Achilles Tendon Ruptures in National Hockey League Players: Return to Sport and Performance Impact

Abstract

Background:

Few studies assess rates of return to play and postinjury performance in National Hockey League (NHL) players who sustain Achilles tendon ruptures. Our objective was to determine the rate of return to play and performance impact among NHL players who undergo surgical repair of Achilles tendon tears.

Methods:

NHL players who sustained an Achilles tendon rupture between 2001 and 2021 were identified using a publicly available injury database. Demographic and outcome data were collected for the 1-year period preceding and the 2-year period following surgery. Our primary outcome was expected wins above replacement per 60 minutes played. A position, draft year, and index season performance matched cohort was created. Pre- and postinjury outcomes were compared between cases and controls with a paired t test.

Results:

We identified 15 cases (9 forwards, 5 defencemen, 1 goaltender). Fourteen of 15 (93%) players returned to play. Preinjury, postinjury year 1, and postinjury year 2 expected wins above replacement were 0.05, 0.05, 0.05 respectively (P > .05). There was no significant difference in performance between cases and controls at any time point.

Conclusion:

Achilles tendon tears are associated with a high rate of return to play in the NHL and are not associated with a significant change in offensive, defensive, or overall performance-based metrics.

Level of Evidence:

Level III, case-control study.

Introduction

Achilles tendon ruptures are significant injuries. Prior work has demonstrated that they are associated with rates of failure to return to play (RTP) of 20.5%, ¹⁷ 29.2%, ²⁴ and 2% ³⁴ in National Basketball Association (NBA) players; 34.4%, ³⁴ 27.6%, ¹⁵ and 38.7% ³⁷ in National Football League (NFL) players; 0% ³⁴ in Major League Baseball (MLB) players; 7.5% ³⁶ in National Collegiate Athletics Association (NCAA) football players; and 13% ¹² and 29.2% ³⁵ in professional soccer players. In NFL players, failure to RTP has been shown to be associated with fewer preinjury games played per season and declining performance in the season prior to injury. ³⁷ Achilles tendon ruptures in professional athletes have been shown to be associated with decreased career duration, ¹⁵ fewer games played,^17,24 fewer minutes played, ³⁵ and decreased performance. ¹⁵

Prior work assessing outcomes following Achilles tendon ruptures in National Hockey League (NHL) players has demonstrated a 78.9% RTP rate with no significant change in performance as measured by traditional statistics including goals scored, assists, plus minus, penalty minutes, shots, and games played. ²⁰ Although traditional statistics does paint a general picture of the effect of Achilles tendon ruptures on NHL performance, these statistics can be influenced by randomness, strategy, usage, teammate performance, and game script and have been shown to be a poor predictor of future team performance. ²

Advanced statistics provide deeper performance insights and exhibit superior predictive capabilities compared with traditional statistics.^2,21,23 Corsi and expected goals are designed to mitigate the influence of randomness by using scoring opportunities as proxies for performance.^2,21 These metrics can be further refined with Regularized Adjusted Plus Minus (RAPM), which accounts for the impact of teammates and opponents on performance by regressing Corsi and expected goals by all players on ice at any given time, and can be included in models that incorporate additional variables such as zone starts, hits, giveaways, takeaways, faceoffs, and blocked shots to quantify a player’s contribution in offensive and defensive scenarios and their overall value to the team expressed as goals above replacement and wins above replacement, respectively.^21,38

Understanding performance changes following Achilles tendon rupture is important for establishing treatment expectations, forecasting performance outcomes, and evaluating the efficacy of various treatment modalities. The purpose of the current study was to determine the rate of RTP following operatively managed Achilles tendon ruptures in NHL hockey players. Additionally, we sought to assess performance following RTP, and to compare performance between players with Achilles tendon injuries and matched controls. Our hypothesis was that performance following Achilles tendon rupture will be unchanged when compared with preinjury and matched controls.

