PEM crash warning signs: early signals in your body and your watch data, and what to do about them
No watch can predict a PEM crash. But the early signs of a crash that is already building can show up in your body and in your own resting heart rate, HRV and sleep, one to three days after a heavier day.
Short answer
There is no reliable way to know for sure that a PEM crash is coming, and no app or wearable has been validated to predict one (6, 7). What many people learn to spot is the start of a crash that is already building. In the body, that often looks like flu-like aches, a sore throat or tender glands, heavier limbs, thicker brain fog, being "wired but tired" at night and sleep that does not refresh (4, 5). In watch data, it can show up as a resting heart rate above your own usual range, HRV below it for several mornings in a row, a higher heart rate through the night, and a busy day one to three days back. Read these against your own normal, not a population chart. Treat two or more of them together as a reason to cut load for the next 24 to 72 hours, not as a verdict.
If you crashed after a week when you "did everything right", you are not careless and you are not imagining it. Post-exertional malaise is delayed by design: the effort and the bill arrive on different days (1, 3). Before you blame yourself, check what your body already recorded.
Note: this article is for general education. It is not medical advice and does not replace a clinician. PEM is recognised clinically, not by an app. New, severe or different symptoms need a medical assessment.
Learn your own crash warning signs with women who get it
Welltory Energy Lab is a paid, moderated online community for women 35 to 65 living with energy-limiting conditions. Peer support and education, not medical care.
Can you tell when a PEM crash is coming?
Partly, and less than you would like. Post-exertional malaise (PEM) is a worsening of symptoms after physical, mental or emotional effort that you could previously manage. It typically starts 12 to 48 hours afterwards and can last days or weeks (1, 2). The Bateman Horne Center puts the window at 12 to 72 hours (5). If you are new to the term, start with our guide to post-exertional malaise.
The timing is not fixed, even for one person. In a survey of 150 people with ME/CFS, about 40% said PEM consistently began within 24 hours of the trigger, 42% said the timing varied, and up to 37% sometimes did not feel it until a day or more later (4). So "a crash is coming" usually means one of two things:
The cost is already incurred and the crash is still building. The effort happened yesterday or the day before. Your body may be showing early signs, and how you spend the next day or two can affect how deep and how long the crash gets.
You are about to spend more than you have. This is the part no sensor can see in advance. It is what pacing and an energy envelope are for.
This article is about the first situation: how to notice the early signs, which numbers on your watch tend to move with them, and what to do. It does not promise a forecast, because no one can honestly give you one.
Early warning signs of a PEM crash in your body
People who live with crashes for a while often describe a recognisable "something is coming" phase. It differs from person to person, and the same person can have different kinds of crashes. These are the signs that come up most often in clinical descriptions and in what people with ME/CFS and Long COVID report (4, 5):
Flu-like feelings. Aches, chills, a sore throat, tender lymph nodes, sometimes feeling feverish. In the ME/CFS survey above, 60% reported at least one of these immune-type symptoms after exertion (4).
Heavy, "expensive" movement. Limbs that feel like wet sand, stairs that suddenly cost more, muscles that ache without a reason.
Thicker brain fog. Losing words, rereading the same line, struggling to follow a conversation you would normally handle.
Wired but tired. Exhausted and buzzing at the same time, a racing mind at bedtime, sleep that will not come or does not refresh.
More sensitivity. Noise, light, smells and movement around you become harder to tolerate (5).
Orthostatic symptoms getting louder. A faster heart on standing, dizziness in queues or the shower, needing to sit down sooner. Orthostatic intolerance often travels with ME/CFS and Long COVID, and POTS has its own criteria (12).
The "unexpectedly good day". This one is a trap rather than a symptom. A day when you feel better than you have in weeks is often the day people overspend, and the bill arrives one to three days later. Our article on the boom and bust cycle explains why.
None of these is specific. A cold, a bad night, heat, a hard week at work or the days before your period can produce several of them. What makes them useful is your own history: which two or three signs tend to show up for you, in what order, before the crashes you remember.
