The participation of athletes in contact sports, particularly

Transcription

1 J Neurosurg Spine 11: , 2009 Mean subaxial space available for the cord index as a novel method of measuring cervical spine geometry to predict the chronic stinger syndrome in American football players Clinical article St e v e n M. Pr e s c i u t t i, M.D., 1 Pe t e r DeLu c a, M.D., 1 Pa u l Marc h e t t o, M.D., 1 Ja r e d T. Wi l s e y, Ph.D., 2 Ch r i s t o p h e r Sh a f f r ey, M.D., 3 a n d Al e x a n d e r R. Va c c a r o, M.D., Ph.D. 1 1 Department of Orthopaedic Surgery, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania; 2 Spinal and Biologics, Medtronic Sofamor Danek, Memphis, Tennessee; and 3 Department of Neurological Surgery, University of Virginia Health System, Charlottesville, Virginia Object. The chronic stinger syndrome is a distinct entity from acute stingers and has been shown to have its own pathophysiology that, unlike acute stingers, may reflect long-standing geometrical changes of the subaxial spinal canal and chronic irritation/degeneration of the exiting nerve root complex. There is no method available, however, to accurately predict these symptoms in athletes. The mean subaxial cervical space available for the cord (MSCSAC) is a novel alternative to the Torg ratio for predicting neurological symptoms caused by cervical spondylosis in elite athletes. It is the goal of this study to determine critical values for this measurement index and to retrospectively correlate those values to neurological symptoms. Methods. Magnetic resonance images obtained in 103 male athletes participating in the 2005 and 2006 National Football League Scouting Combine and a control group of 42 age-matched male nonathletes were retrospectively reviewed. The Torg ratio and SAC values were calculated in triplicate at each cervical level from C3 6 by using lateral radiographs and midsagittal T2-weighted MR images of the cervical spine, respectively. These values were then averaged for each individual to produce mean subaxial cervical Torg ratio (MSCTR) and MSCSAC values. Receiver operating characteristic curves were constructed for each measurement technique and were compared based on their respective area under the curves (AUCs). Results. The MSCSAC difference between athletes with and without chronic stingers was statistically significant (p < 0.01). The difference between athletes with and without chronic stingers compared with controls was also statistically significant (p < and p < 0.001, respectively). The AUC for the MSCSAC was 0.813, which was significantly greater than the AUC for both the MSCTR (p = ) and the individual Torg ratio (p = ). The MSCTR had the second largest AUC (0.676) and the conventional method of measuring individual Torg ratio values produced the lowest AUC (0.661). It was found that using the MSCSAC with a critical value of 5.0 mm produced a sensitivity of 80% and a negative likelihood ratio of 0.23 for predicting chronic stingers. Lowering the cutoff value to 4.3 mm for the MSCSAC resulted in a possible confirmatory test with a specificity of 96% and a positive likelihood ratio of Conclusions. A critical value of 5.0 mm for the MSCSAC provides the clinician with a screening test for chronic stingers and anything < 4.3 mm adds additional confidence as a confirmatory test. These results are ~ 20% more accurate than the classic Torg ratio based on our AUC analysis. It was found that measuring the spinal geometry throughout the length of the subaxial cervical spine produced a more reliable method by which to predict neurological symptoms than the traditional approach of measuring individual levels. This shows that the underlying pathogenesis of the chronic stinger syndrome is best characterized as a process that involves the entire subaxial region uniformly. (DOI: / SPINE08642) Ke y Wo r d s cervical spine space available for the cord stenosis chronic stinger syndrome Torg ratio The participation of athletes in contact sports, particularly American football, predisposes players to injuring the cervical spine and associated nerve Abbreviations used in this paper: AUC = area under their curve; MSCSAC = mean subaxial cervical space available for the cord; MSCTR = mean subaxial cervical Torg ratio; NFL = National Football League; ROC = receiver operating characteristic; SAC = space available for the cord; VB = vertebral body. tracts. One such injury, the stinger syndrome, has received particular attention in the literature. 7,20 A stinger, also called a burner, is classically characterized by unilateral weakness and a burning sensation that radiates This article contains some figures that are displayed in color on line but in black and white in the print edition. 264 J Neurosurg: Spine / Volume 11 / September 2009

