Figures
Abstract
Background
Valproic acid (VPA) has demonstrated potential as a therapeutic candidate for spinal muscular atrophy (SMA) in vitro and in vivo.
Methods
Two cohorts of subjects were enrolled in the SMA CARNIVAL TRIAL, a non-ambulatory group of “sitters” (cohort 1) and an ambulatory group of “walkers” (cohort 2). Here, we present results for cohort 1: a multicenter phase II randomized double-blind intention-to-treat protocol in non-ambulatory SMA subjects 2–8 years of age. Sixty-one subjects were randomized 1∶1 to placebo or treatment for the first six months; all received active treatment the subsequent six months. The primary outcome was change in the modified Hammersmith Functional Motor Scale (MHFMS) score following six months of treatment. Secondary outcomes included safety and adverse event data, and change in MHFMS score for twelve versus six months of active treatment, body composition, quantitative SMN mRNA levels, maximum ulnar CMAP amplitudes, myometry and PFT measures.
Results
At 6 months, there was no difference in change from the baseline MHFMS score between treatment and placebo groups (difference = 0.643, 95% CI = −1.22–2.51). Adverse events occurred in >80% of subjects and were more common in the treatment group. Excessive weight gain was the most frequent drug-related adverse event, and increased fat mass was negatively related to change in MHFMS values (p = 0.0409). Post-hoc analysis found that children ages two to three years that received 12 months treatment, when adjusted for baseline weight, had significantly improved MHFMS scores (p = 0.03) compared to those who received placebo the first six months. A linear regression analysis limited to the influence of age demonstrates young age as a significant factor in improved MHFMS scores (p = 0.007).
Citation: Swoboda KJ, Scott CB, Crawford TO, Simard LR, Reyna SP, Krosschell KJ, et al. (2010) SMA CARNI-VAL Trial Part I: Double-Blind, Randomized, Placebo-Controlled Trial of L-Carnitine and Valproic Acid in Spinal Muscular Atrophy. PLoS ONE 5(8): e12140. https://doi.org/10.1371/journal.pone.0012140
Editor: Isabelle Boutron, University Paris Descartes, France
Received: June 9, 2009; Accepted: July 13, 2010; Published: August 19, 2010
Copyright: © 2010 Swoboda et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was funded by Families of Spinal Muscular Atrophy. The investigation was also supported by grants UL1RR025764 (University of Utah, Center for Clinical and Translational Sciences), UL1RR025005 (Johns Hopkins University, General Clinical Research Center), UL1RR025011 (University of Wisconsin, Clinical and Translational Research Core) M01-RR00034 (Ohio State University, General Clinical Research Center) and the Clinical Research Center, Children's Hospital of Michigan, given by the National Center for Research Resources, National Institutes of Health. The funding sources had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Spinal muscular atrophy (SMA) is an autosomal recessive motor neuron disease and a leading cause of infant and childhood mortality. [1]–[6] More than 95% of affected individuals demonstrate a homozygous deletion/mutation involving exon 7 in SMN1 (survival motor neuron 1), resulting in the biochemical deficiency of the SMN protein, part of a complex that functions in the assembly of small nuclear ribonucleoprotein particles (snRNP). [7], [8] A genomic duplication at this locus has produced a nearly identical gene, SMN2 (survival motor neuron 2) that differs from SMN1 by a nucleotide substitution that promotes exon 7 exclusion, thus producing only a fraction of the identical full length protein. Phenotypic variation in SMA correlates with the number of SMN2 gene copies and the level of SMN protein in cells [9]–[15].
Because all SMA patients have 1 or more SMN2 gene copies, small molecule compounds that target SMN2 to produce increased quantities of full-length SMN protein from the existing SMN2 gene(s) are attractive therapeutic candidates. Valproic acid (VPA), a histone deacetylase (HDAC) inhibitor, directly increases SMN expression in SMA patient-derived cell lines in vitro. [16], [17] Preliminary in vivo data in humans demonstrates up-regulation of SMN2 expression in about one-third of SMA subjects via inhibition of HDAC2 and also appears to alter splicing to increase SMN7 inclusion and thus full-length SMN protein. [18], [19] VPA has also been demonstrated to have neuroprotective properties on glutamate-induced excitotoxicity via up-regulation of alpha-synuclein and increases neurite outgrowth in vitro. [20], [21] VPA treatment has been demonstrated to increase survival in animal models of motor neuron disease including amyotrophic lateral sclerosis and SMA. [22]–[24] In an SMA mouse model, VPA treatment resulted in improved gross motor function, larger evoked motor potentials, less degeneration of spinal motor neurons and improved neuromuscular junction innervation in treated animals compared to age-matched controls [24]. Three open label trials of VPA in human subjects have been published to date, all indicating a possible modest benefit in strength and/or motor function [25]–[27].
As children with SMA may have a limited carnitine synthetic capacity due to significantly diminished skeletal muscle mass, and VPA is known to inhibit carnitine transport and deplete carnitine levels by binding to VPA metabolites, we elected to combine VPA therapy with sufficient supplemental carnitine to avoid concerns about a confounding effect of carnitine depletion. This decision was influenced by data from our open label VPA trial, which indicated an increased suspectibility in SMA subjects to carnitine depletion with VPA treatment [27].
