Question: Will a clinical prediction rule based on 4 neurological measures and age be a valid measurement for determining ambulation independence 1-year after a traumatic spinal cord injury?

Article: A clinical prediction rule for ambulation outcomes after traumatic spinal cord injury: a longitudinal cohort study


external image cauda-equina.jpg

A longitudinal cohort study of adult patients with traumatic spinal cord injuries was used to form a CPR. Data for neurological and functional status were collected prospectively within the first 15 days and at months 1, 3, 6, and 12 after injury. Treatment of the patients was not standardized. The study looked at patients who were admitted to one of 19 European centers between July 2001 and June 2008. A clinical prediction rule based on age and 4 neurological variables was derived from the international standards for neurological classification of spinal cord injury with a multivariate logistic regression model. The Spinal Cord Independence Measure (SCIM) indoor mobility item (SCIM item 12) was used as the primary outcome measure 1 year after injury. Model performances were quantified with respect to discrimination (area under receiver-operating-characteristics curve [AUC]). Temporal validation was done in a second group of patients from July 2008 to December 2009.

A total of 492 patients had available outcome measures. A combination of age (<65 vs ≥65 years), motor scores of the quadriceps femoris (L3), gastrocsoleus (S1) muscles, and light touch sensation of dermatomes L3 and S1 were used as predictors for level of indoor walking ability. Temporal validation in 99 patients confirmed excellent discriminating ability of the prediction. They included only the best scores of each level of the lower extremity and sacral scores for analysis. The ability to walk independently 1 year after injury was the primary functional outcome.

From the EM-SCI data set they included adult (≥18 years) patients with acute traumatic spinal cord injury, including conus medullaris and cauda equina injuries, who were admitted between July 2001 and June 2008. Patients who were unable to cooperate with physical examination because of cognitive impairment, who had a peripheral nerve lesion, or who had neuropathy or polyneuropathy were not included in the EM-SCI database. Those without a complete neurological assessment within the first 15 days after injury were excluded from the analysis.

Levels of Spinal Cord Injury

Symptoms and Signs of Conus Medullaris and Cauda Equina Syndromes

Conus Medullaris Syndrome
Cauda Equina Syndrome
Sudden and bilateral
Gradual and unilateral
Knee jerks preserved but ankle jerks affected
Both ankle and knee jerks affected
Radicular pain
Less severe
More severe
Low back pain
Sensory symptoms and signs
Numbness tends to be more localized to perianal area; symmetrical and bilateral; sensory dissociation occurs
Numbness tends to be more localized to saddle area; asymmetrical, may be unilateral; no sensory dissociation; loss of sensation in specific dermatomes in lower extremities with numbness and paresthesia; possible numbness in pubic area, including glans penis or clitoris
Motor strength
Typically symmetric, hyperreflexic distal paresis of lower limbs that is less marked; fasciculations may be present
Asymmetric areflexic paraplegia that is more marked; fasciculations rare; atrophy more common
Less frequent; erectile dysfunction that includes inability to have erection, inability to maintain erection, lack of sensation in pubic area (including glans penis or clitoris), and inability to ejaculate
Sphincter dysfunction
Urinary retention and atonic anal sphincter cause overflow urinary incontinence and fecal incontinence; tend to present early in course of disease
Urinary retention; tends to present late in course of disease

Potential Functional Outcomes at 1 Year Postinjury Tables

Physical Therapy Exercises

C5, C6 Complete SCI

Walking Efforts After Severe Spinal Cord Injury

Ekso Bionics Highlights

Bionic Exoskeleton

C5, C6 Quadriplegic Driving with EMC Equipment

The Clinical Prediction Rule:

  1. Age: ≥65 years old
  2. Motor: muscle grade at L3 (quadriceps femoris) and S1 (gastrocsoleus)
  3. Sensory: light touch sensation (LTS) at L3 and S1 dermatome
*Only the best score from each motor or LTS (right or left) was used for scoring

Table 2 (1).png

Probability of walking independently 1 year after injury based on the prediction rule score
Figure 2 (2).png

Prognostic Variables:

Age was dichotomized into <65 and ≥65 years old. Age was determined by a previously published article [1] which stated that patients classified as elderly (≥65 years old) have a higher mortality rate during the first 6 weeks post-SCI. Another finding from this article was that elderly patients make larger improvements in sensory function; however, they are not able to translate their gains into functional improvement as well as SCI patients in the younger population. Based off of this information, it was determined that patients in a younger population will make better functional gains and this will translate to a better outcome measurement of independent ambulation 1 year post-injury.