Methods

This was a retrospective cohort study evaluating NHL players who sustained Achilles tendon rupture and underwent surgical repair. Institutional review board (IRB) approval was not required for this study given that the data were publicly available. Consistent with prior work evaluating performance in professional athletes following orthopaedic injury, we identified cases using publicly available data.^{6,16,18,20,22,29} We identified NHL players who underwent surgical repair of Achilles tendon ruptures between the 2001/2002 to 2020/2021 seasons using the NHL Injury Viz Sortable Injury Database. ²⁵ News articles, press releases, injury reports, player profiles, and biographies were obtained for each player via database searching with a web search using the term “(player name) Achilles tendon rupture.” Sources obtained using this search were used to determine the exact injury, date, mechanism, rerupture, complications, and whether affected players underwent surgical repair. We included NHL players who sustained an Achilles tendon rupture and underwent subsequent surgical repair. No eligible players sustained a rerupture or other significant notable complication. Players who sustained an Achilles tendon rupture between the 2008/2009 and 2020/2021 seasons were eligible for inclusion in performance analysis as the NHL began collecting advanced player data in the 2007/2008 season. We included players in the performance analysis who had played at least 1 NHL season prior to injury and returned to play post injury. We excluded players from performance analysis who failed to RTP in the NHL.

We obtained birth date, height, weight, and career duration from Hockey Reference. ³¹ Draft year and round were collected from Hockey Database ¹⁴ alongside a list of NHL draft picks in each entry draft and used to create a position and draft year matched cohort. ⁸ We obtained player position, games played, expected goals, Corsi percentage, RAPM expected goal differential per 60 minutes played, RAPM Corsi differential per 60 minutes played, RAPM expected goals for per 60 minutes played, RAPM expected goals against per 60 minutes played, RAPM Corsi for per 60 minutes played, RAPM Corsi against per 60 minutes played, offensive goals above replacement per 60 minutes played, defensive goals above replacement per 60 minutes played, and wins above replacement per 60 minutes played from Evolving Hockey, ⁸ a website providing access to traditional and advanced hockey statistics that has previously been used in sports science research. ⁴ Performance data were collected for the last full season preceding the Achilles tendon injury and the ensuing first two seasons postinjury.

Outcome Measures

We evaluated postinjury performance with advanced statistics that use shots and scoring chances to quantify performance as shots and chances occur more frequently than goals. These metrics have previously been used in sports research to analyze postinjury performance. ¹ Corsi [comprising Corsi for/against (shot attempts for/against while a player is on the ice) and Corsi percentage (Corsi for/(Corsi for – Corsi against) while a given player is on the ice] are considered a proxy of territorial advantage via puck possession and have been shown to be a good predictor of team performance. ²¹ An expected goal is the shot location–based probability of a scoring chance resulting in a goal. ¹⁰ By using shot attempts and scoring chances to judge performance, Corsi and expected goals ameliorate the effect of randomness due to the low number of goals scored in hockey, with expected goals building on Corsi by quantifying and including the probability of scoring on a given shot. ² Regularized Adjusted Plus-Minus (RAPM) enhances these metrics by regressing them against various player combinations to adjust for the influence of teammate and opponent performance and isolate the impact of any individual on ice player on each metric. ⁹ To further objectify these metrics, RAPM expected goals for and against can be used in a linear regression model alongside shots, hits, zone starts, time on ice, turnovers, takeaways, faceoffs, and blocked shots with adjustment for team quality to quantify a player’s contribution relative to a replacement-level player in terms of goals scored or prevented in offensive and defensive situations. ³⁸ Wins Above Replacement (WAR) further extends this analysis by transforming a player’s offensive and defensive contributions into overall wins added or subtracted compared to a replacement-level player.^38,39 These metrics can then be expressed per unit time to adjust for playing time.

Cohort

We created a 1:1 matched cohort. Each included player’s draft year, round, and selection number were identified alongside a list of players selected in each year of the NHL entry draft using Hockey Database. ¹⁴ For each case, a matched player of the same position who was drafted in the same year and had similar baseline performance metrics (offensive/defensive expected goals above replacement, expected wins above replacement) was selected for inclusion in the cohort.