Our primary objective was to assess potential benefit for improving motor function in a young non-ambulatory cohort of children with SMA in a randomized double-blind placebo-controlled clinical trial. Additional objectives were to further assess the safety of VPA in children with SMA, to assess performance of selected outcome measures for use in a multi-center clinical trials setting and to look for evidence that would support a biologic effect of VPA in these subjects, such as changes in SMN expression, electrophysiologic outcomes, or both.
Methods
Ethics Statement
Written informed parental consent (subjects <18 years) and assent (subjects ≥7 years) were obtained for all subjects. The study was approved by the University of Utah Institutional Review Board (IRB) and at each participating clinical trial site (University of Wisconsin-Madison Health Sciences; Wayne State University; Ohio State Biomedical; Johns Hopkins Medical and Ste Justine Hospital.
Trial Design
The SMA CARNI-VAL trial was a multi-center phase II trial of CARNItine and VALproic acid in patients with spinal muscular atrophy. This trial consisted of two parallel multi-center studies, targeting different SMA cohorts (Clinicaltrials.gov ID NCT00227266). The protocol for this trial and supporting CONSORT checklist are available as supporting information; see Checklist S1 and Protocol S1. Part 1 was a prospective double blind, placebo-controlled randomized intention-to-treat protocol to assess safety and efficacy of VPA and L-carnitine in non-ambulatory SMA type II or III “sitters” 2–8 years of age (cohort 1). Part 2 was a parallel open label trial in SMA type II or III “standers and walkers” 3–17 years of age (cohort 2). In this paper, we review objectives, methods and results only for Part 1.
Study population
We prospectively enrolled 61 non-ambulatory SMA children at six centers in North America. The progress of all participants through the trial is diagrammed in Figure 1. Inclusion and exclusion criteria are indicated in Figure 2.
Study procedures
All subjects completed two baseline visits within a two to six week period. Following the second visit, subjects were randomized to either the placebo arm (group 1) or the treatment arm (group 2). There were 31 subjects randomized to the placebo arm and 30 subjects randomized to the active treatment arm (Figure 1). Placebo was provided for both L-carnitine and VPA. Subjects and investigators remained blinded for the duration of the study. At the six-month visit, subjects in group 1 were switched to active treatment for the second six-month period, and those in group 2 continued on study medication for an additional 6 months. Thus, half of the children received treatment for the last 6 months of the study, and half received active treatment for the full 12 months. This trial design was intended to improve trial enrollment and compliance with the protocol, as focus group discussions with parents of eligible children indicated significant reluctance to enroll in a trial of a licensed drug in which there was a 50% chance of receiving only placebo. Nonetheless, extension of the trial to 12 months in this manner was thought to be potentially useful to study both the benefit, and burden, of longer treatment duration.
Abbott Pharmaceutical provided VPA and placebo, and Sigma-Tau Pharmaceutical provided L-carnitine, at no cost. Divalproex sodium coated particles (Depakote® sprinkle capsules, 125 mg) or matched placebo was administered in divided doses two to three times daily sufficient to maintain overnight trough levels of 50–100 mg/dL. L-carnitine (or matched placebo, 100 mg/ml liquid) was dosed at 50 mg/kg/day, to a maximum of 1000 mg, divided into two daily doses.
Treatment assessments were performed at 3 (V1), 6 (V2) and 12 (V3) months. Laboratories were performed at baseline, 2–3 weeks following initiation, at each treatment visit and midway between V2 and V3 visits, and included a basic chemistry profile, CBC with platelets, transaminases, carnitine profile, amylase, lipase and trough VPA levels. An un-blinded medical monitor, who also reviewed subjects' blood tests and adverse events, performed dosing adjustments. Similar adjustments were made for subjects in the placebo group to preserve blinded status of participants. Contact was possible between the unblinded medical monitor and patients if required by clinical concerns, but efforts were taken to avoid unblinding of the site investigator, parents and patient.
Outcome measures
Primary outcome measures included laboratory safety data, adverse event data and efficacy as measured by change from baseline Modified Hammersmith Functional Motor Scale (MHFMS) scores following six months of treatment versus placebo. Secondary outcome measures included change from baseline MHFMS score following 12 versus 6 months treatment, estimates of innervation via maximum ulnar compound muscle action potential (CMAP), dual-energy X-ray absorptiometry (DEXA), evaluation of body composition and bone density, quantitative assessment of SMN mRNA; evaluation of quality of life using the Pediatric Quality of Life Inventory (PedsQL) and for children five years and older, change from baseline measures of pulmonary function and muscle strength via handheld myometry at 6 and 12 months.
