Intervertebral disc herniation (IVDH) is common, with a 2% of prevalence during the lifetime of dogs (5,20). Disc impingement and stretching of the fibers of the dorsal annulus or ligament cause pain or neurological signs, such as limb ataxia or hypomotility, resulting in muscle wasting (10). Dogs with pain in the lumbosacral junction exhibited significant reductions in the multifidus, psoas, iliopsoas, and sacrocaudal muscle cross-sectional area (CSA) compared to pain-free dogs (6). Similarly, lumbar paraspinal muscles decreased in dogs with degenerative lumbosacral stenosis compared to normal dogs (15). A magnetic resonance imaging (MRI) assessment of the paraspinal muscle in Dachshunds with chronic IVDH confirmed a significant decrease in the ratio of epaxial muscle CSA to disc CSA (4).

Fat infiltration occurs when the muscle fibers are replaced with non-contractile fat tissue, decreasing the functional capacity of that muscle (8). Fat infiltration is a sign of muscle degeneration, which means decreased functional cross-sectional area (FCSA) composed of lean muscle (8,12). Several studies have examined the changes in fat infiltration of the paraspinal muscle in dogs (4,21,25).

Little is known about the paraspinal muscle changes in various breeds of dogs with chronic IVDH. The first aim of this study was to determine if chronic IVDH affects the degeneration of the paraspinal muscles compared to acute IVDH in various breeds of dogs. It was hypothesized that paraspinal muscle CSA would be reduced and intramuscular fat infiltration would be increased, even more at the level of disc herniation in chronic IVDH. Second, the associations between the lateralization of disc herniation and asymmetric paraspinal muscle alteration were examined. It was anticipated that the paraspinal muscles in dogs with one-sided disc herniation would show asymmetrical changes.

Subjects

This study reviewed the medical records at the college of veterinary medicine at Chungnam national university, Ian animal medical center, and Sky animal medical center between February 2018 and February 2021. The inclusion criteria were dogs with the clinical signs of thoracolumbar pain or hindlimb ataxia, supporting MRI evidence of IVDH at the thoracic 12th-13th (T12-T13) vertebral level. The exclusion criteria were (1) a history of vertebral surgery, spinal fracture, or injuries; (2) primary or metastatic spinal tumor; (3) discospondylitis or infectious disease; (4) deformities of the spine, such as hemivertebrae; (5) systemic diseases that can affect the musculoskeletal system, including hyperadrenocorticism and renal disorder; (6) brain or cervical spinal abnormalities; (7) medication which can affect musculoskeletal organs, which includes prednisolone; (8) orthopedic lameness or surgery due to conditions, such as medial patella luxation or cranial cruciate ligament rupture. Patient information retrieved from the clinical records contained breeds, gender, age, body weight, and duration of neurologic signs or pain. Based on the duration of the clinical signs, the dogs were categorized into two groups. Acute IVDH was defined as the onset of clinical signs no more than seven days, and chronic IVDH was defined as lasting more than seven days. In addition, the dogs were categorized into three groups depending on the lateralization of the disc herniation: midline, right, and left. The side of the lesion was confirmed from the transverse MR images.

Measurement and analysis

All MRI examinations were performed using 1.5 Tesla magnets (Vantage Elan^TM; Canon Medical Systems, Japan, GE SIGNA HDXT; GE Healthcare, United States, SIEMENS MAGNETION ESSENZA; Siemens Healthineers, Germany) with the dogs under general anesthesia. Sagittal T2-weighted images of the thoracolumbar spine were obtained to allow planning of the transverse plane of the T12-T13 intervertebral disc space. One transverse image was acquired precisely at the level of the center of T12-T13 intervertebral disc space; this image was used for subsequent measuring CSA or FCSA of paraspinal muscle, spinal canal, and disc. The multifidus, longissimus dorsi, and iliocostal muscles were included in the paraspinal muscle (Fig. 1).

