Injury prevalence among Brazilian Jiu-Jitsu practitioners globally: a cross-sectional study in 881 participants

Abstract

Objectives

Brazilian Jiu-Jitsu (BJJ) is a rapidly growing combat sport characterised by ground fighting and submission techniques. There are limited data regarding BJJ injuries. This study described the incidence and characteristics of injuries in BJJ practitioners.

Methods

This retrospective survey targeted BJJ practitioners globally who trained at least once per week. The survey was distributed online and via posters at multiple Dutch BJJ competitions. Data collection took place from 15 February 2024 to 17 March 2024. We collected training, competition and injury data from the past 12 months. Complaints that required medical attention caused time loss (≥1 week) or led to training modification (≥2 weeks) were defined as an injury.

Results

Of 881 participants, 817 (90%) were male, and the average age was 30.8 years (SD 8.6). Injury information was provided on a total of 888 injuries. The injury incidence was 5.5 (95% CI 4.9 to 6.1) per 1000 hours in training and 55.9 (95% CI 38.8 to 73.0) injuries per 1000 matches, with a higher rate observed among participants with a lower belt level. Most injuries (n=789, 89%) occurred during training, mainly (n=620, 79%) during sparring. The knees (n=223, 25%) and shoulders (n=114, 13%) were the most affected body regions.

Conclusions

This study found an injury incidence of 5.5 per 1000 hours of training and 55.9 per 1000 matches for BJJ practitioners. Future preventative practices should focus on the free sparring part of training to be most effective. Future studies should also monitor BJJ practitioners prospectively to understand injury aetiology better.

WHAT IS ALREADY KNOWN ON THIS TOPIC.

WHAT THIS STUDY ADDS

• This study found that injury incidence in training is 5.5 (95% CI 4.9 to 6.1) per 1000 hours of training and 55.9 (95% CI 38.8 to 73.0) per 1000 matches in competition for BJJ practitioners. Most training injuries happened during free sparring; the knee and shoulder were the most frequently affected body regions.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Based on the data from this study, future preventative efforts should focus on the free sparring part of training since most injuries happen there. Also, more in-depth investigations should be conducted on the mechanisms behind knee and shoulder injuries, ideally in a prospective setting.

Introduction

Brazilian Jiu-Jitsu (BJJ) is a combat sport that gained immense popularity in the last decade due to its use in mixed martial arts competitions like the Ultimate Fighter Championship.^{1 2} While there are no precise data regarding participation, BJJ is commonly seen as one of the most popular combat sports forms globally. The number of tournaments hosted by the International Brazilian Jiu Jitsu Federation has increased from 3 in 1996 to 128 in 2017.¹ Besides, BJJ is not only a combat sport practised by competitive fighters but also a recreational sport practised by millions of people globally.³

BJJ is a grappling sport focused on ground fighting, chokeholds and joint locks to score points or submit one’s opponent.⁴ BJJ can be practised while wearing a kimono (Gi) or tight-fitting, synthetic clothing (No-Gi).⁵ Typical BJJ classes consist of a warm-up, specific technique instruction and a sparring session in which practitioners can be matched on skill and weight.¹ However, academies vary in class structure, duration and prerequisites.¹ Similarly, there are many dissimilarities between BJJ competitions hosted by various federations.⁶ Generally, points can be scored by taking down one’s opponent or successfully implementing chokeholds or locks on the ground.⁷ An athlete wins by one’s opponent submitting or by scoring more points at the end of a 5–10 min match.⁷

Since BJJ’s popularity is relatively new and the technical and physiological demands of the sport are unique, epidemiological data on injuries are scarce and non-conclusive.^{7 8} Previous studies show that injury rates in BJJ range from 9.2 per 1000 hours in state-level competitions to 24.9 per 1000 matches in international competitions.^{9 10} For training and competition injuries combined, Hinz et al¹¹ report an injury rate of 308 injuries per 1000 athletes per year and an injury prevalence of 79%.¹¹ Instead, McDonald et al¹² report an injury point-prevalence of 86% during BJJ training. As for the injury definitions that were used, the studies that focused on competition injuries included all injuries for which help was sought from a medical professional at the tournament. For injuries outside of competitions, Hinz included only injuries that resulted in a minimum time loss of 2 weeks, while McDonald only included injuries that required medical attention or resulted in time loss.^{11 12} The knee, chest and shoulders were the most frequently injured body regions in BJJ athletes.^{1 11}

Few studies have investigated competition and training injuries in BJJ athletes, and the studies that have been done are difficult to compare due to methodological variations. Establishing injury incidences and characteristics and relating these to exposure time is crucial to a better understanding of BJJ injuries and provides the basis for future preventative interventions. Therefore, the current study aimed to describe the incidence of injuries in BJJ athletes during training and competition and the characteristics of BJJ injuries, including body region, onset, timing and mechanism.

