Use and Limitations of Ultrasonography in Sacroiliac Disease

Ultrasound is currently the most sensitive and non-invasive imaging modality available to assess the sacroiliac ligaments and their bony attachments in horses. The examination is relatively easy to perform, and standard ultrasonographic equipment can be used. To differentiate between sacroiliac joint (SIJ) disease and sacroiliac ligament desmitis, the ultrasonographic findings must be interpreted in light of thorough physical and lameness examinations, sacroiliac stress tests, diagnostic sacroiliac analgesia, and nuclear scintigraphy.

1. Introduction

Sacroiliac disease has been recognized as a cause of poor performance and lameness in horses. In the past, the diagnosis has mainly relied on exclusion of other causes of decreased performance and hindlimb lameness [1,2]. Sacroiliac disease can be categorized into sacroiliac joint (SIJ) disease and sacroiliac ligament desmitis [3]. As in lower limb lameness, differentiation between SIJ disease and sacroiliac ligament desmitis is crucial so that the veterinarian can tailor the treatment plan to the specific condition [3]. Regional analgesia and diagnostic imaging tools such as nuclear scintigraphy, radiography, and ultrasonography are currently employed in an attempt to directly diagnose SIJ disease and sacroiliac ligament desmitis [3-7]. Scintigraphy helps to identify SIJ disease and possible enthesopathies associated with the sacroiliac ligaments, but it can produce equivocal results [2,7-9]. Radiography requires dorsal recumbency under general anesthesia, and evaluation of the SIJ is often complicated because of the superimposition of abdominal viscera [10]. Ultrasonographic evaluation of all accessible sacroiliac ligaments and of the ventral aspect of the SIJ can yield valuable information [11-14]. Unfortunately, little information exists about the normal ultrasonographic appearance of the equine sacroiliac region [12].

This manuscript reviews the current literature on sacroiliac ultrasonography and describes ultrasonographic findings in normal horses without the presence of sacroiliac disease.

2. Materials and Methods

Brief Anatomic Review

The sacroiliac anatomy is extensively discussed in other studies [15]. The sacrum is attached to the pelvis by three sacroiliac ligaments [3,16]. The dorsal sacroiliac ligament (DSIL) consists of a dorsal portion (DPDSIL) and a lateral portion (LPDSIL). The strong, cord-like DPDSIL originates at the dorsal aspect of each tuber sacrale (TS) and attaches abaxially at the dorsal apices of the sacral dorsal spinous processes. The LPDSIL forms a triangular sheet that originates from the TS and the caudal edge of the iliac wing and inserts at the lateral sacral crest, where it is continuous with the broad sacrosciatic ligament. The robust interosseous ligament (IL) connects the ventral iliac wing to the dorsal wing of the sacrum. The ventral sacroiliac ligament (VSIL) is thin and blends with the ventral SIJ capsule. It courses between the ventral aspects of the sacral and ilial wings.

Accuracy of Sacroiliac Ultrasonography

Tomlinson et al. [17] performed ultrasonographic measurements of the pelvis and the associated structures in ponies without sacroiliac disease. Those measurements were validated with computed tomography, magnetic resonance imaging (MRI), and frozen transverse section analysis. Among the included measurements were the cross-sectional area of the DPDSIL at S2, the width of the SIJ per rectum, and the width of the VSIL (VSIL and SIJ capsule). Structures measured on each side of the pelvis were not significantly different. The same pelvic structures were measured ultrasonographically in six clinically normal horses. No correlation existed between body weight and the size of the pelvic structures within each study population [17]. Ultrasonography was found to be accurate for measuring and evaluating musculoskeletal structures of the pelvis in both ponies and horses.

