AS THE NUMBER OF interventions proposed for study in

Transcription

1 ORIGINAL ARTICLE Upper- and Lower-Extremity Motor Recovery After Traumatic Cervical Spinal Cord Injury: An Update From the National Spinal Cord Injury Database Ralph J. Marino, MD, Stephen Burns, MD, Daniel E. Graves, PhD, Benjamin E. Leiby, PhD, Steven Kirshblum, MD, Daniel P. Lammertse, MD 369 ABSTRACT. Marino RJ, Burns S, Graves DE, Leiby BE, Kirshblum S, Lammertse DP. Upper- and lower-extremity motor recovery after traumatic cervical spinal cord injury: an update from the National Spinal Cord Injury Database. Arch Phys Med Rehabil 2011;92: Objective: To present upper- (UEMS) and lower-extremity motor score (LEMS) recovery, American Spinal Injury Association Impairment Scale (AIS) change, and motor level change in persons with traumatic tetraplegia from the Spinal Cord Injury Model Systems (SCIMS). Design: Longitudinal cohort; follow-up to 1 year. Setting: U.S. SCIMS. Participants: Subjects (N 1436; age 15y) with tetraplegia with at least 2 examinations, the first within 7 days of injury. Subjects were 80% men injured by vehicular collisions (44%), falls (30%), sports (12%), and violence (11%). Interventions: Not applicable. Main Outcome Measures: Change in AIS, UEMS, LEMS, and motor levels. Results: From a baseline of 7 days or less, 22% of subjects with AIS grade A converted to AIS grade B or better by rehabilitation discharge; and 30%, by 1 year, with 8% to AIS grade C and 7.1% to grade D. Conversion from complete to motor incomplete was not related to timing of the initial examination (P.54) or initial neurologic level (P.96). For AIS grade B, 34% remained motor complete, 30% became AIS grade C, and 37% became grade D by 1 year. Although 82.5% of those with AIS grade C improved to AIS grades D and E, mean 1-year UEMS score was only 35 points. UEMS scores in patients with AIS grade A increased a mean of 9 to 11 points, except for C1 to C3 and C8 to T1 motor levels (gain, 2 3 points). Motor level was unchanged or ascended in 35% and improved 1 level in 42%, 2 levels in 14%, and more than 2 levels in 9%. Motor zone of partial preservation of 2 segments or more was associated with gain of 2 or more motor levels, with a relative risk of 5.0 (95% confidence interval, ; P.001). Conclusions: More patients with cervical complete spinal cord injury may be converting to AIS grade D compared with earlier reports. Motor level recovery in those with AIS grade A and UEMS recovery in those with AIS grade C injuries are potential outcomes for acute clinical trials. Key Words: Outcomes; Recovery; Rehabilitation; Spinal cord injuries; Tetraplegia by the American Congress of Rehabilitation Medicine AS THE NUMBER OF interventions proposed for study in acute SCI has increased, it has become critical to understand the usual course and extent of neurologic recovery. Currently, there are at least 3 large databases that have been used for evaluating neurologic recovery after SCI: 2 longitudinal databases, the EM-SCI database 1,2 and the NSCISC database, which contains longitudinal data from the SCIMS centers, and 1 clinical trial database: the Sygen study database. 3,4 Similarities and differences have been found comparing the Sygen and EM-SCI databases. For example, there are similar rates and amounts of upper-extremity recovery in patients with complete tetraplegia, but differing rates of conversion from complete to incomplete injuries. 3 Neurologic recovery from the SCIMS was reported last in One limitation of that study was that only total motor score recovery was reported. It has since been shown that motor score should be separated into a UEMS and an LEMS. 6,7 This was not possible at the time because the NSCISC database did not include individual key muscle motor scores until late 1993, and there were not enough data to estimate improvements in UEMS and LEMS. Also at that time, there were no From the Departments of Rehabilitation Medicine (Marino) and Pharmacology and Experimental Therapeutics (Leiby), Thomas Jefferson University, Philadelphia, PA; Department of Rehabilitation Medicine, University of Washington, Seattle (Burns); Spinal Cord Injury Service, Veterans Affairs Puget Sound Health Care System, Seattle, WA (Burns); Department of Physical Medicine and Rehabilitation, Baylor College of Medicine, Houston, TX (Graves); Kessler Institute for Rehabilitation, West Orange (Kirshblum); Department of Physical Medicine and Rehabilitation, University of Medicine and Dentistry of New Jersey/New Jersey Medical School, Newark, NJ (Kirshblum); and Department of Physical Medicine and Rehabilitation, University of Colorado, Denver and Craig Hospital, Englewood, CO (Lammertse). Supported by the National Institute on Disability and Rehabilitation Research, Office of Special Education and Rehabilitative Services, Department of Education (grant no. H133N060011). