Ankle Sprains in the National Basketball

Association, 2013-2014 Through 2016-2017

Mackenzie M. Herzog,*yz§ PhD, MPH, Christina DeFilippo Mack,y§ PhD, MSPH,
Nancy A. Dreyer,y§ PhD, MPH, Erik A. Wikstrom,|| PhD, Darin A. Padua,|| PhD,
Mininder S. Kocher,{#** MD, MPH, John P. DiFiori,yyzz MD, and Stephen W. Marshall,yz|| PhD
Investigation performed at the University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina, USA

Background: Ankle sprains are one of the most common injuries in basketball. Despite this, the incidence and setting of ankle
sprains among elite basketball players are not well described.
Purpose: To describe the epidemiology of ankle sprains among National Basketball Association (NBA) players.
Study Design: Cohort study; Level of evidence, 3.
Methods: All players on an NBA roster for 1 NBA game (preseason, regular season, or playoffs) during the 2013-14 through
2016-17 seasons were included. Data were collected with the NBA electronic medical record system. All NBA teams used the
electronic medical record continuously throughout the study period to record comprehensive injury data, including onset, mech-
anism, setting, type, and time lost. Game incidence rates were calculated per 1000 player-games and per 10,000 player-minutes
of participation, stratified by demographic and playing characteristics.
Results: There were 796 ankle sprains among 389 players and 2341 unique NBA player-seasons reported in the league from 2013-
14 through 2016-17. The overall single-season risk of ankle sprain was 25.8% (95% CI, 23.9%-28.0%). The majority of ankle sprains
occurred in games (n = 565, 71.0%) and involved a contact mechanism of injury (n = 567, 71.2%). Most ankle sprains were lateral (n
= 638, 80.2%). The incidence of ankle sprain among players with a history of prior ankle sprain in the past year was 1.41 times (95%
CI, 1.13-1.74) the incidence of those without a history of ankle sprain in the past year (P = .002). Fifty-six percent of ankle sprains did
not result in any NBA games missed (n = 443); among those that did, players missed a median of 2 games (interquartile range, 1-4)
resulting in a cumulative total of 1467 missed player-games over the 4-season study period.
Conclusion: Ankle sprains affect approximately 26% of NBA players on average each season and account for a large number of
missed NBA games in aggregate. Younger players and players with a history of ankle sprain have elevated rates of incident ankle
sprains in games, highlighting the potential benefit for integrating injury prevention programs into the management of initial
sprains. Research on basketball- and ankle-specific injury prevention strategies could provide benefits.
Keywords: ankle; ankle sprain; National Basketball Association; NBA; basketball

Basketball is one of the most popular sports in the United with sprains accounting for approximately 90% of these
States, with 980,673 high school participants15 and 2192 basketball-related ankle injuries.9
teams in the National Collegiate Athletic Association Despite the high rates of ankle sprain among basketball
(NCAA)14 in 2016-2017. In addition, basketball is played players, there is little current information on incidence and
by an estimated 450 million individuals throughout the outcomes of ankle sprains at the elite levels of the sport.
world in recreational and professional settings10,19 and The purpose of this study was to describe the epidemiology
by 1 in 13 people in the United States at least once in a cal- of ankle sprains among National Basketball Association
endar year.16,18 (NBA) players from 2013-2014 through 2016-2017.
Ankle sprains are the most common injury across all lev-
els of basketball participation for both sexes.4,5,9-11,13,17,20,22
Previous studies estimated an incidence rate of 1.0 to METHODS
5.2 ankle injuries per 1000 person-hours in basketball,
We conducted a retrospective cohort study of all players on
an NBA roster for 1 preseason, regular season, or playoff
The American Journal of Sports Medicine games during the 2013-14 through 2016-17 seasons.
1–8
DOI: 10.1177/0363546519864678 Approval was obtained from the University of North Caro-
Ó 2019 The Author(s) lina at Chapel Hill Institutional Review Board, the NBA,

