Whether the newly modified rhizotomy protocol is applicable to guide single-level approach SDR to treat spastic quadriplegia and diplegia in pediatric patients with cerebral palsy?

To reduce muscle tone mainly in particular spastic muscles of lower limbs in spastic cerebral palsy (CP) children via single-level approach selective dorsal rhizotomy (SL-SDR) using a universally applicable rhizotomy protocol still remains challenging [1]. In the current study, we investigated whether our modified rhizotomy protocol-guided SL-SDR which could effectively decrease muscle tone of those particular spastic muscles in hemiplegic cases could be applied as well in spastic diplegic and quadriplegic pediatric CP cases.

In the current study, we retrospectively conducted a cohort review in cases younger than 14 years of age diagnosed with spastic diplegic or quadriplegic CP who undergone our modified protocol-guided SDR in the Department of Neurosurgery, Children’s Hospital of Shanghai since July 2016 to November 2017 with at least 12-month intensive rehabilitation program. Clinical data including demographics, intra-operative electromyography (EMG) recordings, and relevant evaluation results of the included cases were taken from the Database of Pediatric Cerebral Palsy Patients in the Rehabilitation Department of both Children’s Hospital of Shanghai and Shanghai Rehabilitation and Vocational Training Center for the Disabled. Diagnosis of spastic CP for these cases was made by a specialized team including specialists in the rehabilitation department of both institutes.

Inclusion criteria for the cases included in the current study were as follows: (1) with diagnosis of spastic diplegic or quadriplegic CP; (2) age at SDR guided by our modified protocol between 3 and 14 years; (3) after SDR, at least 12-month rehabilitation therapy conducted at the rehabilitation department of Children’s Hospital of Shanghai or Shanghai Rehabilitation and Vocational Training Center for the Disabled; and (4) pre-op and follow-up assessment was conducted by the same rehabilitation team. The exclusion criteria included were as follows: (1) combined with other neurological disorders, such as uncontrolled seizures; (2) spasticity due to head or spinal cord trauma; (3) with tendon contractures or joints deformity in the lower extremities; (4) previously operated on with any orthopedic procedures to release spasticity; (5) with moderate to severe mental retardation; and (6) lost to follow-up.

In a total of 135 cases who undergone our modified protocol-guided SL-SDR, 86 matched our study criteria (62 boys and 24 girls), among which 61.6% (53 cases) were quadriplegic. Pre-operatively, almost 2/3 of our cases were with their GMFCS levels II and III. Mean age at the time of surgery in these cases was 6.2 (3.5–12.0) years. A total of 582 muscles and 413 joints in the lower limbs were marked as target muscles and joints involved in pre-op assessment, respectively. The number of nerve rootlets tested during SDR procedure was between 52 and 84 over our 86 cases with a mean of 66.5 ± 6.7/case. Among those tested (5721 in 86 cases), 47.9% (2740) were identified as sensory rootlets which were related to lower limbs. Based on our protocol, 871 dorsal nerve rootlets determined to be associated with spasticity of those target muscles were sectioned 50%, and 78 were cut 75%. The number of rootlets with their EMG responses matched our rhizotomy criteria varied from 3 to 21 across our cases and was 6.4 ± 4.0, 9.8 ± 4.4, and 12.8 ± 3.2 in cases with their number of target muscle of 3–4, 5–6, and 7–8, respectively. After post-op rehabilitation therapy over a mean duration of 19.9 ± 6.0 months, GMFCS level improved from 3.0 ± 1.0 to 2.6 ± 1.2 (p < 0.05), and GMFM-66 score increased at a mean of 6.1 ± 3.2 in our cases (p < 0.01). No permanent surgery-related complications were recorded in the current study (Tables 1 and 2).

Muscle tone of those target muscles decreased significantly right after SDR and had a tendency to continue to reduce during the follow-up (2.6 ± 0.7, 1.7 ± 0.5, and 1.4 ± 0.5 in all 582 target muscles, pre-op, 3 weeks post-op, and at the time of the last follow-up, respectively). Major improvement of the strength of those target muscles was revealed after intensive rehabilitation therapy when compared with pre-op status with statistical significance achieved (3.9 ± 1.0 vs. 3.3 ± 0.8). ROM of those involved joints improved as well (Table 3).

