Effect of postoperative underwater treadmill therapy on functional recovery in dogs undergoing intervertebral disc surgery: a retrospective matched cohort study

Abstract

Background

Postoperative recovery of neurologic function after surgery for intervertebral disc disease (IVDD) is variable. While underwater treadmill therapy (UWTM) is widely used in practice, objective evidence of its contribution to improved neurologic function remains limited.

Hypothesis/Objectives

To compare the effects of postoperative UWTM vs standard home-based rehabilitation on recovery of neurologic function in dogs undergoing hemilaminectomy or ventral slot surgery.

Animals

Seventeen dogs received supervised UWTM twice weekly for 6 weeks (UWTM), and 17 matched controls received owner-guided exercises (CONTROL).

Methods

Retrospective matched cohort study based on a review of medical records of dogs undergoing surgical decompression for IVDD. Dogs receiving postoperative UWTM matched 1:1 with control dogs based on breed, type of surgery, and body weight (≤10% difference). Neurologic function was assessed preoperatively and weekly (6 weeks) using the Olby Open Field Score (OFS). Improvement was expressed as change from baseline (ΔOFS). The primary outcome was the time to regain independent ambulation (defined as OFS ≥12).

Results

Both groups improved over time. From week 2 onward, ΔOFS was consistently higher in UWTM group than in CONTROL group (week 2: +4 vs +2; week 3: +6 vs +3). Median time to independent walking was shorter in the UWTM group (range 7–35 days) compared with the CONTROL group (range 14–42 days; Mann–Whitney U, P = .0069). No UWTM-related adverse events, including infection or wound complications, were observed.

Conclusions and clinical importance

Incorporation of structured UWTM therapy should be considered as part of postoperative management to enhance recovery outcomes.

Introduction

Intervertebral disc extrusion (IVDE) is one of the most common neurological emergencies in dogs and a leading cause of acute spinal cord injury (SCI). Surgical decompression, frequently via hemilaminectomy or ventral slot, is considered the treatment of choice for dogs presenting with moderate to severe neurological deficits. Although surgery effectively removes extruded nucleus pulposus and relieves mechanical compression, postoperative recovery of neurologic function is highly variable and depends on multiple factors, including severity of preoperative dysfunction, extent of secondary injury, and intensity of rehabilitative interventions.^1,2 Restoration of independent ambulation remains the primary clinical goal, yet the optimal postoperative rehabilitation strategy remains incompletely defined.

Hydrotherapy, particularly underwater treadmill exercise (UWTM), has gained recognition as a rehabilitative modality in dogs with SCI. Its proposed benefits include enhanced neuromotor retraining through buoyancy-assisted stepping, reduced weight-bearing forces, increased sensory input, and improved muscle activation with minimal mechanical loading.^3,4 These effects have been described in experimental and clinical studies, suggesting a potential to accelerate recovery of coordinated locomotion. Despite widespread clinical use, controlled evidence evaluating UWTM specifically in postoperative IVDE remains limited. Most available studies are retrospective, involve heterogeneous populations, lack standardized neurological scoring, or combine UWTM with other rehabilitation modalities, making interpretation difficult. As a result, there is no consensus regarding the true magnitude of benefit, timing of introduction, or which dogs are most likely to respond.

A clear deficit in current knowledge is the absence of prospective, matched-control studies comparing UWTM with standard postoperative management alone, while using validated neurological outcomes such as the Olby Open Field Score (OFS), an ordinal scale from 0 to 14, where 0 is paraplegia with no nociception, 14 is a normal pelvic limb gait.⁵ Little is known about whether UWTM influences not only the overall degree of improvement but also the rate of recovery of neurologic function and the time required to regain functional ambulation, which are the key outcomes of direct clinical relevance to veterinarians and owners.

