Conference Coverage

Biomarker duo rapidly identifies serious bacterial infections



– The combination of serum procalcitonin and C-reactive protein levels upon admission to a pediatric ICU displayed high utility for early diagnosis of serious bacterial infection in critically ill children in a large prospective observational study presented at the annual meeting of the European Society for Paediatric Infectious Diseases.

This winning combination significantly outperformed neutrophil gelatinase-associated lipocalin, activated partial thromboplastin time, and resistin, both individually and in various combinations, for the vital task of making a rapid distinction between infectious and noninfectious causes of pediatric systemic inflammatory response syndrome, reported Enitan D. Carrol, MD, professor of pediatric infection at the University of Liverpool (England).

Dr. Enitan D. Carrol, professor of pediatric infection at the University of Liverpool, England Bruce Jancin/MDedge News

Dr. Enitan D. Carrol

“One of the clinical dilemmas we face in intensive care is being able to differentiate between infectious and noninfectious causes of systemic inflammatory response syndrome. This is important because we need to identify which children have life-threatening infections so that we can promptly initiate antimicrobial therapy,” she explained.

One in four deaths in pediatric ICUs are infection related, Dr. Carrol noted.

“There is an urgent need for infection markers which, firstly, change early in the course of bacterial infection, secondly, correlate with real-time clinical progression, and thirdly, have a rapid turn-around time to allow effective clinical decision making,” she observed.

The combination of procalcitonin and C-reactive protein (CRP) levels measured at admission fits the bill, Dr. Carrol continued. Of the five biomarkers evaluated in her study – all backed by some supporting evidence of efficacy in earlier studies – the top two individual performers in terms of negative predictive value (NPV) were a CRP less than 4.2 mg/dL with a negative NPV of 99%, and a procalcitonin less than 1.52 ng/mL with an NPV of 96%. The positive predictive value of each of the biomarkers was 37%. The sensitivity and specificity of procalcitonin for diagnosis of serious bacterial infection were 78% and 80%, respectively. For CRP, the figures were 93% and 76%.

The combination of procalcitonin and CRP outperformed a multitude of other two-, three-, and four-biomarker combinations tested, with an area under the curve of 93% for combined sensitivity and specificity.

The study included 657 children admitted to the pediatric ICU at Alder Hey Children’s Hospital in Liverpool with systemic inflammatory response syndrome. All had blood samples measured for the five biomarkers on days 1-7. Clinicians were blinded as to the biomarker results. Ninety-two (14%) patients were ultimately found to have a serious bacterial infection – essentially, bacterial meningitis or septic shock – and 565 (86%) had a nonbacterial etiology.

The 28-day mortality rate was 9% in the group with serious bacterial infection, significantly higher than the 2% rate in the group with other causes of their systemic inflammatory response syndrome.

Longitudinal trends in procalcitonin and CRP as evidenced in the study can be used in clinical decision making, according to Dr. Carrol. Mean values of procalcitonin plummeted by 80% from day 1 to day 5 in response to antimicrobial therapy in the group with serious bacterial infections. In contrast, CRP levels rose sharply from day 1 to a peak on day 2, then fell, although the 50% drop from day 2 to day 5 in response to antimicrobial therapy wasn’t as pronounced as the change in procalcitonin.

“There is an additive benefit for both biomarkers compared with CRP alone. The problem with CRP on admission, as I’ve demonstrated in this study, is it often hasn’t risen yet early after admission. So although it gave the best area under the curve of any of the biomarkers, I think that combined with procalcitonin you get a much better descriminator,” Dr. Carrol said.

The median duration of ICU stay in the patients with serious bacterial infection at admission was 5 days, compared with 3 days when the cause of systemic inflammatory response syndrome lay elsewhere. Their median duration of ventilation was significantly longer, too: 4 days versus 2 in children without a serious bacterial infection.

Stepwise logistic regression analysis pinpointed several clinical variables as being associated with prolonged ICU stay.

Clinical factors associated with prolonged ICU stay

In addition, initiation of antibiotic therapy prior to admission to the pediatric ICU was associated with a 50% reduction in the likelihood of a prolonged ICU stay. “This reflects the fact that early antibiotics give you a better prognosis if you have sepsis,” according to Dr. Carrol.

She and her coinvestigators now have embarked on a multicenter U.K. study looking at the impact of procalcitonin to guide duration of antimicrobial therapy in critically ill children.

The Alder Hey study was funded by the U.K. National Institute for Health Research. Dr. Carrol reported having no financial conflicts. Although she serves as a consultant to several health care companies, all remuneration goes directly to the University of Liverpool.

[email protected]

Recommended Reading

Simple tool improves inpatient influenza vaccination rates
The Hospitalist
Antibiotic stewardship in sepsis
The Hospitalist
Severity of sepsis-associated coagulopathy predicts hospital mortality
The Hospitalist
Many hospitals had no mandatory flu vaccine requirements in 2017
The Hospitalist
Trio of blood biomarkers elevated in children with LRTIs
The Hospitalist
CDC concerned about multidrug-resistant Shigella
The Hospitalist
Hospital-acquired conditions drop 8% since 2014, saving 8,000 lives and $3 billion
The Hospitalist
CDC warns of hepatitis A outbreaks in injection drug users
The Hospitalist
CMS sepsis measure a challenge to report
The Hospitalist
Diagnosing pulmonary tuberculosis in the hospital
The Hospitalist
   Comments ()