In 81 cases with their pre-op GMFCS levels II–V, 27 (33.3%) showed improvement with regard to GMFCS level upgrade, among which 4 cases (4.9%) even upgraded over 2 levels. Better results with regard to upgrade in level of GMFCS were observed in cases with pre-op levels II and III when compared with those with level IV and V (24/57 vs. 3/24) with p < 0.01. Upgrading percentage in cases younger than 6 years of age at surgery was revealed greater than in older ones (23/56 vs. 4/25 with p < 0.05). Cases with their pre-op GMFM-66 score more than 50 had a greater score increase when compared with those less (7.1 ± 3.4 vs. 5.1 ± 2.8 with p < 0.01). In the meanwhile, better results were obtained in cases when SDR performed at their younger age, which reached statistical significance with p < 0.01 (6.9 ± 3.3 in case ≤ 6 years vs. 4.7 ± 2.7 in case > 6 years) (Fig. 4).

Preliminary results of our modified rhizotomy protocol-guided SL-SDR to decrease muscle tone of those what we called “target muscles” which were marked during pre-op assessment in spastic hemiplegic CP cases were very encouraging [2]. With the reduction of muscle tone of spastic muscles on their affected lower limbs, their motor function improved dramatically after post-op rehabilitation therapy. Spasticity of lower extremities in spastic diplegic and quadriplegic cases usually affects muscles of adductor, hamstring, gastrocnemius, and soleus with bilateral involvement in a varied degree [3‐6]. Besides the spasticity found in these major muscles, muscle tone prominent increase in rectus femoris is observed as well in some of spastic CP children. During the stage of protocol modification, we had a long discussion with rehabilitation specialists both in our hospital and in the rehabilitation center which were in collaboration with us. The major reason that eventually convinced us to exclude rectus femoris from our target muscle list was the concern of its tension for weight bearing [7, 8]. We were afraid of the deterioration in GMFCS level occurred in those cases after our procedure. In the current study, muscle tone ≥ grade 2 in rectus femoris was observed in 4 cases in their pre-op assessment. All these cases were with their pre-op GMFCS levels III to V. Interestingly, reduction of its muscle tone at an average of 0.5 grade was observed 3 weeks post-op in these cases, though SDR procedures in these cases strictly followed our rhizotomy protocol. We considered that such a muscle tone decreases in rectus femoris might be attributable to two factors: (1) cases with prominent spasticity on their rectus femoris were quite severe. Spasticity in those cases was usually global. Tone decrease in those target muscle groups after SL-SDR would somehow positively influence the muscle tone in rectus femoris; (2) hip adductor was identified as our “target muscle” in all of these cases. EMG responses that first appeared in hip adductor and having reached our rhizotomy criteria actually could also be observed in rectus femoris. Therefore, cutting of such rootlet might also result into the reduction of its muscle tone after the procedure (Fig. 5). The slight improvement of muscle tone in rectus femoris in these cases was quite beneficial for the following post-op standing program for these cases. Nevertheless, the justification of its exclusion from the target muscle list requires to be tested and validated in the following studies with more cases and longer follow-up.

Like how it was applied in hemiplegic cases, our modified rhizotomy protocol was quite applicable as well when it was used to guide SL-SDR to treat CP cases with spastic diplegia and quadriplegia with relatively simple intra-operative EMG interpretation. Among 5721 nerve rootlets tested in all cases in the current study, our protocol successfully identified 949 rootlets considered to be associated with spasticity of 582 target muscles marked pre-operatively, which ranged from 3 to 21 over 86 cases. The number of rootlets that matched our modified rhizotomy criteria increased in cases when their pre-op marked target muscle number increased (6.4 ± 4.0, 9.8 ± 4.4, and 12.8 ± 3.2 in target muscle number of 3–4, 5–6, and 7–8, respectively). Such results suggested that our modified rhizotomy protocol was feasible when it guided SL-SDR to release spasticity of the lower limbs in diplegic or quadriplegic cases.