2 Use of the MSCSAC index to predict chronic stinger syndrome down an upper extremity. 10 Players may experience tingling, burning, or numbing sensations in a circumferential rather than a dermatomal distribution. The condition may last < 1 minute or as long as a couple of weeks. 5 Sallis et al. 21 reported a career incidence as high as 65% in collegiate American football players. Most of these injuries are the result of a one-time acute injury. Proposed injury mechanisms for these acute, single-episode stingers include the following: 1) brachial plexus stretch or traction injury or 2) injury from a direct blow at Erb s point (also known as Erb palsy and Erb paralysis). 1,2,7,21 These acute symptoms characterize those in the majority of athletes in whom stingers are diagnosed. In some players, however, chronic and recurring symptoms develop. Levitz and colleagues 15 were the first to describe chronic stinger syndrome and explore its pathogenesis. They concluded that the pathomechanics of the chronic stinger syndrome were different from those of acute stingers, originating from degenerative spondylotic changes in the cervical spine and associated neural foramina as opposed to the brachial plexus. They defined the chronic stinger syndrome as 1) the chronic, recurring neurapraxia or axonotmesis or both of a nerve root associated with prolonged weakness, 2) symptoms necessitating time loss from practice and/or games, and 3) recurrence. Kartal and asociates 14 have also described advanced cervical spondylosis in premier professional soccer players in their 3rd decade of life and believed that it was the result of the repetitive heading of the ball throughout their careers. Cervical disc bulging, osteophytes in the cervical canal, disc protrusion, and loss of cervical lordosis were the most common MR imaging findings. 14,15 In the study published by Levitz et al., 15 93% of athletes who were experiencing chronic stingers (mean age 23 years) had either disc disease or narrowing of the neural foramina due to degenerative changes. These chronic symptoms are thought to be secondary to the degenerative spondylotic changes in the cervical spine and surrounding soft tissues, which themselves are the result of repetitive micro- and macrotraumas to the head and neck region suffered over years of contactsports participation. 14,15,18 These spondylotic changes cause chronic inflammation and secondary functional narrowing of both the spinal canal and the neural foramen, resulting in geometrical changes of the cervical spinal canal and leading to less area for the spinal nerve root to exit at its respective level. There have been multiple attempts in the literature to characterize the geometry of the cervical spine in athletes and try to predict neurological symptoms such as the chronic stinger syndrome. Pavlov et al. 20 first described and developed the Torg ratio to assess cervical spinal canal stenosis. Subsequent studies demonstrated the inability of the Torg ratio to accurately predict symptomatic cervical spine stenosis in athletes, showing that the Torg ratio has a high sensitivity but poor positive predictive value in predicting neurological symptoms. 12,19 The measurements used to calculate the Torg ratio are perhaps not specific enough to describe the underlying pathogenesis of chronic stingers, possibly explaining why J Neurosurg: Spine / Volume 11 / September 2009 the Torg ratio has been found to have such poor positive predictive values. The SAC is an alternative method for evaluating cervical spinal canal stenosis Unlike the Torg ratio, which is calculated from measurements taken from plain radiographs, SAC measurements are taken from midsagittal MR images, which avoids the magnification errors common with plain radiographs. Also, the SAC index has been proposed to be a better predictor of symptomatic cervical spine stenosis compared with the Torg ratio because the former is able to take into account not only the osseous anatomy of the spine but also the surrounding soft tissues and associated spondylotic changes that may contribute to a functionally stenotic central canal and neural foramen. 22 By convention, the Torg ratio or SAC value assigned to an athlete has been the smallest individual value through the subaxial cervical spine. As mentioned above, however, this method has resulted in a poor positive predictive value for these methods to predict neurological symptoms. 