{"id":16150,"date":"2026-10-25T06:00:00","date_gmt":"2026-10-25T10:00:00","guid":{"rendered":"https:\/\/cov19longhaulfoundation.org\/?p=16150"},"modified":"2026-09-10T13:29:51","modified_gmt":"2026-09-10T17:29:51","slug":"the-biological-signature-of-long-covid","status":"publish","type":"post","link":"https:\/\/cov19longhaulfoundation.org\/?p=16150","title":{"rendered":"The Biological Signature of Long COVID"},"content":{"rendered":"\n<h5 class=\"wp-block-heading\"><em>What the World&#8217;s Largest Clinical and Molecular Datasets Are Beginning to Reveal<\/em><\/h5>\n\n\n\n<p class=\"has-small-font-size wp-block-paragraph\"><strong>John Murphy, CEO<\/strong>, <strong>The COVID-19 Long-haul Foundation<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Abstract<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Long COVID presents one of the most consequential unresolved problems in modern medicine: millions of people develop persistent, relapsing or progressive symptoms after SARS-CoV-2 infection, yet no single laboratory test can reliably identify the condition, define its biological subtypes or predict its trajectory. The question is no longer whether Long COVID exists. The central scientific problem is whether the enormous clinical and molecular datasets accumulated since 2020 can identify reproducible biological signatures that explain why some patients recover while others develop persistent or progressive disease.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This article conducts a cross-dataset synthesis and quantitative analysis of the principal public and published data resources currently capable of addressing this question. These include the NIH RECOVER observational cohorts and BioData Catalyst, the National COVID Cohort Collaborative (N3C), PCORnet, UK Biobank, the U.S. All of Us Research Program, the UK Office for National Statistics Coronavirus Infection Survey, and major published prospective biomarker, proteomic, immunologic and metabolomic cohorts.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Several conclusions emerge.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">First, the evidence does not support a single Long COVID biomarker. A systematic review encompassing 46 studies and 4,415 participants identified <strong>525 candidate blood biomarkers<\/strong>, grouped principally into immune\/inflammatory dysfunction, endothelial\/vascular dysfunction, and metabolic\/coagulation abnormalities. No single marker demonstrated sufficient reproducibility to serve as a diagnostic test.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Second, multiple independent biological studies nevertheless converge on several pathways: persistent immune activation, altered lymphoid and myeloid populations, endothelial and coagulation abnormalities, altered cortisol and serotonin biology, metabolic disturbance, possible viral persistence, and tissue-specific inflammatory signatures.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Third, the strongest prospective evidence that biological measurements can predict later Long COVID comes from a 136-person early-infection cohort. Among 104 participants with post-acute follow-up, 32 developed PASC. Those who developed PASC had greater early viral RNA burden, greater infectious viral shedding, higher plasma nucleocapsid antigen and lower early spike-specific IgG responses. Importantly, conventional inflammatory markers did not distinguish the groups.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Fourth, large population databases demonstrate that biological risk is multidimensional. In an 8,668-person UK Biobank Long COVID cohort, obesity, deprivation, elevated C-reactive protein, triglycerides, vitamin D, HbA1c, cystatin C, urate and alanine aminotransferase, together with lower HDL cholesterol and IGF-1, were associated with Long COVID; chronic kidney disease and COPD were also associated. These associations suggest metabolic, inflammatory and organ-specific susceptibility rather than a single disease pathway.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Fifth, the convergence of these datasets suggests that Long COVID may be best understood as a <strong>network disease<\/strong>: a set of partially overlapping biological states initiated by SARS-CoV-2 infection and sustained by different combinations of viral persistence, immune dysregulation, vascular dysfunction, autonomic disturbance, metabolic disruption and tissue injury.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The next scientific advance should therefore be the construction and prospective validation of <strong>multimodal biological signatures<\/strong>, rather than the search for a solitary biomarker.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Introduction: The Diagnostic Paradox<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Modern medicine is accustomed to diseases that eventually acquire an objective signature.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Diabetes has glucose and HbA1c.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Myocardial infarction has troponin.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Hypothyroidism has thyroid-stimulating hormone and thyroid hormone concentrations.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Chronic kidney disease has creatinine, eGFR and albuminuria.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Many cancers have histopathology, molecular signatures or imaging findings.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Long COVID remains different.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A patient may experience profound fatigue, post-exertional malaise, cognitive dysfunction, dysautonomia, sleep disruption, neuropathic symptoms, cardiopulmonary limitation or multisystem disease, yet routine laboratory testing may be unrevealing.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">An NIH-supported RECOVER analysis of more than 10,000 adults concluded that routine laboratory tests were not sufficiently reliable to diagnose Long COVID. No validated clinical biomarker currently exists.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That conclusion should not be interpreted as evidence that Long COVID lacks biological abnormalities.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It means something more subtle\u2014and scientifically more interesting.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>The biology may be distributed across multiple systems rather than concentrated in one routinely measured laboratory variable.<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">The Data Revolution<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The scale of available Long COVID data is unprecedented.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">RECOVER has made millions of data points available to researchers and reports public access to grouped anonymized data through BioData Catalyst\/PIC-SURE, while participant-level datasets are available through controlled access. The most recent observational snapshots were released in May 2026 and include adult, pediatric and tissue-pathology datasets.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The NIH previously reported that secure data from more than 14,000 adults in RECOVER were available through BioData Catalyst.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">RECOVER also has access through PCORnet and N3C to electronic health records representing <strong>more than 60 million people<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">N3C alone has supported analysis of more than <strong>600 million clinical-condition diagnoses from approximately 14 million patients<\/strong>, producing hundreds of detailed longitudinal clinical phenotypes.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This creates an unusual scientific opportunity.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For the first time, investigators can potentially connect:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>infection \u2192 acute biology \u2192 early immune response \u2192 persistent symptoms \u2192 organ dysfunction \u2192 treatment \u2192 recovery or progression.<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Figure 1. The Emerging Long COVID Data Architecture<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The enormous difference in scale is scientifically important.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The small molecular cohorts provide biological resolution.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The enormous EHR cohorts provide statistical power.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The central challenge is joining the two.