C-Reactive Protein (CRP) levels in influenza and other respiratory viral infections

Background: Viruses are major causes of upper respiratory tract infections (URTIs). Yet, antibiotics are commonly prescribed for URTIs. Point-of-care C-reactive protein (CRP) tests can be used to guide antibiotic prescribing for URTI. We sought to assess for the associations between CRP levels and influenza and other respiratory viral infections. Methods and materials: From June 2016 through November 2018, we screened 715 adults presenting with uncomplicated URTI (ICD10-AM J00-J06) at the emergency department of a 1600-bed acute hospital in Singapore for 15 major respiratory viruses using a multiplex PCR respiratory virus pathogen panel (Seeplex RV15 ACE Detection), via throat and nasal swabs, and CRP levels using a finger-prick point-of-care test (QuikRead go). We compared CRP levels between influenza and other viruses, and assessed for independent associations between CRP levels and influenza and other viral infections using multivariable multinomial logistic regression models. Results: Participants were young (median age 36 [IQR 28–51] years) and predominantly healthy (67.8% had no pre-existing illness). Almost half (48%) of them had a respiratory virus detected, with influenza (20.6%) and rhinovirus (14.4%) being the two most commonly detected viruses. CRP levels were significantly different between patients with influenza, other respiratory viruses, and no virus detected, with 32%, 46%, and 53% respectively having CRP ≤ 5 mg/L (P < 0.001). A CRP level of 6–20 mg/L was 1.7 times as likely to be associated with other respiratory viral infections (AOR 1.73, 95% CI 1.13–2.65) and 3.4 times as likely with an influenza infection (AOR 3.40, 95% CI 2.09–5.52) as a non-viral URTI, after adjusting for age, comorbidity, travel in the past 7 days, prior influenza vaccination within 6 months, month and year of illness episode. A CRP level of 21–40 mg/L was also twice as likely to be associated with an influenza infection (AOR 2.38, 95% CI 1.26–4.47) as a non-viral URTI. Conclusion: A mildly raised CRP level of 6–20 mg/L was more likely to be associated with a viral than non-viral URTI, although a higher CRP level of 21–40 mg/L was more suggestive of an influenza infection. Point-of-care CRP tests could provide clinical decision support for antibiotic use in uncomplicated URTIs.