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated. Correspondence to Ralph J. Marino, MD, Dept of Rehabilitation Medicine, 132 S 10 th St, Ste 375, Main Bldg, Philadelphia, PA 19107, ralph.marino@ jefferson.edu. Reprints not available from the author /11/ $36.00/0 doi: /j.apmr AIS CI EM-SCI ISNCSCI LEMS NSCISC SCI SCIMS TBI UEMS ZPP List of Abbreviations American Spinal Injury Association Impairment Scale confidence interval European Multi-center Spinal Cord Injury International Standards for Neurological Classification of Spinal Cord Injury lower-extremity motor score National Spinal Cord Injury Statistical Center spinal cord injury Spinal Cord Injury Model Systems traumatic brain injury upper-extremity motor score zone of partial preservation

2 370 MOTOR RECOVERY AFTER CERVICAL SPINAL CORD INJURY, Marino other large studies of neurologic outcomes after SCI with which the SCIMS data could be compared. The purpose of this article was to report changes in AIS grade, UEMS, LEMS, and motor levels from injury up to 1 year postinjury in persons with traumatic tetraplegia included in the NSCISC database and compare findings with those from the Sygen and EM-SCI databases. METHODS We extracted demographic and neurologic data from 1994 to 2009 from the NSCISC database. To be included, subjects had to be at least 16 years of age at the time of injury, have a complete neurologic examination recorded within 1 week of injury, and have at least 1 other complete neurologic examination more than 30 days after injury. We excluded children younger than 16 years because the reliability of motor scores may be reduced in these subjects. 8 All centers participating in the NSCISC database had the research approved by their local institutional review boards. The NSCISC database collects neurologic data at system admission (day-1 admisions only), rehabilitation admission, rehabilitation discharge, and 1-year assessment. The criteria and timing of the neurologic assessment changed during this study period. Neurologic data were obtained at system admission for day-1 admissions only, at system discharge for all patients, and at 1-year follow-up. In October 2000, data at rehabilitation admission for all patients were added. The date of the neurologic examination was not added to the database until A complete neurologic assessment in the NSCISC database includes the 20 key muscle scores, right and left sensory and motor levels, AIS grades, and date of examination. Levels and AIS grades are assigned by the individual center. Sensory dermatome scores are not included in the database. In addition, until 2000, when these variables were dropped from the database, sensory and motor ZPPs were included for patients classified as complete. Data Processing During 1994 to 2009, a total of 2393 subjects with tetraplegia admitted within 7 days of injury were entered into the database. Of these, we excluded 250 due to missing date of examination, most of whom were injured in 1994 to 1995, when examination dates were added; 122 because the first neurologic examination was more than 7 days after injury; 209 due to incomplete motor score or sensory level data; 28 due to incomplete baseline data; and 35 because age was younger than 16 years. This left 1749 subjects meeting the inclusion criteria. An additional 313 subjects were excluded because there was not a complete follow-up examination more than 30 days after injury. This left 1436 subjects (82% of those meeting inclusion criteria) with complete neurologic data at admission and at least 1 follow-up examination. Of these, 1393 subjects had data for discharge