Effects of Football Collars on Cervical Hyperextension and Lateral Flexion

Abstract

Objectives:

To evaluate the effectiveness of 3 football collars in reducing cervical range of motion.

Design and Setting:

A repeated-measures design in a controlled laboratory setting.

Subjects:

Fifteen male National Collegiate Athletic Association Division I varsity football athletes.

Measurements:

Cervical hyperextension and lateral flexion were measured with video analysis. Subjects underwent 5 testing conditions: standard football helmet, standard helmet and shoulder pads, and standard pads with the addition of the Cowboy Collar, A-Force Neck Collar, or a foam neck roll. Subjects performed motions both actively and passively.

Results:

All 3 collars reduced hyperextension when compared with the helmet and shoulder pads alone (P < .05); in addition, the Cowboy Collar was superior to the foam neck roll (P < .05) in reducing hyperextension. No collar reduced passive lateral flexion when compared with the helmet and shoulder pads, but the foam neck roll permitted significantly less active lateral flexion (P < .01) than the other 3 brace conditions.

Conclusions:

In a laboratory setting, cervical hyperextension can be controlled through the use of various cervical collars. Cervical lateral flexion (a more common cause of burners in a scholastic population) cannot be controlled with any of the cervical collars tested. Moreover, foam collars may impede active lateral flexion while not providing additional protection when loaded. These results are limited in that they were produced in a controlled situation as opposed to active football play.

Brachial plexus injuries, or burners, are characterized by transient sensory aberrations radiating from the neck and shoulder region with an associated loss of motor function.¹ Common mechanisms of injury in football include excessive cervical hyperextension or lateral flexion or direct compression at the Erb point. The incidence of injury is higher among collegiate athletes than high school athletes.² Repeated episodes can have lasting effects on upper extremity neural function,³ proprioception, and strength.⁴ Football collars are designed to limit the extremes of cervical motion, thus reducing the incidence of brachial plexus injuries. However, this practice is based almost exclusively on empirical data.^5,6 Currently, objective data regarding the ability of these collars to restrict cervical motion and thus prevent the mechanism of burners are inadequate.

Athletes with burners present with lancinating pain and nondermatomal paresthesia that radiates from the shoulder into the arm and may be associated with a loss of motor function in the affected limb.⁷ Clinical weakness in the biceps, spinati, and deltoid muscles may not be apparent initially but can develop hours to days later.⁸ In a follow-up study, athletes with a history of burners had abnormal electromyographic readings and objective strength deficits in 80% and 37% of subjects, respectively.³

Mechanisms of injury are varied but include cervical flexion away from the affected limb, thus placing traction on the brachial plexus,¹ and hyperextension of the cervical spine, which may compress nerve roots exiting the intervertebral foramina.⁹ The injuries reported at different levels of competition may have different causes,^10,11 and the clinical significance of the burner may increase with each episode.⁴ With as many as 65% of collegiate players experiencing a burner at some point in their careers and a recurrence rate reaching 85%,¹² injury prevention is paramount.

Currently, the means of preventing burners fall into 3 categories: neck and shoulder conditioning, coaching of proper blocking and tackling techniques, and protective equipment.^12–15 The purpose of current protective equipment is to limit cervical hyperextension and lateral flexion, thereby reducing the mechanisms of injury. Although empirical data would support the use of football collars, scientific evidence supporting their effectiveness in preventing the mechanisms of brachial plexus injury is limited.^5,6 Our purpose was to evaluate the efficacy of football collars in reducing cervical hyperextension and lateral flexion in a population of National Collegiate Athletic Association Division I collegiate football players by quantifying cervical motion under 5 test conditions in a controlled research setting.

METHODS

Research Design

We used a repeated-measures design. The independent variables were type of motion (active and passive) and brace condition, (a standard helmet, helmet and shoulder pads, and helmet and shoulder pads) with each of the following: Cowboy Collar (McDavid Knee Guard, Inc, Woodridge, IL), A-Force Neck Collar (Active Ankle Systems, Inc, Louisville, KY), or a 5.08-cm foam neck roll (Adams USA Inc, Monterey, TN). Both variables were treated as within-subjects variables. The dependent variables were cervical spinal motion (hyperextension and lateral flexion) in degrees. We counter-balanced directions of motion and the brace conditions within each subject.

