SUBCLINICAL CT ABNORMALITIES IN THE LUMBOSACRAL SPINE OF OLDER LARGE-BREED DOGS

Computed tomography (CT) of the L5-S3 vertebral levels was performed in six, large-breed dogs presented for problems unrelated to the lumbosacral spine. All dogs were asymptomatic for lumbosacral stenosis on neurologic examination. Breeds included German Shepherd, Golden Retriever, Boxermix and Belgian Malinois. Ages ranged from 5-12 years. Five out of six dogs exhibited CT abnormalities. Among the 18 disc levels examined, the most common findings were idiopathic stenosis, loss of vertebral canal epidural fat, and nerve tissue displacement. Less common abnormalities were vertebral canal or foraminal bone proliferation, loss of intervertebral foramen fat, vertebral canal disc bulging, degenerative articular process joint disease, transitional vertebra, dural ossification, foraminal disc bulging, Schmorl's nodes, calcified extruded disc fragment, and sacroiliac joint osteophytes. Vertebral subluxation was absent in all dogs. Findings indicate that some lumbosacral CT abnormalities may be clinically insignificant, especially in older dogs.     

Use of computed tomographic densitometry to quantify contrast enhancement of compressive soft tissues in the canine lumbosacral vertebral canal

OBJECTIVES: To evaluate computed tomography (CT) densitometry as a technique for quantifying contrast enhancement of compressive soft tissues in the canine lumbosacral vertebral canal and to determine whether the degree of contrast enhancement can be used to help predict tissue type or histopathologic characteristics. ANIMALS: 29 large breed dogs with lumbosacral stenosis. PROCEDURE: Contrast-enhanced CT of L5-S3 was performed by use of a previously described protocol. At each disk level, CT densities of a water-filled syringe, epaxial muscles, and 4 vertebral canal locations were measured. Mean tissue enhancement was calculated by vertebral canal location, using water-filled syringe enhancement as a correction factor. Corrected CT enhancement was compared with tissue type, degree of tissue inflammation, and degree of tissue activity. RESULTS: Intravenous contrast administration of contrast medium significantly increased CT densities of water-filled syringes and epaxial muscles. Corrected CT enhancement of vertebral canal soft tissues at stenotic sites was greater than at nonstenotic sites. There was no association between enhancement and tissue type for any vertebral canal location. There was no correlation between enhancement and degree of tissue inflammation. There was a correlation between enhancement and tissue activity in the dorsal vertebral canal only. CONCLUSIONS AND CLINICAL RELEVANCE: A water-filled syringe is a useful calibration tool for CT density measurements. The degree of tissue contrast enhancement, measured by CT densitometry, can be helpful for predicting the location of compressive soft tissues in dogs with lumbosacral stenosis. However, it is of limited value for predicting compressive soft-tissue types or histopathologic characteristics.     

Dynamic MRI is reliable for evaluation of the lumbosacral spine in healthy dogs

Magnetic resonance imaging (MRI) is commonly used to diagnose degenerative lumbosacral stenosis; however, studies show limited correlation between imaging and clinical signs. The purpose of this prospective observer agreement study was to use dynamic MRI of the lumbosacral (LS) spine of healthy dogs to determine reliable reference ranges. Twenty‐two healthy large breed dogs were prospectively enrolled. MRI of the LS spine was performed in T2‐weighted, T1‐weighted, and T2‐weighted SPACE sequences in neutral, flexed, and extended positions. Four observers performed image analyses. Measurements included LS angle, vertebral canal height and area, and LS foraminal areas. Ordinal categorical assessment of loss of fat signal in the foramina, LS compression, intervertebral disc (IVD) degeneration, spondylosis, and IVD protrusion was also performed. The majority of values were significantly larger in flexion versus neutral position, and significantly smaller in extension versus neutral position (P < .05). Subclinical compression and IVD protrusion was noted in a neutral position in 45% and 55% of dogs and in an extended position in 85% and 73% of dogs, respectively. Interobserver agreement was strong (intracluster correlation coefficient [ICC] > .5) except for the L7:LS vertebral canal area ratio (ICC ≤ .03). Intraobserver agreement was high (rho > .5) for all measurements except for the mid‐L6:LS vertebral canal height ratio (rho = .38). There was poor interobserver agreement for loss of fat signal in the foramina and evidence of compression. This study provides the groundwork for future studies using dynamic MRI to evaluate dogs with signs of clinical LS disease.     

