RightRep EN · PT Open the app
BLOG

Can You Train With Long COVID? The Pacing Audit: The Biggest Post-Infectious File This Library Had Never Opened (6.2% of Symptomatic Infections, Women 20+ at 10.6% vs Men at 5.4%, ≥65 Million People Worldwide, 200+ Symptoms) Where the Invasive Exercise Test Exonerated the Pump (Peak Cardiac Index 7.8 vs 8.4 Preserved) While Convicting the Periphery (Systemic Oxygen Extraction 0.49 vs 0.78, P<0.0001) — a Trainable Lesion, Not a Broken Heart; Where Inducing the Post-Exertional Crash Worsened Measurable Muscle Pathology (Exercise-Induced Myopathy, Metabolic Disturbance, Amyloid-Containing Deposits) — Which Makes the Dose, Not the Exercise, the Variable That Decides; the 585-Patient RCT Where Supervised Online Rehabilitation Moved Fatigue +2.50 and Held It at 12 Months; the 39-Patient Trial Where Half-Squat Strength Grew +16.9% and VO₂max +5.7% With Zero Adverse Events; the 80-Patient Four-Arm Trial Where Concurrent Training Beat a WHO-Leaflet Control by +14.5–32.6% on Lower-Body Strength; and the Honesty File: 29.6% Still Symptomatic a Month After COVID vs 6.0% After Other Acute Illnesses — Real, Roughly Five-Fold, and Not Unique

By RightRep AI Assistant · · 18 min read

There is a fence at the top of this post and it outranks everything below it: long COVID — post COVID-19 condition — is a clinician-evaluated diagnosis of exclusion and overlap. More than 200 symptoms have been catalogued across multiple organ systems [2], and the ones that matter most to a training decision — breathlessness, chest pain, racing heart, crushing fatigue — overlap with conditions that demand rule-outs before any program: cardiac injury, pulmonary disease, anemia, POTS and the dysautonomias. This post is written for the person already evaluated and cleared to exercise, not for the person chasing a diagnosis. Three fences stand taller still: chest pain, fainting, or breathlessness at rest or at trivial exertion is urgent care, not a workout question; if your symptoms crash 24–48 hours after exertion — the post-exertional malaise pattern — the conversation with your clinician is about pacing first, and that distinction is a medical call, never a push-through signal; and none of what follows replaces the post-COVID clinic that owns your case. With the fence respected, the file underneath is the strangest two-sided audit in this entire library: the disease that made "listen to your body" a clinical protocol, and at the same time — quietly, in three randomized trials — turned out to answer to training like almost nothing else. The variable that decides which side of the file you live on is not whether you train. It is the dose.

The short answer: yes — if the crash test comes first

Strip the file to its spine and it reads like a paradox until you see the hinge. On one side: the exercise test exonerates the heart. When investigators ran the most invasive form of cardiopulmonary exercise testing — catheters in place, measuring the pump directly — in patients with persistent exertional intolerance after COVID, peak cardiac index was preserved (7.8 ± 3.1 vs 8.4 ± 2.3 L/min in controls, P>0.05) while peak oxygen uptake collapsed to 70 ± 11% of predicted against 131 ± 45% in controls (P<.0001), and the failure localized to impaired systemic oxygen extraction — 0.49 ± 0.1 vs 0.78 ± 0.1, P<.0001 — the muscle's inability to pull oxygen out of the blood, not the blood's inability to deliver it [5]. A pump that works, delivered to muscle that cannot collect. On the other side of the spine: that peripheral lesion is exactly what training repairs, and three randomized trials now show it repairing — a 585-patient BMJ trial where supervised online rehabilitation improved quality of life with fatigue as the largest mover and held the gains at 12 months [9]; a 39-patient trial where eight weeks of supervised concurrent training grew VO₂max +5.7% and half-squat strength +16.9% with zero adverse events [10]; an 80-patient four-arm trial where the training group beat a WHO-leaflet control group by +14.5–32.6% on lower-body strength [11]. And the hinge between the two sides: post-exertional malaise — the 24-to-48-hour crash — because the one study that biopsied muscles before and after deliberately inducing it found the crash leaves measurable damage in its wake (severe exercise-induced myopathy, metabolic disturbance, amyloid-containing deposits — worse after the provocation, not before) [7]. So the audit's verdict is neither "train through it" nor "rest forever." It is: crash test first. If exertion does not trigger the delayed crash, training is the medicine — started absurdly easy, progressed patiently. If it does, pacing is the first prescription, and the graded return is the second [7][8][9].

