key: cord-0040701-7hqk3h1o
authors: Barger, Anne M.
title: Musculoskeletal System
date: 2009-10-30
journal: Canine and Feline Cytology
DOI: 10.1016/b978-141604985-2.50018-9
sha: 8beee2960c71f4a218204c47d16f366c1ba9ae20
doc_id: 40701
cord_uid: 7hqk3h1o

nan

Lameness is the cardinal clinical sign associated with disease of the musculoskeletal system. Other signs include stiffness, ataxia, weakness, pain, fever, limb and joint swelling, and deformity. Depending on the type of disorder, other organ systems may also be involved, including neurologic, endocrine, urologic, hemolymphatic, digestive, respiratory, and cardiovascular systems. Because of this, an animal with musculoskeletal disease may present with a variety of problems and signs.

Cytology may be a component of the workup in an animal with a suspected musculoskeletal disorder. Material that may be sampled include synovial fl uid as well as fi ne-needle aspirates of soft tissue masses involving muscle or proliferative/lytic lesions of the bone. Cytologic evaluation alone is rarely the sole diagnostic test necessary to completely defi ne a musculoskeletal problem. Other important information includes signalment, history, physical examination, radiographs, complete blood count, and biochemistry. In addition, many lesions will require histopathology for defi nitive characterization. Some types of muscle, bone, and joint disease cause changes that cannot be detected by cytologic methods.

Synovial fl uid analysis is part of the minimum database when assessing an animal for joint disease. It is important to recognize that evaluation of the synovial fl uid is only a component of the workup of an animal with suspected joint disease. The data obtained from joint fl uid analysis must be integrated with other clinical and laboratory fi ndings, including appropriate ancillary diagnostic tests (e.g., culture, serology, antinuclear antibody [ANA] titer, rheumatoid factor [RF] titer). Nevertheless, when an animal has suspected joint disease, fl uid evaluation is a critical component in determining the cause of disease.

As with other body cavity effusions, a complete fl uid analysis is helpful when evaluating a synovial effusion. Routine synovial fl uid analysis should include evaluation of color, transparency, protein concentration, viscosity, mucin clot test, nucleated cell count, differential, and cytologic evaluation. These tests are discussed in further detail below. If the sample is limited, the most important component of analysis is the cytology. Typical results for different kinds of joint disease are shown in Table 13 -1.

Collection of synovial fl uid varies to some degree depending on the joint sampled. Descriptions of approaches to various joints have been described. In general, collection of synovial fl uid requires the following materials: 3-to 6-mL syringe, 18-to 22-gauge 1-inch needles, and red-top and/or lavender-top tubes. The amount of restraint and necessary levels of sedation and anesthesia will vary from animal to animal. Enough restraint should be used to minimize struggling during collection. In general, many animals will require at least some degree of sedation or anesthesia. Sterile technique is critical when preparing the site and during aspiration. The fur should be clipped and the area of aspiration scrubbed. Care should be taken not to scratch the articular surface during needle insertion. Palpation and slight fl exion or hyperextension of the joint will help to identify insertion points of the needle. Aspiration sites for specifi c joints are described below. The location of aspiration varies with the joint aspirated. The coxofemoral joint can be aspirated cranioproximal to the trochanter major and slightly ventral and caudal. The stifl e should be fl exed when aspirated. Aspiration can occur medial or lateral to the patellar ligament, midway between the tibia and femur. The tarsocrural joint can be aspirated by hyperextending the joint and inserting the needle lateral or medial to the fi bular tarsal bone. To aspirate the shoulder, insert the needle 1 cm distal and slightly caudal to the acromion process. The elbow should be hyperextended and the needle inserted lateral and along side the olecranon. The carpal joints can be simply aspirated by fl exing the joint and palpating the joint space. The needle should be advanced slowly through the joint capsule into the joint cavity. The amount of fl uid withdrawn depends on the size of animal and joint as well as the amount of C H A P T E R 13