Methods

Study design

We informed all participants about confidentiality and anonymous data management, and each participant provided electronic informed consent. We adhered to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines for reporting cross-sectional studies.¹³

Participants

Participants aged 16 years or older who practised BJJ at least once weekly were eligible for inclusion. Commencing on 15 February 2024, we distributed the survey through social media platforms like Instagram and Reddit, at various national competitions (AJP Tour Amsterdam (16 February Almere, The Netherlands), vs BJJ 22 (17 February, Eindhoven, The Netherlands), Kaizen League Open (24 February, Zwolle, The Netherlands), Grappling Industries (25 February, Amsterdam, The Netherlands) and ADCC Holland (7 March, Amsterdam, The Netherlands) and at two BJJ training venues, namely Amsterdam Training Center and BJJ Academy Amsterdam), and by actively engaging with BJJ schools across the Netherlands. Participants from all countries were eligible for inclusion, and the survey closed on 17 March 2024.

Survey

For the complete survey, see online supplemental materials 1. To facilitate broad accessibility, we distributed the study through Microsoft Forms¹⁴ in English and Dutch. We based the survey on previously used instruments in emerging sports and changed the content to make it relevant to the BJJ context.^{15 16} The survey consisted of three parts: the first part concerned demographic data, the second part concerned data regarding BJJ participation and the third part concerned injury data for the previous year.

Injuries

We asked the participants to register injuries that occurred in the past twelve months. A physical complaint was considered an injury if (1) BJJ practice and competition had to be ceased for a week, (2) BJJ practice or competition had to be modified for 2 weeks or (3) medical attention was sought. Participants registered if the reported injury was due to BJJ training or BJJ competition and further specified the related action. They also registered time loss in days and anatomical injury sites and were asked to provide detailed information per injury. Participants provided detailed information for at least one and maximally three injuries. For injuries without detailed information, we only recorded the number of injuries in competition or training. Given this survey structure, the total number of recorded injuries we used for the incidence calculations differs from the number of injuries we recorded detailed information on, such as onset, time loss and affected body region. We adhered to the International Olympic Committee consensus statement for reporting injury characteristics.¹⁷

Statistical analysis

We used R in R Studio for all data processing and statistical analysis.¹⁸ We summarised demographic data, reporting means and SD or medians and IQRs for continuous variables and counts and percentages for categorical variables. We calculated annual exposure by multiplying the average weekly training hours by 52, then used this to determine injury rates per 1000 hours.¹⁹ For competition injuries, we calculated rates per 1000 matches. We reported injury data with 95% CIs and presented the overall distribution of injuries by anatomical site and mechanism as percentages. To test for differences in the number of injuries between groups, we used t-tests when comparing two categories and analysis of variance when comparing multiple groups if the data were normally distributed. If the data were not normally distributed based on the Shapiro-Wilk test, we used the Wilcoxon rank sum test and the Kruskal-Wallis test to compare two or more groups, respectively.

Patient and public involvement

Neither patients nor the public participated in the design, execution, reporting or sharing of this study.

Results

Participants

900 respondents filled in the survey. Eight did not provide informed consent, 10 had no BJJ experience and 1 respondent filled in the questionnaire twice; these responses were excluded from the analysis, bringing the final sample size to 881 BJJ practitioners. Most of the participants came from the Netherlands (41%), the USA (22%), the UK (6%) and Germany (5%). On average, participants trained 6.8 (SD 5.2) hours per week, of which 3.2 (SD 3.2) was spent sparring. Table 1 provides an overview of the demographic and athletic data of the 881 participants.

Table 1. Demographics and characteristics.