Gross motor function was assessed with the MHFMS. This scale was previously modified for use in a research setting but contains the same 20 gross motor function items as the original clinical scale (complete protocol available at http://smaoutcomes.org). [28]–[29] Degree of innervation by the ulnar nerve in the hypothenar muscle group of the hand was estimated using maximum ulnar CMAP amplitude (complete CMAP protocol available at http://smaoutcomes.org). In children five years and older, myometry measurements were performed three times for right and left elbow flexion, and for right and left knee extension, at each visit using the Lafayette Instrument MMT System Model 01163 myometer with a positioning method as previously described. [30], [31] If any measure was more than 10% different than any of the other two measurements, then a fourth measurement was performed. The summary statistic was the average of either three or four measurements taken during a visit for both the right and left elbow and knee. [32]–[34] In children five years and older, pulmonary function testing (PFT) was performed in an accredited pediatric laboratory and included forced vital capacity (FVC), forced expiratory volume in one second (FEV1) and maximum expiratory and inspiratory pressures (MEP, MIP). Dual-energy X-ray absorptiometry (DEXA) scanning for bone density and body composition was performed at the Salt Lake City, Madison and Columbus sites. Norland DEXA XR-36 software version 3.3.1 for small subjects was used to assess whole body composition and bone mineral density (BMD) and bone mineral content (BMC). Quality of life (QOL) was assessed using the PedsQL [35]. The same parent completed the PedsQL at each visit and children five years of age and older completed the age-appropriate PedsQL. The questionnaire is scored on a 0 to 100 scale such that a higher score indicates better QOL.
SMN2 copy number was determined as previously described. [11] Whole blood was drawn into PAXgene tubes at baseline and each visit and quantitation of full length SMN (flSMN) and SMNΔ7 mRNA transcripts was performed as previously described. [27], [36] Human RPLPO (large ribosomal protein) and PGK1 (phosphoglycerate kinase 1) were run as endogenous controls. Subjects missing baseline data or who had incomplete visit data were excluded from this analysis. Results are reported as relative amounts of flSMN or SMNΔ7 transcripts normalized against the relative amount of RPLPO.
Adverse events data management
Adverse events were elicited systematically via a full review of systems at the time of each visit, and parents were instructed to immediately report symptoms or illnesses occurring between visits via telephone. Study coordinators also reviewed medical records for emergency room or hospital visits. Adverse events were graded using Common Terminology Criteria for Adverse Events v3.0 (CTCAE v3.0). An independent Data and Safety Monitoring Committee provided oversight for the study.
Statistical Analysis
A prior open-label study of SMA type 2 subjects showed the standard deviation for MHFMS-SMA was 3.19. [27] A clinically meaningful change was estimated to be a 3 point change in MHFMS-SMA at 6 months. A sample size of 25 subjects completing both the baseline and six-month evaluations provided at least 90% statistical power with an effect size of 0.941, Type I error of 0.05 (two-sided). Assuming 20% attrition then 60 total subjects were required. Subjects were randomized using a permuted block design with blocks of size 6 with no stratification. The statistician for the study (author CS) created a blinded randomization list using the permuted block algorithm. At the second screening visit the site determines whether the subject was eligible, then completed a randomization request form that was faxed to one of the authors (CS) who would assign the subject per the predetermined randomization list. The treatment assignment was then sent to the central pharmacy to dispense the study medication. The statistician was blinded to treatment assignment, as was the study coordinator, treating investigator, patient and parents.
The assumption of normality for continuous variables was examined using the Shapiro-Wilks tests. Fisher's exact test was used to compare treatment groups in the analysis of categorical variables (e.g., gender). Baseline demographics and function were analyzed as continuous variables with treatment groups compared using a t-test. Analysis of variance (ANOVA) test was used to compare treatment groups for change from baseline data when normally distributed (e.g., change in MHFMS at 6 months). If the data were not normally distributed then the Wilcoxon rank-sum test was used (e.g., change in BMD). Linear regression analysis was used to examine baseline characteristics, potential prognostic factors for normally distributed outcome variables. The Pearson Correlation Coefficient, calculated from the linear regression, was used to assess association between continuous variables. Longitudinal analysis was performed using generalized estimating equations. When the data were normally distributed the identity link function was used and unstructured correlation. We used the small sample adjustment to the Score test, J/(J-1) where J is the number of clusters [37] When both random and fixed effects were analyzed, a mixed effect longitudinal analysis was performed. A p-value less than 0.05 was considered statistically significant. Reliability of the MHFMS in this group of subjects was evaluated using the intraclass correlation coefficient (ICC). This was used to determine the reliability across evaluators at different sites.
All analyses were performed on an intent-to-treat population that was defined as all subjects randomized to receive study medication. There were 61 subjects randomized, 4 subjects did not have a 6-month data: 2 of these dropped out after the 3-month visit, the other 2 missed their 6-month visit due to illness. At 12 months the 2 subjects missing the 6 month visit returned. MHFMS was not performed in one additional patient at 6 months. At 12-months one had a missing MHFMS. No data imputation was performed. CMAP was not performed according to protocol at 2 of the clinical sites accounting for 17 missing values. Parent-proxy QOL was not collected at the 6 month visit in 3 subjects.
Results
Recruitment began in September 2005 and enrollment of the last subject was completed in October of 2006. The last subjects' final visit was in November 2007, and database was locked in March 2008. Participant flow is outlined in figure 1.