The primary hypothesis of this study was that postoperative UWTM accelerates the recovery of neurologic function in dogs undergoing decompressive surgery for thoracolumbar or cervical IVDE compared with standard postoperative care alone. Specifically, we hypothesized that dogs receiving UWTM would demonstrate faster improvement in open field scores over time and a shorter time to regain independent ambulation. Objective of this retrospective matched cohort study was to compare postoperative neurological outcomes between dogs receiving structured rehabilitation including UWTM and matched controls managed conservatively at home.

Materials and methods

This study was designed as a retrospective matched cohort study evaluating the effect of postoperative UWTM on the recovery of neurologic function in dogs undergoing surgical treatment for IVDE. Clinical records from dogs treated between 2018 and 2025 were reviewed. Dogs that received postoperative underwater treadmill therapy (UWTM group) were matched 1:1 to dogs that did not undergo any supervised physical rehabilitation (CONTROL group) (Appendix). Matching was performed based on breed, body weight (≤10%), and type of surgery. All analyses were performed with informed owner consent and in accordance with institutional ethical guidelines for retrospective veterinary clinical research.

Dogs were included if they underwent surgical decompression (hemilaminectomy or ventral slot) due to IVDE, were non-ambulatory at presentation, had preserved deep pain perception, and a Sharp-Wheeler Score (SWS)⁶ of 3 to 4 at baseline (T0—before surgery). Neurological function was evaluated at baseline (T0) and at weekly intervals for 6 weeks (T1–T6) using the OFS, as OFS provides a more sensitive measure of week-to-week changes.⁵

Continuous variables were assessed for normality using the Shapiro–Wilk test. Baseline comparability between groups was evaluated using independent-samples t-tests or Mann–Whitney U tests, and Fisher’s exact test for categorical variables. Longitudinal neurological outcomes (OFS T0–T6) were analyzed using an ordinal mixed-effects regression model with Dog ID as a random effect and Group, Time, and their interaction (Group × Time) as fixed factors. Model-based marginal means were used to estimate between-group differences at each time point. The time to regain independent ambulation was analyzed using Kaplan–Meier survival curves and the log-rank test. Statistical significance was set at P < .05.

Results

A total of 34 client-owned dogs undergoing surgical decompression for thoracolumbar or cervical IVDE were enrolled, with 17 dogs allocated to the underwater treadmill group (UWTM) and 17 to the control group (CONTROL). Baseline characteristics were compared using independent-samples t-tests (continuous variables) and Fisher’s exact test (categorical variables). No statistically significant differences were identified between groups in age, body weight, sex distribution, type of surgery, SWS at admission, or baseline OFS (OFS T0) (all P > .05) (Table 1). All matched pairs remained within ≤7.2% difference in body weight, confirming successful matching.

Table 1.

Baseline signalment and matched preoperative variables for underwater treadmill and control groups of enrolled dogs.

Variable	UWTM (n = 17)	CONTROL (n = 17)	P-value
Age (years)	7.0 ± 2.8 (3–10)	7.0 ± 2.7 (3–12)	.68
Body weight (kg)	12.87 ± 7.12	12.73 ± 7.28	.96
Sex (M/F)	8/9	8/9	1.00
Surgery type	Hemilaminectomy 12 (70%), ventral slot 5 (30%)	Hemilaminectomy 12 (70%), ventral slot 5 (30%)	1.00
Sharp-Wheeler score (T0)	3.47 ± 0.51	3.35 ± 0.49	.82
OFS T0 (Olby Open Field Score)	4.53 ± 1.99	4.59 ± 2.19	.94

Abbreviation: UWTM = underwater treadmill exercise.

Across the 34 enrolled dogs, a total of 12 breeds were represented. The most common were Dachshunds (n = 10; 29%), French Bulldogs (n = 5; 15%), and mixed-breed dogs (n = 6; 18%). Less frequently represented breeds included the Chihuahua, Yorkshire Terrier, Maltese, Cocker Spaniel, Staffordshire Bull Terrier, and Slovakian Hound. The breed distribution did not differ between groups, as each UWTM dog was explicitly matched to a control dog of the same breed (Table 1), ensuring an identical breed composition across treatment arms. The prevalence of chondrodystrophic breeds (Dachshunds, French Bulldogs) was 44% in both groups, reflecting their known predisposition to IVDE. No breed showed disproportionate representation within any neurological severity category at baseline.

