Influenza Transmission Trial (EMIT-2)
Evaluating Modes of Influenza Transmission (EMIT-2): Insights from Lack of Transmission in a Controlled Transmission Trial with Naturally Infected Donors
- Study type:
- Controlled human influenza virus transmission trial (CHIVITT) with naturally infected donors; prospective quarantine cohort study
- Topic:
- Infection & Transmission
- Added:
- 2026-11-10
- Status:
- published
Strength of evidence
Low – Controlled transmission experiment using natural (airborne/contact) routes — satisfying the design requirement for pathogen transmission trials. Direct clinical endpoint (influenza-like illness, ILI) was assessed. However: (1) confirmation of "infection" depends entirely on indirect diagnostic measures (qRT-PCR, cell culture, HAI, ELISA); (2) no pathogen isolation from all culture components is documented; (3) no negative controls with identically processed samples from healthy donors; (4) no blinding documented; (5) extremely small sample (5 donors, 11 recipients); (6) no null control arm. Baseline immunity indicators are internally inconsistent (HAI titres low in 8/11, but higher ELISA binding-antibody AUC in 9/11 versus Donors), which further limits interpretability. The authors' three explanations for the null result (low shedding, cross-reactive immunity, air mixing) are ad-hoc hypotheses without independent evidence and cannot resolve the finding.
Source
Lai J, Sobhani H, Coleman KK, Tai S-HS, Hong F, Sierra Maldonado I, et al. Evaluating modes of influenza transmission (EMIT-2): Insights from lack of transmission in a controlled transmission trial with naturally infected donors. PLoS Pathog. 2026;22(1):e1013153.
Open original studyIf influenza is highly contagious, what happens when five symptomatic patients share a poorly ventilated room with eleven healthy adults for 82 hours? Nothing. Zero secondary cases. This controlled trial (EMIT-2) put naturally infected donors and susceptible recipients together under real-world conditions — and not a single recipient developed influenza-like illness, tested positive, or showed serological evidence of infection.
Key findings
- 0 of 11 Recipients developed influenza-like illness (ILI) after 82.2 cumulative hours of close exposure across 23 events.
- 0 of 89 post-exposure respiratory samples (mid-turbinate swabs and saliva) tested positive by dPCR.
- 0 of 11 Recipients showed serological evidence of infection (no four-fold HAI titre increase; no ELISA AUC increase).
- Donors were confirmed influenza-positive (3x H3N2, 2x H1N1) with low Ct values (12.1-16.1) and 61% culture-positive mid-turbinate swabs, yet exhaled fine aerosol viral RNA was detected in only 44% of samples and culturable virus in just 6%.
- Median cough count during 30-minute exhaled breath collections was 0 (range 0-16).
- Baseline immunity indicators diverged: 8/11 Recipients had LOW HAI titres (<=10, i.e. susceptible by the HAI criterion), yet 9/11 had HIGHER ELISA binding-antibody AUC than the Donors. These two indirect measures point in opposite directions.
- Most Recipients reported more symptoms (mainly malaise) AFTER exposure than before, but none met ILI case definition and all indirect tests remained negative.
Background
- Conventional models hold that influenza "virus" transmits from infected to susceptible individuals via airborne droplets, aerosols or contact. The previous controlled trial (EMIT-1, 2013) used donors inoculated intranasally with egg-adapted virus and also produced minimal transmission (1 serologically positive, asymptomatic control recipient). EMIT-2 was designed to test transmission from naturally infected community cases under controlled conditions.
Methods
- Four quarantine cohorts were attempted (2023-2024); only two (24b, 24c) included both Donors and Recipients. Five Donors (mean age 21, 80% female; 3x H3N2, 2x H1N1) with community-acquired infection confirmed by molecular test (Cepheid GeneXpert Xpress) exposed 11 Recipients (mean age 36, 54% female) in a hotel room with low ventilation (0.25-0.5 air changes/hour) but high air recirculation. 23 exposure events lasted 111-250 minutes each (82.2 cumulative hours). Activities included conversation, games and light physical activity. Recipients were assigned to Intervention (face shields + hand hygiene) or Control groups. Sampling: exhaled breath (G-II), mid-turbinate swabs, saliva, surface swabs, ambient/personal bioaerosol samples. Analysis: dPCR, fluorescent focus assay (FFU), HAI, ELISA. Primary outcome: secondary attack rate (SAR) defined as virological confirmation, symptomatic infection (ILI), or serological evidence.
Results
- Zero secondary cases across all three outcome dimensions. 0/11 ILI, 0/89 PCR-positive post-exposure samples, 0/11 serological conversions. Donor exhaled breath: viral RNA in 7/16 (44%) fine aerosol samples (range 79-8.9 x 10³ copies/30 min), culture-positive in 1/16 (6%). One of 23 surface swabs (marker pen) was culture-positive. Two ambient air samples contained trace viral RNA. Estimated recipient exposure: ~750 RNA copies (24b) and ~290 RNA copies (24c) per person — well below the estimated infectious dose (~1.4 x 10⁵ RNA copies).
- Baseline serology showed a split picture: 8/11 Recipients had low HAI titres (<=10) at admission against the vaccine strains matching the Donor strains (6/8 in cohort 24b, 2/3 in cohort 24c), i.e. susceptible by the HAI criterion. In contrast, ELISA showed higher binding-antibody AUC values in 9/11 Recipients than in Donors at baseline — including all 8 Recipients of cohort 24b and 1/3 of cohort 24c. No Recipient showed a four-fold rise in HAI titre or ELISA AUC after exposure (the single exception had received influenza vaccination two days after discharge and is explained by the vaccination).