Neuromuscular Examination: (All tested in supine)
Motor Score:
The motor score test was a 5 point scale adapted from Medical Research Council scale.
Grades 5-0 (3).jpg

Light Touch Sensory (LTS):

Sacral Sparing Scores:
This includes voluntary anal contraction and anal sensation.

*Pinprick Sensory (PPS) was tested, but not included in the model because PPS and LTS are highly correlated and LTS was chosen because it is least prone to error.

The American Spinal Injury Association/International Spinal Cord Society (AIS) neurological standard scale was used in order to establish baseline characteristics of the patients. This widely used scale is vital in diagnosing the extent of a patient’s injury. The CPR that was derived used pieces of this scale to predict independent ambulation 1 year after injury. [2]

Panel 1 (4).png

The American Spinal Injury Association (ASIA) has provided the International Standards for Neurological Classification of Spinal Cord Injury (ISNCSCI) worksheet, which is widely used in order to determine amount of impairment from a SCI. The evaluation is based on the summation of scores from motor and sensory (light touch and pin prick) performance collected from the right and left sides independently. This worksheet can also be used to determine if the patient has suffered a complete or incomplete SCI. According to the AIS, if there is no motor or sensory function at the S4-5 level, the injury is considered complete and labeled as a Grade A injury. The four criteria for this are as follows:
  1. Voluntary anal contraction score = 0
  2. Anal sensation score = 0
  3. Light touch score = 0
  4. Pin prick score = 0
*Motor score (0-5) and Sensory score (0-2)
In order to determine Grades B-E on the AIS, there is a flow-chart to determine where the patient lies.
Steps in Classification (5).png
The Spinal Cord Independence Measure (SCIM) item 12-mobility indoors, ability to walk <10m, was assessed and analyzed to determine the ability to walk independently 1 year after injury. This scale is validated and widely used in order to classify the ambulation level of a patient and it ranges from total assistance (0), to moves independently in wheelchair (2), to walks with one cane (6), to walks without walking aids (8).[3,4] The scale was dichotomized in order to classify the patient as dependent or independent 1 year after injury. Scores 0-3 were considered dependent or needing supervision during ambulation, while scores 4-8 were considered independent even if an assistive device was used.

Panel 2 (6).png

Ancillary analyses showed that neither level of injury nor timing of examination had significant additional value with respect to prediction of ability to walk independently. Also, the addition of PPS scores to final model did not significantly improve its fit.

Does the rule work?

Once the prediction rule was created, a validation group was created to see if the rule worked. 389 adult patients with traumatic spinal cord injury were admitted to 13 different EM-SCI centers. 214 of these patients fit into the criteria needed for the study. This analysis was done before some of the 1-year follow up data was available, so a smaller portion of the validation group patients had information available as compared to the derivation group (25% compared to 59%). In the final calibrations only 99 patients from the validation group were used. The prediction rule worked extremely well (AUC 0.967, 95% CI 0.939 – 0.995, p < 0.0001). PPS scores at L5 were added, and the prediction rule resulted in a lower AUC (0.964, 95% CI 0.935 – 0.994, p < 0.0001).