1
2 Herzog et al The American Journal of Sports Medicine

the National Basketball Players Association, and the several metrics: player-season (1 player participating in 1
Research Committee of the NBA Physicians Association. NBA season), player-game (1 player participating in 1
NBA game), and player-minute (individual player minutes
of participation in NBA games).
Ankle Sprain Definition
Incident ankle sprains were defined as acute-onset injury Time-Dependent Risk Factors
to any ligamentous structure of the ankle as diagnosed
by the team medical staff. Ankle sprains with no identifi- We examined age, body mass index, history of ankle sprain
able injury event (eg, exacerbations of chronic ankle insta- in the past year, and NBA game minutes played as possible
bility), those that occurred while the player was on G risk factors. When game incidence rates stratified by demo-
League assignment, or other sprains not directly related graphic and playing characteristics were calculated, these
to basketball participation were excluded from the analy- covariates were assessed in a time-dynamic manner at the
sis. Sprain type (lateral, medial, or high) was based on time of the player-game. Age was determined on the date of
the injured ligamentous structure that was entered in the player-game, and body mass index was calculated with
the electronic medical record (EMR) by the certified ath- the most recent height and weight before the player-game,
letic trainer (AT): injuries to the anterior talofibular liga- as recorded in the EMR or the player’s NBA preseason pro-
ment, posterior talofibular ligament, calcaneofibular file. Similarly, history of ankle sprain was based on ankle
ligament, or ‘‘general lateral ligaments’’ were considered sprains (whether or not they were basketball related and
‘‘lateral’’; injuries to the deltoid ligament or tibionavicular those that occurred before 2013-14) reported in the EMR
ligament were considered ‘‘medial’’; and injuries to the dis- cumulative from 365 days before the point in time immedi-
tal tibiofibular syndesmosis, anterior tibiofibular ligament, ately preceding each player-game. Average game minutes
or posterior tibiofibular ligament were considered ‘‘high.’’ played was calculated as the average number of minutes
played in NBA games in the 365 days before the player-
game of interest (ie, game of injury for incident ankle
Data Source
sprains). Players who were in their rookie season were
Data were obtained from the NBA EMR database, a cen- excluded from these calculations.
tralized data collection system that was integrated with
the clinical management of player health for all NBA Playing Time Missed
teams continuously throughout the study period.12 In addi-
tion to the player health information available through the Time missed attributed to ankle sprain was calculated in 2
NBA EMR, player and team participation information was ways: (1) the number of NBA games missed after the game,
integrated in the database, including minutes played per practice, or other activity in which the injury occurred and
game. Injury data, including onset, mechanism, setting, (2) the number of days between the date of injury and the
type, and time lost, were recorded by ATs. Uniform defini- date of the first NBA game in which the player participated
tions were used and accompanied by a detailed quality con- after injury. The number of NBA games missed after injury
trol and audit process. ATs were also required to enter was assessed per the player participation data that ATs
player-game participation data for each player across the entered in the NBA EMR. For our outcome metric of games
entire study period, including associating each missed missed because of injury, we included preseason, regular
game with a particular injury record if the player was season, and playoff games missed to give the full picture
unable to participate because of injury. Entry of practice of the effect of these injuries to the player.
participation information in the EMR was not required
before the 2016-17 season, and practice participation entry Statistical Analysis
was not audited across the study period. Thus, incidence
rate calculations focused on game injuries and game par- Descriptive statistics were calculated for all variables of inter-
ticipation. Playing time exposure was assessed with est, including counts, percentages, means, SDs, medians, and

*Address correspondence to Mackenzie M. Herzog, PhD, MPH, Department of Epidemiology, University of North Carolina at Chapel Hill, McGavran
Greenberg Hall, CB #7435, Chapel Hill, NC 27599-7435, USA (email: [email protected]).
y
Department of Epidemiology, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
z
University of North Carolina Injury Prevention Research Center, Chapel Hill, North Carolina, USA.
§
Injury Surveillance and Analytics, Real-World Analytics Solutions, IQVIA Durham, North Carolina, USA.
||
Department of Exercise and Sport Science, College of Arts and Sciences, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA.
{
The Micheli Center for Sports Injury Prevention, Boston Children’s Hospital Boston, Massachusetts, USA.
#
Division of Sports Medicine, Department of Orthopedics, Boston Children’s Hospital Boston, Massachusetts, USA.
**Department of Orthopaedic Surgery, Harvard Medical School Boston, Massachusetts, USA.
yy
National Basketball Association New York, New York, USA.
zz
Primary Care Sports Medicine, Hospital for Special Surgery New York, New York, USA.
One or more of the authors has declared the following potential conflict of interest or source of funding: M.M.H., C.D.M., and N.A.D. are full-time
employees of IQVIA, a human data science company that is funded for NBA injury surveillance and analytics efforts. D.A.P. is funded by the National Acad-
emy of Sports Medicine. J.P.D. is a paid consultant for the NBA. M.S.K. is a consultant for Smith & Nephew, OrthoPediatrics, and Ossur. AOSSM checks
author disclosures against the Open Payments Database (OPD). AOSSM has not conducted an independent investigation on the OPD and disclaims any
liability or responsibility relating thereto.
AJSM Vol. XX, No. X, XXXX Ankle Sprains in the NBA 3

TABLE 1
Ankle Sprains, Injured Players, Teams, and Players in the NBA by Year, 2013-14 Through 2016-17a

Year, n (%)

