Spinal muscular atrophy (SMA) is a disease of anterior horn cells that results from a homozygous mutation in the survival of the motor neuron (SMN) 1 gene. The incidence of SMA is 1 in 6000 to 1 in 10 000,1,2 and it presents phenotypically with a spectrum of disease severity. Muscle weakness, which is the hallmark of this disease, is the result of diminished levels of SMN, the protein product of the SMN1 gene, which is found on chromosome 5q13.3 The presence of an almost identical homologous centromeric copy of the gene, SMN2, partially compensates for the absence of normal SMN protein production from a SMN1 mutation, and acts to rescue the phenotype by production of some full-length SMN protein, albeit in diminished amounts.4 In general, the greater the number of copies of SMN2 the milder the phenotype observed.5,6 The most severely affected patients with SMA have SMA type I (SMA-I) with severe muscle weakness and impairment of their gross motor skills. In the past, patients with SMA-I have not typically survived beyond the first 2 years of life. In recent years, however, it has become apparent that with the increased use of noninvasive ventilation and supplemental feeding by gastrostomy tube, the natural history of SMA-I has changed. In one recent series, 74% of patients survived beyond 24 months and 50% beyond 10 years of age. Some, but not all, of this increase in survival was thought to be related to the use of chronic ventilation.7 In addition, SMA-I has been recognized to represent a spectrum of disease severity. As a result, some clinicians subdivide classical SMA-I into 3 subcategories, SMA Ia, SMA Ib, and SMA Ic, which correspond to decreasing severity of disease.8
In the SMA-I patient population, tests designed to measure typical motor development present a significant floor effect. Most tests that rely on the normal developmental motor sequence to quantify a patient's skill are not appropriate for this population because patients with SMA-I are, by definition, unable to sit and rarely roll or prop in the prone position. The Test of Infant Motor Performance is one infant examination that was explored as a potential outcome measure for use in this population. However, despite its good reliability,9 it was poorly tolerated in our initial trials, primarily because the items performed in the prone position tended to tax the respiratory reserve of patients with SMA-I, who obligatorily demonstrated diaphragmatic breathing. As a result of this experience, we set out to develop a test that would be well tolerated, reflect the variety of motor skills of patients with SMA-I, provide a tool that would be sensitive to the typical progression of motor skills in this population, and have the potential to serve as a clinically meaningful outcome measure in the context of a clinical trial.
Motor outcome measures that are valid and reliable are important for monitoring the status of patients with SMA-I and are a prerequisite for including this population in clinical trials. Because of the extent of motor impairment seen in these patients, it has been difficult to quantify motor abilities in an objective manner. Because life span is usually limited by respiratory status in patients with SMA-I, the time to reach 16 h/d mechanical ventilation has been proposed as a surrogate measure for time to death, the primary endpoint now considered for SMA-I in clinical trials.8 Valid and reliable motor outcome measures, however, are still needed for this population and will likely be a secondary outcome in any clinical trial. In our previous work10 we described the development of The Children's Hospital of Philadelphia Infant Test of Neuromuscular Disorders (CHOP INTEND) and established its reliability and content validity. An item bank was selected that consisted of a wide variety of items that were felt to represent motor skills of infants who are weak. These test items were both of our own construction and taken from other tests. We tested these items in a group of subjects with SMA-I and in other infants with neuromuscular disorders who were weak. A rigorous statistical investigation of item behavior in the target population was undertaken to assure that the selected items possessed optimal characteristics related to internal consistency, response range, and discrimination, and would capture the full range of motor skill in the target population while avoiding item redundancy. Finally, an expert panel reviewed item construction, content, and applicability to the domain of motor skill, to assure that the selected items fell within the domain of motor skill of patients with SMA-I.10 In this investigation, we have extended this initial effort by examining the validity of the CHOP INTEND through correlation with patient age, bilevel positive airway pressure (BiPAP) usage, and electrophysiological measures. In addition, we have explored the capacity of the CHOP INTEND to differentiate between patients with 2 SMN2 copies and those with 3 SMN2 copies, as well as those on BiPAP and those who did not require BiPAP.
