SARS-CoV-2 is essentially a hybrid: it spreads like a common-cold coronavirus (upper airway infection, pre-symptomatic shedding, R0 ≈ 2–3) but injures the lung like SARS-1 or MERS — that combination of transmissibility and pathogenicity is exactly what made it a pandemic while its predecessors were containable.
2
Antibody immunity to coronaviruses — including common-cold strains — wanes within a year or two; mild or asymptomatic infections generate weaker and shorter-lived immunity, which has direct implications for both natural herd immunity and vaccine durability planning.
3
SARS-1 was eradicated through classic quarantine precisely because it was only contagious after a patient was visibly ill and never spread to animal reservoirs; SARS-CoV-2 has neither of those containment-friendly features, meaning it will never be eradicated — only managed.
4
The infection fatality rate of SARS-CoV-2 is much closer to 1–2% than early estimates of 5–10% once the large denominator of asymptomatic and mild cases is accounted for — making it more comparable to a very severe influenza than to MERS or SARS-1 on a per-infection basis.
Protocols
Concrete recipes — what, when, how much, and why
5 items
Two-phase therapeutic strategy for severe coronavirus disease: antiviral early, immune modulator late
WhatTreat early-stage infection (high viral load, low inflammation) with antiviral therapy to suppress replication; transition to immune-modulating/dampening therapy for late-stage disease dominated by cytokine storm. Do not deploy these in the wrong sequence.
WhenAntiviral: from symptom onset or first positive test through roughly the first 5–7 days. Immune modulator: when cytokine-storm markers (IL-6, ferritin, CRP) are rising and viral load is declining — typically day 7+ in severe cases.
DoseNot specifically dosed in the episode; Perlman references remdesivir (then in trials) as the oral-form hope for early antiviral, and immune-modulating agents in the second phase.
For whomHospitalized patients with moderate-to-severe COVID-19. The boundary between the two phases is a clinical judgment requiring biomarker guidance.
WhySARS-CoV-2 lung injury has two pathologically distinct phases: viral-mediated (high viral load directly damages pneumocytes) and immune-mediated (exuberant, misdirected immune response destroys lung parenchyma). Giving immune suppressors during the viral phase could be deleterious by allowing viral replication to accelerate unchecked.
CaveatsPerlman cautions that giving immune activators (interferon-type agents) once the cytokine storm is established could worsen outcomes — timing is everything. This is not a protocol for outpatient mild illness.
Attia draws the analogy to oncology: in cancer we have first-line, second-line, and third-line therapies; the idea that a single drug is simply 'good or bad' for COVID-19 misses the time-dependence of the biology. Perlman agrees and notes this same biphasic model applies to severe influenza (h5n1 shares the same immune-storm phenotype). The concept also maps onto the SARS-1 era, when researchers tried to balance antivirals with corticosteroids — those early attempts at empirical immune suppression sometimes made outcomes worse in patients who still had high viral loads.
Mechanism
Early phase: viral replication in type-II pneumocytes drives direct cell death and releases PAMPs that trigger innate immune activation. Late phase: excessive cytokine secretion (cytokine storm) drives secondary tissue damage via macrophage activation syndrome-like pathology, even after viral load has peaked.
You need antiviral therapy early on and then maybe an immune modulator later and it becomes an immune activator if you just catch it just at the right time maybe that would help.
Serial biomarker sampling for disease-trajectory stratification
WhatIn patients with documented SARS-CoV-2 infection, take serial blood samples every 48–72 hours to measure a panel of cytokines, metabolites, and standard inflammatory markers. Feed these into a risk model (ideally machine-learning-based) together with demographic and comorbidity data to predict who will progress to pneumonia before it happens.
WhenBeginning at first positive test or first clinical presentation, continuing through the first 10–14 days of illness.
For whomPriority for patients with at least one risk factor: age >60, type 2 diabetes, obesity, or existing cardiopulmonary disease. Younger healthy patients without comorbidities have a lower pretest probability of progression and would require a lower threshold from the model.
WhyThe same clinical presentation (fever, myalgia, malaise) can correspond to completely different biological trajectories: self-limiting upper-respiratory infection in 80% vs. progressive pneumonia in 20%. A stratification tool would allow targeted early intervention for the high-risk group and avoid over-treating the low-risk group with therapies that carry their own risks.