Figure 1. Transverse T2-weighted MR image showing the CSA measurements of the MRI thoracolumbar muscle CSA, spinal canal CSA, and disc CSA measurements. The outlines of the paraspinal muscle, spinal canal, and disc were traced individually. The multifidus muscles were drawn alone, whereas the longissimus muscles and the iliocostal muscles were measured together forming the epaxial muscle. M, left and right multifidus; EPAX, left and right longissimus muscle and iliocostal muscle; S, spinal canal; D, disc.

A commercial picture archival and communication system (ZETTA PACS; Zetta-TY soft, Korea) and ImageJ software (Wayne Rasband, National Institutes of Health, Bethesda, MD) were used for image viewing and measurements. The CSA of the bilateral paraspinal muscles [Multifidus, epaxial muscle (Iliocostalis and longissimus muscle)], spinal canal, and disc at the level of the T12-T13 were measured by constructing polygon points around the outer margins of individual regions (Fig. 1). The multifidus was outlined alone. The longissimus and iliocostalis muscles were combined, forming the epaxial muscle because it was difficult to distinguish the fascial boundary between these muscles in some cases. The ratios of paraspinal muscle CSA to spinal canal CSA or disc CSA (14,16) were calculated for all muscle variables to compensate for differences in body weight and body conformation among dogs. Furthermore, the paraspinal muscle CSA to body weight ratio was obtained.

The FCSA was quantified in the overall paraspinal muscle to investigate the degree of fat infiltration. The maximal signal intensity of lean muscle was obtained based on seven times manual segmentation of the most homogenous muscle within the paraspinal muscle. The threshold was determined by the maximal signal intensity of the lean paraspinal muscle. The fat tissue and FCSA were separated with a threshold using ImageJ software. The percentage of FCSA was calculated and expressed to the value of FCSA/CSA × 100. A smaller value of FCSA/CSA × 100 indicated a greater degree of fat infiltration. Several studies have shown that the measurements of the CSA and FCSA using this threshold are reliable (21).

Left and right paraspinal muscle CSA was calculated as a ratio at the level of the T12-T13 intervertebral disc space.

CSA asymmetry ratio = Right muscle CSA/Left muscle CSA

If ratio ≥ 1, Symmetry % = (Ratio-1) × 100

If ratio < 1, Symmetry % = –([1/Ratio]-1) × 100

With this calculation, an asymmetry index value of 0 would mean perfect symmetry; a value of 100 or –100 indicates a twofold difference in area between the paraspinal muscles on each side. The asymmetry index was used to determine if the paraspinal muscle CSA was altered asymmetrically depending on whether the disc herniation was one-sided (right and left) or midline (6,12,22).

Statistical analysis

All statistics were analyzed using the Statistics Package for the Social Science (SPSS, ver.26; SPSS Inc, IL, USA) with the statistical significance set to p < 0.05. For the normality test, the Shapiro-Wilk test was performed. All data are expressed as the median, minimum, maximum values or mean, and standard deviation using descriptive statistics. The acute and chronic IVDH groups were compared using Mann Whitney u-tests or t-tests depending on the normality test, in terms of age, body weight, muscle CSA, muscle FCSA/CSA, spinal canal CSA, disc CSA, a ratio of muscle CSA to spinal canal CSA or disc CSA, and muscle CSA to body weight ratio. The relationship between body weight and muscle CSA was examined using Spearman correlation analysis. The CSA asymmetry indices were compared among the midline, left-sided, and right-sided disc herniation groups using One-way ANOVA. The CSA asymmetry indices of the midline, right, and left disc herniation were evaluated using a test value of 0 of the One Sample t-test. Three veterinarians with two or three years of experience conducted all measurements. Three raters measured the signal intensity of lean muscle, CSA, and FCSA on each scan slice for an individual subject in random order. One veterinarian with three years of experience measured each variable three times. Reliability analysis was performed to assess the intra-rater and inter-rater reliability of the measurements.