Anne M. Barger 310 CHAPTER 13 • Musculoskeletal System effusion present. Synovial fl uid will be aspirated easily if there is a signifi cant effusion but a few drops may be obtained from joints without an increase in synovial fl uid volume. Before removing the needle from the synovial cavity, the plunger of the syringe should be released to remove any negative pressure. Normal synovial fl uid has a gel-like consistency that should not be mistaken for a clot. The gel-like consistency will become less viscous when shaken and return to the original viscosity upon standing; this property is referred to as thixotropy. Clotting is likely to occur if there is signifi cant blood contamination and infl amed joints may form fi brin precipitates or clots. For these reasons, some joint fl uid should be put into an EDTA tube (lavender-top tube). The EDTA will interfere with tests such as the mucin clot test and culturing. The synovial fl uid should be refrigerated if not immediately evaluated. For samples that may be cultured dependent on the cytologic fi ndings, the fl uid should be put into a red-top tube, left in the sterile syringe, and/or placed in an aerobic culturette. There are advocates of putting fl uid in blood culture media to improve the chances for bacterial growth. The laboratory should be contacted for their recommendations. In many smaller animals, only one or two drops of joint fl uid can be obtained. In these cases, immediate preparation of direct smears is the critical component of sample management (refer to Chapter 1). Regardless of the amount of fl uid collected, it usually is advantageous to make direct smears immediately to best preserve cell morphology. These slides should not be refrigerated before staining.

Normal joint fl uid is typically present in small amounts (Ͻ0.5 mL) and is clear to straw-colored ( Fig. 13-1A ). Redtinged fl uid indicates hemorrhage or peripheral blood contamination. True hemorrhage will be uniformly discolored throughout aspiration, whereas peripheral blood contamination often occurs at the end of aspiration. This may appear as a red tail or wisp in the fl uid. The fl uid should be viscous as evident by stringiness when suspended between fi ngertips, touched by an applicator stick, or expelled from the syringe (see Fig. 13 -1A). The fl uid viscosity is related to the concentration and quality of hyaluronic acid. Normal synovial fl uid has good viscosity and demonstrates thixotropy (see above).

Healthy synovial fl uid should be viscous due to production of mucin. The mucin clot test is done to semiquantitatively assess the amount and/or degree of polym e rization of hyaluronic acid in the joint fl uid. Since EDTA interferes with this test, heparin can be used if an anticoagulant is required before performing this test. One to two drops of undiluted joint fl uid are added to four to eight drops of 2% acetic acid. In a sample with normal hyaluronic acid concentration and quality, a thick, ropy clot will form ( Fig. 13-1B) . As the amount and/or quality of hyaluronic acid decreases in various forms of joint disease, the mucin clot is less well formed. This test is typically interpreted as good, fair, or poor. Normal joints have good mucin clot results.

The direct smear of the synovial fl uid should also be evaluated for presence of windrowing. In a viscous sample, the cells will often line up in rows or windrow ( Fig. 13-2A) . Mucinous material can be identifi ed in the background of the direct smears as eosinophilic granular material ( Fig. 13-2B&C ) or sometimes as proteinaceous crescents.

Cell counts and the differential count are done by routine methods. If enough fl uid is present, cell counts can be made using a hemocytometer. Some reference laboratories use automated cell counters for cell enumeration. Automated cell counters tend to give a higher cell count than the hemocytometer; however, the difference is not usually great enough to affect the clinical interpretation. The cells may occur in clumps and accurate assessment of cell numbers may be diffi cult. In an effort to minimize cell clumping, hyaluronidase can be added to the synovial fl uid. Various methods have been described. The easiest procedure is to add a small amount of hyaluronidase A B 20m C ■ FIGURE 13-2. A, Windrowing. The cells in this fi gure are found in rows. This is referred to as "windrowing" and is commonly seen in fl uids with increased viscosity or increased protein concentration. Infl amed joints may have decreased viscosity by gross visual examination, but prominent windrowing of cells microscopically. (Wright; HP oil.) B, Granular background. Normal synovial fl uid has low nucleated cell numbers with a thick, granular to ropy background material separating the cells. The granular background is related to the mucin content of the fl uid. The low cellularity generally means that fewer than 1 to 2 small to medium-sized mononuclear cells are seen on high-power examination (arrow 