Ultrasonographic Equipment and Technique

The horse should be restrained in stocks and sedated with detomidine hydrochloride (0.01 mg/kg, IV). Using a no. 40 clipper blade, the skin overlying the sacroiliac region is clipped from 10 cm cranial to 35 cm caudal to the TS and 10 cm lateral to both TSs [11,13]. The exposed skin is then cleaned. Ultrasound coupling gel is applied to minimize air between the transducer and the skin. The horse must stand squarely on both hindlimbs to avoid uneven tension on the sacroiliac ligaments and to prevent artifact formation [a,13]. Ultrasonography of the sacroiliac region can be performed with either a portable or non-portable machine [11,12,17]. The highest quality images are achieved by using the highest frequency transducer that penetrates to the area of interest [12,13,17]. A 5- to 7.5-MHz sector or curved- array or linear-array probe is usually used to gain an overview of the sacroiliac region [ a,12]. High-frequency 7.5- to 12-MHz linear-array and/or curved-array transducers are recommended to image the DPDSIL [a,11,12,17]. A stand-off pad can be employed to improve the image quality of superficial structures [11,13]. To image the LPDSIL and the lateral aspect of the sacrum, which are located at a greater depth from the skin, 3.5- to 5-MHz sector probes are required [ a,17].

As a general guideline for a thorough and complete ultrasonographic examination of every anatomic structure of interest, examination in transverse (short-axis) and longitudinal (long-axis) planes should be performed [13,17].

Lumbosacral Junction

To begin the ultrasonographic evaluation, an overview of the lumbosacral junction and the cranial sacroiliac region is obtained in a transverse section with a 5- to 7.5-MHz sector or curved-array probe [12,13]. The transducer is placed on the midline over the caudal lumbar spine and moved caudally to the TSs. Immediately, the supraspinous ligament is seen subcutaneously as a well-defined square or triangular echogenic structure on the midline. It attaches to the apices of the caudal lumbar dorsal spinous processes. The most dorsal aspect of the spinous processes can be identified as a narrow hyperechoic line ventral to the supraspinous ligament [11,17]. The thin hyperechoic left and right thoracolumbar fascia (TLF) joins the supraspinous ligament on midline at a depth of 1 - 2 cm below the skin (Fig. 1) [11].

Figure 1. Normal configuration of the right dorsal portion of the dorsal sacroiliac ligament (DPDSIL). These sequential transverse images (cranial, top left; caudal, bottom right) were obtained from a 4-yr-old Thoroughbred without sacroiliac disease. A scan was obtained every 2 cm starting from 6 cm cranial of the most dorsal point of the tuber sacrale (TS) to the most dorsal aspect of the TS. The thoracolumbar fascia (TLF) joins the supraspinous ligament (SSpL) on midline and inserts on the TS. The thin part (TP) of the DPDSIL originates at the cranial portion of the TS, and the broad part (BP) of the DPDSIL is a continuation of the TLF. At the most dorsal aspect of the TS (bottom right), the BP is located medial to the TP, displaying the most frequently encountered configuration of the DPDSIL. Note that the separation between the BP and the TP is clearly visible.

Tubera Sacrale (TS)

As the evaluation proceeds caudally, the TSs appear as two thin convex hyperechoic lines 2 - 3 cm paramedian to midline (Fig. 1) [12,17]. Both TSs should be evaluated for left/right symmetry. It is not uncommon for one TS to be located more dorsal or cranial compared with the contralateral TS. Slight irregularities of the dorsal surface of the TS have been found in normal horses [17].

A few reports mentioned abnormalities, such as bone surface irregularities, hypoechoic lesions within the insertion, modification of the fiber orientation, and avulsion or chip fractures in horses with sacroiliac pain, at the origin of the DPDSIL at the TS [11-14].

Dorsal Portion of the Dorsal Sacroiliac Ligament (DPDSIL)

A 7.5- or 10-MHz linear-array transducer is placed dorsal to each TS to evaluate the DPDSIL from its origin to its insertion at the apices of the sacral dorsal spinous processes. This ligament is homogenously echogenic with a linear fiber pattern of craniocaudal orientation [11]. There are two parts to the ligament: a thin fibrous band (TP) originating on the dorsal aspect on the TS, and a broader and thicker part (BP) that is a caudal continuation of the TLF and is commonly located medial to the TS (Fig. 1 and Fig. 2) [a,12,17].