from rehabilitation and 705 (40% of original sample) had 1-year data. We used a computerized classification algorithm developed by 1 of the authors (R.J.M.) to check AIS grade designation. This algorithm closely followed the ISNCSCI, 9 but used several key assumptions in making AIS assignment due to the incomplete sacral and sensory data in the NSCISC database. Because there were few subjects with data for the sacral examination and criteria for designating motor level was modified by the American Spinal Injury Association/International Spinal Cord Society in 1996, 10 an examination was classified as motor incomplete only if LEMS was greater than zero. The algorithm resulted in reclassification of 146 of 1436 initial AIS grades (10.2%). Discrepancies between designated and algorithmderived levels and AIS grades were each reviewed by 2 different authors. Motor ZPP was derived for subjects with complete injuries by using key muscle grades. It was not possible to reconstruct sensory ZPP. We used the 2002 ISNCSCI definition for motor level designation. Data Analysis We calculated percentages of change in AIS grades from admission to discharge and to 1 year. For subjects initially classified as AIS grade A, we used chi-square analysis to determine whether the time of the initial examination or initial level of injury influenced conversion to motor incomplete status. Time of initial examination was dichotomized to less than 3 and 3 to 7 days postinjury, whereas level of injury was dichotomized to single neurologic levels C1 to C4 and below C4. The hypothesis was that earlier examinations and higher levels of injuries would be less reliable and have higher rates of conversion than later examinations or those with lower level injuries. Prior studies have indicated that examination at 72 hours or later postinjury is a better predictor of neurologic recovery than the earlier examination. 11,12 Also, patients requiring intubation are more likely to convert to incomplete than those not intubated. 13 We used higher level of injury as a proxy for intubation because intubation status was not available. We also examined whether specific centers had different rates of AIS grade conversion. We calculated mean and median UEMS and LEMS and changes in motor score from injury to 1 year for each initial AIS grade. For subjects initially classified as AIS grade A, we looked at UEMS changes by initial motor level. If the right and left motor levels were not the same, the side with the more rostral level was used for analysis. We also determined change in motor levels from injury to 1 year by right and left side and by initial motor level. We used chi-square analysis to evaluate whether the length of the motor ZPP was related to change in motor level at 1 year. This analysis dichotomized the length of the motor ZPP to 0 to 1 segment or 2 or more segments, and motor level improvement, to 1 or fewer or 2 or more levels. Because only subjects with a motor level of C7 or more rostral can gain 2 or more motor levels as determined by using the upper-limb motor examination, subjects with initial motor levels of C8 or below (caudal) were excluded from the motor level change analysis. RESULTS The 1436 subjects included in the analyses were recruited from 21 different centers. The number of subjects from a single center ranged from 3 to 196. Mean SD age at injury was years. Thirty-four percent of the sample was 30 years or younger and 14% were older than 60 years at the time of injury. Men made up 80% of the sample. Whites made up 61% of the sample; blacks, 27%; Hispanics, 9%; and other racial groups, 3% (table 1). The most common causes of injury were vehicular collisions (44%), falls (30%), sports (12%), and acts of violence (11%). At 1 year, 705 subjects were available for analysis. There was no difference between those with and without 1-year data in terms of sex, race, or initial AIS grade (P.05), but age distribution was different (P.01), primarily due to greater dropout of subjects older than 60 years. There were differences between the initial 1749 subjects with and without follow-up. The 313 without follow-up were more likely to be older than 60 years, have been injured by falling, and be AIS grade D (P.01 for all).