CaveatsNot yet implemented at the time of recording; requires prospective longitudinal cohort design and large labeled datasets to train models. Perlman acknowledges we did not yet know which cytokine signature was most discriminating.
Perlman describes the core problem: cytokine X is elevated in people who will get worse, but not so much that populations separate cleanly — there's a continuous distribution with overlapping tails. A single marker does not work; a multivariate model combining cytokine levels, metabolomics, demographics, and temporal trajectory might. Attia notes this is exactly the structure of a machine-learning problem and would allow, for example, the 40-year-old who feels terrible but will not progress to be spared aggressive treatment, while the 60-year-old with diabetes who looks similar symptomatically is flagged for early antiviral loading.
I'd love to have that we don't have which are biomarkers for different stages of disease. You need as we've talked about — what you'd ideally like to do is take people, sample them every couple of days, see what their numbers look like, put them in this machine learning model that you're talking about.
Pandemic infrastructure preparedness: no-regret investments in PPE, testing capacity, and contact-tracing infrastructure
WhatMaintain national stockpiles of PPE, reagents for PCR and serologic testing, and electronic contact-tracing infrastructure in a state of readiness such that they can be deployed within days of a novel pathogen's sequence being published — not months.
WhenBetween pandemics — these are ongoing maintenance investments, not emergency spending.
For whomNational governments and health agencies. Attia argues this should be viewed as a national security infrastructure investment, analogous to strategic oil reserves, not as reactive emergency healthcare spending.
WhyThe virus genome for SARS-CoV-2 was sequenced by January 11–12, 2020. The delay in testing capacity was not scientific — designing PCR primers takes four days — it was scaling and logistics. A pre-invested infrastructure collapses that delay from months to days, enabling early detection and containment that could truncate the curve.
CaveatsAttia and Perlman acknowledge that NIH funding structures make it hard to fund research for diseases that 'don't exist anymore' (post-SARS-1). The argument is for infrastructure spending that is pathogen-agnostic, not disease-specific research.
Attia makes the analogy to oil reserves: the US keeps enough oil stockpile for 60–90 days of independence. A comparable investment in pandemic infrastructure — perhaps a few billion dollars — would have a far greater return on investment when measured against the economic cost of a full lockdown. He lists the specific components: PPE stockpiles, swab supplies, reagents, a pre-built serologic testing network, and a contact-tracing electronic infrastructure that can be activated without programming from scratch. None of these require knowing the identity of the pathogen in advance.
It's the nasal swabs, it's the reagents, it's all the stuff I just said and about 20 other things. I really hope that when this has said and done this doesn't get forgotten because it's not a staggering investment when you consider what we spend on health care and defense.
Isolation strategy calibration: confine only when symptomatic for SARS-1-type pathogens; assume pre-symptomatic spread for SARS-CoV-2-type pathogens
WhatFor any novel respiratory pathogen, the first surveillance question must be: does transmission occur before symptoms? For pathogens that spread only in the symptomatic phase (SARS-1 model), classic case-isolation-plus-contact-tracing is sufficient. For pathogens with pre-symptomatic spread (SARS-CoV-2 model), that strategy is mathematically inadequate and must be supplemented by population-level measures.
WhenWithin the first weeks of any novel outbreak — before community transmission is confirmed.
For whomEpidemiologists, public-health decision makers, and hospital infection-control teams responding to novel outbreaks.
WhySARS-1 was eliminated with 8,000 total cases using isolation alone. SARS-CoV-2 could not be contained by isolation because the infectious period precedes detectability by 2–3 days. The difference in optimal public-health strategy is determined by this single biological parameter.
Perlman explains that SARS-1's eradication was possible because of two features: (1) no animal reservoir to re-seed the human population; (2) you were not contagious until you were sick — so the sick-contact-isolate loop worked without residual leakage. SARS-CoV-2 breaks both conditions. Attia frames the lesson as a diagnostic protocol: any future pathogen surveillance protocol should estimate the pre-symptomatic infectious window as a first-order priority, not an afterthought.
Because you weren't contagious till you were sick it was easy to look at somebody — that person has SARS, we're gonna stick them in a room by himself, take care of him, and make sure he infects nobody else. Then you would stop the disease.