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">I. The First Major Finding: There Is No Single Biomarker<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">A systematic review of blood biomarkers searched PubMed, Embase, Cochrane and Web of Science through January 2024.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It identified:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>45 observational studies<\/strong><\/li>\n\n\n\n<li><strong>1 interventional study<\/strong><\/li>\n\n\n\n<li><strong>4,415 participants<\/strong><\/li>\n\n\n\n<li><strong>525 candidate blood biomarkers<\/strong><\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The biomarkers clustered into three broad physiological domains:<\/p>\n\n\n\n<ol start=\"1\" class=\"wp-block-list\">\n<li>immune\/inflammatory dysfunction;<\/li>\n\n\n\n<li>endothelial\/vascular dysfunction;<\/li>\n\n\n\n<li>metabolic and coagulation abnormalities.<\/li>\n<\/ol>\n\n\n\n<p class=\"wp-block-paragraph\">Most importantly, the review concluded that <strong>no single biomarker was sufficiently associated with Long COVID prevalence to function as a reliable diagnostic test<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This finding is not a failure of biomarker science.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It may be the clue.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">If 525 candidate markers appear across heterogeneous studies but no solitary marker consistently dominates, the disease may be biologically heterogeneous.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Figure 2. The Biomarker Problem<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The chart above intentionally does <strong>not<\/strong> assign numerical biomarker counts to the three domains because the systematic review did not report them as mutually exclusive numerical categories. The important quantitative observation is the <strong>525-marker universe<\/strong>, not a fabricated subdivision.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This distinction matters in scientific reporting.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">II. The Immune System Leaves a Distinctive Fingerprint<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">One of the most influential biological studies examined 273 participants using detailed immune profiling.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Investigators found differences in both myeloid and lymphoid compartments.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Compared with matched controls, participants with Long COVID showed:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>higher non-conventional monocytes;<\/li>\n\n\n\n<li>higher activated B-cell populations;<\/li>\n\n\n\n<li>higher double-negative B cells;<\/li>\n\n\n\n<li>higher exhausted CD4 T cells;<\/li>\n\n\n\n<li>lower cDC1 populations;<\/li>\n\n\n\n<li>lower central-memory CD4 T cells;<\/li>\n\n\n\n<li>stronger antibody responses against SARS-CoV-2;<\/li>\n\n\n\n<li>higher antibody responses against Epstein-Barr virus and varicella-zoster virus;<\/li>\n\n\n\n<li>lower morning cortisol.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The study&#8217;s machine-learning analysis identified cortisol as the strongest individual predictor in its model, with cortisol alone producing an AUC of approximately <strong>0.96<\/strong> in that specific study population.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That number is impressive.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It is also dangerous if misinterpreted.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">An AUC of 0.96 in one carefully characterized research cohort does <strong>not<\/strong> mean cortisol is a validated clinical diagnostic test.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Replication in independent populations is essential.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Figure 3. The Biological Domains That Recur Across Studies<\/h5>\n\n\n\n<pre class=\"wp-block-preformatted\"> <code>                   LONG COVID\n                        \u2502\n        \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n        \u2502               \u2502                \u2502\n        \u25bc               \u25bc                \u25bc\n     IMMUNE          VASCULAR         METABOLIC\n   DYSREGULATION     DYSFUNCTION      DYSFUNCTION\n        \u2502               \u2502                \u2502\n   \u250c\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2510      \u250c\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2510       \u250c\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2510\n   \u25bc    \u25bc    \u25bc      \u25bc   \u25bc    \u25bc       \u25bc   \u25bc    \u25bc\n T\/B  Myeloid  EBV  Endo- Clot  Micro- Glucose Lipids\ncells cells   reactivation thelial ting vascular     |\n                              \u2502             \u2502\n                              \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u252c\u2500\u2500\u2500\u2500\u2500\u2500\u2518\n                                     \u25bc\n                              ORGAN-SPECIFIC\n                                  INJURY\n                                     \u2502\n                    \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n                    \u25bc                \u25bc               \u25bc\n                  Brain            Heart            Lung\n                    \u2502                \u2502               \u2502\n                    \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\n                                     \u25bc\n                                CLINICAL\n                                SYMPTOMS<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">The diagram represents a mechanistic hypothesis synthesized from multiple studies\u2014not a proven single causal pathway.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">III. Viral Persistence: The Most Important Prospective Signal<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Perhaps the most informative evidence comes not from a study of patients already ill for years, but from people studied <strong>during the first days of infection<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Lu and colleagues enrolled <strong>136 participants within five days of their first positive SARS-CoV-2 PCR test<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">They collected:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>up to 21 nasal samples during the first 28 days;<\/li>\n\n\n\n<li>blood at enrollment;<\/li>\n\n\n\n<li>blood at days 9, 14, 21 and 28;<\/li>\n\n\n\n<li>follow-up at four and eight months.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Among 104 participants completing post-acute evaluation:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>32 (31%) developed PASC.<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The comparison is revealing.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Patients who later developed PASC demonstrated:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>higher maximum SARS-CoV-2 RNA levels;<\/li>\n\n\n\n<li>higher plasma nucleocapsid antigen;<\/li>\n\n\n\n<li>greater cumulative RNA shedding;<\/li>\n\n\n\n<li>greater infectious viral shedding;<\/li>\n\n\n\n<li>lower early spike-specific IgG.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The differences were especially apparent during the first nine days.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Figure 4. Early Viral Burden as a Potential Predictor<\/h5>\n\n\n\n<pre class=\"wp-block-preformatted\"> <code>                   ACUTE INFECTION\n                          \u2502\n                          \u25bc\n              Higher early viral burden\n                          \u2502\n             \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2534\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n             \u25bc                         \u25bc\n       Greater RNA shedding      Higher N-antigen\n             \u2502                         \u2502\n             \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u252c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\n                          \u25bc\n                  Slower effective\n                   viral control\n                          \u2502\n                          \u25bc\n                 Increased probability\n                     of PASC<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">This does not prove that viral persistence causes Long COVID.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It does, however, provide an important temporal clue:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>biology measurable during acute infection may contain information about future disease.