Data Analysis

We included all eligible players in the database who sustained Achilles tendon injuries. As such, an a priori sample size calculation was not performed. We performed data analysis with IBM SPSS Statistics, version 29. Outcome data were compared among individual players between each paired season (injury – 1 and injury + 1, injury – 1 and injury + 2) with a paired t test assuming equal variance. Data at each time point (injury – 1, injury + 1, injury + 2) were compared between cases and controls with a paired t test assuming equal variance. Demographic and outcome data was compared between the players who sustained an Achilles tendon laceration and players who sustained a noncontact rupture with the independent samples t test not assuming equal variances. P <.05 was used as a cutoff for a statistically significant result. Players who only played one season following injury were only included in analysis for the postinjury season in which they played.

Results

We identified 16 eligible players listed as having an Achilles tendon injury through database searching. We included 15 NHL players (9 forwards, 5 defencemen, and 1 goaltender) who underwent surgical Achilles tendon repair between the 2001/2002 and 2020/2021 seasons. Mean age, height, weight, and body mass index were 29.8 years, 186 cm, 90.4 kg, and 26.1, respectively. Players sustained injury an average of 7.9 seasons following the season in which they played their first NHL game. The injury mechanism was laceration in 5 cases and a noncontact injury while training in 10 cases. Fourteen of 15 (93%) players returned to play at least 1 game in the NHL with an median recovery time of 197 days (IQR 105) from the injury to their next regular-season game. Of the 10 players who are no longer active NHL players, mean career duration following the injury season was 4.4 seasons (Table 1).

Table 1.

Displays Demographic, Draft, and Career Length Data for Included Cases.

Player	Position a	Shot	Draft Round	Age at Injury	Height (cm)	Weight (kg)	BMI	Mechanism (1 = Laceration, 2 = noncontact)	Seasons Prior to Injury	Seasons Following Injury	Currently Active (2023/2024)
1	F	L	3	33	178	86	27	1	10	7	N
2	F	R	1	25	185	83	24	2	7	11	N
3	F	R	5	24	190	98	27	1	14	1	N
4	D	R	9	26	190	97	27	2	11	3	N
5	F	R	1	27	188	83	23	2	5	9	N
6	D	R	1	24	183	86	26	1	3	11	Y
7	F	R	2	34	185	97	28	2	14	3	N
8	G	L	7	27	198	97	25	2	5	0	N
9	F	R	1	27	185	92	27	1	5	8	Y
10	F	L	6	33	185	92	27	2	9	5	N
11	F	L	1	30	185	90	26	1	4	7	Y
12	D	L	3	36	183	90	27	2	12	3	N
13	F	R	4	31	180	83	26	2	6	5	Y
14	D	L	3	29	190	95	26	2	4	4	Y
15	D	L	U	33	185	87	25	2	4	1	N

Abbreviation: BMI, body mass index.

^aD, defenceman; F, forward; G, goaltender.

Thirteen players who sustained their injury between the 2008/2009 and 2020/2021 seasons and returned to play were included in statistical analysis. Table 2 displays index and postinjury data; of note, there was no significant difference between index, postinjury year 1, and postinjury year 2 for any outcome variables. Table 3 displays a comparison of index outcome variables for cases and controls; of note, there was no significant difference in pre- and postinjury outcome measures between groups.

Table 2.

Comparison of Preinjury and Postinjury Performances Among Cases at Baseline and 1 and 2 Years Postsurgery.

	Baseline a Performance (n = 13)	Performance 1 y Postinjury (n = 13)	P b	Performance 2 y Postinjury (n = 11)	P b
Games played	66	62	.46	70	.50
Expected goals per 60 min played c	0.57	0.55	.53	0.62	.21
Corsi percentage d	50.6	51.2	.67	50.8	.70
Offensive goals above replacement per 60 min played	0.25	0.19	.53	0.25	.99
Defensive goals above replacement per 60 min played	0.05	0.10	.24	0.04	.90
Wins above replacement per 60 min played	0.05	0.05	.92	0.05	.95

^aMost recent full season before Achilles tendon rupture/repair.