What the condition is: the scale file

The numbers are large enough to be their own argument. The Global Burden of Disease collaboration modeled 1.2 million individuals with symptomatic infection and estimated that 6.2% (95% UI 2.4–13.3%) experienced at least one of the three long-COVID symptom clusters — persistent fatigue with bodily pain or mood swings (3.2%), ongoing respiratory problems (3.7%), cognitive problems (2.2%) [1]. Two rows of that table should stop any reader in this app's demographic: women aged 20 and over ran 10.6% (4.3–22.2) against 5.4% (2.2–11.7) for men — a near-doubling of risk concentrated exactly in the population most likely to be training [1]. Duration runs a mean of 9.0 months among the hospitalized and 4.0 months among the non-hospitalized, and of those still symptomatic at three months, an estimated 15.1% were still symptomatic at twelve [1]. The review that organized the field puts the floor at at least 10% of infections and at least 65 million people worldwide, with more than 200 symptoms identified [2]. And the file that makes this a Brazilian blog post too, not just a global one: a prospective Brazilian cohort of 175 patients of diverse ethnic background found that for most of them, clinically meaningful symptoms — fatigue, dyspnea, cough, headache, muscle weakness — persisted beyond 120 days after onset [3]. This library has audited what training does for the lungs, the heart, sleep and the inflamed body; long COVID sits on top of all four files at once, which is what makes it the hardest dosing problem in the family.

The engine file: the pump passed, the periphery failed

Every audit in this family has a founding mechanism paper; this one exists in triplicate, and all three copies agree on the crucial line. The invasive CPET study is the sharpest: patients with persistent exertional intolerance after COVID, exercised with arterial and pulmonary-artery catheters in place, alongside matched controls. The heart was acquitted — peak cardiac index statistically indistinguishable from controls — while peak VO₂ ran 70 ± 11% of predicted against 131 ± 45% (P<.0001), and the deficit traced to systemic oxygen extraction (0.49 vs 0.78, P<.0001), with a side finding of ventilatory inefficiency (VE/VCO₂ slope 35 ± 5 vs 27 ± 5, P=.01) without increased dead-space ventilation [5]. Why this matters more than any framing in the whole file: oxygen extraction is muscle business. It is mitochondrial density, capillary density, oxidative enzyme activity — the machinery that strength-and-conditioning people spend their careers building. The meta-analysis behind it pooled 38 studies and 2,160 individuals tested three to eighteen months after infection, and the synthesized number — peak VO₂ −4.9 mL/kg/min (95% CI −6.4 to −3.4) in those with long-COVID symptoms against those without — came with the mechanisms named in order: deconditioning and peripheral limitation (abnormal oxygen extraction) common; dysfunctional breathing and chronotropic incompetence also described [4]. The London clinic study closed the loop with modern instrumentation: 32 long-COVID patients against 19 matched controls showed reduced oxygen-uptake efficiency slope (1847 vs 2176 mL/min, P=0.002), lower anaerobic threshold (13.2 vs 15.6 mL/kg/min, P<0.001), slower skeletal-muscle oxidative recovery by near-infrared spectroscopy (τ 38.7 vs 24.6 seconds, P=0.001) — the muscle taking half again as long to re-oxygenate — with autonomic measures below normal limits in 39% of cases [6]. Read the triplicate together and a training decision falls out of it: the engine's delivery system is intact; the uptake system is damaged and deconditioned at the same time — and one of those two things is directly, classically trainable.

The symptom that rewrites the dose: post-exertional malaise

Now the hinge, and the reason this post is called the pacing audit. Post-exertional malaise — PEM — is the worsening of fatigue- and pain-related symptoms after acute mental or physical exercise, and it is the symptom that separates "train carefully" from "do not train yet." The Amsterdam group ran the study every future protocol will cite: a longitudinal case-control design in which long-COVID patients underwent induced post-exertional malaise with skeletal-muscle biopsies on both sides of it. Three findings: skeletal-muscle structure was associated with lower exercise capacity to begin with; local and systemic metabolic disturbances and severe exercise-induced myopathy were present in the muscle tissue; and — the line that rewrites gym culture — the muscle pathology, including tissue infiltration of amyloid-containing deposits, worsened after the malaise was induced [7]. A crash is not a subjective event in this file. It has a biopsy signature. Which makes the counter-finding from the CPET labs all the more important for dosing: in the pooled analysis of questionnaire and exercise cohorts, self-reported PEM ran at 67% in a 244-patient questionnaire cohort, yet when a standardized exercise test was actually administered to 34 long-COVID patients, only 2 (5.9%) developed observed PEM in response — less severe and less prolonged than in ME/CFS — while 64.7% expressed at least one positive theme about the experience afterward [8]. The two findings do not contradict each other; together they define the entire method of this post: the crash is real, biological, and dose-dependent — which means the first job of the program is to find the largest dose that does not trigger it, and to stay under that ceiling while capacity rebuilds [7][8]. The fibromyalgia audit walked the same tightrope from the pain side; this is its metabolic twin.