Subjects

Fifteen male National Collegiate Athletic Association Division I varsity football athletes (height = 185 ± 5.99 cm, mass = 99.91 ± 17.68 kg, neck girth = 41.27 ± 2.69 cm, age = 19.67 ± 0.87 years) volunteered to participate in the study. The following exclusionary criteria for potential subjects were used: (1) symptoms of current neck or brachial plexus injury, (2) ranges of motion less than 90% of accepted norms using standard goniometry,¹⁶ or (3) positive Spurling sign.¹⁷ All participants completed an appropriate human subjects' consent form and health history questionnaire. The Temple University Institutional Review Board approved the study protocol.

Instrumentation

Helmet and Shoulder Pads

Subjects used the helmet (VSR-4, Riddell, Elyria, OH) and shoulder pads (Douglas Protective Equipment, Houston, TX) issued and fitted by the university equipment managers. The principal investigator (J.A.G.) evaluated the equipment and confirmed fit following accepted guidelines.¹⁸

Braces

The Cowboy Collar is a neck-roll system that combines a molded collar of polyethylene foam with a padded vest (Figure 1). The collar is worn under the shoulder pads and is designed to engage the sides and rear of the helmet and reduce extreme cervical motion.¹⁰ We used the large size in this study, which was appropriate for all subjects.

Figure 1.

Cervical collars. A, Cowboy Collar. B, A-Force Neck Collar. C, Foam neck roll.

The A-Force Neck Collar incorporates a molded cervical collar held in place by 2 straps that pass posteriorly under the axillae and fasten in the back with a plastic buckle (see Figure 1). The system is reinforced anteriorly by 2 horizontal straps. The restraint system is adjustable for exact fitting to each participant's chest size. This design fixes the equipment to the body, allowing it to function independently of shoulder-pad movement. The Pro size for players weighing more than 84 kg was used in the study.

The 5.08-cm foam neck roll is a traditional neck roll (see Figure 1) that was incorporated into the player's shoulder pads according to the manufacturer's instructions.

Chair

A Biodex Adjustable Accessory Chair (model 875-0190, Biodex Medical Systems, Shirley, NY) was used during data collection. The chair back was locked in a vertical position. The base of the chair contained a hydraulic lift, which was used to aid the vertical alignment of the subject within the data-collection field. The chair had a braking system, which permitted the chair to be stabilized to the floor.

Dynamometer

The MicroFet Hand-Held Dynamometer (Hoggan Health Industries, West Draper, UT) is a battery-operated and microprocessor-controlled transducer used to measure force application. The hand-held transducer weighs less than 0.5 kg and fits into an examiner's palm. It has a cantilevered arm with 3 strain gauges arranged to measure perpendicular as well as nonperpendicular force vectors simultaneously. The strain gauges measure movement in a single plane and transfer that coded movement to the microprocessor, where it is converted into pounds. The MicroFet was set at the low threshold setting, registering values in 0.09-kg (0.2-pound) increments. The MicroFet has a test range of 0.36 to 45.36 kg (0.8–100 pounds) and a liquid crystal display that shows force being applied, peak force applied at any point during a test, and the time of force application to 0.1 second. The MicroFet has been demonstrated as a reliable tool when used by a single examiner.¹⁹

Motion Analysis

We used a video camera (Panasonic model AG-456, Matsushita Electric Corp of America, Secaucus, NJ) to record the single-plane movements at a shutter speed of 60 Hz. The camera was centered to the subject and perpendicular to the plane of data collection in order to minimize parallax error in the 2-dimensional movements.²⁰ We hung a plumb line from the ceiling to establish a vertical along the background wall. A 400-cm line was marked on the floor as a continuation of the vertical on the wall. The floor line was bisected at right angles by a second 400-cm line parallel to the wall. The intersection of the floor lines identified the location of the subject, with 1 line denoting the plane of motion of the subject, whereas the other indicated the midline of the data-collection area. During testing, we placed the camera perpendicular to the plane of motion at a distance of 400 cm along the midline of the data-collection area. A level was then used to confirm that the camera was at a right angle in all planes. Reflective markers placed on the body were digitized so that cervical movement relative to the trunk could be tracked and measured by determining angular displacement.

