Acute flaccid myelitis: Rehabilitation challenges and outcomes in a pediatric cohort

3.1 Clinical description of cohort

Between March 15, 2005 and January 30th, 2017, 31 children with AFM (17 males and 14 females) were seen at our facility for intensive SCI rehabilitation. Ages ranged from as young as 7 months to 16 years. Most ( n = 18) were admitted from home and almost half ( n = 14) were diagnosed retrospectively using the CDC 2015 revised criteria (see Table 1). Ten of these children were from the 2014 outbreak, 7 from the 2016 outbreak, and 14 had onset of illness prior to 2014; in this cohort, the earliest onset of paralysis was recorded as occurring in October 2006, with the child starting ABRT 5.5 years post onset (longest time between paralysis onset and ABRT start). In the whole cohort, time to start of ABRT ranged from 0.5 to 77 months.

In terms of presenting complaints for the prodromal illness, most had fever ( n = 29) and upper respiratory infection ( n = 17). A small number ( n = 3) presented with gastrointestinal illness. Neurologically, 26 presented initially with monoplegia, with 15 progressing to more fulminant neurological involvement; five eventually developed complete respiratory failure as a result of their paralysis. Of those presenting with monoplegia, 11 involved a lower limb and 15 involved an upper limb. In 24 of the cases, a lumbar puncture was performed as part of the diagnostic workup. Pleocytosis was noted in 19 cases, with lymphocytic predominance in seven and an elevated protein in eight (see Table 2).

Central nervous system imaging of both brain and spinal cord was performed in all patients. On spinal cord MRI, all patients showed anterior horn involvement, five of them demonstrated additional polyradicular extension, and six demonstrated nonspecific white matter tract involvement.

All patients had extensive evaluation to find the etiology of the AFM. This included testing for infectious causes such as West Nile virus, Western Equine virus, St. Louis encephalitis virus, Herpes Simplex virus, Coxsackie virus, Lyme, syphilis, and Bartonella. Enterovirus D-68 was confirmed in four cases (nasopharyngeal testing), while two patients tested positive for rhinovirus, one for Human Herpesvirus 6, and one for Mycoplasma. Testing for various autoimmune etiologies including lupus, neuromyelitis optica, paraneoplastic syndromes, hereditary neuropathies, endocrine dysfunctions (i.e., thyroid) and demyelinating diseases (i.e., oligoclonal bands, myelin basic protein) were negative.

EMG/nerve conduction studies were documented in seven of the 31 individuals in the cohort; they showed a predominance of anterior horn involvement, with abnormal or unobtainable late response F waves and H reflex. All showed normal sensory conduction.

Acute treatment was varied and reflected the diagnostic challenge which was present with the early outbreaks in 2014. All patients in our cohort were given high dose corticosteroids ( n = 31) with unclear effect on symptoms. Nineteen had intravenous immunoglobulin (IVIG) and plasmapheresis was used in ten patients. Due to severity of illness, two patients required extracorporeal membrane oxygenation (ECMO). Fluoxetine was trialed in four patients due to reports of utility in motor recovery in stroke patients, as well as its in vitro inhibition of enterovirus replication [22].

Sensation was noted to be intact in almost all patients (28/31) both at admission and discharge. The three patients deemed to have impaired sensation were under four years of age, so accurate assessment of sensory function was unreliable and difficult. However, they were felt to have decreased reaction to pin prick assessment versus expected.

Pain was reported in 19/31 patients at admission and 12/31 at discharge. In general, the nature of pain at admission was reported as neuropathic – thus, 11/19 patients were treated with gabapentin initially – whereas the pain at discharge was reported as more neuromuscular in nature. In the 14 patients that were evaluated as outpatients only, there were no reports of pain at the time of their clinic evaluation.

Six out of the 31 children were less than four years of age, making accurate assessment of bowel and bladder continence difficult. Of the remaining 25, on admission 10 were incontinent of bowel and 8 were incontinent of bladder. By discharge, only one child continued to have bowel incontinence and none had bladder incontinence.

About a third of the children were found to have joint range of motion limitations due to muscle contracture in the affected limb(s) (10/31) or scoliosis (11/31). Both limb contractures and scoliosis were documented in four patients.

