LUMBOSACRAL (LOWER BACK) DISEASE IN WORKING DOGS

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: COMPLETE
• Funding Source: HATCH
• Reporting Frequency: Annual
• Accession No.: 0230018
• Project Start Date: May 18, 2012
• Project End Date: Jan 28, 2016
• Program Code: [(N/A)]- (N/A)

Recipient Organization
WEST VIRGINIA UNIVERSITY
886 CHESTNUT RIDGE RD RM 202
MORGANTOWN,WV 26505-2742

Performing Department
Animal & Veterinary Science

Non Technical Summary
Working dogs are high-performance athletes that perform vital functions such as protecting livestock, screening for agricultural pathogens, assisting people with psychological or mobility limitations, searching for and rescuing people who are lost or injured, assisting in disaster relief, and supporting police and military operations for the U.S. and around the world. Commonly used breeds include German Shepherds, Belgian Malinois, Labrador Retrievers and Golden Retrievers. Lumbosacral (lower back) disease is an important cause of disability in these valued animals, often affecting them in middle age when they are at the peak of their performance. Signs include lameness, sensitivity in the lower back region, abnormal tail carriage, and/or reluctance to do working tasks that require movement of the lower back region. As with human athletes, lumbosacral disease in working dogs is often a multi-factorial problem and causes are incompletely understood. Proposed mechanisms include soft tissue injury, degenerative joint disease, degenerative disc disease, developmental malformation of skeletal structures, and spinal instability. If left untreated, the dogs often compensate by shifting their weight and altering how they do their working tasks. This can place additional strain on other spinal regions and joints. Symptomatic treatments for lumbosacral disease in working dogs initially include rest and non-steroidal anti-inflammatory medications. However, some dogs fail to respond to these treatments. Long term rest is not a viable option for these dogs, as teams depend on them and long term rest would cause loss of muscle mass that would further delay return to duty. Sports medicine treatments and preventative measures developed for human athletes show promise for helping working dogs return to active duty sooner and maintain active duty for a longer duration. The proposed study will use non-invasive diagnostic imaging tests, computerized image analyses, and genetic tests to determine the causes of and risk factors for lumbosacral disease in working dogs. Results will help primary care veterinarians and researchers develop better treatments and preventative measures for use by working dog owners, handlers, and trainers.

Animal Health Component

100%

Research Effort Categories

Basic

(N/A)

Applied

100%

Developmental

(N/A)

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
303	3830	1160	50%
311	3830	1160	50%

Knowledge Area
303 - Genetic Improvement of Animals; 311 - Animal Diseases;

Subject Of Investigation
3830 - Pets (companion animals);

Field Of Science
1160 - Pathology;

Keywords

working dog

lumbosacral disease

lumbosacral stenosis

lower back pain

computed tomography

motion analysis

genetic testing

Goals / Objectives
Research objectives: Evaluate the causes of and risk factors for lumbosacral disease in working dogs using clinical examination, computed tomography, motion analysis, occupational exposure analysis, and genetic analysis. Educational objectives: Provide research training for 1 current PhD student, 2 new MS students, and 10-20 undergraduate students. Timeline: Year 1: Submit abstract for presentation at ACVR meeting. Recruit 1st MS student. Submit grant proposals. Collect pilot data for PhD student project. Submit publications for previously completed studies. Year 2: Collect and analyze data for PhD and MS student projects. Submit grant proposals. Submit PhD student abstract for national meeting. Submit publications for Year 1 studies. Year 3: Continue data collection and analysis for MS and PhD student projects. Submit grant proposals. Submit MS student abstract for national meeting. Recruit 2nd MS student. Submit publications for Year 2 studies. Year 4: Continue data collection and analysis for MS and PhD student projects. Submit grant proposals. Submit PhD student abstract for national meeting. Submit publications for Year 3 studies. Year 5: Continue data collection and analysis for MS and PhD student projects. Submit grant proposals. Submit MS student abstract for national meeting. Submit publications for Year 4 studies.