An artifactual hypoechoic appearance is created if the ultrasound beam is not perpendicular to the ligamentous fibers [13]. Because the DPDSIL is crescent shaped over the dorsal aspect of the TS (Fig. 1), the echogenicity over the entire width of the ligament cannot be evaluated with one transverse ultrasonographic image [a,13]. Therefore, it is imperative that echogenicity and linear fiber pattern of the thin and broad part (TP and BP) of the DPDSIL be evaluated separately by angling the transducer in a transverse plane on transverse and longitudinal scans [a,13]. The thickness of the DPDSIL reported by Denoix [11] is 6 - 10 mm. The total cross-sectional area of the DPDSIL measured at S2 is reported to be 0.89 - 1.59 cm2 [17].

Figure 2. Left oblique view of a necropsy specimen depicting the lumbosacral junction of a 5-yr-old Thoroughbred. The thoracolumbar fascia (TLF) inserts dorsally at the tuber sacrale (TS) but continues caudally to become the broad part (BP) of the dorsal portion of the dorsal sacroiliac ligament (DPDSIL). The thin part (TP) of the DPDSIL originates from the TS and consolidates with the BP caudally. Note the lateral portion of the dorsal sacroiliac ligament (LPDSIL).

Desmitis of the DPDSIL has been reported to be the most common soft-tissue lesion of the equine spine [14]. Diffuse and discrete heterogeneous echogenicity and hypoechoic areas with decreased linear fiber pattern in the DPDSIL have been described [11-14]. Tomlinson et al. [12] identified unilaterally decreased cross-sectional area measurements in horses with chronic sacroiliac disease as well as unilateral enlargement of the DPDSIL associated with hypoechoic areas in acutely affected animals. In one horse with acute DPDSIL desmitis, resolution of the aforementioned hypoechoic areas, decrease in ligament size, and development of TS height asymmetry were reported after 9 mo of rest [12]. No change in ultrasonographic appearance was seen in four horses with decreased DPDSIL size on a follow-up evaluation after completion of an exercise program and resolution of hindlimb lameness [12].

Interosseous Ligament (IL)

The IL cannot be imaged ultrasonographically because of its inaccessible location between the ilial and sacral wings, on the medial aspect of the SIJ.

Iliac Wing and Lateral Portion of the Dorsal Sacroiliac Ligament (LPDSIL)

The iliac wing and the LPDSIL are best imaged by positioning the 3- to 5-MHz sector probe lateral to the TS in a ventral direction. The wing of the ilium can be evaluated for incomplete or complete fractures, often located just dorsal or dorso-medial to the SIJ, and can involve the ilial articular surface [18]. For a more detailed examination, the iliac wing can be thoroughly imaged by following the bone from the TS to the tuber coxae [17,18]. The LPDSIL appears at the caudal edge of the ilial wing where it originates. On a transverse image with the transducer just paramedian to midline, the sacral dorsal spinous processes and the lateral aspect of the sacrum are seen as two hyperechoic lines jointly forming a J-shape (Fig. 3). The LPDSIL is seen as a thin echoic line oriented parallel to the sacral dorsal spinous processes and attaching at the lateral sacral crest [a,17]. The LPDSIL is examined for irregularities and abnormal thickness, whereas the sacral processes and the lateral sacral crest should be evaluated for any roughening or irregularities.

Figure 3. Transverse percutaneous sonogram of the lateral portion of the left dorsal sacroiliac ligament (LPDSIL) in a 10-yr-old Quarter horse free of sacroiliac disease. The sacral spinous processes and the fused transverse processes form a J shape. The LPDSIL attaches at the lateral sacral crest and has a thin and regular appearance.