2013-14 2014-15 2015-16 2016-17 4-Season Average

Ankle sprains 203 215 190 188 199

Ankle sprains resulting in 1 games missed 85 (42) 103 (48) 80 (42) 85 (45) 88.3 (44)
Injured players 156 164 140 145 151.3
Injured players with 1 games missed 77 (49) 91 (55) 70 (50) 79 (54) 79.3 (52)
Team-games 2861 2847 2841 2816 2841.3
Playersb 570 580 598 593 585.3
Player-games 30,320 30,465 30,708 30,517 30,502.5
Player-minutesc 692,148 688,848 686,705 681,062 687,190.8

a
Ankle sprains, n = 796. NBA, National Basketball Association.
b
Players include all those who were on an NBA team roster for at least 1 game during each season. Three additional players (2013-14,
n = 1; 2014-15, n = 1; 2016-17, n = 1) who sustained an ankle sprain during the season but were not on an NBA team game roster during
the season of injury are included in the total number of players.
c
Player-minutes are available only for NBA game participation.

interquartile ranges where appropriate. The single-season Among all ankle sprains reported, a majority were lat-
risk of ankle sprain was calculated as the number of injured eral sprains (n = 638, 80.2%) and involved a contact mech-
players in a season divided by the total population at risk anism of injury (n = 567, 71.2%) (Table 3). Medial ankle
for injury in that season, which was defined as the total num- sprains were slightly more likely to occur because of a non-
ber of players on an NBA roster for 1 preseason, regular sea- contact mechanism than lateral or high sprains (37.1% vs
son, or playoff games in each season. Game incidence rates 28.8% and 33.3%, respectively). Of the sprains that
were calculated per 1000 player-games and per 10,000 occurred because of contact, most were reported as general
player-minutes of participation. Incidence rates per 1000 contact with another player (n = 240, 42.3%) or from step-
player-games included all games in which the player partici- ping on an opponent’s foot (n = 189, 33.3%). The most com-
pated irrespective of duration of participation. Incidence rates mon mechanism of noncontact sprains was inversion (n =
per 10,000 player-minutes included the total number of 141, 61.6%). Most ankle sprains occurred in games (n =
minutes in which each player participated during each 565, 71.0%) or practices (n = 158, 19.9%). The remaining
game. Models were created to calculate risks and rates with 73 ankle sprains occurred during training (eg, condition-
95% CIs, with log-link/binomial-residual regression and Pois- ing, weight lifting) or other basketball-related events (eg,
son regression, respectively. charity game, exhibition game). Common activities
Stratification was used to assess the risk and rate of reported at the time of the ankle sprain were general
ankle sprain among subgroups of players with cut points defense (n = 150, 18.8%), layup driving (n = 115, 14.5%),
defined by the median value of the variable. Comparisons and rebounding (n = 97, 12.2%).
between subgroups were performed with the models Fewer ankle sprains were reported in the first quarter as
described earlier with a robust standard error estimator compared with the second, third, or fourth quarter (Table 3).
to stabilize the variance in the log-link/binomial-residual Most ankle sprains that occurred in games were reported
models. We examined the overdispersion parameter (Pear- during the regular season. Game ankle sprain incidence
son chi-square divided by its degrees of freedom) to ensure rates were slightly higher in the preseason than in the reg-
appropriateness of the Poisson regression models. ular season and playoffs (4.8 vs 4.5 and 4.0 ankle sprains
per 1000 player-games, respectively) (Table 2). When counts
of ankle sprains reported by week were assessed, ankle
RESULTS sprain incidence peaked at the start of preseason games
and generally decreased over time (Figure 1).
There were 796 ankle sprains among 389 players reported
across 2341 unique NBA player-seasons among the 30
teams from 2013-14 through 2016-17 (Table 1).
Demographics

Risk, Rate, and Injury Characteristics Player characteristics at the time of ankle sprain were fairly
consistent across ankle sprain types, although players who
The overall single-season risk of ankle sprain was 25.8% sustained high ankle sprains had fewer prior ankle sprains
(95% CI, 23.9%-28.0%). The single-season risk of an ankle (mean 6 SD: 1.6 6 2.8 vs 2.4 6 2.8 for all injured players)
sprain resulting in 1 NBA game missed for injury was (Table 3). The ankle sprain incidence rate in games,
13.6% (95% CI, 12.1%-15.2%). Both the single-season risk particularly lateral sprains, was higher among players
and the game incidence rate of ankle sprain remained rel- aged \26 years versus 26 years (incidence rate ratio
atively stable across the study period (Tables 1 and 2). [IRR], 1.54; 95% CI, 1.27-1.86; P \ .001) (Table 4).
4 Herzog et al The American Journal of Sports Medicine

TABLE 2
Incidence of Ankle Sprain in NBA Games by Year, 2013-14 Through 2016-17a

Ankle Sprains, nb Player-Games, n Incidence Rate per 1000 Player-Games 95% CI Confidence Limit Ratioc

