A newly developed tool can identify bacterial acute sinusitis in preliminary studies, researchers report in an article published in the July/August issue of the Annals of Family Medicine. The investigators now plan to test the clinical rule in prospective studies to determine its accuracy and effect on clinical outcomes.
“The clinical rule uses six signs and symptoms, plus a C-reactive protein [CRP] greater than 15 mg/L. The signs and symptoms include preceding respiratory symptoms, no recent sinus infection, tender unilateral maxillary sinus, maxillary toothache, and purulent nasal discharge,” first author Mark Ebell, MD, from the University of Georgia in Athens, told Medscape Medical News via email.
Although further work is needed to validate the rule, it has the potential to decrease unnecessary use of antibiotics.
“Testing the tool in Danish primary care patients showed that it’s possible to classify about half of patients with sinus symptoms as low risk for a bacterial infection, as well as a much smaller group that’s very likely to have a bacterial infection,” Dr Ebell said.
Research has shown that sinusitis represents the most common reason for prescription of antibiotics in primary care: more than 70% of patients with symptoms of sinusitis receive antibiotics. However, studies also suggest that only about 30% of patients actually have a bacterial infection that requires antibiotics. Therefore, a tool that distinguishes between patients at low vs high risk for bacterial sinusitis may help decrease inappropriate antibiotic use.
To develop the clinical decision tool, the researchers tested various models against each of three reference standards: abnormal computed tomography findings, presence of pus in the sinuses, or a positive bacterial culture.
They also used two different approaches to develop the tools. They first used a logistic regression model to come up with a point-scores system. The second technique uses an algorithm developed from a classification and regression tree (CART) model. The CART algorithm asks about unilateral tender maxillary sinus tenderness, CRP, preceding sinusitis, and maxillary toothache.
They tested each of the models on data from 175 adults seen at primary care clinics in Denmark and for whom data on signs, symptoms, CRP, and reference standard tests had been collected prospectively.
The point-score rule split the patient group into three groups on the basis of their risk for a bacterial infection: low (16%), moderate (49%), and high (73%).
When compared against a bacterial culture reference standard, the clinical rule had good accuracy, with an area under the receiver operating characteristic curve between 0.721 and 0.767.
In this population, the researchers estimate that antibiotic use would have decreased to 34% if the tool had been used to guide treatment, and all high-risk and half of the intermediate-risk patients received the drugs. That rate of prescribing is consistent with current practice guidelines, suggesting just 27% of patients with sinusitis actually need antibiotics.
In other words, using the point score with bacterial culture as the reference would probably lead to more conservative use of antibiotics, according to the authors.
The CART model with bacterial culture as the reference could provide a good alternative for providers who prefer algorithms. The CART model classified a similar number of patients as low, moderate, and high likelihood for bacterial sinusitis at 6%, 31%, and 59%, respectively. Area under the receiver operating characteristic curve results for the CART model were between 0.783 to 0.827, similar to the point scores rule.
Results from the logistic regression analysis also showed that the strongest individual predictor of bacterial sinusitis was CRP, which strengthens the case for making CRP testing available in primary care.
“The CRP test was the strongest predictor, and is widely used in Europe to guide decisions about antibiotics, but hasn’t been approved for routine office use by the US [Food and Drug Administration]. That’s unfortunate, because we need all the tools we can get to help reduce inappropriate antibiotic use,” Dr Ebell stressed.
The authors mentioned several limitations, including incomplete data on fever and no data on certain types of computed tomography findings associated with sinusitis. Also, the findings may not apply to children, because this study only included adults.
The authors have disclosed no relevant financial relationships.
Ann Fam Med. 2017;15:347-354. Abstract
For more news, join us on Facebook and Twitter
Tidak ada komentar:
Posting Komentar