Variable	Category	n (%)/mean±SD
Gender, n (%)
	Female	60 (6.8)
	Male	817 (92.7)
	Neither	2 (0.2)
	Prefer not to say	2 (0.2)
Age, mean±SD (years)		30.8±8.6
Height, mean±SD (centimetres)		179.3±8.1
Weight, mean±SD (kilograms)		80.3±12.2
Body mass index, mean±SD		24.9±3.0
Experience, mean±SD, (years)		5.5±4.7
Belt level, n (%)
	White	309 (35.1)
	Blue	316 (35.9)
	Purple	134 (15.2)
	Brown	61 (6.9)
	Black	61 (6.9)
Weekly training hours		6.8±5.2
Weekly sparring		3.2±3.2
Warming up, n (%)
	Never	31 (3.5)
	Every now and then	74 (8.4)
	Sometimes	107 (12.1)
	Often	187 (21.3)
	Always	482 (54.7)
Strength and conditioning, n (%)
	Yes	589 (66.9)
	No	292 (33.1)
Instructor, n (%)
	Yes	184 (20.9)
	No	697 (79.1)

Injury incidence

In total, 714 (81%) participants sustained at least one training or competition injury in the past year, 506 (57%) sustained more than two injuries and 288 (33%) sustained more than three injuries. 695 participants (79%) sustained at least one injury during training, and 149 (17%) participants sustained at least one injury during competition. Participants reported a total of 1913 injuries, of which 1711 (89%) were sustained during training and 202 (11%) during competition (table 2). Combined with a total of 310 830 hours of BJJ training during the past year, this resulted in an injury incidence of 5.5 (95% CI 4.9 to 6.1) per 1000 hours of training and 202 competition injuries over 3604 matches, resulting in a competition injury incidence of 55.9 (95% CI 38.8 to 73.0) per 1000 matches.

Table 2. Injury characteristics.

All injuries (n=1913)
Training or competition, n (%)
	Training	1711 (89.4)
	Competition	202 (10.6)
Injuries with detailed description (n=888)
Time loss, mean±range (days)		33.9 (0–365)
Competition, n (%)		99 (11.1)
Part of training, n (%) (n=789)
	Warming up	18 (2.3)
	Technique training	69 (8.7)
	Position sparring	79 (10.0)
	Free sparring	620 (78.6)
	Cool down	3 (0.4)
Onset, n (%)	Acute	511 (57.5)
	Repetitive	89 (10.0)
	Repetitive with sudden onset	200 (22.5)
	Missing data	88 (9.9)
Mechanism, n (%)
	Direct contact	563 (63.4)
	Indirect contact	178 (20.0)
	No contact	59 (6.6)
	Missing data	88 (9.9)

Injury characteristics

Out of the 1913 reported injuries, participants offered detailed information for 888, which accounts for 46%. 511 were acute injuries (58%), and 563 were a consequence of direct contact with a training partner or opponent (63%). Of these injuries for which detailed information was provided, 99 happened during a competition and 789 happened during training. Of the 789 training-related injuries, 620 happened during the free sparring portion of the training (n=620, 70%).

We made comparisons for all and training injuries only because there were insufficient competition injuries in the current sample to employ meaningful statistical tests. The number of injuries and the injury incidence were non-normally distributed across all comparison categories. Hence, we used the non-parametric statistical tests to compare groups (table 3). Overall, males sustained significantly more injuries than females (p=0.026), but this difference was not statistically significant for training injuries only or injury incidence (per 1000 hours). Participants who were also BJJ instructors sustained fewer injuries on average (p=0.049) than non-instructors. When comparing the different belt levels, we found an overall significant difference in the number of injuries and the injury incidence. A pairwise comparison using a Bonferroni correction showed that injury incidence was significantly lower for brown or black belts than for white or blue belts. We found no difference in the number of injuries between participants who did and did not perform strength and conditioning to support their BJJ training. How often participants performed a warm-up was also not associated with the number of injuries.

Table 3. Group-wise comparison for training and all injuries.

