A score to predict Pseudomonas aeruginosa infection in older patients with community-acquired pneumonia

Abstract

Background: In Thailand, the incidence of community-acquired pseudomonal pneumonia among 60- to 65-year-olds ranges from 10.90% to 15.51%, with a mortality rate of up to 19.00%. Antipseudomonal agents should be selected as an empirical treatment for elderly patients at high risk for developing this infection. The purpose of this study was to identify risk factors and develop a risk predictor for Pseudomonas aeruginosa infection in older adults with community-acquired pneumonia (CAP).

Methods: A retrospective data collection from an electronic database involved the elderly hospitalized patients with P. aeruginosa- and non-P. aeruginosa-causing CAP, admitted between January 1, 2016, and June 30, 2021. Risk factors for P. aeruginosa infection were analysed using logistic regression, and the instrument was developed by scoring each risk factor based on the beta coefficient and evaluating discrimination and calibration using the area under the receiver operating characteristic curve (AuROC) and observed versus predicted probability (E/O) ratio.

Results: The inclusion criteria were met by 81 and 104 elderly patients diagnosed with CAP caused by P. aeruginosa and non-P. aeruginosa, respectively. Nasogastric (NG) tube feeding (odd ratios; OR = 40.68), bronchiectasis (B) (OR = 4.13), immunocompromised condition (I) (OR = 3.76), and other chronic respiratory illnesses (r) such as atelectasis, pulmonary fibrosis, and lung bleb (OR = 2.61) were the specific risk factors for infection with P. aeruginosa. The "60-B-r-I-NG" risk score was named after the 4 abbreviated risk variables and found to have good predicative capability (AuROC = 0.77) and accuracy comparable to or near true P. aeruginosa infection (E/O = 1). People who scored at least two should receive empirically antipseudomonal medication.

Conclusions: NG tube feeding before admission, bronchiectasis, immunocompromisation, atelectasis, pulmonary fibrosis and lung bleb were risk factors for pseudomonal CAP in the elderly. The 60-B-r-I-NG was developed for predicting P. aeruginosa infection with a high degree of accuracy, equal to or comparable to the existing P. aeruginosa infection. Antipseudomonal agents may be started in patients who are at least 60 years old and have a score of at least 2 in order to lower mortality and promote the appropriate use of these medications.

Keywords: Community-acquired pneumonia; Elderly; P. aeruginosa; Predictive risk score.

Fig. 1

60-B-r-I-NG risk score for predicting the risk of P. aeruginosa infection in elderly CAP patients

Fig. 2

AuROC and calibration curve of 60-B-r-I-NG risk score for P. aeruginosa infection prediction. a AuROC of the 60-B-r-I-NG risk score. b Calibration curve of the 60-B-r-I-NG risk score

Fig. 3

Calibration curve of refitting 60-B-r-I-NG risk score

Fig. 4

Probability of P. aeruginosa infection at various scores of the 60-B-r-I-NG model

Fig. 5

The risk score to the 60-B-r-I-NG checklist