Also said
“The combination of there being no reservoir so it's not like camels which could continue to introduce into human populations and the fact that because you weren't contagious till you were sick — it was a classic kind of quarantine and identification approach.”— Identifies the two features that made SARS-1 eradicable, neither of which SARS-CoV-2 shares.
Post-COVID clinical monitoring: track lung function and neurological symptoms in recovered patients
WhatFor patients who recovered from significant COVID-19 illness (including those who were never hospitalized but reported weeks of severe symptoms), conduct follow-up pulmonary function testing (spirometry, DLCO) and cognitive/neurological assessment at 3 and 12 months post-recovery.
When3 months post-acute illness and 12 months post-acute illness.
For whomAny patient with more than one week of significant COVID-19 symptoms, regardless of hospitalization status. Higher priority for those with measurable acute hypoxia.
WhyPerlman notes that SARS-1 patients likely had lasting pulmonary fibrosis but this was not well-documented. MERS patients almost certainly had residual lung damage. SARS-CoV-2 is causing prolonged respiratory limitation in people who were never sick enough to be hospitalized — a pattern Attia illustrates with two personal examples of fit 40-somethings still unable to run a 9-minute mile three months post-infection.
CaveatsAt the time of recording, the mechanism was uncertain: Perlman notes fibrosis, prolonged hypoxia from ventilation, and immune-mediated neurological damage (analogous to Kawasaki disease) are all plausible contributors and hard to disentangle in retrospect.
Perlman draws on his SARS-1 experience to explain that neurological symptoms in SARS-1 survivors were attributed to prolonged ventilation, corticosteroid use, and potentially direct viral neuroinvasion — but the relative contributions were never sorted out. He predicts the same difficulty will arise with SARS-CoV-2 long-haulers. The MIS-C/Kawasaki-like syndrome in children is the clearest example of immune-mediated post-viral injury with cardiac consequences; in adults the analogous mechanism could produce subclinical myocarditis or neurological dysfunction.
I suspect that there was a better bit but you always have its in Arkansas as people always talk about neurological disease without actually ever finding the virus in the brain — it was attributed to being on ventilators and corticosteroids for long periods of time contributing to a cognitive dysfunction.
What's new
Personal practice updates, fresh positions, predictions
6 items
SARS-CoV-2 is a functional hybrid of a cold coronavirus and SARS-1
~slice-3
Perlman frames SARS-CoV-2 as a mixture of a common-cold coronavirus (explains upper-airway replication and pre-symptomatic transmissibility) and a SARS/MERS-type virus (explains the severe lower-airway pneumonia in a minority of patients). Neither precursor alone could have caused a pandemic.
Why this matters: Explains in one biological frame why SARS-CoV-2 achieved something SARS-1 and MERS never did — it can spread widely before anyone knows they are sick, and it retains the lung-damaging capacity of its more lethal cousins.
Background
SARS-1 was confined to the deep lungs, so transmission required aerosol-generating medical procedures; MERS had a very low human-to-human R0; common-cold coronaviruses spread freely but cause only nuisance illness. SARS-CoV-2 occupies both niches simultaneously.
Perlman notes that SARS-1 used the ACE2 receptor but only efficiently infected the lower airway, which is why it was essentially non-contagious until a patient required intubation or suctioning. SARS-CoV-2 uses the same ACE2 receptor but also efficiently infects the upper respiratory tract — the same anatomical site that allows common-cold coronaviruses (including NL63, which also uses ACE2) to spread so readily. The result is a virus that is transmissible during the incubation period and from people who will never develop pneumonia, while still causing severe lung disease in roughly 20% of infected individuals. Attia puts it succinctly: the overall IFR may be ~1–2%, but if only 20% of infected people get true lung disease, the mortality among that sub-group approaches the 10–25% seen in SARS-1 and MERS.
I think about it as SARS Coby to being a mixture of a common cold corona virus and then SARS or MERS coronavirus in the lungs so that's why you have the transmissibility and the severe disease because those both.
