The recommendations from more than 20 different EPs on how to manage the same patient with typical influenza symptoms varied from obtaining no testing whatsoever to ordering a full cardiac workup plus additional testing. In coming to my decision about what testing I would perform on this patient, I tried to organize some medical literature around several themes regarding the patient’s presentation.
Did the patient’s symptoms represent a serious illness? In a 1995 Annals of EM study by Marco, et al., elderly patients with fever of greater than 103 degrees were 7 times more likely to have a serious illness– most often pneumonia or urinary tract infections. To further risk stratify patients in this study, lab studies including a WBC count and a chest X-ray were required. WBC counts > 11,000 nearly doubled the odds of a serious illness and WBC counts > 20,000 increased the chances of serious illness nearly twentyfold. Infiltrates on chest X-ray made the presence of a serious illness 2.5 times as likely. To determine whether the patient had a “serious” illness, then, some studies suggest that laboratory testing is helpful.
Does the patient have pneumonia? If the patient in this scenario did have a serious illness, his symptoms were most suggestive of pneumonia. Unfortunately, pneumonia can be difficult to diagnose clinically.
Clinical findings may increase the likelihood of pneumonia, but are not definitive. A 1997 JAMA study by Metlay et al. showed that historical features of fever, chills, and myalgias all made the diagnosis of pneumonia somewhat more likely while presence of runny nose and sore throat made the diagnosis of pneumonia less likely. This same study showed that presence of a fever increased the likelihood of pneumonia between 1.4 and 4.4 times while the absence of abnormal chest findings decreased the likelihood of pneumonia by nearly half.
Even when using clinical findings, physicians still have difficulty diagnosing pneumonia. For example , in a 1992 Scandinavian study in the Journal of Primary Health Care, only 5% of adult patients (20 of 402) with respiratory tract symptoms had findings suggestive of pneumonia on chest X-ray. Of those 20 patients who had pneumonia on their X-rays, physicians correctly diagnosed pneumonia in only one third of patients based on history and physical examination alone. Additionally, physicians misdiagnosed an additional 22 patients as having pneumonia based on their clinical examinations when the patients’ chest x-rays were normal. Another study by Lieberman, et al., reported that pneumonia could be reliably diagnosed clinically 74% of the time. (Scand J Prim Health Care. 2003 Mar;21 (1):57-60).
The patient presented in this scenario had nonspecific clinical findings that are unlikely to be helpful in determining whether pneumonia was present.
What About Influenza? The scenario noted that there was a moderate amount of influenza in the community when the patient was seeking care. While only 5-20% of the population gets influenza each year, the effects of influenza can be significant. Influenza is estimated to cause 226,000 hospitalizations and 36,000 deaths in the United States each year. The CDC’s influenza web site (www.cdc.gov/flu) notes that it is sometimes difficult to distinguish influenza from other respiratory infections based on signs and symptoms alone. However, in areas with confirmed influenza virus circulation, the history of fever along with an acute onset of cough in a generally healthy adult is 79%-88% accurate in diagnosing influenza. The value of fever and cough in diagnosing influenza is not as accurate in children and older adults, predicting about 75% of influenza cases in children 5-12 years of age, 64% of influenza cases in children less than 5 years old, and only 30% of influenza cases in patients more than 65 years old. If elderly patients have received an influenza immunization, then cough has no value in predicting influenza infection. The only useful clinical signs in the vaccinated elderly are myalgias and presence of a fever which, when combined, were only 41% sensitive for diagnosing influenza.
Rapid influenza testing was available in the case presented, but the CDC notes that rapid influenza tests are 90-95% specific and only 50-75% sensitive. A positive influenza test clinches the diagnosis, but a negative test by no means rules out influenza. In addition, false negative testing may cause a clinician to perform significant amounts of additional testing to rule out other etiologies for the patient’s symptoms - when the patient really has influenza. Based on these CDC data, during an influenza outbreak, a clinical history of cough and fever may be more accurate than a rapid influenza test at diagnosing influenza in healthy adults.
The diagnosis of influenza increases the risk of morbidity and mortality in certain populations such as the very young, the old, and those with pre-existing lung disease. For example, from 1990-1999, patients aged 65 and older were nearly 200 times as likely to die from influenza than were patients aged 0-49 years. While influenza may have serious consequences, the patient presented in this case scenario did not fall into a high-risk classification.
The Utility of Chest X-rays Many responses to the scenario recommended obtaining a chest x-ray. While chest x-rays may be considered the “standard” for diagnosing pneumonia, chest x-rays are not as sensitive for diagnosing pneumonia as some would believe. In one 2008 Journal of Emergency Medicine study, plain chest radiographs missed 27% of pneumonia cases that were later diagnosed on chest CT, and in another 2006 American Journal of Emergency Medicine study presented at the 2005 ACEP Research Forum, 17.7% of normal chest x-rays later showed pneumonia on CT scans. Relying on a normal chest x-ray to exclude pneumonia could result in failing to diagnose more than a quarter of patients who actually have pneumonia.
The benefit of a “positive” chest x-ray confirming the need for antibiotic use may also be overestimated. Several sources note that half of all pneumonias are of viral etiology, and a majority of the cases of viral pneumonias are caused by the influenza virus. If the patient in this case has pneumonia when influenza is present in the community, it is likely that the patient has influenza pneumonia and will therefore unresponsive to antibiotic therapy. However, keep in mind that a small subset of influenza patients will develop secondary bacterial pneumonia.
The Bottom Line There is no “right” or “wrong” answer to the question of how to treat febrile patients with symptoms of pneumonia. To a reasonable degree of medical certainty, a healthy adult patient less than 65 years old who has symptoms of influenza during an influenza outbreak can be diagnosed with influenza and requires no further workup. Patients at the extremes of age and patients of any age who are toxic appearing may need further evaluation. The patient in this scenario is likely past the window of benefit from antiviral medications. My belief is that it is reasonable for a physician to treat this patient symptomatically without any testing at all. However, it would be important to inform the patient that influenza is likely, that rapid influenza testing is both expensive and lacks sensitivity, and that pneumonia, if present on chest x-ray, would likely be viral in etiology and would not respond to antibiotic treatment anyway. Allowing the patient to take part in decision making presents a win-win situation for providing evidence based medical care.
So, what is the standard of care? Clearly, with the data presented, regarding radiographic and clinical diagnoses of pneumonia and the limitations to influenza testing, there is no gold standard for diagnosis, and thus no consensus can be reached to establish a standard of care for the evaluation of such patients. There is no “right” or “wrong” answer to the question of how to treat febrile patients with symptoms of pneumonia. To a reasonable degree of medical certainty, a healthy adult patient less than 65 years old who has symptoms of influenza during an influenza outbreak can be diagnosed with influenza and requires no further workup. Patients at the extremes of age and patients of any age who are toxic appearing may need further evaluation. The patient in this scenario is likely past the window of benefit from antiviral medications. My belief is that it is reasonable for a physician to treat this patient symptomatically without any testing at all. However, it would be important to inform the patient that influenza is likely, that rapid influenza testing is both expensive and lacks sensitivity, and that pneumonia, if present on chest X-ray, would likely be viral in etiology and would not respond to antibiotic treatment anyway. If further diagnostics won’t be performed, it is wise to document you have informed the patient about signs and symptoms suggesting a more serious or worsening medical condition. Allowing the patient to take part in decision making presents a win-win situation for providing evidence based medical care.