Predictors of disease severity in children hospitalized for pertussis during an epidemic

Background: Australia recently experienced its worst pertussis epidemic since introduction of pertussis vaccine into the National Immunisation Program. This study aimed to determine factors associated with severe pertussis in hospitalized children during an epidemic using a novel pertussis severity scoring (PSS) system. Methods: This prospective, observational, multicenter study enrolled children hospitalized with laboratory confirmed pertussis from 8 tertiary pediatric hospitals during a 12 month period (May 2009-April 2010). Variables assessed included demographics, clinical symptoms and relevant medical and immunization history. Cases were scored using objective clinical findings with cases classified as either severe (PSS > 5) or not severe (PSS ≤ 5). Logistic regression models were used to predict variables associated with severe disease. Results: One hundred twenty hospitalized children 0-17 years of age were enrolled with a median PSS of 5 (interquartile range 3-7). Most (61.7%) were classified as not severe with 38.3% (46/120) severe. Most severe cases (54.3%) were <2 months of age. Presence of coinfection [odds ratio (OR): 4.82, CI: 1.66-14.00], <2 months old (OR: 4.76, CI: 1.48-15.32), fever >37.5°C (OR: 5.97, CI: 1.19-29.96) and history of prematurity (OR: 5.00, CI: 1.27-19.71) were independently associated with severe disease. A total of 70 cases in children ≥2 months of age, almost a third (n = 23) had not received pertussis vaccine. Conclusions: Most severe pertussis occurred in young, unimmunized infants, although severe disease was also observed in children >12 months of age and previously vaccinated children. Children admitted with pertussis with evidence of coinfection, history of prematurity or fever on presentation need close monitoring.