Heavy Strength Training and Knee OA: What START Found

Heavy strength training did not worsen knee joint compressive forces during walking in a major 18-month trial of adults with knee osteoarthritis. That finding weakens a common reason to avoid lifting: the belief that heavier resistance work will automatically accelerate joint damage. It does not prove that heavy training repairs cartilage, prevents progression, or beats lighter training for pain.

The distinction matters. In the START randomized clinical trial, high-intensity and low-intensity strength training produced pain outcomes that were statistically and clinically similar. High intensity also caused more nonserious adverse events. For many people with knee OA, the useful message is simpler and better supported: strength training belongs on the table; the right intensity depends on tolerance and goals.

Why clinicians worried about load

Knee osteoarthritis involves pain, stiffness, and impaired movement, but pain does not provide a direct readout of cartilage damage. Clinicians have often worried that heavy resistance exercise raises tibiofemoral compressive force: the force pressing the thighbone and shinbone together at the knee. The concern had a reasonable biomechanical basis. A 2023 prospective cohort study by Amiri et al. found that higher tibiofemoral contact and muscle forces during gait were associated with radiographic knee OA progression over three years.

Association during walking, however, does not establish that supervised resistance training causes progression. Walking and a squat, leg press, or knee-extension exercise have different joint positions, movement patterns, and muscle activation. START tested whether an 18-month training program changed the knee loads participants used during everyday walking.

What the START trial tested

Researchers randomly assigned 377 community-dwelling adults with symptomatic, radiographically confirmed knee OA to high-intensity strength training (127 people), low-intensity strength training (126), or an attention-control program (124). Participants had a mean age of 65 years; 40% were women, and BMI ranged from 20 to 45. The study followed them for 18 months, and 320 participants (85%) completed it.

The trial used two primary outcomes: WOMAC pain, a patient-reported knee OA pain scale, and maximal tibiofemoral contact force along the tibia during walking. This design matters because it tested both the outcome people feel and the proposed harm pathway during daily ambulation.

Heavy training did not beat lighter training for pain

At 18 months, the high-intensity group had a mean WOMAC pain score of 5.1, compared with 4.9 in the attention-control group. The adjusted difference was 0.2 points (95% CI, −0.6 to 1.1; P=0.61). Against low-intensity training, high intensity scored 5.1 versus 4.4, an adjusted difference of 0.7 points (95% CI, −0.1 to 1.6; P=0.08).

Neither comparison reached statistical significance, and both sat below the trial's prespecified minimum clinically important difference of 2 WOMAC points. The numerical result favored low-intensity training over high intensity, not the reverse, but the difference was too small and imprecise to establish a meaningful advantage. If pain relief is the sole goal, START gives no reason to assume more load will deliver more relief.

Walking knee forces stayed similar

High-intensity training also did not increase maximal tibiofemoral compressive force during walking. The high-intensity group averaged 2,453 N, compared with 2,512 N in the attention-control group: an adjusted difference of −58 N (95% CI, −282 to 165; P=0.61). Compared with low-intensity training, which averaged 2,475 N, the adjusted difference was −21 N (95% CI, −235 to 193; P=0.85).

Those estimates do not show higher walking loads after heavy training. They also do not establish that the small differences are clinically meaningful, because the investigators reported that the minimum clinically important difference for this force endpoint was unknown. The clean conclusion is narrow: in this population and program, high-intensity training did not measurably worsen compressive forces during gait.

The price of higher intensity

Safety was not identical across the training arms. The high-intensity group reported 53 nonserious adverse events, compared with 30 in the low-intensity group and four in the control group. That is a 77% higher burden than low-intensity training. Thirteen serious adverse events occurred across the study—five in high intensity, three in low intensity, and five in control—and adjudicators considered none related to the intervention.

The serious-event finding is reassuring, but it does not erase the difference in day-to-day tolerability. The available summary does not specify the types of the 53 high-intensity events, so readers should not guess whether they were muscle soreness, joint flares, or something else. For an older adult managing a painful knee, a program that delivers similar pain results with fewer minor problems has a practical advantage.

What this trial cannot answer

START measured compressive force while participants walked, not while they performed heavy resistance exercises. That is its central limitation. The trial challenges the idea that heavy training makes habitual gait mechanics more damaging, but it did not directly measure knee forces during a heavy leg press, squat, or knee extension. It therefore cannot settle whether repeated exercise-specific loading affects cartilage over decades.

The gait concern itself remains biologically plausible. Amiri et al. linked higher gait forces with radiographic progression, although that observational study cannot prove those forces caused progression. Research in post-traumatic OA also connects altered loading after ACL reconstruction with later structural outcomes. A 2025 systematic review examined that question after traumatic knee injury, while studies of hopping and patellofemoral loading after ACL reconstruction describe a younger, surgically distinct population. Those findings should not be treated as direct evidence about the older adults with primary knee OA enrolled in START.

Other research supports exercise generally, not a heavy-loading prescription. A 2025 randomized trial reported improved pain, stiffness, and physical function after 12 strength-training sessions over roughly four weeks in older adults with knee OA. Its short duration cannot confirm START's 18-month safety picture, and it did not compare intensities. A 2024 meta-analysis of resistance circuit training in older adults also found improvements in body composition, strength, endurance, blood pressure, and function, but it was not specific to knee OA or to intensity comparisons.

Low-load blood-flow-restriction training is worth watching because it aims to create a strong muscle stimulus with less external load. But the 2023 BFR knee OA paper is a trial protocol, not a results paper. It cannot yet tell patients whether BFR matches conventional training for pain, function, or safety.

How to use this evidence

Fear of loading should not keep someone with knee OA sedentary. A protocol for the EPIPHA-KNEE trial notes that nine out of ten people with knee OA are inactive, often because they believe activity will damage the joint. START provides direct randomized evidence against a blanket ban on heavier strength training in adults like those studied.

Heavy strength training was not superior for pain, and it did not raise walking knee compressive forces over 18 months. Low-intensity training delivered comparable pain outcomes with fewer nonserious adverse events.

If you enjoy heavier lifting, progress gradually, tolerate it well, and train under appropriate supervision, this trial does not support the claim that you are inevitably grinding your knees into dust. If heavy sessions aggravate symptoms or repeatedly produce setbacks, lower-intensity resistance training remains an evidence-based option with comparable pain outcomes in START. Choose intensity for comfort, function, muscle and metabolic goals, and adherence—not because heavier loads have shown superior OA pain relief.