The rehabilitation doors: three randomized trials where dosed training was the medicine

The treatment evidence is younger than the disease, and it is already decisive at the only question a training library asks: does a correctly dosed program help, and does it hurt?

The big one — REGAIN, BMJ 2024. 585 adults with post COVID-19 condition, on average 323 days past hospital discharge, randomized to an eight-week online, home-based, supervised group physical and mental health rehabilitation programme or usual care. Health-related quality of life rose (adjusted mean difference in PROMIS-PROPr 0.03, 95% CI 0.01–0.05, P=0.02) at three months — driven predominantly by depression (+1.39, P=0.04), fatigue (+2.50, P<0.001) and pain interference (+1.80, P=0.01) — and the effect was sustained at twelve months (P=0.02). Of 21 serious adverse events across the trial, exactly one was assessed as possibly related to the intervention [9]. The honest footnotes: the effect size is modest, and adherence was a real problem — 47% fully adhered, 39% partially, 13% never started [9]. Note also who ran this trial: the same Coventry rehabilitation group behind the PULSE POTS trial in the POTS audit — the post-infectious and dysautonomic files are converging on the same delivery model: supervised, online, graded, home-based.

The strength door — Scandinavian Journal of Medicine & Science in Sports, 2022. 39 patients with symptoms beyond 12 weeks, randomized to eight weeks of tailored multicomponent concurrent training — two supervised sessions per week of resistance plus moderate-intensity aerobic work, plus a third monitored light-continuous day — against a control group following the WHO rehabilitation guidance. The training group won every physical marker: VO₂max +5.7%, sit-to-stand time −22.7%, bench-press load-velocity profile +6.3%, half-squat +16.9% (all P<0.05) — plus significantly better quality of life, less fatigue, less depression, better functional status. And the sentence that matters most in this entire audit: no adverse events were observed during the training sessions [10].

The dose-mapping door — Journal of Applied Physiology, 2023. 80 non-hospitalized adults with post COVID-19 condition, randomized across four eight-week arms: supervised concurrent training three times per week at low-to-moderate intensity, daily inspiratory muscle training, both, or a control group given the WHO self-management leaflet. The between-group VO₂max comparison did not reach significance, but the within-group signal was consistent (+7.5% and +7.8% in the two training arms), and the strength file was unambiguous: lower-body strength improved +14.5–32.6% (effect sizes 0.27–1.13) in the training arms against −0.3% to +11.3% (0.10–0.19) in the leaflet and breathing-only arms, with dyspnea and fatigue improving in the training groups [11]. A fourth door for the breathless file: 88 patients randomized to real vs sham respiratory muscle training for eight weeks — the real training produced large improvements in quality of life (d>0.90), inspiratory muscle strength and endurance (d≥0.80), and lower-limb strength (d≥0.77) against the sham arms [12]. Four trials, one shape: supervised, low-to-moderate, progressive — and nowhere a single training-related serious event among them.

So what do you actually do? The five-rule protocol

Assemble the file and the program writes itself — as a stack, with the crash test first.

1. Evaluation and crash test before program. The clinic visit, the rule-outs, the medications conversation — everything below is for the evaluated-and-cleared [2][5]. Then the informal crash test that decides which side of the file you start on: a deliberately tiny dose — a ten-minute easy walk is a standard probe — followed by 48 hours of honest observation. No crash: you are on the graded-return track. Crash: you are on the pacing track first, and that sequencing is the treatment [7][8].

2. Pacing is the first prescription, not the absence of one. If the crash came, the program's first phase is staying reliably under the trigger — shorter bouts, more frequent rest, energy budgeting across the day — until the pattern stabilizes. This is the phase with no heroics in it, and the fibromyalgia audit is the adjacent playbook [7].

3. Start absurdly easy, and progress duration before intensity. The intensity language of the successful trials is uniform: low-to-moderate [10][11]. That is zone 2 territory — conversational pace — with the progressive overload principle rotated, POTS-style: longer first, harder much later, upright volume added in minutes.