Thoracolumbar disc extrusions predominated in the study population, accounting for 25 of 34 cases (74%), while cervical disc extrusions (ventral slot procedures) represented 9 cases (26%). The localization did not differ between the 2 studied groups, with each UWTM dog matched to a CONTROL dog of the same surgical category (12 hemilaminectomies and 5 ventral slot procedures per group). Within the thoracolumbar cohort, the most frequently affected regions were T12–T13 and T13–L1, consistent across both groups. The most frequently affected intervertebral disc in the cervical segment was the C3–C4 disc, observed in three of nine dogs (33.3%). No significant association was observed between lesion localization and baseline neurological severity, and the distribution of cervical vs thoracolumbar lesions remained balanced between treatment groups.

No adverse events associated with underwater treadmill therapy, such as surgical site infection, wound dehiscence, or neurological deterioration, were recorded in the UWTM group. No adverse events of any kind related to home-based exercise were observed.

Neurological function improved in both groups across the 6-week postoperative period; however, dogs receiving underwater treadmill sessions demonstrated earlier and better recovery. At week 1 (T1), before the initiation of underwater treadmill therapy, dogs in the UWTM group demonstrated greater early neurological improvement, as reflected by a greater median increase in OFS relative to baseline (ΔOFS). When evaluating ΔOFS, the UWTM group demonstrated consistently higher median values compared with the control group from week 2 onward. At week 2, median ΔOFS was +5 (interquartile range IQR 4–6) in UWTM vs +3 (IQR 2–4) in CONTROL; at week 3, +6 (5–8) vs +4 (3–5); and at week 4, +7 (6–8) vs +5 (4–6). Mann–Whitney U tests demonstrated significant between-group differences at multiple time points, and significance persisted at key improvement weeks after the Holm–Bonferroni correction (Table 2).

Table 2.

Between-group comparison of neurological improvement (ΔOFS) at each postoperative time point (P > .05—n.s., P < .05^*).

Time point	ΔOFS UWTM (median, IQR)	ΔOFS CONTROL (median, IQR)	Mann–Whitney U	P-value	Adjusted P (Holm-Bonferroni)
Week 1 (T1)	+3 (2–4)	+2 (1–3)	83	.031	.062 (n.s.)
Week 2 (T2)	+5 (4–6)	+3 (2–4)	67	.008	.024 (*)
Week 3 (T3)	+6 (5–8)	+4 (3–5)	54	.002	.006 (*)
Week 4 (T4)	+7 (6–8)	+5 (4–6)	59	.004	.012 (*)
Week 5 (T5)	+8 (7–9)	+6 (5–7)	61	.005	.015 (*)
Week 6 (T6)	+9 (8–10)	+7 (6–8)	58	.003	.009 (*)

Abbreviations: IQR = interquartile; OFS = Olby Open Field Score; UWTM = underwater treadmill exercise.

Time to independent ambulation (defined as OFS ≥12) differed notably between groups. The median time to walking was 14 days (IQR 7–21) in the UWTM group compared with 28 days (IQR 21–35) in the CONTROL group. This difference was statistically significant (Mann–Whitney U, P = .0069). No dogs were censored in the survival analysis, as all individuals regained ambulation within the study period. The Kaplan–Meier curve (Figure 1) illustrates that 50% of the UWTM dogs achieved functional ambulation by week 3, whereas the CONTROL group reached the same threshold closer to week 4.

Figure 1.

Median Olby Open Field Score (OFS) over the 6-week postoperative period in dogs receiving underwater treadmill therapy (UWTM group) and those receiving standard postoperative care (CONTROL group) except for UWTM. The UWTM group demonstrated faster recovery of neurologic function than the CONTROL group, with higher OFS scores from week 1 onward and a consistently steeper improvement slope. Shaded areas represent the interquartile range (IQR).