Pre-symptomatic transmission is the central feature that distinguishes SARS-CoV-2 from all prior dangerous coronaviruses
~slice-2 and slice-3
SARS-1 was containable because you could only transmit it when you were visibly, severely ill — the classic quarantine-and-isolate model worked perfectly. SARS-CoV-2 breaks that model entirely because shedding begins before symptoms appear.
Why this matters: The single-greatest structural difference in pandemic risk is not virulence but the pre-symptomatic window. Perlman makes clear this is why even ideal public-health execution could not stop SARS-CoV-2 the way it stopped SARS-1.
Background
SARS-1 R0 was 2–3 in-hospital but much lower in the community because community spread required visible severe pneumonia. The official R0 hid a bimodal distribution: near-zero in community settings, much higher in hospitals during aerosol-generating procedures.
Perlman explains that SARS-1 caused 8,000 total cases globally before being eliminated in July 2003. The mechanism of elimination was simple: once a person was sick enough to transmit, they were sick enough to be identified and isolated. SARS-CoV-2's pre-symptomatic infectious window makes that strategy mathematically impossible — by the time you find a case, the next generation of infections is already underway.
You don't really spread, you're very unlikely to spread it if you're not symptomatic. You don't have to shut the world down but not only that you get to isolate people when they're sick and treat them before they treat others.
Also said
“That's why SARS was not so contagious because it only stayed in the deep lungs until you went to the hospital and had procedures done.”— Mechanistic explanation of why SARS-1 was containable while SARS-CoV-2 is not.
Antibody immunity to coronaviruses wanes within 1–2 years and is weaker after mild infection
~slice-3
Human volunteer studies with common-cold coronaviruses showed measurable neutralizing antibody at one year, but immunity declined steadily. Mild infections produce briefer, weaker responses — mirroring exactly what was beginning to emerge for mild COVID-19 cases at the time of recording.
Why this matters: Directly informs whether natural infection or a single-dose vaccine can produce durable herd immunity. Perlman suggests the answer is probably no for mild infections — a critical planning assumption.
Background
Common-cold coronaviruses (229E, OC43, NL63, HKU1) circulate annually; people get reinfected roughly every 12–24 months. The durability data for SARS-CoV-2 was nascent at recording but Perlman predicted the same pattern based on common-cold coronavirus biology.
Perlman distinguishes IgG durability from IgA (mucosal) durability and notes both seem to wane. He adds that T-cell responses to common-cold coronaviruses were not well-studied before COVID-19 because common colds resolve so quickly — often before a full T-cell response even develops. Contrast with smallpox: people infected in 1918 still had measurable antibody in 1995. The difference likely reflects the magnitude of antigen exposure and the degree of systemic immune activation. A deep pneumonia that forces systemic inflammation generates lasting immune memory; a two-day cold does not.
We know that there's an antibody response to these viruses it seems to wane it goes away. The IgA response that seems to help we don't have that much information about IgA responses they seem to wane as well.
Also said
“Here we have these common cold corona viruses a year later they waned in a couple years later they're probably almost gone so we don't really understand that — that's really a key question.”— Quantifies the durability gap compared to highly immunogenic viruses like smallpox.
Cross-reactive T-cell responses to SARS-CoV-2 in never-infected people — intriguing but not yet proven protective
~slice-4
A 2020 Cell paper (Sette lab) found T-cell activation against SARS-CoV-2 peptides in ~50% of blood donors who had never been exposed to COVID-19, suggesting possible cross-reactivity from prior common-cold coronavirus infections. Perlman applies significant caveats: the assay measured activation markers, not cytokine production or killing function, and the target peptides showed little sequence homology with known common-cold coronavirus epitopes.
Why this matters: If real and functional, cross-reactive T-cell immunity could partially explain why some never-infected individuals have very mild disease. But Perlman warns against over-interpreting activation-only readouts.
Background
The finding was simultaneously reported by several independent groups, lending some credibility. At the time it generated widespread media coverage about pre-existing immunity.
Perlman lays out a three-part critique of the cross-reactive T-cell evidence: (1) the assay measured surface activation markers, not cytokine secretion or cytolytic killing — standard flu-vaccine T-cell studies use cytokine response as their read-out, which is more functionally meaningful; (2) when researchers looked for homology between the SARS-CoV-2 peptides that activated these T-cells and known common-cold coronavirus sequences, they found surprisingly little overlap, raising the question of where the response actually came from; (3) cytokine levels in the naive people were extremely low, approaching background noise. He concludes: The jury is just out on how important it is and what it means.