What your watch data can show before a crash
A watch cannot see PEM. It can see a few things that tend to change when your body is under extra load: your heart rate at rest and during sleep, your heart rate variability (HRV, the small variation in time between heartbeats), how long and how well you slept, and how much you moved. The useful question is not "is this number good?", but "is this number unusual for me?".
| Signal | What a shift often looks like | What else moves it | How much to trust it |
|---|---|---|---|
| Morning resting heart rate | Several beats above your own usual range, two or more mornings in a row | Infection, poor sleep, alcohol, heat, dehydration, luteal phase, stimulants | Most robust of the set; read as a trend |
| Morning HRV | Below your usual range for several mornings, or an oddly high reading in the middle of feeling awful | Same as above, plus breathing, posture, measurement timing, beta-blockers | Useful as a direction, weak for a single day |
| Night heart rate | Higher through the night, no clear dip in the early hours | Late meals, alcohol, room heat, illness | Good companion to resting heart rate |
| Sleep | Shorter, broken, "wired" nights, waking unrefreshed | Pain, stress, noise, cycle | A sign of load, also a cause of the next one |
| Steps and active time | A busy day one to three days back, often a "good day" | Errands, appointments, travel | Shows the cause, not the warning; misses mental load |
| Stationary stress | More time with a raised heart rate while sitting or lying | Standing, emotional load, heat | Catches load that steps miss |
Resting heart rate: a few beats above your own normal
Resting heart rate is the steadiest signal most wearables give. Across 92,457 adults, average daily resting heart rate ranged from 40 to 109 bpm between people, but within one person it stayed much more consistent over time (8). That is why a population "normal" says little and your own line says a lot. In a wearable study of people after COVID-19, those who reported Long COVID had a resting heart rate that stayed above their own pre-infection baseline for 133 days on average, against 71 days in those without Long COVID (9). A morning resting heart rate that sits above your usual range for two or three days, without an obvious cold or bad night, is a pattern some people describe when they look back at their data before a crash. Our guide to resting heart rate with a chronic illness covers how chronic conditions shift this number in the first place.
HRV: falling for several mornings, or oddly good during a crash
HRV is lower on average in ME/CFS and Long COVID than in healthy people (10, 13), so your normal may look "bad" on any standard chart. What matters is the direction against your own range. In a 2026 study of 121 people with Long COVID, HRV stayed below that of healthy controls for 24 hours after exercise near the first ventilatory threshold, and night-time HRV dropped further after harder and longer sessions (7). That fits the delayed pattern of PEM: the cost of a big day can show up in the next night and the next morning.
There is a catch that surprises people. In some bodies with ME/CFS, POTS, Long COVID or fibromyalgia, an unusually high or "good-looking" HRV reading can appear in the middle of a crash rather than during recovery. A calmer number in that context can mean the body has less left to react with, not more (Welltory Help Center). If your HRV says "recovered" and your body says "hit by a truck", believe your body. Our article on morning HRV for pacing explains how to take a reading that means something.
Sleep and night heart rate
Many people notice that sleep goes first: they cannot fall asleep despite exhaustion, wake at 3 a.m. with a pounding heart, or sleep nine hours and wake up worse. On a watch this often looks like a night heart rate that never really drops and shorter or more broken sleep. A consistently raised night pulse can also simply be part of your baseline with a chronic condition, which is why one night read alone says little (Welltory Help Center). A bad night after a busy day is worth noticing; a run of them is worth acting on.
Steps and active time: look one to three days back
Steps rarely warn you. They point to the cause. The busy day, often the good day, is usually one to three days before the crash, not the day of it. And steps miss a lot: standing in a queue, a long phone call, a hard conversation, a noisy room. That is how people crash with zero steps. Stationary stress, time with a raised heart rate while you are not moving, catches some of that invisible load.
Why the signal is usually one to three days upstream
The most useful habit is to stop looking at yesterday and start looking two or three days back. Effort and payback are separated by hours to days (1, 4). The day you feel the crash is rarely the day that caused it.
Morning readings carry information about the same day. In a 2026 study of 4,244 people with complex chronic illness who took a 60-second morning reading and reported symptoms each evening, mornings with a higher heart rate and lower HRV than the person's own usual were followed by worse crash, fatigue and brain fog reports that evening. Adding the morning reading to the previous day's symptoms improved how well a model could tell worse days from better ones (AUC 0.82–0.85 against 0.73–0.83 with symptoms alone) (6). That is a group-level finding about the same day, with a model, not a personal two-day forecast.