Overall 554 122,010 4.5 4.2-4.9 1.2

Year
2013-14 143 30,320 4.7 4.0-5.6 1.4
2014-15 141 30,465 4.6 3.9-5.4 1.4
2015-16 133 30,708 4.3 3.7-5.1 1.4
2016-17 137 30,517 4.5 3.8-5.3 1.4
Season
Preseason 53 10,997 4.8 3.7-6.3 1.7
Regular season 472 103,817 4.5 4.1-5.0 1.2
Playoffs 29 7196 4.0 2.8-5.8 2.1

a
Players include all those who participated in at least 1 NBA game during each season. Three additional players (2013-2014, n = 1; 2014-
2015, n = 1; 2016-2017, n = 1) who sustained an ankle sprain during the season but did not participate in at least 1 NBA game during the
season of injury are included in the total number of players. NBA, National Basketball Association.
b
Injuries were excluded if they were reported to have occurred in a game but could not be linked to a specific game in which the player
participated (n = 11 ankle sprains).
c
Upper bound of the 95% CI / lower bound of the 95% CI (provides an indicator of the precision of the estimate).

TABLE 3
Player Characteristics at Time of Ankle Sprain by Sprain Type, 2013-14 Through 2016-17

Ankle Sprains, n (%) or Mean 6 SD

All (n = 796) Lateral (n = 638) Medial (n = 62) High (n = 33) Unspecified (n = 63)

Setting
Game 565 (71) 457 (72) 44 (71) 20 (61) 44 (70)
Practice 158 (20) 132 (21) 6 (10) 11 (33) 9 (14)
Other 73 (9) 49 (8) 12 (19) 2 (6) 10 (16)
Mechanism
Contact 567 (71) 454 (71) 39 (63) 22 (67) 52 (83)
Noncontact 229 (29) 184 (29) 23 (37) 11 (33) 11 (17)
Timing in gamea
Pregame 3 (\1) 2 (\1) 0 (0) 0 (0) 1 (2)
First quarter 66 (12) 53 (12) 5 (11) 4 (20) 4 (9)
Second quarter 150 (27) 117 (26) 13 (30) 4 (20) 16 (36)
Third quarter 148 (26) 126 (28) 8 (18) 7 (35) 7 (16)
Fourth quarter 160 (28) 127 (28) 15 (34) 4 (20) 14 (32)
Overtime 10 (2) 7 (2) 2 (5) 1 (5) 0 (0)
Unknown/missing 28 (5) 25 (5) 1 (2) 0 (0) 2 (5)
Age, y 25.8 6 3.9 25.8 6 3.9 25.5 6 4.0 25.0 6 3.6 26.6 6 4.2
Body mass index,b kg/m2 24.8 6 1.8 24.8 6 1.8 24.3 6 1.5 24.5 6 1.6 24.9 6 1.6
Prior ankle sprains, n 2.5 6 2.8 2.5 6 2.8 2.3 6 2.3 1.6 6 2.0 2.7 6 3.3
Lateral 2.1 6 2.6 2.3 6 2.6 1.2 6 1.7 0.8 6 1.5 1.6 6 2.7
Medial 0.2 6 0.4 0.1 6 0.3 0.8 6 0.6 0.1 6 0.3 0.2 6 0.5
High 0.1 6 0.3 0.1 6 0.2 0.0 6 0.2 0.6 6 0.5 0.1 6 0.3

a
Timing in game is available for the 565 ankle sprains that were reported in games (n = 457, lateral; n = 44, medial; n = 20, high; n = 44,
unspecified).
b
Body mass index is missing for 72 ankle sprains (n = 61, lateral; n = 3, medial; n = 5, high; n = 3, unspecified).

History of Ankle Sprain ankle sprain was particularly notable for lateral ankle
sprain occurrence (IRR, 1.44; 95% CI, 1.16-1.80; P = .001).
The ankle sprain incidence rate in games was 41% higher
(IRR, 1.41; 95% CI, 1.13-1.74; P = .002) among players
with a history of ankle sprain in the past year as compared Playing Time Lost
with those without a history of ankle sprain in the past year
(Table 4). The absolute rate difference between players with Most ankle sprains did not result in any NBA games
and without a history of ankle sprain was 0.68 excess ankle missed (n = 443, 56%) beyond the game, practice, or other
sprains per 10,000 player-minutes. The effect of previous activity in which they occurred. Among ankle sprains that
AJSM Vol. XX, No. X, XXXX Ankle Sprains in the NBA 5