Group	Category	Median number of injuries (IQR)	P value	Median number of training injuries (IQR)	P value	Training injury incidence per 1000 hours of training (median (IQR))	P value
Gender
	Male*	2 (1; 3)	0.026	2.5 (1.8; 3.2)	0.061	5.5 (2.4; 9.6)	0.094
	Female	1 (0; 2.2)		1 (0; 2.2)		4.6 (0; 8.5)
	Neither	2.5 (1.8; 3.2)	–	2.5 (1.8; 3.2)		7.1 (5.2; 9)
	Not Provided	0 (0; 0)	–	0 (0; 0)		0 (0; 0)
Belt level			0.009		0.029		0.001
	White	2 (1; 3)		2 (1; 3)		6.4 (1.9; 11.5)
	Blue	2 (1; 3)		2 (1; 3)		6.4 (3.1; 10.8)
	Purple	2 (1; 3)		2 (1; 3)		4.8 (2.7; 8.2)
	Brown	2 (0; 2)		2 (0; 2)		3.8 (0; 6.4)
	Black	2 (1; 3)		2 (1; 2)		3.8 (1; 7.7)
Strength and conditioning
	Yes	2 (1; 3)	0.114	2 (1; 3)	0.113	5.5 (2.4; 9.6)	0.745
	No	2 (1; 3)		2 (1; 3)		4.8 (1.9; 9.6)
BJJ instructor
	Yes	2 (1; 3)	0.049	2 (1; 3)	0.105	4.8 (1.9; 8.1)	0.057
	No	2 (1; 3)		2 (1; 3)		5.8 (2.4; 9.6)
Warm-up frequency			0.641		0.600		0.643
	Never	2 (1; 3)		2 (1; 3)		3.8 (0.8; 9.6)
	Rarely	2 (1; 3)		1 (1; 3)		5.6 (1.5; 9.6)
	Sometimes	2 (1; 3)		2 (0.5; 2.5)		5.1 (0.4; 9.6)
	Often	2 (1; 3)		2 (1; 3)		5.8 (2.8; 9.6)
	Always	2 (1; 3)		2 (1; 3)		4.8 (2.4; 9.6)

^*Comparison is between male and female to due lack of data in other categories.

BJJ, Brazilian Jiu-Jitsu.

Submission holds accounted for 28% of injuries (n=247), particularly armlocks (n=92, 37% of submission injuries) and leglocks (n=92, 37% of submission injuries). Differentiating further into specific submission holds showed that the armbar accounted for many of the submission injuries (n=52, 21%), followed by the toe hold (n=24, 10%) and inside heel hook (n=20, 8%). Approximately 42% (n=373) of injuries were diagnosed by a medical professional. The knee was the most injured anatomical site (n=223, 25%), followed by the shoulder (n=114, 13%), hand (n=71, 8%) and chest (n=62, 7%) (figure 1).

Figure 1. Injury frequency by body region.

Discussion

The yearly injury prevalence among the 881 BJJ participants was 81%. We found an injury incidence rate of 5.5 (95% CI 4.9 to 6.1) per 1000 hours of training and 55.9 (95% CI 38.8 to 73.0) per 1000 matches. The most affected body regions were the knee and shoulder. Most injuries happened during free sparring and were acute-onset injuries due to direct contact with an opponent.

Comparison with existing literature

The injury incidence in this study was difficult to compare with other studies due to the scarcity of research and the wide variation in injury definitions and study designs.^{11 12 20} Hinz et al¹¹ reported an incidence of 308 injuries per 1000 athlete years and a 3-year injury incidence of 69% in their survey study of 1140 participants.¹¹ However, their stringent injury definition: ‘Injuries necessitating an absence from training for at least 2 weeks’, only entails complete time loss injuries and fails to include injuries that were substantial enough to modify training or seek medical attention but not to cease training altogether,¹⁷ which may have resulted in underreporting. McDonald et al¹² used a more inclusive injury definition, including injuries requiring medical attention. They reported a yearly prevalence of 85%, similar to what we found in the current study.¹²

The current study indicates that while males sustained more injuries in general, the number of training injuries per 1000 hours was similar between the sexes. Currently, there are limited data regarding sex differences in BJJ injuries. Considering the apparent sex-specific factors involved in sports injuries in general,²¹ future BJJ studies should examine sex more thoroughly as a risk factor for injury. Interestingly, we found that participants with higher belt levels (brown and black) had a significantly lower training injury incidence than those with lower belt levels (white and blue). This contradicts what Kreiswirth et al¹⁰ found regarding injuries at the BJJ world championship, where a higher belt level was associated with increased injury risk.¹⁰ Hinz et al¹¹ also found that a higher belt level was associated with increased injury risk for training and competition injuries.¹¹ Since we only compared training-related injury incidence, a possible explanation is that the additional experience of practitioners with higher belt levels is a protective factor during training, where the stakes are typically lower than in competition.