In most of these papers the T-cell responses are measured not by functionality but rather by being activated in a certain way and these activations to my mind are a surrogate for actual functionality and the functionality was not well demonstrated in any of these studies.
Also said
“The targets for these viruses for the T-cell response is not the usual response that you get in terms of targets and you see after the wild-type SARS-CoV-2 infection and because of this difference the mix is also a little unclear to me where this response is coming from.”— Explains why the sequence-homology argument for common-cold cross-reactivity is weaker than it sounds.
MERS mortality of ~35% is a denominator artifact, not a true biological superiority in virulence
~slice-3
Perlman argues that the apparent 10x gap between MERS (35%) and SARS-CoV-2 (1–2%) IFR collapses when you apply a consistent denominator: because SARS-CoV-2 infects 5x more people subclinically (upper-airway-only or asymptomatic), the lung-disease mortality among those who actually develop pneumonia is roughly equivalent across all three dangerous coronaviruses.
Why this matters: Reframes the apparent safety gap between SARS-CoV-2 and its predecessors — the denominator, not biology, is doing most of the work. If SARS-CoV-2 had MERS-level transmissibility, it would also have appeared to have 35% mortality.
Perlman's back-of-envelope: if 1,000 people get SARS-CoV-2, roughly 200 develop lung disease, and of those 200 maybe 10–15 die (5–7.5% of the sick subset). For SARS-1 or MERS, 100% of identified cases had lung disease, so the CFR and IFR were nearly identical — no iceberg of mild cases to hide the deaths. The implication is that the lung is equally vulnerable in all three, and the outcome for a person who develops SARS-CoV-2 pneumonia is statistically similar to SARS-1 or MERS pneumonia.
If you have with SARS and MERS there were a hundred people infected they all get some variability some variation in pneumonia and you have a certain mortality rate ranging from 10% to 30%. SARS-CoV-2: 2,000 people, maybe 20 of them are going to get the lung disease, the other 1,800 are going to be asymptomatic, subclinical, have a cold.
MERS geographic confinement to the Arabian Peninsula remains unexplained
~slice-2
MERS coronavirus has caused ongoing camel-to-human and limited human-to-human transmission in Saudi Arabia since at least 2012 — with cases still occurring once or twice a week at the time of recording — but virtually no cases in Africa despite camels carrying the identical or very similar virus there. The reason is genuinely unknown.
Why this matters: Highlights that even after a decade of study, the molecular and epidemiological determinants of zoonotic spillover remain poorly understood — a cautionary note for pandemic preparedness models.
Perlman notes that MERS in camels causes a common cold; in humans it causes severe pneumonia with a ~35% CFR. The same virus in African camels does not spill over to humans in Africa, North Africa, or the Gulf countries of Africa despite close human-camel contact in those regions. The viral sequence differences between Arabian-Peninsula camels and African camels are subtle and hard to map to the spillover phenotype. This means the spillover determinants could be host-genetic, environmental, or just a matter of specific human behaviors in the Arabian Peninsula that are not yet characterized.
We don't know the buyers have been in camels since probably earlier. We know it doesn't jump in Africa or other parts of Asia so this is a real mystery in this virus — why is it only in the Arabian Peninsula when you look at the MERS virus in the camels.
Recommendations
Products, supplements, and tools mentioned in the episode
2 items
Listen to the Peter Attia / David Watkins immunology episode before this one
Practice
Attia and Perlman repeatedly reference the immunology foundations covered in the Watkins episode (innate vs. adaptive immunity, B-cell / T-cell distinction, IgM vs. IgG vs. IgA) as prerequisites for fully understanding the COVID-19 immune-response discussion in this episode.
Perlman specifically mentions that B-cell / T-cell distinctions, neutralizing vs. binding antibodies, and the IgA mucosal response are concepts the listener needs before the cross-reactive immunity and antibody durability sections make sense. Attia pauses the conversation twice to recommend the Watkins episode for listeners who feel lost.
Actually if you're listening to this now and you have not listened to the interview with David Watkins this would be a great time to hit pause, go back and do that because in that discussion we really do the immunology tour de force.