A total of 27 subjects with SMA-I were evaluated on the CHOP INTEND as part of an ongoing institutional review board–approved multicenter natural history study. Subjects were recruited from the neuromuscular clinics of each Pediatric Neuromuscular Clinical Research site, as previously reported,10 and represent a balance of patients with newly diagnosed or chronic SMA-I. The study was institutional review board approved at each site and each parent or subject provided informed consent or assent as appropriate. Subjects were evaluated with the CHOP INTEND as part of a comprehensive longitudinal natural history evaluation. The subject's age at symptom onset was determined by history. The need for BiPAP respiratory support was recorded for each subject. Subjects who consented to venipuncture had their SMN2 copy number determined by a previously described methodology11 using a centralized molecular genetic core laboratory (Wendy K. Chung). The compound motor action potential (CMAP) was determined by supramaximal stimulation of the ulnar nerve at the wrist with subsequent determination of single motor unit action potentials. Motor unit number estimation (MUNE) was determined by the multiple point stimulation method and was derived by dividing the amplitude of the CMAP by the average amplitude of approximately 10 single motor unit action potentials.12
Pearson product-moment correlations were used to correlate the electrophysiological parameters with the INTEND scores, and t tests were used to compare differences between the means of subgroups based on BiPAP usage. An analysis of covariance model was used to determine the relationship between the CHOP INTEND score and age. The model included SMN2 copy number, age, and the interaction between SMN2 copy number and age as independent variables.
A total of 27 subjects with SMA-I were evaluated on the CHOP INTEND. Sixteen of the subjects were male and 11 were female, with a mean (± SD) age of 48.2 months ± 69 (range, 3.8–260) or an average age of just more than 4 years. The mean age of symptom onset was 3.3 months ± 1.5 (range, 1–7). Fifty-two percent of the subjects were classified with the more severe type Ib (n = 14) and 48% with the relatively milder type Ic (n = 13). None had the congenital type Ia form. Fifteen of the subjects required BiPAP support at the time of the CHOP INTEND evaluation and 12 did not. No subject required invasive respiratory support.
Significant Pearson product-moment correlations were noted between the CHOP INTEND score and age (Figure 1) (n = 27; r = −0.51; P = .007) and months since symptom onset (n = 27; r = −0.49; P = .005). When patients were categorized by SMN2 copy number, the correlation between age and copy number was stronger. In patients with 2 copies of SMN2, r = −0.60 (n = 16; P = .015), and in those with 3 copies, r = −0.83 (n = 9; P = .005).
An analysis of covariance model for the CHOP INTEND score that included SMN2 copy number (n = 25) and age as well as their interaction as independent variables accounts for approximately half of the variance in CHOP INTEND (R2 = −0.49) with the difference between the slope of the regression lines between CHOP INTEND and age for patients with 2 SMN2 copies and 3 SMN2 copies approaching significance with a P value of .0711(Figure 2). The equation for predicting the CHOP INTEND score for patients with 2 copies of SMN2 is y = −3.61 (age) + 29.6, and for patients with 3 copies, it is y = −1.22 (age) + 31.4.
Patients who required BiPAP (n = 15, mean = 15.2 ± 10.2) and those who did not have a requirement for BiPAP (n = 12, mean = 31.2 ± 4.2) scored significantly differently on the CHOP INTEND (P < .0001) (Figure 3). The difference between patients requiring more than 8 hours of BiPAP and those needing less than 8 hours was also significant (n = 26, P < .0001). The duration of BiPAP required had a significant correlation with the CHOP INTEND score (r = −0.74; P < .0001). This difference was not due to group differences in age (Figure 4). All subjects over the age of 5 years (n = 6) were using BiPAP and had CHOP INTEND scores of 21 or less, with the exception of 1 subject with a CHOP INTEND score of 32. In subjects younger than 5 years (n = 21), CHOP INTEND scores were 26 or higher for those not using BiPAP (n = 12), and most (6 of 9) were less than 15 in those using BiPAP. Correlations between the CHOP INTEND score and the electrophysiological measures were not significant with Pearson product-moment correlations between the INTEND and the MUNE (n = 17; r = −0.04, P = .087) and CMAP (n = 17; r = −0.16, P = .53), failing to reach significance.
