Intensive Functional Neurorehabilitation and Follow-up of 84 Paraplegic Dogs Affected by Intervertebral Disc Disease

The objectives of this study were to verify whether the functionality obtained with functional neurorehabilitation intensive protocols (FNRIP) improve ambulation, promoting a new therapeutic approach, and understand the expected time for functional recovery. Furthermore, to know whether “spinal reex” locomotion could be a functional locomotory pattern, which may improve the quality of life. A controlled prospective clinical study using a large cohort of 84 dogs comprising mostly chondrodystrophic-breeds. The dogs were diagnosed with T10-L3 Hansen Type I, using computed tomography or magnetic resonance imaging, and treated with hemilaminectomy. All had postsurgical neurological stage 0 or 1, according to the Open Field Score (OFS), and showed either an absent or decreased exor peripheral reex. All patients were subjected to FNRIP within a maximum of 3 months, data were recorded on days 1,3,7,15,30,45,60,75, 90 and patients were followed-up after 8-10 days, at 1 and 6 months, and in some cases, after 1 and 2 years.


Background
Thoracolumbar disc disease is frequently observed in chondrodystrophic dogs (Bergknut et al., 2011;Grossbard et al., 2014). The extrusion of disc material (Type I intervertebral disk disease, IVDD) causes contusion and compression of the spinal cord at different magnitudes in relation to the extrusion velocity and the mass of extruded disc material (Jeffery et al., 2013). Contusive events could lead to a chain of biochemical and vascular events that could produce myelomalacia under extreme conditions (Brisson,  Surgery removes the extruded disc material, but the neuronal circuits persist in an interrupted manner. Functional neurorehabilitation (FNR) is a eld of physical medicine and rehabilitation associated with restorative neurology. It is based on evidence of signal transmission through the lesion, both caudally and rostrally, as detected by electromyography (Tansey, 2010;Dimitrijevic, 2012;Kakulas & Kaelan, 2015).
Therefore, the goal is to activate the central axon pathways traversing the lesion by synaptic stimulation (Dimitrijevic, 2012).
Re ex conditioning protocols, such as locomotor training, promote motor function restoration (Wolpaw, 2006;Chen et al., 2010;Thompson & Walpow, 2014a, 2014b, 2015. Furthermore, neuronal plasticity, owing to neurogenesis, gene activation, dendritic modi cations, and changes in physiological properties, as well as glial and vascular plasticity (Wolpaw, 2010) probably to some degree also promote the restoration of motor function secondary to sodium-channel voltage activation (Wolpaw, 2007). This type of step-training increases the e cacy of synaptic connections from residual descending pathways to interneurons located near the lesion (Côté et al., 2017). Therefore, the use of multidisciplinary approaches may improve the functionality (Gomes-Osman et al., 2016; Gant et al., 2018).
To the best of our knowledge, no previous study has been conducted on FNRIP in dogs from grade 0 (open eld test score, OFS 0) and 1 (OFS 1), that is, paraplegic dogs DPP negative (-) and DPP positive (+), respectively.
The primary aim of this study was to examine whether the functionality obtained with the FNRIP could improve ambulatory status with minimum coordination de cits, and thereby provide a therapeutic alternative.
The second aims were to verify the expected time for functional recovery and understand whether "spinal re ex" locomotion (SRL) could be a functional locomotor pattern that permits automaticity and a high quality of life.

Participants
This was a controlled prospective clinical study using a large cohort of dogs (n = 84). Most dogs were chondrodystrophic breeds, and all had T10-L3 intervertebral degenerative disc disease (Hansen Type I IVDD) diagnosed with either computed tomography or magnetic resonance imaging.
Prior surgery dogs had OFS scores of 1 (with DPP) or 0 (without DPP) (Olby et al., 2001) (Fig. 1). All OFS 0 dogs went to surgery, three to ve days after DPP loss. They were all treated with hemilaminectomy and had postsurgical OFS scores of 1, with absent/decreased exor peripheral re exes, or 0. A total of 67 dogs were paralyzed for less than 7 days after surgery, and 17 dogs for more than 7 days after surgery. They all lacked other concomitant diseases.
Dogs OFS 0 were excluded if they presented all the criteria above but with surgical approach before 3 days or more than 5 days after injury. Also participants were excluded if they presented other SCIs, lesions outside of the T10-L3 neurolocation, higher OFS values (> 1) postoperatively, or lower OFS values admission. The rst day was allocated for adaptation and the patients´ rst encounter with locomotor training and rehabilitation modalities.

