Does acute bronchitis really exist?

The logistic regression model had an R2 of 0.37, indicating that all the signs and symptom variables that we considered explained only approximately a third of the differences between these 2 conditions. To illustrate the limitations of using just these 2 variables to differentiate acute bronchitis and URI, if the presence of a cough was used as the sole determinant of the diagnosis, the sensitivity for acute bronchitis would be 98% (132 out of 135), but specificity would be only 29% (120 out of 409). Wheezing on its own had a sensitivity of 30% with a specificity of 97%. Adding in the absence of runny nose or nausea to these other predictors was no better than wheezing alone. Leaving out wheezing and including cough, absence of runny nose, and nausea gave a sensitivity of 82% with a specificity of 65%. Thus, clinical factors alone are insufficient to explain why some patients were given the diagnosis they received.

We also explored whether clinicians used a diagnosis to influence treatment decisions. It appeared that treatments did differ according to the diagnosis. Patients with acute bronchitis were 8 times more likely to receive a bronchodilator than those with a URI (70% vs 8%, P <.001) and were more likely to receive antibiotics (26% vs 4%, P <.001). Decongestants ere prescribed more often when the diagnosis was a URI (33% vs 16%, P <.001).

Discussion

Our study demonstrates that physicians rely on only a few clinical factors to differentiate acute bronchitis from URI. These factors are poor predictors of which patients will receive a diagnosis of acute bronchitis and which URI, since significant overlap exists between these 2 disorders. Despite a lack of clear differences between these conditions, physicians use these diagnoses to make treatment decisions. The low degree of specificity for clinical signs and symptoms in differentiating the 2 conditions implies that physicians use other cues such as the patient’s desire for treatment, personal thresholds for prescribing, and a general gestalt when labeling the condition.

These findings mirror our previous study that shows significant clinical overlap between sinusitis and URI.³ As in our current findings with acute bronchitis, “sinusitis” may simply be an acute respiratory infection predominantly in the head with moderate to high severity. Computed tomography scans of patients with clinical syndromes diagnosed as “colds” show sinus inflammation in 47% of patients.¹⁶ The condition frequently labeled as “sinusitis” may constitute a desire of physicians to use antibiotics to treat symptoms caused by this virally induced inflammation, even though scant evidence exists that antibiotics improve outcomes even when sinusitis is demonstrated by radiologic techniques. In most clinical situations, a majority of patients, clinicians, and clinical pharmacists appear to assign the diagnosis of sinusitis or believe antibiotics are helpful for any respiratory infection with discolored nasal discharge.^12,17,18 However, discolored discharge on its own is a poor predictor of antibiotic response and is a common manifestation of a URI.

The same situation seems to exist with acute bronchitis. Acute bronchitis may be nothing more than an acute respiratory infection that is predominantly in the chest. This would explain why there is such a high degree of overlap between the signs and symptoms of this condition and illnesses labeled as URIs. Except for the observation that albuterol is effective at hastening symptom resolution in patients with productive coughs and wheezes^19,20 but ineffective for patients with nonspecific coughs associated with upper airway symptoms,²¹ there is little value in labeling patients as having acute bronchitis rather than a cold, since these 2 entities are probably manifestations of the same clinical condition.

Taken together, it may be more useful to reconceptualize respiratory infections as a single clinical entity rather than by anatomically specific diagnoses. Patients may differ in the degree of severity of the illness and the anatomic area that produces most of their complaints. Clinically and in our research, we may be better served by representing all viral respiratory conditions as a single clinical diagnosis with severity and anatomic indications. All patients with acute symptoms involving the sinuses, nose, pharynx, and bronchial tree could be labeled as having an “acute respiratory infection,” either “sinus predominant,” “bronchial predominant,” or even “generalized” when involving all areas of the respiratory tract.

Most evidence suggests that the vast majority of all these infections are caused by viruses. This does not imply that patients do not have infections and are not ill. Not only do viruses vary in the degree in which they infect individuals, but patients also differ in their abilities to cope with their cold symptoms. Rather than attempting to eradicate the cause of the cold, the goals of treatment for these disorders should be improving outcomes that matter to patients, such as reducing their symptoms, improving their ability to function at work and at home, and relieving their anxiety. Instead of trying to eradicate bacteria that do not exist in the majority of patients, physicians should focus on treatments proven to alleviate the predominant symptoms and the underlying reasons patients seek care for this self-limited problem. These latter reasons could include loneliness, anxiety about the severity of the condition (for example, in a young child), or frustration that the symptoms are interrupting other important activities. Attention to these important psychosocial events may be even more important than the type of drug written on the prescription pad. Dismissing a patient with these problems as being uninformed, bothersome, or inappropriately seeking care may not be productive in addressing the underlying reasons the patient has consulted a physician. Failing to address these underlying psychosocial issues or prescribing an antibiotic to meet physicians’ desire to help and meet an unstated patient expectation may result in repetitive care-seeking for the same self-limited symptoms,^22,23 ultimately benefiting no one and consuming health care resources that could be better invested elsewhere.