Ice Hockey Injuries: Who Gets Hurt and Why Does it Matter?

Ice hockey is an inherently physical sport and as such creates situations where injury is possible and often likely. With the influence that injuries can have on a team’s success, research has started to focus on our ability to assess injury risk and prevent injuries before they occur. However, these strategies cannot be effectively laid out until we first understand who gets hurt, when they get hurt, and how they get hurt.

Who gets hurt?

According to a prospective cohort study performed by Flik et al., there is a significant disparity in injury incidence depending upon player position in NCAA Division 1 ice hockey. Only 6.2% of injuries were sustained by goaltenders, whereas 32.7% were defensemen and 61.1% were forwards. In agreement with this distribution, Agel et al found 9.6% of injuries effected goaltenders, 40.8% were defensemen, and 48.3% were forwards. This information also correlated with a recent study investigating injury incidence during the World Championship and Olympic tournaments, where goaltenders were once again the least injured followed by defensemen and forwards. However, when evaluating injuries in the National Hockey League (NHL), the only significant difference found was a higher likelihood of defensemen missing gametime due to injury in comparison to forwards. This study also found that a significant predictor of missing at least 5 games due to injury included being a goaltender (odds ratio = 1.68). So, while goaltenders do not get injured as often, their return to play is often more extended in comparison to other players. These numbers are likely not as drastically position dependent due to the disparity of position players on the ice at one time (3 forwards, 2 defensemen, and 1 goaltender), which may account for a significant amount of the variance reported in the literature at the collegiate, international, and professional levels.

When do they get hurt?

Understanding when injuries occur during the course of an individual game and throughout the course of the season allows us to understand when an athlete is at an increased risk of injury. With regards to the collegiate level, 65.5% of injuries occur during games and 35.5% during practice. Additionally, preseason practice rates were more than twice as high as in season injury rates. Two studies looked into the distribution of injuries during game play and found drastically different results. Agel and colleagues found that most injuries occurred in the 2nd and 3rd periods, which is likely due to athlete fatigue and intensity of gameplay increasing as the game progresses. In contrast to this report, amongst NHL regular season injuries, the vast majority occur in the first period (48.1%). Intuitively, as the NHL season progresses, the likelihood of injury also increases. This increase in injury rate is likely due to player fatigue and increased intensity of play as teams are fighting for a playoff spot.

How do they get hurt?

Due to the physical nature of ice hockey, the vast majority of injuries are contact-related, however non-contact injuries tend to make up a higher proportion of practice injuries. According to Agel and colleagues, non-contact injuries make up 9.7% of injuries during games, whereas they make up 32% of injuries during practice. During games, approximately 50% of injuries are due to contact with another player and 39.6% are due to contact with another object (boards, puck, etc.). In agreement with these trends, in the NHL body checking made up 28.6% of injuries, while incidental contact (14.3%), hit by puck (13.5%), contact with environment (9.4%), and other intentional player contact (7.4%) made up the bulk of injuries incurred. Aside from these mechanisms, non-contact injuries made up 14.8% of all injuries and accounted for 11.7% of man games lost due to injury (1,921 games) between 2009 and 2012. This gives good insight into the general cause of injury, but what types of injury are most common amongst ice hockey players?

Over a 16 season timeframe, injures most often sustained during gameplay included internal derangement of the knee (13.5%), concussion (9.0%), acromioclavicular joint injury (8.9%), upper leg contusion (6.2%), and musculotendinous strain of the hip/groin region (4.5%). However, distribution of injuries during practice shows a slightly different distribution. The most common injuries during practice were musculotendinous strain of the hip/groin region (13.1%), internal derangement of the knee (10.1%), ankle ligament sprain (5.5%), concussion (5.3%), and acromioclavicular joint injury (4.4%). Looking further at this data, during games, the highest prevalence of severe injuries was knee internal derangement and the most common mechanism of injury was due to player contact. Of the most common severe injuries reported, musculotendinous injuries of the pelvis/hip was the only pathology (6.2%) that had a non-contact mechanism as the most common cause. In agreement with these findings, amongst NHL players, the most common sites of injury include the head (17%), thigh (14%), knee (13%), and shoulder (12%). Additionally, with regards to man games lost, these regions also comprised the largest impact on their respective teams. Unfortunately, this study did not break the body regions into specific injuries/pathologies.

Given the fairly vague description of the injuries reported in these large epidemiological studies, it is also important to look more in depth to determine which specific injuries are actually reported in the literature. With regards to knee ligamentous injuries, Sikka and colleagues found that between 2006 and 2010, only 47 players sustained an anterior cruciate ligament tear in the NHL, which is significantly lower than most professional contact team sports. These injuries included 3 goaltenders, 8 defensemen, and 36 forwards. Of these 47 injuries, the reported mechanism for all but one injury was contact with another player and/or with the boards. In addition to the primary ACL rupture, 68% of injuries reported a concomitant meniscal or medical collateral ligament injury as well. Looking at the more commonly injured MCL, from 2003-04 to 2010-11, 13 MCL injuries were reported within one collegiate ice hockey program. This resulted in ten different players being injured (12.7%) and an incident rate of 0.44 per 1,000 athlete exposures. Of these injuries, 77% were contact related and an acute non-contact injury was reported in 15% of cases.