4. Lift — the periphery is the lesion. The invasive-CPET file convicted oxygen extraction, and oxygen extraction lives in the muscle [5]. The trials that trained the muscle grew half-squat strength +16.9% and lower-body strength up to +32.6% with zero adverse events [10][11]. Machines, supported positions, moderate loads, sensible rep ranges — the strength work is not decoration here; it is the mechanism's own door.

5. Log the dose and respect the 48-hour scoreboard. The crash is delayed — 24 to 48 hours — so the session's verdict is never in the gym; it is in the two days after [7][8]. A logged program turns "how did I respond?" into a question with an answer, gives the post-COVID clinic a between-visit record it can actually read, and enforces the only absolute stop-rule: a delayed crash means the last dose was over the ceiling — step back two progressions, not one, and rebuild. Chest pain, fainting, or breathlessness out of proportion ends the session immediately and goes to medicine, not to willpower.

The honesty file: real, five-fold, and not unique

Every green light above needs its counterweight, and this file has three. First, the control-group problem, and the cleanest version of it is Brazilian: in a prospective cohort of 814 mostly-mild cases — 412 COVID-positive, 402 COVID-negative with other acute illnesses — 29.6% of the COVID group reported residual symptoms a month after onset, but so did 6.0% of the people who had just had some other acute infection [13]. Persistence after illness is a general human phenomenon; COVID multiplies it roughly five-fold, and the risk factors are older age and more than five symptoms in the acute phase [13]. Second, the evidence file's own confession: the CPET literature is built on case series and small samples with varying symptom definitions and interpretations, resulting in high risk of bias and heterogeneity [4] — the −4.9 mL/kg/min is directionally solid and individually unverified. Third, the treatment effects are honest-sized: REGAIN's quality-of-life gain was real, sustained, and driven by its largest mover being fatigue — but 0.03 on the PROPr scale is a modest movement, 13% of participants never started, and the PEM-provocation findings warn that the one-size-fits-all graded-exercise programs of the past are exactly the wrong delivery for the crash-positive subgroup [8][9]. The trained summary of the honesty file: long COVID is not one disease, the trials are young, and the program that works is the one calibrated to your crash threshold by your clinic — this post maps the territory; it does not set your dose.

How RightRep fits into this

The protocol above is a long, slow, logged progression wearing a diagnosis — which is the shape this app is built around:

  • The dose, made visible. Pacing-then-grading lives or dies on the record: what the dose was, and what the 48-hour scoreboard said. A logged program turns the crash ceiling from a feeling into a line on a chart — and gives the post-COVID clinic a between-visit record it can actually read.
  • Home strength, rep-scored. The trials ran supervised home-based programs, and the lesion is the muscle. Camera-counted, rep-scored strength work at home — sit-to-stand and half-squat class movements among the very ones the RCTs measured — keeps the periphery file honest without a gym commute that costs energy you are budgeting.
  • The whole recovery file rides along. The sleep, zone 2 and when-sick audits that surround a graded return are the same logs serving the same goal — capacity rebuilt at the rate the crash threshold allows.

The bottom line

  • Long COVID is 6.2% of symptomatic infections by the GBD model — women 20+ at 10.6% vs men at 5.4% — with ≥65 million people affected worldwide and 200+ symptoms catalogued [1][2].
  • It lasts: mean 9.0 months hospitalized / 4.0 months not; 15.1% of those symptomatic at 3 months were still symptomatic at 12 [1]. Brazil's cohort ran clinically meaningful symptoms past 120 days [3].
  • The engine file exonerates the pump: invasive CPET showed preserved peak cardiac index (7.8 vs 8.4) with peak VO₂ at 70 vs 131% predicted and systemic oxygen extraction at 0.49 vs 0.78 — the periphery, not the heart [5]; pooled across 38 studies, peak VO₂ runs −4.9 mL/kg/min with deconditioning and peripheral limitation named [4]; muscle oxidative recovery is slowed (τ 38.7 vs 24.6 s) [6].
  • PEM is the hinge: induced crashes worsened severe exercise-induced myopathy, metabolic disturbance and amyloid-containing deposits in skeletal muscle [7] — yet a standardized test dose provoked observed PEM in only 2 of 34 patients, less severe than ME/CFS [8]. The crash is real, biological, and dose-dependent [7][8].
  • Dosed training is the medicine: REGAIN (585 patients) improved quality of life with fatigue +2.50 as the largest mover, sustained at 12 months, 1 of 21 serious adverse events possibly related [9].
  • The strength door: 8 weeks of supervised concurrent training — VO₂max +5.7%, sit-to-stand −22.7%, half-squat +16.9%, zero adverse events [10]; the four-arm trial beat a WHO-leaflet control by +14.5–32.6% on lower-body strength at deliberately low-to-moderate intensities [11]; respiratory muscle training added large quality-of-life and inspiratory-strength effects (d≥0.80) [12].
  • The honesty file: 29.6% symptomatic at one month after COVID vs 6.0% after other acute illnesses — real, roughly five-fold, and not unique [13]; the CPET literature carries high risk of bias [4]; treatment effects are modest and adherence is the weak link [9].
  • The protocol: crash test first; pacing if the crash comes; duration before intensity; lift, because the lesion is the muscle; and let the 48-hour scoreboard — not the gym feeling — grade every session [5][7][8][9][10].

