THERE may be nothing more anxiety-provoking for a physician than caring for a previously healthy infant or young child who presents in shock. Once a child's condition has progressed to this point, it can be very difficult to determine the exact cause. Shock is a common pathway for a multitude of life-threatening illnesses and injuries. As the child's condition worsens, the similarities among the clinical presentations of the divergent causes of shock overwhelm the differences. Fortunately, there are fundamental principles applicable to multiple causes of shock in children. In this issue of Pediatric Emergency Medicine PRACTICE, we will present an approach to pediatric shock based, as far as possible, on the available evidence.
APC — activated protein C
ARDS — adult respiratory distress syndrome
ATN — acute tubular necrosis
CVP — central venous pressure
DIC — disseminated intravascular coagulation
ECMO — extracorporeal membrane oxygenation
ED — emergency department
FDP — fibrin degradation product
FRC — functional residual capacity
GFR — glomerular filtration rate
IO — intraosseous
IVC — inferior vena cava
LPS — leukopolysaccharide
MODS — multiple organ dysfunction syndrome
MOSF — multiple organ system failure
PEEP — positive end-expiratory pressure
PIP — peak inspiratory failure
RSI — rapid sequence induction
SIRS — systemic inflammatory response syndrome
SVC — superior vena cava
VALI — ventilator-associated lung injury
It is impossible to create a purely evidence-based approach to pediatric shock. The reasons for this are quite straightforward. First, "pediatric shock" is a heterogeneous clinical entity. Multiple etiologies lead to shock. It is impossible to compare treatments, for example, when a study population includes children with hemorrhagic shock from trauma, hypovolemic shock from a diarrheal illness, cardiogenic shock in chronically ill children with congenital heart disease, septic shock, and distributive shock from anaphylaxis. Second, individual cases of pediatric shock are not common. A single institution would have to study data spanning many years to have a reasonably sized study. Third, the cause of shock is often not immediately apparent on presentation to the ED or intensive care unit. Therefore, studies tend to be retrospective and rely on information that is only available as the case unfolds over time. This leads to studies that have limited applicability to ED care. Fourth, children in shock are often critically ill, and some clinicians consider interventional or experimental studies unethical.1-3 Performing a study that substantially risks a child's death is unappealing, to say the least, to many researchers, patients, and families.3 This leads to paucity of relevant studies. Fifth, given the severity of illness, exceptions from informed consent may be needed to allow the performance of a study. Obtaining an exception from informed consent is an arduous process that few researchers have the resources or willingness to endure.3-5
Given the difficulties associated with performing studies on pediatric shock, physicians are left to act on very incomplete information. This can lead to a continued use of ineffective or even harmful therapies, simply because evidence is not available to refute their use.6,7 Reasons cited for using these ineffective therapies include: "love of [a] pathophysiological model (that is wrong)," "a need to do something," and "clinical experience."7
Another problem arises when the results of studies involving adults only are applied to the care of children. A recent example illustrates this point nicely. There have been studies and reports demonstrating that activated protein C (APC) is an effective therapy for adults in septic shock.8-10 However, a recent multicenter study of APC for the treatment of children in septic shock was suspended due to excessive complications and a lack of demonstrated benefit over placebo.11 In this case, there was an increase in intracranial bleeding, particularly in children younger than 2 months. Reliance on adult data to guide the care of children in this instance would have been harmful.
Finally, some of our most fundamental concepts are supported by very small studies. For example, any clinician who has been practicing for a few years knows that critically ill children are often found to be hypoglycemic on presentation. Studies that directly address this, however, are rare. Probably the best known is by Losek, who reported on 49 children undergoing "resuscitation," 9 of whom were discovered to be hypoglycemic.12 Another example involves fluid resuscitation. Although nearly universally recommended, few studies have directly explored whether or not fluid resuscitation is beneficial. The most widely cited of these is probably the study by Carcillo et al, which included only 34 children.13 Systematic reviews regarding fluid resuscitation seldom evaluate cherished, unproven "facts" and instead compare two similar therapies. 14,15
There are currently few data on the incidence of children presenting with shock to the ED. The evidence that does exist is predominantly related to septic shock. Based on data from children admitted to hospitals in 7 states, the national age-adjusted annual incidence of pediatric sepsis was found to be 0.56 cases per 1000 children, or 42,364 cases per year.16
The incidence of severe sepsis was found to be highest among infants, particularly low and very low birth weight babies. Boys were also found to have a significantly higher incidence compared to girls, approximating an additional 3300 boys per year nationally.16 Hospital mortality was 10.3% — an estimated 4300 or more deaths nationally from severe sepsis. Half of those deaths were in patients with a chronic comorbidity.17 Mortality in critically ill children is highly associated with multiple organ dysfunction syndrome (MODS) — it is common for multiple organs to fail early, acutely, and simultaneously.18 Data in children with septic shock and organ failure are limited, and most data analyze the incidence of sepsis, septic shock, and MODS in the pediatric intensive care unit rather than in the ED.19 Gram-negative septic shock comprises 50% of total cases of culture-proven bacterial sepsis, with approximately 115,000 deaths/year.16,17 As a group, gram-negative bacteria cause most of the deaths due to sepsis. Recently, more gram-positive cases of septic shock have been seen, likely due to the increased use of intravascular devices. The remainder of sepsis cases can be attributed to fungal, viral, and idiopathic causes.
Probable factors contributing to the increasing incidence of sepsis are the widespread use of corticosteroid and immunosuppressive therapies for organ transplants and inflammatory diseases, and the longer lives of patients predisposed to sepsis. This rise in bacteremia and sepsis is also related to the increased use of invasive devices, such as surgical prostheses, home mechanical ventilatory equipment, and percutaneous intravenous catheters. The indiscriminate use of antibiotics — creating conditions for overgrowth, colonization, and subsequent infection by aggressive, antimicrobial-resistant organisms — contributes, as well. The most frequent sites of infection include the lungs, abdomen, and urinary tract. Other sources include the skin, soft tissue, and the central nervous system.
1. "He wasn't hypotensive, so I figured he wasn't in shock."
2. "The pulse ox reading was normal. Why would I have given oxygen?"
3. "I didn't want to fluid overload the kid!"
4. "I gave 60 mL/kg of normal saline. How could that possibly not be enough?"
5. "What do you mean, she decompensated in the CT scanner? She looked fine 2 hours ago!"
6. "I didn't give antibiotics because I couldn't find a source of infection."
7. "The chest x-ray was normal. There weren't any infiltrates or effusions. But I guess, now that I look at it, the heart does look big."
8. "I've never given dopamine to a child, so I just kept giving fluids."
9. "Hydrocortisone? No, I didn't give any. Why should I have given hydrocortisone?"
10. "I wanted to make sure I knew what was going on before I called for transfer."
Evidence-based medicine requires a critical appraisal of the literature based upon study methodology and number of subjects. Not all references are equally robust. The findings of a large, prospective, randomized, and blinded trial should carry more weight than a case report.
To help the reader judge the strength of each reference, pertinent information about the study, such as the type of study and the number of patients in the study, will be included in bold type following the reference, where available. In addition, the most informative references cited in the paper, as determined by the authors, will be noted by an asterisk (*) next to the number of the reference.