Reliability

We established investigator reliability a priori for cervical spine range of motion and measurement protocols used in this study. Ten subjects of convenience, not associated with the study, underwent repeated testing using the protocols in the experimental phase of the study. The testing protocol was found to be reliable with intraclass correlation coefficient (3,k) values of .94 for active lateral flexion and hyperextension range and .75 and .94 for passive lateral flexion and hyperextension, respectively.

Testing Procedures

Subjects were assessed for evidence of exclusionary criteria. All subjects completed the medical history questionnaire, measurement of neck girth, and assessment of cervical spine range of motion. The order of joint movement and brace conditions was counterbalanced across subjects. For the analysis of lateral flexion, subjects were fitted with the appropriate brace condition and seated sideways in the test chair against a dark screen background and facing the camera. Subjects placed the left arm over the back support of the chair, and a 15.24-cm–wide hook-and-loop strap was secured around the torso and the chair back to stabilize the upper body to the chair. Reflective markers were placed on the subject's xiphoid process and suprasternal notch and midline of the chin, forehead, and top of the helmet. The plumb line was used to align all markers in the frontal plane; this served as the starting position for each repetition in all trials (Figure 2). Camera height was set at the level of the subject's suprasternal notch to standardize subject placement in the field of view.

Figure 2.

Starting position for lateral flexion.

For the analysis of hyperextension, subjects were fitted with the appropriate brace condition and were seated facing forward in the chair. The chair was rotated 90° counterclockwise so the right arm of the subject was toward the camera. The hook-and-loop fastener was once again secured around the body to stabilize it to the chair. Subjects folded their arms across their chests to increase visibility of markers placed along the mid-axillary line. The subjects were instructed to look at a mark at eye height while reflective markers were placed on the seventh and 10th ribs along the midaxillary line, 2 cm above the ear hole of the helmet, and the top of the helmet. The plumb line was used to align all markers in a sagittal plane; this served as the starting position for each repetition in all trials (Figure 3). Camera height was set at the height of the subject's shoulder to standardize subject placement in the field of view.

Figure 3.

Starting position for hyperextension.

All trials started with the markers aligned with the plumb line. Active-motion trials were performed by having the subject move his head in the appropriate direction. The subject gave a verbal cue when he reached the end of his available motion, at which point a visual marker was placed in the video field of view. Passive-motion trials were induced by the principal investigator (J.A.G.) using the MicroFet to standardize force application. For hyperextension, the subject was instructed to relax in an extended cervical position while a posteriorly directed overpressure of 133.5 N (30 pounds) was applied along the midsagittal line of the helmet just above the face mask (Figure 4). This value was determined during pilot testing as the maximum force that did not initiate discomfort. For lateral flexion, the subject relaxed in a laterally flexed position while a laterally directed overpressure was applied at a point 2 cm above the ear hole of the helmet (Figure 5). When the appropriate force was noted on the MicroFet, the examiner placed a visual marker in the video field of view. After every trial, the subject actively returned his head to the starting position. During both active and passive trials, the movements were repeated 3 times for a total of 30 lateral-flexion and 30 hyperextension movements.

Figure 4.

Force application for passive trial, hyperextension.

Figure 5.

Force application for passive trial, lateral flexion.

We used the Peak Motus 2000 software (Peak Performance Technologies, Inc, Englewood, CO) to capture and digitize video clips. For lateral flexion, a line segment was created on the trunk (xiphoid to suprasternal notch) and a second segment on the head (chin to top of helmet). The excursion from the vertical position to maximum flexion was the criterion measure. For hyperextension, a line segment was created on the trunk (rib 10 to rib 7) and a second segment on the head (ear hole to top of helmet). The excursion from the vertical position to maximum hyperextension was the criterion measure. Eight frames at the end range of motion were averaged to measure maximum angular displacement of the head relative to the body for each trial. The 3 trials that composed each condition were measured and averaged for use in the statistical analysis.