3 MOTOR RECOVERY AFTER CERVICAL SPINAL CORD INJURY, Marino 371 Variable Table 1: Demographics and Comparison of Subjects With and Without 1-Year Data Total 1-y Present 1-y Missing N n % n % Sex Men Women Race White Black Hispanic Other Age group (y) Initial AIS grade A B C D NOTE. N P Changes in AIS Grades There were 1393 subjects with admission and discharge AIS grades (table 2a). Median time from injury to rehabilitation discharge was 74 days (interquartile range, ). For those initially classified as AIS grade A, 78.3% remained AIS grade A at discharge from inpatient rehabilitation, whereas 12.1% converted to AIS grade B and 9.6% converted to motor incomplete. Conversely, of those initially classified as AIS grade C, 77.2% improved to AIS grade D, whereas 22% remained AIS grade C. Those originally classified as AIS grade B had a mixed picture: 49.2% remained motor complete (AIS grade A or B), whereas 27.2% and 23.6% converted to AIS grades C and D, respectively. For the 1-year data, there were 705 subjects with AIS grades (table 2b). There appeared to be continued evolution in AIS grades at 1 year compared with discharge from rehabilitation. The percentage of subjects remaining AIS grade A decreased to 70.2%, whereas the percentage converting to motor incomplete increased to 15.1%, with 7.1% to AIS grade D. For those originally AIS grade B, 33.6% remained motor complete, 29.6% improved to AIS grade C, and 36.8% improved to AIS grade D. For those classified as AIS grade C at baseline, 82.5% improved to AIS grade D or E. At 1 year, 14.1% of those originally AIS grade D had converted to AIS grade E compared with only 1.5% conversion at the time of discharge. Conversion from complete (AIS grade A) to motor incomplete classification was not related to time of examination within the first week of injury ( 2.38; P.54) or to higher single neurologic level of injury ( 2.003; P.96). To examine possible effects of loss to follow-up on conversion from AIS grade A to D, we looked at the first and last available examination for subjects initially classified as AIS grade A. There were 646 subjects, of whom 35 (5.4%) converted to AIS grade D. For the 310 subjects whose last exam- Initial Table 2: AIS Grades From Injury to Discharge and to 1 Year After Injury A B C D E Total n % n % n % n % n % n a. From injury (rows) to discharge (columns) A B C D Total b. From injury (rows) to 1 y after injury (columns) A B C D Total NOTE. Percentages represent row percents.

4 372 MOTOR RECOVERY AFTER CERVICAL SPINAL CORD INJURY, Marino Table 3: Motor Scores at Injury and 1 Year Initial 1 y AIS Grade N Motor Score Mean SD Median IQR Mean SD Median IQR A 336 UEMS LEMS B 125 UEMS LEMS C 109 UEMS LEMS D 135 UEMS LEMS Abbreviation: IQR, interquartile range. ination was within 6 months of injury, the conversion rate to AIS grade D was 3.5% compared with 7.1% for those with an examination later than 6 months postinjury. The conversion rate from AIS grade A to D differed by center from 0.0% to 16.4%, although only 2 centers had rates higher than 7% (rates, 15.4% and 16.4%). Without these 2 centers, the conversion rate would have been 3.5% by using the last available examination or 4.1% by using the 1-year examination. Changes in Motor Scores and Motor Levels For the 1-year data, changes in UEMS or LEMS varied by initial AIS grade (table 3; fig 1). Mean change in UEMS ranged from points for AIS grade A to points for AIS grade C. Mean change in LEMS ranged from points for AIS grade A to points for AIS grade C. Subjects who initially were AIS grade D gained fewer points than those with AIS grade C due to a ceiling effect, but ended up with the highest scores (see fig 1). Final mean UEMS and LEMS for AIS grade D were 44 and 45 points compared with 35 and 38 points for AIS grade C, respectively. There were 315 subjects with complete injuries at admission and analyzable data for motor level and UEMS changes at 1 year. Of these, 46 subjects converted to motor incomplete at the 1-year examination based on LEMS higher than zero. Mean motor scores at the time of injury and at 1 year by motor level are shown in figure 2. Mean change in UEMS was 9 to 11 points, except for C1 to C3 and C8 to T1, for which the change was only 2 to 3 points. A 50 Mean Upper Limb Motor Scores 40 Motor Score Motor score change Initial Motor Score B 0 50 A B C D Initial AIS Grade Mean Lower Limb Motor Scores 40 Motor Score Motor score change Initial Motor Score 0 A B C D Initial AIS Grade Fig 1. (A) Upper- and (B) lower-limb motor scores at 1-year by initial AIS grade.

