Patients and Methods

Methods

Virus isolation and titration. Clinical specimens (throat swabs) that were found to be virus-positive by use of rapid diagnostic kits were stored at −80°C until virus isolation. Madin-Darby canine kidney cells were used for viral isolation and plaque assay. Viruses were subtyped by conventional hemagglutinin- and neuraminidase-inhibition assays.

Plaque assay. A modified plaque-forming assay was used [8]. Monolayer cell cultures of Madin-Darby canine kidney cells in 6 wells were infected with 10-fold serial dilution of the clinical sample. After adsorbing at 35°C in 5% carbon dioxide for 1 h, the infected cells were overlayered with 2% agarose-minimal essential medium and were incubated at 35°C in 5% carbon dioxide for 3 days. After being fixed with 10% formalin, the plaques were stained with crystal violet.

Statistical methods. We used Student's t test for all statistical comparisons of mean values, and we used the χ² test to evaluate differences of fever duration in patients with influenza. Data are expressed as mean values (±SD).

Verbal informed consent was obtained from each child's parent or guardian and was recorded on the chart. Informed consent was also obtained from children >7 years of age who were able to understand the concepts and procedures of the protocol. This study was conducted with the approval of the ethics committees of our hospitals.

Results

Comparison of the effectiveness of oseltamivir and zanamivir against influenza A (H3N2). There was no statistically significant difference in total febrile period or duration of fever after the start of treatment between the patients treated with oseltamivir and the patients treated with zanamivir. During the 2006–2007 epidemic, there were 8 patients with influenza A (H3N2) infection who did not receive treatment with any antiviral agents (mean age ± SD, 10.6±2.6 years; range, 4.9–14.2 years), and their mean total febrile period (±SD) was 4.31±2.07 days. Thus, the total febrile period was shortened by ∼2 days in both the group treated with oseltamivir (P<.05) and the group treated with zanamivir (P<.05).

In both groups, the body temperature of approximately one-half of the patients returned to normal within 24 h, and in ∼90%, the body temperature returned to normal within 48 h (table 1). There were no statistically significant differences in the percentages between the groups.

Comparison of the effectiveness of oseltamivir and zanamivir against influenza A (H1N1). There was no statistically significant difference in total febrile period or the duration of fever after the start of treatment between the group treated with oseltamivir and the group treated with zanamivir. In both groups, the body temperature of one-third of the patients returned to normal within 24 h, and the body temperature of ∼80% of the patients returned to normal within 48 h. The differences between the groups was not statistically significant.

Comparison of the effectiveness of oseltamivir and zanamivir against influenza B. There was no statistically significant difference in total febrile period or duration of fever after the start of treatment between the group treated with oseltamivir and the group treated with zanamivir. During the 2006–2007 epidemic, there were 47 patients with influenza B virus infection who did not receive treatment with any antiviral agents (mean age ± SD, 9.8±3.0 years; range, 4.3–13.4 years), and their mean total febrile period (±SD) was 4.02±1.26 days. Thus, the total febrile period was shortened by ∼1 day in both the group treated with oseltamivir (P<.01) and the group treated with zanamivir (P<.01).

In both groups, body temperature returned to normal within 24 h in ∼40% of patients and within 48 h in 70%–80% of patients. There were no statistically significant differences in the percentages between the groups.

Effectiveness of oseltamivir against influenza A (H1N1), influenza A (H3N2), and influenza B virus. Among the patients treated with oseltamivir, the duration of fever after the start of treatment was significantly shorter in patients with influenza A (H3N2) infection than in patients with influenza A (H1N1) infection (mean duration, 1.35 days vs. 1.79 days; P<.05) or in patients with influenza B (mean duration, 1.35 days vs. 1.86 days; P<.01) (table 1). The total febrile period was statistically significantly shorter in patients with influenza A (H3N2) virus infection than in patients with influenza B (mean duration, 2.40 days vs. 2.95 days; P<.01), and it was shorter, although the difference was not statistically significant, in patients with influenza A (H3N2) infection than in patients with influenza A (H1N1) infection (mean duration, 2.40 days vs. 2.60 days).

The percentages of patients whose temperature became normal were statistically significantly higher among patients with influenza A (H3N2) infection than among patients with influenza A (H1N1) infection (within 24 h, 56.0% vs. 33.3% [P<.05]; within 48 h, 93.4% vs. 79.2% [P<.05]) or patients with influenza B (within 24 h, 56.0% vs. 38.3% [P<.01]; within 48 h, 93.4% vs. 71.1% [P<.01]) (table 1). The total febrile period and duration of fever after the start of treatment were shorter among patients with influenza A (H1N1) infection than among patients with influenza B, although the difference was not statistically significant.