No significant cognitive impairments were identified in any of the children included in this cohort.

There were 17 unique inpatients in the relevant time period, with two admitted twice each, for a total of 19 admissions. On admission, seven of those were on mechanical ventilators, nine required no ventilatory support, and the remaining three were mixed (on CPAP or BiPAP, or on vent at night only). Of the ten patients who admitted with any degree of support, eight were able to decrease that support (either to lower ventilator settings, PAP only during the day, or weaned off ventilator completely), while two discharged at the same degree of respiratory support as on admission (see Table 3).

There were some trends in predicting ability to wean ventilator support. For example, on average, those patients who weaned to any degree were older on admission (average of 93.6 months) than those who did not (average of 67 months). The average length of time between onset of AFM and admission in those that weaned was 9.2 months; this included one significant outlier who was admitted twice, at 25 and 35 months after illness onset. When this patient was excluded, the average dropped to 2.2 months after onset. In contrast, those who were unable to be weaned at all presented an average of 12.75 months after illness onset. This could represent the natural course of the illness, with relatively rapid improvement soon after onset, as well as the increased benefit of early therapy. The outlier mentioned above did make progress but in the setting of a diaphragmatic pacer placement.

Two of the children in our series, both of them belonging to the pre-2015 group, underwent diaphragmatic pacing. One child had no response to stimulation at the time of implantation. His ventilator settings – which we had not been able to wean prior to pacer placement – have since been decreased, but he has not yet been able to pace independent of the ventilator. The second patient, a teenager, had response to stimulation in one hemidiaphragm at implantation. He has been able to pace off the ventilator for short periods of time.

The Spinal Cord Independence Measure (SCIM) was utilized for 22 of the 31 patients. Out of the 22, 14 were seen for intensive inpatient admissions; the remaining eight were seen as outpatients. Of these, five were seen for a short term bout of therapy, which consists of intensive daily therapy for one to two weeks, limiting their functional changes in such a short period of time. Paired sample t -tests for the 11 inpatients with both admission and discharge scores suggested statistically significant change from admission to discharge. Only 5 of the outpatients had both admission and discharge scores which precluded statistical analysis; however, all 5 patients either maintained or increased raw scores on the SCIM.

All seventeen inpatients were evaluated using the Physical Ability and Mobility Scale (PAMS) during their inpatient admission. This is a 20 question scale created by the Kennedy Krieger Institute to measure 20 skills including: tolerance to positioning, tolerance to sitting in chair, tolerance to orthoses or splint, support for seating system, head control, trunk control, rolling supine to/from prone, transitioning from supine to sit, transitioning from sit to stand, standing, transitioning from floor to stand, environmental transfers, transfers into and out of a car, walking on level ground – assistive device, walking on level ground – distance, walking on level ground – level of assistance, community skills, wheelchair mobility, standing balance, and stairs. Paired sample t -tests showed statistically significant gains from admission to discharge on this measure (Table 4).

The Functional Independence Measure for Children (WeeFIM ® ) was also utilized for 13/17 of the inpatients. Given the wide range of ages, developmental quotient (DQ) scores were used, which represent percent typical for age. As a group, paired t -tests showed statistically significant improvements in self-care, mobility, cognition, and total WeeFIM ® DQ scores, as noted in Table 4.

Clinical muscle testing by therapists demonstrated that improvement in motor function occurred in a distal to proximal pattern, with return of lower limb function seen before upper limb and core/cervical musculature. Despite lower limb motor return, most children continued to require mobility devices for home and community ambulation, although 11 patients did have only upper limb deficits at discharge. These patients were able to ambulate independently, but continued to demonstrate weakness/paralysis in one or both upper extremities and were at high risk for developing shoulder subluxation. The unique presentation of paralysis for those patients who did require assistance with mobility required highly complex or alternative drive controls for wheelchairs. Shoulder paralysis limited manual mobility use; therefore, it was recommended that patients utilize either manual mobility with a power assist mechanism or a power wheelchair. Some examples include: a compact proportional joystick with foot control, Switch-it Microguide light touch miniature joystick with deflection, e-motion power assist wheels, and e-fix power base with proportional joystick. Of the seven patients who were ventilator-dependent, four were able to ambulate short distances with hospital staff, but they continued to demonstrate significantly diminished head control and continued to require ventilator support; therefore, those individuals required wheelchair mobility for community distances and to maximize independence.