5 MOTOR RECOVERY AFTER CERVICAL SPINAL CORD INJURY, Marino UEMS Initial 1-year 10 Fig 2. Mean UEMS for subjects with AIS grade A by initial motor level. 0 C1-3 C4 C5 C6 C7 C8-T1 Initial Motor Level For subjects initially classified AIS grade A, the right-side motor level deteriorated 1 to 3 levels in 5.4%, remained the same in 29.5%, and improved 1 level in 41.6%, 2 levels in 14.3%, and more than 2 levels in 9.2%. Results for left-side motor level changes were similar. Longer length of the motor ZPP was associated with a greater likelihood of 2 or more segments of motor level improvement. For subjects with initial motor levels C4 to C7, the relative risk of gaining 2 or more motor levels if motor ZPP was 2 or higher versus 0 to 1 segment was 5.0 (95% CI, ) on the right side ( ; P.001) and 5.0 (95% CI, ) on the left ( ; P.001). DISCUSSION Our understanding of the natural recovery of neurologic function after SCI has evolved with analyses of the NSCISC, EM-SCI, Sygen, and other databases that have tracked key outcomes over time. The knowledge gained from such analyses has served to inform the design of interventional clinical trials, as well as the discussion of prognosis with patients and clinicians. In this report, we have updated neurologic outcomes from the NSCISC database by using data from patients injured between 1994 and It should be noted that a number of variables may influence neurologic recovery in SCI that were not included in this analysis. These include age and plausible, if yet unproved, treatment variables, such as timing and adequacy of decompressive spine surgery, blood pressure management, and neuroprotective steroid therapy. For this reason, application of these data should be undertaken with an understanding of the specific clinical or research context. It also should be noted that neurologic variables in the NSCISC database have changed slightly over the years and data are transcribed from treatment records created by clinicians with varying degrees of experience and who may not have been systematically trained and tested in the application of the ISNCSCI. A number of studies have documented the potential reliability concerns and the benefit of training in the application of the ISNCSCI Based on these concerns, clinical trials such as the Sygen study have required prospective formal training and reliability testing of neurologic examiners, as well interpreters of neurologic data. To address the classification issues, the present study used a computerized neurologic classification algorithm that was cross-checked by experienced SCI medicine clinicians. Although not widely adopted, this approach recently has been reported, 17,18 with reclassification rates similar to our experience. 19 For the most part, results were consistent with prior reports from the NSCISC database. 5 However, we found a higher rate of conversion from complete to incomplete status (30% overall, 15% to motor incomplete) at 1 year postinjury compared with the earlier report (15% overall, 8% to motor incomplete). 5 In particular, 7.1% converted from AIS grade A to AIS grade D, much higher than the 2.3% reported in 1999, 5 but similar to the 7% reported by the EM-SCI. 18 This may occur in part because the present report was limited to subjects with tetraplegia, rather than all levels of SCI. Maynard et al 11 reported that 8% of patients with complete traumatic tetraplegia converted to Frankel grade D or E by 1 year. The recent International Campaign for Cures of Spinal Cord Injury Paralysis systematic review of multiple existing SCI databases, including Sygen and EM-SCI, concluded that only 20% of patients with complete injuries on initial acute examination converted to incomplete status during the first postinjury year, with about half the conversions improving to AIS grade B and the other half improving to motor incomplete status (AIS grade C or D). 