Looking at the most common non-contact injury, hip/pelvic pathology has also been investigated in the literature for this population. Over a four year span, 890 hip or groin injuries were reported in the NHL. Of those reported, 10.6% were found to be intra-articular in nature. There was a very small difference between injury occurring during games (44.6%) and during practice (41.4%), but the vast majority occurred during the regular season (71.2%). The most frequent intra-articular hip diagnosis made in this cohort was hip labral tear (69.1%), followed by hip osteoarthritis (13.8%), hip loose body (6.3%), femoroactebular impingement (5.3%), other hip injury (3.1%), and hip chondromalacia (2.12%). With regards to player position, injuries per 1000 player-game appearances were significantly higher in goaltenders compared with all other players.

Why does it matter?

Having a successful team depends on multiple factors, but regardless of coaching or talent, injuries can have a significant impact on a team’s ability to win. Unfortunately, the NHL does not readily provide injury data to the general public, however has provided data supporting the negative impact of injuries on a team’s success. Over the past five seasons, the Stanley Cup champion has been among the top five least injured teams throughout the regular season. Along those same lines, this season’s President’s Trophy (best regular season record) winning Washington Capitals also had the least man games lost due to injury. Additionally, with the exception of the Pittsburgh Penguins, none of the top five most injured teams have had consistent success over the past five years. In fact, two of those included in the top five most injured also had the two of the worst overall records over this timeframe (Columbus Blue Jackets and Edmonton Oilers).

Screen Shot 2016-04-24 at 3.29.05 PM

In addition to the obvious impact on team performance, when a player returns from a significant injury, their productivity and durability has the potential to decline in the seasons following their return to play. The presence of a meniscal injury was associated with a decreased length of career for all positions. Furthermore, for wings and centers, the number of games played decreased in the first full season after ACL injury from 71.2 to 58.2 and in the second full season to 59.29. With regard to offensive production, there was a 31% reduction in goals scored per season, 60% reduction in assists, and 42% reduction in total points compared with an uninjured control group. Only 37.5% of players who were previously selected as All-Stars were able to regain this honor upon their return to play.

Finally, there is also a financial impact as well. According to Donaldson and colleagues, between the 2009-10 and 2011-12 seasons, 50.9% of NHL players missed at least one game, which translated to a total salary cost of $218 million per year.

With the substantial impact on team performance and the associated financial implications, understanding how to identify those at risk and develop programs to lessen the likelihood of injury are paramount to a successful organization. Part 2 of this series will delve into the evidence regarding injury risk assessment and prevention in this population of athletes.


1. Agel J, Dompier T, Dick R, Marshall S. Descriptive Epidemiology of Collegiate Men’s Ice Hockey Injuries: National Collegiate Athletic Association Injury Surveillance System, 1988–1989 Through 2003–2004. Journal of Athletic Training. 2007;42(2):241-248.

2. Flik K. American Collegiate Men’s Ice Hockey: An Analysis of Injuries. American Journal of Sports Medicine. 2005;33(2):183-187. doi:10.1177/0363546504267349.

3. Currier, Nathan. “The Most Injured NHL Teams Since the 2009-2010 Season.” N.p., 12 Apr. 2016. Web. 22 Apr. 2016.

4. Donaldson L, Li B, Cusimano MD. Economic burden of time lost due to injury in NHL hockey players. Inj Prev. 2014;20(5):347-349. doi:10.1136/injuryprev-2013-041016.

5. Epstein DM, McHugh M, Yorio M, Neri B. Intra-articular Hip Injuries in National Hockey League Players: A Descriptive Epidemiological Study. American Journal of Sports Medicine. 2013;41(2):343-348. doi:10.1177/0363546512467612.

6. Grant JA, Bedi A, Kurz J, Bancroft R, Miller BS. Incidence and Injury Characteristics of Medial Collateral Ligament Injuries in Male Collegiate Ice Hockey Players. Sports Health. 2013;5(3):270-272. doi:10.1177/1941738112473053.

7. McKay CD, Tufts RJ, Shaffer B, Meeuwisse WH. The epidemiology of professional ice hockey injuries: a prospective report of six NHL seasons. British Journal of Sports Medicine. 2014;48(1):57-62. doi:10.1136/bjsports-2013-092860.

8. Sikka R, Kurtenbach C, Steubs JT, Boyd JL, Nelson BJ. Anterior Cruciate Ligament Injuries in Professional Hockey Players. American Journal of Sports Medicine. 2016;44(2):378-383. doi:10.1177/0363546515616802.

9. Tuominen M, Stuart MJ, Aubry M, Kannus P, Parkkari J. Injuries in men’s international ice hockey: a 7-year study of the International Ice Hockey Federation Adult World Championship Tournaments and Olympic Winter Games. British Journal of Sports Medicine. 2015;49(1):30-36. doi:10.1136/bjsports-2014-093688.

VMO: An Update

The following is an article written for the online, video-based physical therapy continuing education company MedBridge

Patellofemoral pain syndrome (PFPS) is among the most common sports injuries and yet the current treatment protocols are not optimal. In particular, the latest research questions our ability to selectively recruit vastus medialis obliquus (VMO) and affect its timing and suggests that VMO may have nothing to do with the PFPS.

Prevalence of Patellofemoral Pain Syndrome

PFPS has an astounding prevalence according to a retrospective case-control analysis by Taunton et al. They analyzed 2,002 running-related injuries seen at a primary care sports injury facility – and 42.1% (842/2,002) were knee injuries. Of these knee injuries, 39.3% (331/842) were due to patellofemoral pain syndrome (PFPS), which made PFPS by far the most common diagnosis found in this large-scale study.