6,22,25 Castro and coworkers 6 argued that the original definition of cervical stenosis by Pavlov et al. 20 was based on the sensitivity of the relative operating characteristic curve for the average Torg ratio value over the entire cervical canal, not the smallest individual cervical level measured. It has been shown in the literature that the pathogenesis of the chronic stinger syndrome is cervical spine spondylosis, 14,15 a problem that has the potential to cause long-term neurological problems for these athletes, yet currently there is no way to accurately predict these symptoms in these high-level athletes. The purpose of this study was to compare the Torg ratio and SAC and their capacity for positively predicting the chronic stinger syndrome in elite athletes. This will be done in 2 ways: 1) assigning the smallest individual value for C-3 through C-6 for the Torg ratio, and 2) averaging the individual Torg ratio and SAC values through the subaxial cervical spine and assigning a composite index value to each athlete. These composite indices are referred to as the MSCTR and the MSCSAC, respectively. It is our hypothesis that the SAC will provide a higher sensitivity and specificity for predicting neurological symptoms compared with the Torg ratio due to the SAC measurement index having an intuitive geometrical relationship to the pathogenesis of chronic stingers. We believe that the predictive value of the SAC will be further enhanced when taking the geometrical changes throughout the subaxial cervical spine into account (MSCSAC). Methods Patient and Control Populations Approval for this retrospective study was obtained from both our institutional review board and the NFL. Cervical radiographs and MR images obtained in 103 male athletes and 42 male nonathletes (control group) were reviewed. The athlete group was composed of a random cross-section of American collegiate football players who entered into the 2005 and 2006 NFL draft and in 265

3 S. M. Presciutti et al. whom imaging studies were acquired as part of the NFL Scouting Combine evaluation. The average weight and height of the Combine participants was 233 lb and 74 inches, respectively. In the Scouting Combine medical history is evaluated and collegiate football players who are candidates for the NFL draft undergo a preliminary physical examination. In addition, many teams get information from other sources such as direct discussions with college athletic trainers. When the athlete is examined at the Combine, the physicians will question the athlete about the circumstances of his reported injury, perform their own physical examination, combine all of the available history from various sources, and determine if the palyer truly had a stinger. If deemed necessary, the physician at the Combine orders whatever medical diagnostic tests deemed appropriate. In 2005 and 2006, for example, with 326 total athletes participating in the Combine, > 600 MR imaging studies were performed. For the purposes of this study, all identifying markers were eliminated prior to data review, with the exception of the history of stingers in the athlete group. It is important to note that for this study, an athlete was not designated as having a positive history for chronic stingers if he had only a single stinger episode in his past. Multiple episodes of stingers had to have been reported and confirmed by the Combine medical staff, and all portions of the chronic stinger definition, as outlined above, had to be satisfied. The control group of 4 individuals was assembled from male nonathletes matched for age (age range years, mean weight 182 lb, and mean height 68 inches). The control individuals were randomly chosen from a list of patients in whom cervical radiographs and MR images were obtained at our institution between 2006 and These general search criteria brought up 164 patients. Most of the control individuals had such images acquired to investigate a single episode of trauma, not to the head or neck in particular, but when the cervical spine needed to be cleared during routine evaluation in our emergency department. The 164 possible control individuals were significantly restricted because we excluded from this study any control individual in whom cervical spine problems (such as discogenic radiculopathy) developed during the 6-month period after their imaging studies or in whom any cervical anatomical aberration (disc protrusion/herniation or fracture) was documented on plain radiography or MR imaging. Radiographic Measurements Before undergoing radiography and MR imaging, the individuals were positioned supine with no occipital padding. All radiographs obtained in the athlete (103 individuals) and control nonathlete (42 individuals) were evaluated sagittally at C3 6. The sagittal diameters of the VB and spinal canal were measured manually in millimeters and assessed using Phillips ISite software. The Torg ratio was calculated by measuring the shortest distance from the midpoint between the superior and inferior endplates of the posterior aspect of the VB to the nearest point on the corresponding spinolaminar line and dividing this value by the anteroposterior diameter at the midpoint between the superior and inferior endplates of the VB. The anteroposterior measurement of the VB was made at the midportion of the VB to avoid the effect of osteophytes. The athlete s MR images were acquired using one of two 1.5-T units at Methodist Hospital in Indianapolis, Indiana: either a GE Healthcare or Siemens Medical Solutions imager. The control individuals MR images were obtained in one of our institutions, 