Effectiveness of zanamivir against influenza A (H1N1), influenza A (H3N2), and influenza B. Among the patients treated with zanamivir, there was no statistically significant difference between patients with influenza A (H1N1) and patients with influenza A (H3N2) with regard to total febrile period or the duration of fever after the start of treatment (table 1). However, the total febrile period was statistically significantly shorter among patients with influenza A (H3N2) infection than among patients with influenza B (2.39 days and 2.84 days, respectively; P<.05), and the duration of fever after the start of treatment was also shorter among patients with influenza A (H3N2) than among patients with influenza B, but the difference in duration of fever was not statistically significant.

There was no statistically significant difference between the patients whose temperature became normal within 24 h and those whose temperature became normal within 48 h. However, the percentages for both groups were higher for patients with influenza A (H3N2) infection than for patients with influenza A (H1N1) infection or patients with influenza B (table 1).

Influenza A (H3N2) virus isolation rate after the start of neuraminidase inhibitor therapy. On the first day of treatment, influenza A (H3N2) virus was detected in throat specimens obtained from every patient (100%). On the third day of treatment, there was no statistically significant difference between the virus isolation rate in the group treated with oseltamivir and the rate in the group treated with zanamivir. However, on the fifth day of treatment, the isolation rate in the group treated with oseltamivir was 41.2%, compared with only 6.3% in the group treated with zanamivir, a statistically significantly lower rate (P<.05) (table 2).

Influenza A (H1N1) virus isolation rate after the start of neuraminidase inhibitor therapy. Influenza A (H1N1) virus persisted in throat specimens obtained from a high percentage of the patients after the start of treatment, although the total number of patients was small. Even on the fifth day of treatment, the isolation rate was 30%–40% in both the group treated with oseltamivir and the group treated with zanamivir (table 2).

Changes in virus shedding by patients with influenza A (H3N2) infection. Before neuraminidase inhibitor therapy was initiated (i.e., on day 1), mean infection titer (±SD) was similar—4.23±0.67 in the group treated with oseltamivir (20 patients) and 4.10±0.98 in the group treated with zanamivir (18 patients) (figure 1). After the start of therapy, the virus titers of the specimens from both treatment groups decreased rapidly. On the third day of treatment, the mean virus titers (±SD) in the group treated with oseltamivir (15 patients) and the group treated with zanamivir (17 patients) were almost the same (1.21±1.17 vs. 1.20±1.25, respectively). On the fifth day of treatment, the group treated with oseltamivir (17 patients) had a statistically significantly higher infection titer than did the group treated with zanamivir (16 patients) (0.76±1.08 vs. 0.07±0.29; P<.05). On the fifth day of treatment, 15 of the 16 patients in the group treated with zanamivir did not shed viruses.

Changes in virus shedding by patients with influenza A (H1N1) infection. Before neuraminidase inhibitor therapy was initiated (i.e., on day 1), mean virus infectivity (±SD) was 4.25±0.58 in the oseltamivir group (7 patients) and was 3.80±0.93 in the group treated with zanamivir (7 patients). After the start of treatment, there was no statistically significant difference between the group treated with oseltamivir and the group treated with zanamivir (mean virus infectivity ± SD on the third day of treatment, 1.56±1.02 [7 patients] vs. 1.77±1.77 [6 patients]; mean virus infectivity ± SD on the fifth day of treatment, 1.10±1.39 [6 patients] vs. 0.88±1.18 [7 patients]) (figure 2).

Discussion

To our knowledge, this is the first study to compare the clinical effectiveness of oseltamivir with that of zanamivir against influenza virus infection in children. The comparison between the oseltamivir-treated group and the zanamivir-treated group in this study confirmed that both neuraminidase inhibitors—oseltamivir and zanamivir—are equally effective in reducing the febrile period of children with influenza A (H1N1), influenza A (H3N2), and influenza B infection (table 1).

However, the results showed that oseltamivir had lower clinical effectiveness against influenza A (H1N1) than against influenza A (H3N2). Among patients with influenza A who were treated with oseltamivir, the duration of fever after the start of treatment was significantly longer for patients with influenza A (H1N1) infection, and the percentage of patients who became afebrile was statistically significantly lower for the patients with influenza A (H1N1) infection (table 1). These results were obtained in a double-blind fashion; the diagnosis of influenza A was based on the results of rapid diagnostic tests, so neither the patients nor the pediatricians knew whether the patients had influenza A (H1N1) or influenza A (H3N2) infection.