Locomotor training
All dogs were acclimated to the land treadmill and began with higher body weight support (60%-80% body weight) (Dietz & Muller, 2004), which was decreased with the load tolerance (Millis & Ciuperca, 2015). The treatment sessions were supervised by a rehabilitator and commenced 2 days after admission to the clinical study.
While being supported in the harness, the patients started with quadrupedal step-training as part of their daily protocol. However, when some resistance was offered, a change to hindlimb bipedal step-training was required (Maier et al., 2009;Shah et al., 2016). During bipedal training, the forelimbs rested on a platform raised above the treadmill belt (de Leon et al., 1998) while the perineal area was stimulated, by suspending and crimping the tail or with assisted bicycle limb movements (Alluin et al., 2015).
For each training session, variables such as the walking speed and duration were increased and recorded, starting from 0.8 km/h to a maximum of 1.

Electrical stimulation
Electrical stimulation protocols were used to manage pain (interferential electrical stimulation -IES), increase muscular contraction, and possibly, increase neural connections (functional electrical stimulation -FES) and descending pathway depolarization (transcutaneous electrical spinal cord stimulation -TESCS).

Interferential electrical stimulation
This technique is a form of stimulation that has two separate channels and uses alternating currents (Sherman & Olby, 2004) through four electrodes placed on the skin near the region of spinal hyperesthesia and crossed at a 90° angle with the following parameters: acute pain, 80-150 Hz and 2-50 ms; chronic pain, 1-10 Hz, 100-400 ms (Levine & Bockstahler, 2014;Hady & Schwarz, 2015), once a day.

Functional electrical stimulation
This neuromodulation modality that uses a short electrical pulse sequence results in spinal re exes. Its aim is to stimulate the lower motoneuron near the motor region or through peripheral afferent stimulation (Kralj et al., 1983;Holsheimer, 1998;Hamid & Hayek, 2008). This modality was performed in all the patients with super cial electrodes, using a segmental technique. The cathode was placed on the skin region corresponding to L7-S1 and the anode was placed near the ventromedial motor region of the hindlimb exor muscle group.

Transcutaneous electrical spinal cord stimulation
All the patients underwent TESCS three times/day (5 days/week), which was gradually discontinued when a exion-extension locomotor pattern appeared. The surface electrodes were placed on the paravertebral muscles (cathode at T11-T12 and anode at L7-S1, dorsal to the Iliac crest) (

Pharmacological management
During the fourth week of the described protocol, if the exion-extension locomotor pattern was present with a DPP negative result (tested on the medial and lateral digits of the hindlimb bilaterally, on the tip and base of the tail, and on the S1-S2 dermatomes), it was added to the training protocol, and with the owner's consent, pharmacological management. It was administered 4-aminopyridine, a K + channelblocking compound (Hayes, 2004 If any side effects (seizures, diarrhoea and vomiting) occurred in any patient, they were immediately treated and withdrawn from the clinical study. The guidelines of the FNRIP in different phases are described in Table 1. Its application was consistently performed within the patient's cardiorespiratory capacity and according to the improvements observed on the functional neurorehabilitation examination and OFS assessment over a 3-month period. The patients in the clinical study had neurogenic bladders. Thus, their bladders were expressed manually 3-4 times/day (Martins & Ferreira, 2018). The urine was monitored daily for odor and color changes. If there was a suspected urinary tract infection, urine culture (cystocentesis) and speci c antibiotic treatment were administered.
The dogs were maintained under a full-time hospitalization regime. They were able to rest on soft beds with multiple disposable absorbent pads and encouraged to maintain sternal recumbency. Dogs were fed three times per day with an intake increase of 30% and hydric support of 100-120 ml/kg was orally administered after resistance training alternated with strength training, according to the patient's needs. At the end of the day, class IV laser therapy was administered to reduce pain at trigger points (Bennaim et al., 2017).
All dogs were trained during the day, starting at 9:00 a.m. and nishing at 7:00 p.m. They were assisted only by veterinarians and veterinary nurses, who had taken the CCRP course.