CT measures of lumbosacral paraspinal muscle size are not correlated with CT measures of lumbosacral stability in military working Labrador Retrievers

Management of degenerative lumbosacral stenosis in military working dogs more frequently utilizes core conditioning exercise programs. Future research on the effectiveness of these programs may benefit from an improved understanding of relationships between paraspinal muscle size and lumbosacral stability. The aim of this retrospective, secondary analysis, cross‐sectional study was to test the following hypotheses related to CT measures: (a) transverse paraspinal muscle area ratios differ between hip flexion and extension, (b) lumbosacral angle and lumbosacral range of motion differ by lumbosacral angle measurement technique, and (c) transverse paraspinal muscle area ratios are correlated with CT measures of lumbosacral stability (parasagittal and parasagittal oblique foraminal area changes) regardless of hip position and with lumbosacral range of motion within each hip position. Lumbosacral CT scans in hip flexion and extension were retrieved from a previous prospective study of military working Labrador Retrievers. A single observer performed triplicate measurements of transverse paraspinal muscle area ratios, parasagittal foraminal area, and parasagittal oblique foraminal area for each hip position and another observer independently performed triplicate measurements of lumbosacral angle and lumbosacral range of motion at L7‐S1 using two published techniques. Thirty‐nine dogs were analyzed and significant differences were identified between hip flexion and extension for all mean transverse paraspinal muscle area ratio values (P ≤ .05). Mean lumbosacral angles also significantly differed between the published techniques in both hip flexion and extension. When comparing mean lumbosacral range of motion values, one of the published techniques produced significantly smaller values. No significant correlation was found between transverse paraspinal muscle area ratios and parasagittal foraminal area changes, parasagittal oblique foraminal area changes, or lumbosacral range of motion. These results should be considered when designing studies using these CT measures in military working dogs.     

COMPARISON BETWEEN COMPUTED TOMOGRAPHIC AND SURGICAL FINDINGS IN NINE LARGE-BREED DOGS WITH LUMBOSACRAL STENOSIS

In a three-year prospective study, computed tomographic (CT) and surgical findings were compared for nine large breed dogs with lumbosacral stenosis. Surgically-excised tissue was examined histologically in seven dogs and additional necropsy evaluation was performed in one dog. The CT abnormalities observed at sites of confirmed cauda equina compression were: loss of epidural fat, increased soft tissue opacity, bulging of the intervertebral disc margin, spondylosis, thecal sac displacement, narrowed intervertebral foramen, narrowed vertebral canal, thickened articular process, articular process subluxation, articular process osteophyte, and telescoped sacral lamina. The CT characteristics of lumbosacral degenerative disease and discospondylitis were similar to those described in humans. In three dogs, CT findings at the site of cauda equina compression were consistent with congenital or developmental spinal stenosis, but the method of surgical exposure precluded confirmation. Epidural fibrosis (eight dogs) and multi-level CT abnormalities (six dogs) were identified but the cause(s) and significance were unknown.     