<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That is fundamentally different from merely finding a biomarker after Long COVID has already developed.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Quantitative Analysis of the Early-Infection Study<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The 104 participants with post-acute follow-up divided into:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>32 PASC;<\/li>\n\n\n\n<li>72 without PASC.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Thus:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">[<br>P(PASC)=\\frac{32}{104}=30.8%<br>]<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The approximate standard error for this proportion is:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">[<br>SE=\\sqrt{\\frac{p(1-p)}{n}}<br>]<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">giving approximately:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">[<br>SE \\approx 0.045<br>]<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">and an approximate 95% interval of:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">[<br>30.8% \\pm 8.8%<br>]<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">or approximately:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>22.0%\u201339.6%.<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The estimate is therefore substantial but imprecise because the cohort is small.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The strength of the study is not statistical size.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Its strength is <strong>temporal resolution<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Biological specimens were collected before investigators knew which patients would later develop persistent symptoms.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That greatly reduces one important source of reverse causation.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">IV. The Unexpected Negative Finding<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The same study produced an equally important negative result.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Seven inflammatory markers were examined:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>IL-6;<\/li>\n\n\n\n<li>IL-10;<\/li>\n\n\n\n<li>TNF-\u03b1;<\/li>\n\n\n\n<li>MCP;<\/li>\n\n\n\n<li>IP-10;<\/li>\n\n\n\n<li>IFN-\u03b1;<\/li>\n\n\n\n<li>IFN-\u03b3.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The investigators found <strong>no statistically significant differences<\/strong> in their acute trajectories between participants who later developed PASC and those who did not.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This matters.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">If Long COVID were simply the consequence of an excessively high generalized inflammatory response during acute COVID, one might expect these markers to provide a strong early signal.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">They did not.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The result points toward a more complicated model:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>viral burden + immune control + tissue response + later immune persistence<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">rather than simply:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>more inflammation = Long COVID.<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">V. Proteomics Reveals Subtypes Rather Than One Disease<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Another major investigation examined <strong>657 participants<\/strong> at least three months after hospitalization and measured <strong>368 plasma proteins<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Among them:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>426 had at least one Long COVID symptom;<\/li>\n\n\n\n<li>233 had fully recovered.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Elevated markers of myeloid inflammation and complement activation were associated with Long COVID.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">But the remarkable observation was the appearance of different protein signatures associated with different symptom domains.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Examples included:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>IL-1R2, MATN2 and COLEC12 associated with cardiorespiratory symptoms, fatigue and anxiety\/depression;<\/li>\n\n\n\n<li>MATN2, CSF3 and C1QA associated with gastrointestinal symptoms;<\/li>\n\n\n\n<li>C1QA associated with cognitive impairment;<\/li>\n\n\n\n<li>SPON1 and NFASC associated with cognitive impairment and nerve-tissue repair;<\/li>\n\n\n\n<li>SCG3 associated with gastrointestinal symptoms and possible brain-gut disturbance.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">This is one of the strongest arguments against treating Long COVID as biologically homogeneous.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Figure 5. From One Disease to Multiple Biological Phenotypes<\/h5>\n\n\n\n<pre class=\"wp-block-preformatted\"> <code>                        LONG COVID\n                             \u2502\n          \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n          \u2502                  \u2502                 \u2502\n          \u25bc                  \u25bc                 \u25bc\n    CARDIORESPIRATORY    COGNITIVE          GI\n       PHENOTYPE         PHENOTYPE       PHENOTYPE\n          \u2502                  \u2502                 \u2502\n     IL-1R2\/MATN2        C1QA\/SPON1        MATN2\/CSF3\n     COLEC12             NFASC             SCG3\n          \u2502                  \u2502                 \u2502\n          \u25bc                  \u25bc                 \u25bc\n      Pulmonary          Neuroimmune       Brain-gut\n      \/vascular          dysfunction       disturbance\n       biology<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">The investigators explicitly cautioned that the study was designed to understand inflammatory processes rather than establish definitive diagnostic biomarkers.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That distinction should be retained.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">VI. The UK Biobank Provides Population-Level Biological Confirmation<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The UK Biobank provides a fundamentally different kind of evidence.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A 2025 <em>Nature Communications<\/em> study analyzed:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>8,668 participants in a Long COVID cohort<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">and separately:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>108,407 participants in a post-acute complications cohort.<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Approximately 32% of the Long COVID cohort were classified as cases, compared with approximately 2% of the PACS cohort.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The Long COVID phenotype was associated with:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>obesity;<\/li>\n\n\n\n<li>socioeconomic deprivation;<\/li>\n\n\n\n<li>elevated C-reactive protein;<\/li>\n\n\n\n<li>elevated triglycerides;<\/li>\n\n\n\n<li>vitamin D;<\/li>\n\n\n\n<li>HbA1c;<\/li>\n\n\n\n<li>cystatin C;<\/li>\n\n\n\n<li>urate;<\/li>\n\n\n\n<li>alanine aminotransferase;<\/li>\n\n\n\n<li>reduced HDL cholesterol;<\/li>\n\n\n\n<li>reduced IGF-1;<\/li>\n\n\n\n<li>chronic kidney disease;<\/li>\n\n\n\n<li>chronic obstructive pulmonary disease.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Most of these associations were also observed in PACS, while alanine aminotransferase and vitamin D showed more phenotype-specific behavior.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Figure 6. Biological Systems Implicated by UK Biobank Findings<\/h5>\n\n\n\n<pre class=\"wp-block-preformatted\"> <code>             UK BIOBANK ASSOCIATIONS\n                       \u2502\n       \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n       \u25bc               \u25bc                 \u25bc\n   METABOLIC       INFLAMMATORY       ORGAN \/\n   DISTURBANCE       SIGNALING       COMORBIDITY\n       \u2502               \u2502                 \u2502\n   HbA1c             CRP              CKD\n   TG                \u2502                COPD\n   HDL               \u2502                obesity\n   IGF-1             \u2502\n   urate             \u2502\n   vitamin D         \u2502\n                     \u2502\n                     \u25bc\n               SYSTEMIC HOST\n                VULNERABILITY<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">The UK Biobank findings are particularly important because they shift the discussion from isolated molecules toward <strong>host susceptibility<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Long COVID may partly depend upon the biological state of the host before and during infection.