3 For the Sygen database, 7.2% of subjects with complete tetraplegia converted to AIS grade D (Benzel 4) at 1 year. 3 A subanalysis of the EM-SCI database in the International Campaign for Cures of Spinal Cord Injury Paralysis review noted that the conversion rate from complete to incomplete status was greater for tetraplegia than paraplegia, with twice as many subjects with tetraplegia converting to AIS grade B or AIS grade D. A more recent EM-SCI database analysis reporting AIS conversion status in the first year after SCI concluded that 70% of patients initially classified as AIS grade A did not convert, whereas 17% improved to AIS grade B; 6%, to AIS grade C; and 7%, to AIS grade D. 18 This report also noted that many of the conversions from AIS grade A were due solely to recovery in the sacral segments. In our study, the definition of recovery to motor incomplete status required motor sparing in at least 1 lower-extremity key muscle. The higher rate of conversion from AIS grade A to D could be due to the quality of the initial examination. On chart review, Maynard 11 found that all patients who converted from Frankel A to Frankel D or E had some degree of TBI, which may have resulted in incorrect initial classification. The SCIMS do not exclude subjects with TBI or other conditions that may impede the initial examination, which may have increased the percentage converting to AIS grade D. In addition, there is no

6 374 MOTOR RECOVERY AFTER CERVICAL SPINAL CORD INJURY, Marino requirement that examiners for the SCIMS undergo training or reliability testing on the ISNCSCI. SCIMS data did not include raw data for sacral sparing; therefore, we cannot be sure that all subjects designated AIS grade A were classified correctly based on sacral sparing. However, EM-SCI and Sygen excluded patients with TBI and required training of examiners, yet had an overall conversion rate similar to ours. Taking into consideration the different nature of the respective databases, our data are consistent with the European and Sygen trial experience. Rates of change in AIS grade for subjects initially classified as incomplete were similar to other reports. For AIS grade B subjects, the EM-SCI group reported that at 1 year, 32.5% remained motor complete, 35% improved to AIS grade C, and 32.5% improved to AIS grade D. 18 Both Sygen and EM-SCI showed that about 80% of subjects initially AIS grade C improved to AIS grade D, 3 which closely matched our results. Motor score recovery in subjects with tetraplegia differed by original AIS grade and for upper and lower extremities. For AIS grade A subjects, upper extremities gained a mean of 8.8 points (median, 7 points), which is in general agreement with prior data from Waters et al 20 (mean, 8.7 points) and Fisher 21 (mean, 8.2 points), but somewhat lower than the Sygen (12 points) and EM-SCI data (14 points). 3 The gain in motor score in the lower extremities also was greatest for subjects who initially were AIS grade C, although the average final LEMS score was higher in AIS grade D subjects. Separating the UEMS from the LEMS showed that persons with motor incomplete injuries (AIS grades C and D) continue to have noticeable weakness in the upper limbs, with a mean residual deficit of 30% (15 of 50 points) for AIS grade C subjects. It was been noted that patients with motor incomplete injuries had greater functional deficits in self-care activities than in ambulation. 22 The UEMS and self-care function may be appropriate outcome measures for neuroprotective or neuroregenerative clinical trials in cervical incomplete SCI. Further examination of the residual UEMS deficit in this and other databases would be valuable to determine its viability as an outcome measure. A number of studies looked at changes in motor levels in persons with neurologically complete lesions, although these studies used different times for the initial examination. In a multicenter study, Ditunno et al 23 found that 70% to 85% of persons with complete tetraplegia recovered to antigravity strength in the next motor segment, depending on the initial motor level. Fischer et al, 21 in a small study, reported that 67% of persons with AIS grade A injury would regain 1 motor level. Our data for motor level recovery (65% gain at least 1 level) were similar to those reported by Fisher, 21 but lower than those reported by others. This may be due to different definitions of motor level recovery, different timing of the initial examination, or a different distribution of motor function in the segment below the initial motor level. Many earlier studies defined recovery as the next key muscle reaching at least grade 3 strength, without consideration of the more rostral muscles. 23 We used the current ISNCSCI definition, which requires the key muscle above the antigravity muscle to be grade 5. Our definition would tend to lower the percentage of subjects recovering a motor level. Timing of the initial examination also can influence observed recovery in the ZPP. There can be relatively large changes in UEMS during the first 2 weeks after SCI, with scores obtained more than 24 hours postinjury better at predicting 6-month scores. 