Study objective: To assess whether C-reactive protein (CRP) is a sensitive marker of pneumonia and to evaluate whether it may be used as an index of treatment response.
Design: A retrospective casenote review was carried out on 40 patients admitted with simple community acquired pneumonia and 20 patients admitted with purulent bronchitis (infective exacerbations of chronic obstructive airways disease). Serum CRP levels, in addition to other traditional markers of infection, were measured in all patients on the first day. In 21 cases of pneumonia, a second CRP measurement was available after 3 to 7 days of antibiotic therapy.
Results: Temperature and WBC count showed considerable overlap between the pneumonia and bronchitic groups, whereas there was no overlap in serum levels of CRP. C-reactive protein levels were above 100 mg/L in all but two cases. In the bronchitic group only 7 out of 20 had levels above the normal range (< 10 mg/L). Mean +/- Standard Error of the Mean and lower/upper quartiles for CRP (mg/L) were as follows: pneumonia 217 +/- 16 mg/L, 130/275; purulent bronchitis, 18 +/- 3 mg/L, 10/18; [95% confidence interval (CI) for difference 153, 244 mg/L]. A CRP above 70 mg/L in pneumonia on day 1 occurred in association with a WBC count < 12 x 10(9)/L in 45% of cases and with a temperature < 37.0 degrees C in 32%. CRP levels fell to < 100 mg/L in all cases of pneumonia after antibiotic treatment: pretreatment 213 +/- 21 mg/L +/- 2, 138/270; posttreatment 31 +/- 5 mg/L, 14/47; [95% CI for difference 141, 221 mg/L].
Conclusion: Serum CRP may be a useful adjunctive test in pneumonia, both in terms of distinguishing parenchymal from endobronchial infection, as well as being a marker of treatment response.