Similarly, in an older study (1984), Devereaux et al found that over a five year period, 137 patients presented with PFPS – accounting for 25% of all knee injuries seen at this sports injury clinic.

These two studies were conducted 17 years apart, giving support to the consistently high prevalence of this disorder. Now the most important question is, how are we treating these patients?

How are we treating patients with PFPS?

Imbalance of LV and VMO

The biomechanical study by Lieb et al made the VMO the mainstay of most physical therapy protocols and treatment approaches for PFPS.

Conducted in 1968, the study found that VMO’s fibers are oriented at 55° from the longitudinal axis of the femur. This orientation alone makes VMO the primary restraint to lateral subluxation of the patella. The study further postulated that VMO was able to counterballance the pull of the much larger vastus lateralis (VL) due to the discrepancy in mechanical advantages.

As a result of this study, an insufficient balance between the VL and VMO has long been considered the primary contributing factor in developing patellar subluxation, or maltracking.

VL:VMO timing

A study by Cowan et al found that subjects with PFPS have an imbalance of VL:VMO timing. The VL typically begins to fire approximately 15-20 ms prior to the VMO.

Due to this understanding of the biomechanics, the treatment strategy typically involved correcting the potential VMO atrophy, hypoplasia, inhibition, and impaired motor control.

Now, this all seems logical in theory, but can we actually selectively train the VMO? Is this relatively small muscle affected differently than the rest of the quadriceps musculature in the presence of pain?

New research questions…

Does VMO atrophy relative to the rest of quadriceps?

A recent study by Giles and colleagues refutes one of the cornerstones of the VMO theory; namely, that this small muscle tends to atrophy relative to the rest of the quadriceps during or following surgery. They performed a cross-sectional study of 35 participants diagnosed with PFPS.

The results showed atrophy of all portions of the quadriceps muscles – and no selective atrophy of the VMO – present in the affected limb of people with unilateral PFPS.

Can we selectively activate VMO?

Even if the atrophy was present, the literature is not very kind to the ability to preferentially activate this musculature either.

Cerny et al evaluated the ability to preferentially recruit the VMO during 22 different quadriceps exercises. Through electromyographic analysis, they determined that VMO activity was not higher in any exercise compared to VL.

In a randomized controlled trial, Song et al found no evidence that VMO can be activated separately. They compared the change in VMO cross-sectional area after 8 weeks of unilateral leg press and unilateral leg press with subsequent hip adduction. The two groups showed no significant difference between the change in VMO cross-sectional area (the standard leg press actually yielded better results).

Thus, selective isolation of the VMO in everyday clinical practice is highly unlikely. In all reality, if we consider inability to selectively recruit their target, most of the ‘VMO programs’ are merely strengthening the quadripceps as a whole. If we could selectively recruit these fibers, according to Grabiner et al, it would take approximately 60% of maximal voluntary contraction to stimulate hypertrophy of the VMO.

Does VMO training have advantages over general quadriceps strengthening?

In 2010, Bennell et al investigated how VMO retraining compares to a general quadriceps strengthening program in relation to vasti onset.

The VMO retraining group used EMG biofeedback during the following series of exercises:

  • isometric VMO contractions at 90° of knee flexion
  • standing mini squats to 40°
  • isometric contraction of the VMO in combination with hip abduction and hip external rotation during an isometric wall contraction in standing
  • step downs

The quadriceps group performed:

  • isometric quad sets
  • straight-leg raises
  • SAQs
  • side-lying hip abduction

At the conclusion of the training programs, the retraining group actually did create more significant changes in stair descent activation in the short-term. However, at the 8-week follow-up, both values were nearly identical. The initial improvement may have been due to the use of ‘step downs’ in the retraining group, which most closely simulates the functional and muscular demands of stair descent.

During stair ascent, on the other hand, the quadriceps strengthening group caused a much more significant alteration in VMO:VL timing and was the only group that caused the VMO to fire prior to the VL.

A study conducted by Laprade et al showed similar results using isometric exercise. This study compared the EMG activity in individuals with PFPS and asymptomatic controls during 5 isometric exercises. There was no significant difference in the ratio of VMO:VL firing between the two groups.

Given these results, I find it hard to support the use of VMO training in everyday clinical practice.

Does VMO strengthening make a difference for PFPS?

Suppose it was possible to selectively recruit the VMO. Would it reduce patellofemoral contact stress sufficiently to relieve the pain?

Sawatsky et al say no. They conducted a biomechanical study using New Zealand white rabbits. Although this is not a direct human study, the muscular alignment and pull of the quadriceps is very similar: the fibers of the VMO and VL are oriented at 45-50° and 14-19°, respectively, in relation to the longitudinal axis of the femur.

They transected VMO at varying levels of knee flexion (30°, 60°, and 90°), measuring patellofemoral joint contact pressures before and after the transections.

There were no significant differences between peak pressures, average pressures, contact areas, or contact shapes before and after transection.

If the contact area and pressure are not altered when the muscle is removed from the equation, then why do we continue thinking VMO training is the gold standard in PFPS treatment?

Should we stop blaming the glutes for everything?