Our own search through 700,000 days of wearable data for a PEM signal pointed the same way. Averages barely separated people who reported PEM from people who did not. What differed was stability: how predictably the body recovered from day to day, how quickly the nervous system shifted gears after effort, and how much cardiac work each unit of walking cost. One bad recovery night also tended to be followed by another. That research algorithm is not in the Welltory app, it misses more than half of self-reported cases, and it does not tell anyone that a crash is coming. What it gives you is a way to look: at the spread of your numbers across a fortnight, not at one morning's value.
How to build your own "normal line"
A warning sign only means something against a baseline. Building one is simple, but it takes a few weeks and some consistency.
Measure the same way, at the same time. Morning resting heart rate and HRV taken on waking, before coffee and before getting up, in the same position. Wear the same device. Changing devices or posture changes the numbers.
Collect two to four weeks. Two weeks is the minimum to see your usual range; four is better. Welltory's Crash Tracker, for example, needs at least 14 reliable days out of the previous 30 before it shows a personal boundary at all (Welltory Help Center).
Look at the range, not just the average. Note your typical value and how much it usually moves. A steady, mediocre number is a different situation from an unstable good one.
Mark the obvious disruptors. Infections, travel, alcohol, heat waves, new medicines, poor nights and your menstrual cycle. Cardiac vagal activity, which HRV reflects, changes measurably across the cycle and is usually lower in the luteal phase (11). Some medicines, beta-blockers in particular, flatten heart rate and HRV and can hide load.
Write the lagged column. Each evening, note how you felt and whether anything worsened compared with one and two days ago. This column is what connects the numbers to a crash. Our activity and symptom diary guide has a template.
Re-check your line every season. As your illness changes, so does your normal. A baseline from six months ago may be too generous or too strict now.
A simple amber check: when signals line up
You do not need a score to decide. A practical way to organise what you see is three states, built from your own data and body:
Green: your numbers are inside your usual range and your body feels like your normal. Follow your usual pacing plan.
Amber: two or more signals are off your line for two or more mornings, or one signal is off and your body is showing one of your personal early signs. Treat the next 24 to 72 hours as a reduced day.
Red: you are in a crash. Stop planning and protect rest.
This is a way to organise decisions, not a validated rule. Some amber days will pass without a crash, and some crashes will come on days that looked green in every number. Over time, compare your amber days with what happened next and adjust which signals you weight. If your heart rate goes over your pacing ceiling during the day, our guide to heart rate pacing covers how to set and use that limit.
What to do at the first signs of a crash
The early phase is the part you can influence. Guidance for ME/CFS recommends that people reduce activity and rest when symptoms flare, and adjust plans until they return to their usual level (1). In practice, an amber plan often looks like this:
Drop the biggest demand. Postpone the appointment, the social visit, the shopping trip or the hard conversation if you can. One cancelled demand is worth more than ten small savings.
Cut upright time. Sit to shower, sit to cook, lie down between tasks. Standing is an energy cost of its own.
Rest before you need to. Short, planned rests through the day, not one collapse at the end.
Lower sensory load. Dimmer light, quieter rooms, fewer screens and fewer messages.
Do not treat a good-looking number as permission. A green score on an amber morning is not a reason to catch up.
Keep the plan for the whole window. Payback runs 24 to 72 hours. Ending the amber plan the first evening you feel slightly better is the most common way to turn a mini crash into a long one.
Have a script ready. A pre-agreed message for work, family or friends ("I'm in a low-energy patch, I need to move today's plan") saves you the energy of explaining while foggy.
Log it. Note the signs, what you did and how it went. Next time you will know more about which signs matter for you.
Salt and fluid changes, medicines and compression are worth discussing with your clinician rather than changing on your own during a flare. Extra salt changes blood pressure and the work your heart does, so it belongs in a plan agreed with a clinician.
What watch data cannot tell you
It is worth being blunt, because the opposite promise is common in apps and marketing.
No validated crash prediction exists. As of 2026, there is no personal model that tells you on Tuesday that Thursday will go badly. Studies show group-level associations and models that improve same-day estimates (6, 7).
The signals are non-specific. Resting heart rate and HRV move with infections, alcohol, heat, poor sleep, the menstrual cycle, stress and medicines as readily as with overexertion. A shifted value says something is off, never what.