Baseline Characteristics of Study Population
The analysis of demographic data (Table 1) indicates that the 2 treatments were well balanced for most factors. More males were enrolled in the placebo group (64.5%) and more females were enrolled in the treatment group (56.7%) but overall, there was no statistical difference (p = , 0.126). Body Mass Index (BMI) assessments indicated no significant difference between treatment arms (Supplemental Table S1). Total body BMD, BMC, lean mass and fat mass prior to the start of treatment were assessed via DEXA scans performed at three sites (Supplemental Table S2). Baseline pulmonary function tests (PFTs) in subjects five years of age and older and baseline CMAP values were equivalent across groups (Supplemental Tables S3 and S4). Myometry data from children five years of age and older indicated that upper extremity strength was greater in the placebo group (Supplemental Table S5, p≤.03), but lower extremity myometry data demonstrated no difference between groups. The PedsQL parent assessment indicates no statistical difference between the placebo and treatment group on any of the subscales, psychosocial or total QOL (Supplemental Table S6). Only a subset of subjects was able to complete their own PedsQL (Supplemental Table S7).
Reliability of the primary efficacy outcome measure, MHFMS-SMA, during the screening period
We used the ICC to determine the reliability of MHFMS from the first screening visit to the second, rated within their institution. Evaluating all subjects, the reliability is 0.97. The reliability at each institution was excellent and similar to the reliability previously reported (0.95–0.99). [27], [29]
Primary efficacy outcome
Impact of treatment on gross motor function as measured via MHFMS.
The primary endpoint was change from baseline in the MHFMS score at 6 months. The distribution of the change from baseline data was normally distributed. There was no difference in change from baseline at 6 months between placebo and treatment groups (p = 0.492, Table 2). A linear regression analysis for the change from baseline to 6 months of the MHFMS, including age at study entry and treatment group as independent variables, indicates that age has a borderline effect on change (p = 0.0564), but treatment was not significant at p = 0.0912. The coefficient for age is -0.47 (SE 0.238) indicating that as age increases the change in MHFMS decreases. Thus, subjects that were older at the start of the study had less of an increase in scores than younger subjects. Gender was examined as a possible independent variable but was excluded from the model (p = 0.92).
During the last six months, starting from V2, all subjects received active treatment. There was no statistical difference in the change from baseline between the two groups; however, for subjects on treatment for the entire 12 months, there was a small increase in the MHFMS score compared to baseline (Table 3, p = 0.185). Based upon the indication that age had a negative relationship on MHFMS change over time, we investigated splitting the subjects into two groups: over age three years and under age three years. Generalized estimating equations were used to examine the change from baseline MHFMS score over a 12 month period in children ages two to three years of age (n = 23). There was a negative effect of baseline weight over the 12 month interval (estimate = −0.81, p = 0.0066) but a positive effect of treatment over that period (estimate = 1.21, p = 0.03). Analysis of children three years of age and older (n = 33) demonstrated a similar negative weight effect (estimate = −0.16, p = 0.0005) without the apparent treatment effect (p = 0.153). Table 4 displays the change from baseline in MHFMS within each age group by treatment arm. Subjects two to three years of age in the treatment group have an increasing slope through each visit, indicating a treatment benefit, while no such effect was observed in children greater than 3 years.
Ancillary analyses
Impact of weight, BMI and body composition on primary motor outcomes.
The BMI z-score and BMI percent were adjusted for subject age. BMI, BMI z-score and BMI percent were available at each follow-up visit. A generalized estimating equation analysis of change from baseline MHFMS indicated that increasing BMI over the year of study was negatively associated with outcome. BMI was treated as a time-varying covariate in the generalized estimating equation model. As BMI increased over time, change in MHFMS decreased (estimate = −0.05, p<0.0001). Linear regression analysis indicated that a change in fat mass as measured by DEXA is negatively related to a change in MHFMS (p = 0.0409). As fat mass increases by the 6-month time-point, MHFMS decreases. Figure 3 presents the change from baseline fat mass versus the baseline fat mass by treatment arm from the three sites that obtained DEXA measurements. BMI percent based upon z-score correlated well with baseline fat mass (correlation = 0.475, p = 0.0004). A regression analysis indicates that as weight at baseline increases, change from baseline MHFMS score worsens (p = 0.0058).
Impact of treatment on bone density.
The change in total body BMD and BMC at 6 months was not normally distributed. There were no differences between the two treatment groups in the change in BMD (p = 0.82) or BMC (p = 0.41).
Impact of age on primary motor outcome.
As subjects get older, their weight increases, and the relationship between weight and age causes a multi-colinearity problem in the analysis. However, linear regression analysis of change in MHFMS and age as the only independent predictor indicated that age is statistically significant at p = 0.0069. The coefficient for age is −0.62 (SE 0.219) indicating that as age increases the change in MHFMS score decreases.
Impact of treatment on electrophysiologic measures of innervation.
There was no significant change in CMAP negative peak amplitude or area measurements between treatment groups after 6 months (Table 5). However, there was a good correlation between maximum CMAP amplitude and our primary outcome measure, the MHFMS (figure 4).
Correlation = 0.5643. The red lines are 95% confidence bands. Note that the majority of the data are contained in the linear relationship between these two factors with only one outlier.