The mixed-effects ordinal regression model (cumulative logit link with random intercepts for Dog ID and fixed effects for Group, Time, and Group × Time interaction) revealed a significant interaction effect (P < .001). This indicates that the trajectory of recovery of neurologic function differed between groups, with the underwater treadmill producing a significantly steeper improvement slope over time. Post hoc estimated marginal means confirmed higher OFS values in the UWTM group from week 1 onward (all P < .01) (Figure 2). Ordinal regression analysis further quantified the influence of the underwater treadmill on the recovery of neurologic function. Dogs in the UWTM group had significantly higher odds of achieving better OFS scores over time compared with controls (β = 1.06 ± 0.22, P < .001), corresponding to an odds ratio (OR) of 2.88 (95% CI, 1.88-4.40). This indicates that, at any given postoperative week, dogs receiving underwater treadmill were almost 3 times more likely to demonstrate improvement within a higher OFS category relative to matched controls. Time was also a significant predictor of recovery (β = 0.685 ± 0.030, P < .001), with an OR of 1.98 (95% CI, 1.87-2.10) per week, confirming a steady and biologically expected improvement in locomotor function across both groups. These findings complement the results of the mixed-effects model, reinforcing that the underwater treadmill not only accelerates recovery but also increases the likelihood of improved recovery of neurologic function at every time point.

Figure 2.

Kaplan–Meier survival curve for return to independent ambulation (defined as OFS ≥12). Abbreviation: OFS = Olby Open Field Score.

Discussion

This study evaluated the effect of structured postoperative underwater treadmill on the recovery of neurologic function after surgical decompression for thoracolumbar or cervical IVDE. The principal finding was that dogs receiving underwater treadmill recovered significantly faster, as evidenced by higher OFS scores, a greater ΔOFS progression, and a markedly shorter time to independent ambulation. These results indicate that the underwater treadmill can meaningfully enhance functional recovery beyond standard postoperative home care alone, an observation with substantial clinical implications.

Improvement of neurologic function occurred in both groups, but the underwater treadmill group demonstrated earlier gains beginning in the first postoperative week, with statistically significant differences from baseline emerging from the end of week 2 onward. These findings are consistent with studies demonstrating the benefits of structured rehabilitation after IVDE. In a controlled study by Martins et al.,⁷ dogs subjected to intensive postoperative neurorehabilitation achieved ambulatory status faster than controls receiving conventional care. Our results support these conclusions and extend them by using a matched design controlling for breed, body weight, surgical approach, and neurological severity, which are variables known to influence the outcome.⁸ Conversely, Zidan et al. found no significant difference in functional recovery between basic and intensive physical rehabilitation after thoracolumbar intervertebral disc herniation,⁹ highlighting potential variability in protocol efficacy. The reported variability in protocol efficacy among studies involving UWTM is likely attributed not only to whether UWTM was applied, but also to substantial differences in how the therapy was performed. Published studies in canine neurological and orthopedic rehabilitation demonstrate considerable heterogeneity in UWTM protocols, including variation in session frequency (ranging from multiple sessions per week to less frequent applications), session duration, walking speed and intensity, water level (from distal limb immersion to mid-thoracic or cervical levels), and the degree of therapist assistance or manual support provided during exercise. Additionally, differences in timing of UWTM initiation relative to surgery, overall rehabilitation program integration, and dog neurological severity can further influence outcomes. These protocol-related factors likely affect biomechanical loading, proprioceptive stimulation, and neuromuscular activation during UWTM, thereby contributing to the variability in reported therapeutic efficacy across studies.

The median time to independent ambulation was 14 days in the underwater treadmill group vs 28 days in controls. Previous studies report median recovery times of 10–21 days for dogs with thoracolumbar IVDE presenting with ambulatory paraparesis and 14–42 days for those presenting with deep-pain-positive paraplegia treated surgically.^8,10 Thus, underwater treadmill shifted recovery toward the faster end of the expected range, whereas matched control dogs recovered toward the slower end. Our Kaplan–Meier analysis further demonstrated that underwater treadmill dogs reached 50% ambulation (OFS 9–10) approximately 1 week earlier, supporting the conclusion that early functional gains were not random but part of a consistent pattern of accelerated recovery.