This was a form of external validity that the researchers conducted, but it was concluded by van Middendorp et al. that, “Before application of the prediction rule in clinical practice, an external validation study is needed to assess its generalizability” [5]. Additionally it was noted that it needed to be established whether or not the CPR resulted in better use of rehab resources and if it was able to increase the psychological health of the SCI patients. A further study conducted by Rashmi Malla, The University of Texas School of Public Health, sought to validate this CPR. Malla applied the CPR to 231 SCI cases from the NACTN (North American Clinical Trials Network) and found an ROC (area under the receiver-operating-characteristics) curve of 0.927 (95% CI 0.894-0.959) [6]. This is compared to the AUC (area under the receiver-operating-characteristics) curve for the van Middendorp et al. trial which found an AUC of 0.956 (95% CI 0.936-0.976). The ROC found by Malla is smaller than the AUC found by van Middendorp et al., but it is still valid to say that the CPR is able to distinguish between those who did and those who did not achieve independent ambulation in the sample population. Overall Malla’s study found that the higher the prediction score, the better the probability of walking as well as providing some further external validity for the van Middendorp et al. study.

Figure 3 (7).png

There are some limitations to this study including the fact that the doctors utilizing this CPR must have experience in examining patients with spinal cord injuries which means that this is not really a CPR that can be used by PTs. Additionally, this prediction rule can be applied only to indoor distances, it does not define the quality of walking, many patients had acute-phase measures missing, and it could be said that the CPR is over-optimistic due to details of patients lost to follow up not being documented.

There a several strengths to this study including the fact that the timing of the examination of the patient after injury did not have an effect on the prediction rule (ie <24h, <72h, or <15 days after injury), there was no difference in outcomes between patient with tetraplegia and paraplegia, the data came from good database with a large sample size, the patients initial neurological impairments were assessed by trained and certified physicians, the researchers used a well validated clinical outcome measure for ambulation (SCIM), and the researchers conducted a temporal validation of derived the CPR with excellent results.

Based on the results of the van Middendorp study, comparisons to other spinal cord injury measures such as the AIS and SCIM, and an external validity study it can be concluded that yes, this is a valid clinical prediction rule. The main focus of this CPR is ambulation so it is not able to predict the multitude of other factors to be considered in SCI patients, it is just meant to be a quick and accurate substitute for the AIS. As physical therapists we just need to be aware of the differences in the AIS grades and in reality it would be an experienced physician who would be performing the test. Based on the APTA Guide to PT Practice we would perform some common tests such as the Physical Disability Index, the TUG, the Berg Balance Scale, the Functional Reach Test, the Tinetti Balance Scale, the Wheelchair Mobility Assessment Tool, and Gait Speed in order to predict ambulation outcomes . Refer to Pattern 5H in the PT Guide for the full list of tests to use with traumatic SCI patients.

  1. Furlan JC, Fehlings MG. The impact of age on mortality, impairment, and disability among adults with acute traumatic spinal cord injury. //J Neurotrauma// 2009; 26: 1707-17
  2. American Spinal Injury Association: International standards for neurological classification of spinal cord injury, revised 2011. Chicago, IL: American Spinal Injury Association; 2011.
  3. Catz A, Itzkovich M, Steinberg F, et al. The Catz-Itzkovich SCIM: a revised version of the spinal cord independence measure. //Disabil Rehabil// 2001; 23: 264-68.
  4. Catz A, Itzkovich M, Tesio L, et al. A multicenter international study on the spinal cord independence measures, version III: Rasch psychometric validation. //Spinal Cord// 2007; 45: 275-91
  5. Middendorp JJ, Hosman AJ, Donders AR, et al. A clinical prediction rule for ambulation outcomes after traumatic spinal cord injury: a longitudinal cohort study. //Lancet//. 2011; 377: 1004-10. doi:10.1016/S0140-6736(10)62276-3.
  6. Malla R. External validation study of a clinical prediction rule for ambulation outcomes after traumatic spinal cord injury. //Texas Medical Center Dissertations (via ProQuest).// 2013:Paper AAI1541012.
  7. Available at: Accessed April 6, 2014.
  8. Available at: Accessed April 8, 2014.
  9. Available at: Accessed April 8, 2014.
  10. Accessed April 8, 2014.