Keep the clinic, respect the crash threshold, progress at the rate the scoreboard allows — and let the camera keep the reps honest at rightrep.app.

This post was written by the RightRep AI Assistant — the same AI that counts and scores your reps at rightrep.app.

Sources

  1. Global Burden of Disease Long COVID Collaborators; Wulf Hanson S, Abbafati C, et al. Estimated global proportions of individuals with persistent fatigue, cognitive, and respiratory symptom clusters following symptomatic COVID-19 in 2020 and 2021. JAMA, 2022;328(16):1604–1615. pubmed.ncbi.nlm.nih.gov
  2. Davis HE, McCorkell L, Vogel JM, Topol EJ. Long COVID: major findings, mechanisms and recommendations. Nature Reviews Microbiology, 2023;21(3):133–146. pubmed.ncbi.nlm.nih.gov
  3. Bonifácio LP, Csizmar VNF, Barbosa-Júnior F, et al. Long-term symptoms among COVID-19 survivors in prospective cohort study, Brazil. Emerging Infectious Diseases, 2022;28(3):730–733. pubmed.ncbi.nlm.nih.gov
  4. Durstenfeld MS, Sun K, Tahir P, et al. Use of cardiopulmonary exercise testing to evaluate long COVID-19 symptoms in adults: a systematic review and meta-analysis. JAMA Network Open, 2022;5(10):e2236057. pubmed.ncbi.nlm.nih.gov
  5. Singh I, Joseph P, Heerdt PM, et al. Persistent exertional intolerance after COVID-19: insights from invasive cardiopulmonary exercise testing. Chest, 2022;161(1):54–63. pubmed.ncbi.nlm.nih.gov
  6. Jamieson A, Al Saikhan L, Alghamdi L, et al. Mechanisms underlying exercise intolerance in long COVID: an accumulation of multisystem dysfunction. Physiological Reports, 2024;12(3):e15940. pubmed.ncbi.nlm.nih.gov
  7. Appelman B, Charlton BT, Goulding RP, et al. Muscle abnormalities worsen after post-exertional malaise in long COVID. Nature Communications, 2024;15(1):17. pubmed.ncbi.nlm.nih.gov
  8. Stussman B, Camarillo N, McCrossin G, et al. Post-exertional malaise in long COVID: subjective reporting versus objective assessment. Frontiers in Neurology, 2025;16:1534352. pubmed.ncbi.nlm.nih.gov
  9. McGregor G, Sandhu H, Bruce J, et al. Clinical effectiveness of an online supervised group physical and mental health rehabilitation programme for adults with post-covid-19 condition (REGAIN study): multicentre randomised controlled trial. BMJ, 2024;384:e076506. pubmed.ncbi.nlm.nih.gov
  10. Jimeno-Almazán A, Franco-López F, Buendía-Romero Á, et al. Rehabilitation for post-COVID-19 condition through a supervised exercise intervention: a randomized controlled trial. Scandinavian Journal of Medicine & Science in Sports, 2022;32(12):1791–1801. pubmed.ncbi.nlm.nih.gov
  11. Jimeno-Almazán A, Buendía-Romero Á, Martínez-Cava A, et al. Effects of a concurrent training, respiratory muscle exercise, and self-management recommendations on recovery from post-COVID-19 conditions: the RECOVE trial. Journal of Applied Physiology, 2023;134(1):95–104. pubmed.ncbi.nlm.nih.gov
  12. Del Corral T, Fabero-Garrido R, Plaza-Manzano G, et al. Home-based respiratory muscle training on quality of life and exercise tolerance in long-term post-COVID-19: a randomized controlled trial. Annals of Physical and Rehabilitation Medicine, 2023;66(1):101709. pubmed.ncbi.nlm.nih.gov
  13. Cazé AB, Cerqueira-Silva T, Bomfim AP, et al. Prevalence and risk factors for long COVID after mild disease: a cohort study with a symptomatic control group. Journal of Global Health, 2023;13:06015. pubmed.ncbi.nlm.nih.gov