The mechanisms by which underwater treadmill can enhance recovery of neurologic function include buoyancy-assisted gait enabling earlier stepping, hydrostatic pressure reducing edema and discomfort, controlled resistance promoting symmetrical limb use, and increased proprioceptive stimulation of spinal locomotor networks. These physiologic effects have been described in both veterinary and human SCI rehabilitation.^9,11–13

Evidence from experimental and clinical research in other species suggests that repetitive locomotor training can facilitate engagement of central pattern-generating circuits,¹⁴ but such mechanisms have not yet been conclusively demonstrated in canine IVDE patients. Therefore, while activation of central pattern generators remains a plausible theoretical explanation, our findings should be interpreted in light of the established mechanical and proprioceptive advantages of underwater treadmill rather than unverified neurophysiological claims.

A meaningful consideration in interpreting these results is the difference in early postoperative management between groups. Dogs allocated to the UWTM group remained hospitalized until suture removal and received structured, daily supervised rehabilitation, including passive range-of-motion exercises, assisted standing, weight-shifting exercises, and controlled positioning under trained staff supervision. In contrast, dogs in the control group were discharged earlier and transitioned sooner to owner-performed home-based physical rehabilitation. A limitation of our study was that the home-based rehabilitation program was not objectively monitored, the compliance was assessed only based on owner verbal reports during follow-up visits.

Early postoperative recovery of neurologic function is influenced not only by preoperative neurological severity but also by the consistency, timing, and quality of sensory and motor input during the initial recovery phase.^15,16 At week 1 (T1), before initiation of underwater treadmill therapy, dogs in the UWTM group already demonstrated greater neurological improvement, reflected by a higher ΔOFS relative to baseline. This early difference is likely attributable to the longer period of supervised in-hospital rehabilitation, which provides more consistent and controlled therapeutic input compared to home-based programs that are inherently variable in compliance, technique, and frequency. Accordingly, the difference observed between groups at T1 most likely reflects differences in early postoperative management rather than a true baseline imbalance.

However, the most pronounced and statistically robust divergence in functional recovery emerged from the second postoperative week onward, coinciding with the initiation and continuation of underwater treadmill therapy. Taken together, these findings suggest that a combination of extended hospitalization with structured rehabilitation and underwater treadmill therapy can contribute to improved recovery of neurologic function, with underwater treadmill representing an important component of this multimodal approach.

The balanced distribution of thoracolumbar and cervical extrusions ensured that lesion localization did not bias the results. Cervical disc extrusions can recover differently than thoracolumbar lesions,¹⁷ but the matched study design equalized this variable. The identical representation of chondrodystrophic breeds, including Dachshunds and French Bulldogs, which constituted nearly half of both groups also minimized confounding associated with breed-specific disease expression and recovery patterns.¹⁸

Several limitations must be acknowledged. First, the sample size was modest, limiting statistical power for subgroup analyses. Nevertheless, significant findings across multiple independent analyses (ΔOFS, mixed-effects modeling, Kaplan–Meier survival) strengthen confidence in the observed effect. Second, complete blinding was not feasible, although neurological scoring was performed by an experienced clinician blinded to group allocation. Third, early postoperative care differed between groups; while this reflects real-world clinical practice, it likely contributed to early variation in progress. Fourth, weekly OFS scoring cannot capture rapid changes occurring between visits. Finally, the study was conducted at a single center, and generalizability to other rehabilitation facilities requires further validation.