24 Wu et al 25 found that zero grade muscles in the first segment below the motor level were more likely to reach antigravity strength if they were at least grade 1 strength at 1 month compared with muscles that remained grade 0. Persons with sparing of motor function in the first segment below the motor level have a greater chance of recovery to grade 3 or higher in that segment. Waters 20 reported that only 27% of subjects improved to grade 3 or higher at 1 year in a muscle that was initially (at 1 month) grade 0, whereas 97% of muscles graded 1 or 2 did so. Fisher 23 noted that subjects with some motor function in the first segment below the original motor level had a greater chance of that segment reaching antigravity strength than those with a grade 0 key muscle in that segment. The greater the strength in the first level muscle, the greater the chances of improvement. Burns and Ditunno 26 reported that 100%, 75% to 80%, and 25% to 30% of muscles in persons with initial strength of grades 2.5, 1 to 2.5, and 0 would improve to grade 3 or higher at 1 year, respectively. Study Limitations The analyses in this study were limited by the lack of complete neurologic data in the NSCISC database. The lack of complete neurologic data limited our ability to check the classification of sensory and motor levels and AIS grade for errors of interpretation. Our reclassification rate of 10.2% was similar to the 13% reported by Chafetz et al, 19 who also used a computer classification algorithm. The sacral sparing definition of complete/incomplete status remains controversial; however, this data set did not have the raw data for ZPP, S4 to S5 sensation, anal sensation, and voluntary contraction of the anal sphincter to examine the influence of sacral sparing on classification or AIS grade change. We could not confirm completeness of injury by using the sacral-sparing definition and accepted the complete or incomplete designation in our classification algorithm. Incomplete subjects were designated AIS grade B if LEMS was zero and AIS grade C or D if LEMS was greater than zero. The small number of subjects ( 2% of the sample) with initial AIS grade listed as A or unknown but with LEMS greater than zero were dropped from the analysis. The decision to classify subjects as motor incomplete only if they had lower-extremity key muscle function may have underestimated the number of motor-incomplete subjects. Subjects with only voluntary anal contraction or sensory sacral sparing and nonkey muscle function more than 3 levels below the motor level would have been classified as motor complete. However, this decision would avoid the unusual circumstance reported by Spiess et al, 18 in which a change in sensory level (from C4 to C3) resulted in a patient with no LEMS score changing from AIS grade B to D. Although our loss-to-follow-up rate after discharge was approximately 50%, it is similar to prior 1-year data from the SCIMS 5 and to the 6-month follow-up of 72% and 12-month follow-up of 52% reported by the EM-SCI study group. 27 Lack of follow-up was not related to initial AIS grade, but older subjects were less likely to have 1-year data. Given the similar rates of conversion of AIS grades in our data and the Sygen and EM-SCI data, we do not believe that dropouts meaningfully affected the results. CONCLUSIONS Motor recovery of traumatic tetraplegic SCI in the NSCISC database is similar to other published data in terms of AIS grade conversion and motor score improvement. Separating motor scores in upper- and lower-limb components shows different patterns of motor recovery based on AIS grade. Deficits in UEMS persist at 1 year in most motor incomplete patients. Motor level recovery of 2 or more levels in those with

7 MOTOR RECOVERY AFTER CERVICAL SPINAL CORD INJURY, Marino 375 cervical complete SCI occurs in only 23.5% of patients. Motor level recovery in those with AIS grade A SCI and UEMS recovery in AIS grade C SCI may be viable outcome measures for clinical trials. References 1. Curt A, Van Hedel HJA, Klaus D, Dietz V. Recovery from a spinal cord injury: significance of compensation, neural plasticity, and repair. J Neurotrauma 2008;25: Curt A, Schwab ME, Dietz V. Providing the clinical basis for new interventional therapies: refined diagnosis and assessment of recovery after spinal cord injury. Spinal Cord 2004;42: Fawcett JW, Curt A, Steeves JD, et al. Guidelines for the conduct of clinical trials for spinal cord injury as developed by the ICCP