Ventral Sacroiliac Ligament (VSIL) and Sacroiliac Joint (SIJ)

A transrectal ultrasonographic evaluation of the SIJ and VSIL is indicated if an abnormal response to transrectal palpation of the SIJ region is identified. A 5-, 7.5-, or 10-MHz sector or linear-array (endorectal) probe can be used [11,12,17]. The VSIL is an echogenic structure and covers the ventral SIJ aspect, blending with the ventral SIJ capsule [11,12,17]. Denoix [11] reported successful ultrasonographic examination of the VSIL per rectum. The ligament is often difficult to separate from the ventral SIJ capsule and surrounding fascia [12,17]. In anatomic specimens, no distinct separation can be appreciated between these three structures [a,17]. The VSIL (or the soft tissues at the ventral SIJ aspect) has been reported to be about 2 - 5 mm thick [11,17]. The ventral aspect of the SIJ consisting of the ventral surfaces of the sacral and ilial wings are seen as smooth hyperechoic lines, but the full width of the SIJ cannot be imaged [11,17].

VSIL injury and rupture associated with pelvic trauma and pelvic or sacral fractures have been observed on post-mortem examination but have not been documented ultrasonographically [11]. However, ventral periarticular remodeling of the SIJ has been reported [11]. In one study, ultrasonography of the VSIL per rectum in 20 horses with sacroiliac disease did not reveal any abnormalities [12].

3. Results

Sacroiliac ultrasonography was performed in nine normal horses at Cornell University College of Veterinary Medicine in 2001. The mean age was 9.4 ± 7 yr (mean ± SD; range, 4 - 24 yr), and mean weight was 521 ± 53 kg (range, 445 - 636 kg). Selection of all individuals was based on clinical evaluation, which consisted of a visual inspection of the croup for bony or muscular asymmetry, a gait evaluation to rule out hindlimb lameness, sacroiliac stress tests, and a transrectal palpation. After ultrasonography, all animals were euthanized, and detailed dissections of the pelvic specimens were performed. The necropsy findings were correlated with the pre-mortem ultrasonographic images.

Tubera Sacrale (TS)

Ultrasonographically, all evaluated horses had a mild to moderate irregular appearance of the caudal aspect of the TS. In addition, hyperechoic areas that cast acoustic shadows were occasionally seen at the caudal edge of the TS of younger individuals and were associated with irregular mineralization of the cartilaginous TS epiphyses. The presence of secondary apophyseal ossification centers was confirmed at necropsy. The mean distance from the dorsal aspect of the TS to the skin was 2.04 ± 0.44 cm (range, 1.42 - 3.33 cm) with minimal variation between the left and right side.

Dorsal Portion of the Dorsal Sacroiliac Ligament (DPDSIL)

The thin part (TP) of the DPDSIL is crescent shaped. It follows the contour of the dorsal aspect of the TS at its origin and inserts at the apices of the 3rd to 6th sacral dorsal spinous processes. The broad part (BP) of the DPDSIL is a continuation of the TLF that inserts with or without fold formation onto the dorsal aspect of the TP at the TS (Fig. 1 and Fig. 4). In most horses, the BP is located medial to the TS and runs caudally to insert on the second and third sacral dorsal spinous processes. The shape and location of the BP at the level of the TS is subject to variation depending on the spatial relationship of the TS and the lumbar and sacral dorsal spinous processes.

Figure 4. Normal configuration of the right dorsal portion of the dorsal sacroiliac ligament (DPDSIL). These sequential transverse images (cranial, top left; caudal, bottom right) were obtained from a 4-yr-old Thoroughbred without sacroiliac disease. A scan was obtained every 2 cm starting from 6 cm cranial of the most dorsal point of the tuber sacrale (TS) to the most dorsal aspect of the TS. The thoracolumbar fascia (TLF) joins the supraspinous ligament on midline and inserts on the TS. The thin part (TP) of the DPDSIL originates at the cranial aspect of the TS. The broad part (BP) of the DPDSIL is a continuation of the TLF. At the most dorsal aspect of the TS (bottom right), the BP is located dorsal to the TP, displaying the less frequently seen configuration of the DPDSIL. Note the fold formation and the barely visible separation between the BP and the TP.

