OR WAIT null SECS
This newly-developed prediction model evaluating CAP risk was noted as having demonstrated accuracy, strong discriminatory qualities, and practicability in clinical settings.
The use of a nomogram model is effective for prediction of community-acquired pneumonia (CAP) risk in hospitalized individuals with acute asthma exacerbations, according to recent findings.1
These findings resulted from a retrospective study designed to assess the clinical characteristics of acute asthma exacerbations among those with CAP and to work toward a prediction model for CAP for those who have been hospitalized with such exacerbations.
This research was viewed as invaluable, given that pneumonia has been shown to be an independent risk factor for mortality among those hospitalized for asthma exacerbations, and given the varied prevalence of CAP in different studies.2 This retrospective analysis was led by Jinxiang Wang, from the Department of Pulmonary and Critical Care Medicine at Capital Medical University’s Beijing Luhe Hospital in China.
“We identified some of the clinical characteristics of adult AEs patients with CAP,” Wang and colleagues wrote. “We developed a prediction model to evaluate the CAP risk in adults with (asthma exacerbations).”
The research team conducted a retrospective examination of the records of patients kept in Beijing Luhe Hospital at Capital Medical University, gathering them between December 2017 - August 2021. All of the study subjects had had diagnoses of asthma exacerbations, specifically bronchial asthma following the GINA diagnostic criteria.
Asthma exacerbations, as defined by the team’s criteria, involve sudden symptoms among patients like shortness of breath, wheezing, coughs, tightness in the chest, or a sharp worsening of existing symptoms, often requiring treatment changes. The investigators noted that severity of such exacerbations was separated into ‘mild-to-moderate,’ as ‘severe,’ or as ‘life-threatening’ based on the guidelines used in the study.
The team’s criteria for inclusion were subjects being aged 18 years or older and hospitalized for asthma exacerbations. Their criteria for subject exclusion were being discharged within 24 hours that also had incomplete information, being under 18, pregnancy, those with pulmonary vasculitis, having human immunodeficiency virus infection, or immunosuppressive therapy.
The investigators used Lasso regression and used multivariate logistic regression methods to point out the most successful predictors. Subsequently, the team crafted their predictive nomogram utilizing these key predictors.
For internal validation of the nomogram model, the research team implemented the bootstrap technique. They sought to assess the nomogram's quality of performance by utilizing metrics such as the calibration curve, the area under the receiver operating characteristic curve (AUC), and the decision curve analysis (DCA).
Overall, the investigators ended up with data on 308 subjects who had all been admitted to the hospital for asthma exacerbations. They reported that 21% of these individuals were shown to have had CAP.
The research team noted that several elements were independently linked with CAP in the group, including higher levels of fibrinogen, previous utilization of systemic corticosteroids, increased C-reactive protein levels, greater count of white blood cells, the observed existence of fever, and early-onset asthma.
The team looked into the nomogram’s predictive accuracy for CAP, and they found that it displayed an AUC of 0.813 (95% CI: 0.753 - 0.872). Their internal validation ended up confirming the model’s reliability with a concordance index of 0.794.
The investigators also found that the calibration curve in their findings showed strong alignment with the study’s diagonal reference line. Additionally, their DCA indicated that the nomogram's clinical utility was shown to be the highest when the threshold probability of CAP in study subjects ranged from 3% - 89%.
The investigators in summary reported that there were distinctive clinical characteristics observed in this group, and the co-occurrence of such characteristics were shown to be relatively common.
“Using the nomogram, clinicians can determine the risk of CAP in patients hospitalized with an AE,” they wrote. “With this prediction model, we can accurately predict the risk of CAP in patients with acute AE and play a role in early screening and diagnosis of CAP. This has an important role in guiding the use of antimicrobial drugs in patients with acute AEs.”