Anti-inflammatory effects of adjunctive macrolide treatment in adults admitted to hospital with influenza: an open-label, randomised controlled trial

Abstract Background Influenza virus infections are causing substantial morbidity and mortality, despite availability of antiviral treatments. Macrolides have been shown to ameliorate inflammation in respiratory diseases and provide clinical benefits. Data in influenza, however, are scarce. We aimed to assess the anti-inflammatory effects of macrolide treatment in patients with influenza, and its effects on viral clearance and symptom resolution. Methods In this open-label, multicentre, randomised controlled trial, we recruited adults admitted to hospital for laboratory-confirmed influenza in Hong Kong. Key inclusion criteria were age 18 years or older, influenza A and B virus infections confirmed by PCR or immunofluorescence assays, and presentation within 4 days from illness onset. Patients were randomly assigned (1:1) using a computer-generated sequence to oseltamivir (75 mg twice daily) plus azithromycin (500 mg per day) or oseltamivir (75mg twice daily) alone, both given orally for 5 days. The primary outcome was change in plasma cytokine and chemokine concentration over time (day 0–10), analysed by intention to treat. Generalised estimating equation (GEE) models were used to analyse longitudinal data, and were adjusted for potential confounders. Ethics approvals were obtained from the institutional review bodies of all participating institutes. All patients provided written informed consent. This trial is registered with ClinicalTrials.gov, number NCT01779570. Findings During the influenza seasons beginning from 2013–14 through to 2015–16, 50 patients were randomly assigned to the oseltamivir-azithromycin (n=25) or oseltamivir (n=25) groups, with similar baseline characteristics (mean age 54·7 years [SD 18·5] in the oseltamivir-azithromycin group vs 58·6 years [18·1] in the oseltamivir group; 16 [64%] of 25 patients in the oseltamivir-azithromycin group were men vs 15 [60%] of 25 in the oseltamivir group). Three key pro-inflammatory cytokines declined faster in the oseltamivir-azithromycin group than in the oseltamivir group: interleukin (IL)-6 (GEE β=–0·037 [95% CI −0·067 to −0·007], p=0·016; change from baseline −83·4% vs −59·5%), IL-17 (β=–0·064 [–0·117 to −0·012], p=0·015; −74·0% vs −34·3%), and CXCL9 (β=–0·010 [–0·020 to 0·000], p=0·043; −71·3% vs −56·0%). Non-significant differences in the following cytokines were observed between treatment groups: CXCL8 (β=–0·018 [–0·037 to 0·000], p=0·056; −80·5% vs −58·0%), sTNFR-1 (β=–0·003 [–0·006 to 0·000], p=0·084; −40·1% vs −24·8%), IL-18 (β=–0·006 [–0·015 to 0·003], p=0·197; −29·1% vs 30·2%), and C-reactive protein (β=–0·033 [–0·088 to 0·022], p>0·10; −77·5% vs −48·2%). Two serious adverse events (SAEs) occurred in the oseltamivir-azithromycin group (Pseudomonas aeruginosa pneumonia, post-influenza vestibular neuronitis onset after stopping treatment for 1 week) versus one SAE in the oseltamivir group (increased ascites; p>0·99); all SAEs were considered unrelated to treatment. Other common adverse events were gastrointestinal or hepatic symptoms (five [20%] of 25 in the oseltamivir-azithromycin group vs four [16%] of 25 in the oseltamivir group; p>0.99) and dizziness or hearing symptoms (two [8%] vs two [8%]; p>0·99). All events were transient and reversible, and no participants died in this study. Interpretation We found significant anti-inflammatory effects with adjunctive macrolide treatment in adults with severe influenza infection. The clinical benefits of a macrolide-containing regimen deserve further study. Funding Research Grant Council of the Government of the Hong Kong Special Administrative Region, China (468112).