The mechanism is real. The window is three days.

A recurring claim in the popular medical-news cycle is that hydroxychloroquine, ivermectin, and ribavirin are effective against hantavirus cardiopulmonary syndrome (HPS) and have been suppressed by health authorities. The pharmacology underlying the claim is not fabricated. The mechanisms are real. The animal signal — chloroquine against Andes virus in a hamster model, ribavirin in the Safronetz 2011 study — is real. The direction of effect is not wrong.

The problem is not the drug. The problem is the disease.

Hantavirus cardiopulmonary syndrome does not kill through unchecked viral replication. It kills through a self-sustaining immune cascade that, once triggered, decouples from the virus that started it. You can block every new virion from entering a cell, and the patient will still die of pulmonary oedema. You can cut viral load by four orders of magnitude, and the lungs will still fill.

We built a coupled model of hantavirus cardiopulmonary syndrome — eight differential equations covering viral load, endothelial infection, capillary leak, pulmonary oedema, lactic acidosis, IL-6, platelets, and bradykinin — and ran the relevant drug regimens through it. A single patient sick enough that the immune system offers nothing on its own (host resilience = 0). Then a cohort of one hundred patients with variable resilience, age, and exposure dose.

What the model says, by treatment-start day

Calibration baseline: supportive care alone yields 41% mortality in the cohort. That sits inside the CDC surveillance range (36%) and matches the design baseline of the Mertz 2004 ribavirin trial, so the model reproduces real-world HPS outcomes before any drug is added.

• HCQ, prophylactic (day −1): 12% mortality. Real effect — the drug blocks viral entry through pH-dependent endosomal inhibition, and the model gives it full credit for that.
• HCQ at symptom day 1: 14% mortality.
• HCQ at symptom day 5: 24% mortality. Benefit roughly halved.
• HCQ at 800 mg: 14% — identical to 400 mg. The entry-block mechanism saturates; more drug buys nothing.
• Ivermectin alone at day 1: 23% mortality. Weaker than HCQ at the same timing, consistent with its narrower mechanism (importin-mediated nuclear transport block — important for viral production but a single point of leverage).
• Ivermectin at day 5: 30%. Barely better than doing nothing.
• HCQ + ivermectin combined, day 1: 4% mortality. This is the strongest signal in the model and it is not additive — it is synergistic. The two mechanisms cover different entry points: HCQ blocks viral entry; ivermectin blocks production from cells already infected. Together they suppress viral replication enough that the cascade never fully engages in most patients.
• HCQ + ivermectin combined, day 5: 20% mortality. Synergy gone. Cascade already firing.
• Ribavirin at day 1: 0% mortality. This reproduces the Safronetz 2011 hamster result in a human-physiology setting — give it before the cascade fires, and the patient clears.
• Ribavirin at day 4.5 (median enrolment time in Mertz 2004): 13% mortality. The model reproduces the Mertz null — the drug still lowers viral load, but the downstream endpoint (pulmonary oedema, mortality) no longer moves. The wrong-stage thesis, confirmed in silico.
• ECMO only: 11% mortality, matching the modern ECMO-era HPS baseline.

What the numbers say

Every regimen that works early loses most or all of its benefit by symptom day five. Hantavirus has an aggressive incubation (1–4 weeks) but a deceptively generic prodrome: fever, myalgia, headache — indistinguishable from any viral illness. The patient does not know they have HPS. Neither does their GP. It looks like flu.

The first symptom specific enough to trigger a hantavirus test is shortness of breath. That arrives four to seven days after the prodrome starts. By the time a clinician has drawn blood, confirmed serology, and considered ribavirin — by the time anyone knows this is hantavirus and not influenza — the cascade is already at or past the point where these drugs can stop it.

The model says: a prophylactic or day-1 HCQ + IVM regimen would produce close to 4% mortality in HPS. That is not a small number. It is a meaningful, clinically significant survival benefit.

The problem is that there is no patient population to give it to. No endemic region screens for hantavirus exposure weekly. No diagnostic predicts which febrile patient in the prodrome has HPS versus a thousand other viruses. The drugs work in a window that closes before the disease announces itself.

The asymmetry

You can be right about the drug and wrong about the disease. The model gives HCQ, ivermectin, and ribavirin full credit for their documented in-vitro and animal activity against hantavirus. The model also shows that the credit expires in roughly three days from symptom onset, and HPS does not get diagnosed in three days.

This is not a unique pattern. The same physics applies to every antiviral tested against a disease whose pathology is driven by a post-replication cascade — SARS-CoV-2, dengue, Ebola, the list goes on. An upstream drug cannot fix a downstream process once the process has committed. The bottleneck is not drug availability or potency. It is the gap between the first symptom and the cascade becoming self-sustaining. No oral drug regimen tested so far closes that gap.

Where the regulatory position is, and isn’t, wrong

The WHO’s position on ivermectin for hantavirus — broadly dismissive — is defensible when you constrain the question to what a clinician can do for a patient who walks into a hospital with HPS. At that stage, the model says, ivermectin alone buys 11 percentage points of mortality reduction (41% → 30%). That is not nothing. It is also not a miracle.

The WHO does not address HCQ prophylaxis or post-exposure prophylaxis for hantavirus, because HPS is not a continuous epidemic. There is no standing indication. The question that matters — whether a HCQ + IVM regimen in an endemic community with a known outbreak could cut mortality meaningfully — is a question the regulator has not been asked to answer, because the clinical trial data does not exist. The model suggests the answer would be yes, with caveats about timing, saturation, and cardiac risk.

That is not suppression. That is absence of evidence, which is a different thing.

What the field could do

The constructive version of this argument is not “the regulators are wrong.” It is: the trial design needed to answer the timing question has not been run. A prospective trial of post-exposure prophylaxis in an outbreak-endemic community, with HCQ + ivermectin started within 24 hours of a known exposure event, is a study the field could do. The model suggests the effect size would be large. Whether the logistics of running it inside an endemic community are feasible is a question for the public-health community, not for the modelling community.

Until that trial exists, the honest statement is the model’s: the mechanism is real, the window is three days, and the bottleneck is diagnostic.