Outcome measures
All patients were assessed by neurological examination every 5-7 days by the same certi ed CCRP examiner/instructor. The measured outcomes, including the OFS values, were evaluated at baseline (day of admission) and on days 3, 7, 15, 30, 45, 60, 75, and 90 after FNRIP implementation. The presence of DPP, the exor re ex, exion-extension locomotor pattern, and postural standing were investigated. This facilitated the establishment of an accurate and systematic evaluation of functionality among patients.
Patients were considered functional if they showed functional "spinal re ex" locomotion (FSRL), and OFS of 13/14), and were discharged.
We de ned functionality as the patient's ability to stand up, maintain postural standing, take at least three steps, and engage in voluntary or automatic micturition and defecation. Based on neural reorganization, FSRL can be de ned as an "involuntary" movement with the autonomous ability to stand, take steps, and engage in voluntary or automatic micturition and defecation, thereby giving the patient independence and autonomy. Autonomous ability in movement control suggests that parts of the brain and spinal cord can probably activate movements with some conscious control (Shik & Orlosky, 1976).
However, with nonfunctional "spinal re ex" locomotion (NSRL), although the patient demonstrates the presence of a exion-extension locomotor pattern, they do not have the ability to stand or promote the step-cycle.
After the end of the study, dogs that become DPP +, or DPP -but with FSRL or NSRL, were discharged and released to owner´s guardianship. In case of progressive myelomalacia (ascending/descending), or in DPP -dogs that don´t recover, and upon owner's request, euthanasia was considered.
The method of euthanasia was performed using induction with Propofol through intravenous (IV) administration (cephalic vein), followed by Pentobarbital IV administration after the dog fall asleep, within a quiet room.

Statistical analysis
It is possible to assume normality of the data with an appropriately large sample size (> 40) (Ghasemi & Zahediasl, 2012). The quantitative, qualitative, and categorical data were analyzed using the IBM SPSS Statistics software, Version 22 (International Business Machines Corporation), and the results were interpreted at a level of signi cance of p 0.05. The categorical data were presented as frequencies and proportions (95% con dence interval).
The clinical study was designed to investigate the outcomes and the changes in the OFS. The means and medians were calculated for the OFS at each time point. Thus, all 84 dogs were grouped and analysed according to the time of their discharge. Chi-squared tests and t-tests were used for inferential statistics regarding variables with a normal distribution.

