Catching Serious Bacterial Infections in Infants
Second-generation Procalcitonin Assay Provides Better Discrimination
By Gina Rollins
Diagnosing serious bacterial infections (SBI) in young infants is a vexing, yet vital clinical challenge, as the very young are more susceptible to SBI and have a higher risk of mortality. There has been considerable effort to identify babies at risk for SBI so that clinicians can act promptly and appropriately to initiate what could be life-saving interventions. A number of biomarkers have been proposed. In a recent study, investigators tested serum procalcitonin using a second-generation assay for that purpose. This issue of Strategies examines those findings..
Fever is the reason for a large proportion of pediatric emergency department visits, but up to 20% of cases have no identifiable cause. Most ultimately prove to be benign in nature and resolve on their own. However, as many as 9% of infants 1 to 3 months of age and 12% of those less than 29 days old who present with fever of unknown origin turn out to have SBIs like bacteremia, meningitis, urinary tract infection, or pneumonia. Clinical examination alone is not an adequate diagnostic method, as about 12% of well-appearing babies actually are seriously ill. As a result, febrile infants must undergo a battery of tests, ranging from CBC and blood culture to cerebrospinal fluid analyses, while waiting for a definitive diagnosis.
Procalcitonin (PCT) has been examined extensively in recent years as a marker for systemic inflammation, infection, and sepsis, both singularly and in combination with other markers such as CRP, in adults and children. The predominant assay used in most studies has been an immunoluminometric assay, called the LUMItest, manufactured by Brahms. However, particularly in the pediatric population, the utility of PCT has been limited by less-than-ideal sensitivity and specificity, with reported sensitivity ranging from 30% to 100% and specificity ranging from 50% to 90%. Now in a new study published in Pediatrics (2008; 122:701-710), researchers examined the performance of a second-generation immunofluorescent assay that measures the signal emitted from an immunocomplex using time-resolved amplified cryptate emission, a method known as TRACE, in detecting SBI in young infants. The automated test, called Kryptor and manufactured by Brahms, was granted FDA clearance in March 2008, and has a functional sensitivity in the range of 60 pg/mL, according to the authors.
In addition to assessing the performance of the new assay, this study adds to the body of knowledge regarding PCT as a marker of SBI in very young children. The study is important because it is one of only a few that have explored PCT levels in children in the emergency department setting, and it is the first to assess the performance of PCT solely in infants less than 90 days old, the authors report.
The prospective cohort study involved 234 infants 90 days old or younger who presented at the emergency department with a fever of at least 38 ºC. Infants who received antibiotics or immunizations within 48 hours were excluded, as were those with a previously identified immunodeficiency or chronic disease, focal bacterial infection on examination, or surgery within the past 7 days. Before they had knowledge of PCT levels the authors classified patients as having definite, possible, or no SBI based on the results of various other tests. In all, 12.8% of subjects had definite SBI, 5.1% had possible SBI, and 82% had no SBI. Among those with definite SBI, there were four cases of bacteremia, two with bacteremia and UTI, and 24 with UTI. Among patients with possible SBI, there were five cases of pneumonia and seven UTIs.
Overall, patients with definite, or definite or possible SBI had higher mean levels of PCT than those with no SBI, 2.21 ng/mL and 2.48 ng/mL versus 0.38 ng/mL, respectively. The researchers also subdivided results based on age and found that mean PCT levels were higher in all classes of younger infants who were less than 29 days old, 3.41 ng/mL and 4.53 ng/mL for those with definite or definite or possible SBI; 0.40 ng/mL for those with no SBI. In contrast, older infants between 29 and 90 days had lower mean PCT levels: for those with definite or definite or possible SBI, the levels were 1.69 ng/mL and 1.45 ng/mL, respectively; for those with no SBI, the level was 0.38 ng/mL.
In discriminating definite or possible SBI and no SBI, the ROC curve had an AUC of 0.76. A cutoff value of 0.12 ng/mL had a sensitivity of 95.2%, specificity of 25.5%, and negative predictive value of 96.1%. This cutoff value identified all six cases of bacteremia, with PCT level ranges from 0.25 ng/mL to 7.3 ng/mL. However, it would have misclassified two patients with low-colony count UTIs who also had negative urinalysis results as being at low-risk for SBIs.
The study’s significance is that it demonstrated that PCT was sensitive enough both to detect all cases of SBI and to establish a cutoff to identify very young children at low-risk for SBI. “Below the 0.12 ng/mL cutoff you can be fairly certain the child does not have an infection, and so you can withhold antibiotics,” explained coauthor Kenneth Becker, MD, PhD, professor of medicine and physiology and director of the division of endocrinology at the George Washington University Medical School in Washington, DC. “If a child has a high PCT level—1, 2, 3 ng/mL or higher—you have cause to worry that the bacterial infection is quite serious.” Becker also is director of endocrinology at the VA Medical Center in Washington, DC.
That type of information is very helpful to emergency physicians trying to determine the best course of treatment for very young children, according to Allen Hsiao, MD, FAAP, assistant professor of pediatrics at Yale University. “So often you’re waiting for results to know whether to admit or discharge the child. In the meantime, these very young babies are being exposed to germs from other kids. The sooner you can get them to a safer, cleaner environment, the better,” he said.
Lab directors who are interested in adding the test to their lab’s menu will find the new PCT assay easy to run. Once the Kryptor system has been calibrated, it can process the PCT assay in about 20 minutes, according to Becker. “It’s almost real-time, and would be very useful in the emergency setting,” he noted.