Five years of monitoring for the emergence of oseltamivir resistance in patients with influenza A infections in the Influenza Resistance Information Study

Background and objectives The Influenza Resistance Information Study (IRIS) was initiated in 2008 to study the emergence of neuraminidase inhibitor (NAI) resistance and the clinical course of influenza in immunocompetent treated and untreated patients. Methods Patients had throat/nose swabs collected on days 1, 3, 6 and 10 for analyses of influenza type, subtype and virus susceptibility to NAIs. RT‐PCR‐positive samples were cultured and tested for NAI resistance by specific RT‐PCR and phenotypic testing. Scores for influenza symptoms were recorded on diary cards (Days 1‐10). This study focuses on influenza A‐infected cases only. Results Among 3230 RT‐PCR‐positive patients, 2316 had influenza A of whom 1216 received oseltamivir monotherapy within 2 days of symptom onset (9 seasonal H1N1; 662 H3N2; 545 H1N1pdm2009). Except for 9 patients with naturally resistant seasonal H1N1 (2008/9), no resistance was detected in Day 1 samples. Emergence of resistance (post‐Day 1) was detected in 43/1207 (3.56%) oseltamivir‐treated influenza A‐infected patients, with a higher frequency in 1‐ to 5‐year‐olds (11.8%) vs >5‐year‐olds (1.4%). All N1‐ and N2‐resistant viruses had H275Y (n = 27) or R292K (n = 16) substitutions, respectively. For 43 patients, virus clearance was significantly delayed vs treated patients with susceptible viruses (8.1 vs 10.9 days; P < .0001), and 11 (23.2%) remained RT‐PCR positive for influenza at Day 10. However, their symptoms resolved by Day 6 or earlier. Conclusions Oseltamivir resistance was only detected during antiviral treatment, with the highest incidence occurring among 1‐ to 5‐year‐olds. Resistance delayed viral clearance, but had no impact on symptom resolution.


| INTRODUCTION
Neuraminidase inhibitors (NAIs) are the mainline therapy of influenza. 1 Through binding in the conserved catalytic domain of the enzyme, these drugs can inhibit all types and subtypes of influenza neuraminidase, but to varying degrees. 2 In recent years, the human influenza A viruses have developed complete resistance to an older class of drugs, the adamantanes, indicating the ability of these viruses to develop and subsequently maintain resistance to antivirals. 3 In the first years of NAIs usage, following their introduction in 1999, naturally occurring resistance was sporadically reported and a very limited number of cases were described. [4][5][6][7] However, in 2008, naturally occurring oseltamivir resistance was detected among seasonal H1N1 viruses in Norway. 8 This resistant virus eventually displaced the NAI-susceptible H1N1 virus rendering virtually all seasonal H1N1 viruses highly resistant to oseltamivir. 8,9 This emergence was not related to the use of antivirals. 10,11 The resistant H1N1 virus was then replaced during the 2009-2010 pandemic by the influenza A H1N12009pdm virus, which was oseltamivir sensitive. 12 As a consequence of this emergence and dissemination of an NAIresistant virus, surveillance systems have been implemented to monitor antiviral susceptibility to NAIs. In this context, a global observational study was initiated in 2008, the Influenza Resistance Information Study (IRIS), to study the emergence of NAI resistance and the clinical course of influenza in immunocompetent treated and untreated patients.
The primary objective of the IRIS study was to assist with early detection of influenza resistance to antivirals and describe the clinical course and outcome of patients with influenza according to subtype and antiviral susceptibility.
Influenza Resistance Information Study is a prospective, multicentre, information-gathering study (NCT00884117). It is the largest study of its type that has collected sequential clinical and virological data during the course of infection, using sensitive RT-PCR detection methods for both detection of the virus and follow-up of substitutions associated with oseltamivir resistance in H1N1 and H3N2 viruses. Major findings of the first 3 years of this study have already been reported. 13 This article reports the first 5 years of surveillance carried out through IRIS, with a specific focus on the description of the emergence of influenza A-resistant viruses in treated patients, including the timeline of the emergence of the resistant viruses and the identification of the substitutions associated with this resistance.