1.5-T Siemens units. In obtaining all studies standard imaging protocols were used. It is important to note that the resolution of 1.5-T units is ~ 500 µ, which is more than powerful enough to make measurements in the submillimeter range. The SAC index was calculated from midsagittal T2- weighted MR images of the cervical region. The SAC is calculated by subtracting the sagittal diameter of the spinal cord from the disc-level sagittal diameter of the spinal canal. The disc-level sagittal spinal canal diameter was measured as the shortest distance between the intervertebral disc and the spinolaminar line. The sagittal spinal cord diameter was measured at the midline of the VB at the appropriate disc level. These measurements were traced manually in millimeters and assessed using Phillips ISite software. All individual measurements for both the athlete and control groups were taken in triplicate by a single reviewer at each of the 4 cervical levels, and a mathematical mean and corresponding SD were calculated for both Torg ratio and SAC values at each cervical level. The smallest individual Torg ratio and SAC value from C-3 through C-6 was the player s assigned respective value. Finally, Torg ratios and SAC values from all 4 cervical levels were averaged for each individual, yielding composite values for the MSCTR and MSCSAC, respectively. The reviewer was completely blinded to the category of the patient (controls, athletes who had chronic stingers, and athletes who had no stingers). Statistical Analysis The MSCSAC was evaluated for 3 groups: athletes with chronic stingers, athletes without chronic stingers, and controls. The differences between the groups were assessed by 1-way ANOVA with a Tukey s post hoc. All athletes were also compared with controls by a 2-tailed unpaired t-test. Receiver operating characteristic curves plot a test s true-positive rate (sensitivity) against its false-positive rate (1-specificity) for continuously changing cutoffs over the entire possible range of test results. The ROC curve is used to evaluate the tests that result from these cutoffs with the entire spectrum of pairs of true-positive rates and false-positive rates and so gives a global description of a test s classification accuracy. 17 A completely random guess would give a point along a diagonal line, called the line of no discrimination, from the left bottom to the top right corner. The ROC curve is one of the most popular measures used for clinical diagnostic tests and has been widely used in different areas of medicine. 9 In this study we chose the area under the ROC curve (the AUC), the most popular summary index of the ROC curve, as our measure of a given test s performance. The AUC measures the probability that test values from a ran- 266 J Neurosurg: Spine / Volume 11 / September 2009

4 Use of the MSCSAC index to predict chronic stinger syndrome domly selected pair of symptomatic and nonsymptomatic individuals are correctly ordered and is thus a convenient global measure for the quantification of diagnostic accuracy. 29 For all 3 measurement techniques, ROC curves were generated by combining the measurement data from all 103 of the athletes and the 42 control individuals with their clinical history of either having chronic stinger syndrome or not, as already defined. These curves were created using SPSS computer software (version ), and the AUCs were calculated. The comparison of the AUCs was performed using MedCalc software for Windows (version ). The value of α was set at A criticalratio z-test was calculated that compared the ROC curves derived from the different measurement tests. Values > 1.96 were considered sufficient evidence to reject the null hypothesis (that is, that ROC AUCs were the same). Results Table 1 shows the mean MSCSAC for the 3 study groups: athletes with chronic stingers (N = 28), athletes without chronic stingers (N = 75), and controls (N = 42). It is evident that those athletes who reported a history of chronic stingers had a statistically significantly lower MSCSAC than athletes without chronic stingers (p < 0.01) and the controls (p < 0.001). Also, athletes without chronic stingers had a significantly smaller MSCSAC index than controls (p < 0.001). It is important to note that none of the control individuals had a history of stingers in their history. Also, all athletes were compared with controls (with respect to their respective group s mean MSC- SAC) by a 2-tailed unpaired t-test, and the difference was again found to be significant (p < ). Figure 1 shows the ROC curves for all 3 measurement indices with sensitivity plotted against 1-specificty. The AUC calculations for the 3 measurement indices are summarized in Table 2. The AUC for the MSCSAC ROC was significantly greater than both those for MSCTR (p = ) and the individual Torg ratio (p = ). The MSCSAC achieved an area under the ROC curve of 0.813, a value that exceeded the previously accepted