Effect of hind limb position on the craniocaudal length of the lumbosacral space in anesthetized dogs

Objective To investigate whether rostral extension of the hind limbs increases the cranio‐caudal dorsal interlaminar distance between the seventh lumbar vertebra and the sacral bone (LS distance) in sternally recumbent anesthetized dogs. Study design Prospective clinical study. Animals Eighteen dogs (eight neutered males, three intact males, six spayed females, one intact female) of various breeds, weighing 4–34 kg and ranging in age from 1 to 13 years. Methods Each dog was grouped by size: small (≤10 kg), medium (15–20 kg) or large (≥25 kg). Each dog was anesthetized and positioned in sternal recumbency. Computed tomography (CT) of the lumbosacral area was performed with the hind limbs resting on the stifle and the feet extended posteriorly, and then with the hind limbs extended rostrally. LS distance, craniocaudal dorsal interlaminar distance between sixth and seventh lumbar vertebra (L6–L7 distance), length of L7 vertebral body and lumbosacral angle (LS angle) were measured on a reconstructed mid‐sagittal CT image from the two hind limb positions. The measurements from the two hind limb positions for the whole dog population and by size were compared using Student’s T tests. Diagnostic interpretation of the CT images was performed. Results The length of L7 was taken as the reference value as it was not affected by hind limb position. LS distance, L6–L7 distance and LS angle were significantly higher when the hind limbs were extended rostrally in all three size groups. The CT images of ten dogs showed clinically undetected osteoarthrosis of the ileo‐ and lumbosacral area. Conclusions and clinical relevance Rostral extension of the hind limbs significantly increases LS and L6–L7 distance and LS angle even in dogs with clinically undetected osteoarthrosis of the ileo‐ and lumbosacral area, and may enhance the ease of lumbosacral epidural injection in sternally recumbent anesthetized dogs.     

Radiographic measurement of the proximal and distal mechanical joint angles in the canine tibia

Objective— To describe a method for radiographic measurement of proximal and distal mechanical joint angles of the canine tibia. Normal ranges were established for a population of dogs, and specifically, Labrador retrievers, with cranial cruciate ligament rupture.
Study Design— Retrospective study.
Sample Population— Tibiae (n=105) of dogs evaluated for cranial cruciate ligament rupture; 70 tibiae were from Labrador retrievers, 35 tibiae were from other breeds.
Methods— Anatomic landmarks were established and craniocaudal radiographs were used to measure the mechanical medial proximal (mMPTA) and distal (mMDTA) tibial angles in the frontal plane.
Results— Means (±SD) for all tibiae were mMPTA, 93.30±1.78°, and mMDTA, 95.99±2.70°. For Labrador retrievers, means were mMPTA, 93.38±1.81°, and mMDTA 96.34±2.51°. No significant differences were detected between Labrador Retrievers and non-Labrador retrievers with respect to mMPTA (power=0.5) or mMDTA (power=0.342). Labrador Retrievers were significantly younger than non-Labradors ( P =.003).
Conclusion— A method for measurement of the mechanical joint angles of the canine tibia in the frontal plane was established and reference ranges for a population of dogs and Labrador Retrievers with cranial cruciate ligament rupture are reported.
Clinical Relevance— The established method of measurement and references ranges can be used to aid in diagnosis, determining indications, and surgical planning for angular limb deformities of the tibia, especially when affected bilaterally. The methodology and reference values may also be used for postoperative critique.     

EVALUATION OF CANINE LUMBOSACRAL STENOSIS USING INTRAVENOUS CONTRAST-ENHANCED COMPUTED TOMOGRAPHY

The objective of this study was to evaluate intravenous contrast-enhanced computed tomography as a technique for predicting the within-level location(s) of compressive soft tissues in the canine lumbosacral spine. Pre-operative intravenous contrast-enhanced computed tomography of the L5-S3 vertebral levels was performed in 12 consecutive large breed dogs with lumbosacral stenosis. The images were evaluated for enhancement of soft tissues by two radiologists who were unaware of the surgical findings. For each within-level location (dorsal canal, ventral canal, right lateral recess, left lateral recess) enhancement was classified as present, absent or equivocal. The results were compared with the results of surgical exploration and histopathology of excised tissues. The positive predictive values of intravenous contrast-enhanced computed tomography for compressive soft tissues involving the dorsal canal, ventral canal and lateral recesses were 83%, 100%, and 81% respectively. Negative predictive values for compressive soft tissues involving these locations were 29%, 50%, and 40% respectively.     