Offseason Preseason Regular Season Playoffs

40 All-Star Breakb
40

Averaged Number of Ankle Sprains

Cumulativea Number of Ankle

35 35
30 30

per Season
25 25
Sprains

20 20
15 15
10 10
5 5
0 0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

1
2
3
4

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25

1
2
3
4
5
6
7
8
Week of the NBA Seasonc

Figure 1. Cumulative (bars) and average (line) incidence of ankle sprain by week, 2013-14 through 2016-17. aThe cumulative
number of ankle sprains is the 4-year sum of ankle sprains reported each week. bThe All-Star break was shorter in 2013-14
than the other seasons (2013-14: 4 days; 2014-15, 2015-16, 2016-17, and 2017-18: 6 days). cThe first day of preseason week
1 is the first day of training camp (2013-14: September 27, 2013; 2014-15: September 26, 2014; 2015-16: September 24,
2015; 2016-17: September 23, 2016). dThe average is calculated as the cumulative number of ankle sprains reported each
week divided by 4. NBA, National Basketball Association.

resulted in at least 1 missed game, players missed a median 1988-89 through 1997-98 seasons (n = 874, 92.8%), possibly
of 2 NBA games (interquartile range [IQR], 1-4) and took driven by an increase in the proportion of high ankle
a median 8 days (IQR = 5-15) to return to NBA game sprains reported in our study (n = 33 [4.1%] vs n = 8
play (Table 5). High ankle sprains had a higher median [0.8%]). While it is possible that these differences in
number of games missed for injury (median, 5.0; IQR, reported incidence rates and sprain types across the liter-
3-16) and a longer median number of days to return to ature on NBA players represent true changes in the inci-
NBA game play (median, 24.0; IQR, 7-37). NBA games dence of ankle sprain, potentially because of changes in
missed and time lost for injury did not differ between con- the pace and style of play in the league over time, it is
tact and noncontact ankle sprains. A cumulative total of also important to note the significant differences in data
1467 NBA player-games were missed because of ankle collection methods and inclusion criteria.
sprains in the league over the 4-season study period. The previously published studies on NBA injuries uti-
lized a convenience sample of data reported from 1988-89
through 2004-05. During that period, only 88.2% of NBA
DISCUSSION teams were represented in the database, and the report-
able injury definition included the following: ‘‘injury
Approximately 1 in 4 NBA players experienced at least 1 required physician referral, prescription medication, or
incident ankle sprain each season, leading to a season both; resulted in a game or practice being missed, or it
average of nearly 200 ankle sprains reported in the NBA. necessitated emergent care.’’8 Ankle sprains reported in
Among these injuries, 44% resulted in the player missing the present study likely include injuries from a broader
a subsequent NBA game, and the cumulative effect was distribution of basketball activities, given the transition
1467 games missed attributed to ankle sprains that of data collection to the EMR, expansion of reporting to
occurred across 4 seasons (an average of approximately all injuries rather than those within a more narrow report-
370 player-games missed per season). able injury definition, and data quality enhancements over
time. Subsequently, the ability to accurately compare find-
Incidence of Ankle Sprain ings across studies is limited.
Although similar limitations associated with differences
Overall, the game incidence rate across the 4-season study in injury surveillance systems, sampling, and access to
period was 4.5 ankle sprains per 1000 player-games. In medical care also make it difficult to draw definitive con-
comparison, previous studies describing injuries and ill- clusions about the differences in rates across levels of
nesses in the NBA estimated game incidence rates of 3.2 men’s basketball participation, the incidence rate of ankle
to 3.5 ankle sprains per 1000 player-games.6,8,21 Consis- sprains identified among NBA players in this study
tent with the prior literature, a majority of the ankle appears higher than that in other levels of men’s basket-
sprains reported in the NBA from 2013-14 through 2016- ball participation. Previous studies based on the NCAA
17 were lateral sprains (n = 638, 80.2%); however, this pro- injury surveillance database reported ankle sprain inci-
portion was lower than that reported by Starkey21 from the dence rates among men’s basketball players of 2.3 per
6 Herzog et al The American Journal of Sports Medicine

TABLE 4
Game Injury Incidence Rates per 10,000 Player-Minutes by Player Characteristics
and Sprain Type, 2013-14 Through 2016-17a

Game Ankle Sprains

Group: Metric Allb (n = 554) Lateral (n = 447) Medial (n = 43) High (n = 20)