Below is an article written for… 

Anterior cruciate ligament (ACL) rupture1,2 and patellofemoral pain syndrome (PFPS)3,4,5 are two of the most common lower extremity complaints that physicians or physical therapists will encounter. In addition to the high incidence of these pathologies, with regards to ACL injury, very high ipsilateral re-injury and contralateral injury have also been reported6,7,8. With the importance of treating and/or preventing these injuries, several researchers have taken it upon themselves to determine what movement patterns predispose athletes to developing these injuries. This research indicates that greater knee abduction moments9,10, peak hip internal rotation11 and hip adduction motion12 are risk factors for PFPS development. Whereas, for ACL injury, Hewett and colleagues13 conducted a prospective cohort study identifying Knee abduction angle at landing as predictive of injury status with 73% specificity and 78% sensitivity. Furthermore, as the risk factors for developing both disorders are eerily similar, Myer et al performed a similar prospective cohort study finding that athletes demonstrating >25 Nm of knee abduction load during landing are at increased risk for both PFPS and ACL injury14. With a fairly robust amount of research supporting a hip etiology in the development of these injuries, it would make sense that weakness of the hip musculature would also be a risk factor, right?

A recent systematic review found very conflicting findings on the topic. With regards to cross-sectional research, the findings were very favorable with moderate level evidence indicating lower isometric hip abduction strength with a small effect size (ES) and lower hip extension strength with a small ES15. Additionally, there was a trend toward lower isometric hip external rotation and moderate evidence indicates lower eccentric hip external rotation strength with a medium ES in individuals with PFPS15. Unfortunately, the often more influential prospective evidence told a different story. Moderate-to-strong evidence from three high quality studies found no association between lower isometric strength in hip abduction, extension, external rotation or internal rotation and the risk of developing PFPS15. The findings of this systematic review indicated hip weakness may be a potential consequence of PFPS, rather than the cause. This may be due to disuse or fear avoidance behaviors secondary to the presence of anterior knee pain. Regardless of its place as a cause or consequence, hip strengthening has proved beneficial in patients with both PFPS16,17,18 and following ACL Reconstruction19, but does it actually help to change the faulty movement patterns?

Gluteal strengthening can cause several favorable outcomes, from improved quality of life to decreased pain, unfortunately however marked changes in biomechanics is not one of the benefits. Ferber and colleagues20 performed a cohort study analyzing the impact of proximal muscle strengthening on lower extremity biomechanics and found no significant effect on two dimensional peak knee abduction angle. In slight contrast however, Earl and Hoch21 found a reduction in peak internal knee abduction moment following a rehabilitation program including proximal strengthening, but no significant change in knee abduction range of motion was found. It should be noted that this study included strengthening of all proximal musculature and balance training, so its hard to conclude that the results were due to the strengthening program. Potentially, gluteal endurance may be more influential than strength itself, so it would make sense that following isolated fatigue of this musculature, lower extremity movement patterns would deteriorate.

Once again, this belief is in contrast to the available evidence. While fatigue itself most definitely has an impact on lower extremity quality of movement, isolated fatigue of the gluteal musculature tells a different story. Following a hip abductor fatigue protocol, patients only demonstrated less than a one degree increase in hip-abduction angle at initial contact and knee-abduction angle at 60 milliseconds after contact during single-leg landings22. In agreement with these findings, Geiser and colleagues performed a similar hip abductor fatigue protocol and found very small alterations in frontal plane knee mechanics, which would likely have very little impact on injury risk23. The biomechanical explanation for why weakness or motor control deficits in the gluteal musculature SHOULD cause diminished movement quality makes complete sense, but unfortunately, the evidence at this time does not agree.

While the evidence itself does not allow the gluteal musculature to shoulder all of the blame, this does not mean we should abandon addressing these deficits in our patients. As previously stated, posterolateral hip strengthening has multiple benefits, but it is not the end-all-be-all for rehabilitation or injury prevention of lower extremity conditions. Proximal strength deficits should be assessed through validated functional testing in order to see its actual impact on lower extremity biomechanics on a patient-by-patient basis. Following this assessment, interventions should be focused on improving proximal stability, movement re-education, proprioception, fear avoidance beliefs, graded exposure, and the patient’s own values, beliefs, and expectations.

Research Review: Immediate Effects of Real-Time Feedback on Jump-Landing Kinematics

In the next instalment of the Research Review Series, we discuss the impact of real-time feedback in addition to post-response feedback compared to post-response feedback alone on jumping-landing mechanics in a young female population1

Study Design

Randomized controlled trial.


Thirty-six pain-free females were recruited from the general student population at the University of Toledo and assigned to either the real-time feedback (RTF), RTF plus post-response feed-back (RTF+), or the no feedback control group (CG).

Inclusion Criteria: (1) Female gender, (2) No current or previous lower extremity musculoskeletal complaints

Exclusion Criteria: (1) Male gender, (2) previous history of fracture, surgery, or significant orthopaedic injury to the lower extremity


Outcome Measures: Kinematic and kinetic data were collected as participants performed 3 trials of a jump-landing task from a 30-cm box, positioned at a horizontal distance of 50% of the participant’s height from 2 force platforms. Sagittal plane moments and angles at the knee and hip, frontal plane angles at the knee, and vertical ground reaction forces during the jump-landing task were quantified at baseline and post-intervention.

Randomization: Block randomization was used with concealed allocation to assign participants to 1 of the 3 groups. An opaque envelope was used to conceal group assignment until after baseline testing.