Age-related thoracic radiographic changes in golden and labrador retriever muscular dystrophy

Golden retriever and Labrador retriever muscular dystrophy are inherited progressive degenerative myopathies that are used as models of Duchenne muscular dystrophy in man. Thoracic lesions were reported to be the most consistent radiographic finding in golden retriever dogs in a study where radiographs were performed at a single-time point. Muscular dystrophy worsens clinically over time and longitudinal studies in dogs are lacking. Thus our goal was to describe the thoracic abnormalities of golden retriever and Labrador retriever dogs, to determine the timing of first expression and their evolution with time. To this purpose, we retrospectively reviewed 390 monthly radiographic studies of 38 golden retrievers and six Labrador retrievers with muscular dystrophy. The same thoracic lesions were found in both golden and Labrador retrievers. They included, in decreasing frequency, flattened and/or scalloped diaphragmatic shape (43/44), pulmonary hyperinflation (34/44), hiatal hernia (34/44), cranial pectus excavatum (23/44), bronchopneumonia (22/44), and megaesophagus (14/44). The last three lesions were not reported in a previous radiographic study in golden retriever dogs. In all but two dogs the thoracic changes were detected between 4 and 10 months and were persistent or worsened over time. Clinically, muscular dystrophy should be included in the differential diagnosis of dogs with a combination of these thoracic radiographic findings.     

Computed tomographic evaluation of cervical vertebral canal and spinal cord morphometry in normal dogs

The height, width, and cross-sectional area of the vertebral canal and spinal cord along with the area ratio of spinal cord to vertebral canal in the cervical vertebra were evaluated in images obtained using computed tomography (CT). Measurements were taken at the cranial, middle, and caudal point of each cervical vertebra in eight clinically normal small breed dogs (two shih tzu, two miniature schnauzers, and four mixed breed), 10 beagles, and four German shepherds. CT myelography facilitated the delineation of the epidural space, subarachnoid space, and spinal cord except at the caudal portion of the 7th cervical vertebra. The spinal cord had a tendency to have a clear ventral border in the middle portion of the vertebral canal and lateral borders near both end plates. The height, width, and area of the vertebral canal and spinal cord in the cervical vertebra were increased as the size of dog increased. However, the ratio of the spinal cord area to vertebral canal area in the small dogs was higher than that of the larger dogs. Results of the present study could provide basic and quantitative information for CT evaluation of pathologic lesions in the cervical vertebra and spinal cord.     

Diffuse idiopathic skeletal hyperostosis of the spine in a nine‐year‐old cat

A nine‐year‐old intact female domestic shorthair cat was evaluated for paraparesis, ataxia and severe spinal hyperaesthesia. Neurological examination indicated a T3‐L3 spinal cord segment lesion. Computed tomography of the thoracolumbar and lumbosacral vertebral column was performed. This showed contiguous smooth new bone formation ventral and lateral to the vertebrae extending from the cranial thoracic area to the lumbosacral junction and appearing similar to canine diffuse idiopathic skeletal hyperostosis. There was also marked dorsolateral stenosis of the vertebral canal at the level of T4‐T5 because of degenerative changes of the facet joints. To the authors’ knowledge, this is the first published report of feline diffuse idiopathic skeletal hyperostosis.     

