The warning usually arrives one pop too late: keep doing that and you will get arthritis. It sounds plausible because the noise is sharp, the movement is deliberate, and osteoarthritis is often imagined as the sum of small mechanical insults. But plausibility has been carrying more weight than the evidence.

The best-supported answer is narrower than either family folklore or internet absolution. Ordinary, painless knuckle cracking has not been shown to cause hand osteoarthritis. The sound is associated with a rapid cavitation event inside a synovial joint, not a piece of cartilage snapping. Yet the research base is small, mostly observational, and offers evidence against an arthritis association rather than proof that every technique, frequency, and lifetime of cracking is harmless.

That distinction matters. “No demonstrated arthritis link” is a defensible conclusion. “Nothing could ever go wrong” is not.

The sound was visible before its cause was settled

In April 1947, anatomists J. B. Roston and R. Wheeler Haines published an experiment that treated a cracked knuckle as a mechanical event rather than a bad habit. They applied traction along a finger while measuring force and taking serial radiographs of the metacarpophalangeal joint—the large knuckle where a finger meets the hand. At a critical pull, the crack sounded and the joint surfaces sprang farther apart. A radiograph taken immediately afterward, while traction was maintained, showed a clear space within the joint. The same joint then entered a refractory period: it could not immediately produce another crack and generally needed about 20 minutes before the cycle could be repeated.[1]

Those observations already weaken the folk picture. If the noise were simply two rough surfaces grinding, it should not require a pressure threshold, create a visible gas space, and then temporarily refuse an encore.

In July 1971, Anthony Unsworth, Duncan Dowson, and Verna Wright built a machine to measure the load–separation behavior more closely. Their work placed the event within cavitation: pulling the joint surfaces apart lowers pressure in the synovial fluid until a gas-filled cavity forms. The crack belongs to that sudden transition.[2] “Cavitation” can sound destructive because collapsing cavities damage ship propellers and pumps. A biological joint, however, is not a propeller blade, and sharing a physical word does not establish shared tissue injury. Mechanism can generate a hypothesis; it cannot supply the missing health outcome.

The mechanism was refined again in April 2015. Gregory Kawchuk and colleagues recorded ten finger joints from one adult with cine MRI at 3.2 frames per second. At joint release, a cavity appeared and remained visible afterward. That sequence led the authors to favor cavity formation, or inception, rather than the previously popular story that the pop came from a pre-existing bubble collapsing.[5]

It did not end the acoustic argument. In March 2018, V. Chandran Suja and A. I. Barakat published a mathematical model in which an idealized 200-micrometre bubble partially collapsed. The simulated sound's magnitude and dominant frequency resembled recorded cracks, while a smaller bubble persisted—showing how collapse could remain plausible despite the post-crack cavity on MRI.[6] This is a model, not a faster view of the same cavity inside the joint. Taken together, the studies securely connect the pop to cavitation but leave the sound's split-second source—cavity inception, partial collapse, or some combination—less settled than the 2015 image alone suggested. Simultaneous internal imaging and acoustic recording fast enough to resolve that instant, repeated across more than one person, would distinguish the alternatives more directly.

This is the first myth to discard: a loud crack is not an acoustic report of cartilage breaking. The images show a space forming as joint surfaces separate, and the competing mechanism accounts both concern the behavior of that cavity. Neither shows osteoarthritis beginning.

The arthritis question produced a less theatrical record

The outcome evidence arrived through modest studies, not one decisive trial. A randomized experiment assigning people to crack one set of joints for decades would be neither practical nor especially ethical. Researchers instead used two kinds of snapshot: one study paired recalled habits with a clinical hand examination; another paired recall with existing radiographs. These observational designs can detect an association, but they must contend with imperfect memory, self-selection, and other differences between people who crack and people who do not.

In May 1990, Jorge Castellanos and David Axelrod evaluated 300 consecutive patients aged 45 or older: 74 habitual knuckle crackers and 226 non-crackers. They found no greater prevalence of hand arthritis in the cracking group. They did, however, report more hand swelling and lower grip strength among habitual crackers, and the habit clustered with manual work, nail biting, smoking, and alcohol use.[3]

That paper therefore split the story in two. Its arthritis result argued against the familiar warning. Its functional findings left open a different concern. But the study was cross-sectional: exposure and outcome were assessed at the same encounter, while several potentially relevant behaviors traveled together. It could show an association with swelling and grip; it could not establish that cracking caused either one.

A 2011 retrospective case-control study aimed more directly at osteoarthritis. Kevin Deweber and colleagues identified 329 people aged 50 to 89 who had undergone a right-hand radiograph; 215 returned the survey, a 65% response rate. Among those respondents, 135 had radiographic hand osteoarthritis and 80 did not. Habitual cracking was reported by 18.0% of the osteoarthritis group and 23.2% of controls, a difference that was not statistically significant (P = .361). For the distal and proximal finger joints and the large knuckles, neither years of cracking nor “crack-years,” a cumulative measure combining frequency and duration, was associated with osteoarthritis. There were too few data to make the same assessment at the thumb's first carpometacarpal joint.[4]

This is stronger than asking only whether someone can remember ever cracking a knuckle. The investigators looked for a dose pattern: if repeated cracking were a meaningful cause, heavier lifetime exposure should tend to leave more radiographic disease. The study did not detect either the association or the dose response in the joints with enough data. Still, it relied on recalled habit, examined one hand radiographically, and began with a selected clinical population in which one third of those contacted did not respond. It does not measure every subtle tissue change over a lifetime.