- Recipients reported sporadic mild symptoms (mainly malaise), with most reporting more symptoms after exposure, but none met ILI criteria and all indirect tests remained negative.
Interpretation
- The authors interpret the null result as reflecting three factors: (1) low donor contagiousness (mild season, infrequent coughing); (2) potential cross-reactive immunity in middle-aged recipients; (3) rapid air mixing reducing concentrated breath plume exposure. These are post-hoc explanations without independent supporting evidence. The directly observed result is straightforward: prolonged close contact with symptomatic individuals did not produce illness in any recipient.
- It is worth noting that the two indirect immunity markers used (HAI, ELISA) point in opposite directions: HAI classified 8/11 Recipients as susceptible (titre <=10), while ELISA showed 9/11 with higher binding-antibody AUC than the Donors. The authors build their "cross-reactive immunity" argument largely on the ELISA finding. Per strict evidence criteria, this internal inconsistency between indirect measures further limits the interpretability of the immunity hypothesis.
- The "influenza virus" and "transmission" framework used to interpret the findings is a theoretical model; the study's detection methods (dPCR, cell culture, HAI, ELISA) are indirect measurements per established evidence criteria.
Limitations
- Extremely small sample: 5 Donors, 11 Recipients across only 2 effective cohorts (2 of 4 attempted).
- No negative controls with identically processed samples from healthy donors (required for pathogen transmission experiments).
- No pathogen isolation from all culture components documented.
- No blinding (single or double) documented.
- No null control arm (IR vs CR compares interventions, not presence/absence of exposure).
- Primary outcome definition mixes direct (ILI) and indirect (PCR, serology) endpoints; "infection" attribution depends on indirect measures.
- Baseline immunity assessment is internally inconsistent: 8/11 Recipients had LOW HAI titres (<=10, susceptible by HAI criterion), yet 9/11 had HIGHER ELISA binding-antibody AUC than the Donors at baseline. The two indirect measures disagree, and the authors' cross-reactive-immunity argument rests mainly on the ELISA finding.
- Recipients were middle-aged (mean 36) with potential cumulative exposure to prior infections and vaccinations.
- Donors had mild symptoms with median 0 coughs per collection period; exhaled viral loads were at the low end of previous observations.
- Environmental controls (fan coils, dehumidifiers) created rapid air mixing that may have reduced short-range plume exposure — a confounding factor acknowledged by the authors.
- Three ad-hoc explanations for the null result lack independent evidence and cannot be accepted as resolution without further controlled data.
- Cohort 23a (2023) and 24a (2024) produced no usable data (no influenza Donors recruitable).
WorthDay perspective
- This study provides a striking directly observable result: 11 people spent 82 hours in a confined space with 5 symptomatic individuals, and not one became ill. That is a hard clinical endpoint.
- The conventional interpretation — that "influenza virus" simply failed to transmit due to low shedding, pre-existing immunity or air mixing — rests on three unconfirmed ad-hoc hypotheses. Each explanation is plausible within the germ-theory model, but none is supported by independent evidence within this study. Per strict evidence criteria, post-hoc explanations cannot resolve a null finding.
- The baseline immunity data deserve particular attention: the two serological measures used (HAI and ELISA) point in opposite directions. HAI classified 8/11 Recipients as susceptible (titres <=10), while ELISA showed 9/11 with higher binding-antibody AUC than the Donors. The authors' "cross-reactive immunity" argument is built on the ELISA result. When two indirect measures of the same hypothetical construct disagree, neither can be treated as established — and the immunity-based explanation becomes even more tentative.
- Notably, most Recipients DID report increased symptoms (mainly malaise) after exposure. These clinical observations — direct endpoints — were dismissed as non-infection because indirect diagnostic tests (PCR, serology) remained negative. This illustrates the endpoint hierarchy problem: clinical symptoms (direct) are subordinated to surrogate markers (indirect).
- The study also confirms what EMIT-1 (2013) already showed: controlled transmission experiments with either inoculated or naturally infected donors have consistently failed to demonstrate the predicted secondary attack rate. The "influenza transmission" hypothesis remains unconfirmed under controlled experimental conditions.
- Note: Evidence level is low (🟠) because the experimental design includes a direct clinical endpoint (ILI), but "infection" confirmation depends entirely on indirect measures, baseline immunity markers disagree internally, no proper negative controls exist, and the sample is extremely small. The findings are consistent with a lack of demonstrated transmission but cannot confirm or refute the broader theoretical framework without more rigorous controlled experiments.
Notes
Controlled human influenza transmission trial with naturally infected donors (5 donors, 11 recipients, 82.2 h exposure). Zero secondary cases: no ILI, no PCR-positive samples, no serological infection. 8/11 Recipients had low baseline HAI titres (<=10), but 9/11 showed higher ELISA binding-antibody AUC than Donors. Authors offer three ad-hoc explanations (low shedding, cross-reactive immunity, air mixing) without independent evidence. Direct clinical endpoint (ILI) is negative; infection confirmation relies on indirect measures. Low evidence (🟠).
Permalink: https://www.worth-day.com/science-compass/s0066-influenza-transmission-trial-emit-2
Related studies
Science meets practice