Interventions: Prior to the intervention, both the RTF+ and PRF feedback groups were presented with a PowerPoint presentation explaining the goals of the jump-landing task. This presentation outlined the need to (1) land with both feet at the same time, (2) land in neutral valgus/varus position, (3) land with feet shoulder width apart, (4) land on toes and rock onto heels, (5) land with increased bending at the knees and hips, and (6) land softly. After viewing the presentation, participants in both intervention groups performed 3 sets of 6 jumps from the box. Following each set of 6 jumps, the investigator reviewed the goals that the participant failed to accomplish in the previous jumps and showed the participant the corresponding PowerPoint slides to reinforce the correct form.

For the RTF+ group, participants received a live, digital representation of their body segments and were able to see a reference line to assist in making biomechanical corrections in the frontal plane. The RTF+ group was provided the following explanation,

You will now be able to see markers representing your knee and toe on the screen in real time; start with your toe marker in line with the reference line and then line your knee marker up with the reference line. This is the way the markers should line up when you land; we want you to watch the video monitor, focusing on keeping the shank segment in line with the reference line when you land from your jump. You can aim to land with your foot on the tape line, but your main focus should be to keep the shank segment lined up with the line when landing from the jump.

Participants in the no-feedback control group performed the same jump landing sequence as the other two groups, however they received no feedback or PowerPoint presentation on the major goals of jump landing.


Post hoc testing revealed that the RTF+ and PRF groups had a greater increase in knee flexion, hip flexion, and greater decrease with regards to vertical ground reaction forces compared to the control group, but no differences were found between the two groups. Neither the RTF+ or PRF groups demonstrated a significant change in the knee extensor moment, hip extensor moment, or knee abduction angle post-intervention.


This study had several significant limitations that may have impaired the impact of the interventions provided. First of all, there was no “pre-screening” for excessive valgus and due to this fact, the included subjects had very low knee abduction angles at baseline. This limited the ability to make any kind of meaningful impact post-intervention. This is where utilizing a simple lower extremity functional test would have aided in providing a better patient population. Additionally, this study only measured immediate impact of the interventions without looking into long-term retention of the movement pattern, which would be more important in injury risk reduction. With a fairly small, albeit adequately powered, sample-size, the statistical significance of some of the changes may have been hampered.

Clinical Implications

Anterior cuciate ligament injuries have significantly increased from 40.9 to 47.8 per 10,000 patients according to a study looking into trends and demographics of ACL injuries in the United States4. This rising incidence coupled with very high re-injury rates to the contralateral and ipsilateral limb3,5,6 make developing injury risk reduction programs paramount. Due to its importance, several researchers have investigated faulty movement patterns and the incidence of ACL injury. These studies have identified prospective evidence linking decreased knee flexion and increased knee abduction angles as predictive of future injury2. While the aforementioned study did not produce a change to the knee abduction angle (likely due to a ceiling effect), there was a significant change in knee flexion angle, which may aide in decreasing the likelihood of future injury. While this is promising preliminary evidence, in order to have a significant impact, an injury prevention program must be repetitive and participants must have good compliance7,8 to create a long-term change in movement pattern. This study, due to its design, does not capture this aspect.


1. Ericksen HM, Thomas AC, Gribble PA, Doebel SC, Pietrosimone BG. Immediate Effects of Real-Time Feedback on Jump-Landing Kinematics. Journal of Orthopaedic & Sports Physical Therapy. 2015; 45(2): 112–118. doi:10.2519/jospt.2015.4997.

2. Hewett TE, et al. Biomechanical Measures of Neuromuscular Control and Valgus Loading of the Knee Predict Anterior Cruciate Ligament Injury Risk in Female Athletes: A Prospective Study. American Journal of Sports Medicine. 2005;33(4):492–501. doi:10.1177/0363546504269591.

3. Kamath GV. Anterior Cruciate Ligament Injury, Return to Play, and Reinjury in the Elite Collegiate Athlete: Analysis of an NCAA Division I Cohort. Am J Sports Med. 2014;42(7):1638–1643. doi:10.1177/0363546514530866.

4. Leathers MP, Merz A, Wong J, Scott T, Wang JC, Hame SL. Trends and Demographics in Anterior Cruciate Ligament Reconstruction in the United States. Journal of Knee Surgery. 2015. [Epub ahead of print]

5. Paterno MV, Rauh MJ, Schmitt LC, Ford KR, Hewett TE. Incidence of Second ACL Injuries 2 Years After Primary ACL Reconstruction and Return to Sport. Am J Sports Med. 2014;42(7):1567–1573. doi:10.1177/0363546514530088.

6. Rugg CM, Wang D, Sulzicki P, Hame SL. Effects of prior knee surgery on subsequent injury, imaging, and surgery in NCAA collegiate athletes. Am J Sports Med. 2014;42(4):959–964. doi:10.1177/0363546513519951.

7. Sugimoto D, Myer GD, Bush HM, Hewett TE. Effects of Compliance on Trunk and Hip Integrative Neuromuscular Training on Hip Abductor Strength in Female Athletes. Journal of Strength and Conditioning Research. 2014;28(5):1187–1194. doi:10.1097/JSC.0000000000000228.

8. Sugimoto D, Myer GD, Bush HM, Klugman MF, McKeon JMM, Hewett TE. Compliance With Neuromuscular Training and Anterior Cruciate Ligament Injury Risk Reduction in Female Athletes: A Meta-Analysis. Journal of Athletic Training. 2012;47(6):714–723. doi:10.4085/1062-6050-47.6.10.