| Patient inclusion and virological analysis
During this study period, the criteria for inclusion were as previously described. 13 Briefly, patients >1 year of age, presenting within 48 hours after disease onset of influenza-like illness and/or a positive rapid test result for influenza were eligible for enrolment.
Patients had throat or nasal swabs collected on days 1, 3 (self-swab), 6 and 10 for real-time reverse transcription PCR (RT-PCR) analyses of influenza type, subtype and susceptibility to NAI. NAI susceptibility was determined according to the IC 50 values performed on the viruses by a chemiluminescent assay (NA-Star), and the measure of the fold increase observed as compared to IC 50 values of susceptible strains, according to the common procedure. 14 gene segments) and phenotypically tested for NAI resistance when possible, as previously described. 16  ) were recorded daily on diary cards by the patient (fever, sore throat, nasal congestion, cough, myalgia, fatigue and headache), checked by the physician (days 1-10), and summed to produce a total symptom score as previously described. 13

| Biostatistical analysis
Kaplan-Meier plots were generated for time to viral RNA clearance and time to resolution of all diary card symptoms (no single symptom score of >1 on diary card). The statistical analysis was carried out by either Kruskal-Wallis or Wilcoxon signed-rank tests.

| RESULTS
During the 5-year surveillance period, 3230 patients with a single influenza strain infection were recruited. All study centres enrolled patients, the majority of whom (75.3%) were from the Northern Hemisphere ( Figure 1 As reported previously, the detection of resistance was significantly more frequent in the 1-to 5-year-old age group as compared with the combined older age groups (30/253 [11.85%] vs 13/909[1.43%]; P < .0001; Table 2). This was also observed when comparing the rate of resistance by subtype (16.1% vs 1.7% for H1N1pdm09 and 7.7% vs 1.2% for H3N2). In addition, we observed an increase in detection of

| DISCUSSION
In years 1-5 of the IRIS study, resistance to oseltamivir in influenza A viruses was not detected in the 2316 Day 1 samples analysed (except and R292K, no other substitution associated with resistance was detected in the NA. This suggests that as opposed to observations in immunocompromised, these 2 positions are almost the exclusive "hot spots" for changes associated with antiviral resistance in the context of oseltamivir pressure in immunocompetent patients. 14,20 In our study, the proportion of R292K substitutions in H3N2 viruses isolated (16/662, 2.4%) was relatively low as compared to Kiso et al. 21 The latter reported 9/50 (18%) of emerging resistant viruses during treatment in children with a first detection at day 4 of treatment. In our study, we report only 7.7% (10/129) of emerging resistance. This difference may be due to the recruitment of cases (majority of hospital cases in the Kiso study), the sampling procedures (nasal washes in some patients of the Kiso study vs swabs in our study) and the geographical distribution of our patients. However, we both support the idea that monitoring the resistance of influenza viruses requires analysis of sequential specimens collected in patients treated with oseltamivir. Surveillance with D1 samples only cannot provide a clear picture for an emerging resistance risk assessment, especially for H3N2 viruses.
According to analysis of the sequential specimens collected in these patients, resistant viruses emerged by Day 3 of treatment from susceptible strains and were selected by oseltamivir. It is known that the fitness of R292K H3N2 viruses is putatively severely impaired. 22 This detection of the R292K substitution was performed by a specific snip RT-PCR, a sensitive method that can detect down to 5% of a minority species. 15 It confirmed also that in most cases, a mixed population is detected, supporting the hypothesis that impaired NA activity of 292K viruses may be trans-complemented by the NA activity of R292 bystander viruses, as it has been reported for mutations at position 119. 23 The mutation-specific real-time RT-PCR used in the current study and HA or NA sequencing were performed on the original clinical specimen and not culture based or genetically assessed by sequencing on culture material. In the present study, of the 16 samples in which the R292K mutation was detected, only 5 could be cultured and subsequently tested phenotypically for NAI susceptibility. Two of them showed an increased IC 50 , while 3 had a normal IC 50 . This lack of detection of reduced inhibition may be due to the use of the Na-Star system (chemiluminescent assay) that is less sensitive than the gold standard MUNANA-based fluorescent assay. 24 The H275Y resistance in H1N1pdm09 was the most reported in this study . This is consistent with the various reports about oseltamivir resistance in N1 viruses in general. This resistance emerged from Day 3 in 1/3 of the 27 cases, not only in children (Table 2b). When, measured, the IC50 values were mostly reduced inhibition, and some were normal inhibition, due to mixed genotypes (H275 & 275Y) and the use of the chemiluminescent assay. The