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">VII. All of Us Adds Another Dimension: Baseline Health<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The U.S. All of Us Research Program provides an important complementary perspective.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A recent analysis included <strong>65,464 adults<\/strong> with evidence of COVID-19:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>40,655 had at least one post-infection Long COVID symptom;<\/li>\n\n\n\n<li>24,809 did not.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The investigators found that greater pre-infection symptom burden was associated with greater odds of subsequent Long COVID symptoms.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Compared with the median pre-infection symptom burden, participants reporting zero symptoms had an adjusted odds ratio of approximately:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>0.15 (99% CI 0.04\u20130.61).<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Lower pre-infection self-rated mental health was also associated with greater odds of subsequent Long COVID.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The important implication is not that pre-existing symptoms \u201ccause\u201d Long COVID.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It is that the biological phenotype of the patient <strong>before infection<\/strong> may matter.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">VIII. N3C Shows That the Phenotype Is Multisystem<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The N3C\/RECOVER EHR analysis is perhaps the most powerful demonstration of scale.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Investigators analyzed more than:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>600 million clinical-condition diagnoses<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">from approximately:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>14 million patients.<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The analysis generated hundreds of detailed clinical phenotypes and showed that disease patterns after COVID differed according to:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>age;<\/li>\n\n\n\n<li>sex;<\/li>\n\n\n\n<li>infection wave;<\/li>\n\n\n\n<li>Long COVID status.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Earlier PCORnet work independently identified increased post-COVID diagnoses involving:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>dementia;<\/li>\n\n\n\n<li>hair loss;<\/li>\n\n\n\n<li>pulmonary fibrosis;<\/li>\n\n\n\n<li>dyspnea;<\/li>\n\n\n\n<li>pulmonary embolism;<\/li>\n\n\n\n<li>chest pain;<\/li>\n\n\n\n<li>abnormal heartbeat;<\/li>\n\n\n\n<li>malaise;<\/li>\n\n\n\n<li>fatigue.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Several findings were replicated across both the INSIGHT and OneFlorida+ networks.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This is important because a biological explanation must ultimately account for the clinical phenotype.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A mechanism that explains fatigue but not pulmonary disease, autonomic dysfunction, gastrointestinal symptoms or cognitive impairment is unlikely to explain the entire syndrome.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">IX. What the Biomarker Literature Actually Says<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The evidence can now be divided into five levels.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Level 1 \u2014 Epidemiological association<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Large databases establish that Long COVID is associated with identifiable demographic, clinical and metabolic risk factors.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Strong evidence.<\/strong><\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Level 2 \u2014 Biological association<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Multiple studies identify differences in immune, metabolic, vascular and hormonal measurements.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Strong and increasingly replicated evidence.<\/strong><\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Level 3 \u2014 Predictive association<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Some early biological variables appear capable of predicting subsequent PASC.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Promising but incompletely validated evidence.<\/strong><\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Level 4 \u2014 Mechanistic causation<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Whether those biological abnormalities actually cause persistent disease remains incompletely established.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Insufficient evidence for most pathways.<\/strong><\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Level 5 \u2014 Therapeutic reversibility<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The strongest proof of causality would be:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">biomarker abnormality \u2192 targeted treatment \u2192 biomarker normalization \u2192 clinical recovery.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>This remains the major unmet requirement.<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Figure 7. The Evidence Pyramid<\/h5>\n\n\n\n<pre class=\"wp-block-preformatted\"> <code>                        \u25b2\n                         \u2502\n                  THERAPEUTIC PROOF\n                         \u2502\n               Biomarker changes AND\n                 clinical recovery\n                         \u2502\n                    CAUSALITY\n                         \u2502\n               Mechanism demonstrated\n                         \u2502\n                    PREDICTION\n                         \u2502\n             Early biomarker predicts\n                 later phenotype\n                         \u2502\n                    ASSOCIATION\n                         \u2502\n              Biomarker differs between\n                    groups\n                         \u2502\n                   EPIDEMIOLOGY\n                         \u2502\n               Risk factors correlate\n                     with disease\n                         \u25bc<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">Much of Long COVID research currently resides in the middle three levels.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The field&#8217;s greatest challenge is moving upward.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">X. A Convergence of Biological Mechanisms<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Despite substantial disagreement among individual studies, several biological domains recur with surprising consistency.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">1. Viral persistence or delayed viral clearance<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Evidence includes:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>viral RNA burden;<\/li>\n\n\n\n<li>prolonged shedding;<\/li>\n\n\n\n<li>circulating viral antigen;<\/li>\n\n\n\n<li>tissue persistence observations;<\/li>\n\n\n\n<li>early viral burden predicting later PASC.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The 136-person prospective study provides especially important temporal evidence.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">2. Immune dysregulation<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Evidence includes:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>altered monocyte populations;<\/li>\n\n\n\n<li>altered B-cell populations;<\/li>\n\n\n\n<li>altered T-cell phenotypes;<\/li>\n\n\n\n<li>complement abnormalities;<\/li>\n\n\n\n<li>persistent inflammatory protein signatures;<\/li>\n\n\n\n<li>altered antiviral antibody profiles.<\/li>\n<\/ul>\n\n\n\n<h5 class=\"wp-block-heading\">3. Endothelial and coagulation abnormalities<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Multiple studies and systematic reviews identify vascular and coagulation abnormalities among the recurring biological domains.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This is compatible with the multisystem vascular manifestations observed clinically.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">4. Neuroendocrine dysfunction<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Reduced cortisol has emerged repeatedly enough to merit serious investigation, although it is not yet a validated diagnostic marker.