What Are We Missing? The Influence of Fatigue.

The following is another article written for the online, video-based physical therapy continuing education company MedBridge

Recently, a lot of attention has been paid to re-injury and return to sport following anterior cruciate ligament reconstruction (ACLR) and the results continue to be less than exceptional. A recent case series of elite collegiate athletes who suffered ACL injuries prior to and during their college careers continually found difficulty returning to sports participation (Kamath et al., 2014). Of the 35 athletes who had undergone ACLR prior to enrollment in college, the rate of re-operation on the involved limb was 51.4%, the rate of re-rupture of the ACL graft was 17.4%, and contralateral ACL rupture was 20.0% within this population of athletes. Similarly, those who underwent ACLR during college had a 20.4% re-operation rate, 1.9% suffered re-rupture of the ACL graft, and 11.1% of these athletes underwent ACLR on the contralateral limb. In agreement with these findings, a prospective cohort study of 456 collegiate athletes conducted by Rugg and colleagues found that athletes entering college with a history of ACLR had a 892.9-fold increase in knee surgery compared to those who entered college without undergoing surgery. Unfortunately, these findings are not isolated to collegiate athletes as professional (Busfield et al., 2009) and high school athletes (McCullough et al., 2012) alike have similar statistics. Considering these numbers, it points to inadequate or premature return to athletic participation, which may be because we are overlooking a very important aspect of athletic competition.

First, let’s look at what factors have been shown to predispose these athletes to injury. Hewett et al conducted a prospective cohort study that identified factors that may put athletes at risk for initial ACL injury. After screening 205 female collegiate athletes with a drop-jump task, 9 athletes went on to suffer an ACL injury during the following season. These 9 athletes had several important factors in common in comparison to those who did not go on to suffer injury. Knee abduction angle at landing was 8° greater, knee abduction moment was 2.5 times greater, and there was a 20% higher ground reaction force in ACL–injured than in uninjured athletes. More importantly, the authors determined that injury could be predicted in those with an increased knee abduction moment (dynamic valgus) with 73% specificity and 78% sensitivity.

Prior to return to sport, many athletes will undergo functional testing (hop testing, Y-Balance Test, etc.), but do these tests, done under optimal circumstances, tell the full story or are we missing something?

Fatigue has been shown repeatedly to have negative affects on lower extremity biomechanics. A systematic review recently examined the literature pertaining to lower extremity biomechanics and neuromuscular fatigue during single-leg landings (Santamaria et al., 2010). After analyzing 8 studies and 141 total subjects, kinematic data revealed greater knee and hip flexion and increased dorsiflexion post-fatigue. More importantly, following the introduction of fatigue, there was no change in peak knee valgus angles. However, as anticipated/practiced drop-landings are performed primarily in the sagittal plane, these specific procedures may not be sufficient to determine movement patterns during athletic competition. When an unanticipated landing was used, the results were drastically different with a significant increase in peak knee valgus angle post-fatigue compared to pre-fatigue. This unanticipated landing would seem to represent the demands of athletic competition more accurately and thus demonstrates an increased risk of injury with neuromuscular fatigue. In agreement with these findings, Brazen et al found no change in frontal plane biomechanics during an anticipated drop-landing task after neuromuscular fatigue, however they did find a higher anterior-posterior time to stabilization (TTS) and vertical TTS, which once more increases the likelihood of injury.

More specific to patients following ACLR, Webster et al conducted a study comparing the response to neuromuscular fatigue between uninjured control subjects and athletes following ACLR. This study once again utilized an anticipated drop-landing task with data collected pre and post fatigue. Fatigue led to reduced flexion in the lower limb, increased hip and knee abduction, increased knee rotation, and reduced knee joint moments. The response to fatigue was similar with no significant differences between the ACL-reconstructed limb and the control group as well as the reconstructed limb and the contralateral limb. To further investigate the lower extremity biomechanics of athletes following ACLR, the Lower Extremity Error Scoring System (LESS) was developed. Padua et al determined the LESS to be a valid and reliable tool in assessing jump-landing biomechanics with good inter-rater reliability (ICC= 0.84) and excellent intra-rater reliability (ICC= 0.91). This evaluation tool involves counting the number of faulty movement patterns during a jump-landing task with < 4 errors being an excellent score, ≤ 5 being good, ≤ 6 being moderate, and > 6 being poor landing mechanics. When evaluating the influence of fatigue on LESS scores, Gokeler et al found significant differences between patients status-post ACLR and uninjured control subjects. The initial median score pre-fatigue for ACLR patients was 6.5 (poor) and 7.0 following fatigue, whereas the uninjured control subjects scored 2.5 (excellent) pre-fatigue and drastically increased to 6.0 (poor) post-fatigue. This shows an obvious decline in movement quality following fatigue, which may place both post-ACLR patients and uninjured controls at risk for injury.

Fatigue is an often-neglected aspect in the decision to return an athlete to sport or to assess an athlete’s initial risk for injury. This data should be used to further evolve our testing procedures to account for these potentially injurious movement patterns secondary to neuromuscular fatigue. Trent Nessler, DPT developed a fatigue protocol and concomitant testing procedure for return to sport and injury risk assessment purposes as part his Dynamic Movement Assessment. Albert Einstein was quoted saying, “The definition of insanity is to continually do the same thing over and over expecting a different result”. If we are to improve these return to sport and re-injury numbers, fatigue cannot be overlooked anymore and must be included in our clinical decision making process.