published method (measuring individual Torg ratio values) by > 18%. Table 3 provides a summary of the statistical comparisons between the ROC curves for the 3 measurement indices. Critical-ratio z-test values are provided as a way to evaluate statistical significance between the ROC curves. The z-test values > 2.0 or < 2.0 are often taken as indicating a statistically significant difference. The MSCSAC ROC curve was found to be statistically superior to both measuring individual Torg ratio values and the MSCTR measurement, with z-test values of 2.20 and 1.99, respectively. A statistical breakdown of the 3 measurement indices tested with varying critical values is provided in Table 4. We found that using the MSCSAC, with a critical value of 5.0 mm, produced a reasonable screening test for predicting the chronic stinger syndrome with a sensitivity of 80% and a negative likelihood ratio of Lowering the cutoff value to 4.3 mm for the MSCSAC resulted in a J Neurosurg: Spine / Volume 11 / September 2009 Table 1: Mean MSCSAC values for the 3 study groups deviations Group No. of Cases MSCSAC* athletes ± 1.1 w/ chronic stingers ± 0.8 w/o chronic stingers ± 0.9 controls ± 0.8 *Values are presented as the mean ± SD. possible confirmatory test with a specificity of 96% and a positive likelihood ratio of Discussion The Torg ratio has classically been used in the literature to predict neurological symptoms based on cervical spine geometry. 6,20,23 25 Lateral radiographs, on which the Torg ratio are based, however, cannot detect the softtissue changes, which are important in the underlying pathogenesis of syndromes like chronic stingers. This has perhaps resulted in the poor positive predictive values for the Torg ratio in the literature. 6,23,25 Indeed MR imaging technology and the SAC index are much more adept tools at detecting both soft-tissue and osseous abnormalities. Additionally, it could be reasoned that the degenerative spondylotic changes that are responsible for chronic stingers would be seen throughout the subaxial spine and not only at an individual level. In fact, Levitz et al. 15 found evidence of multiple levels being involved in players with chronic stinger syndrome. This is the reasoning for averaging the SAC from C-3 through C-6 to produce the composite MSCSAC index. It appears that these spondylotic changes affect the cervical spine rather uniformly. Changes in the neural foramina (disc herniation, facet joint degeneration) that cause the chronic stinger syndrome are a part of the same pathological process that causes degenerative narrowing of the central canal. Although they affect different locations in the cervical spine, the processes are identical and are due to the same repetitive trauma in these athletes. Thus, any narrowing in the central canal picked up by these measurements is a direct reflection of the same phenomenon occurring in the neural foramina and not merely an epiphenomenon. Three different measurement indices were compared and evaluated for their ability to predict neurological symptoms from geometrical changes in the cervical spine of elite American football athletes by constructing ROC curves for each of the measurement techniques: the MSC- SAC, MSCTR, and SAC. Our results show that the MSCSAC index (examining the geometry of the subaxial cervical spine) has the largest AUC (0.813) and is therefore the best measurement index for predicting chronic stinger syndrome. This means that the MSCSAC achieves the highest diagnostic accuracy among all approaches evaluated. The MSCTR had the second-largest AUC and therefore outperformed 267

5 S. M. Presciutti et al. Fig. 1. The ROC curves for the 3 measurement indices evaluated (the MSCSAC, MSCTR, and the individual Torg ratio [TR]) for predicting chronic stinger syndrome in elite athletes. Sensitivity is plotted against the 1-specificity, and the no-discrimination line is given in each of the plots, represented as the solid diagonal line. the conventional method of taking individual Torg ratio measurements. This demonstrates that the underlying pathogenesis of chronic stinger syndrome is perhaps best characterized as a process that involves the entire subaxial region rather uniformly. This is a confirmation of the findings by Levitz et al. 15 and Kartal et al. 14 Also, because the AUC for the MSCSAC index is ~ 18% greater than the 2 methods that rely on the Torg ratio, it can be reasoned that MR images are able to detect changes in the cervical spine that are not seen on plain radiographs. This is inherent to the anatomical and pathological courses of chronic stingers, which result from a series of repetitive traumatic events in conjunction with osseous and soft-tissue degenerative changes. The comparison of the measurement indices was further evaluated and broken down into simpler, more utilitarian numbers. In clinical practice it is essential to know how a particular test result predicts the risk of abnormality. Sensitivities and specificities 3 do not do this: they describe instead how abnormality (or normality) predicts particular test results. Predictive values 4 do give probabilities of abnormality for particular test results, but they depend on the prevalence of abnormality in the study sample and can rarely be generalized beyond the study. Likelihood ratios provide a solution, however, as they can be used to calculate the probability of abnormality while adapting for varying prior probabilities of the chance of abnormality from different contexts. 