The aim of this investigation was the validation of a motor assessment tool that accurately represents the motor skills of infants and children who are weak across the phenotypic spectrum of SMA-I and would reflect the variation of other measures of disease severity. In our initial investigation,10 which included test development, we proceeded through an iterative process of item selection and analysis to assure that the resulting test would fully reflect the domain of motor skill while avoiding item redundancy. Here we have looked at factors that were felt to reflect or correlate with overall disease severity. The first, SMN2 copy number, which is a genetic modifier, remains static irrespective of the patients' age and the progression of the disease process, but tends to have an inverse relationship with disease severity at least across the full disease spectrum from type I to type III. We explored whether such a relationship exists within the subtypes of SMA Ib and Ic across the age spectrum represented in this cross-sectional sample of patients with newly diagnosed and chronic SMA-I. The other factors that we used to correlate with disease severity were patient age, BiPAP need requirement, and electrophysiological measures.
The patients with the most severe form of SMA-I typically do not survive to older ages, and as a result, any cross-sectional sample will be biased by selection of those who survive to complete the study. The presumed overrepresentation of patients with more severe disease at younger ages and patients with less severe disease at older ages will tend to diminish the correlation of the CHOP INTEND with age. Nonetheless, there was still a significant correlation between age and CHOP INTEND score (r = −0.51) that increased when patients were divided into groups based on copy number (2 copies, r = −0.60; 3 copies, r = 0.83). When subjects with 2 copies of SMN2 and subjects with 3 copies of SMN2 were compared and age was taken into account, the regression line between age and CHOP INTEND approached statistical significance (P = .0711). This analysis was somewhat limited by the fact that the patients with 2 copies of SMN2 are spread over a smaller age range, with only 1 patient with 2 SMN2 copies more than 45 months of age, as compared with 4 patients with 3 SMN2 copies more than 12 years of age. The selection bias that is created by the terminal nature of SMA-I limits this type of analysis. A visual inspection of the data (Figure 2) provides additional information regarding the relationship between age, CHOP INTEND score, and SMN2 copy number, and the effect of the age distribution of patients with 2 SMN2 copies as compared with those with 3 SMN2 copies.
The CHOP INTEND was able to be used to differentiate between patients with and without a BiPAP requirement, and also correlated with hours of BiPAP needed. This provides 1 measure of concurrent validity. CMAP and MUNE did not significantly correlate with the CHOP INTEND scores. This likely reflects the uniformly low baseline values for the CMAP and MUNE in SMA-I and, in effect, indicates a floor effect for these electrophysiological measures of motor unit number and activation. Swoboda et al13 have shown that CMAP and MUNE correlate with SMA type and SMN2 copy number across the spectrum of all 3 types of SMA, where there is significantly wider functional range and more variability in the MUNE and CMAP than what is seen in the SMA-I population. Subjects with SMA-I have limited variability of the MUNE and CMAP, especially after the initial precipitous decline at and just before symptom onset, and increasing age has not been found to be significantly associated with these measures within the type I population.
