Limits...
Interleukin-27: a novel biomarker in predicting bacterial infection among the critically ill.

Hanna WJ, Berrens Z, Langner T, Lahni P, Wong HR - Crit Care (2015)

Bottom Line: Infected status was determined by a chart review by an intensivist blinded to biomarker results.Similar AUCs were found for the subset of immunocompromised patients.In a CART-derived analysis taking immunocompromised status into consideration, a combination of IL-27 and PCT yielded an AUC of 0.81 (0.75 to 0.86), statistically improved from either IL-27 or PCT alone.

View Article: PubMed Central - PubMed

Affiliation: Division of Critical Care Medicine, Cincinnati Children's Hospital Medical Center and Cincinnati Children's Research Foundation, 3333 Burnet Avenue, Cincinnati, OH, 45229, USA. william.hanna@sickkids.ca.

ABSTRACT

Introduction: A continued need exists for effective diagnostic biomarkers in bacterial sepsis among critically ill patients, despite increasing use of available biomarkers such as procalcitonin (PCT). Interleukin-27 (IL-27) has shown early promise in a recent preliminary study, exhibiting high specificity and positive predictive values for bacterial infection in critically ill children. This validation study was performed to assess the value of IL-27 in predicting bacterial infection among patients admitted to the pediatric intensive care unit and to compare its performance with that of PCT.

Methods: A single-center (n = 702) prospective study was performed comparing both IL-27 and PCT levels between bacterially infected and uninfected cohorts in the pediatric intensive care unit. Infected status was determined by a chart review by an intensivist blinded to biomarker results. Formal performance comparisons included calculations of receiver operating characteristic (ROC) curves for IL-27 and PCT individually in addition to a combination strategy using a decision tree generated by classification and regression tree (CART) methodology. Secondary analysis focusing on subjects with documented bloodstream infections was performed.

Results: The overall infection rate was 27 %. ROC curves for the primary analysis yielded areas under the curve (AUCs) of 0.64 (0.59 to 0.68) for IL-27 and 0.61 (0.56 to 0.65) for PCT. Secondary analysis defining infected status exclusively through positive blood cultures yielded AUCs of 0.75 (0.68 to 0.81) for IL-27 and 0.64 (0.57 to 0.71) for PCT, with a specificity of 95 % (92 % to 97 %) for the prior established IL-27 cut-point value of at least 5.0 ng/ml. Similar AUCs were found for the subset of immunocompromised patients. In a CART-derived analysis taking immunocompromised status into consideration, a combination of IL-27 and PCT yielded an AUC of 0.81 (0.75 to 0.86), statistically improved from either IL-27 or PCT alone.

Conclusions: Despite having a modest predictive value for infection independent of source, IL-27 may serve as a useful biomarker in estimating risk of bacterial infection among critically ill pediatric patients with bloodstream infections. In particular, among immunocompromised subjects, this diagnostic biomarker may be helpful either alone or using a combination strategy with other available biomarkers.

No MeSH data available.


Related in: MedlinePlus

The classification tree includes interleukin-27 (IL-27), procalcitonin (PCT), and immunocompromised status (I.C.) as predictor variables. The biomarker concentrations are presented in nanograms per milliliter. The root node provides the total number of subjects, and the number of subjects with negative and positive blood cultures, with the respective rates. Each daughter node provides the respective decision rule criterion and the number of subjects with negative and positive blood cultures, with the respective rates. Terminal nodes (TN) TN1, TN3, and TN5 are considered low-risk terminal nodes (≤12 % risk of positive culture). TN2, TN4, TN6, TN7 and TN8 are higher-risk terminal nodes (≥12 % risk of positive) compared with the root node
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
getmorefigures.php?uid=PMC4627377&req=5

Fig3: The classification tree includes interleukin-27 (IL-27), procalcitonin (PCT), and immunocompromised status (I.C.) as predictor variables. The biomarker concentrations are presented in nanograms per milliliter. The root node provides the total number of subjects, and the number of subjects with negative and positive blood cultures, with the respective rates. Each daughter node provides the respective decision rule criterion and the number of subjects with negative and positive blood cultures, with the respective rates. Terminal nodes (TN) TN1, TN3, and TN5 are considered low-risk terminal nodes (≤12 % risk of positive culture). TN2, TN4, TN6, TN7 and TN8 are higher-risk terminal nodes (≥12 % risk of positive) compared with the root node

Mentions: A subsequent CART analysis was generated by using this same modified definition which included IL-27 levels, PCT levels, and immunocompromised status as predictor variables (Fig. 3). Starting with the “root” node that included all patients (n = 579) and dividing into two subsequent “daughter” nodes, eight terminal nodes were ultimately generated, each with specific cutoff points for IL-27 and PCT. Terminal nodes 1, 3, and 5 were designated as low-risk nodes with 3 %, 4 %, and 0 % of patients being infected, respectively. Terminal nodes 2, 4, 6, 7, and 8 were designated as higher-risk nodes compared with the “root” node, with rates of 12 %, 29 %, 25 %, 38 %, and 86 %, respectively. Using these designations, the CART algorithm was then tested, yielding an AUC of 0.81 (0.75, 0.86). This was statistically improved from both the individual IL-27 (0.81 versus 0.74, P < 0.01) and PCT (0.81 versus 0.64, P <0.01) AUCs (Additional file 2). Test characteristics (95 % CI) for the tree yielded a sensitivity of 84 % (72–91), specificity of 63 % (59–67), negative predictive value of 97 % (94–98), positive predictive value of 23 % (18–29), positive likelihood ratio of 2.3 (1.9–2.6), and negative likelihood ratio of 0.3 (0.2–0.4).Fig. 3


Interleukin-27: a novel biomarker in predicting bacterial infection among the critically ill.