^bP values for comparison of performance at 1 and 2 years postsurgery with presurgery.

^cExpected goals refers to the shot location–based probability of a scoring chance resulting in a goal.

^dCorsi percentage refers to a team’s shot attempts for / (shot attempts for + shot attempts against) while a given player is on the ice.

Table 3.

Comparison of Preinjury and Postinjury Performances Between Cases and Matched Controls at Baseline and 1 and 2 Years Postsurgery.

	Baseline Performancea			Performance 1 y Postinjury		P	Performance 2 y Postinjury		P
	Cases	Controls		Cases	Controls		Cases	Controls
Games played	66	72	.38	62	63	.89	70	67	.30
Offensive goals above replacement per 60 min played	0.25	0.18	.47	0.19	0.08	.29	0.25	0.22	.66
Defensive goals above replacement per 60 min played	0.05	−0.04	.26	0.10	0.03	.41	0.04	−0.01	.56
Wins above replacement per 60 min played	0.05	0.03	.29	0.05	0.02	.19	0.05	0.04	.58

^aMost recent full season before Achilles tendon rupture/repair.

Table 4 displays a comparison of baseline data, career length, and games played per season among players who sustained an Achilles tendon laceration compared to players who sustained a noncontact rupture. Of note, players who sustained a noncontact rupture were on average older, had a shorter mean career duration, and played significantly fewer games both pre- and postinjury compared with players who sustained a laceration.

Table 4.

Comparison of players Who Sustained a Laceration to the Achilles tendon vs Players Who Sustained a Noncontact Injury.

	Laceration (n = 5)	Noncontact (n = 10)	P
Mean draft round	2	4	.17
Mean age at injury	27.6	30.1	.27
Mean height (cm)	184	187	.31
Mean weight (kg)	90	90	>.99
Mean BMI	27	26	.19
Mean seasons prior to injury	7.2	7.7	.84
Mean seasons following injury	6.8	4.4	.25
Mean GP 1 season preinjury	75.0	58.7	.04
Mean GP 1 season postinjury	77.8	55.6	.01
Mean GP 2 seasons postinjury	80.5	65.9	.03

Abbreviations: BMI, body mass index; GP, games played.

Discussion

Our primary objective was to determine the effect of Achilles tendon rupture on the rate of return to play and performance in NHL hockey players. We found a high rate of RTP (93%) and no significant differences in performance based on numbers available compared with preinjury and matched controls. Overall, our data suggest that NHL players can return to sport and perform at a high level following surgical management of Achilles tendon ruptures.

Return to sport following Achilles tendon ruptures has been reported to be 63% to 71%^17,24,34 in NBA basketball players, 61% to 72%^15,34,37 in NFL football players, 92.5% in NCAA football players, ³⁶ 100% ³⁴ in MLB baseball players, and 71% to 82%^12,34 in professional soccer players. Achilles tendon rupture is associated with shorter career duration and decreased performance in NFL running backs and linebackers, ¹⁵ fewer games played and decreased player efficiency rating in NBA basketball players,^17,24 no clinically significant postoperative performance change in NCAA football players, ³⁶ and decreased games and minutes played in professional soccer players. ³⁵ Two previous studies have assessed performance following Achilles tendon repair in NHL hockey players. Trofa et al ³⁴ assessed public records from 1989 to 2013 and found 1 player with an Achilles tendon rupture who did not return to play after injury and so performance following repair was unable to be assessed. Lubbe et al ²⁰ assessed performance after RTP following multiple orthopaedic procedures including Achilles tendon repair, finding that NHL players returned at a rate of 78.9% with no difference in performance score (a mathematical transformation of goals scored, assists, plus minus, penalty minutes, shots, and games played) following injury. Our finding of no significant performance change compared to preinjury and matched controls are similar to Lubbe et al, ²⁰ although our analysis is based on advanced statistics that address limitations associated with traditional statistics.