Impact of treatment on quality of life outcome assessments.
Quality of life was measured using the PedsQL. There were no statistically significant differences between treatment groups for the parent-proxy QOL at 6 months (Table 6). The parent-proxy total QOL was compared to significant change (3 point change) in MHFMS at 6 months. Total QOL did not improve as MHFMS improved, but there was evidence of deterioration in QOL as MHFMS declined (Table 7).
Impact of treatment on myometry measurements.
Myometry was assessed in fifteen subjects at least 5 years of age whose assessments were not limited by contractures. Tables 8 and 9 present the 6 month and change from baseline data for upper and lower extremity strength assessments. There were no statistically significant differences between the two treatment groups (Table 10).
At 12 months, there were no statistically significant differences between groups initially assigned to placebo or to treatment. In the group receiving treatment for the full year, there was no continual improvement over the second six-month treatment period (data not shown). A generalized estimating equation model indicates that there is no period by arm effect (p = 0.73) and no BMI effect (p = 0.7983).
Impact of treatment on pulmonary function.
PFTs were performed only in subjects five years of age and older (n = 24), thus limiting the power to observe any relationship with treatment. Treatment was not associated with changes in any of the PFT outcomes over the treatment period (data not shown).
Impact of treatment on quantitative measures of SMN mRNA in whole blood.
Results of mRNA analysis at 6 months are in Table 11. There was no difference in actual baseline values or change from baseline values between the 2 treatment groups at 6 months. In addition, no difference was apparent between treatment groups through 12 months for either flSMN or Δ7 SMN. A mixed effect model analysis over the entire year indicated that baseline flSMN was highly predictive of flSMN at one year (p<0.0001), indicating very little change in flSMN. The correlation between Δ7 SMN baseline and Δ7 SMN at one year was less than flSMN, but still highly significant (p<0.0001). The minimal change over time and lack of correlation with baseline function provides little indication of a relationship between these mRNA parameters and outcome measures.
VPA trough levels
Eight participants (13%) were less than fully compliant with recommended VPA dosing, and eight participants (13%) were less than fully compliant with carnitine during one or more study visits during phase 1. Mean overnight VPA trough level in Group 2 (subjects receiving treatment) during phase 1 was 58.5 +/− 63.3 SD (16.8 to 101.0, range), which falls within the targeted trough level range of 50–100 mg/dL. VPA levels at 6 months were not associated with change in MHFMS.
Adverse Events
Adverse events occurred in 58% of those receiving placebo and in 77% of treated subjects during the randomized treatment phase. Although there were no statistically significant differences between the treatment groups for adverse events, numbers of subjects reporting adverse events, serious adverse events, treatment-related adverse events, and severe adverse events were all greater in the treatment group (Table 12). A detailed listing of all adverse events during the treatment phase is presented in Supplemental Table S8. Gastrointestinal symptoms and respiratory symptoms were more frequent in the treatment group. Four subjects (13%) on active treatment had a severe adverse event, most commonly pneumonia (Table 13). Treatment-related adverse events were nearly equal between groups. There was no pattern in the placebo group, while gastrointestinal AE were the most common in the treatment group. Overall adverse events for the entire 12- month period in each treatment group are presented in Supplemental Table S9.
Discussion
We chose to perform a placebo-controlled study of VPA in non-ambulatory children with SMA based on encouraging preliminary observations of improved motor function during an open label study [27]. Our choice to target children ages 2–8 years reflected a compromise between concerns regarding the lower limit of age where reliable assessment of motor function was possible versus concerns that improvement in older children may be constrained by other factors that would confound assessment of an SMN-enhanced treatment effect. A combined treatment regimen of VPA and carnitine was selected to avoid a potential confounding effect of carnitine depletion, because we had previously demonstrated an apparent susceptibility for more rapid depletion of carnitine with VPA treatment in this population. [27] Overall, enrollment was targeted towards a more homogeneous younger cohort of subjects who theoretically might be more likely to respond to therapy.
We did not detect a statistically significant improvement after six months of treatment in our primary outcome measure - the change in overall gross motor function assessed with the MHFMS. In addition, we did not detect a significant change in secondary endpoints of strength and function, which included assessment of maximum upper and lower extremity strength via hand-held myometry and PFTs. Adverse event frequency, although not statistically different between groups, was somewhat higher in the treatment group, particularly with regard to gastrointestinal and respiratory symptoms. Although we did not observe any clinical or laboratory evidence of serious hematologic or hepatic toxicity in this study, excessive weight gain was clearly more prevalent in the active treatment group. DEXA measurements confirmed that the associated weight gain was due largely to an increase in total body fat mass in the absence of an increase in lean mass.