8 Each test result has its own likelihood ratio, which summarizes how many times more (or less) likely patients with the disease are to have that particular result than patients without the disease. For these reasons, likelihood ratios were chosen to compare the various combinations of testing technique and critical value. Because the MSCTR and MSCSAC were the best measurement techniques as determined by their respective AUC, each index was assessed at 2 different cutoff points (Table 4). The MSCTR measurement index was evaluated with critical values of 0.8 and 0.7, which are the 2 most popular cutoff points for the Torg ratio, as proposed by Torg s group 20 and Castro et al., 6 respectively. The MSCSAC was assessed with cutoff points of 4.3 and 5.0 mm. The 4.3-mm critical value was chosen because it is a previously proposed cutoff (S. Presciutti et al., unpublished data, 2008) and the 5.0-mm was selected due to its correspondence with the point of the curve at the minimum distance from the upper left corner (100% sensitivity, 100% specificity). These breakdowns provided insight into ranges of values that would produce good screening and confirmatory tests. We found that using 5.0-mm as a critical value for the MSCSAC produced the best results as a screening test for determining which athletes may experience chronic stinger syndrome due to cervical spondylosis. This critical value achieved a high sensitivity (80%), negative predictive value (89%), accuracy (85%), and negative likelihood ratio (0.23). Such a high sensitivity helps to avoid missing those athletes in whom there is potential to develop chronic stinger syndrome. In screening, one must also be confident that a negative test result or in this case an athlete with an MSCSAC > 5.0 mm translates into a patient in whom the probability of developing clinical symptoms is low. We found that this test and cutoff produces the lowest negative likelihood ratio and provides significant confidence to the patient and clinician when a negative test result is achieved. Lowering the cutoff to 4.3 mm for the MSCSAC pro- Table 2: Summary of AUCs for the measurement indices* Measurement Index Area Under ROC Curve individual TR ± MSCTR ± MS-SAC ± * Values are presented as the mean ± SEM. Abbreviation: TR = Torg ratio. 268 J Neurosurg: Spine / Volume 11 / September 2009

6 Use of the MSCSAC index to predict chronic stinger syndrome Table 3: Statistical breakdown of the AUC analyses between study groups* Variable MSCTR vs MSCSAC MSCTR vs Individual TR MSCSAC vs Individual TR AUC difference SEM % CI to z value p value * Difference refers to the respective AUCs. Critical-ratio z values and p values are provided as a way to evaluate statistical significance between the ROC curves. vided an extremely high positive likelihood ratio (13.25) and specificity (96%). This combination could therefore be used as a rule-in test if there were any reason to doubt cervical spondylosis and/or canal stenosis being the reason for a player s chronic stinger symptoms, following a positive test with 5.0 mm as a critical value for the MSCSAC. The high positive likelihood ratio of this testing combination means that, when a positive test is observed, the clinician could conclude that the athlete has a > 13-fold increase in the odds of developing the chronic stinger syndrome than if a negative test result had occurred. These results, for the first time, show that it may be possible to accurately and reliably assess cervical spine geometry as a way to predict the development of the chronic stinger syndrome in elite American football athletes. Screening tests must be done with purpose, however, and to be cost effective there should be a treatment for the condition itself or its prevention. The management of stingers has classically revolved around addressing predisposing factors, 11,26,27 correcting strength-related deficits, 28 and enhancing protective equipment that limit neck movement. 13 These interventions have focused on treating the acute stinger syndrome, however, and it is unclear whether these management options will have similar success with chronic stinger syndrome because they do not share the same underlying pathogenesis. 