Hanna WJ, Berrens Z, Langner T, Lahni P, Wong HR - Crit Care (2015)

The classification tree includes interleukin-27 (IL-27), procalcitonin (PCT), and immunocompromised status (I.C.) as predictor variables. The biomarker concentrations are presented in nanograms per milliliter. The root node provides the total number of subjects, and the number of subjects with negative and positive blood cultures, with the respective rates. Each daughter node provides the respective decision rule criterion and the number of subjects with negative and positive blood cultures, with the respective rates. Terminal nodes (TN) TN1, TN3, and TN5 are considered low-risk terminal nodes (≤12 % risk of positive culture). TN2, TN4, TN6, TN7 and TN8 are higher-risk terminal nodes (≥12 % risk of positive) compared with the root node
© Copyright Policy - OpenAccess
Related In: Results  -  Collection

License 1 - License 2
Show All Figures
getmorefigures.php?uid=PMC4627377&req=5

Fig3: The classification tree includes interleukin-27 (IL-27), procalcitonin (PCT), and immunocompromised status (I.C.) as predictor variables. The biomarker concentrations are presented in nanograms per milliliter. The root node provides the total number of subjects, and the number of subjects with negative and positive blood cultures, with the respective rates. Each daughter node provides the respective decision rule criterion and the number of subjects with negative and positive blood cultures, with the respective rates. Terminal nodes (TN) TN1, TN3, and TN5 are considered low-risk terminal nodes (≤12 % risk of positive culture). TN2, TN4, TN6, TN7 and TN8 are higher-risk terminal nodes (≥12 % risk of positive) compared with the root node
Mentions: A subsequent CART analysis was generated by using this same modified definition which included IL-27 levels, PCT levels, and immunocompromised status as predictor variables (Fig. 3). Starting with the “root” node that included all patients (n = 579) and dividing into two subsequent “daughter” nodes, eight terminal nodes were ultimately generated, each with specific cutoff points for IL-27 and PCT. Terminal nodes 1, 3, and 5 were designated as low-risk nodes with 3 %, 4 %, and 0 % of patients being infected, respectively. Terminal nodes 2, 4, 6, 7, and 8 were designated as higher-risk nodes compared with the “root” node, with rates of 12 %, 29 %, 25 %, 38 %, and 86 %, respectively. Using these designations, the CART algorithm was then tested, yielding an AUC of 0.81 (0.75, 0.86). This was statistically improved from both the individual IL-27 (0.81 versus 0.74, P < 0.01) and PCT (0.81 versus 0.64, P <0.01) AUCs (Additional file 2). Test characteristics (95 % CI) for the tree yielded a sensitivity of 84 % (72–91), specificity of 63 % (59–67), negative predictive value of 97 % (94–98), positive predictive value of 23 % (18–29), positive likelihood ratio of 2.3 (1.9–2.6), and negative likelihood ratio of 0.3 (0.2–0.4).Fig. 3

Bottom Line: Infected status was determined by a chart review by an intensivist blinded to biomarker results.Similar AUCs were found for the subset of immunocompromised patients.In a CART-derived analysis taking immunocompromised status into consideration, a combination of IL-27 and PCT yielded an AUC of 0.81 (0.75 to 0.86), statistically improved from either IL-27 or PCT alone.

View Article: PubMed Central - PubMed

Affiliation: Division of Critical Care Medicine, Cincinnati Children's Hospital Medical Center and Cincinnati Children's Research Foundation, 3333 Burnet Avenue, Cincinnati, OH, 45229, USA. william.hanna@sickkids.ca.

ABSTRACT

Introduction: A continued need exists for effective diagnostic biomarkers in bacterial sepsis among critically ill patients, despite increasing use of available biomarkers such as procalcitonin (PCT). Interleukin-27 (IL-27) has shown early promise in a recent preliminary study, exhibiting high specificity and positive predictive values for bacterial infection in critically ill children. This validation study was performed to assess the value of IL-27 in predicting bacterial infection among patients admitted to the pediatric intensive care unit and to compare its performance with that of PCT.

Methods: A single-center (n = 702) prospective study was performed comparing both IL-27 and PCT levels between bacterially infected and uninfected cohorts in the pediatric intensive care unit. Infected status was determined by a chart review by an intensivist blinded to biomarker results. Formal performance comparisons included calculations of receiver operating characteristic (ROC) curves for IL-27 and PCT individually in addition to a combination strategy using a decision tree generated by classification and regression tree (CART) methodology. Secondary analysis focusing on subjects with documented bloodstream infections was performed.

Results: The overall infection rate was 27 %. ROC curves for the primary analysis yielded areas under the curve (AUCs) of 0.64 (0.59 to 0.68) for IL-27 and 0.61 (0.56 to 0.65) for PCT. Secondary analysis defining infected status exclusively through positive blood cultures yielded AUCs of 0.75 (0.68 to 0.81) for IL-27 and 0.64 (0.57 to 0.71) for PCT, with a specificity of 95 % (92 % to 97 %) for the prior established IL-27 cut-point value of at least 5.0 ng/ml. Similar AUCs were found for the subset of immunocompromised patients. In a CART-derived analysis taking immunocompromised status into consideration, a combination of IL-27 and PCT yielded an AUC of 0.81 (0.75 to 0.86), statistically improved from either IL-27 or PCT alone.

Conclusions: Despite having a modest predictive value for infection independent of source, IL-27 may serve as a useful biomarker in estimating risk of bacterial infection among critically ill pediatric patients with bloodstream infections. In particular, among immunocompromised subjects, this diagnostic biomarker may be helpful either alone or using a combination strategy with other available biomarkers.

No MeSH data available.


Related in: MedlinePlus