Effectiveness of Manual Therapy for Knee Pain

The following is another article written for the online, video-based physical therapy continuing education company MedBridge Education

Knee osteoarthritis (Knee OA) is one of the most prevalent and debilitating orthopedic complaints for 28% of adults over 45 years old and 37% of those over 65 years old in the United States. In addition, 1.6% of adults over the age of 60 have undergone total knee arthroplasty (Dillon et al). Improving the underlying mobility, strength, pain, and functional limitations associated with this pathology is a critical component of patient care. There are a number of interventions employed by physical therapists for individuals suffering from knee pain – some more effective than others. Amongst one of the more common, albeit controversial, is the use of manual therapy, or more specifically, joint mobilization.

In 2000, Deyle et al published an initial investigation into the potential effectiveness of manual therapy techniques in combination with exercise in the treatment of knee OA. In this randomized controlled trial, patients in the intervention group received manual therapy techniques based on their specific impairments, which potentially included passive physiologic and accessory joint movements, muscle stretching, and soft-tissue mobilization, applied primarily to the knee. However, if any additional deficits were found in other regions (i.e. hip, foot/ankle, lumbar spine), manual therapy techniques were directed at these areas. At the completion of the study, the intervention group achieved significant improvements in 6-minute walk distance and WOMAC score at 4 weeks and 8 weeks. Additionally, only 5% of those in the intervention group underwent total knee arthroplasty (TKA) in comparison to 20% in the control group. This study was a great first step, however as the control group only received subtherapeutic ultrasound for 10 minutes to the area of knee symptoms, further investigation was warranted.

Later in 2005, Deyle et al conducted a similar study with a control group, which included a standardized home exercise program. In this study, patients in the intervention group received 8 sessions of manual therapy treatment, which consisted of passive physiological and accessory movements, manual muscle stretching, and soft-tissue mobilization. These techniques were primarily applied to structures in the knee region opposed to the holistic approach previously used in the 2000 study. At the 4 week follow-up, WOMAC scores had improved by 52% in the clinic treatment group compared to 26% in the home exercise group, whereas both groups improved by approximately 10% in their 6-minute walk distances. Additionally, at the one-year follow-up, there was no significant difference between groups in either measure. This study gives credence to short-term functional improvements for manual therapy techniques, but not necessarily walking speed or capacity. While this does offer some evidence to support the inclusion of manual therapy, it also puts into question whether a home exercise program is an adequate comparison group.

More recently, Abbott et al conducted a randomized controlled trial comparing manual therapy, exercise, and combined manual therapy and exercise, and a usual care group in the treatment of hip and knee OA. The findings of this study were interesting, at the one year follow-up, both the manual therapy and exercise groups achieved statistically significant improvements with regards to reduction in WOMAC scores. Whereas, combined manual therapy and exercise did not meet this same significant improvement. Along with these findings, following the intention to treat analysis, all intervention groups improved but only usual care plus manual therapy and usual care plus exercise therapy achieved clinically significant reductions of >28 WOMAC points from baseline. Once again, manual therapy and exercise plus usual care improved, but did not meet the 28-point improvement threshold. In a secondary analysis of this trial by Pinto et al, it was determined that within the New Zealand healthcare system, both manual therapy and exercise offer a significant cost savings over usual care for OA treatment.

With this recent research, there does seem to be a fairly significant benefit to the utilization of manual therapy, however a multi-modal program consisting of manual therapy and exercise seems to be less effective than manual therapy in isolation. It should be taken into consideration that this conclusion was derived from one study and may not be a true representation of the patient population as a whole. In agreement with the benefits of manual therapy found by Abbott et al, a systematic review published by Jansen et al found a greater effect size with manual therapy and exercise (0.69) in comparison to either exercise therapy (0.38) or strength training (0.34) in isolation. Additionally, recent works by Rhon et al and Ko et al found significant increases in proprioception and functional performance when manual therapy was combined with exercise and perturbation exercises, respectively.

In addition to the varying degrees of effectiveness found in the aforementioned studies, it must be considered that not every patient with knee OA will respond similarly to any given therapeutic intervention. In order to help delineate those patients with knee OA who will respond favorably to hip mobilization, Currier et al proposed a Clinical Prediction Rule (CPR) to make this distinction. This particular study found that of those individuals who 2+ of the 5 variables present, following hip mobilization, the positive likelihood ratio was 12.9 and probability of success was 97% (success defined as a decrease of at least 30% on composite Numerical Pain Rating Scale score obtained during functional tests or a Global Rating of Change Scale score of at least 3). This CPR should be used with caution, however, as no validation study has been conducted to this date. While this CPR provides some idea as to which patients will respond favorably to manual therapy interventions, it should be understood that this decision must be made in conjunction with sound clinical reasoning following a thorough patient history and physical examination.

Alexis Wright, PT, PhD, DPT, FAAOMPT goes into great detail with regards to evidence-based decision making when deciding whether joint mobilization or manipulation will benefit your patient in her course, “Evidence-Based Examination of the Hip“. So, while the research is far from definitive regarding this specific intervention, manual techniques do appear to provide significant improvements in proprioceptive capacity, perceived physical disability, and pain levels for patients presenting with knee osteoarthritis.

Research Review: Manual Physical Therapy and Perturbation Exercises in Knee Osteoarthritis

Rhon et al, 2013

Study Design

Prospective, observational cohort study.