The shape and location of the thin part of the DPDSIL does not seem to vary. Two different configurations of the two parts occurred in healthy horses: either the BP was located medial to the TP, which was the predominant configuration (seen in seven of nine horses; Fig. 1) or the BP was seen dorsal to the TP (seen in two of nine horses; Fig. 4). The separation between the two parts of the DPDSIL is not always clearly visible. Caudal to the TS, both parts of the DPDSIL gradually consolidate to form the cord-like, mostly oval or rectangular-shaped ligament that attaches at the abaxial aspect of the sacral dorsal spinous process apices. At the level of S2, the division between the two parts is sometimes still seen. The thickness and cross-sectional area of the DPDSIL was measured at a location halfway between the most dorsal point and the caudal edge of the TS. Measurement of the cross-sectional area is subjective because of difficulties in delineating all borders of the DPDSIL on one image. The mean thickness was 0.45 ± 0.29 cm (range, 0.14 - 1.07 cm), and the cross-sectional area was 1.82 ± 0.41 cm2 (range, 1.12 - 2.50 cm2).

Bilateral, small focal hypoechoic areas (type 1 of 4 echogenicity = slightly less echogenic than normal, minimal fiber pattern disruption [19]) were identified ultrasonographically in the DPDSIL in three of nine normal horses (4 - 6 yr of age). A bilateral general decrease in linear fiber pattern (fiber score 1 = 50 - 75% fiber alignment [20]) was seen at the dorsal spinous process of S2 in a 24-yr-old Quarter horse mare. Histology of DPDSIL sections of the 24-yr-old mare revealed locally extensive mild to moderate, multifocal cartilaginous metaplasia with multifocal, mild myofiber mineralization. This is compatible with age-related changes. Sections of ultrasonographically normal appearing DPDSIL from other horses were also submitted for histological analysis, and no significant histologic lesions were identified.

Lateral Portion of the Dorsal Sacroiliac Ligament (LPDSIL)

In a small number of horses, difficulties were encountered when imaging the LPDSIL because of the decreased skin penetration of the ultrasound waves. The sacral dorsal spinous processes and the lateral sacral crest were found to be smooth in all animals. The LPDSIL appeared as a regular and thin line, measuring <4 mm in thickness.

Sacral Dorsal Spinous Processes Apices

Ultrasonography of the sacral dorsal spinous processes apices is best performed with a 5- to 8-curved- array transducer with a small radius. The first sacral dorsal spinous process was only seen in four of nine horses because of its deep location and the echogenic tissues present between both TSs on the midline. Mild to moderate diffuse irregularities were found in eight of nine horses. The spinous process apices of the third to sixth sacral vertebrae were often roughened dorsally.

Ventral Sacroiliac Ligament (VSIL) and Sacroiliac Joint (SIJ)

Transrectal ultrasonographic evaluation of the VSIL and SIJ is limited to the ventral aspect of the joint and was not found to yield significant information because of the inability of the ultrasonographer to distinguish the VSIL from surrounding tissues.

4. Discussion

A complete and systematic approach to the ultrasonographic examination of the sacroiliac region combined with a detailed knowledge of the normal appearance of the sacroiliac ligaments will aid the clinician in detecting subtle lesions.

A mild to moderate roughened appearance to the caudal aspect of the TS should be considered normal in horses of any age [a,17]. Secondary apophyseal ossification centers that cast acoustic shadows at the caudal edge of the TS were seen in animals ≤5 yr of age and have not previously been described [a]. The aforementioned findings should not be confounded with avulsion fractures of the DPDSIL origin.