Results
All 84 dogs met the inclusion criteria, and the results of their evaluations were recorded after FNRIP implementation. Most dogs were of chondrodystrophic breeds. The French Bulldog was the most ≤ common (26 dogs), followed by a high variability of breeds, with 17 Dachshunds; nine Yorkshire terriers; seven Pekingese; six Poodles; ve small-legged, long-bodied mixed breeds; four Beagles; and two each of Portuguese Podengos, Jack Russell Terriers; Pinschers, and Cocker Spaniels; and one Basset Hound and Chihuahua each. There were 31 females and 53 male participants with no signi cant differences in age (p = 0.756) or weight (p = 0.453). The mean age was 4.15 years and the mean weight was 9.131 kg ( Table 2). The most frequently affected neurolocation was T11-T12 in 27.4% (23/84) of the subjects (Fig. 2). At the time of presentation, all dogs were admitted with an OFS of 0 or 1, and 79.8% (67/84) had been paralyzed for less than 7 days and 20.2% (17/84) for more than 7 days. Abbreviations: CI, con dence interval; SE, standard error Fifty-one dogs, or 60.8% (51/84) of the participants, were admitted with an OFS of 1 and were discharged when they achieved an OFS of 13. A total of the 33 dogs, or 39.2% of the participants were admitted with an OFS of 0 (33/84), among which, 16 were discharged in an ambulatory state: Ten dogs had an OFS of 0 -FSRL, and six dogs had an OFS of 13 with recovered DPP after 4 weeks. In dogs that were DPP +, the OFS facilitated evaluation at each time point and throughout the progressive stages essential for achieving functionality. The details are presented in the ow diagram in Fig. 3, as well as the representative evaluation graphics in Fig. 4.
According to the OFS, no improvements were observed from days 1 to 3 in any of the 84 dogs. After one week of FNRIP implementation, all 51 dogs admitted with an OFS 1 had a classi cation 9. By day 15, all dogs had completed FNRIP phase A. Twenty dogs that were admitted with an OFS of 1, less than 7 days post-injury, were medically discharged with an OFS of 13. They showed only mild residual proprioceptive de cits, with clear improvements (Fig. 4A).
Among those 20 dogs, during an 8-10-day follow-up consultation, 100% showed no sensory decline. However, only 60% (12/20) presented at the next follow-ups (at 1 month and 6 months) with improved proprioceptive de cits, as manifested by a 1-point increase in their OFS (14). Three dogs presented at the rst two follow-ups and showed an improved OFS of up to 14. Five dogs maintained an OFS of 13, although four of those missed the last follow-up consultations.
On day 30, 21 dogs that were admitted with an OFS of 1 received a medical discharge with an OFS of 13 (Fig. 4B). Furthermore, 28.6% (6/21) of that number showed improvements in their OFS (from 13 to 14) at the rst follow-up. Three dogs missed the 1-month follow-up and eight dogs missed the 6-month followup The remaining 33.3% (7/21) showed a one-point increase in their OFS.
≤ Five dogs that remained at FNRIP phase C showed improvements in their OFS at the time of medical discharge, with an OFS of 13 on day 60 (Fig. 4C). Follow-up on these dogs indicated improvements at the rst follow-up in three dogs. Only one dog missed the last follow-up. On day 90, ve dogs were discharged with an OFS of 13, all of which were re-evaluated during the three follow-ups. Four of those ve dogs showed an increase in the OFS (14).
All dogs that were admitted with an OFS 0 (n = 33) showed no change in the OFS until day 15. Four dogs left the study within the 15-day period, (represented by the soft blue line in Fig. 4A), one euthanized and 3 presented descending myelomalacia. However, the 29 dogs that remained showed the presence of a exion-extension locomotor pattern until day 30.
At the same time point (day 30), two dogs recovered DPP, and were medically discharged on day 60 with on OFS of 13 (Fig. 4C). Three dogs recovered DPP on day 45, and one on day 60. Each of those dogs was discharged on day 90 with an OFS of 13, after 4-amynopiridine administration (Fig. 4D). Thus, a total of 21% (6/29) dogs recovered DPP.
On day 60, four dogs had medical discharge with an OFS of 0 -FSRL, three dogs with an OFS of 0 -NSRL and two dogs with an OFS of 13, as mentioned before (Fig. 3). On day 90, six dogs were discharged with an OFS of 0 -FSRL, 10 dogs with an OFS of 0 -NSRL, and four dogs with an OFS of 13 (Fig. 3).
All four dogs with FSRL that were discharged on day 60 were re-evaluated at the 8-10-day and 30-day follow-ups, and they maintained autonomous functionality. In addition, the six dogs that were discharged with FSRL on day 90 attended the three follow-ups and maintained an ambulatory state with minimal coordination de cits.
Outcome measures regarding the OFS mean and median scores show no obvious increase in continuous scores at each time point, given that some dogs remained in treatment and others were discharged on days 15, 30, 60, and 70 (Table 3). However, regarding the transition time points during treatment (days 7, 45 and 75), an increase was observed at the maximum range classi cation, with OFS scores of 9, 11, and 12, respectively. The OFS means at day 15 (5.79) and day 30 (5.95) were higher than those at day 60 (4.52) and day 90 (4.64) ( Table 3). Abbreviations: CI, con dence interval; SE, standard error Chi-squared tests revealed a statistically signi cant relationship between a DPP positive test result at admission and the ability to achieve functionality (p = 0.000). The same type of relationship (p = 0.000) was observed between DPP positivity and the time to discharge. Very few dogs admitted with an OFS 1 required up to 3 months before discharge (0.1%).