panel: spontaneous recovery after spinal cord injury and statistical power needed for therapeutic clinical trials. Spinal Cord 2007;45: Geisler FH, Coleman WP, Grieco G, Poonian D. Measurements and recovery patterns in a multicenter study of acute spinal cord injury. Spine 2001;(26 Suppl 24S):S Marino RJ, Ditunno JFJ, Donovan WH, Maynard FJ. Neurologic recovery after traumatic spinal cord injury: data from the Model Spinal Cord Injury Systems. Arch Phys Med Rehabil 1999;80: Marino RJ, Graves DE. Metric properties of the ASIA motor score: subscales improve correlation with functional activities. Arch Phys Med Rehabil 2004;85: Graves DE, Frankiewicz RG, Donovan WH. Construct validity and dimensional structure of the ASIA motor scale. J Spinal Cord Med 2006;29: Mulcahey MJ, Gaughan J, Betz RR, Johansen KJ. The International Standards for Neurological Classification of Spinal Cord Injury: reliability of data when applied to children and youths. Spinal Cord 2007;45: American Spinal Injury Association. International Standards for Neurological Classification of Spinal Cord Injury, revised 2000, reprinted Chicago: American Spinal Injury Association; American Spinal Injury Association. International Standards for Neurological and Functional Classification of Spinal Cord Injury, revised Chicago: American Spinal Injury Association; Maynard FM, Reynolds GG, Fountain S, Wilmot CB, Hamilton R. Neurologic prognosis after traumatic quadriplegia: three-year experience of California regional spinal cord injury care system. J Neurosurg 1979;50: Brown PJ, Marino RJ, Herbison GJ, Ditunno JF Jr. The 72-hour examination as a predictor of recovery in motor complete quadriplegia. Arch Phys Med Rehabil 1991;72: Burns AS, Lee BS, Ditunno JFJ, Tessler A. Patient selection for clinical trials: the reliability of the early spinal cord injury examination. J Neurotrauma 2003;20: Chafetz RS, Vogel LC, Betz RR, Gaughan JP, Mulcahey MJ. International Standards for Neurological Classification of Spinal Cord Injury: training effect on accurate classification. J Spinal Cord Med 2008;31: Mulcahey MJ, Gaughan J, Betz RR, Vogel LC. Rater agreement on the ISCSCI motor and sensory scores obtained before and after formal training in testing technique. J Spinal Cord Med 2007; (30 Suppl 1):S Priebe MM, Waring WP. The interobserver reliability of the revised American Spinal Injury Association standards for neurological classification of spinal injury patients. Am J Phys Med Rehabil 1991;70: Rupp R, Schweidler J, Curt A, Dietz V, Gerner HJ. An electronic tool for multi-center administration, assessment and analysis of clinical trials in spinal cord injury. Biomed Tech 2005; Spiess MR, Müller RM, Rupp R, Schuld C, Van Hedel HJA. Conversion in ASIA Impairment Scale during the first year after traumatic spinal cord injury. J Neurotrauma 2009;26: Chafetz RS, Prak S, Mulcahey MJ. Computerized classification of neurologic injury based on the international standards for classification of spinal cord injury. J Spinal Cord Med 2009;32: Waters RL, Adkins RH, Yakura J, Sie I. Motor and sensory recovery following complete tetraplegia. Arch Phys Med Rehabil 1993;74: Fisher CG, Noonan VK, Smith DE, Wing PC, Dvorak MF, Kwon B. Motor recovery, functional status, and health-related quality of life in patients with complete spinal cord injuries. Spine 2005;30: Penrod LE, Hegde SK, Ditunno JF Jr. The effect of age on prognosis for functional recovery in acute traumatic central cord syndrome (CCS). Arch Phys Med Rehabil 1990;71: Ditunno JFJ, Cohen ME, Hauck WW, Jackson AB, Sipski ML. Recovery of upper-extremity strength in complete and incomplete tetraplegia: a multicenter study. Arch Phys Med Rehabil 2000;81: Herbison GJ, Zerby SA, Cohen ME, Marino RJ, Ditunno JFJ. Motor power differences within the first two weeks post-sci in cervical spinal cord-injured quadriplegic subjects. J Neurotrauma 1992;9: Wu L, Marino RJ, Herbison GJ, Ditunno J. Recovery of zerograde muscles in the zone of partial preservation in motor complete quadriplegia. Arch Phys Med Rehabil 1992;73: Burns AS, Ditunno JF. Establishing prognosis and maximizing functional outcomes after spinal cord injury: a review of current and future directions in rehabilitation management. Spine 2001; (26 Suppl 24S):S Van Middendorp JJ, Hosman AJF, Pouw MH, Van De Meent H. Is determination between complete and incomplete traumatic spinal cord injury clinically relevant: validation of the ASIA sacral sparing criteria in a prospective cohort of 432 patients. Spinal Cord 2009;47: