How Many Push-Ups Should You Be Able to Do? What Your Max Reps Actually Predict

There's a test you can run in your living room, in about sixty seconds, that requires no equipment, no subscription and no lab. It costs nothing to repeat every three months, and in one of the more surprising studies of the last decade it was a better predictor of future heart problems than a treadmill test.

It's the push-up test. And the honest answer to "how many should I be able to do" is more interesting than a number, because the number itself is a proxy for something bigger.

TL;DR: In a Harvard-led cohort of 1,104 male firefighters followed for ten years, men who could do more than 40 push-ups had a 96% lower rate of cardiovascular events than men who could do fewer than 10 (incidence rate ratio 0.04, 95% CI 0.01–0.36), after adjusting for age and BMI. Even 11 or more push-ups was associated with meaningfully lower risk. The test was run to a metronome at 40 push-ups per minute. The headline number needs context: the cohort was occupationally active men, there were only 37 events in total, and the confidence interval is very wide. Push-up capacity isn't magic. It's a cheap proxy for total-body fitness, and it correlates well enough with cardiorespiratory fitness that it carries real signal. Use it as a quarterly checkpoint, not a diagnosis.

YouTube: Related video

What the study actually did

The study was published in JAMA Network Open in February 2019 by a team from Harvard's T.H. Chan School of Public Health and Cambridge Health Alliance. It's a retrospective cohort, which means researchers went back through existing medical records rather than recruiting people prospectively.

The records came from 1,562 male career firefighters across 10 Indiana fire departments who had periodic medical exams at a single clinic between 2000 and 2010. Of those, 1,104 had usable push-up data. Mean age was 39.6, mean BMI 28.7. They were followed for up to ten years, and researchers counted new diagnoses of coronary artery disease and other major cardiovascular events, including heart failure and sudden cardiac death.

The test protocol matters, because most people do it differently:

• Push-ups were performed to a metronome set at 80 beats per minute, which works out to 40 push-ups per minute. That's a steady, controlled cadence, not a frantic sprint.
• The count stopped at 80 repetitions, or when the person missed three or more beats, or when they stopped from exhaustion or symptoms like dizziness, chest pain or shortness of breath.
• Participants were then grouped into five categories in increments of 10: fewer than 10, 11 to 20, 21 to 30, 31 to 40, and more than 40.

Over the follow-up period there were 37 cardiovascular events across 8,601 person-years. That's a small number, and it's the single most important thing to hold in your head when you read the results.

What it found

Push-up capacity at baseline was inversely associated with cardiovascular events over the next decade, and the association held after adjusting for age and BMI.

The headline comparison: men who could do more than 40 push-ups had an incidence rate ratio of 0.04 compared with men who could do fewer than 10. That's the "96% lower risk" figure you've seen in headlines, and it's technically what the paper reported.

The more useful finding is buried a paragraph later. Participants who could do 11 or more push-ups already had significantly reduced risk. The relationship was a gradient, not a cliff at 40.

And here's the part that makes the study genuinely interesting: the researchers also ran submaximal treadmill tests on the same men, using a modified Bruce protocol, and the crude incidence rate ratios were similar whether you split the cohort by push-up capacity or by estimated cardiorespiratory fitness. A free bodyweight test was carrying roughly the same information as a lab-grade treadmill protocol.

What the 96% number does not mean

This is where most articles about this study go wrong, so let's be precise.

It's not a 96% reduction in your personal risk. An incidence rate ratio of 0.04 compares event rates between two groups in one cohort over one decade. Your absolute risk depends on your age, blood pressure, lipids, family history, smoking status and a dozen other things.

The confidence interval is enormous. 0.01 to 0.36. That's the range consistent with the data. With 37 events split across five groups, the top category might contain a handful of events. A wide interval means "probably a real and substantial association, precise magnitude unknown."

The cohort was male firefighters. Firefighting is a physically demanding job with fitness standards for employment. These men were occupationally active, mostly middle-aged, and all male. The authors say plainly that larger studies in more diverse cohorts are needed. There's no equivalent published study of this size in women.