Given the increased fat mass in the treatment group compared to the placebo group during the phase 1 period, it is notable that we did not observe a decline in gross motor function in the treatment group as a whole. As in the open label study [27], weight gain was not uniform across the population. Non-ambulatory subjects greater than five years of age, and those having a higher BMI at baseline, were at greatest risk. Children with the greatest improvement in gross motor function as assessed by the MHFMS gained less fat mass than did those with stable or worsening MHFMS scores. The fact that the MHFMS did not decline in light of substantial weight gain in the active treatment group suggests that VPA may indeed have a modest yet measurable biologic effect that is outweighed by its confounding effect of weight gain. Increases in fat mass may not always be deleterious in this population, where both over- and under-nourished states are common. Some children may benefit from the improved nutritional intake stimulated by an effect of VPA on appetite, perhaps putting them into a more favorable anabolic state. However, in other cases, due to increased gastrointestinal symptoms related to VPA and/or L-carnitine, weight loss could be a negative confounding factor.
Although no treatment effect was detected in the primary outcome measure as assessed by the MHFMS, a post hoc analysis demonstrated a statistically significant benefit in the youngest cohort of children, those two to three years of age. In addition, individuals who received the active treatment over a full year showed modestly improved function, gaining just over 2 points on the MHFMS, although this did not reach statistical or obvious functional significance. Age proved to be a significant factor as to whether or not subjects demonstrated an increase in MHFMS scores in association with treatment. We have previously demonstrated that increased age is associated with increased severity of denervation, so it is not surprising that the youngest non-ambulatory SMA children appear more likely to demonstrate a benefit with intervention [12]. Duration of treatment also appeared to be a factor and will need to be carefully considered in the design of future clinical trials in non-ambulatory SMA subjects. The ideal trial duration clearly depends on the expected mechanism of the treatment to be studied and relies heavily on the primary outcome measure chosen. This study affirms the value of the MHFMS in the multicenter pediatric clinical trial setting as a simple, reliable and easy to use outcome measure for documenting gross motor function in children as young as two years of age. However, if reinnervation is considered the most likely mechanism for a given therapeutic intervention, a trial duration of at least one year or longer may be more ideal to prove a treatment benefit using the MHFMS.
Possible disease-related biomarkers assessed in this study included quantitative SMN mRNA and maximum ulnar CMAP. Levels of flSMN and Δ7 SMN mRNA did not change with VPA. In contrast, VPA-response has been observed in SMA patient cell lines (16,17), one-third of SMA subjects (18) and in spinal cord of type III-like SMA mice (24). This result suggests that SMN expression in blood may not be a good surrogate biomarker to track VPA response. Because VPA targets are not specific, VPA response in SMA subjects may be much more complex than simply altering SMN expression. Indeed, recent data in a mild SMA model suggests that VPA may function by increasing SMN mRNA and protein, decreasing apoptosis and enhancing neuroprotection (24). Alternatively, whole blood may not be the appropriate cell type to detect changes in SMN expression, or such changes may be too small to be detected by relative quantification that relies on endogenous controls to normalize data. Finally, the assay we used has limitations, because it is not based on absolute quantification of SMN mRNA.
The strength of correlation between maximum CMAP amplitude and our primary outcome measure, the MHFMS, suggests that in non-ambulatory patients with SMA the CMAP may be a robust surrogate outcome measure for assessing treatment interventions if the putative mechanism of effect is to enhance innervated muscle mass by reinnervation or other trophic effects on nerve or muscle. However, baseline severity of denervation overall in a given cohort will profoundly impact whether or not improvement is anticipated and should be a key consideration in any trial design incorporating this as an outcome measure, as ceiling and floor effects may limit the usefulness of this assay in those of greater or lesser strength or age.
In summary, this study demonstrated no benefit from six months of treatment with VPA and L-carnitine in a young non-ambulatory cohort of subjects with SMA. Any modest biologic impact of VPA on individual outcome measures over the duration of treatment in this study was clearly outweighed by weight gain. This trial confirmed observations made in the open label study that the effects of age and weight gain with VPA are critical confounding factors that will have to be considered in the design of future clinical trials. Whether or not a subset of younger non-ambulatory SMA subjects might benefit from VPA treatment early in the course of their disease over a longer treatment period remains an outstanding question. Treatment intervention with VPA in non-ambulatory SMA children outside of a clinical study should be discouraged given the lack of clear benefit from this trial, and the potentially serious adverse events associated with VPA treatment. However, these results should be interpreted in the context of the limited treatment duration examined, an issue which should be carefully considered in the design of future clinical trials based on proposed mechanism of therapeutic action. We anxiously await the completion of ongoing studies of VPA in an ambulatory adult population (the VALIANT trial, clinicaltrials.gov ID NCT00481013) to help us further expand our knowledge about possible biologic impact of VPA in subjects with SMA. The ability to obtain strength data from the entire cohort in the VALIANT trial should help to answer outstanding questions about potential biologic effects of VPA in subjects with SMA, and whether this agent merits further study as a therapeutic agent in this disorder.
Supporting Information
Table S1.
Baseline Body Mass by Treatment Arm.
https://doi.org/10.1371/journal.pone.0012140.s001
(0.05 MB DOC)
Table S2.
Body Composition at Baseline by Treatment Arm.
https://doi.org/10.1371/journal.pone.0012140.s002
(0.04 MB DOC)
Table S3.
Pulmonary Function Testing at Baseline by Treatment Arm.
https://doi.org/10.1371/journal.pone.0012140.s003
(0.06 MB DOC)
Table S4.