Project Methods
Dogs: We will recruit 50 working dogs through partnerships with four major referral veterinary hospitals. Licensed veterinarians will perform all clinical examinations and supervise all data collection. Dogs will be excluded if diseases that preclude safe anesthesia are identified or if there is CT evidence of fractures, previous lumbosacral surgery, infection or neoplasia. Clinical and occupational exposure data: Primary care veterinarians will determine presence or absence of clinical signs of lumbosacral disease using previously published criteria. Handlers or owners will complete an occupational exposure questionnaire describing working tasks, and any observed deficits in performance. Genetic data: Blood samples will be collected and genetic assays will be performed for included dogs using previously established techniques in Dr. Jianbo Yao's laboratory in the Davis College of WVU. CT data: Computed tomography scans of the lumbosacral region will be acquired at partnering hospitals. Dogs will be sedated or anesthetized as directed by the primary care veterinarian and then placed in a dorsally recumbent position with hind legs extended and secured to the table. A repeat scan of the lower back region will be acquired with the hindlegs positioned in flexion. CT scan data will be interpreted by a board-certified veterinary radiologist (JJ) who will be unaware of clinical findings at the time of review. Qualitative data recorded will include the presence or absence of CT lesions involving the sacroiliac joints, coxofemoral joints, lumbosacral spine, and paraspinal soft tissues. Presence of concurrent lesions in other anatomic regions will also be recorded. Quantitative measures such as lumbosacral paraspinal muscle area, paraspinal muscle density, vertebral body density, lumbosacral angle, lumbosacral foraminal areas, vertebral canal area/vertebral body area ratios, and articular process joint angles will be recorded by student researchers using computerized image analysis. Motion data: Motion capture video data will be recorded for selected dogs. Animals will first be given a chance to acclimate to the room and to practice walking alongside their trainer in the motion capture area. Obstacles will also be set up to mimic those encountered by dogs during their regular working tasks. The dogs will be fitted with reflective markers at each of the principal limb joints (e.g., hip, knee, paw) and at 4-inch intervals along the dorsal spine. Additional markers will be used on the head, nose, and tail. Once the markers are all secured, the handler will lead the dog walking across the room and through these working tasks. Motion data will be recorded at 120 frames/second and analyzed using tracking software. Data analyses: Associations between CT, clinical, occupational exposure, genetic, and motion analysis findings will be determined and odds ratios for risk of lumbosacral disease will be calculated.

Progress 05/18/12 to 01/28/16

Outputs
Target Audience:Nothing to report, PI has left the university. Changes/Problems:Nothing to report, PI has left the university. What opportunities for training and professional development has the project provided?Nothing to report, PI has left the university. How have the results been disseminated to communities of interest?Nothing to report, PI has left the university. What do you plan to do during the next reporting period to accomplish the goals?Nothing to report, PI has left the university.

Impacts
What was accomplished under these goals? Nothing to report, PI has left the university.

Publications

Progress 10/01/13 to 09/30/14

Outputs
Target Audience: Veterinarians, military and law enforcement agencies, working dog training/breeding organizations, canine and human sports medicine researchers Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? One PhD student and two undergraduate students gained research experience, presentation experience, and first author abstract publications in a major veterinary scientific journal. The PI and other project members gained national recognition. How have the results been disseminated to communities of interest? Scientific poster session at ACVR national meeting and 2 published abstracts in the Veterinary Radiology & Ultrasound journal. Scientific poster session at the PAG conference. What do you plan to do during the next reporting period to accomplish the goals? Submit publications based on work completed in 2013-14. Complete data analysis for PhD student genetics study and submit paper. Complete 2 new undergraduate research projects and submit abstracts for poster sessions.

