Antiviral Treatment for Patients Hospitalized with Severe Influenza Infection May Affect Clinical Outcomes

To the Editor—We congratulate McGeer et al. [1] on their prospective study of hospitalized patients with influenza, which shows that use of antiviral treatment (72% of their patients received treatment within 96 h after symptom onset) is associated with reduced mortality. McGeer et al. [1] clearly describe the burden of severe influenza and highlight the importance of early recognition and therapeutic intervention. We have reported that timely initiation (i.e., ≤48 h after illness onset) of oseltamivir treatment is independently associated with reduced length of hospitalization, after adjustment for age, comorbidities, and complications, among patients with severe influenza (n=356) [2]. In contrast with McGeer et al. [1], our diagnosis and treatment of influenza were based on a standard protocol [2]. Cox proportional hazards models were used to analyze the data, and we assumed that treatment initiated within 2 days of illness onset resulted in optimal efficacy [3, 4]. In our cohort, an overall mortality of 3.4% (12 deaths) was observed. Patients who died were not significantly different in age (mean age±SD, 65.4±19.1 years vs. 70.4±18.3 years; P>.05, by Student's t test) and sex (male sex, 58.3% vs. 51.2%; P>.05, by χ2 test), when compared with survivors. Mortality among patients who had major systemic comorbid illnesses as defined by the Pneumonia Severity Index classification (i.e., congestive heart failure and cerebrovascular, neoplastic, long-term liver, and renal diseases, [2]) was significantly higher (8.3% vs. 1.2%; P=.002, by Fisher's exact test) than mortality among patients with no major comorbid illnesses. Notably, we observed that the death rate among 232 patients who received oseltamivir within 96 h after illness onset was 2.2% (n=5), compared with 5.6% (n=7) among the 124 patients who did not receive such treatment (P=.120, by Fisher's exact test). Results of multivariate logistic regression analysis showed that the presence of a major comorbidity (OR, 9.3; 95% CI, 2.4–36.3; P=.001) and oseltamivir initiated within 96 h after illness onset (OR, 0.26; 95% CI, 0.08–0.87; P=.029) were independently associated with increased and decreased mortality, respectively. In a subgroup analysis comparing only patients who received “late” oseltamivir (i.e., initiated >48 h after illness onset) with those who received no treatment, the treated group appeared to have lower mortality (1.2% vs. 5.5%), although the difference did not reach statistical significance (P=.140, by Fisher's exact test). These observations support the conclusion of McGeer et al. [1] that antiviral treatment for severe human influenza infection may favorably affect clinical outcome. Efforts to ensure early diagnosis and therapeutic intervention are warranted, particularly among high-risk patients [1–4]. We admit that the number of deaths was small, and the results are limited by the retrospective nature of the analysis. Finally, although the data of McGeer et al. [1] and our own analysis presented above suggest the benefit of late antiviral therapy (initiated >48 h after illness onset) in this patient group, the issue was not directly addressed and thus remains inconclusive. However, we may add that, in a separate study involving hospitalized patients with influenza, a relatively high viral concentration (mean concentration, 4.48±0.65 log10 copies/mL) was detected among the late presenters (i.e., those presenting on day 3–4; n=8), which correlated with hypercytokinemia [5]. Whether antiviral treatment for these patients will provide benefit is not yet clear. More studies on viral load profiles among hospitalized patients with influenza and clinical trials of late antiviral therapy are needed to resolve the issue. Potential conflicts of interest All authors: no conflicts.