CMAP Values at Baseline by Treatment Arm.
https://doi.org/10.1371/journal.pone.0012140.s004
(0.04 MB DOC)
Table S5.
Myometry Scores at Baseline by Treatment Arm.
https://doi.org/10.1371/journal.pone.0012140.s005
(0.06 MB DOC)
Table S6.
PedsQL Parent Proxy Sub scores at Baseline by Treatment Group.
https://doi.org/10.1371/journal.pone.0012140.s006
(0.05 MB DOC)
Table S7.
PedsQL Child Assessment at Baseline by Treatment Arm.
https://doi.org/10.1371/journal.pone.0012140.s007
(0.05 MB DOC)
Table S8.
Adverse events occuring during Phase I arm.
https://doi.org/10.1371/journal.pone.0012140.s008
(0.10 MB DOC)
Table S9.
Overall adverse events during all 12 months.
https://doi.org/10.1371/journal.pone.0012140.s009
(0.08 MB DOC)
Checklist S1.
CONSORT Statement 2001 Checklist.
https://doi.org/10.1371/journal.pone.0012140.s011
(0.19 MB DOC)
Acknowledgments
Charles Scott, Ph.D with CBS Squared, Inc, Fort Washington, Pennsylvania conducted the statistical analysis. The authors would like to thank Elzbieta Slominski, Desiree Nickle and Dana Conrad for technical assistance with regard to SMN mRNA quantification. We also gratefully acknowledge all the site clinical coordinators, research nurses, and evaluators who were critical to the success of this study. Finally, we would like to thank all of the children and families that participated in this study.
Author Contributions
Conceived and designed the experiments: KJS CBS TOC SPR KJK GA MKS GD JAM MBB JTK. Performed the experiments: KJS TOC LRS GA BE MKS GD TWP GMC JTK. Analyzed the data: CBS LRS BL SLS GMC. Contributed reagents/materials/analysis tools: LRS KJK BL TWP SLS JAM. Wrote the paper: KJS CBS SPR KJK MBB JTK.
References
- 1. Brahe C, Bertini E (1996) Spinal muscular atrophies: recent insights and impact on molecular diagnosis. J Mol Med 74(10): 555–62.
- 2. Roberts DF, Chavez J, Court SD (1970) The genetic component in child mortality. Arch Dis Child 45(239): 33–8.
- 3. Pearn J (1978) Incidence, prevalence, and gene frequency studies of chronic childhood spinal muscular atrophy. J Med Genet 15(6): 409–13.
- 4. Czeizel A, Hamula J (1989) A hungarian study on Werdnig-Hoffmann disease. J Med Genet 26(12): 761–3.
- 5. Emery AE (1991) Population frequencies of inherited neuromuscular diseases--a world survey. Neuromuscul Disord 1(1): 19–29.
- 6. Merlini L, Stagni SB, Marri E, Granata C (1992) Epidemiology of neuromuscular disorders in the under-20 population in Bologna Province, Italy. Neuromuscul Disord 2(3): 197–200.
- 7. Lefebvre S, Burgien L, Reboullet S, Clermont O, Burlet P, et al. (1995) Identification and Characterization of a Spinal Muscular Atrophy-determining Gene. Cell 80: 155–165.
- 8. Wan L, Battle DJ, Yong J, Gubitz AK, Kolb SJ, et al. (1995) The Survival of Motor Neurons Protein Determines the Capacity for snRNP Assembly: Biochemical Deficiency in Spinal Muscular Atrophy. Mol Cell Biol 25(13): 5543–51.
- 9. Lorson CL, Hahnen E, Androphy EJ, Wirth B (1998) A single nucleotide in the SMN gene regulates splicing and is responsible for spinal muscular atrophy. Proc Natl Acad Sci USA 96: 6307–6311.
- 10. Feldkotter M, Schwarzer V, Wirth R, Wienker TF, Wirth B (2002) Quantitative Analyses of SMN1 and SMN2 Based on Real-Time LightCycler PCR: Fast and Highly Reliable Carrier Testing and Prediction of Severity of Spinal Muscular Atrophy. Am J Hum Genet 70(2): 358–368.
- 11. Mailman MD, Heinz JW, Papp AC, Snyder PJ, Sedra MS, et al. (2002) Molecular Analysis of Spinal Muscular Atrophy and Modification of the Phenotype by SMN2. Genet Med 4: 20–26.
- 12. Swoboda KJ, Prior TW, Scott CB, McNaught TP, Wride MC, et al. (2005) Natural history of denervation in SMA: relation to age, SMN2 copy number, and function. Ann Neurol 57: 704–712.
- 13. Wirth B, Brichta L, Schrank B, Lochmüller H, Blick S, et al. (2006) Mildly affected patients with spinal muscular atrophy are partially protected by an increased SMN2 copy number. Hum Genet 119: 422–428.
- 14. Prior TW, Swoboda KJ, Scott HD, Hejmanowski AQ (2004) Homozygous SMN1 deletions in unaffected family members and modification of the phenotype by SMN2. Am J Med Genet A 130: 307–310.