Impacts
What was accomplished under these goals? Abstract 1: Digital CT scans and corresponding gross pathology slice photographs of the sacroiliac region were retrieved from data previously acquired for a cadaver study on ultrasound sacroiliac-guided injection in dogs. Fifteen, adult, mixed breed dogs were evaluated (30 sacroiliac joints). A pre-veterinary student (MC) and a veterinary radiologist (JJ) reviewed CT images and recorded presence or absence of the following CT lesions previously reported to be characteristics of clinically significant sacroiliac joint disease in humans: subchondral cyst, subchondral sclerosis, subchondral erosions, intra-articular ankylosis, and para-articular ankylosis. Findings were recorded in three independent review sessions separated at least a week apart, with the order of dogs randomized between reading sessions. A fourth reading session was then used to resolve any discordant interpretations. A veterinary pathologist (PS) reviewed gross pathology slice photographs using the same protocol and recorded presence or absence of the same lesions as those described above. Data recorded during the 4th reading session for each test were used for statistical comparisons. A statistician (IH) compared agreement between CT and gross pathology for sacroiliac joint lesion detection using McNemar's test and Kappa statistics. In this sample of 15 dogs, prevalence of lesions consistent with sacroiliac joint disease was 100% based on CT and 100% based on gross pathology slices. More subchondral cyst and intra-articular ankylosis lesions were detected using CT versus gross pathology. More subchondral sclerosis and subchondral erosion lesions were detected using gross pathology versus CT. Para-articular ankylosis was detected only with CT. Proportions of detected subchondral cyst lesions differed between CT and gross pathology for left and combined right/left sacroiliac joints (p < 0.05, McNemar's). Overall agreement was poor between CT and gross pathology for detection of subchondral cyst, subchondral sclerosis, subchondral erosion, and intra-articular ankylosis lesions in combined right/left sacroiliac joints (Kappa < 0.20). Abstract 2: Whole body computed tomography (CT) scans were acquired for 10 military working dogs presented to the Department of Defense Military Working Dog Veterinary Service at Lackland, TX. Five dogs were presented for lower back pain, and five dogs were presented for other reasons and had no clinically detected lower back pain. All CT scans were acquired using the same 64-slice CT scanner. Scans were archived, stored and transferred to an image analysis workstations (Apple Inc. Cupertino, CA) for analysis using CT image analysis freeware (OsiriX DICOM Viewer, Version 4.2). A board-certified veterinary radiologist unaware of lower back pain status viewed the scans and recorded location and characteristics for all observed musculoskeletal lesions. Lesions were grouped into one of fifteen anatomic region categories (cervical spine, shoulder, carpus, thoracic spine, etc.). After all scans were read, the total number of lesions was determined for each dog and each anatomical region. A statistician performed comparisons between dogs in the lower back pain group versus dogs in the no low back pain group using one-way analysis of variance (ANOVA). Each group included 1 female and 4 male. Ages ranged from 2 to 10 years, with even distribution of ages between groups. All dogs in the lower back pain positive group were German Shepherds. Dogs in the lower back pain group included 1 German Shepherd, 1 Labrador Retriever, and 3 Belgian Malinois. On average, dogs with lower back pain had a greater number of musculoskeletal lesions (60.6) throughout their whole body than dogs without lower back pain (21) (P=0.003). A significant difference was found in the number of lesions between the two groups for all regions of the spine except the cervical: thoracic (P=0.015), lumbar (P=0.026), lumbosacral (P=0.002), caudal (P=0.010). A significance difference was also detected in the tarsus (P=0.049). No differences were identified between groups for the remaining anatomic regions Abstract 3: Computed tomography scans of the lumbosacral region were taken for 40 Labrador retriever military working dogs at Lackland Air Force Base. Using previously established qualitative and quantitative criterion for LS phenotype, each dog was designated as either positive or negative for LS at each location of the lumbosacral spine. Qualitative assessment was made by board-certified veterinary radiologist (JJ) and the quantitative assessment by an individual (MM) trained in CT image analysis. Using logistic regression (IH), 8 dogs were selected which had the best agreement between the qualitative and quantitative assessments - 4 positive for LS and 4 negative for LS. Genomic DNA was isolated using GenSolve DNA Recovery kit (GenTegra, Pleasanton, CA, USA) from peripheral blood drops collected on FTA cards (Whatman Plc., Maidstone, UK) for the 8 dogs. The whole exome of these 8 dogs will be sequenced using a whole exome capture kit (Nextera, Illumina, San Diego, CA, USA). Sequencing data will be analyzed and compared with each other, as well as with the canine reference genome to identify any sequence variation in the exome that might be significantly associated with LS in Labrador retrievers.