We have reported elsewhere the finding that oseltamivir is less effective against influenza B infection than it is against influenza A (H3N2) infection [1], and lower effectiveness of oseltamivir against influenza B was confirmed in the present study (table 1). Thus, the present study demonstrated that oseltamivir is clinically more effective against influenza A (H3N2) infection than it is against influenza A (H1N1) or influenza B virus infection. No statistically significant difference was found between oseltamivir's effectiveness against influenza A (H1N1) infection and its effectiveness against influenza B.

Zanamivir has been reported to be equally effective against influenza A and influenza B [9]. However, we found a statistically significantly shorter total febrile period in patients with influenza A (H3N2) infection than in patients with influenza B infection (table 1). Zanamivir is probably less effective against influenza B than it is against influenza A (H3N2) infection, as is the case with oseltamivir. Zanamivir was also less effective against influenza A (H1N1) infection than it was against influenza A (H3N2) infection, although the difference was not statistically significant (table 1).

The effectiveness of neuraminidase inhibitors is basically determined by the susceptibility (IC₅₀) of influenza viruses [1]. Our results for the clinical effectiveness of oseltamivir apparently reflect the reported oseltamivir IC₅₀ values for influenza viruses: 0.3 nmol/L for influenza A (H3N2) [2], 0.66 nmol/L [10] and 1.34 nmol/L [11] for influenza A (H1N1), and 75.4 nmol/L for influenza B [1].

Although there have been small differences in the reported susceptibility (IC₅₀) of influenza A (H1N1) virus to zanamivir—for example, 0.37 nmol/L [10] and 0.92 nmol/L [11] for influenza A (H1N1) virus and 1.6 nmol/L [3] for influenza A (H3N2) virus—because the concentration of zanamivir in the respiratory tract is much higher [12], differences in IC₅₀ values may not be evident. On the other hand, there was a significant difference in the total febrile period between patients with influenza A (H3N2) infection and patients with influenza B, probably because of much lower susceptibility of influenza B virus to zanamivir (10.1 nmol/L) [3].

The isolation rate of influenza A (H3N2) virus from the throat on the fifth day of treatment was statistically significantly higher in the oseltamivir-treated group than in the zanamivir-treated group (43.8% vs. 6.3%), suggesting that zanamivir is more effective than oseltamivir in clearing influenza A (H3N2) virus from the respiratory tract. Because influenza A (H3N2) virus is more susceptible to oseltamivir than to zanamivir (0.3 nmol/L [2] vs. 1.6 nmol/L [3]), the difference in isolation rate of influenza A (H3N2) virus between the oseltamivir-treated group and the zanamivir-treated group is probably attributable to the higher concentration of zanamivir in the respiratory tract [12].

It is unknown whether the influenza viruses at such low infection titers as 0.76 plaque-forming units/mL in the oseltamivir-treated group on the fifth day of treatment are capable of causing infection in others (figure 1). However, prolonged viral shedding is an indispensable factor for induction of resistance to oseltamivir, because oseltamivir-resistant strains emerge 3–5 days after the start of therapy [2, 3]. Even on the fifth day of treatment, influenza A (H3N2) virus was isolated from throat specimens obtained from 41.2% of the oseltamivir-treated patients in this study, whereas most subjects in the zanamivir-treated group had no virus isolated from throat specimens on day 5. We found that influenza A (H3N2) developed resistance to oseltamivir easily. In contrast, we did not detect influenza A (H3N2) viruses that were resistant to zanamivir. A detailed analysis of strains that are resistant to oseltamivir and zanamivir will be reported elsewhere.

There was no statistically significant difference in the isolation rate of influenza A (H1N1) virus between the oseltamivir-treated group and the zanamivir-treated group (table 2), and viral shedding persisted in both groups, which possibly led to resistance or lower susceptibility to both neuraminidase inhibitors (figure 2). Because of the small number of patients, more data will be needed before the clinical implications of this finding become clear.

In conclusion, we compared the clinical effectiveness of oseltamivir and zanamivir in children with influenza and found that both neuraminidase inhibitors were equally effective in reducing the febrile period in influenza A (H1N1), influenza A (H3N2), and influenza B virus infection.