Data Analysis

We performed two 2 × 5 analyses of variance with repeated measures on both factors to determine the effect of the type of motion (active and passive) and 5 brace conditions on cervical hyperextension and lateral flexion. Post hoc analyses were performed using Tukey Honestly Significant Difference testing to determine where significant differences existed. The .05 level of probability was considered significant. Data were analyzed using the SPSS statistical package (version 10.0, SPSS Inc, Chicago, IL).

RESULTS

We found significant main effects were found for both movement (active versus passive: F_1,114 = 48.90, P < .001) and brace condition (F_4,56 = 47.01, P < .001) in hyperextension (Table 1). The passive overpressure (84.87 ± 9.95) resulted in significantly greater hyperextension than the active movement (68.47 ± 12.58, P < .001). Post hoc comparisons for brace condition (Figure 6) revealed that the Cowboy Collar, A-Force Neck Collar, and neck roll permitted significantly less hyperextension than the shoulder pads alone. In addition, the Cowboy Collar allowed significantly less hyperextension than did the neck roll. Also worth noting is that the shoulder pads did not significantly reduce motion when compared with the helmet alone.

Table 1.

Hyperextension Movement (Mean ± SD)

Figure 6.

Main effect for hyperextension (mean ± standard error). *Significantly different from the helmet condition (P < .01). †Significantly different from the shoulder-pads condition (P < .05). ‡Significantly different from the shoulder-pads condition (P < .01). §Significantly different from the neck-roll condition (P < .05).

Significant main effects for both movement (active versus passive: F_1,14 = 90.96, P < .001) and brace condition (F_4,56 = 26.87, P < .001) in lateral flexion were noted (Table 2). Brace-by-movement interaction was also significant (F_4,56 = 12.51, P < .001). Post hoc analysis revealed that although the 3 brace and shoulder-pad conditions for the passive overpressure condition were not significantly different, the neck roll permitted significantly less active range of motion than the other 2 brace and shoulder-pad conditions (Figure 7, P < .01).

Table 2.

Lateral-flexion Movement (Mean ± SD)

Figure 7.

Brace-by-pressure interaction for lateral flexion (mean ± standard error). *Significantly different from the helmet and all other conditions (P < .01). †Significantly different from the passive neck-roll condition (P < .01). ‡Significantly different between the active and passive motions within the brace condition (P < .01).

DISCUSSION

Recommendations for the use of football collars to prevent burners are well documented in the literature.^12–15 Collars are hypothesized to reduce the amount of cervical motion induced by the collisions that accompany blocking and tackling in football,²¹ namely hyperextension and lateral flexion, which have both been cited as mechanisms for acute burners.^22,23 Hyperextension of the cervical spine can significantly decrease neuroforaminal space²⁴ and pressure within this space to pathologic levels.⁹ Lateral flexion places tension on the upper trunk of the brachial plexus by increasing the interval between the cervical spine and the shoulder girdle.¹ The subsequent effects of excessive compression or traction on peripheral nerves include interruptions of neural blood flow, deformation of the myelin sheaths, and slowing of axonal transport and nerve conduction. Long-term sequelae can include Wallerian degeneration and subsequent denervation of the muscle.^25,26 Preventing these injuries by limiting end-range cervical motion with minimal loss of function is the ideal scenario for both the athlete and the medical staff.