The 1990 functional warning did not reproduce cleanly

In a prospective study published online in January 2017, Robert Boutin and colleagues studied 400 metacarpophalangeal joints in 40 asymptomatic adults. Thirty participants reported daily habitual cracking. Two orthopaedic surgeons performed the examinations without knowing the participants' cracking histories or sonographic outcomes. Two musculoskeletal radiologists, unaware of the cracking histories and whether a crack had been heard, interpreted the ultrasound images.[7]

The investigators found no swelling after cracking and no meaningful group differences in disability scores, grip strength, or laxity. Joints that audibly cracked gained an average 7.7 degrees of passive total range of motion immediately afterward (95% confidence interval 3.8 to 11.7 degrees). In a separate comparison, total range of motion was 9.0 degrees greater among habitual crackers than non-crackers (95% confidence interval 2.9 to 15.1 degrees). The study was designed to observe immediate examination and ultrasound findings—not to settle arthritis risk decades later.[7]

These results do not erase the 1990 report. They show why its strongest causal reading is unstable. One study associated habitual cracking with swelling and weaker grip in an older clinical sample; a later, smaller study of asymptomatic adults, using blinded examinations and dynamic imaging, did not reproduce those abnormalities. Differences in age, selection, definitions, and timing could explain the mismatch. A large prospective cohort that recorded cracking behavior before symptoms began, then followed standardized radiographs and hand function for years, would do more to resolve it than another snapshot.

What the myth gets wrong—and what the evidence cannot promise

The strongest interpretation is that habitual knuckle cracking is not an important cause of hand osteoarthritis. Across the available outcome studies, the predicted excess arthritis does not appear; in the 2011 work, greater reported exposure did not reveal a hidden dose response.[3][4] The mechanism studies also replace the image of audible cartilage damage with a pressure-driven cavitation event.[1][2][5][6]

The strongest caution is methodological. The arthritis studies were not large longitudinal cohorts, cracking exposure was self-reported, and the 2017 ultrasound study examined immediate findings in only 40 asymptomatic adults.[4][7] Rare injuries from forceful manipulation, effects in already damaged joints, and small long-term functional changes are different questions. The literature reviewed here does not grant them a universal safety certificate.

It also does not make every joint noise equivalent. The research concerns voluntary distraction of finger knuckles that produces a discrete crack. Repeated clicking during motion can arise from other structures, while pain, persistent swelling, warmth, stiffness, or loss of function changes the clinical question. NHS guidance treats those symptoms—not sound alone—as reasons to seek assessment, urgently when a joint is hot and swollen or illness accompanies it.[8]

So the useful boundary is simple. A painless knuckle pop is poor evidence of arthritis and has not been shown to create it. A joint that hurts, swells, feels hot or stiff, or becomes difficult to use—especially after an injury—should not be waved away as “just cracking.” In that setting, the sound is no longer the diagnosis; it is one detail in a symptomatic joint.[8]

The myth survived because it joined a memorable noise to a familiar disease and supplied an intuitively tidy mechanism. More than seventy years of increasingly direct observation have untied those pieces. A cavity can now be watched as it appears during the pop, even as the exact acoustic instant remains debated. The predicted arthritis signal has repeatedly failed to materialize. What remains is not a license for aggressive manipulation, but a better-calibrated conclusion: sound is vivid; damage requires evidence.

Sources

1. J. B. Roston and R. Wheeler Haines, “Cracking in the Metacarpo-Phalangeal Joint,” Journal of Anatomy 81 (April 1947) — early traction-and-radiography study of joint separation, the clear space, and the refractory period.
2. Anthony Unsworth, Duncan Dowson, and Verna Wright, “‘Cracking Joints’: A Bioengineering Study of Cavitation in the Metacarpophalangeal Joint,” Annals of the Rheumatic Diseases 30 (July 1971) — load–separation experiments and the cavitation model.
3. Jorge Castellanos and David Axelrod, “Effect of Habitual Knuckle Cracking on Hand Function,” Annals of the Rheumatic Diseases 49 (May 1990) — 300-patient cross-sectional study of arthritis, swelling, grip, and correlated habits.
4. Kevin Deweber, Mariusz Olszewski, and Rebecca Ortolano, “Knuckle Cracking and Hand Osteoarthritis,” Journal of the American Board of Family Medicine 24 (March–April 2011) — radiographic case-control study of 215 survey respondents and cumulative cracking exposure.
5. Gregory N. Kawchuk et al., “Real-Time Visualization of Joint Cavitation,” PLOS ONE 10 (15 April 2015) — cine-MRI evidence linking the crack to cavity inception rather than collapse.
6. V. Chandran Suja and A. I. Barakat, “A Mathematical Model for the Sounds Produced by Knuckle Cracking,” Scientific Reports 8 (29 March 2018) — partial-collapse model and comparison with recorded acoustic signatures.
7. Robert D. Boutin et al., “‘Knuckle Cracking’: Can Blinded Observers Detect Changes with Physical Examination and Sonography?” Clinical Orthopaedics and Related Research 475 (2017) — blinded examination and dynamic ultrasound of 400 joints in 40 adults.
8. NHS, “Joint Pain” (reviewed 26 February 2026) — symptom and escalation boundaries for painful, swollen, hot, stiff, or impaired joints.
9. Jaysin Trevino, “Cracking Knuckles” (12 June 2010), via Wikimedia Commons — provenance page for the article's documentary photograph.