Influenza Challenge Models: Ready for Prime Time?

To The Editor—We wish to thank Dr Memoli and his team for their efforts to revive influenza challenge models. In their recent article they describe a new H3N2 influenza challenge model [1]. We certainly agree that controlled human infection models for influenza and other important infectious diseases can provide critical information for the development of novel and effective drugs and vaccines.

However, we believe controlled human infection models should employ protocols that assess clinical and virological endpoints that mimic the disease target. The current definition of mild to moderate influenza disease (MMID) as requiring only 1 positive viral polymerase chain reaction test (including on Day 1) and the report of only 1 influenza-specific or nonspecific symptom on a single day may not be sufficiently stringent to allow meaningful assessments of novel drugs and/or vaccines [2]. For example, symptoms included in the current MMID definition contain such nonspecific findings as headache, stomachache, nausea, and lack of appetite. Table 2 in their present publication [1] reports that although 6 of 8 (75%) subjects challenged with low doses of H3N2 developed “influenza” symptoms, none shed the virus. This suggests that these symptoms were unrelated to the challenge. In subjects given the higher H3N2 doses, only 55% had a detectable virus by influenza-specific polymerase chain reaction, while the frequencies of clinical symptoms were no higher than the frequencies in the low-dose challenge. Therefore, it is unclear whether the H3N2 challenges resulted in clinically significant influenza infections or whether the symptoms reported truly represent influenza infections. In our experience, subjects housed in challenge units often have headaches, stomachaches, and other nonspecific symptoms even when administered placebos. We are concerned that the use of this illness definition will not provide meaningful assessments of new drugs and/or vaccines or susceptibility to rechallenge [3].

The Memoli group has compared shedders (infected) to nonshedders (no infection) after challenge with their H1N1 challenge strain, and identified differences [4]. However, we suggest that using more stringent definitions of MMID will improve the use of this model. In earlier flu challenge studies, we have required subjects to have a fever (≥100°F) and at least 2 influenza symptoms on 1 or more evaluations, in addition to identifying nasal shedding of the challenge influenza virus strain by culture [5]. Other investigators using influenza challenges have required either multiple symptoms or Grade 2 events for participants to qualify as being ill. Others do not define illness but compare a compilation of signs and symptoms. Similarly, challenge studies of norovirus [6–8] have required specific gastrointestinal symptoms of either diarrhea, vomiting, or more than 1 symptom.

There is a clear need to develop influenza challenge strains that cause more significant illness [9]. However, it is also useful to explore alternate definitions of clinical illness that include multiple symptoms and, perhaps, eliminate nonspecific symptoms. This will increase confidence that the signs and symptoms are secondary to infection and are, indeed, modifiable by new treatments or prior infections. We recommend that the current data be reevaluated to determine a more stringent definition of MMID that will better differentiate infected from noninfected subjects postchallenge.

Note

Potential conflicts of interest. The authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.