No signi cant relationships were observed among the weight and age categories, and the ability to reach an ambulatory state, p = 0.219 and p = 0.844 for the Chi-squared tests and t-tests, respectively. Furthermore, no signi cant relationships were observed in the time post-injury to reach functionality (p = 0.861).
Total functionality was considered in dogs that recovered to an OFS of 13 (n = 57) and an OFS of 0 -FSRL (n = 10), which was observed in 79.8% of the patients (67/84). In the sub-group of dogs admitted with an OFS of 0 that remained DPP -until the fourth week after admission, the functionality was 55% Twenty-nine dogs with an OFS of 0 showed no recovery of DPP until the fourth week. Two recovered on day 30 after FNRIP implementation (Fig. 4C) and the remaining four recovered following the administration of 4-aminopyridine (Fig. 4D). Therefore, 21% (6/29) recovered owing to the synaptic and anatomical neuroplasticity achieved by the possible combination of locomotor training, electrostimulation, and in some cases, 4-aminopyridine, which unlocked a possible lack of connection between the brain stem and spinal locomotor network silenced by inhibitory effects (Mehrholz et al., 2012(Mehrholz et al., , 2017).
In the present study, an overall functionality of 55% (16/29) was observed in dogs with an OFS of 0, 21% (6/29) of the dogs recovered DPP at a later time point (after day 30), and 35% (10/29) achieved FSRL, within a maximum of 3 months. Recovery of DPP after 30 days is re ective of the pre-selected population, and is suggestive of "complete" spinal cord injury, or unsuccessful neurological cases.
According to Jeffery et al. (2016), DPP recovery is subjective. Thus, it is essential to restrict the evaluation at each time point and, in order to decrease subjectivity, to ensure that it is assessed by the same assessor, at the same time of the day and in the same environment, as performed in the present study. Also, it is important to mention that the authors believe that in the cases of FSRL (10/29), and given the One of the possible explanations lies in the association between locomotor training and FES that may have facilitated a possible conversion of type II bers to type I (Côté et al., 2017). This is essential for postural support while standing (Postans et al., 2004). Depending on the cathode-anode orientation, it can also lead to the potential regeneration and activation of new connections (Thompson & Wolpaw, 2015).
In addition, the authors believe that TESCS could promote the stimulation of residual motor descending pathways, and therefore, the spinal locomotor network Hofstoetter et al., 2014Hofstoetter et al., , 2015.
The results obtained in the present study were neither age nor weight-dependent, which was inconsistent with most previous studies ( Considering the functional outcome, this study found close results compared to those presented by Langerhuus & Miles (2017) ´ systematic review and meta-analysis, but they were completed within a time frame of 2-3 months.
No cases of self-mutilation were observed, as there was an absence of paraesthesia through the FNRIP that allowed neuromodulation (Smania et al., 2010 There were several clinical study limitations. One was the experimental design without inter-observer and intra-observer examination. Nevertheless, images were reviewed by independent two observers (neurologist and CCRP). Additionally, no scale was applied to dogs without DPP. Although a speci c scale exists, it could not have been included in the present study because it was not used consistently for all cases . Furthermore, it is important to note the need for future studies that would more accurately specify various scores that are applicable to dogs without DPP (OFS of 0), which could be related to their possible functionality.

Conclusions
The present clinical study demonstrated that intensive, multidisciplinary, functional neurorehabilitation protocols, including locomotor training and electrostimulation programs, applied to a pre-selected group of patients with an OFS of 1, may result in functional clinical recovery (100%). However, these methods were unable to reduce the time of recovery. In addition, for patients with an OFS of 0, the association of these protocols with the administration of 4-aminopyridine led to a 55% recovery rate over a period of up to 3 months. These ndings may facilitate our determination of the answer to dog owners´ most frequent questions about the time that is necessary to achieve functionality and urinary ability in cases of a lack of neurological success.
The FNRIP showed the possibility of helping dogs with functional "spinal re ex locomotion". Subjects demonstrated quadruped locomotion, balance, and autonomous micturition in higher numbers and within a shorter time frame. This indicated possible motor recovery, based on the level of independence and animal welfare. We concluded, therefore, that FNRIP is an asset in the discipline of restorative neurology in veterinary medicine. All owners were informed and signed a consent form before patients' admission.

Consent for publication
Not applicable Availability of data and materials The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests     This is a list of supplementary les associated with this preprint. Click to download.