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">5. Metabolic dysfunction<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">UK Biobank findings implicate glucose metabolism, lipids, urate, vitamin D and other metabolic variables.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">6. Serotonin biology<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Experimental and human studies have implicated reduced serotonin availability in PASC, providing a possible connection among inflammation, gastrointestinal physiology, platelets, the vagus nerve and cognitive function.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XI. Why the Studies Sometimes Contradict One Another<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The apparent contradictions in Long COVID biomarker research are often scientifically informative.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">One study identifies abnormal inflammatory markers.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Another does not.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">One reports autoantibodies.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Another finds little difference.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">One identifies viral persistence.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Another does not find virus in the compartment it examined.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">These findings need not be mutually exclusive.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Long COVID may be <strong>heterogeneous<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A patient whose principal pathology involves persistent viral antigen may not have the same biomarker profile as a patient whose dominant pathology is autonomic, vascular, neurological or metabolic.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The appropriate question therefore becomes:<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>Which biological signature corresponds to which clinical phenotype?<\/strong><\/p>\n<\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\">That question is substantially more sophisticated than asking:<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\">\u201cWhat is the Long COVID biomarker?\u201d<\/p>\n<\/blockquote>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XII. The Multi-Omic Model<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The most promising future approach is therefore not one test but a multimodal signature.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A hypothetical diagnostic vector might be:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">[<br>X =<br>{V,I,E,M,N,H,G,F}<br>]<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">where:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>(V) = viral persistence;<\/li>\n\n\n\n<li>(I) = immune state;<\/li>\n\n\n\n<li>(E) = endothelial\/coagulation state;<\/li>\n\n\n\n<li>(M) = metabolic state;<\/li>\n\n\n\n<li>(N) = neurological\/autonomic state;<\/li>\n\n\n\n<li>(H) = hormonal state;<\/li>\n\n\n\n<li>(G) = gastrointestinal\/microbiome state;<\/li>\n\n\n\n<li>(F) = functional phenotype.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The objective would be:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">[<br>P(\\text{Long COVID phenotype}\\mid X)<br>]<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">rather than relying upon:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">[<br>P(\\text{Long COVID}\\mid \\text{single biomarker})<br>]<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This is a fundamentally different diagnostic philosophy.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Figure 8. The Proposed Long COVID Biological Signature<\/h5>\n\n\n\n<pre class=\"wp-block-preformatted\"> <code>                   PATIENT\n                       \u2502\n          \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n          \u25bc            \u25bc            \u25bc\n       VIRUS         IMMUNE       HOST\n          \u2502            \u2502            \u2502\n          \u25bc            \u25bc            \u25bc\n      Antigen       T\/B cells     Metabolism\n      RNA           Cytokines     Hormones\n      Shedding      Complement    Genetics\n          \u2502            \u2502            \u2502\n          \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\n                       \u25bc\n                  MULTI-OMIC\n                    PROFILE\n                       \u2502\n        \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n        \u25bc              \u25bc               \u25bc\n     Neural         Vascular        Metabolic\n     phenotype      phenotype       phenotype\n        \u2502              \u2502               \u2502\n        \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\n                       \u25bc\n                 LONG COVID\n                   SUBTYPE\n                       \u2502\n                       \u25bc\n                PREDICTED COURSE\n                       \u2502\n            \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n            \u25bc          \u25bc          \u25bc\n         Recovery   Persistent  Progressive<\/code><\/pre>\n\n\n\n<p class=\"wp-block-paragraph\">This is a proposed research architecture, not a validated clinical algorithm.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XIII. The Most Important New Research Question<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The scientific community has spent considerable effort asking:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>What biological abnormalities occur in Long COVID?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The next question should be:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Which biological abnormalities occur before Long COVID develops, which occur after it develops, and which persist when the disease becomes chronic?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This requires repeated sampling.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A single blood draw is inadequate.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The ideal study would obtain:<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Acute phase<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Day 0\u20137<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Early recovery<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Day 14\u201328<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Early post-acute phase<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Month 2\u20133<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Established Long COVID<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Month 6<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Chronic disease<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Month 12<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Long-term disease<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Years 2\u20135<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The objective would be to determine which abnormalities:<\/p>\n\n\n\n<ol start=\"1\" class=\"wp-block-list\">\n<li>precede symptoms;<\/li>\n\n\n\n<li>accompany symptoms;<\/li>\n\n\n\n<li>disappear when symptoms resolve;<\/li>\n\n\n\n<li>persist after clinical recovery;<\/li>\n\n\n\n<li>worsen before clinical deterioration.<\/li>\n<\/ol>\n\n\n\n<p class=\"wp-block-paragraph\">The fourth and fifth categories would be particularly informative.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XIV. The Biomarker That Changes Should Matter More Than the Biomarker That Merely Correlates<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Consider two hypothetical biomarkers.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Biomarker A is elevated in 60% of Long COVID patients but remains elevated after recovery.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Biomarker B rises when symptoms worsen, falls when symptoms improve and changes before clinical deterioration.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Biomarker B is potentially far more valuable.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It could become:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>a diagnostic marker;<\/li>\n\n\n\n<li>a prognostic marker;<\/li>\n\n\n\n<li>a disease-activity marker;<\/li>\n\n\n\n<li>a treatment-response marker.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The distinction is critical.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The field should therefore prioritize <strong>dynamic biomarkers<\/strong>.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XV. The Ultimate Experiment<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The strongest test of a proposed mechanism is intervention.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Suppose persistent viral antigen is associated with a specific Long COVID phenotype.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Then:<\/p>\n\n\n\n<ol start=\"1\" class=\"wp-block-list\">\n<li>identify patients with that phenotype;<\/li>\n\n\n\n<li>document the biomarker;<\/li>\n\n\n\n<li>administer a targeted antiviral strategy;<\/li>\n\n\n\n<li>measure whether antigen falls;<\/li>\n\n\n\n<li>measure whether symptoms improve;<\/li>\n\n\n\n<li>repeat in an independent cohort.<\/li>\n<\/ol>\n\n\n\n<p class=\"wp-block-paragraph\">The same logic applies to:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>immune modulation;<\/li>\n\n\n\n<li>autonomic dysfunction;<\/li>\n\n\n\n<li>coagulation abnormalities;<\/li>\n\n\n\n<li>cortisol abnormalities;<\/li>\n\n\n\n<li>metabolic dysfunction;<\/li>\n\n\n\n<li>serotonin pathways.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The chain must become:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>biomarker \u2192 mechanism \u2192 intervention \u2192 biomarker response \u2192 clinical response.<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Only then does association approach causation.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XVI. Why Routine Blood Tests Can Be Normal<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The apparent contradiction between substantial biological abnormalities in research studies and normal routine clinical laboratory testing is therefore understandable.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Routine tests measure a limited physiological space.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A standard CBC may not measure:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>T-cell exhaustion;<\/li>\n\n\n\n<li>abnormal monocyte phenotypes;<\/li>\n\n\n\n<li>viral antigen reservoirs;<\/li>\n\n\n\n<li>tissue-specific inflammation;<\/li>\n\n\n\n<li>endothelial signaling;<\/li>\n\n\n\n<li>microbiome-derived metabolites;<\/li>\n\n\n\n<li>altered serotonin physiology;<\/li>\n\n\n\n<li>subtle autonomic dysfunction.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Likewise, a normal creatinine does not prove that endothelial or neurological systems are normal.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The absence of an abnormal routine laboratory result should therefore not be interpreted as evidence that no biological abnormality exists.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The NIH-supported RECOVER analysis makes precisely this distinction: routine laboratory testing is presently insufficient to diagnose Long COVID.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XVII. The Problem of Causality<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">A major danger now confronts Long COVID research.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">As the number of candidate biomarkers increases, so does the probability of false-positive associations.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">If 525 biomarkers have been reported, researchers must distinguish:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>association<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">from<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>causation.<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A biomarker can be:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>a cause;<\/li>\n\n\n\n<li>a consequence;<\/li>\n\n\n\n<li>a compensatory response;<\/li>\n\n\n\n<li>a marker of tissue injury;<\/li>\n\n\n\n<li>a marker of treatment;<\/li>\n\n\n\n<li>a marker of another disease;<\/li>\n\n\n\n<li>or a coincidental correlate.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Longitudinal sampling and intervention are therefore indispensable.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XVIII. A Proposed Five-Year Biomarker Consortium<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The existing datasets could support a much more ambitious project.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Phase 1: Phenotyping<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Enroll 10,000\u201320,000 newly infected individuals.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Collect standardized:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>symptom inventories;<\/li>\n\n\n\n<li>functional measurements;<\/li>\n\n\n\n<li>cognitive assessments;<\/li>\n\n\n\n<li>autonomic measurements;<\/li>\n\n\n\n<li>wearable data.<\/li>\n<\/ul>\n\n\n\n<h5 class=\"wp-block-heading\">Phase 2: Biological sampling<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">At predefined intervals collect:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>blood;<\/li>\n\n\n\n<li>plasma;<\/li>\n\n\n\n<li>serum;<\/li>\n\n\n\n<li>saliva;<\/li>\n\n\n\n<li>stool;<\/li>\n\n\n\n<li>urine;<\/li>\n\n\n\n<li>nasal specimens.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Where scientifically justified, specialized tissue studies could be added.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Phase 3: Multi-omics<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Measure:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>proteomics;<\/li>\n\n\n\n<li>metabolomics;<\/li>\n\n\n\n<li>transcriptomics;<\/li>\n\n\n\n<li>immune-cell phenotyping;<\/li>\n\n\n\n<li>cytokines;<\/li>\n\n\n\n<li>autoantibodies;<\/li>\n\n\n\n<li>viral antigen;<\/li>\n\n\n\n<li>viral RNA;<\/li>\n\n\n\n<li>microbiome composition.<\/li>\n<\/ul>\n\n\n\n<h5 class=\"wp-block-heading\">Phase 4: EHR integration<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Link:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>diagnoses;<\/li>\n\n\n\n<li>prescriptions;<\/li>\n\n\n\n<li>emergency visits;<\/li>\n\n\n\n<li>hospitalization;<\/li>\n\n\n\n<li>imaging;<\/li>\n\n\n\n<li>laboratory results;<\/li>\n\n\n\n<li>rehabilitation;<\/li>\n\n\n\n<li>disability;<\/li>\n\n\n\n<li>mortality.<\/li>\n<\/ul>\n\n\n\n<h5 class=\"wp-block-heading\">Phase 5: Prediction<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Train models to predict:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>recovery;<\/li>\n\n\n\n<li>persistent disease;<\/li>\n\n\n\n<li>relapsing disease;<\/li>\n\n\n\n<li>progressive disease.<\/li>\n<\/ul>\n\n\n\n<h5 class=\"wp-block-heading\">Phase 6: Intervention<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Test whether modifying the predicted biological pathway changes the trajectory.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Figure 9. The Complete Research Pipeline<\/h5>\n\n\n\n<pre class=\"wp-block-preformatted\"> <code>            SARS-CoV-2 INFECTION\n                      \u2502\n                      \u25bc\n              EARLY BIOLOGY\n                      \u2502\n        \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n        \u25bc             \u25bc             \u25bc\n      Virus         Immune        Host\n      burden        response      factors\n        \u2502             \u2502             \u2502\n        \u2514\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2518\n                      \u25bc\n                 BIOLOGICAL\n                  SIGNATURE\n                      \u2502\n                      \u25bc\n                 PHENOTYPE\n                      \u2502\n          \u250c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u253c\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2510\n          \u25bc           \u25bc            \u25bc\n       Recovery    Persistent   Progressive\n                      \u2502\n                      \u25bc\n                 TARGETED\n                INTERVENTION\n                      \u2502\n                      \u25bc\n                BIOMARKER CHANGE\n                      \u2502\n                      \u25bc\n                CLINICAL CHANGE\n                      \u2502\n                      \u25bc\n                  CAUSALITY<\/code><\/pre>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XIX. What the Current Data Already Establish<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The cumulative evidence now supports several propositions with reasonable confidence.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">First<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Long COVID is biologically heterogeneous.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Second<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Multiple biological systems are implicated.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Third<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The abnormalities are not restricted to psychological or subjective phenomena.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Fourth<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Some biological differences are measurable during acute infection before Long COVID becomes clinically established.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Fifth<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The strongest evidence currently points toward combinations of viral, immune, vascular, metabolic and neuroendocrine abnormalities rather than one universal mechanism.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Sixth<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Different symptom clusters may correspond to different biological signatures.<\/p>\n\n\n\n<h5 class=\"wp-block-heading\">Seventh<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Large-scale EHR and biobank datasets now make systematic validation possible.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XX. What Has Not Yet Been Proven<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">Several claims remain scientifically premature.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">There is not yet a universally validated blood test for Long COVID.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">No single biomarker has demonstrated adequate sensitivity and specificity across populations.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The presence of SARS-CoV-2 antigen or RNA in a tissue does not automatically prove that it is causing symptoms.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Autoantibodies have been observed, but their causal significance varies across studies.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Microclots and endothelial abnormalities remain active areas of investigation rather than universally established clinical diagnostic criteria.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Low cortisol is biologically intriguing but is not presently a stand-alone diagnostic test.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The existence of biological abnormalities does not establish that every symptom in every Long COVID patient arises from the same mechanism.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">These distinctions are essential if biomarker research is to progress from hypothesis generation to clinical medicine.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XXI. The Central Scientific Conclusion<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The evidence accumulated since 2020 has changed the nature of the Long COVID question.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The first question was:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Does a persistent post-COVID illness exist?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The answer is now unequivocally yes.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The second question was:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>How common is it?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Large population studies have provided increasingly precise estimates.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The third question is now:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>What biological state distinguishes patients who recover from those who remain ill?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The available evidence suggests that the answer will not be a single molecule.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It is likely to be a <strong>biological constellation<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That constellation may include:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>persistent or inadequately controlled viral material;<\/li>\n\n\n\n<li>altered innate and adaptive immunity;<\/li>\n\n\n\n<li>abnormal complement activity;<\/li>\n\n\n\n<li>endothelial dysfunction;<\/li>\n\n\n\n<li>coagulation abnormalities;<\/li>\n\n\n\n<li>altered cortisol physiology;<\/li>\n\n\n\n<li>altered serotonin metabolism;<\/li>\n\n\n\n<li>metabolic dysfunction;<\/li>\n\n\n\n<li>tissue-specific inflammatory responses;<\/li>\n\n\n\n<li>neurological and autonomic abnormalities;<\/li>\n\n\n\n<li>microbiome disruption.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">The relative importance of these mechanisms may vary dramatically from patient to patient.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">XXII. The New Paradigm: From \u201cLong COVID\u201d to Biological Phenotypes<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The term Long COVID is useful clinically but may eventually prove too broad scientifically.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Medicine may ultimately recognize several overlapping post-COVID biological syndromes:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Post-viral inflammatory phenotype<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Vascular\/endothelial phenotype<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Neuroautonomic phenotype<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Metabolic phenotype<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Persistent-antigen phenotype<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Post-infectious immune phenotype<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Multisystem phenotype<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">These categories should not be imposed prematurely.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">They should emerge from reproducible data.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Conclusion<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The world&#8217;s largest clinical databases and increasingly sophisticated biological studies have begun to transform Long COVID from an epidemiological mystery into a tractable biological problem.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The evidence is now too extensive to dismiss the illness as merely a collection of nonspecific symptoms, yet too heterogeneous to justify the opposite error of claiming that one mechanism explains every patient.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The data point toward something more complex.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Long COVID appears to be a network disorder with multiple biological entry points and multiple sustaining mechanisms.<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The strongest prospective evidence indicates that the biological story may begin during acute infection. Higher early viral burden, prolonged shedding, higher nucleocapsid antigen and weaker early spike-specific antibody responses were associated with later PASC in a carefully followed outpatient cohort.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The strongest immunological studies demonstrate altered immune-cell populations, antibody patterns and neuroendocrine signaling.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Proteomic studies identify distinct biological signatures associated with cardiorespiratory, gastrointestinal, cognitive and fatigue phenotypes.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Population-scale UK Biobank analyses identify metabolic, inflammatory and organ-specific risk markers.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">EHR analyses involving millions of patients demonstrate that the clinical consequences are genuinely multisystem.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">And the systematic biomarker literature has identified hundreds of candidate abnormalities without finding one sufficiently reliable universal marker.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Taken together, these observations suggest that the search for <strong>the<\/strong> Long COVID biomarker is probably asking the wrong question.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The more productive question is:<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>Can a multidimensional biological signature identify which biological pathway is dominant in a particular patient, predict that patient&#8217;s clinical trajectory, and identify a treatment capable of changing it?<\/strong><\/p>\n<\/blockquote>\n\n\n\n<p class=\"wp-block-paragraph\">That is the experiment that should define the next phase of Long COVID research.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The objective should no longer be merely to identify abnormalities.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It should be to connect:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>virus \u2192 biology \u2192 phenotype \u2192 trajectory \u2192 treatment \u2192 recovery.<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Only when that chain becomes experimentally demonstrable will Long COVID move from a disease recognized primarily through clinical history and symptom reporting to a disease that can be <strong>measured, stratified, predicted and treated according to its underlying biology.