The return of motor function from distally to proximally in the majority of patients put them at increased risk for orthopedic deformities including shoulder subluxation, hip subluxation or dislocation, and scoliosis. The majority of patients did not respond well enough to FES using the alternative parameters described above to adequately immobilize the affected joint and prevent chronic damage. Therefore, therapeutic taping and bracing were also used to manage the upper limb subluxations. Many of the commercially available subluxation slings deter use of the arm, which was not appropriate for many of these patients as they had functional use of distal musculature including finger movement. Therefore, it was important to stabilize the shoulder and still allow patients to utilize their existing function in the arm for activities of daily living. Patients were followed closely for hip subluxations. Positioning was adjusted in bed and in the wheelchair per the orthopedic recommendations to decrease the progression of the hip subluxation. Parents were trained in appropriate weight bearing and positioning for home to limit this progression. Bracing was utilized to decrease the progression of scoliosis, although tolerance to the braces was limited. Alternative bracing to the hard Thoracic Lumbar Sacral Orthoses (TLSOs) that are typically used was encouraged in order to increase tolerance to wear time. Patients were able to use the strength in their feet and hands to contour their bodies into less then optimal positions, increasing the progression of these deformities at a higher rate than seen with typical paralysis.

Although not common in our cohort, prior reviews have shown that most patients with AFM display cranial nerve involvement, most commonly in cranial nerves VI, VII, IX, and X [11], which can cause oral sensory and motor problems as well as pharyngeal dysphagia. Individuals with AFM may need alternative forms of communication if tracheostomy placement or mechanical ventilation limit their verbal communication. Assistive technology may be needed to help someone with upper extremity impairments access their environment.

Of the 17 patients admitted to the inpatient unit with AFM, nine required speech-language pathology (SLP) services. The remaining eight were screened by an SLP and did not require services. Of the patients receiving SLP intervention, eight required mechanical ventilation and had g-tubes placed before or during their acute rehabilitation admission. The remaining patient had experienced respiratory failure requiring intubation for 21 days but was extubated prior to transfer to our rehabilitation facility; this patient had an NG tube.

All nine patients requiring SLP services had dysphagia at the time of their admission and received total or supplemental tube feedings. Two patients initially silently aspirated during a videofluoroscopic swallow study (VFSS), but were later cleared for regular diets by a repeat VFSS prior to their discharge. Five patients were not cleared to have any oral intake (NPO) at discharge from the inpatient rehabilitation stay having not undergone VFSS; of these five, four were later cleared for regular diets by a VFSS and one was deceased at follow up (not as a direct result of his AFM). Reasons for not completing a VFSS prior to discharge included refusal behaviors and lack of readiness for oral intake based on clinical and medical evaluation. Two patients who were evaluated in our institute about two years (23 and 26 months) after their illness had already been cleared for a regular diet based on VFSS, but information on when they began eating by mouth was not available. Both had g-tubes for supplemental nutrition and hydration, as they did not meet all of their needs by mouth. Food refusal was a common observation among patients with AFM and ventilator dependence, which affected their participation in swallowing therapy.

Information on dental occlusion was available for six of the patients with long-term ventilator dependence, of which five developed a noticeable underbite. Based on exam, these were considered to be likely the result of weak resting tone for CN V, which can give a slack jaw, and open mouth appearance. In the long term, this can result in a permanent malocclusion (anterior open bite) which requires orthodontic and/or surgical correction once the facial skeleton stops growing in the late teens. Further research is warranted to confirm the cause and functional impact of these underbites and to develop prevention or remediation methods.

Of the eight patients with tracheostomies, information about speaking valve use was available for six patients. One did not tolerate speaking valve placement due to breath stacking. The remaining five patients tolerated the valve from a medical perspective but verbalized discomfort or resisted via nonverbal means (e.g., crying, kicking).