- 15. Lefebvre S, Burlet P, Liu Q, Bertrandy S, Clermont O, et al. (1997) Correlation between severity and SMN protein level in spinal muscular atrophy. Nat Genet 16: 265–269.
- 16. Brichta L, Hofmann Y, Hahnen E, Siebzehnrubl FA, Raschke H, et al. (2003) Valproic acid increases the SMN2 protein level: a well-known drug as a potential therapy for spinal muscular atrophy. Hum Mol Genet 12: 2481–2489.
- 17. Sumner CJ, Huynh TN, Markowitz JA, Perhac JS, Hill B, et al. (2003) Valproic acid increases SMN levels in spinal muscular atrophy patient cells. Ann Neurol 54: 647–654.
- 18. Brichta L, Holker I, Haug K, Klockgether T, Wirth B (2006) In Vivo Activation of SMN in Spinal Muscular Atrophy Carriers and Patients Treated with Valproate. Ann Neurol 59: 970–975.
- 19. Kernochan LE, Russo ML, Woodling NS, Huynh TN, Avila AM, et al. (2005) The role of histone acetylation in SMN gene expression. Hum Mol Genet 14: 1171–1182.
- 20. Leng Y, Chuang DM (2006) Endogenous α-Synuclein is Induced by Valproic Acid through Histone Deacetylase Inhibition and Participates in Neuroprotection against Glutamate-Induced Excitotoxicity. J Neurosci 26(28): 7502–7512.
- 21. van Bergeijk J, Haastert K, Grothe C, Claus P (2006) Valproic acid promotes neurite outgrowth in PC12 cells independent from regulation of the survival of motoneuron protein. Chem Biol Drug Des 67: 244–247.
- 22. Sugai F, Yamamoto Y, Miyaguchi K, Zhou Z, Sumi H, et al. (2004) Benefit of valproic acid in suppressing disease progression of ALS model mice. Eur J Neurosci 20: 3179–3183.
- 23. Tsai LK, Tsai MS, Lin TB, Hwu WL, Li H (2006) Establishing a standardized therapeutic testing protocol for spinal muscular atrophy. Neurobiol Dis 24: 286–295.
- 24. Tsai LK, Tsai MS, Ting CH, Li H (2008) Multiple therapeutic effects of valproic acid in spinal muscular atrophy model mice. J Mol Med 86: 1243–1254.
- 25. Weihl CC, Connolly AM, Pestronk A (2006) Valproate may improve strength and function in patients with type III/IV spinal muscular atrophy. Neurology 67: 500–501.
- 26. Tsai LK, Yang CC, Hwu WL, Li H (2007) Valproic acid treatment in six patients with spinal muscular atrophy. Eur J Neurol 14: e8–e9.
- 27. Swoboda KJ, Scott CB, Reyna SP, Prior T, Chan G et al (2009) Phase I/II Open Label Study of Valproic Acid in Spinal Muscular Atrophy. PLoS ONE 4(5): e5268.
- 28. Main M, Kairon K, Mercuri E, Muntoni F (2003) The Hammersmith functional motor scale for children with spinal muscular atrophy: a scale to test ability and monitor progress in children with limited ambulation. Eur J Paediatr Neurol 7(4): 155–9.
- 29. Krosschell KJ, Maczulski JA, Crawford TO, Scott C, Swoboda KJ (2006) A modified Hammersmith functional motor scale for use in multi-center research on spinal muscular atrophy. Neuromuscul Disord 16(7): 417–26.
- 30. Beenakker EA, van der Hoeven JH, Fock JM, Maurits NM (2001) Reference values of maximum isometric muscle force obtained in 270 children aged 4-16 years by hand-held dynamometry. Neuromuscul Disord 11(5): 441–446.
- 31. Merlini L, Mazzone ES, Solari A, Morandi L (2002) Reliability of hand-held dynamometry in spinal muscular atrophy. Muscle Nerve 26(1): 64–70.
- 32. Stuberg WA, Metcalf WK (1988) Reliability of quantitative muscle testing in healthy children and in children with Duchenne muscular dystrophy using a hand-held dynamometer. Phys Ther 68(6): 977–982.
- 33. Dawson C, Croce R, Quinn T, Vroman N (1992) Reliability of the Nicholas Manual Muscle Tester on upper body strength in children ages 8-10. Pediatric Exercise Science 4: 340–350.
- 34. Sloan C (2002) Review of the reliability and validity of myometry with children. Phys Occup Ther Pediatr 22(2): 79–93.
- 35. Varni JW, Limbers C, Burwinkle TM (2007) Literature review: health-related quality of life measurement in pediatric oncology: hearing the voices of the children. J Ped Psychol 32(9): 1151–1163.
- 36. Simard LR, Belanger M-C, Morrissette S, Wride M, Prior TW, Swoboda KJ (2007) Pre-clinical validation of a multiplex real-time RT-PCR assay to quantify SMN transcripts in whole blood from SMA patients. Neurology 2007; 68: 451–456.
- 37. Guo X, Pan W, Connett JE, Hannan PJ, French SA (2005) Small-sample performance of the robust score test and its modifications in generalized estimating equations. Stat Med 24: 3479–3495.