At the level of the TS, the TP and BP of the DPDSIL can assume two different configurations depending on the spatial relationship of the TS and the lumbar and sacral spinous processes. Previous descriptions of the normal DPDSIL did not mention the two possible configurations and simply described the ligament as crescent shaped or oval [11,12]. In the majority of horses evaluated in our study, the BP of the ligament was located medially to the TP. In a few animals, the BP was situated dorsally to the TP. The shapes of the left and right DPDSIL are identical if both TSs are level. All horses evaluated in our study had level TSs, and therefore, showed no shape incongruency between the left and right DPDSIL. Conversely, if the left TS is elevated relative to the right TS, the BP of the right DPDSIL is pulled into a more dorsal position. Differences in shapes between the left and right DPDSIL can develop where the BP of the right DPDSIL is displaced dorsally to the TP, whereas the BP of the left DPDSIL remains medial to the TP. This can lead to an increased thickness of the right DPDSIL compared with the left ligament. Dorsal height differences between the TSs is common in normal horses, and this is theorized to be secondary to chronic asymmetric muscular or ligamentous forces acting on the malleable bony pelvis [2,3]. Incongruent shapes between the left and right DPDSIL as the sole abnormal findings on ultrasonographic evaluation of this ligament is not necessarily indicative of desmitis [a]. As an example, shape differences were observed in several horses diagnosed with SIJ disease or ilial wing fracture. The DPDSIL shape and thickness differences between the left and right side in these cases were attributed to the unleveled TS [a].

A bilateral regional decrease in linear fiber pattern (up to fiber score 1 = 50 - 75% parallel fiber alignment [20]) or the presence of diffuse heterogenic appearance of the DPDSIL can be related to age in older horses and is not necessarily a sign of DPDSIL desmitis. Bilateral occurrence of one or two small focal hypoechoic areas (type 1 of 4 echogenicity = slightly less echogenic than normal, minimal fiber pattern disruption [19]) in the DPDSIL without general enlargement of the ligament may also be an incidental finding; this was seen in three of nine horses [a]. Therefore, the bilateral presence of a fiber score of 0 (=75 - 100% linear fiber alignment) and a mild diffuse heterogeneous echogenicity (type 1 lesion), seen occasionally in horses of all ages, may be normal. The previously published cross-sectional area of the DPDSIL was measured at S2 (0.89 - 1.59 cm2) and is smaller compared with our values of 1.12 - 2.50 cm2, which were measured halfway between the most dorsal point and the caudal edge of the TS [17]. We found that this specific location was best suited for concurrent evaluation of the shape, thickness, and cross-sectional area of the DPDSIL. The range of values for the thickness of the DPDSIL obtained in our study (0.14 - 1.07 cm) showed a broader distribution than those mentioned in the literature (6 - 10 mm) [a,11].

Mild to moderate diffuse bone surface irregularities at the sacral dorsal spinous process apices without decreased linear fiber pattern or hypoechoic lesions at the insertion site of the DPDSIL is commonly seen in normal horses.

On transrectal ultrasonographic evaluation, only the ventral aspect of the SIJ can be evaluated for periarticular remodeling [11,17]. Various degrees of degenerative SIJ changes, including periarticular modeling changes seen especially at the caudomedial border, have been reported in normal horses [1,21]. Ultrasonographic evidence of SIJ new bone formation without concurrent focal pain on transrectal palpation is, therefore, of questionable clinical relevance [11]. Ultrasonographic evaluation of the soft tissues located at the ventral aspect of the SIJ is also of uncertain clinical value because of the difficulties in visualizing and/or differentiating between the VSIL and SIJ capsule [12,17]. It is conceivable that moderate to severe trauma to the VSIL and SIJ capsule associated with fluid accumulation, edema formation, and generalized soft-tissue thickening may be identified on transrectal ultrasonography.

In conclusion, sacroiliac ultrasonography is relatively straight forward and can be performed in the field or hospital setting with standard ultrasound equipment and high-frequency transducers. This is currently the most sensitive imaging modality to assess the dorsal and lateral portion of the DSIL, the VSIL, and their bony attachments. The obtained information must be interpreted in light of thorough physical and lameness examinations, sacroiliac stress tests, periarticular SIJ analgesia, and nuclear scintigraphy in an attempt to differentiate between SIJ disease and sacroiliac ligament desmitis.

This study was funded by the Dean’s Fund for Clinical Excellence at the College of Veterinary Medicine (Cornell University, Ithaca, NY) and by the New York Thoroughbred Horsemen’s Association (Jamaica, NY). The authors thank Rebecca Lin for reviewing the manuscript.

Footnote

1. Engeli E and Yeager AE. Unpublished data, 2001-2002.