Our study’s finding of a failure to RTP rate of 7% is notably lower than rates of failure to return in other sports (20.5%-32%^17,24,34 in NBA players and 27.6%-38.7%^15,34,37 in NFL players). This discrepancy may be attributed to variations in injury mechanism, biomechanical demands, and career longevity among different sports. Achilles tendon ruptures are often caused by a combination of tendinosis and an acute loading stress. ¹⁹ Video analysis shows that in basketball^17,27and soccer players, ⁵ these ruptures typically occur during loading or forced dorsiflexion in noncontact situations. In contrast, 33% of Achilles tendon injuries in our cohort resulted from lacerations. Players who sustained lacerations may experience a higher RTP rate because of the absence of preinjury tendinosis and thus a higher-quality tendon for repair. Additionally, we found that these players were younger than players who sustained a noncontact Achilles tendon rupture and played significantly more games per season both pre- and postinjury compared to those with noncontact ruptures.

The rigid boot worn by ice hockey players partially restricts ankle dorsiflexion and plantarflexion (which may prevent the typical dorsiflexion loading pattern of Achilles tendon rupture seen in other sports) and minimizes plantarflexion activity (which may reduce the effect of Achilles tendon rupture on RTP and subsequent performance). ²⁶ NHL players have a mean career duration of 9.68 seasons, ¹¹ whereas NFL players average 3.3 seasons (among running backs[2.57], wide receivers [2.81], tight ends [2.85], cornerbacks [2.94], linebackers [2.97], and defensive lineman [3.24], having a mean career duration below the NFL average), ³² and NBA players average 4.5 years. ⁷ Thus, the higher propensity to return in NHL players may reflect their longer career expectancy and extended period of peak performance compared with athletes in other sports.

NHL players have RTP rates of 95% ¹⁸ following anterior cruciate ligament reconstruction and 87.5% to 100% after orthopaedic shoulder procedures.^3,20,28,33 Although some studies report no performance decline following ACL reconstruction, ⁶ others have noted decreased performance.^18,30 Orthopaedic shoulder procedures have been reported to be associated with decreased performance score 1 year postsurgery with a return to baseline by year 2 ¹³ and fewer games played postinjury, ²⁰ although other authors have reported no change in shots on goal per game ²⁸ ; goals, assists, points, and point shares ³³ ; and games played and performance score. ²⁰ Our finding of a 93% RTP rate and no significant changes in performance following Achilles tendon repair suggest that NHL players may return to play at similar rates and with similarly maintained performance metrics as NHL players returning from other common orthopaedic injuries.

A large advantage of our study was the use of Corsi, expected goals, and goals or wins above replacement. These metrics address limitations associated with traditional hockey statistics used in prior research. Games played, goals, assists, and points are influenced by randomness, usage, strategy, teammate performance, game script, and other factors not directly related to player’s innate ability; reliance on these metrics to evaluate player performance can lead to systemic misjudgement. ² On a team level, these outcomes are a less accurate predictor of team performance than advanced statistics. ² Prior research has demonstrated that expected goals better predict subsequent team performance than prior wins and losses or goals scored. ² Corsi is a proxy of territorial advantage through puck possession and has been shown to be a good predictor of team performance. ²¹

This study has several limitations. The use of publicly available data induces possible observer bias as there is the possibility that some players were missed who had differing performance outcomes compared with included players. However, publicly available data are frequently used in studies comparing outcomes following orthopaedic surgical procedures^6,18,22 and was improved on in our case using a publicly available database previously compiled in a systematic fashion. ²⁵ Furthermore, we did not have access to information on surgical technique, concomitant injuries, procedures, or rehabilitation protocols used, precluding further stratification and analysis. The power of our study was limited by our small sample size. Although we included all eligible players, an ex ante power calculation demonstrated that assuming 15 minutes of ice time per game over a 72-game season, we would need a sample of 26 cases to ascertain a performance difference of 1 win above replacement per season assuming an alpha of .05 and beta of 0.2.

Conclusion

In conclusion, Achilles tendon tears are associated with a high return to play in the NHL and are not associated with a significant change in offensive, defensive, or overall performance based on the numbers available.