It's observational. Men who can do 40 push-ups differ from men who can do 8 in a hundred ways that the study didn't measure: training history, diet, sleep, smoking, income, how recently they saw a doctor. Adjusting for age and BMI doesn't fix that.

So: a real signal, from a specific population, with imprecise numbers. That's still useful. It's just not a diagnosis.

Why push-up capacity tracks heart risk at all

If push-ups were purely an arm exercise, this association would make no sense. Three things explain it.

Push-up capacity is a total-body effort. A strict push-up is a moving plank. Your core, glutes and legs hold the line while your chest, shoulders and triceps move the load. Past about 20 reps, the limiting factor stops being your arms and starts being your ability to keep producing work while breathing hard. You're doing submaximal cardiovascular work with a strength component.

Muscular strength predicts mortality, and it isn't only about the muscle. A prospective cohort of 8,762 men followed for an average of 18.9 years, published in the BMJ in 2008, found that men in the middle and top thirds of combined leg and bench press strength had lower all-cause mortality than the bottom third (hazard ratios 0.72 and 0.77). The pattern held after adjusting for cardiorespiratory fitness. Interesting nuance: the association with cardiovascular death weakened once fitness was accounted for, which suggests strength partly matters because stronger people tend to be fitter people.

Strength work moves the metabolic markers that drive risk. Muscle is glucose disposal tissue. More of it, and more of it trained, tends to mean better insulin sensitivity, better blood pressure regulation and a healthier lipid profile. None of that shows up in a push-up count directly. It shows up in the years of training that produce the count.

Put together: your push-up number is a rough, free readout of how much total-body capacity you're carrying. That's why it correlates with outcomes.

How to run the test properly

Do this once, write the number down, and don't retest for at least a month. A test you run weekly becomes a training session, and you'll start sandbagging or peaking for it.

Setup

1. Warm up for two or three minutes first. Arm circles, a few slow shoulder rotations, ten bodyweight squats, twenty seconds of marching in place. Don't test cold.
2. Use a metronome app at 80 beats per minute. Each rep is two beats: down on one, up on the next. That's the study's cadence and it stops you from bouncing.
3. Get into a plank with hands slightly wider than your shoulders, fingers forward, elbows tracking back at roughly 45 degrees rather than flared to 90.
4. Head neutral, gaze at a spot on the floor about 30 cm ahead of your hands. Squeeze glutes, tuck ribs down so your lower back doesn't sag.

Execution

1. Lower until your chest is about a fist's height off the floor. A rolled-up towel or a water bottle works as a depth target.
2. Press back up until your elbows are straight, not hyperextended.
3. Stop when you miss three beats, or when your hips sag or your head leads the way up. Those are the study's failure criteria in spirit: form breakdown ends the set.

What invalidates a set: hips sagging, half reps, head craning forward, hands drifting toward your head, or resting in the up position for several seconds. Write down only strict reps. An honest 12 beats a flattering 25 every time, because the point is the trend.

What your number means in practice

The study gives us five published categories. Here they are, with an honest interpretation of each.

Two honest warnings about this table. First, "fewer than 10" is not a death sentence. It's a fitness floor, and it's among the most trainable numbers in this entire field. Untrained people routinely double it in six weeks.

Second, there is no clean, validated, age-and-sex-specific norm table for push-ups the way there is for VO2 max. You'll find plenty online, mostly derived from school fitness testing or military standards, and they don't generalize well. If you want a reference point that's actually yours, your own quarterly trend is better than a stranger's chart.

If you're a woman reading this and feeling excluded, that's fair. The study was men only, and I'd rather say so than pretend the numbers transfer. The principle likely does: muscular endurance is trainable, and low capacity is worth fixing. The specific cut-offs are unproven for you.

What your watch can and can't tell you

You can run this test with Apple Watch recording a "Functional Strength Training" workout, and it's worth doing once to see how your heart responds. A set to failure will spike your heart rate into the 140s or higher, even though you're not moving your feet.

Don't read too much into that number. Optical wrist heart rate lags badly during short, high-intensity efforts with wrist flexion, and a 40-second set is exactly the worst case for it. The strength training heart rate explainer covers why your numbers look wrong and when a chest strap is worth it.