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Jula C, Jones J, Holaskova I, Childs K. Whole body musculoskeletal lesions in military dogs with versus without lower back pain. In Abstracts of the ACVR Annual Scientific Meeting Poster Session, St. Louis, MO. Oct. 22-24, 2014. Vet Radiol Ultras 2014; 55(6): 696.
• Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Carnevale M, Jones J, Sponenberg P, Holaskova I. Comparison between CT and gross pathology for detection of sacroiliac lesions in dogs. In Abstracts of the ACVR Annual Scientific Meeting Poster Session, St. Louis, MO. Oct. 22-24, 2014. Vet Radiol Ultras 2014; 55(6): 695.
• Type: Conference Papers and Presentations Status: Published Year Published: 2014 Citation: Mukherjee M, Yao J, Welsh A, Jones J, Holaskova I. Genetic factors associated with lumbosacral stenosis in Labrador retrievers. Poster session of the XXIII International Plant and Animal Genome Conference, Jan. 10-14, 2015. San Diego California.

Progress 01/01/13 to 09/30/13

Outputs
Target Audience: Veterinarians, military and law enforcement agencies, working dog training/breeding organizations, canine and human sports medicine researchers Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? One PhD student and two undergraduate students gained research experience, presentation experience, and first author abstract publications in a major veterinary scientific journal. The PI and other project members gained national recognition. How have the results been disseminated to communities of interest? Scientific poster session at ACVR national meeting and 3 published abstracts in the Veterinary Radiology & Ultrasound journal. What do you plan to do during the next reporting period to accomplish the goals? Submit publications based on work completed in 2013. Complete data analysis for PhD student genetics study and submit abstract for presentation at PAG genetics national meeting in Jan. 2015. Complete data collection and analyses for three additional projects and submit abstracts for poster session of ACVR meeting in October. Travel to military dog hospital during summer to collect new data.

