The window between two big efforts is the most underrated skill in ultra running. Anyone can suffer through one race. The runners I respect most are the ones who finish Saturday, sleep four hours in a strange bed, and toe the line on Sunday with a knee that still tracks straight. That is not toughness. That is triage, and triage is a skill you can learn.
I have been on both sides of this. I spent years as a physiotherapist in Bulgaria, and the questions I asked patients on a clinic plinth are the same questions I now ask myself at 11 p.m. in a damp tent with a sore patella and a 5 a.m. alarm. The frame is identical. Only the consequences are different — if I get the call wrong, I do not just lose a race. I lose months. So here is the protocol I actually use, in the order I use it, when knee pain shows up between back-to-back ultras.
First: Is This Mechanical or Load-Related?
Before any treatment, you need a diagnosis — even a rough one. Knee pain mid-stage event falls into two big buckets, and they ask for very different responses.
Mechanical pain is the kind that has a structural cause: a twisted step on a root, a rock you rolled off, a stumble where you felt the joint shift in a way it should not have. It often has a precise moment. You can usually point to it within ten minutes. The pain has a specific location, sometimes a click, sometimes a sense that the knee is "not sitting right." It can hurt at rest, not just under load. Effusion (swelling around or behind the kneecap) often appears within two to six hours.
Load-related pain is the slow burn. No single event triggered it. It built across hours. It is usually diffuse — under the kneecap, around the medial fat pad, along the IT band insertion, or at the patellar tendon. It eases when you sit down and gets worse on the first few stairs after a break. There is no swelling, or only a vague puffiness.
The reason this distinction matters: load-related pain almost always responds to the protocol I am about to describe. Mechanical pain may not, and pushing into it is how meniscus irritation becomes a meniscus tear and a mild MCL strain becomes a six-week problem.
Quick triage at the finish
Sit down. Take the sock off. Look at the knee from the front and from the side. Compare it to the other one. Then run through this checklist:
- Visible swelling or warmth compared to the other knee? Flag.
- Pain at rest, lying flat with the leg straight? Flag.
- Can you fully straighten the knee with the heel propped on a small pack? If you cannot reach full extension passively, that is a flag — and a big one. I have written before about why restoring full knee extension is the single most important milestone after knee injury. The reason is the same here: a knee that will not extend is a knee that is guarding something.
- A single point of pain you can cover with one fingertip? Flag.
- Pain on a single-leg quarter squat, slow and controlled? If yes, what kind — sharp catching pain (flag) or familiar burning fatigue pain (usually safe)?
Two or more flags, and I treat it as mechanical until proven otherwise. Zero or one flag with diffuse soreness, and I run the load-protocol below.
The 12-24 Hour Protocol for Load-Related Knee Pain
This is what I do, in order, when the pain reads as load-related and I want to be on the start line tomorrow with a knee that can carry me.
Hour 0-2: Decompress, do not freeze
The old advice was ice, ice, ice. That has shifted. Heavy icing in the immediate post-effort window can slow the very inflammatory signaling that drives repair. I use brief cold — five to ten minutes of a wet cold towel or a stream crossing — only if there is real heat or sharp irritation. Then I stop.
What I do instead:
- Walk for ten minutes, easy and flat, in soft shoes. Not because it heals anything, but because it pushes lymphatic fluid out of the leg and stops the joint from stiffening into a position it does not want.
- Get the legs above the heart for fifteen minutes. Lying on the floor with legs up the wall is enough. This is the single best move for reducing the swollen, heavy-leg feeling that fools you into thinking the knee is more injured than it is.
- Compression on the calves, not the knee itself. Knee compression overnight tends to interfere with circulation and sleep. Calf compression supports venous return and helps reduce overall lower-leg congestion.
Hour 2-6: Eat for inflammation, not against it
This part gets ideological online. I keep it boring:
- Protein first. Aim for 0.4 g per kg body weight in the first meal post-finish, then the same again before bed. For me at 56 kg that is roughly 22 g per meal. Yogurt, eggs, fish, lentils — whatever you can stomach.
- Carbohydrates in real amounts. Glycogen-depleted tissue heals worse. This is not the night to skip rice.
- Fluids with electrolytes, slowly, not chugged. A litre over three hours is better than a litre in twenty minutes.
- Polyphenols help, mostly because they help sleep. Tart cherry juice has decent evidence for reducing muscle soreness. I am not religious about it, but I take it when I have it.
- Skip the NSAID reflex. Ibuprofen and similar drugs feel like a shortcut. They mask pain that is giving you real information, they thin the protective gut lining you are about to ask to absorb gels for another twelve hours, and there is a case to be made that they impair connective tissue adaptation. If pain is so bad you are considering NSAIDs to sleep, that is itself a flag — see the pull-out section below.
Hour 6-10: Mobility that calms, not mobility that drains
The mistake here is overdoing it. A 40-minute foam-rolling session the night before a race leaves you feeling worse, not better. The goal of evening mobility between ultras is not to "fix" anything. It is to give the knee a clear message that it is safe.