A quiet watch is not a clean bill of health. Some crashes come after days that looked ordinary in every number, especially after cognitive or emotional load.
Readiness and recovery scores were not built for this. They were designed and validated mostly on healthy and athletic people, and can show green in the middle of a crash.
Wrist sensors have limits. Wrist-worn heart rate is good for trends but lags short spikes. A chest strap is more accurate for real-time pacing. Camera readings depend on stillness and light.
More checking is not more safety. If watching numbers raises your anxiety, check less, or track in seasons rather than every day. A tool that makes your days worse is not earning its place.
Before you blame yourself for the crash
The thought after a crash is often "I should have known" or "I must have done something wrong". Usually the crash was set up one to three days earlier, on a day that felt fine or even good, by a load you could not feel at the time: a long queue, an emotional phone call, a warm room, a shorter night, the week before your period. That is not a failure of willpower. It is how PEM works.
What your data can sometimes do is make that delayed cost easier to see. Scroll back two or three days before your last crash and look for the busy day, the short night, the morning when resting heart rate sat above your line. You are looking for a trace, not a culprit. Change one thing, such as a rest after every appointment or a shorter shopping trip, and watch your trend for one to two weeks. You are not broken; there is a reason, and sometimes it shows up in what your body already recorded.
Living with crashes: the questions people ask next
Why did I crash when my numbers looked fine?
A PEM crash can follow days that look normal in every wearable metric. Cognitive, emotional and sensory effort move heart rate and HRV much less than walking does, and a wrist sensor smooths short spikes. Infections and hormonal changes can also lower your capacity without changing your morning numbers much. Normal numbers before a crash do not mean the crash was imagined or self-inflicted; they mean the load was one the sensors could not see. Look at your diary instead: appointments, conversations, standing and noise one to three days back.
Can a PEM crash start without any warning signs?
Yes. Some people describe crashes that arrive with no noticeable early phase, especially after a large effort, and timing varies even within one person (4). Early signs are a pattern many people learn over time, not a guarantee. If your crashes tend to come without warning, pacing by plan matters more than watching for signals: planned rests, time limits and stopping before you feel tired, as described in our guide to the 50% rule.
Why does my HRV look great when I'm crashing?
For some people with ME/CFS, POTS, Long COVID or fibromyalgia, HRV can spike or look unusually calm during a crash. Possible reasons include a blunted heart response to effort and a nervous system that reacts less because it has less left to react with. A high HRV during a crash is not a sign of recovery. Read it together with resting heart rate, sleep and how you feel, and down-weight HRV when everything else disagrees.
Does my menstrual cycle change my crash warning signs?
It can. HRV tends to be lower and resting heart rate slightly higher in the luteal phase, the one to two weeks before a period (11). Many women with energy-limiting conditions also report lower capacity before their period. That means the same numbers can mean different things at different points of the cycle. Note your cycle days next to your baseline, and compare amber days with the same phase in previous months rather than with your whole-month average.
Do beta-blockers or ivabradine hide crash warning signs?
They can blunt them. These medicines lower heart rate and can change HRV, so the usual rise in resting heart rate under load may be smaller or absent. That does not make them a problem; it means your baseline needs to be built while you are on a stable dose, and body signs matter more. If you start, stop or change the dose, treat the next few weeks as a new baseline period, and talk to your prescriber before making changes.
How often should I check my data?
Less often than you might think. Once a morning is enough for resting heart rate and HRV, and a weekly look back at the trend is enough for patterns. Checking every hour mostly adds worry. Many people find it works best to track closely for a few weeks, learn their pattern, then step back and return to closer tracking during changes, such as a new medicine, a flare season or after an infection.
How do I explain an amber day to work or family?
Keep it short and practical. Something like: "My body is showing early signs of a crash. If I rest today, I lose one day. If I push, I can lose a week." Pre-agree with the people closest to you what an amber day means: which plans move, who covers what, and that you will not explain again each time. If they want to understand more, our guide on how to support someone with a chronic illness is written for partners, family and friends.
How to bring this up with your doctor
Bring dates, not conclusions. Two to four weeks of notes: what you did, how you felt one, two and three days later, and when crashes started and ended. The delay is the key feature, and a record is the only way to show it.
Use the term "post-exertional malaise" and describe it as delayed and disproportionate: "ordinary effort makes me worse a day or two later, and it lasts for days".
