Evidence Paper Review

Visual summary (figure interpretation)

The log‑scale chart shows a clear experimental result from the paper: amantadine/rimantadine IC50s for influenza A are in the sub‑µg/mL range (≈0.1–0.4 µg/mL) while the same drugs require >40–100 µg/mL to inhibit influenza B — a >100‑fold shift — whereas ribavirin shows mid‑µg/mL activity against both A and B (≈2.6–6.8 µg/mL). These values are directly from the study's reported means and example isolates and determine whether therapeutic blood concentrations are likely to be inhibitory in vivo (authors compare to human PK). Hayden et al., IC50 values and methods

Concise critical appraisal (evidence + limits)

• Strengths: clear method with MDCK + trypsin enabling both A and B plaques; triplicate assays and repeated measures reported; IC50s are compared to human pharmacokinetic levels to judge clinical relevance (paper links in vitro numbers to achievable blood/nasal levels and to then‑available clinical trial outcomes).
• Transparent limitation acknowledged by authors: results depend on cell line and passage history (e.g., A/PR/8/34 highly passaged strain is less susceptible), and in vitro IC50s do not capture host immune effects or in vivo pharmacodynamics — the paper explicitly discusses these caveats and compares to clinical data where possible.
• Methodological cautions: adsorption time and presence of drug during adsorption change plaque counts; authors standardized adsorption to 60 min with drug present to reduce variability, but other labs using different adsorption/overlay conditions may obtain different IC50s (a known issue with plaque inhibition–type assays).
• Reproducibility: authors present repeated assays (means ± SD) showing acceptable within‑laboratory reproducibility; raw per‑isolate tables are summarized rather than fully enumerated in text, and there is no public dataset, which limits cross‑lab reproduction without repeating experiments.

Hayden et al., assay reproducibility & limitations

Where the paper sits in practical terms

Applied implication: by reporting IC50s near human steady‑state amantadine blood levels (0.3–1.2 µg/mL quoted in the paper) the assay correctly predicted amantadine's clinical utility for influenza A and lack of utility for influenza B — the authors explicitly use PK/PD context to interpret the in vitro data rather than overclaiming translational certainty. Hayden et al., PK/PD interpretation

Concrete reproducibility checklist (to replicate the assay)

1. Use MDCK monolayers (approx. 10^5 cells/dish) and ensure confluent monolayers (3–5 days after seeding).
2. Prepare drug dilutions freshly in double‑distilled water and in double‑concentrated MEM for adsorption step; include drug in adsorption fluid at same concentration as overlay.
3. Adsorb virus (50–150 PFU) for 60 min at 36°C with frequent shaking, then overlay with 0.6% agarose containing trypsin (2 µg/mL) and drug at target concentrations.
4. Incubate 36–48 h, stain with neutral red, count plaques; calculate % inhibition vs drug‑free infected control and determine IC50 by probit or other logistic dose–response analyses.
5. Include cytotoxicity controls for each drug (drug only, no virus) to identify concentrations that produce cell toxicity (authors observed toxicity at 100–200 µg/mL for adamantanes in MDCKs).

(All above details, including adsorption time and overlay composition, are taken from the methods and results sections of the paper.) Hayden et al., methods

Main critical caveats & blindspots

• Cell model dependence: MDCK cells are convenient and permissive; however, receptor distribution and drug uptake differ from human airway epithelium — IC50s can shift with different cell systems or overlays (authors and later literature note cell‑type dependence).
• Passage history effect: highly passaged laboratory strains (e.g., A/PR/8/34) can acquire altered susceptibility; the paper documents this and warns about generalizing from lab strains.
• PK/PD simplification: comparing in vitro IC50 to plasma concentrations is helpful but incomplete — tissue distribution, intracellular concentrations, active metabolites, and immune response all shape clinical outcomes; the authors correctly avoid overclaiming beyond these constraints.
• Data sharing: raw per‑isolate tables are summarized in the paper; no public datasets / accession numbers are provided, which was standard for 1980 but limits modern meta‑analysis without repeating experiments.

Hayden et al., caveats

What would change my confidence?

Evidence that would substantially change the interpretation: cross‑lab reproducibility showing large shifts in IC50s when identical strains/conditions are used (would undermine assay reliability), or clinical trials demonstrating that drugs predicted inactive by the assay are nonetheless clinically effective (would falsify assay predictive value). Conversely, multi‑center standardization studies reproducing the reported IC50 ranges would strengthen external validity.

Hayden et al., PK/clinical comparison