy needs during critical illness is crucial for provision of appropriate nutrition support. Severity of illness, fluctuations in physiologic state, and clinical interventions such as alterations in ventilatory support are some of the factors that result in mean daily and intraday variability in energy expenditure in critically ill children.1 Consistent provision of nutrients is frequently challenging in critically ill children because of intolerance, interruption, and access to feeding route in addition to variability in energy expenditure.2,3 Individual assessment of energy requirements and provision of optimal nutrition support should be the standard of care, as both underfeeding and overfeeding have been associated with complications.4,5 However, accurate estimates of energy expenditure are difficult to predict during acute illness.6,7 Several studies have documented discrepancies in measured vs equation-estimated energy expenditure.8-12

Indirect calorimetry (IC) has emerged as the gold standard for the measurement of resting energy expenditure (REE).8,13,14 We report a case of dramatic hypermetabolism detected by IC during dysautonomic storms in an adolescent with anoxic brain injury. The case highlights the role of IC in energy expenditure measurements in the intensive care unit (ICU), and illustrates a challenging complication of brain injury and the importance of using actual measured energy expenditure in children with unexplained weight loss.

Case Report

B.H., a 1 4-year-old African American girl, awoke with chest pain and shortness of breath during an acute exacerbation of asthma. She subsequently collapsed and was found to be unresponsive by her mother, who called the emergency medical services (EMS) and initiated cardiopulmonary resuscitation. Upon arrival approximately 20 minutes later, EMS found the patient pulseless with no respiratory effort. She required prolonged resuscitation, including intubation for mechanical ventilatory support, and was transported to the emergency department by ambulance. At the emergency department, she was noted to have transient cardiac rhythm disturbances with stable blood pressure but minimal neurologic response. She was transferred to the multidisciplinary pediatric ICU for further management. She was noted to have profound neurologic impairment and spastic quadriparesis and was nonverbal with poor response to stimuli. An electroencephalogram showed severe diffuse encephalopathy, likely resulting from ischemic brain damage.

She was gradually weaned from mechanical ventilatory support and was hemodynamically stable a week after hospitalization, when she developed episodes of dysautonomia in the form of paroxysmal hypertension, tachycardia, fever, agitation, diaphoresis, limb rigidity, and abnormal posturing. These episodes recurred up to 10 times daily, each lasting from 15 to 90 minutes. Despite trials with multiple medications including antiseizure drugs, atenolol, Clonidine, hydralazine, morphine, fentanyl, methadone, bromocriptine, and intrathecal baclofen, the paroxysmal autonomic storms continued intermittently for 8 weeks. A tracheostomy tube was placed surgically for airway protection.

Nutrition parameters on admission included a weight of 79 kg and height of 157 cm. Her body mass index of 32 kg/m^sup 2^ was at the 98th percentile for age (Centers for Disease Control and Prevention National Center for Health Statistics 2000 growth charts). Enteral feeding was initiated via a nasojejunal tube using an intact protein formula at a goal rate of 80 mL per hour for 24 hours to provide 1920 kcal and 77 g of protein per day. This was estimated to provide approximately 140% of REE, estimated by the Schofield formula, for her adjusted body weight of 55.8 kg.15 It is currently our practice to use adjusted body weight when calculating energy requirements in obese subjects. Although the data are limited, calculations of energy requirements using adjusted body weight have been shown to be accurate.16 However, during the next 8 weeks of her ICU stay, she was noted to have a dramatic weight loss of 20 kg. On review of her actual caloric intake, we noted that she received approximately 1760 kcal per day, largely because of interruptions in her feeding regimen for bedside and operative interventions. Despite falling short of her goal value, her caloric intake was still 125% of her estimated needs, and she had no signs of malabsorption.

Hypermetabolism due to frequent sympathetic storms and the failure to meet increased energy requirements was a potential reason for her weight loss. IC was performed to measure REE in the patient, who was now breathing humidified air spontaneously through a tracheostomy tube. A clear, plastic canopy was placed around her head and neck for IC measurement using the portable Vmax Encore IC cart (Sensormedics, Loma Linda, CA). After equilibration, the patient had one of her typical autonomic storms with sudden onset of spontaneous tachycardia, increased work of breathing, diaphoresis, and decerebrate posturing. This episode was confirmed by the bedside nurse as one of the typical autonomic storms experienced by the patient. A steady state, defined as a period of at least 5 minutes with <10% fluctuation in oxygen consumption (VO^sub 2^) and carbon dioxide production (VCO^sub 2^) and <5% fluctuation in respiratory quotient, was achieved, allowing real-time measurement of energy expenditure during this storm. Energy expenditure measurement obtained at this time was 4322 kcal/d, 309% of the predicted REE for adjusted weight, indicating extreme hypermetabolism (Table 1). The episode lasted 12 minutes and subsided after administration of IV morphine sulfate. Soon after resolution of the autonomic storm and return of vital parameters to baseline, another distinct 10-minute steady-state period was observed as IC was continued. The measured REE at this time point was nearly the same as the estimated REE and was significantly lower (P < .0001) than the previously measured REE (Figure 1). The respiratory quotient was low during both periods of time, consistent with inadequate provision of energy.

