Fever and hypoxia predict radiographic pneumonia among children with wheezing

Mathews B, Shah S, Cleveland RH, Lee EY, Bachur RG, Neuman MI. Clinical predictors of pneumonia among children with wheezing. Pediatrics 2009;124:e29-36.

Question 

Among children with wheezing in the emergency department (ED) setting, what factors predict radiographically-confirmed pneumonia?

Design 

Prospective cohort study.

Setting 

Single ED in Boston, Massachusetts.

Participants 

A total of 526 children ≤21 years of age (median age 1.9 years) who were evaluated in the ED were found to have wheezing on examination and underwent chest radiography because of possible pneumonia.

Intervention 

Historical features and examination findings were collected by treating physicians before knowledge of the chest radiography results. Chest radiographs were read independently by 2 blinded radiologists.

Outcomes 

Accuracy of the history and physical findings were assessed by use of likelihood ratios.

Main Results 

Thirty-six percent of patients were hospitalized. A history of wheezing was present for 247 patients (47%). Twenty-six patients (4.9% [95% confidence interval [CI]: 3.3-7.3]) had radiographic pneumonia. History of fever at home (positive likelihood ratio [LR]: 1.39 [95% CI: 1.13-1.70]), history of abdominal pain (positive LR: 2.85 [95% CI: 1.08 -7.54]), triage temperature of ≥38°C (positive LR: 2.03 [95% CI: 1.34-3.07]), maximal temperature in the ED of ≥38°C (positive LR: 1.92 [95% CI: 1.48-2.49]), and triage oxygen saturation of <92% (positive LR: 3.06 [95% CI: 1.15- 8.16]) were associated with increased risk of pneumonia. Among fever-free children (temperature of <38°C) with wheezing, the rate of pneumonia was very low (2.2% [95% CI: 1.0-4.7]).

Conclusions 

Radiographic pneumonia among children with wheezing is uncommon. Historical and clinical factors may be used to determine the need for chest radiography for wheezing children. The routine use of chest radiography for children with wheezing but without fever should be discouraged.

Commentary 

Mathews et al are to be commended for putting together this large, well-designed, prospective cohort of wheezing children to evaluate features that predict the presence of pneumonia on radiography. This study overcomes some of the limitations in prior studies that excluded children with known asthma or those older than 2 years of age. The most important finding from this study is that radiographic pneumonia in children with wheezing is very low (4.9%), and even lower in those younger than 2 years old (3%). This article includes a very nice table of multilevel likelihood ratios for different historical features and physical findings. For example, the strongest predictor of radiographic pneumonia was the presence and height of fever in the ED. A temperature in the ED ≥ 38°C was associated with a likelihood ratio of 1.92 and a temperature ≥39°C had a likelihood ratio of 3.92. However, because of the small number of patients with radiographic pneumonia, the confidence intervals around these ratios were quite large. In addition, ruling out the presence of pneumonia by use of these findings would be difficult; the absence of fever on arrival to the ED is associated with a likelihood ratio of only 0.66. Although it may be difficult to use any one of these findings to rule out pneumonia, these data are the first step in the development of a clinical prediction rule for pneumonia in children with wheezing—such a rule would be a welcome addition to our diagnostic armamentarium.