15 Deciding which athletes in whom management should be instituted based on radiological studies, however, can be a challenge. Many clinical and laboratory tests do not have a simple positive or negative result but rather give a range of values. Clinical experience and common sense tell us that the more abnormal a test result, the more likely that it reflects actual disease than a test result that is mildly abnormal. This is why we chose the likelihood ratio as a means of interpreting the results of the MSC- SAC. For example, as opposed to saying that a player with an MSCSAC index of < 5.0 mm has a positive result, one can say that the athlete has a > 6-fold odds of having neurological symptoms related to chronic stingers than if the MSCSAC had measured > 5.0 mm. It is also important to note that Lurie et al. 16 showed that the effect of degenerating discs and facet joints on the central spinal canal can be evaluated effectively with MR imaging, which yields a moderate to substantial intra- and interreader reliability. 16 This strengthens the utility of this measurement index as one that can be applied widely. In this era of professional sports and their attendant economical matters, it was a concern that these athletes may be underreporting their injuries to their team trainers. This would certainly be a real concern considering the study design if we were reporting acute onetime stingers. The nature of chronic stingers, however, makes it much less likely for a player to be able to hide his symptoms due to the chronicity of the problem and the requirement of the definition of chronic stingers that holds that a player must have lost time from a game and/ or practice. One limitation of this study is its clinical relevance because plain radiography now is more frequently available as a screening mechanism than MR imaging. Today, preparticipation radiography is rarely conducted as a screening tool, however. Radiographic examinations are normally performed after an athlete renders a complaint such as a single stinger episode. After such a complaint is made, the use of MR imaging to calculate the MSCSAC index would be more of a viable option. If MR imaging were performed and it demonstrated an MSCSAC value below the critical 5.0 mm, then the athlete could be better counseled on the possible risks of developing recurring chronic stingers. Table 4: Statistical breakdown of the effectiveness of measurement indices predicting neurological symptoms* Percentage Likelihood Ratio Measurement Index (critical value) Sensitivity Specificity PPV NPV Accuracy Positive Negative individual TR (0.8) MSCTR (0.7) MSCTR (0.8) MSCSAC (4.3 mm) MSCSAC (5.0 mm) * NPV = negative predictive value; PPV = positive predictive value. J Neurosurg: Spine / Volume 11 / September

7 S. M. Presciutti et al. Another limitation of the study is that the data were calculated retrospectively and therefore only shows a strong statistical relation between measuring spinal geometry with the MSCSAC index and athletes experiencing chronic stinger syndrome in the past. Direct conclusions regarding screening and actually predicting future neurological symptoms from altered cervical geometry cannot be made. As mentioned, however, acute and onetime stingers result from acute foraminal narrowing or brachial plexus injuries, 1 whereas chronic stinger syndrome results from spondylotic changes in the cervical spine. 15 It must be stressed then that the conclusions of this research pertain only to predicting the development of chronic stinger syndrome after a player has already experienced an acute stinger episode in the past. Our results cannot be applied to routine screening of asymptomatic athletes for the first-time development of acute stinger syndrome. Conclusions The MSCSAC measurement index provides the clinician with the greatest ability to date to predict the development of chronic stinger syndrome in elite American football athletes, based on changes in cervical spine geometry. The previous method of measuring the Torg ratio at individual cervical levels only yielded limited success because of a low sensitivity and poor positive predictive value. 6,23,25 An MSCSAC value of 5.0 mm provides the best results as a screening test for chronic stinger syndrome in elite athletes. A value < 4.3 mm provides excellent results as a confirmatory test. These critical values and measurement index are ~ 20% more accurate than the classic Torg ratio based on its AUC analysis. We also demonstrated, for the first time, that measuring spinal canal geometry throughout the subaxial cervical spine produces results superior to those of the classic method in which individual levels are measured, possibly showing that the degenerative changes that underlie the chronic stinger syndrome occur rather uniformly throughout the subaxial cervical spine. With the very strong correlation shown between chronic stingers and the MSCSAC index retrospectively, it seems prudent that this measurement be further investigated in a prospective fashion. Future research must also be conducted to determine what management options exist for athletes