OBJECTIVE: To identify factors associated with malignant pertussis. DESIGN: A retrospective case notes review from January 2003 to August 2013. Area under the receiver-operator characteristic curve was used to determine how well vital sign and white cell characteristics within 48 hours of hospital presentation identified children with malignant pertussis. SETTING: The national children's hospital in Auckland, New Zealand. PATIENTS: One hundred fifty-two children with pertussis. MEASUREMENTS AND MAIN RESULTS: There were 152 children with confirmed pertussis identified, including 11 children with malignant pertussis. The area under the receiver-operator characteristic curve was 0.88 (95% CI, 0.78-0.97) for maximum heart rate. The optimal cut-point was 180 beats/min, which predicted malignant pertussis with a sensitivity of 73% and a specificity of 91%. The area under the receiver-operator characteristic curve was 0.92 (95% CI, 0.81-1.0) for absolute neutrophil count, 0.85 (95% CI, 0.71-0.99) for total WBC count, 0.80 (95% CI, 0.63-0.96) for neutrophil-to-lymphocyte ratio, and 0.77 (95% CI, 0.58-0.92) for absolute lymphocyte count. All children with malignant pertussis had one or more of heart rate greater than 180 beats/min, total WBC count greater than 25 × 10/L, and neutrophil-to-lymphocyte ratio greater than 1.0 with an area under the receiver-operator characteristic curve of 0.96 (95% CI, 0.91-1.0) for a multivariate model that included these three variables. CONCLUSIONS: Clinical predictors of malignant pertussis are identifiable within 48 hours of hospital presentation. Early recognition of children at risk of malignant pertussis may facilitate early referral to a PICU for advanced life support and selection for trials of investigational therapies.
OBJECTIVE: To identify factors associated with malignant pertussis. DESIGN: A retrospective case notes review from January 2003 to August 2013. Area under the receiver-operator characteristic curve was used to determine how well vital sign and white cell characteristics within 48 hours of hospital presentation identified children with malignant pertussis. SETTING: The national children's hospital in Auckland, New Zealand. PATIENTS: One hundred fifty-two children with pertussis. MEASUREMENTS AND MAIN RESULTS: There were 152 children with confirmed pertussis identified, including 11 children with malignant pertussis. The area under the receiver-operator characteristic curve was 0.88 (95% CI, 0.78-0.97) for maximum heart rate. The optimal cut-point was 180 beats/min, which predicted malignant pertussis with a sensitivity of 73% and a specificity of 91%. The area under the receiver-operator characteristic curve was 0.92 (95% CI, 0.81-1.0) for absolute neutrophil count, 0.85 (95% CI, 0.71-0.99) for total WBC count, 0.80 (95% CI, 0.63-0.96) for neutrophil-to-lymphocyte ratio, and 0.77 (95% CI, 0.58-0.92) for absolute lymphocyte count. All children with malignant pertussis had one or more of heart rate greater than 180 beats/min, total WBC count greater than 25 × 10/L, and neutrophil-to-lymphocyte ratio greater than 1.0 with an area under the receiver-operator characteristic curve of 0.96 (95% CI, 0.91-1.0) for a multivariate model that included these three variables. CONCLUSIONS: Clinical predictors of malignant pertussis are identifiable within 48 hours of hospital presentation. Early recognition of children at risk of malignant pertussis may facilitate early referral to a PICU for advanced life support and selection for trials of investigational therapies.
Authors: Clark D Russell; Arun Parajuli; Hugo J Gale; Naomi S Bulteel; Philipp Schuetz; Cornelis P C de Jager; Anne J M Loonen; Georgios I Merekoulias; J Kenneth Baillie Journal: J Infect Date: 2019-02-22 Impact factor: 6.072