Impacts
What was accomplished under these goals? Abstract 1: Computed tomography (CT) databases from the Department of Defense Military Working Dog Veterinary Service were searched for Belgian Malinois that had scans of the lumbosacral region. Digital CT images were uploaded directly to an image analysis workstation. Hard copy CT images were converted to DICOM format using a digital scanner system (Vidar Sierra Advantage, Sound Eklin, Carlsbad, CA) and then transferred to the image analysis workstation. All images were analyzed using the same software (Osirix version 4.1.2). Centimeter scales were used for calibration of area measurements in hard copy images. A single observer unaware of clinical status traced triplicate regions of interest (ROIs) around multifidus, psoas, quadratus, longissimus, and serratus dorsalis lateralis muscles where visible at the L5-6, L6-7, L7-S1, and S1-2 vertebral levels. Regions of interest were also traced around vertebral bodies at the same locations as muscle ROI’s and used to calculate muscle/vertebral body area ratios. After all CT measurements were recorded, medical record data were reviewed and dogs were divided into two groups: dogs with lumbosacral pain and dogs without lumbosacral pain. Mean CT area ratios for each muscle were calculated for each dog group and compared between groups. Nine digital and 8 hard copy CT studies were included in analyses. Of the 17 Belgian Malinois included in the study, 13 were males and 4 were females. Twelve dogs were in the lumbosacral pain positive group, and 5 dogs were in the lumbosacral pain negative group. All paraspinal muscle groups in dogs with lumbosacral pain had graphically lower mean area ratios, except the quadratus lumborum, which was higher by 0.06. Findings were consistent with previous human studies and supported the theory that decreased paraspinal muscle mass may be a quantitative indicator of lower back pain in working Belgian Malinois. Future studies are needed to determine the effect of therapeutic interventions on paraspinal muscle mass in dogs with lower back pain. Abstract 2: Three golden retrievers were recruited from the Hearts of Gold service dog training program. Clinical examination and positional computed tomography (CT) findings were recorded by observers unaware of motion analysis findings. Dogs were trained to wear a body suit (K9 Topcoat, LeadDog, Inc., Talent, OR) and walk on a treadmill prior to motion analysis. Each dog participated in two motion analysis sessions, with at least 3 walking trials per session and at least 5 minutes of rest between trials. The treadmill velocity for the first trial in each dog was determined based on the velocity that allowed a steady walking pace, and was repeated for subsequent trials. For each motion analysis session, one author (JJ) placed markers at the following palpable bony landmarks using double sided wig tape: caudodorsal angles of scapulas, T1 spinous process, T6 spinous process, mid-lateral thoracic wall rib margins, T13 spinous process, L6 spinous process, caudodorsal angles of ilial wings, and ischiatic tuberosities. Another author (BS) placed tracking markers. Marker data were recorded and visualized using an eight-camera optical motion capture system (MX Series, Vicon Motion Systems, Oxford, UK) and visualization software (Vicon Nexus 1.7.1). For the second session, all markers were removed and replaced using the same landmarks as above. Observers unaware of clinical and imaging findings (BS,CM,AN) used kinematic computation software (C-MotionVisual3D 4.8.9. C-Motion Inc., Germantown, MD) to model 3D marker data and compute joint kinematics for each trial in each dog. Mean ROM’s for the 3 dogs were compared for each XYZ axis and each session. Dog ages ranged from 4-7 years. There was 1 male and 2 females. Two dogs had abnormal gait on clinical exam and 1 dog was normal. All dogs had degenerative sacroiliac joint disease, 2 dogs had degenerative coxofemoral joint disease, and 1 dog had degenerative LS disease with positional instability. Correlation coefficients for ROM’s between motion analysis sessions ranged from 0.84-0.99. Findings indicated that lumbosacral motion analysis is repeatable for golden retrievers with varying clinical and CT status. Abstract 3: Computed tomography (CT) archives at the Holland Military Working Dog Hospital and the Virginia Tech Veterinary Teaching Hospital were searched for CT scans of Labrador retrievers that included the lumbosacral spine. Scans with evidence of previous surgery, neoplasia, fractures or infection were excluded. The remaining scans were transferred to image analysis workstations (Apple Inc. Cupertino, CA) for analysis using CT image analysis freeware (OsiriX DICOM Viewer, Version 4.2). A single observer made measurements unaware of clinical findings. Quantitative variables recorded for each dog were spinal canal area, canal fat area, canal density and vertebral body area. Each variable was measured in triplicate and at each of 5 vertebral locations (caudal L5, cranial L6, caudal L6, cranial L7 and caudal L7). All area measurements were in cm2 and densities in Hounsfield units (HU). Means for triplicate measures and canal/vertebral body area ratios were calculated for each dog and each location. Quantitative variable means and ranges for all dogs were graphed and described. A total of 23 dogs met inclusion criteria. Across all 5 locations in all dogs, mean canal areas ranged from 0.56 to 1.74; mean canal/vertebral body area ratios ranged from 0.21 to 0.79; mean canal fat areas ranged from 0.00 to 1.12; and mean canal densities ranged from -48.67 to 157.02. Spinal canal, canal fat, and vertebral body area measurements had higher intra-reader repeatability than canal density measurements. Findings indicated that quantitative CT phenotypic characteristics of the lumbar spinal canal vary widely within the Labrador retriever breed. Follow-up studies are in progress to describe qualitative phenotypic characteristics of the lumbar spinal canal for this breed.

