Proposed rules increase accuracy of acute rhinosinusitis and acute bacterial rhinosinusitis diagnoses.
Scientists have developed an efficient way of distinguishing between acute rhinosinusitis and acute bacterial rhinosinusitis to reduce the inappropriate prescribing of antibiotics.
For the study, investigators used a multivariate analysis and a classification and regression tree (CART) analysis to develop proposed clinical decision rules for the diagnosis of acute rhinosinusitis. Included in the study were 175 Danish adults aged 18 to 65 years who were seeking treatment for suspected cases of acute rhinosinusitis. The investigators prospectively recorded signs, symptoms, C-reactive protein (CRP), and reference standard tests.
Point scores based on logistic regression models, and algorithms based on classification and CART were used to develop the clinical decision rules, which were used for each of the 3 reference standards used to diagnose patients: an abnormal finding on a CT scan; purulent or mucopurulent fluid detected by an antral puncture of the maxillary sinus; and a positive bacterial culture from the antral puncture fluid. Using these diagnostic rules, patients were placed in either low, moderate, or high-risk groups for acute rhinosinusitis or acute bacterial rhinosinusitis.
According to the study, point scores based on logistic regression was the most common approach to developing clinical decision rules. Patients who received more points were more likely to have a correct diagnosis. Furthermore, the ease of this approach makes it more appealing to physicians.
The CART algorithm uses the predictor variable to distinguish symptoms between patients, creating a tree. Each terminus, or leaf, of the tree shows the probability of the outcome of interest. Three CART models were developed for each of the 3 reference standards. The accuracy of each model correctly categorizes patients in the low, moderate, or high-risk classifications based on the probability of acute rhinosinusitis and acute bacterial rhinosinusitis in each group.
Anacute rhinosinusitisdiagnosis is defined by the presence of purulent or mucopurulent antral puncture fluid. Signs and symptoms include preceding upper respiratory tract infection, previous diagnosis of sinusitis, toothache, purulent nasal discharge, and any unilateral tenderness of the maxillary sinus.
The results of the study, published in the Annals of Family Medicine, showed that the CRP and ESR tests were the strongest predictors of acute rhinosinusitis and bacterial rhinosinusitis. However, researchers believe that the best reference standard is the positive bacterial culture of antral puncture fluid, with the point score system classifying patients into the 3 categories. Patients with a positive bacterial culture as the reference standard for low, moderate, and high had a 16%, 49%, and 73% likelihood of acute bacterial rhinosinusitis. The authors noted that CT scans are not recommended for clinical use because they place a minimal number of people in the low-risk group.
Overall, about 70% of patients are prescribed antibiotics when only 30% are suffering from acute bacterial rhinosinusitis. The sinus scores placed almost half of the participants in the low-risk group for acute bacterial rhinosinusitis, indicating that antibiotics should not be prescribed to these individuals.
Patients categorized as high-risk should be offered antibiotics, while moderate-risk patients should seek further testing.
“We have developed a point score and a CART-derived algorithm for diagnosing acute bacterial rhinosinusitis,” the authors wrote. “Use of such a score can help patients and physicians more confidently avoid antibiotics if patients are at low risk for a bacterial infection, reducing overuse of antibiotics.”