Study offers guidance for telling anthrax from influenza

Sep 3, 2003 (CIDRAP News) – Researchers who compared the reported signs and symptoms of anthrax and of common viral respiratory tract infections have published guidelines that could help clinicians distinguish inhalational anthrax from more benign infections in the aftermath of an anthrax attack.

In a study published in Annals of Internal Medicine, the researchers say that neurologic symptoms other than headache, such as confusion, blurred vision, and dizziness, are the clinical markers most useful for distinguishing possible anthrax from viral respiratory tract infections such as influenza.

Dyspnea, nausea, and vomiting also are helpful for differentiating anthrax from the viral respiratory ailments, says the report by Nathaniel Hupert and colleagues from Cornell University and New York Presbyterian Hospital. Rhinorrhea and sore throat are more likely to suggest a viral respiratory tract infection, while fever and cough are common in both types of illness.

The authors collected data from 13 reports on a total of 28 inhalational anthrax cases, including the 11 cases that occurred in 2001 and 17 cases dating as far back as 1920. They also examined five studies reporting on the clinical features of 2,672 cases of influenza and 1,932 cases of other viral respiratory infections, such as respiratory syncytial virus, parainfluenza, and rhinoviruses. Anthrax cases were also compared with ambulatory cases of community-acquired pneumonia (CAP).

All but one of the anthrax patients had fever, chills, or cough, and more than half had dyspnea, chest discomfort or pain, and nausea or vomiting, the report says. Forty-three percent of the patients had nonheadache neurologic symptoms, and 81% had abnormalities on lung auscultation. Sore throat and rhinorrhea were reported in only 18% and 14% of patients, respectively.

The researchers calculated the "positive likelihood ratio" for various signs and symptoms—the ratio of the prevalence of the feature in inhalational anthrax patients to its prevalence in the other respiratory illnesses. Any ratio greater than 3 or less than 0.3 was considered clinically important. The ratio could not be determined for nonheadache neurologic symptoms because no relevant data were collected in the viral illness studies.

The analysis showed that any abnormality on lung auscultation is the most discriminating clinical feature of anthrax relative to influenza, with a positive likelihood ratio of 8.1 (95% confidence interval [CI], 5.3 to 12.5). Dyspnea is the most discriminating symptom (positive likelihood ratio, 5.3 [95% CI, 3.7 to 7.4]), followed by nausea or vomiting (ratio, 5.1 [95% CI, 3.0 to 8.5]). The ratios for anthrax versus other viral respiratory infections were similar. But the researchers found no major differences between the presenting signs and symptoms of anthrax and CAP. Fever and cough were common in both anthrax and the viral illnesses.

The report states, "Three symptoms in particular hold promise for differentiating anthrax from influenza and other viral respiratory infections in a mass screening setting: nonheadache neurologic symptoms, such as dizziness and confusion; dyspnea; and upper gastrointestinal tract symptoms, such as nausea and vomiting." The combination of any of these three symptoms with fever or chills or a cough "should raise suspicion of inhalational anthrax, as should any abnormality on auscultation of the lungs in a patient with fever/chills or cough in this setting."

On the other hand, the combination of fever or chills, cough, and ear, nose, and throat symptoms is unlikely to suggest inahalational anthrax, even after anthrax exposure, the report says. The authors used these findings to create an algorithm for identifying potential anthrax cases. The first step of the algorithm is an assessment for nonheadache neurologic symptoms, which, if present, indicate the patient should be sent to a healthcare facility for definitive testing for anthrax.

In an editorial accompanying the article, Harold C. Sox, MD, editor of the journal, gives the screening algorithm mixed marks. He says the algorithm seems "safe" in that it would probably send most patients with actual anthrax to hospital care, but it probably would also send many patients with other diagnoses to hospitals. The authors' "analysis of alternative diagnoses is relatively weak because published studies do not report the frequency of combinations of findings that they found useful," Sox writes. "As a result, we do not know the efficiency of triage; possibly, the algorithm would triage most patients with alternative diagnoses to hospital care."

Hupert N, Bearman GML, Mushlin AI, et al. Accuracy of screening for inhalational anthrax after a bioterrorist attack. Ann Intern Med 2003;139(5):337-46 [Abstract]

Sox HC. A triage algorithm for inhalational anthrax. Ann Intern Med 2003;139(5):379-81[Full text—requires subscription]

This week's top reads