The most frequently injured body parts in the current study were the knee and shoulders, which is in line with Kreiswirth et al¹⁰ and Scoggin et al⁹ but not with Stephenson and Rossheim.^{22 9 10 22} The latter found that the head, neck and trunk were the most frequently injured sites. This discrepancy may be explained by the fact that Stephenson and Rossheim.²² included only injuries that were reported to the emergency room, and head or neck trauma is likely to prompt an immediate visit to the emergency room. Furthermore, BJJ-sparring aims to submit one’s opponent; this is most frequently done or attempted via leg or arm locks, which put the extremities in extreme positions, forcing the opponent to submit. However, extremity injuries often arise not only from submission techniques but also from unplanned, rapid movements during sparring and takedowns, characterised by high physical unpredictability and quick directional changes.⁸ These two factors combined may explain the large proportion of knee and shoulder injuries in the current sample. In line with Scoggin et al,⁹ the armbar was the submission that led to the most injuries in this study.⁹

Strengths and limitations

This study has several noteworthy strengths. We included a relatively large (n=881) international sample, recorded a substantial number of injuries (n=1913), and participants provided detailed information for 888 injuries. This allows us to draw meaningful conclusions regarding injury incidence and characteristics. We also gathered detailed information regarding participation and injury characteristics, including experience level, typical focus on warm-up, injury mechanism and injury onset. Here, this study provides insights into various factors that may be included in future aetiological investigations.

Despite our large sample size and robust methodological design, this study has several limitations. As researchers noted previously, participants who have had injuries may be more likely to fill in a survey regarding injuries, resulting in sampling bias.²³ Furthermore, the survey could only be completed digitally, which limits the responses to participants who are comfortable using digital devices and poses another potential source of sampling bias. Recall bias is inherent to retrospective studies. However, we minimised recall bias by only inquiring about the previous year, and Gabbe et al²⁴ showed that athletes can adequately recall injuries that happened in the past months.²⁴ The survey only required participants to register detailed injury information about one injury, with the optionality to do this for three injuries. This limit was set to avoid respondent fatigue and encourage completion of the survey while minimising the risk of under-reporting the most severe cases. Finally, we used an injury definition that only captured injuries that caused (partial) time loss or required medical attention. This resulted in an inability to capture injuries that did not result in time loss or require medical attention. However, we reasoned that the physical nature of BJJ results in many physical complaints, many of which are not serious enough to be considered an injury by the athlete. To avoid over-reporting of injuries, we used the current injury definition based on previous similar research in other sports.¹⁵

Practical implications and future research

Although the injury incidence is highest during competition, most injuries occurred during training due to the significantly larger volume of training hours compared with competition time. Considering that a large portion of BJJ class time is spent on free sparring and that the goal in free sparring is typically submitting an opponent, it is understandable that most injuries happen during free sparring and in submissions. While injuries are unavoidable, the high incidence of training injuries during free sparring may indicate a need for better preventative practices. Injuries during sparring are a major challenge in combat sports training in general. Other combat sports like boxing and judo have solved this by implementing protective equipment or modifying the rules.²⁵ Some BJJ schools have implemented rules where submissions are no longer finished by tapping out but rather by holding the end range of motion for three seconds. This method may prove beneficial during armbars where the joint is at significant risk of hyperextension trauma when the submission is applied quickly. This prevents extending extremities past safe ranges of motion to get the opponent to tap out and may create a safer sparring environment while maintaining the ability to practise submissions.

Further preventative work should be based on modifiable risk factors. Few studies have investigated these risk factors for injuries in a BJJ population. Prospective research to identify injury risk factors would be a logical next step in BJJ injury prevention research.²⁶

Conclusions

This cross-sectional study in 881 BJJ practitioners found an injury incidence of 5.5 (95% CI 4.9 to 6.08) per 1000 hours of training for training injuries and an incidence of 55.9 (95% CI 38.8 to 73.0) per 1000 matches for competition injuries. Injury incidence was also lower among participants with a brown or black belt level than those with a blue or white belt level. The knee and shoulder were the most injured body regions, and most injuries occurred during free sparring training. Therefore, preventative practices should focus on the free sparring part of training to be most effective. Future research should investigate injury risk factors prospectively in BJJ participants to provide more accurate information on injury aetiology, which may be used to optimise future preventative programmes.

Data availability statement

Data are available on reasonable request.