What your watch is genuinely good for is the recovery side. A hard set to failure is a real training stress. If your overnight HRV dips and resting heart rate climbs the next morning, that's the expected response and it should resolve within a day or two. If it takes four days, you're either detrained or under-recovered, and either way the answer isn't more push-ups. That's the same HRV down, resting heart rate up pattern that shows up after any hard session.

How to improve your number

Push-ups respond to frequency and consistency more than to heroic single sessions. Three sessions a week, 20 to 30 minutes apart from your other training, is plenty.

Week 1 to 2: build the base. Three sets at 50 to 60% of your max, three times a week, with 90 seconds rest. If you can't do 8 strict floor push-ups, do them on an incline: hands on a bench, a table, or a stair. Incline push-ups are not a lesser exercise. They're how you build the volume to earn floor reps.

Week 3 to 4: add density. Same three sets, but every session add one rep per set, or shave 10 seconds off the rest. Slow the lowering phase to three seconds; it roughly doubles the time under tension without adding reps.

Week 5 to 6: one hard set, then back off. Test your max once. Then take a lighter week. Repeat the cycle.

Two things that quietly cap progress:

• Skipping the core work. Most people fail a push-up test when their hips sag, not when their arms give out. Planks, dead bugs and hollow holds fix the actual limiter.
• Testing too often. Max efforts are expensive. Once every four to six weeks is enough.

If you're starting from zero, this is a strength project with a cardiovascular side effect, and it pairs well with the simplest strength plan that works. The WHO physical activity guidelines recommend muscle-strengthening activity on at least two days a week for adults, alongside 150 to 300 minutes of moderate aerobic activity. Push-ups count toward the strength side. They don't replace the aerobic side.

Pair it with one other free test

One number is a snapshot. Two or three cheap tests, repeated quarterly, become a trend you can actually act on. The best companion to push-ups is the sitting-rising test, which measures the other end of the same idea: how well your body handles getting off the floor, which is strength, mobility and balance in one movement.

Add a cardio recovery reading from a brisk walk or a hard effort, and you've got a three-part picture of capacity that costs nothing and takes fifteen minutes a quarter.

Put all three in the same place, with the same date attached, and let the trend do the talking. That's what Century AI is built for: your recovery score and health score already reflect sleep, HRV and resting heart rate, and a quarterly strength checkpoint sits neatly next to them. A number that goes up over twelve months means more than any single day's reading.

When not to test

Skip the test, and see a clinician instead, if any of these apply:

• current shoulder, elbow or wrist pain, or a history of shoulder dislocation
• recent chest pain, unusual breathlessness, or a heart condition without medical clearance
• uncontrolled blood pressure
• recent abdominal or hernia surgery
• pregnancy, without checking with your provider first

Nothing about this test is so valuable that it's worth provoking an injury for.

The short version

• In a ten-year cohort of 1,104 male firefighters, push-up capacity predicted cardiovascular events, and 11+ reps was already associated with lower risk.
• More than 40 reps vs fewer than 10 gave an incidence rate ratio of 0.04, but that's 37 events, a wide confidence interval, and a very specific population.
• Push-up capacity works as a proxy for total-body fitness, which is why it carries heart signal at all.
• Run the test to a metronome at 40 per minute, to strict form, and count only clean reps.
• There's no reliable age-and-sex norm table. Use the study's five categories as a rough map and your own quarterly trend as the real benchmark.
• Train it three times a week with incline progressions, core work, and slow eccentrics.
• Watch your next-day HRV and resting heart rate, not your heart rate during the set.
• Retest every four to six weeks, not every week.

The appeal here is the same as with any good metric: it's free, it's repeatable, and it tells you something true about your body without needing to be interpreted by a lab. Sixty seconds, once a quarter. That's a low price for a number that predicts something real.

This isn't medical advice. If you have a heart condition, joint problems or you're returning from injury, check with a clinician before testing yourself.

Century AI helps you understand your body with a daily health score, recovery score, and sleep insights — using the watch you already wear.