Muscle tone reduction of those target muscles (1.0 ± 0.6 grade in 582 muscles) was satisfactory 3 weeks after SL-SDR, without impairment of their muscle strength observed generally in our cases. In our study, muscle tone decrease had a tendency to continue (0.3 ± 0.4 grade) over the following 19.9 months, which was consistent with previous studies when SDR was guided by the traditional EMG response grading system [6, 9‐12]. Significant improvement of spasticity in lower the limbs after our SL-SDR procedure created a better condition for the following intensive rehabilitation program. After a mean duration of 19.9 months of rehabilitation therapy, along with the major strength increase of those 582 target muscles (0.5 ± 0.5 grade), GMFCS level and GMFM-66 score of our cases improved dramatically (0.4 ± 0.6 and 6.1 ± 3.2, respectively). In CP population, GMFM-66 score hardly improves more than 3.0/year when a child is older than 6 years of age treated using the conservative measures [13]. Therefore, such a score increase obtained in our study when SL-SDR was guided using our modified rhizotomy protocol could be taken as a favorable outcome. Similar results were reported from other centers when SDR was performed in the traditional manner [14‐17]. Further investigation in our study revealed that cases when SDR performed to younger than 6 years old had much better outcomes than when performed to older ones, no matter to assess them using GMFCS level or GMFM-66 score. These results were comparable to those SDR managed via the traditional way. Not surprisingly, like what earlier studies concluded [1, 14, 15, 17‐19], our data showed as well that mild cases improved much better than those severe ones. Cases with their pre-op GMFCS level of II and III had more chance to have their level upgrade than those with IV and V (24/57 vs. 3/24) after the rehabilitation therapy. Interestingly, when we used GMFM-66 score to assess the outcomes of our cases, we found that, after post-op rehabilitation therapy, score increased more in cases having greater pre-op GMFM-66 scores (Fig. 4c). Statistical significance was obtained with p < 0.01 when our cases were categorized into two groups (pre-op GMFM-66 scores more than 50 vs. less). When we used score of 60 to divide them, significant difference between two groups still achieved with p < 0.05. We considered that two factors might contribute to such a phenomenon: (1) relatively restricted brain damage in mild cases, leaving more motor control involved neuronal network intact; (2) mild cases might be more cooperative during the post-op rehabilitation program, thus having better outcomes. Traditionally, the purpose of treatment for CP cases with their GMFCS levels IV and V was to ease the burden of caretakers of those children and to reduced pain caused by severe spasticity. In the current study, we surprisingly found that 3 cases with their pre-op level IV presented improvement of GMFCS level after a duration of almost 2 years of post-op rehabilitation program. Whether such results were associated with their age (SL-SDR performed around the age of 5 years) and relatively high pre-op GMFM-66 score (above 45 in all) requires to be further investigated in the future.

In the current study, we found that EMG responses of most of those sectioned dorsal rootlets reached the traditional rhizotomy criteria (EMG response grading 3+ or 4+), which means that when SL-SDR is performed in the traditional manner, those rootlets would be cut as well. However, there were up to 35% of rootlets in which EMG responses matched our rhizotomy criteria that did not reach 3+ or 4+ of the traditional grading system. About 7% of rootlets tested during surgery reached the rhizotomy criteria of the traditional one that failed to meet ours. Such phenomenon suggested that dorsal rootlet selection during SDR surgery was quite different when EMG responses were interpreted in a different way. Intro-operative EMG responses are considered to reflect neuroelectrophysiological characteristics of neuronal circuits in the spinal cord of spastic CP children. Though our results suggested that SL-SDR could mainly reduce muscle tone of those particular spastic muscles in the lower limbs of CP children when EMG interpreted in our way, the exact relations of EMG responses evoked via dorsal rootlet stimulation with clinical status of spasticity in CP cases still remained unclear. Therefore, further investigation in the following studies to address how neuroelectrophysiological circuits in the spinal cord exactly work in CP children are required.

In the current study, we found that 2 cases that experienced gastrocnemius weakness on one side 3 weeks after the surgery. Muscle strength decreased over one grade when its muscle tone decreased from grade 4 to 2 in both cases. At the time of the last follow-up, improvement was observed with muscle strength recovering to its pre-op status after rehabilitation therapy in these two cases. Muscle weakness after SDR was frequently reported in the literature [20, 21]. The exact mechanism beneath such phenomenon still remains unclear. Whether it is due to the input significant decrease from muscle spindles resulting into initial motor control difficulty requires to be investigated in the future.

Several shortcomings existed in the current study. In addition to its nature of retrospective cohort review performed in one center, the heterogeneous nature of CP may affect the reliability of our conclusions, though patient selection criteria were set. Second, the follow-up of this study is still short. Whether spasticity would recur in more cases in the future remains unclear. However, since this is the first study trying to propose a standardized rhizotomy protocol which could be universally applied in SL-SDR to treat all types of spastic CP cases via decreasing muscle tone mainly in particular muscles, we hope our study could provide an approach when treating spastic CP patients. In the meanwhile, EMG recordings interpreted in SL-SDR have the potential to help clinicians to understand further about the neuroelectrophysiological characteristics of the spinal cord in CP cases.

Our modified rhizotomy protocol was applicable to guide SL-SDR to treat pediatric cases with spastic diplegia and quadriplegia. Its efficacy to decrease muscle tone of particular muscles in our cases was favorable. With the significant reduction of muscle tone right after SL-SDR, motor function improved dramatically after a mean duration of 19.9 months of rehabilitation program in our cases.