The severe weight loss in our patient during her pediatric ICU course could be multifactorial. Her caloric intake was deemed adequate based on estimations with standard equations using adjusted body weight. The use of adjusted body weight in calculations of energy expenditure has been shown to be accurate but lacks sufficient data, and the practice has been debated. However, this is unlikely to have played a major role in her weight loss, as her measured resting energy expenditure in the absence of hypermetabolic storms was more closely related to the estimation using adjusted body weight.

Since the patient had endured many days of frequent autonomic storms over the previous 8 weeks, her total energy requirements were substantially influenced by the hypermetabolism during the storms. We calculated that the amount of time she likely spent in the dysautonomic and hypermetabolic stage ranged from 150 minutes to 900 minutes per day. Based on the energy expenditure measurements by IC and the documented actual daily intake, we estimated energy deficits of approximately 1000 kcal/d during the 8 weeks of poorly controlled autonomic storms. The profound weight loss was therefore due to unexpectedly high energy requirements that could not be met despite energy intake that exceeded the estimated REE.

Surgical replacement and restoration of a malfunctioning baclofen pump finally resulted in complete resolution of her dysautonomia. Once her storms were controlled, energy expenditure was presumed normal, and weight maintenance ensued. In the subsequent 7 weeks during her stay at a rehabilitation facility, her weight remained stable on an energy intake of approximately 1920 kcal/d.

Discussion

Energy requirements in sick or hospitalized subjects are determined by the nutrition status of the patient as well as by the nature and severity of the underlying illness. IC is used intermittently to measure REE over short periods of time, with an aim to extrapolate this value to total daily REE and guide nutrient intake.2,17 Strict adherence to sound technique, stringent patient selection criteria, and the use of steady-state data help to minimize errors and eliminate the influence of artifact on the measurement.18 Stress factors are routinely added to the steady-state REE estimates in an attempt to improve their correlation with true 24-hour energy expenditure.19 However, REE can vary significantly from day to day and even within a 24hour period because of metabolic changes and therapeutic interventions.20,21 When variability in REE represents true metabolic instability, as shown in our patient with autonomic storms, isolated periods of steady-state REE cannot be extrapolated to represent total daily REE. A previously recorded REE in our patient would not have been helpful unless it captured a dysautonomic event. The number of hypermetabolic events per day influenced the total energy requirement in our patient.

Accurate estimation of energy requirements in children with brain injury is difficult, with significant differences between measured and estimated values.22 In patients with traumatic head injury, a hypermetabolic acute phase is followed by variable energy expenditure during the subsequent clinical course.23 In contrast, patients with atraumatic brain injury may have decreased energy expenditure.24 The nutrition requirements of these children may be overestimated by routinely used estimation equations, which in turn increase the incidence of overfeeding and obesity in this cohort.25 In a subgroup of patients with neurologic impairment, the clinical course may be complicated by episodes of profound hypermetabolism. Paroxysmal sympathetic storm is a well-defined but poorly understood complication of brain injury from different etiologies such as trauma, infection, near drowning, or hypoxemic-ischemic insult.26,27 Patients may present with a combination of clinical signs and symptoms including hyperthermia, tachycardia, hypertension, agitation, hyperhydrosis, and extensor or decerebrate posturing. These episodes may last for several minutes to hours, may recur multiple times during the day, and are often resistant to treatment.26,28 Heightened awareness of this complication in children with brain injury may allow prompt initiation of treatment strategies. Various treatment modalities have been tried in patients with such storms with variable success.28

In such cases, increased energy expenditure must be suspected, and energy provision must be guided by REE measurements whenever possible. Continuous measurement of REE may be feasible with recent advances in continuous gas exchange measurement technology and is desirable in this group of patients. Newer compact and portable IC devices have the ability to measure gas fractions integrated with flow signal and use complex software to provide REE with reasonable accuracy and agreement with standard IC machines.29' Careful attention to daily energy balance will prevent cumulative energy deficits, which result in weight loss and anthropometric changes during hospitalization.2

In summary, we report the case of a child with severe paroxysmal dysautonomia associated with ischemic brain injury who suffered severe weight loss in the ICU. Hypermetabolism and increased energy expenditure during these episodes was quantified with IC, as was a return to baseline energy expenditure with short-term medical therapy of the autonomic storm. Awareness of the severity of the energy deficit caused by this complication will allow prompt institution of treatment. Finally, this case illustrates the importance of regular weight monitoring during the ICU course and the need to measure energy expenditure in patients with unexpected weight loss and potential metabolic fluctuations. Unless energy requirements during the acute stage are accurately measured and matched by adequate intake, cumulative energy deficits will ensue with negative nutrition consequences. Both underfeeding and overfeeding have been associated with increased morbidity in hospitalized patients. IC allows measurement of energy expenditure and guides nutrition intake, especially in patients with wide metabolic fluctuations.