Fifteen participants (7 male, 8 female) with a mean age of 55 years old were recruited from a convenience sample of consecutive patients evaluated for knee osteoarthritis (OA) at the Physical Therapy Clinic, Brooke Army Medical Center, San Antonio, Texas. With regards to severity, ten patients had bilateral symptoms, all 15 patients had radiographic signs of knee OA, and 10 had visible boney enlargement of the knee joint. Additionally, four of the included patients were active duty military personnel.

Inclusion Criteria: Utilizing criteria proposed by Altman (1991) and Altman et al (1986), the participants were included if they met at least one of the following three clinical clusters.

1. Knee pain for most days of the prior month: AND Crepitus with active motion and Morning stiffness in knee 38 years
2. Knee pain for most days of the prior month: AND Crepitus with active motion and Morning stiffness in knee > 30 minutes and Bony enlargement
3. Knee pain for most days of the prior month: AND No crepitus and Bony enlargement

Additional inclusion criteria include being eligible for care in a military medical treatment facility, minimum age 38 years old, and the ability to read, write, and speak sufficient English to complete the outcome tools.

Exclusion Criteria: Only periarticular pain or pain referred from another region (no joint pain), injections to the knee within the last 30 days, history of knee joint replacement surgery on involved limb, evidence of other systemic rheumatic condition (lupus, rheumatoid arthritis, psoriasis, or gout), and balance deficits from other non-musculoskeletal conditions (neurologic impairments, diabetic neuropathy, cerebellar disorders, or Parkinson disease)


Outcome Measures: The Western Ontario and McMaster Universities arthritis index (WOMAC), Global Rating of Change (GROC), Functional Squat Test (FST) evaluated with numerical pain rating scale (NPRS) and range of motion (ROM), and the Step-Up Test (SUT). Additionally, tolerance to treatment was determined by asking the participants a series of questions regarding whether their symptoms had gotten significantly worse at five different time points since their last visit. Time points included were immediately following treatment, several hours following treatment, that evening prior to bed, the following morning, and from the following morning until the follow-up (approximately 72 hours later).

Evaluation: The initial evaluation included a detailed history, review of systems, and physical examination. The history included questions regarding the duration, severity, location, and distribution of symptoms. The physical examination included functional tests, palpation of bony landmarks, ROM measurement, muscle length tests, and manual assessment of the joints and soft tissues of the lower extremity.

Interventions: Each patient was treated two times per week for four weeks and received both manual therapy and perturbation interventions. Visits included joint and soft-tissue mobilization, which was supplemented with stretching, ROM, and strengthening exercises. Additionally, each patient was provided with a home exercise program targeting their specific functional limitations. The manual therapy techniques were tailored to the impairments of each individual patient, however these interventions included varying grades of knee flexion, knee extension, and patella mobilizations. With regards to perturbation training, each patient was progressed based upon clinical reasoning and as tolerated by the individual patient. Each program generally started with more emphasis on manual therapy interventions and towards the end of the program, the focus switched to perturbation exercises.


WOMAC: The mean WOMAC score demonstrated a statistically significant improvement from baseline to 6 months with a 46% improvement, which was well above the minimal clinically important difference (MCID) of 12%. Additionally, the total WOMAC score was significantly improved at the end of the 4 week intervention period and remained improved at the 6 month follow-up. Finally, the only WOMAC sub-scale that did not remain improved at the 6 month follow-up was the ‘Stiffness’ sub-scale.

GROC: At the one month follow-up, 87% of patients reached the 3 point change in GROC to identify a clinically important change. Changes decreased over time with 80% of patients still maintaining this threshold of change at 3 months and only 60% at the final 6 month follow-up. Additionally, and probably more importantly, 47% of patients met the threshold for ‘dramatic change’ (GROC > 6) at all time points.

FST: Following the 4 week intervention period, statistically significant improvements in NPRS and ROM during the FST were documented. An average decrease from 5 to 3 on the NPRS and an improvement from 29° to 35° with regards to ROM.

SUT: The Step-Up Test values also significantly improved at the 4 week evaluation with a mean improvement of 4-5 steps during the 15 second test. This translated to an average increase from 9 to 14 steps completed during the test.


Due to the prospective cohort design of this study, no comparison group was included, thus no cause and effect relationship can be identified. Additionally, some of the improvements seen in this study could be attributed to other medical treatment many of the patients received. By 6 months five patients had received knee joint injections of either corticosteroid or viscosupplementation and two of those same patients received arthroscopic surgery. Of these patients receiving either injection or arthroscopic surgery, none reported improvement in symptoms immediately following the aforementioned procedures. Pain medication was used by 12 patients initially (10 patients daily; 2 patients as needed), including non-steroidal anti-inflammatory drugs and/or acetaminophen. However, it should be pointed out that at each of the follow-up points, fewer patients were taking medications than at baseline.

Clinical Implications

While no cause and effect relationship can be determined, this study does demonstrate theoretical effectiveness of a combined manual therapy and perturbation training approach to the treatment of knee osteoarthritis. This approach was associated with significant improvements in pain, function, and balance measures. There were several limitations evident within the study, however the potential positive impact of the interventions provided add to the current literature supporting perturbation and manual therapy techniques for patients suffering from knee osteoarthritis.

Rhon D, et al. Manual physical therapy and perturbation exercises in knee osteoarthritis. Journal of Manual & Manipulative Therapy. 2013; 21(4): 220–228.