QUANTITATIVE SURVEY RADIOGRAPHIC EVALUATION OF THE LUMBOSACRAL SPINE OF NORMAL DOGS AND DOGS WITH DEGENERATIVE LUMBOSACRAL STENOSIS

Survey radiographic studies of the lumbosacral region for 93 normal dogs and for 26 dogs with confirmed degenerative lumbosacral stenosis were reviewed. Normal dogs were divided into 9 groups based on age and body weight. For normal dogs, increasing age and body weight were associated with a decreased ability to extend the lumbosacral joint and with increased incidence and severity of spondylosis. Transitional lumbosacral vertebrae and evidence of lumbosacral disc space collapse were very infrequent findings, and the pivot point for lumbosacral motion was consistently centered over the lumbosacral disc space. Relative to an age/weight matched sub-population of normal dogs, dogs with degenerative lumbosacral stenosis had similar mean normalized lumbosacral vertebral canal height, larger mean neutral lumbosacral angle, decreased extension of the lumbosacral joint, increased flexion of the lumbosacral joint, reduced lumbosacral range of motion, increased lumbosacral dynamic malalignment, higher incidence and severity of spondylosis, higher incidence of transitional vertebrae, and higher incidence of lumbosacral disc space collapse. A logistic model based strictly on radiographic parameters was able to discriminate normal from affected dogs with an overall accuracy rate of 86%.     

COMPUTED TOMOGRAPHY OF THE NORMAL CANINE LUMBOSACRAL SPINE: A MORPHOLOGIC PERSPECTIVE

The lumbosacral spine of six normal dogs weighing 4.5 to 24.5kg was imaged by computed tomography in 5.0 mm & 10.0 mm transverse planes. The vertebral canal and thecal sac (including emerging nerve roots not distinguished as separate structures from the spinal cord) were measured along dorsoventral and transverse dimensions at cranial, middle and caudal levels within each vertebra from transverse tomographic images. Linear measurements were standardized to the dorsoventral dimension of the L₆ vertebral midbody to permit comparison and averaging of the vertebral and thecal sac dimensions among different sized dogs. The dorsoventral and transverse vertebral canal size progressively increased from cranial to caudal within each vertebra from L₁?L₆ (p ≤ 0.05). The transverse dimension of the thecal sac image increased caudally within each vertebra from L₁?L₄ (p ≤ 0.05). The vertebral canal dorsoventral and transverse dimensions were largest in the midlumbar area (p ≤ 0.05). The transverse, but not the dorsoventral, imaged dimension of the thecal sac peaked in the L₄ vertebra (p ≤ 0.05). The dorsoventral thecal sac image was observed to fill the vertebral canal in the cranial and middle vertebral levels in vertebrae L₁ through L₅ in over 60% of these normal dogs. However, epidural fat could almost always be seen lateral to the thecal sac regardless of what lumbar vertebra or vertebral level was imaged. Cranial to the lumbosacral junction, the dorsal intervertebral disk margin was almost always concave relative to the thecal sac. However, at the L₇-S₁ junction, some dogs had flat or even slightly convex dorsal intervertebral disk margins. The dorsal and ventral longitudinal ligaments and the ligamentum flavum could not be identified as distinct structures on the 5.0 mm transverse tomographic images.     

Association of cauda equina compression on magnetic resonance images and clinical signs in dogs with degenerative lumbosacral stenosis

Magnetic resonance imaging (MRI) was used to examine the lumbosacral spine of 27 dogs with degenerative lumbosacral stenosis. Four normal dogs were also similarly imaged. Compression of the soft-tissue structures within the vertebral canal at the lumbosacral space was assessed in two ways: by measuring dorsoventral diameter on T1-weighted sagittal images and cross-sectional area on transverse images. The severity of the clinical signs was compared to the severity of cauda equina compression. No significant correlation was found. It is concluded that degree of compression as determined by MRI at time of presentation is independent of disease severity.     