Show your own trend, not a single number. A chart of morning resting heart rate across a month, with crashes marked, says more than one reading.
Ask what should be ruled out. Thyroid problems, iron deficiency (including ferritin), sleep apnoea, coeliac disease, heart rhythm problems and orthostatic intolerance can look similar or add to the picture.
Ask how to adjust medicines and fluids during a flare, and what changes you should not make on your own.
If exercise is proposed, ask about a stopping rule. Current UK guidance does not recommend fixed incremental exercise programmes for ME/CFS (1).
Seek urgent care for chest pain, fainting, severe shortness of breath, new weakness or numbness on one side, confusion, a high fever, or any symptom that is new or very different from your usual crashes. Those are not PEM warning signs, and they should not wait for a pattern.
Start understanding your body
Learn what affects your energy, stress, sleep, and daily state. Get the app.
How Welltory helps — and what it cannot do
The limit first. Welltory is a general wellness app. It does not diagnose PEM, ME/CFS, Long COVID or POTS, does not predict crashes, and does not tell you what caused one. What it can do is keep the record that a foggy brain cannot, and line it up against your own baseline.
Log the crash. On iOS, Crash Tracker lets you choose the day a crash began, add start and end times or "Still ongoing", severity, symptoms, what came before, what you did afterward, whether it helped, and a note. Logging is free. Crash Tracker is currently offered in English to new iOS users who reported ME/CFS or chronic fatigue, Long COVID or POTS during onboarding (Welltory Help Center).
See what changed from your baseline. With a subscription, Crash Tracker shows Total Active Time, Stress Time and Today's HRV (lnRMSSD) next to your own boundaries. Activity and stress boundaries are built from your previous 30 days and need at least 14 reliable days. A blank space means no reliable comparison, not a normal day. The Help Center says it plainly: the charts cannot predict a crash or identify its cause, so look for patterns across several days rather than treating one shifted value as a warning.
Look one to three days back. Before each logged crash, check sleep, stress minutes, Battery, resting heart rate and HRV. When the app asks "What happened?" during a stress or rest episode, answer by tap, typing or voice. Those tags build My Patterns on iOS: which tags tend to come with stress, which reliably help you recover, and rare events that hit harder than their frequency suggests (Welltory Help Center).
Take it to your doctor. In the web app, open Dashboard, choose a chart and time range, and export it as CSV (Welltory Help Center).
Devices. HRV plus activity works with Apple Watch, Samsung Watch or Google Pixel Watch (series 2, 3 and later), or a chest strap that records RR intervals over Bluetooth 4.0; most other trackers and rings send activity and sleep only (Welltory Help Center).
Where you can work this out with other people. Welltory Energy Lab is a paid, moderated online community for women aged 35 to 65 living with energy-limiting conditions such as ME/CFS, Long COVID, POTS, fibromyalgia and MCAS. Members compare how their crashes start and what their data looked like, and the team helps them read their own data. It is education and peer support, not medical care. You can join Energy Lab here; what the community is, who it is for and what it costs is explained on a separate page.
Patterns are associations to discuss with your clinician, not proof of what caused a crash.
Compare how your crashes start with women who live with them too
Welltory Energy Lab is a paid, moderated online community for women 35 to 65 living with energy-limiting conditions. Peer support and education, not medical care.
How we made it
We based this article on current clinical guidance for ME/CFS (NICE NG206, CDC, the 2015 Institute of Medicine report and the Bateman Horne Center), on a survey of how PEM starts and how long it lasts, and on recent wearable studies in Long COVID and complex chronic illness, including a 2026 study of 4,244 people with daily morning readings. For the product section, we checked every function against the Welltory Help Center on October 21, 2026. Descriptions of early signs reflect themes that people with energy-limiting conditions commonly describe, summarised in general terms; no individual stories or quotes were used. We did not run a new Welltory data analysis for this article.
Discounts for blog readers: up to 36% off
See what affects your energy, stress, sleep, and daily state with Welltory
This article is for educational purposes only and is not medical advice. Post-exertional malaise is recognised clinically, not by an app. No app or wearable, including Welltory, can predict a PEM crash or identify its cause; Welltory is a general wellness product and does not diagnose, predict, monitor or treat ME/CFS, Long COVID, POTS or any other condition. Seek urgent care for chest pain, fainting, severe shortness of breath or symptoms that are new for you. Sources were retrieved on October 21, 2026.