I do, in this order, about fifteen minutes total:
No deep tissue work on the IT band. No aggressive trigger-point ball under the quad. Both feel productive and both leave you walking to the start line with a leg that does not know what it is supposed to do.
Hour 10-14: Isometrics, the most underused tool in the ultra runner's kit
This is the part I want every back-to-back runner to know about. Isometrics — holding a muscle contraction without movement — have two effects that are perfect for the night between races. They produce a real analgesic effect that lasts thirty to forty-five minutes, sometimes more. And they restore neural drive to a quad that has gone quiet from inhibition.
A quad that has gone quiet is the biggest cause of "my knee feels weak and floppy on the descents" on day two. The quad has not lost strength overnight. It has lost the willingness to fire because the knee is mildly irritated. Isometrics wake it back up without adding fatigue.
My evening protocol, about ten minutes:
- Wall sit at 60 degrees, holding for 45 seconds, five rounds, with 30 seconds rest. Do not go deeper. Deeper hurts more knees than it helps. If 60 degrees is painful, try a higher angle.
- Spanish squat if you have a band, otherwise wall sit. Same dose.
- Straight-leg raises, 10 slow reps each side, focusing on locking the quad fully before lifting. If you cannot lock the quad fully, that is the most useful drill of the night — keep practicing the lock without lifting.
The pain should be a 3 or 4 out of 10 during the hold, no more, and it should ease in the thirty seconds after each round. If it climbs into 6 or 7 territory or sharpens, stop. That is the knee telling you this is not load-related after all.
This same principle is something I lean on heavily in my patellar tendonitis rehab progression with runners, and the dose mid-race is essentially the same as the early-rehab dose. It works because tendon and patellofemoral pain both respond to isometric load in a way that is almost stubbornly consistent.
Hour 14-22: Sleep is the leverage point nobody respects
Sleep does more for joint inflammation than every supplement, sock, brace, and kinesio-tape job combined. It is also the thing ultra runners are worst at protecting between races.
What I do:
- Cool the room to around 17-18°C if I can control it. Cool rooms produce deeper sleep, and deeper sleep is where growth hormone and the bulk of tissue repair happens.
- Pillow between the knees when side sleeping. This keeps the upper knee from internally rotating and dragging on the irritated tissue. Sounds small, makes a real difference.
- No screens for the last 45 minutes. Not for the blue light reason, for the cortisol reason — the mental loop of checking splits and reading the weather forecast is what keeps your nervous system in race mode at midnight.
- Caffeine cutoff at 2 p.m. the day before, even if it means a sleepy late afternoon. The half-life of caffeine is roughly five to six hours, and you cannot afford it pulling at your sleep architecture.
If I have done my sleep job, the knee that read as a 5/10 at bedtime is often a 2/10 at wake-up. Inflammation drains overnight in ways nothing during the day can match.
The Morning of Day Two
You wake up, you swing your legs out of bed, you take ten steps to the bathroom. This is the most honest assessment moment you will get. Trust it.
- Pain under 3/10 with no swelling, knee tracks normally — race.
- Pain 3-5/10 that warms up within the first 15 minutes of an easy walk — race, with a conservative pacing plan and a clear pull-out trigger.
- Pain that gets worse, not better, in the first 15 minutes — do not race.
- Pain plus visible swelling, or any sense the knee is unstable — do not race.
Before the start, ten minutes of progressive loading: walk, then walk-jog, then an easy jog. Climb a curb. Step down off it slowly on the affected leg. If you can do a single-leg quarter squat without sharp pain, you have enough quad control to start.
When to Pull Out — A Framework
I have pulled out of races. Not as many as I should have, honestly. The runners I see go wrong are the ones who treat the decision as a binary on the morning of the race. It is not. It is a set of pre-committed triggers you decide on the night before, when your judgment is still clean.
My triggers, written in my notes app the night before:
Any one of these, and I am out. Not because one race matters that much, but because the back half of an ultra on a compromised knee is how a five-day problem becomes a five-month problem. I have learned this watching too many friends limp through a finish line in May and miss their A-race in September.
The Honest Anecdote
Last spring at a stage race in the Rhodope mountains, I finished day one feeling fine and woke up day two with a knee that would not fully extend. About 5 degrees short. No swelling, no specific moment of injury, but the quad would not switch on cleanly. I gave it the full morning protocol — heel slides, isometrics, easy walking — and the extension came back within forty minutes. I started, ran conservatively for the first two hours, and the knee settled completely by hour three.
If the extension had not come back, I would have been out. That was my pre-committed trigger and I had told Deso before bed what it was. Having someone you trust hold the line for you is more useful than any brace.
The Short Version
The protocol, in one paragraph: decompress with walking and elevation, eat protein and carbs, skip NSAIDs, do fifteen minutes of calm mobility and ten minutes of isometric holds in the evening, sleep cold and dark, then trust the first ten steps of day two more than anything your race-brain tells you. If the knee reads mechanical, not load-related, the protocol does not apply and you make the harder choice. Most of the time the harder choice is the right one. The race will run again next year. The knee you race it on is the only one you have.