<\/strong><\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Footnotes and References<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>1.<\/strong> Thaweethai T, Donohue SE, Martin JN, et al. <em>Long COVID trajectories in the prospectively followed RECOVER-Adult US cohort<\/em>. Nature Communications. 2025;16:9557. The prospective cohort included 3,659 participants and identified eight longitudinal symptom trajectories.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>2.<\/strong> RECOVER COVID Initiative. <em>RECOVER Data<\/em>. The program provides public grouped\/anonymized data through BDC-PIC-SURE and controlled participant-level access through NHLBI BioData Catalyst. The latest observational data snapshots were released in May 2026.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>3.<\/strong> NIH. <em>NIH RECOVER makes long COVID data easier to access<\/em>. NIH reported that secure data from more than 14,000 adults were available to authorized researchers through BioData Catalyst.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>4.<\/strong> RECOVER. <em>Electronic health records are an important tool to address Long COVID questions<\/em>. RECOVER reports access to EHRs from more than 60 million people through PCORnet and N3C.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>5.<\/strong> O&#8217;Neil ST, Madlock-Brown C, Wilkins KJ, et al. <em>Finding Long-COVID: temporal topic modeling of electronic health records from the N3C and RECOVER programs<\/em>. npj Digital Medicine. The analysis included more than 600 million diagnoses from approximately 14 million patients.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>6.<\/strong> Lu S, Peluso MJ, Glidden DV, et al. <em>Early biological markers of post-acute sequelae of SARS-CoV-2 infection<\/em>. Nature Communications. 2024;15:7466. The study prospectively enrolled 136 people within five days of PCR positivity.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>7.<\/strong> Klein J, Wood J, Jaycox JR, et al. <em>Distinguishing features of Long COVID identified through immune profiling<\/em>. Nature. 2023. The investigators characterized immune-cell, antibody, hormonal and cytokine differences and developed machine-learning models of Long COVID status.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>8.<\/strong> Liew F, et al. <em>Large-scale phenotyping of patients with long COVID post-hospitalization reveals mechanistic subtypes of disease<\/em>. Nature Immunology. 2024. The study profiled 368 plasma proteins in 657 participants.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>9.<\/strong> Persistent serum protein signatures define an inflammatory subcategory of Long COVID. Nature Communications. The study identified inflammatory protein signatures involving interferon, NF-\u03baB, neutrophil and cytokine pathways.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>10.<\/strong> Blood Biomarkers of Long COVID: A Systematic Review. The review included 45 observational and one interventional study involving 4,415 participants and identified 525 candidate blood biomarkers.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>11.<\/strong> Alcalde-Herraiz M, Iqbal S, Wallin JJ, et al. <em>Sociodemographic factors, biomarkers and comorbidities associated with post-acute COVID-19 sequelae in UK Biobank<\/em>. Nature Communications. 2025. The study included 8,668 participants in the Long COVID cohort and 108,407 in the PACS cohort.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>12.<\/strong> Kehl-Floberg K, Freisberg E, Pop-Vicas A, et al. <em>Long COVID risk by pre-infection symptoms and functional status: A retrospective cohort study of data from the All of Us Research Program<\/em>. The analysis included 65,464 adults. The current publication reports associations between pre-infection symptom burden and later Long COVID symptoms.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>13.<\/strong> <em>Data-driven analysis to understand Long COVID using electronic health records from the RECOVER initiative<\/em>. The analysis used INSIGHT and OneFlorida+ and identified multiple replicated post-COVID diagnoses.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>14.<\/strong> <em>Risk Factors Associated with Post-Acute Sequelae of SARS-CoV-2 in an EHR Cohort: A National COVID Cohort Collaborative Analysis<\/em>. The N3C study included 8,325 PASC cases and 41,625 matched controls and identified associations with age, sex, hospitalization and several comorbidities.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>15.<\/strong> NIH. <em>Routine lab tests are not a reliable way to diagnose Long COVID<\/em>. The NIH-supported RECOVER study enrolled more than 10,000 adults and concluded that routine laboratory testing does not currently provide a reliable diagnostic biomarker.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>16.<\/strong> Wong AC, et al. <em>Serotonin reduction in post-acute sequelae of viral infection<\/em>. Cell. 2023. Subsequent reviews describe persistent serotonin reduction in participants with PASC and experimental evidence linking altered tryptophan\/serotonin physiology with cognitive effects.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>17.<\/strong> The immunology of Long COVID. Nature Reviews Immunology. The review synthesizes evidence concerning immune perturbation, endothelial dysfunction, coagulation, autoantibodies, viral reservoirs and herpesvirus reactivation.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h5 class=\"wp-block-heading\">Data-Analysis Note<\/h5>\n\n\n\n<p class=\"wp-block-paragraph\">The quantitative calculations in this article were derived from published aggregate results. Where individual-level datasets were not directly available, the analysis does <strong>not<\/strong> claim an independent reanalysis of patient-level records. Published cohort sizes, percentages, odds ratios, confidence intervals, AUC values and biomarker counts are reported as published. Derived proportions are explicitly identified as calculations. The figures describing mechanisms and proposed biological architecture are conceptual rather than empirical.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The term \u201call public and published databases\u201d is used operationally here to mean the major publicly accessible or published datasets presently capable of informing Long COVID biological research. It does not imply that every institutional, proprietary, unpublished or inaccessible dataset worldwide has been exhaustively searched.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><\/p>\n","protected":false},"excerpt":{"rendered":"<p>What the World&#8217;s Largest Clinical and Molecular Datasets Are Beginning to Reveal John Murphy, CEO, The COVID-19 Long-haul Foundation Abstract Long COVID presents one of the most consequential unresolved problems [&hellip;]<\/p>\n","protected":false},"author":2,"featured_media":16154,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[100,252,1320,289,290,1572,1571,421],"tags":[],"class_list":["post-16150","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-cortisol","category-inflammation","category-long-cov","category-long-haul-disease","category-long-term-effects","category-multimodal-biological-signatures","category-network-disease","category-pasc"],"_links":{"self":[{"href":"https:\/\/cov19longhaulfoundation.org\/index.php?rest_route=\/wp\/v2\/posts\/16150","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cov19longhaulfoundation.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/cov19longhaulfoundation.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/cov19longhaulfoundation.org\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/cov19longhaulfoundation.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=16150"}],"version-history":[{"count":2,"href":"https:\/\/cov19longhaulfoundation.org\/index.php?rest_route=\/wp\/v2\/posts\/16150\/revisions"}],"predecessor-version":[{"id":16152,"href":"https:\/\/cov19longhaulfoundation.org\/index.php?rest_route=\/wp\/v2\/posts\/16150\/revisions\/16152"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/cov19longhaulfoundation.org\/index.php?rest_route=\/wp\/v2\/media\/16154"}],"wp:attachment":[{"href":"https:\/\/cov19longhaulfoundation.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=16150"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/cov19longhaulfoundation.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=16150"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/cov19longhaulfoundation.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=16150"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}