Bill Gates 2015 TED Talk: 'The Next Outbreak? We're Not Ready'
Book
Attia and Perlman cite the Gates TED Talk as a prescient warning — painful to watch in retrospect — that coronaviruses and emerging pathogens were on the threat list and that pandemic infrastructure was dangerously under-invested.
The talk was cited in a broader discussion about whether scientists and policy-makers should have predicted COVID-19. Perlman notes the US Department of Defense had also published a 2010–2011 report listing coronaviruses as a major emerging-virus threat. Neither the DoD report nor the Gates talk translated into the infrastructure investments Attia argues were no-regret moves — a failure both speakers frame as a solvable policy problem rather than an intelligence failure.
Bill Gates talked about this — it's become a very well-known TED talk that is painful to watch now because of how accurate it was.
Stanley Perlman Lab / University of Iowa — COVID-19 immune-response durability research
Service Sponsored · disclosed
Attia mentions the collaboration as the reason he was introduced to Perlman and why they had regular contact in the months before this recording. The study was examining the durability and nature of immune responses in COVID-19-recovered individuals.
DisclosureAttia and Perlman were actively co-authoring a study on immune-response durability at the time of recording — disclosed multiple times.
The collaboration is mentioned briefly at the start and end of the episode, positioned as the 'other project' that gave Attia regular access to Perlman during the peak of the pandemic. The research design — serial sampling, measuring both antibody titers and T-cell responses over time — is precisely the type of longitudinal cohort study Perlman later describes as the ideal approach to answering the immune-durability question.
We've been part of a collaboration that is working on a longer study trying to understand the durability of immune response and the impacts of that, which we discussed very briefly at the end of this podcast.
Lines worth pulling out — contrarian, specific, or perfectly phrased
6 items
I think about it as SARS-CoV-2 being a mixture of a common cold coronavirus and then SARS or MERS coronavirus in the lungs so that's why you have the transmissibility and the severe disease because those both.
The single most concise biological explanation of why COVID-19 became a pandemic while SARS-1 and MERS did not — in one sentence from the world's foremost coronavirus expert.
Because you weren't contagious till you were sick it was easy to look at somebody — that person has SARS, we're gonna stick them in a room by himself, take care of him, and make sure he infects nobody else.
The SARS-1 containment lesson in its purest form — and the implicit explanation of why it could never work for SARS-CoV-2.
The key question: if you got to a point where there's enough herd immunity or that hundred million turns into six billion people have seen the infection so nobody gets pneumonia anymore — it turns into a common cold.
Perlman's endgame scenario for SARS-CoV-2: not eradication, but endemic coexistence at common-cold severity once population immunity is sufficiently broad.
I spoke to my friends in China and I got off the phone and told my wife this is a big deal — so I'm not sure what I was facing that on because it wasn't so much evidence of human-to-human transmission but it was pretty clear to me then that this was going to be a major problem.
Perlman's early-December 2019 gut call — weeks before official pandemic warnings — based on 40 years of coronavirus research pattern recognition.
We have these common cold corona viruses a year later they waned, a couple years later they're probably almost gone. We don't really understand that. That's really a key question and it impacts the ability of people to be reinfected, impacts the vaccine responses, impacts general herd immunity as we try to get rid of this virus.
Perlman's summary of the central durability uncertainty — stated months before the term 'waning immunity' became a household phrase.
If you have with SARS and MERS there were a hundred people infected they all get some variability in pneumonia and you have a certain mortality rate ranging from 10% to 30%. SARS-CoV-2: 2,000 people, maybe 20 of them are going to get the lung disease, the other 1,800 are going to be asymptomatic, subclinical, have a cold.
The denominator argument that explains why SARS-CoV-2's IFR is lower than MERS while its lung pathology is equally severe — the key to understanding all three virus mortality comparisons.
Sign in to share feedback
Tell us if this brief hit the mark or missed it — feedback feeds back into the next iteration of the prompt.
Reading is free for everyone. A free account adds the personal layer: save protocols, follow experts, and see how the other experts weigh in on this same topic.
Educational summary of the cited expert source — not medical advice. Open the source recording linked above and consult a qualified physician before acting on any protocol.