The IVDH group was composed of 126 dogs of various breeds and ages. Thirty-one dogs met the criteria for inclusion in the study. There were eleven Maltese, five Poodles, five mongrel dogs, three Dachshunds, two Welsh Corgis, two Pekingese, two Cocker Spaniels, and one Shih Tzu (Table 1). The study included 20 dogs with acute IVDH and 11 dogs with chronic IVDH. There were no significant differences in age and body weight between the acute and chronic IVDH groups (Table 2). The castrated male was the most common gender in the acute and chronic IVDH groups. The discs were almost herniated to the midline (Table 1). The body weight had a significant positive linear correlation with CSA of bilateral multifidus muscles and epaxial muscles (multifidus; p-value = 0.000, epaxial muscle; p-value = 0.000) (Fig. 2).

Table 1 . Distribution of the breeds, genders, and IVDH lateralization in patients.

		Acute IVDH (n = 20)	Chronic IVDH (n = 11)	Total (n = 31)
Breed	Mongrel	2	3	5
	Welsh Corgi	1	1	2
	Pekingese	0	2	2
	Shih Tzu	1	0	1
	Dachshund	2	1	3
	Maltese	8	3	11
	Poodle	5	0	5
	Cocker Spaniel	1	1	2
Gender	Female	1	2	3
	Spayed female	7	3	10
	Male	2	0	2
	Castrate male	10	6	16
Side of disc herniation	Midline	17	6	23
	Left & right	3	5	8

Table 2 . Median, minimum, and maximum values of age, body weight, and duration of clinical signs in acute and chronic IVDH groups.

Clinical data	Acute IVDH (n = 20)				Chronic IVDH (n = 12)				Total (n = 32)				p-value
	Med	Min	Max		Med	Min	Max		Med	Min	Max
Age (years)	6.5	1.0	12.0		8.0	3.0	11.0		7.0	1.0	12.0		0.555
Body weight (kg)	5.0	2.0	16.0		8.0	3.0	10.0		5.0	2.0	16.0		0.087
Duration of clinical signs (days)	1.0	0.0	7.0		28.0	10.0	180.0		3.0	0.0	180.0		0.000

Med, median; Min, minimum; Max, maximum..

p-value: Less than 0.05 is significantly valuable..

Figure 2. Linear relationship between body weight and CSA of paraspinal muscles.

Cross-sectional area (CSA) ratio

The left and right epaxial muscle CSA to the body weight ratios were significantly higher in acute IVDH patients compared to the chronic group (left epaxial muscle to body weight ratio: p-value = 0.007, right epaxial muscle to body weight ratio: p-value = 0.000) (Table 3). CSA of the paraspinal muscle and ratio of paraspinal muscle CSA to spinal canal CSA or disc CSA showed no significant difference between patients with acute and chronic groups (p-value > 0.05).

Table 3 . Median, minimum, and maximum values of the epaxial muscle CSA to spinal canal CSA, disc CSA, and ratio to the body weight.

Clinical data		Acute IVDH (n = 20)				Chronic IVDH (n = 11)				Total (n = 31)				p-value
		Med	Min	Max		Med	Min	Max		Med	Min	Max
EM CSA/SC CSA	Lt	7.20	2.90	11.60		6.34	3.89	11.20		7.08	2.90	11.60		0.549
	Rt	7.19	4.04	14.80		6.63	4.40	11.42		6.75	4.04	14.80		0.451
EM CSA/D CSA	Lt	2.10	0.95	3.42		1.80	1.29	2.41		1.97	0.95	3.42		0.095
	Rt	2.15	1.26	3.43		1.87	1.46	2.46		1.95	1.26	3.43		0.087
EM CSA/BW	Lt	41.43	19.38	52.25		33.40	14.26	41.43		38.64	14.26	52.25		0.007
	Rt	39.15	19.33	57.73		34.34	16.14	45.95		37.72	16.14	57.73		0.000

Med, median; Min, minimum; Max, maximum; EM, epaxial muscle; SC, spinal canal; D, disc; BW, body weight; Lt, left; Rt, right..

p-value: Lesser than 0.05 is significantly valuable..

Fat infiltration

Fat infiltration was evaluated by the percentage of FCSA to CSA for the individual muscle. The degree of fat infiltration was significantly higher in dogs with chronic IVDH than the acute IVDH (Table 4). Fig. 3 shows representative images of fatty infiltration in the multifidus and epaxial muscles with acute and chronic IVDH groups.