Conclusions

Despite these limitations, this study provides evidence that postoperative underwater treadmill accelerates recovery of neurologic function in dogs undergoing surgery for thoracolumbar or cervical IVDE. Underwater treadmill was associated with earlier and steeper improvements in locomotor function and a clinically significant reduction in time to independent ambulation. These findings support the integration of underwater treadmill into multimodal postoperative rehabilitation protocols and warrant larger, multicenter trials to define optimal timing and intensity of therapy. Underwater treadmill therapy was well tolerated in this cohort and was not associated with any detectable adverse events, supporting its safety as part of postoperative rehabilitation after spinal surgery.

Abbreviations

DPP

deep pain perception

IVDE

intervertebral disc extrusion

OFS

Olby Open Field Score

PROM

passive range of motion exercise

SCI

spinal cord injury

SWS

Sharp-Wheeler score

UWTM

underwater treadmill exercise

Appendix

Material and methods

Dogs in the UWTM group underwent structured rehabilitation from the first day postoperatively and they underwater treadmill protocol starting when surgical wounds had healed sufficiently (typically 7 days after surgery). Underwater treadmill sessions were performed two to three times weekly, with the majority of dogs receiving two sessions per week. Duration and intensity of sessions were adapted to each dog’s neurological progression. Sessions typically lasted 10–20 minutes and were adjusted according to neurological status, gait quality, and signs of fatigue, with therapist assistance provided as needed.

Dogs in both groups received standard postoperative medical management, including analgesia, restricted activity, and bladder management where indicated. In addition, owners were instructed on basic home-based supportive care, to place the dog in lateral recumbency with appropriate support of the trunk and limbs to maintain a neutral, symmetrical body alignment. This included the use of padding or rolled towels to prevent trunk rotation or limb malposition. Dogs were regularly (every 4–6 hours) rotated from side to side to minimize prolonged pressure on bony prominences and reduce the risk of pressure sores and secondary musculoskeletal complications.

Owners were also instructed in a standardized home-based physical rehabilitation program, which was considered part of physiotherapy. This program included passive range-of-motion exercises (PROM) of all four limbs and gentle soft-tissue massage of the pelvic limbs, avoiding the surgical site. These exercises were performed twice daily in short sessions lasting approximately 10–15 minutes.

As neurological function improved, owners were progressively instructed to perform active exercises aimed at stimulating voluntary motor function and postural control. These included assisted standing and weight shifting, controlled leash walking, sit-to-stand exercises, and slow figure-of-eight walking. Exercises were performed once to twice daily, with intensity and complexity adjusted according to each dog’s neurological status, tolerance, and progression. In addition, during assisted standing and walking, owners were instructed to support the dog to maintain symmetrical weight-bearing and proper limb alignment.

The home-based physical rehabilitation program was identical for both groups. Dogs in the UWTM group received UWTM therapy in addition to the standardized home-based physical rehabilitation program described above, which was initiated during postoperative hospitalization. Dogs in the CONTROL group received the same standardized rehabilitation program during hospitalization only, without subsequent underwater treadmill therapy. After discharge, dogs in the UWTM group continued supervised rehabilitation in the clinic, whereas dogs in the CONTROL group performed rehabilitation exclusively at home. Compliance with the home-based physical rehabilitation program was assessed based on owner verbal reports during in-clinic follow-up visits and was not objectively monitored.

Author contributions

Mária Kuricová (Conceptualization, Data curation, Methodology, Visualization, Writing—original draft, Writing—review & editing), Magdaléna Török (Conceptualization, Data curation, Investigation, Methodology, Validation, Writing—original draft, Writing—review & editing), Patrik Zelezník (Conceptualization, Data curation, Investigation, Writing—original draft, Writing—review & editing), and Tomáš Lipták (Conceptualization, Data curation, Methodology, Visualization, Writing—original draft, Writing—review & editing)

Conflicts of interest

The authors declare no conflicts of interest.

Funding

The authors received no specific funding for this work.

Off-label antimicrobial declaration

Authors declare no off-label use of antimicrobials.

Institutional animal care and use committee or other approval declaration

Authors declare no institutional animal care and use committee or other approval was needed.

Human ethics approval declaration

Authors declare human ethics approval was not needed.