Currently, there is no validated motor assessment tool specifically designed for patients with SMA-I. Our earlier report of the reliability of the CHOP INTEND and this initial validation study provide support for continued exploration of this test instrument both as a clinical measure and as an outcome measure for clinical trials in patients with SMA-I. This study is limited by the relatively small sample size and the mixture of patients with newly diagnosed and more chronic SMA-I. As only baseline data were analyzed here, there was no opportunity to examine the relative sensitivity of the CHOP INTEND to change over time in this population. Larger long-term natural history studies are necessary to address these concerns. As SMA types II and III have been shown to be remarkably stable over periods of years, it is unclear how long the observation period would need to be to demonstrate a decline of motor skill in the SMA-I population. The use of the CHOP INTEND in a clinical trial, however, could provide a continuous measure that could be tracked longitudinally to determine a response to treatment as compared with the dichotomous measure of time to mechanical ventilation that can be determined only when the end point has been reached.
The initiation of BiPAP, used in this study as a marker of pulmonary disease severity, is subject to potential variation in clinical judgment or in duration of usage. Practice standards at the 3 participating clinical centers of the Pediatric Neuromuscular Clinical Research are similar, however, and the effect of this factor is therefore felt to be small. The need for noninvasive ventilation is currently the best marker of pulmonary status in patients with SMA-I.
This cross-sectional study provides support for the validity of the CHOP INTEND. Here we have provided initial validity data based on the relationship between the CHOP INTEND and subject age, BiPAP usage, and SMN2 copy number. As future interventions and, hopefully, targeted drug treatments become available for SMA-I, the CHOP INTEND will serve as a useful tool for measuring their effect.
This work would not have been possible without the generous participation of the children and the commitment of their families, whom the authors would also like to thank. The authors appreciate the support of the many physicians who referred patients to authors' centers.
1. Burd L, Short SK, Martsolf JT, Nelson RA. Prevalence of type I spinal muscular atrophy in North Dakota. Am J Med Genet. 1991;41(2):212–215.
2. Emery AE. Population frequencies of inherited neuromuscular diseases—a world survey. Neuromuscul Disord. 1991;1(1):19–29.
3. Lefebvre S, Burglen L, Reboullet S, et al. Identification and characterization of a spinal muscular atrophy–determining gene. Cell. 1995;80(1):155–165.
4. Vitali T, Sossi V, Tiziano F, et al. Detection of the survival motor neuron (SMN) genes by FISH: further evidence for a role for SMN2 in the modulation of disease severity in SMA patients. Hum Mol Genet. 1999;8(13):2525–2532.
5. Sumner CJ. Molecular mechanisms of spinal muscular atrophy. J Child Neurol. 2007;22(8):979–989.
6. Elsheikh B, Prior T, Zhang X, et al. An analysis of disease severity based on SMN2 copy number in adults with spinal muscular atrophy. Muscle Nerve. 2009;40(4):652–656.
7. Oskoui M, Levy G, Garland CJ, et al. The changing natural history of spinal muscular atrophy type 1. Neurology. 2007;69(20):1931–1936.
8. Bertini E, Burghes A, Bushby K, et al. Outcome measures and treatment of spinal muscular atrophy. Presented at: 134th ENMC International Workshop; February 11–13, 2005; Naarden, The Netherlands. Neuromuscul Disord. 2005;15(11):802–816.
9. Finkel RS, Hynan LS, Glanzman AM, et al. The test of infant motor performance: reliability in spinal muscular atrophy type I. Pediatr Phys Ther. 2008;20(3):242–246.
10. Glanzman AM, Mazzone E, Main M, et al. The Children's Hospital of Philadelphia Infant Test of Neuromuscular Disorders (CHOP INTEND): test development and reliability. Neuromuscul Disord. 2010;20(3):155–161.
11. Stratigopoulos G, Lanzano P, Deng L, et al. Plastin 3 expression is associated with SMA disease severity only in post-pubertal females. Arch Neurol. 2010;63(10):1252–1256.
12. Gooch C, Kaufmann P. Multiple point stimulation motor unit number estimation with single motor unit tracking in a theraputic ALS trial. Suppl Clin Neurophysiol. 2003a;55:284–285.
13. Swoboda KJ, Prior TW, Scott CB, et al. Natural history of denervation in SMA: relation to age, SMN2 copy number, and function. Ann Neurol. 2005;57(5):704–712.