Our results indicate that in a controlled laboratory setting, cervical hyperextension can be limited through the use of all 3 neck rolls as compared with shoulder pads alone. This point is tempered by the fact that passive overloading of the braces still results in an additional 18.9° of hyperextension (on average) across the 3 brace conditions. The Cowboy Collar provided the most protection, followed by the A-Force Neck Collar and the standard neck roll. The Cowboy Collar has the added benefit of providing additional padding over the shoulder and chest, which may help prevent compression at the Erb point. The standard neck roll had the largest increase in hyperextension from active motion to passive motion (20.2°). This information can be used by clinicians when determining the amount of cervical restriction required by a specific athlete. It is interesting to note that hyperextension between the helmet alone and the helmet and shoulder-pad combination did not differ. Given the high incidence of burners, shoulder-pad manufacturers might consider providing additional posterior restraint to the standard shoulder pads.

None of the 3 braces tested restricted passive lateral flexion better than the shoulder pads alone. At the same time, the standard neck roll displayed a significant reduction in active lateral flexion. This is contrary to what one would expect of an ideal collar, which should allow full active motion but restrict excessive passive motion. The standard neck roll may provide the athlete with unwanted restriction or a false sense of security. The athlete may assume the neck roll will control motion, which is true until a large force is placed upon it, at which time it fails. The findings of restriction in hyperextension but not in lateral flexion are consistent with the findings of Hovis and Limbird.²⁷ These authors found similar results when using standard neck rolls or custom-made neck orthoses.

Due to the limited amount of research in this area, our examination was performed in a controlled setting where cervical motion could be evaluated under known force applications. The active motions examined reflect the movements necessary for performance and function on the field of play. Any significant reduction in active motion that is not accompanied by a similar passive effect would indicate equipment that limits function without protective benefits. The passive motions in this study were meant to imitate contact-induced motions of the head, during which the applied force is greater than that generated by the player's neck musculature. In these cases, control of the head and neck is lost, and the resultant cervical motions at or beyond the active end range become the mechanism for burners and other cervical injuries.

According to Levitz et al,¹¹ chronic recurrent burner syndrome is due to nerve-root compression in the intervertebral foramen, secondary to degenerative changes or disk disease. The mechanism associated with this injury is hyperextension or extension and rotation of the cervical spine. This mechanism is more common in collegiate and professional players, but these players have greater access to equipment that can decrease this mechanism. Brachial plexus stretch, stemming from lateral flexion, occurs more commonly in scholastic athletes.¹⁰ This mechanism is not controlled with the equipment tested. In addition, these athletes are more likely to have access to the less expensive standard neck roll. This may result in the impression of protection during active neck motion, but the protection fails the athlete when overpressure is applied.

Given the high rate of acute and chronic burners, the long-term sequelae of the injury,^3,4 and the apparent inability of the neck roll to limit lateral flexion, preventing this injury may require a combination of equipment improvements and other methods. One area to consider during rehabilitation is neuromuscular control of the head-neck complex. Although this concept is commonly considered after injury to other joints,²⁸ it is not normally listed as part of standard rehabilitation after a burner.^29,30 Long-term proprioceptive deficits in the shoulder can result after repeated, low-grade burner injuries.⁴ At other joints, such deficits have been shown to leave athletes susceptible to reinjury,³¹ but we have been unable to find data on neuromuscular control around the neck.

Our findings are limited by the use of the controlled laboratory setting. It remains to be seen if these braces will provide similar results during full, live contact. During play, not only is the stiffness of the brace a concern but also movement of the brace and the brace–shoulder-pad complex. We were able to negate extraneous shoulder-pad movement during the controlled trials. This study was also limited in that rotational or combination movements were not investigated. Brace effects on hyperextension combined with lateral flexion, as replicated with the Spurling sign, need to be explored.

CONCLUSIONS

These results indicate that football collars may be useful in reducing injuries caused by cervical hyperextension and that the Cowboy Collar provided the greatest amount of restriction. Lateral flexion was not significantly reduced beyond the addition of shoulder pads, except with the standard neck roll, which reduced active neck movement only. Therefore, all 3 braces would need to be modified to control lateral flexion. The decision to use a football collar should be based upon the mechanism of injury specific to the athlete, the head and neck motion required by the player's position, and an evaluation of the collar's ability to reduce excessive motion without limiting function. Additional preventive and rehabilitative protocols need to be addressed, especially for injury from lateral flexion.