Age, y
\26
Injuries 297 246 20 9
Rate (95% CI) 2.51 (2.20-2.88) 2.08 (1.78-2.43) 0.17 (0.11-0.26) 0.08 (0.04-0.15)
26 (reference)
Injuries 257 201 23 11
Rate (95% CI) 1.63 (1.41-1.89) 1.28 (1.09-1.51) 0.15 (0.09-0.23) 0.07 (0.04-0.13)
\26 vs 26: IRR (95% CI) 1.54 (1.27-1.86) 1.62 (1.31-2.01) 1.15 (0.63-2.10) 1.09 (0.46-2.59)
Body mass index,c kg/m2
\25
Injuries 277 220 27 7
Rate (95% CI) 2.04 (1.75-2.37) 1.62 (1.36-1.94) 0.20 (0.13-0.30) 0.05 (0.03-0.11)
25 (reference)
Injuries 241 196 16 10
Rate (95% CI) 1.95 (1.66-2.30) 1.59 (1.34-1.88) 0.13 (0.08-0.22) 0.08 (0.04-0.15)
\25 vs 25: IRR (95% CI) 1.04 (0.84-1.30) 1.02 (0.80-1.31) 1.54 (0.81-2.92) 0.64 (0.25-1.64)
Prior ankle sprain in past yeard
1 prior sprain
Injuries 211 175 17 4
Rate (95% CI) 2.35 (1.99-2.78) 2.03 (1.70-2.43) 0.18 (0.11-0.29) 0.05 (0.02-0.12)
No prior sprain (reference)
Injuries 272 211 23 13
Rate (95% CI) 1.67 (1.46-1.92) 1.41 (1.23-1.62) 0.15 (0.10-0.22) 0.08 (0.05-0.14)
1 vs no prior: IRR (95% CI) 1.41 (1.13-1.74) 1.44 (1.16-1.80) 1.22 (0.64-2.32) 0.62 (0.23-1.70)
Average NBA game minutes in past yeard
\15
Injuries 57 48 2 1
Rate (95% CI) 2.44 (1.83-3.25) 2.05 (1.52-2.78) 0.09 (0.02-0.34) 0.04 (0.01-0.30)
15-30 (reference)
Injuries 259 204 25 12
Rate (95% CI) 1.86 (1.61-2.14) 1.47 (1.25-1.72) 0.18 (0.12-0.27) 0.09 (0.05-0.15)
.30
Injuries 167 134 13 4
Rate (95% CI) 1.87 (1.52-2.30) 1.50 (1.18-1.92) 0.15 (0.08-0.26) 0.05 (0.02-0.12)
IRR (95% CI)
\15 vs 15-30 1.31 (0.95-1.82) 1.40 (0.99-1.98) 0.48 (0.11-2.03) 0.50 (0.06-3.78)
.30 vs 15-30 1.01 (0.79-1.29) 1.02 (0.77-1.36) 0.81 (0.42-1.57) 0.52 (0.17-1.59)

a
IRR, incidence rate ratio; NBA, National Basketball Association.
b
Injuries were excluded if they were reported to have occurred in a game but could not be linked to a specific game in which the player
participated (n = 11 ankle sprains).
c
Body mass index is missing for 8591 (7.0%) player-games, including 37 ankle sprains that were linked to a game.
d
Players who were injured during their rookie season were excluded (n = 14,250 player-games, n = 71 ankle sprains).

1000 player-games7 from 1988-1989 through 2003-2004 and Injury History as a Risk Factor
1.4 per 1000 athlete-exposures23 (which includes 1 player
participating in 1 game or practice) from 2009-2010 through A key finding of this study was that the incidence rate of
2014-2015. Furthermore, there is some indication that inci- ankle sprain in games was 41% higher among players
dence rates increase as the level of participation increases, with a history of ankle sprain in the past year as compared
with a reported incidence rate of ankle sprains in high with those without a history of ankle sprain. History of
school boys’ basketball lower than that reported in college ankle sprain was reported to be one of the strongest risk
at 0.8 per 1000 athlete-exposures.17 Again, these compari- factors for future ankle sprain in other physically active
sons provide some indication of trends across levels of par- populations.2,3,13 Agel et al1 reported in 2007 that women’s
ticipation, but more research with strong and consistent basketball players with a history of ankle sprain were 5.5
methodology could provide insight into ankle sprain occur- times more likely to sustain another ankle sprain as com-
rence and outcomes across the participation spectrum. pared with those with no history of ankle sprain, and
AJSM Vol. XX, No. X, XXXX Ankle Sprains in the NBA 7

TABLE 5
NBA Games Missed Because of Ankle Sprain and Days to Return to Game Play
by Sprain Type and Mechanism, 2013-14 Through 2016-17a

NBA Games Missed Days to Return to NBA Game Play

Ankle Sprains, n Missed 1 Game, n (%) Total Median (IQR) Mean (Range) Median (IQR) Mean (Range)

Overall 1467 2 (1-4) 4.2 (1-59) 8 (5-15) 20.4 (1-283)