who experience the chronic stinger syndrome and what opportunities exist to intervene at an earlier stage in the disease process. Acknowledgment We would like to give a special thanks to the NFL for allowing us to pursue this research. Disclosure Drs. Vaccaro and Shaffrey receive Medtronic support for the Spine Trauma Study Group. The authors report no other conflicts of interest concerning the materials or methods used in this study or the findings specified in this paper. References 1. Albright JP, McAuley E, Martin RK, Crowley ET, Foster DT: Head and neck injuries in college football: an eight-year analysis. Am J Sports Med 13: , Albright JP, VanGilder J, El-Khoury G, Crowley E, Foster D: Head and neck injuries in sports, in Scott WN, Nisonson B, Nicholas JA (eds): Principles of Sports Medicine. Maryland: Williams & Wilkins, 1984, pp Altman DG, Bland JM: Diagnostic tests 1: sensitivity and specificity. BMJ 308:1552, Altman DG, Bland JM: Diagnostic tests 2: predictive values. BMJ 309:102, Cantu RC, Bailes JE, Wilberger JE Jr: Guidelines for return to contact or collision sport after a cervical spine injury. Clin Sports Med 17: , Castro FP Jr, Ricciardi J, Brunet ME, Busch MT, Whitecloud TS 3rd: Stingers, the Torg ratio, and the cervical spine. Am J Sports Med 25: , Clancy WG Jr, Brand RL, Bergfeld JA: Upper trunk brachial plexus injuries in contact sports. Am J Sports Med 5: , Deeks JJ, Altman DG: Diagnostic tests 4: likelihood ratios. BMJ 329: , Hanley JA: Receiver operating characteristic (ROC) methodology: the state of the art. Crit Rev Diagn Imaging 29: , Hershman EB: Brachial plexus injuries. Clin Sports Med 9: , Hershman EB: Injuries to the brachial plexus, in Torg JS (ed): Athletic Injuries to the Head, Neck, and Face, ed 2. St. Louis: Mosby, 1991, pp Herzog RJ, Wiens JJ, Dillingham MF, Sontag MJ: Normal cervical spine morphometry and cervical spinal stenosis in asymptomatic professional football players: plain film radiography, multiplanar computed tomography, and magnetic resonance imaging. Spine 16:S178 S186, Hovis WD, Limbird TJ: An evaluation of cervical orthoses in limiting hyperextension and lateral flexion in football. Med Sci Sports Exerc 26: , Kartal A, Yildiran I, Senkoylu A, Korkusuz F: Soccer causes degenerative changes in the cervical spine. Eur Spine J 13: 76 82, Levitz CL, Reilly PJ, Torg JS: The pathomechanics of chronic, recurrent cervical root neurapraxia. The chronic burner syndrome. Am J Sports Med 25:73 76, Lurie JD, Tosteson AN, Tosteson TD, Carragee E, Carrino JA, Kaiser J, et al: Reliability of readings of magnetic resonance imaging features of lumbar spinal stenosis. Spine 33: , McClish DK: Analyzing a portion of the ROC curve. Med Decis Making 9: , Meyer SA, Schulte KR, Callaghan JJ, Albright JP, Powell JW, Crowley ET, et al: Cervical spinal stenosis and stingers in collegiate football players. Am J Sports Med 22: , Odor JM, Watkins RG, Dillin WH, Dennis S, Saberi M: Incidence of cervical spinal stenosis in professional and rookie football players. Am J Sports Med 18: , Pavlov H, Torg JS, Robie B, Jahre C: Cervical spinal stenosis: determination with vertebral body ratio method. Radiology 164: , Sallis RE, Jones K, Knopp W: Burners: offensive strategy for an underreported injury. Phys Sportsmed 20:47 55, Tierney RT, Maldjian C, Mattacola CG, Straub SJ, Sitler MR: Cervical spine stenosis measures in normal subjects. J Athl Train 37: , Torg JS: Epidemiology, pathomechanics, and prevention of 270 J Neurosurg: Spine / Volume 11 / September 2009

8 Use of the MSCSAC index to predict chronic stinger syndrome football-induced cervical spinal cord trauma. Exerc Sport Sci Rev 20: , Torg JS, Corcoran TA, Thibault LE, Pavlov H, Sennett BJ, Naranja RJ Jr, et al: Cervical cord neurapraxia: classification, pathomechanics, morbidity, and management guidelines. J Neurosurg 87: , Torg JS, Naranja RJ Jr, Pavlov H, Galinat BJ, Warren R, Stine RA: The relationship of developmental narrowing of the cervical spinal canal to reversible and irreversible injury of the cervical spinal cord in football players. J Bone Joint Surg Am 78: , Warren RF: Neurologic injuries in football, in Jordan BD, Tsairis P, Warren RF (eds): Sports Neurology. Rockville, MD: Aspen Publishers, 1989, pp Watkins RG, Dillin WM: Cervical spine and spinal cord injuries, in Fu FH, Stone DA, (ed): Sports Injuries: Mechanisms, Prevention, Treatment. Baltimore: Williams & Wilkins, 1994, pp Weinstein SM: Assessment and rehabilitation of the athlete with a stinger. A model for the management of noncatastrophic athletic cervical spine injury. Clin Sports Med 17: , Zweig MH, Campbell G: Receiver operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clin Chem 39: , 1993 Manuscript submitted September 25, Accepted March 5, Address correspondence to: Steven M. Presciutti, M.D., 331 Old Kings Highway, Downingtown, Pennsylvania SPresciutti@resident.uchc.edu. J Neurosurg: Spine / Volume 11 / September