Ultrasonographic Adrenal Gland Measurements in Healthy Yorkshire Terriers and Labrador Retrievers

An upper threshold of 7.4 mm for maximal adrenal gland diameter is commonly used to detect pituitary‐dependent hyperadrenocorticism ultrasonographically in dogs. There is a substantial overlap between adrenal gland diameter of healthy dogs and of those with pituitary‐dependent hyperadrenocorticism. The aim of this study is to determine the measurements of both adrenal glands, in particular, of the height at the caudal glandular pole in a longitudinal plane, in the Labrador retriever and Yorkshire terrier, two breeds widely represented in the population suspected of hyperadrenocorticism. Seventeen Labrador retrievers and 24 Yorkshire terriers considered healthy were included in the study. Adrenal gland measurements were taken on static images and comprised in measurements of the length in a longitudinal plane (L), of the height at the cranial (CrHLG) and caudal pole (CdHLG) in a longitudinal plane and in a transverse plane (CrHTR and CdHTR, respectively), and of the width at the cranial and caudal poles in a transverse plane (CrWTR and CdWTR, respectively). This study established new upper thresholds for the left and right height at the caudal pole measured in a longitudinal plane: 7.9 mm (left) and 9.5 mm (right) for the Labrador retrievers and 5.4 mm (left) and 6.7 mm (right) for the Yorkshire terriers. All the measurements were significantly different between the two breeds. There was a significant relationship between CdHTR and CdHLG, and the age of the dogs for both breeds.     

Comparison of tibial plateau angles in normal and cranial cruciate deficient stifles of Labrador retrievers

OBJECTIVE: To investigate tibial plateau angles (TPA) in normal and cranial cruciate ligament (CCL) deficient stifles of Labrador retrievers. STUDY DESIGN: Prospective clinical study. ANIMALS: Eighty-one client-owned purebred Labrador retrievers. METHODS: Lateral radiographs of the tibia were obtained from 2 groups of dogs. Group I (42 dogs) had CCL rupture diagnosed by arthrotomy or arthroscopy. Group II (39 dogs) had no history of orthopedic problems, no radiographic evidence of CCL rupture, and dogs were >8 years of age. The tibial axis and the tibial plateau were determined on the radiographs, and the TPA was measured using image measurement software. The TPA measurement results of groups I and II were compared. RESULTS: Group I (CCL rupture) had a mean TPA (+/-SD) of 23.5 (+/-3.1) degrees, and group II (normal) had a mean TPA (+/-SD) of 23.6 (+/-3.5) degrees. With a P value of.97, no statistical difference was detected between the 2 groups. CONCLUSIONS: No correlation between the magnitude of TPA and CCL rupture was identified in this group of Labrador retrievers. CLINICAL RELEVANCE: In Labrador retrievers, TPA should not be used as a predictor of CCL rupture.     

Evaluation of the internal vertebral venous plexus, vertebral canal, dural sac, and vertebral body via nonselective computed tomographic venography in the cervical vertebral column in healthy dogs

OBJECTIVE: To evaluate nonselective computed tomographic (CT) venography for evaluating the cervical internal vertebral venous plexus (IVVP), define the diameter and area dimensions of the IVVP, and determine the relationship between dimensions of the cervical IVVP and other vertebral components in medium-sized dogs. Animals-6 healthy dogs that weighed 18 to 27 kg. Procedure-Helical CT scans were performed from C1 to C7 before and after IV injection of contrast medium (480 mg of iodine/kg) and a continuous infusion (240 mg of iodine/kg). Image data were transferred to a CT workstation, and measurements were performed on displayed transverse images. Diameter and area measurements of the vertebral canal, dural sac, IVVP, and vertebral body were obtained at C3 to C7. RESULTS: Opacification of vertebral venous structures was achieved in all dogs with no adverse reactions. Sagittal diameters of the IVVP for C3 to C7 ranged from 0.6 to 3.2 mm. Transverse diameters ranged from 2.32 to 5.74 mm. The IVVP area represented 12.4% of the mean vertebral canal transverse area and 30.61% of the mean vertebral epidural space area. Area measurements of the IVVP were significantly correlated with vertebral canal area and dural sac area. CONCLUSIONS AND CLINICAL RELEVANCE: Results indicated that nonselective CT venography is a safe, sensitive method for performing morphometric assessments of the cervical IVVP in dogs. Findings support the theory that there may be a physiologic or developmental relationship between cervical vertebral canal components.     

