REF-1489 · Knapik, 2022 · Journal of Special Operations Medicine · Paper · PubMed

How often do people get hurt in functional training?

Across 28 studies, 36% reported an injury and the rate was 4.3 per 1000 h, but 9.9 in the 5 prospective cohorts; single studies varied widely. Says nothing on what causes it.

Main caveat: Abstract only. Definitions, quality and heterogeneity are unknown; single author, without a second reviewer noted in the abstract.
moderate evidenceMeasured on recreational to highly trained athletes

Knapik reviewed studies that gave quantitative injury prevalence or rates in high-intensity functional training in people aged 18 or over. Twenty-eight studies with 11,089 participants qualified. Overall injury prevalence was 36% and the rate 4.3 per 1000 h, but the five prospective cohorts gave 9.9 per 1000 h, with a wide interval.

Most commonly injured sites were the shoulder and back/spine (26% each), knee (14%), wrist, hand and fingers (12%), arm and elbow (10%), and ankle and foot (6%). Because prospective rates ran higher, the author says the true rate cannot yet be appropriately quantified.

Our reading: the honest figure is a range; the prospective estimate sits about twofold above the overall one (9.9 vs 4.3, our arithmetic), with an interval (3.3-16.4) that includes 4.3 and no test of the difference, so a per-1000-h number should never be quoted without saying how it was collected.

Nothing here separates novices, loads or programmes, and no comparison with another activity is given.

Protocol

Selection: individuals 18 years or older; quantitative data on injury prevalence and/or rates during high-intensity functional training.
Studies and participants: 28 studies; n = 11089 participants.
Design mix: five prospective cohort studies; the abstract does not give the designs of the others.
Not given in the abstract: databases and dates, the injury definition(s), quality assessment, heterogeneity statistics, the share of military participants.
Funding and interests: not available from the abstract; single author.

Key Results

MeasurePooled estimate (95% CI)Range of single studies
Injury prevalence36% (32-41%)12% to 74%
Injury rate, all studies4.3 per 1000 h (3.35-5.23)0.04 to 18.90 per 1000 h
Injury rate, five prospective cohorts9.9 per 1000 h (3.3-16.4)Not given
Site% of total injuries
Shoulder26%
Back/spine26%
Knee14%
Wrist, hands, fingers12%
Arm, elbow10%
Ankle, foot6%

Mechanisms

Measured: frequency and site only. No mechanism, load or movement is linked to injury in the abstract.
Authors' interpretation: the higher prospective rates likely reflect better tracking of injuries over time.
Inference: retrospective recall loses minor injuries, so the pooled 4.3 is more likely too low than too high; the prospective CI is wide enough that this is not settled.

Practical Application

How often do people get hurt in functional training? Across 28 studies, 36% reported an injury and the rate was 4.3 per 1000 h, but 9.9 in the 5 prospective cohorts; single studies varied widely. Says nothing on what causes it.
Reading a number: give the rate with its design: 4.3 per 1000 h overall, 9.9 prospective. Do not set either against another sport without a like-for-like definition; this abstract gives no comparison.
Where to look first: shoulder and back/spine carried half of the injuries in this set. In HYROX athletes followed weekly, the knee and lower leg carried the most burden (REF-904, B), so this site list does not transfer as it stands.
What it does not establish: that functional training is more or less dangerous than other training, that any progression, load jump or rep scheme is safer, or that any site follows from a movement.

Limitations

Abstract only. Definitions, quality and heterogeneity are unknown; single author, without a second reviewer noted in the abstract.
Very wide range of studies. A spread of 0.04 to 18.90 per 1000 h makes one pooled number weak on its own.
Definitions vary across this literature (REF-1490 states it for CrossFit); a "time-loss" and a "any problem" count are not the same rate.
No comparison reported. The abstract gives prevalence and rates only and links no exposure or intervention to injury.
Interests. Not available from the abstract.
How this entry was checked and changed
Why this level

a meta-analysis, but of prevalence and rate, not of a controlled comparison, with a very wide spread between studies; read from the abstract only (OpenAlex marks it closed, no repository copy), so heterogeneity statistics, study quality and definitions are not available