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2013 Citation: 1. Cain B, Jones J, Childs K, Mukherjee M. CT morphometry of paraspinal muscles in working Belgian malinois with versus without lumbosacral pain. In: Abstracts for the 2013 ACVR Scientific Conference: Savannah, Georgia. Vet Radiol Ultras 2013: 54(6); 680-698. 2. Stewart B, Jones J, Moore C, Nimbarte A, Meade J. Repeatability of lumbosacral motion analysis in golden retriever service dogs. In: Abstracts for the 2013 ACVR Scientific Conference: Savannah, Georgia. Vet Radiol Ultras 2013: 54(6); 680-698. 3. Mukherjee M, Jones J, Childs K, Pierce B, Grimm P. CT phenotypic characteristics of the lumbar spinal canal in Labrador retriever dogs. In: Abstracts for the 2013 ACVR Scientific Conference: Savannah, Georgia. Vet Radiol Ultras 2013: 54(6); 680-698.

Progress 01/01/12 to 12/31/12

Outputs
OUTPUTS: Human studies have identified decreased computed tomographic (CT) areas and densities for multifidus, psoas, quadratus, and longissimus muscles in patients with lower back pain and proposed that disuse atrophy of these stabilizing muscles may contribute to recurrence or persistence of lower back pain (Daneels et al, 2000; Kamez et al, 2007, Bouche et al, 2011). The purpose of the current study was to describe CT areas and densities of similar muscles in Labrador Retrievers with versus without lower back pain (lumbosacral pain, LSP). Medical record and CT databases at the Virginia-Maryland Regional College of Veterinary Medicine and Department of Defense Military Working Dog Veterinary Service were searched for Labrador Retrievers that had CT scans including the lumbosacral region. Computed tomography data were imported into image analysis freeware (Osirix version 4.1.2). A single observer unaware of medical record findings used hand-traced regions of interest to measure transverse CT areas and mean CT densities for multifidus, psoas/iliopsoas, quadratus, and longissimus muscles where visible at L5-6, L6-7, L7-S1, and S1-2 vertebral levels (Smallwood and Thomas, 1982). Vertebral body areas were measured at the same locations as muscle area measurements and used to calculate muscle/body area ratios. After all CT measurements were recorded, medical record data were reviewed and dogs were divided into four groups: working dogs with LSP, working dogs without LSP, companion dogs with LSP, and companion dogs without LSP. Lumbosacral pain was defined as medical record evidence of a painful response to palpation of the lumbosacral region and/or elevation of the tail. Mean CT area ratios and CT densities for each muscle were calculated for each dog group and graphically compared. Results were presented in a poster session of a national meeting and published as an abstract. PARTICIPANTS: B. Francis1, J. Jones1, B. Pierce2, K. Childs3, P. Grimm3, M. Mukherjee1. 1West Virginia University, WV, 26506-6108; 2Virginia-Maryland Regional College of Veterinary Medicine, VA, 24061-0442; 3Department of Defense Military Working Dog Veterinary Service, TX, 78236 Brandy Francis was an undergraduate research student at the time of the study. She has since gradated. Meenakshi Mukherjee is a PhD student and currently employed as a graduate research assistant. TARGET AUDIENCES: Veterinarians, military and law enforcement agencies, and working dog training/breeding organizations. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.

Impacts
Eighteen Labrador Retrievers were included in the study, 12 males and 6 females. Groups consisted of 1 companion dog without LSP, 14 companion dogs with LSP, 2 working dogs without LSP, and 1 working dog with LSP. Mean area ratios were lower for all paraspinal muscles in working and companion dogs with LSP, except for the longissimus lumborum. Mean CT densities were lower for all paraspinal muscles in working dogs with LSP, except for the quadratus lumborum. Mean CT densities were lower for longissimus and quadratus lumborum muscles and higher for psoas/iliopsoas and multifidus muscles in companion dogs with LSP. Most of the findings in our study were similar to those reported in human studies. Higher area ratios for longissimus muscles and higher mean densities for psoas/iliopsoas, quadratus lumborum, and multifidus muscles in dogs with LSP were unexpected and may warrant further investigation.

Publications

• Francis B, Jones J, Pierce B, Childs K, Grimm P, Mukherjee M. CT morphometry of paraspinal muscles in Labrador Retrievers with versus without lumbosacral pain. In: Abstracts for the 2012 Scientific Conference: Las Vegas, Nevada, Octopber 18-21, 2012. Vet Radiol Ultras 2012: 53:6; 690.