Sprain type
Lateral 638 268 (42) 1036 2 (1-4) 3.9 (1-59) 8 (5-15) 20.6 (1-220)
Medial 62 32 (52) 147 3 (2-4.5) 4.6 (1-18) 8 (5-14) 29.5 (2-283)
High 33 17 (52) 173 5 (3-16) 10.2 (1-31) 24 (7-37) 24.9 (4-79)
Unspecified 63 36 (57) 111 2 (1-3) 3.1 (1-19) 6 (4-10) 9.3 (3-42)
Mechanism
Contact 567 258 (46) 1104 2 (1-4) 4.3 (1-59) 7.5 (5-15) 18.1 (1-220)
Noncontact 229 95 (41) 363 2 (1-4) 3.8 (1-25) 8 (5-18.5) 26.7 (1-283)

a
IQR, interquartile range; NBA, National Basketball Association.

a study of recreational and elite basketball players in Aus- Strengths and Limitations
tralia noted 4.9-times higher odds (95% CI, 2.0-12.5) of
ankle sprain among athletes with a history of ankle injury This study used the NBA EMR system, which contains
as compared with those without.13 To our knowledge, this detailed player health and participation information. The
is the first study to assess this association among NBA NBA EMR system included consistent reporting guidelines
players. Given the consistency in results suggesting that across the study period, and these likely enhanced accuracy
ankle sprain history is an important risk factor for injury and completeness of the data reported. However, it is possible
across multiple sports populations, this area may warrant that reporting changes in data entry over time may affect
further investigation into factors that may mitigate the comparisons across seasons of the study. It is also important
increase in risk, as well as opportunities to predict and pre- to note that rosters change throughout the course of the sea-
vent future injury occurrence. son, and some players who are on an NBA team roster in the
preseason are released from the league before the start of the
regular season. Thus, not all players who were on a roster
Setting, Timing, Player Demographics, have the same amount of exposure to NBA play. Addition-
and Playing Characteristics ally, data on player injury history outside of the NBA are lim-
ited, potentially resulting in some underascertainment of
In addition to supporting calculations of risk and incidence sprains that occurred outside of the player’s NBA career.
rates, the NBA EMR data used in this study provide rich Reporting of mechanism and activity at the time of
information on setting, timing, player demographics, and injury relies on AT observation or athlete recall and was
playing characteristics associated with ankle sprain occur- not validated. Only game participation data were available
rence. Our results suggest that younger players have higher for this study; therefore, incidence rate calculations are
rates of ankle sprains in game than do players aged 26 presented for game injuries only, and game results may
years. Within games, ankle sprains are also more commonly not be generalizable to other settings of participation. We
reported in the second, third, and fourth quarters. The rea- also were not able to assess practice or conditioning time
son for this is unclear but may relate to intensity of competi- missed because of the injury, and players who did not
tion, changes in type of player (starters vs rotation players), miss a game may have missed other NBA team-related
or reduction in joint stability over the course of the game, per- activities because of the ankle sprain. Missed game calcu-
haps secondary to the effectiveness of tape wearing off. In lations focus only on complete games missed after the onset
addition, the majority of reported ankle sprains were due to of injury and do include time missed in the game where the
a contact mechanism, such as stepping on an opponent’s injury occurred; in cases where a player was removed from
foot or other player-to-player contact. These findings provide participation in a game because of an incident injury but
insight into which players may be at increased risk of sus- was able to participate in future NBA games, the partially
taining an injury, which can be useful for targeting future missed game was not counted in these metrics. In addition,
prevention initiatives. In addition, these data help identify there may be a different threshold for return to play
potential intrinsic and extrinsic injury risk factors that could throughout the season (eg, players may be more likely to
benefit from further research. In particular, future studies miss preseason games than playoffs games) that may influ-
should consider further assessing mechanisms of ankle ence missed time calculations. Missed time calculations
sprains, perhaps including video review. are also affected by the timing of injury within the NBA
8 Herzog et al The American Journal of Sports Medicine

