Assessment of Influenza-like Illness Criteria in Hospitalized Adults with Non-influenza Respiratory Virus Infections

Background: The Centers for Disease Control and Prevention (CDC) defines influenza-like illness (ILI) as a fever ≥ 100.0F associated with cough and/or sore throat in the absence of an alternative diagnosis.  Studies have shown that these criteria are non-specific and are often present among patients with acute non-influenza respiratory virus infections (RVIs).

Objective: To determine the incidence of ILI among hospitalized adults with non-influenza RVIs and to assess the impact of ILI diagnosis on clinical outcomes.

Methods: We performed a retrospective cohort study of 550 patients hospitalized from Dec 1, 2008 to Sep 30, 2010 who had real time reverse transcription polymerase chain reaction (RT-PCR) assays positive for non-influenza respiratory viruses.  Patients less than 18 years of age and those diagnosed with influenza were excluded. 

Results: The cohort included 10 (1.8%) with adenovirus, 82 (14.9%) human metapneumovirus, 57 (10.4%) parainfluenza virus, 94 (17.1%) respiratory syncytial virus, 296 (53.8%) rhinovirus, and 11 (2%) with mixed viral infections.  Overall, 272 (49.5%) met criteria for ILI. Of these, 265 (97.4%) presented with cough and 79 (29%) presented with sore throat.  Only patients with rhinovirus were statistically less likely to present with ILI [134 (45%) vs. 162 (54%), p = 0.03]. There was no difference in the frequency of underlying medical illnesses including diabetes mellitus, human immunodeficiency virus, chronic lung disease, tobacco use, obesity, pregnancy, chronic kidney or liver disease, transplant history, malignancy or use of immunosuppressive agents between the two groups.  On presentation, patients who met ILI criteria were more likely to have an elevated white blood cell count, 11.1 vs. 9.1 K/µL (p<0.0001), a higher absolute lymphocyte count, 2375.3 vs. 1396.9 mm³ (p<0.0001), and a lower serum albumin, 2.96 vs. 4.07 g/dL (p<0.0001).  The incidence of bacterial co-infection did not differ between those with ILI vs. those without ILI [56 (20.6%) vs. 67 (24.1%), p=0.32].  Patients who met the ILI definition were less likely to require mechanical ventilation (p=0.05) and were more likely to report symptom resolution during their hospitalization (p=0.002).  Overall, patients who met ILI criteria had shorter hospital lengths of stay (8.6 vs. 10.1 days, p<0.001) and significantly lower mortality rates (2% vs. 9%, p<0.0001).

Conclusions: ILI criteria are a non-specific screen for influenza met by a large proportion of patients with non-influenza RVIs.  In adults with non-influenza RVIs requiring hospitalization, meeting ILI criteria was associated with less severe disease, shorter lengths of hospital stay and decreased mortality despite no observed difference in underlying co-morbidities.  Further study is needed to determine virus and/or host factors that influence worsening outcome in patients with non-influenza RVIs who do not meet ILI criteria.