Table 4 . Multifidus and epaxial muscle FCSA/CSA (%) of acute and chronic IVDH groups.

		Acute IVDH (n = 20)				Chronic IVDH (n = 11)				Total (n = 31)				p-value
		Med	Min	Max		Med	Min	Max		Med	Min	Max
M FCSA/CSA	Lt	76.35	58.80	95.40		61.40	36.70	86.70		72.30	36.70	95.40		0.014
	Rt	79.45	65.80	99.00		61.10	39.90	76.40		75.60	39.90	99.00		0.000
EM FCSA/CSA	Lt	87.10	59.20	96.90		63.90	33.20	79.70		79.00	33.20	96.90		0.001
	Rt	86.75	42.50	98.00		64.90	41.30	87.60		77.60	41.30	98.00		0.001

M, multifidus; EM, epaxial muscle; Med, median; Min, minimum; Max, maximum; FCSA, functional cross-sectional area; CSA, cross-sectional area; SC, spinal canal; BW, body weight; Lt, left; Rt, right..

p-value: Lesser than 0.05 is significantly valuable..

Figure 3. (A, B) Cross-sectional area images of an acute IVDH dog with sudden hindlimb ataxia 1 day ago. (C, D) The cross-sectional scan image of chronic IVDH patient during three weeks with mild hindlimb ataxia. (B, D) The red area surrounded by a white outline is muscle FCSA. When comparing FCSAs, a patient with chronic IVDH has more fat infiltration than a dog with acute IVDH.

Muscle asymmetry indices

The location of most IVDH lesions was the midline: 21 dogs of the midlines and 6 dogs of right-sided and 4 dogs of left-sided disc herniation. The asymmetry indices of the muscle CSA of the midline, right-sided, and left-sided groups were not significantly asymmetric across all muscles (p-value > 0.05, test value = 0). But there was a difference between right-sided and left-sided disc herniation groups. In the right-sided disc herniation group, the right epaxial muscle CSA was significantly smaller than the left-sided disc herniation group (Table 5). The asymmetry indices of the epaxial muscle CSA of the right-sided disc herniation group were significantly lower than the left-sided disc herniation group.

Table 5 . Mean value of the muscle asymmetry indices and difference among the midline, right-sided, and left-sided disc herniation groups.

	Midline mean	Right-sided mean	Left-sided mean	Mid vs Rt p-value	Mid vs Lt p-value	Rt vs Lt p-value
Rt. EPAX CSA	202.56	162.60	311.55	0.554	0.056	0.024
Lt. EPAX CSA	200.15	164.96	271.10	0.635	0.276	0.136
EPAX-ASY	2.82	–1.50	18.85	0.814	0.077	0.049

Mid, midline; Rt, right; Lt, left; Rt. EPAX CSA, right epaxial muscle CSA; Lt. EPAX CSA, left epaxial muscle CSA; EPAX-ASY, asymmetry indices of the epaxial muscle CSA..

p-value: lesser than 0.05 is significantly valuable..

Reliability analysis

The inter-observer agreement for the CSA and the fat content from signal intensities of MRI of paraspinal muscles were excellent (ICC, 0.745-0.995). The measurements by one veterinarian with three years of experience showed good reliability (ICC, 0.886-0.999).

The anatomy, nerve innervation, and function of the canine multifidus and longissimus muscles are not significantly different from that in humans (15,18,19,23,24). Despite the apparent differences in gait posture between bipeds and quadrupeds, humans and dogs share several spinal biomechanical features, including similarities in axial compressive loads and the pathogenesis of disc degeneration in non-chondrodystrophic breeds (3,7). Therefore, changes in the paraspinal muscle of dogs with disc herniation are expected, as identified in human patients with lower back pain.