Was this helpful?
Ask AI for a summary of page
Written by Jane Smorodnikova
The founder and CEO of Welltory. A recognized tech leader with two Master's degrees and experience at MIT, she has scaled Welltory to over 17 million users.
Written by Tatsiana Yashyna
Deputy COO at Welltory. With a background in medicine and years of working with health data, she translates research and real physiological signals — sleep, stress, heart rate, and hormones — into clear, evidence-based explanations that help people understand what their bodies are telling them.
References
- National Institute for Health and Care Excellence. Myalgic encephalomyelitis (or encephalopathy)/chronic fatigue syndrome: diagnosis and management. NICE guideline NG206, 2021. https://www.nice.org.uk/guidance/ng206
- Centers for Disease Control and Prevention. ME/CFS: strategies to prevent worsening of symptoms. https://www.cdc.gov/me-cfs/hcp/clinical-care/treating-the-most-disruptive-symptoms-first-and-preventing-worsening-of-symptoms.html
- Institute of Medicine. Beyond Myalgic Encephalomyelitis/Chronic Fatigue Syndrome: Redefining an Illness. National Academies Press, 2015. https://www.ncbi.nlm.nih.gov/books/NBK274235/
- Chu L, Valencia IJ, Garvert DW, Montoya JG. Deconstructing post-exertional malaise in myalgic encephalomyelitis/chronic fatigue syndrome: a patient-centered, cross-sectional survey. PLOS ONE 2018;13(6):e0197811. https://doi.org/10.1371/journal.pone.0197811
- Bateman Horne Center. ME/CFS Crash Survival Guide, chapter 1: what is ME/CFS and understanding a crash. 2023. https://batemanhornecenter.org/what-is-me-cfs-and-understanding-a-crash/
- Aitken A, Sawyer A, Iwasaki A, et al. Digital physiological biomarkers predict within-person symptom changes in complex chronic illness. npj Digital Medicine 2026. https://doi.org/10.1038/s41746-026-02543-3
- Ruijgt TM, Slaghekke A, Ellens A, Janssen KW, Wüst RCI. Wearable heart rate variability monitoring, autonomic dysfunction and post-exertional malaise in long COVID: an observational study. Sports Medicine 2026 (online first). https://doi.org/10.1007/s40279-026-02487-4
- Quer G, Gouda P, Galarnyk M, Topol EJ, Steinhubl SR. Inter- and intraindividual variability in daily resting heart rate and its associations with age, sex, sleep, BMI, and time of year: retrospective, longitudinal cohort study of 92,457 adults. PLOS ONE 2020;15(2):e0227709. https://doi.org/10.1371/journal.pone.0227709
- Radin JM, Vogel JM, Delgado F, et al. Long-term changes in wearable sensor data in people with and without Long Covid. npj Digital Medicine 2024;7(1):246. https://doi.org/10.1038/s41746-024-01238-x
- Nelson MJ, Bahl JS, Buckley JD, Thomson RL, Davison K. Evidence of altered cardiac autonomic regulation in myalgic encephalomyelitis/chronic fatigue syndrome: a systematic review and meta-analysis. Medicine 2019;98(43):e17600. https://doi.org/10.1097/MD.0000000000017600
- Schmalenberger KM, Eisenlohr-Moul TA, Würth L, et al. A systematic review and meta-analysis of within-person changes in cardiac vagal activity across the menstrual cycle: implications for female health and future studies. Journal of Clinical Medicine 2019;8(11):1946. https://doi.org/10.3390/jcm8111946
- Sheldon RS, Grubb BP, Olshansky B, et al. 2015 Heart Rhythm Society expert consensus statement on the diagnosis and treatment of postural tachycardia syndrome, inappropriate sinus tachycardia, and vasovagal syncope. Heart Rhythm 2015;12(6):e41–e63. https://doi.org/10.1016/j.hrthm.2015.03.029
- da Silva ALG, Vieira LP, Dias LS, et al. Impact of long COVID on the heart rate variability at rest and during deep breathing maneuver. Scientific Reports 2023;13:22695. https://doi.org/10.1038/s41598-023-50276-0