Biomechanical flexion-extension forces in normal canine lumbosacral cadaver specimens before and after dorsal laminectomy-discectomy and pedicle screw-rod fixation

Objective— To determine biomechanical flexion–extension forces in cadaveric canine lumbosacral specimens, before and after dorsal laminectomy with partial discectomy, and after dorsal pedicle screw–rod fixation of L7 and S1.
Study Design— Biomechanical cadaver study.
Animals— Cadaveric spine specimens without lumbosacral pathology from mature, intact Labrador retrievers (n=12).
Methods— Lumbosacral spine segments were subjected to a constant bending moment from L6 to S1 in a hydraulic 4-point bending materials testing machine. Force and displacement were recorded during each loading cycle constituting 1 complete flexion–extension cycle of the spine. Each spine segment had 3 series of recordings of 5 loading cycles each: (1) intact spine, (2) after surgical destabilization by dorsal laminectomy and partial discectomy, and (3) after surgical stabilization using dorsal pedicle screw–rod fixation.
Results— After dorsal laminectomy and partial discectomy, the neutral zone and range of motion were not different from those in the native spine specimen. After pedicle screw–rod fixation, the neutral zone and range of motion of the instrumented specimen significantly ( P <.0001) decreased compared with the native specimen and the specimen after dorsal laminectomy.
Conclusion— Dorsal laminectomy and partial discectomy does not lead to significant spinal instability in flexion and extension whereas pedicle screw and rod fixation effectively stabilizes the lumbosacral spine.
Clinical Relevance— Dorsal laminectomy and partial discectomy does not lead to significant spinal instability. Pedicle screw–rod fixation of L7 and S1 may be used to stabilize an unstable L7–S1 junction in dogs with degenerative lumbosacral stenosis.     

Investigation of breeding strategies to increase the probability that German shepherd dog and Labrador retriever dog guides would attain optimum size

An optimum-sized dog guide weighs 18 to 32 kg and measures 53 to 64 cm in height at the withers when mature body size is attained. Effects of selection index with and without restrictions, independent trait selection, directional selection, stabilizing selection, and negative assortative mating were modeled using data from German shepherd dogs and Labrador retrievers raised by the Seeing Eye, Inc., Morristown, NJ from 1979 to 1997. The selection goals were to decrease mature weight and mature height in German shepherd dogs and to decrease mature weight and increase mature height in Labrador retrievers. Mature weights were recorded for 1,333 German shepherd dog offspring and their 69 dams and 17 sires, and 1,081 Labrador retriever offspring and their 51 dams and 13 sires. Mature heights also were recorded for offspring and parents, including 871 German shepherd dogs from 70 dams and 15 sires, and 793 Labrador retrievers from 40 dams and 13 sires. Selecting on mature weight alone produced the highest aggregate genetic-economic gain for German shepherd dogs compared with the selection indices with and without restrictions, generating a 2.10-kg decrease in mature weight and a correlated 0.36-cm decrease in mature height. In Labrador retrievers, selecting for mature height alone produced the highest aggregate genetic-economic gain but caused an increase in mature weight. Weighting the two traits equally but in the opposite direction without restrictions was the only index that produced the desired effect of decreasing mature weight and increasing mature height in Labrador retrievers. Response to selection for one generation of directional selection for a single trait included a 0.50-kg decrease in mature weight for German shepherd dogs, a 0.59-kg decrease in mature weight for Labrador retrievers, a 0.18-cm decrease in mature height for German shepherd dogs, and a 0.91-cm increase in mature height for Labrador retrievers. Increasing the percentage of dogs attaining optimum size may decrease the cost of production for the Seeing Eye, Inc., because fewer dogs would need to be raised and trained to provide assistance to the same number of blind individuals.