The causal relationship between the paraspinal muscle and IVDH is unclear (7). In humans with chronic lower back pain, nerve root compression or irritation may exist and induce muscle denervation (1,7). Reduced activation of the paraspinal muscles leading to disuse atrophy may be caused by pain-guarding behavior, reflex inhibition, or inflammation in affected humans (7). Asymmetry of the bilateral muscle CSA and increased fat infiltration can be induced by pain as an inhibitor of the motion on the symptomatic side (26). Compensatory hypertrophy could occur on the non-painful side, which causes an imbalance of the paraspinal muscles (26). The decreased CSA of the multifidus muscle at the problem level may lead to local muscle weakness and instability of the spine, and then instability of the adjacent vertebral levels makes the muscle more vulnerable to atrophy (26). Therefore, a study to identify the causal relationship between the paraspinal muscle and IVDH would be valuable.

In this study, epaxial muscle CSA in dogs with chronic IVDH was less than in the acute group when compensating for the patient’s body shape or weight. The epaxial muscle may be an important part of identifying changes throughout the IVDH in dogs. The epaxial muscle plays a major role in proper locomotion and spinal stability (9). In humans, significant relationships between the multifidus and lower back pain have been investigated (13), while there was an unclear connection between the multifidus and IVDH in dogs. This could be because 1) the area of the multifidus in dogs is too small for accurate measurement and evaluation; 2) the epaxial muscle occupies the largest area of the same level of muscle in the spine, compared to the multifidus mainly doing the role of fixation; 3) the epaxial muscle may be affected further by IVDH owing to the role of vertebral fixation, extension, and lateral movement in dogs (11).

Degenerative alteration of the paraspinal muscles with fat infiltration was detected in dogs with chronic IVDH. While atrophy of the muscle CSA was identified only in the epaxial muscle, reduced FCSA was confirmed in the multifidus and epaxial muscle in this study. Muscle composition may be a better indicator of the pathology of paraspinal muscle wasting than the muscle CSA. Human research on the response of the multifidus to IVDH, with symptoms lasting for less than six weeks, demonstrated changes in fat infiltration at the lesion side without a decrease in the muscle CSA (2). There have been reports of a significant increase in fat tissue of multifidus and epaxial muscle with chronic neurological symptoms over three months (17).

Ipsilateral atrophy of the paraspinal muscles in human patients with low back pain was observed on the symptomatic side. The extent of the change was significantly greater in chronic low back pain in the erector spinae muscle (26). This study exhibited no remarkable asymmetry in muscle CSA of IVDH patients dependent on the location of disc herniation. But there was a significant difference in asymmetry indices of epaxial muscle CSA between right and left-sided disc herniation. Left or right-sided lateralization of a disc herniation may be associated with an asymmetric change of epaxial muscle. Right epaxial muscle CSA in the right-sided disc herniation group was significantly smaller than in the left-sided disc herniation group. There are factors to consider that each group is made up of a small population and differences in the degree of clinical signs among the patients. The ratio of chronic and acute disc herniation patients in each group is the same at 1:1, so the period of clinical signs is less relevant to the previous results.

The limitations of this study include its retrospective nature and a small number of cases. Another weakness is that the assessors drawing the ROIs were not blinded to the side of the disc lesion on the MR images. On the other hand, the two observers did not know the history, signalment, or diagnosis of the patients at the time of the measurements. Third, each dog has a different body shape, and no factors other than body weight have been found to compensate for it. Identifying the appropriate factors to recompense various dog-specific body types and body weight will be needed. Further studies are required in a larger and more standardized population of dogs. Finally, objective evaluation of patient obesity was difficult and could not be reflected in the patient comparison. An evaluation of the FCSA with the body conditional score was considered more useful.

MRI is useful for evaluating muscle CSA and composition quantitatively because of its good soft-tissue contrast. Using MRI, this study found that dogs with chronic IVDH over seven days had significantly greater fat infiltration and reduced CSA of epaxial muscle compared to dogs with acute IVDH. The present findings showed asymmetric degeneration of the paraspinal muscles that are specific to the symptomatic side in one-sided disc herniation groups.

This study was supported by the research fund of Chungnam National University.

The authors have no conflicting interests.