season, as players with injuries that occur toward the end Athletic Association Injury Surveillance System, 1988-1989 through
of the season have reduced opportunity to miss games. 2003-2004. J Athl Train. 2007;42(2):194-201.
8. Drakos MC, Domb B, Starkey C, Callahan L, Allen AA. Injury in the
Medians and interquartile ranges were emphasized for
National Basketball Association: a 17-year overview. Sports Health.
descriptive statistics instead of means and ranges, as they 2010;2(4):284-290.
were less likely to be influenced by these censored observa- 9. Fong DT, Hong Y, Chan LK, Yung PS, Chan KM. A systematic review on
tions. We had a low sample size to assess medial and high ankle injury and ankle sprain in sports. Sports Med. 2007;37(1):73-94.
sprains, and data on shoe type, bracing, and taping were 10. Gribble PA, Bleakley CM, Caulfield BM, et al. Evidence review for the
not available. Finally, this study assesses a unique popula- 2016 International Ankle Consortium consensus statement on the
tion of elite athletes who participate in the NBA; results of prevalence, impact and long-term consequences of lateral ankle
sprains. Br J Sports Med. 2016;50(24):1496-1505.
this study may not be generalizable to other populations. 11. Hosea TM, Carey CC, Harrer MF. The gender issue: epidemiology of
ankle injuries in athletes who participate in basketball. Clin Orthop
Relat Res. 2000;372:45-49.
CONCLUSION 12. Mack CD, Meisel P, Herzog MM, et al. The establishment and refine-
ment of the National Basketball Association player injury and illness
database. J Athl Train. 2019;54(5):466-471.
Ankle sprains affect approximately 26% of NBA players on
13. McKay GD, Goldie PA, Payne WR, Oakes BW. Ankle injuries in basket-
average each season and account for a large number of ball: injury rate and risk factors. Br J Sports Med. 2001;35(2):103-108.
missed NBA games in aggregate. Younger players and play- 14. National Collegiate Athletic Association. Student-athlete participation
ers with a history of ankle sprain have elevated rates of inci- 1981-82–2016-17: NCAA sports sponsorship and participation rates
dent ankle sprains in games, highlighting the potential report. http://www.ncaa.org/about/resources/research/sports-sponsor
benefit of integrating injury prevention training programs ship-and-participation-research. Accessed June 12, 2018.
into the management and rehabilitation of initial sprains, 15. National Federation of State High School Associations. 2016-17 high
school athletics participation survey results. http://www.nfhs.org/
as well as the potential benefits of future research on ParticipationStatistics/PDF/2016-17_Participation_Survey_Results
basketball- and ankle-specific injury prevention strategies. .pdf. Accessed June 12, 2018.
16. National Sporting Goods Association. Sports Participation in 2005:
Series 1. Mt Prospect, IL: National Sporting Goods Association; 2006.
REFERENCES 17. Nelson AJ, Collins CL, Yard EE, Fields SK, Comstock RD. Ankle inju-
ries among United States high school sports athletes, 2005-2006.
1. Agel J, Olson DE, Dick R, Arendt EA, Marshall SW, Sikka RS. J Athl Train. 2007;42(3):381-387.
Descriptive epidemiology of collegiate women’s basketball injuries: 18. Pappas E, Zazulak BT, Yard EE, Hewett TE. The epidemiology of
National Collegiate Athletic Association Injury Surveillance System, pediatric basketball injuries presenting to US emergency depart-
1988-1989 through 2003-2004. J Athl Train. 2007;42(2):202-210. ments: 2000-2006. Sports Health. 2011;3(4):331-335.
2. Barker HB, Beynnon BD, Renstrom PA. Ankle injury risk factors in 19. Pasanen K, Ekola T, Vasankari T, et al. High ankle injury rate in ado-
sports. Sports Med. 1997;23(2):69-74. lescent basketball: a 3-year prospective follow-up study. Scand J
3. Beynnon BD, Murphy DF, Alosa DM. Predictive factors for lateral Med Sci Sports. 2017;27(6):643-649.
ankle sprains: a literature review. J Athl Train. 2002;37(4):376-380. 20. Roos KG, Kerr ZY, Mauntel TC, Djoko A, Dompier TP, Wickstrom EA.
4. Borowski LA, Yard EE, Fields SK, Comstock RD. The epidemiology of The epidemiology of lateral ligament complex ankle sprains in
US high school basketball injuries, 2005-2007. Am J Sports Med. National Collegiate Athletic Association sports. Am J Sports Med.
2008;36(12):2328-2335. 2017;45(1):201-209.
5. Cumps E, Verhagen E, Meeusen R. Prospective epidemiological study 21. Starkey C. Injuries and illnesses in the National Basketball Associa-
of basketball injuries during one competitive season: ankle sprains and tion: a 10-year perspective. J Athl Train. 2000;35(2):161-167.
overuse knee injuries. J Sports Sci Med. 2007;6(2):204-211. 22. Waterman BR, Owens BD, Davey S, Zacchilli MA, Belmont PJ Jr. The
6. Deitch JR, Starkey C, Walters SL, Moseley JB. Injury risk in profes- epidemiology of ankle sprains in the United States. J Bone Joint Surg
sional basketball players: a comparison of Women’s National Basket- Am. 2010;92(13):2279-2284.
ball Association and National Basketball Association athletes. Am J 23. Zuckerman SL, Wegner AM, Roos KG, Djoko A, Dompier TP, Kerr
Sports Med. 2006;34(7):1077-1083. ZY. Injuries sustained in National Collegiate Athletic Association
7. Dick R, Hertel J, Agel J, Grossman J, Marshall SW. Descriptive epi- men’s and women’s basketball, 2009/2010-2014/2015. Br J Sports
demiology of collegiate men’s basketball injuries: National Collegiate Med. 2018;52(4):261-268.

For reprints and permission queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.