powder (amount adherent to an applicator stick) directly into the sample tube, which may result in more accurate cell counts. If only slides are prepared, cell numbers can be roughly estimated by counting the number of cells per low-power fi eld (10ϫ) and multiplying the count by 100 to give an approximate number per μl. However estimates from smears are less accurate and tend to be higher than counts from automated counters. Normal joints will have low nucleated cell numbers, usually fewer than 3000 cells/μl in the dog and 1000 cells/μl in the cat (Pacchiana et al., 2004) , although more typically the count is fewer than 500 cells/μl in both species. These counts may vary slightly based on breed, age, body weight, and joint sampled. Consequently, only 1 to 2 cells per high-power fi eld (40ϫ) will be observed depending on the thickness of the direct smear (see Fig. 13 -2B). Gibson et al. (1999) demonstrated the variability in performing these estimates by a group of clinicians on synovial fl uid. Cells commonly observed in synovial fl uid include lymphocytes, macrophages (clasmatocytes), neutrophils, and, occasionally, synovial lining cells that produce glycosaminoglycans. Neutrophils typically account for less than 5% to 10% of nucleated cells in normal joints. If fl uid is obtained, both direct smears and concentrated preparations can be evaluated. If available, a cytocentrifuge is useful in preparing concentrated preparations. Concentrated preparations can also be prepared by centrifuging the fl uid, pouring of the supernatant, and resuspending the fl uid in one or two drops of supernatant. Smears can then be prepared from this concentrated preparation. Concentrated preparations are useful in synovial fl uid particularly if the cell count is low (Ͻ500 cells/μl).

Protein concentration is often measured by refractometry, which usually provides a value that is useful for routine clinical classifi cation and interpretation of the synovial fl uid. The most accurate measurement of protein requires chemical methods. Normal synovial fl uid generally has a low protein concentration (Ͻ2.5 g/dl) or commonly between 1.5 to 3.0 g/dl (MacWilliams and Friedrichs, 2003) . Protein concentration will increase with infl ammatory disease. False increases in protein can occur with EDTA, especially if a short sample is submitted or if the patient has received an intra-articular injection.

The primary goal in synovial fl uid evaluation is to distinguish infl ammatory joint disease from degenerative joint disease (see Table 13 -1). Other types of joint disease that may be distinguished include hemarthrosis and neoplastic disease. Further defi ning the disease process, as noted above, requires integrating the synovial fl uid fi ndings with other historical, physical, and laboratory fi ndings including imaging techniques. It is important to note that synovial fl uid analysis alone rarely differentiates or identifi es the specifi c cause from among the multiple etiologic factors involved in infl ammatory and noninfl ammatory joint diseases.

Infl ammatory joint disease is characterized by fi nding increased numbers of white blood cells, particularly neutrophils ( Fig. 13 -3), in the joint fl uid. Absolute numbers of segmented neutrophils are often moderately to markedly increased. However, the infl ammatory process appears to cytologically wax and wane with time and, if polyarticular, involve other joints with varying intensity. Consequently, repeating joint sampling and, more importantly, sampling multiple joints, even if not clinically affected, has diagnostic value. A key point is that infl ammatory joint disease has both infectious and noninfectious causes.

Some cases of joint disease are caused by bacterial ( Fig. 13 -4) or fungal infection . In general, septic joints have very high cell counts. In most cases, the cells are primarily segmented neutrophils. It is important to evaluate the condition of the neutrophils. Degenerative or karyolytic neutrophils are more commonly observed with septic joints. Degenerate neutrophils have a pale, swollen nucleus with some loss of nuclear segmentation. However, often the majority of the neutrophils will appear nondegenerative in septic arthritis. In one study, Staphylococcus sp. was the most common bacterial agent isolated in septic joints (Marchevsky and Read, 1999) . Organisms may gain access to joints either hematogenously or via direct inoculation. In addition, there may be infection elsewhere in the body (e.g., endocarditis) with immune complex deposition in the synovial tissue and resultant nonseptic infl ammation in the fl uid. Bacterial and fungal arthritis most commonly present with solitary joint involvement but on occasion may have multiple joint involvement, especially in young animals. (Santos et al., 2006; Harvey and Raskin, 2004) . In one study of E. ewingii joint infection cases in which the diagnosis was confi rmed by polymerase chain reaction testing of peripheral blood, nucleated cell counts ranged from 16,000 to 125,000/μl with 63% to 95% neutrophils (Goodman et al., 2003) .

Many animals with infl amed joints have nonerosive disease . Causes of nonerosive polyarthritis include infl ammation secondary to infection or neoplasia elsewhere, breed-specifi c polyarthritis ■ FIGURE 13-4. Bacterial arthritis. Bacterial arthritis may be caused by direct inoculation or hematogenous spread. Infected joints typically have high neutrophil counts (Ͼ50,000/μl). In this example, the neutrophils display degenerative changes, including nuclear swelling (short arrows), and cytoplasmic vacuolization. The presence of degenerative changes strongly supports infection; however, the lack of degenerative changes or observable microorganisms does not rule out the possibility of infection. The bacteria may be located in the joint tissue and not present in the synovial fl uid. Rare bacteria were observed after prolonged searching (long arrow). (Wright; HP oil.).

■ FIGURE 13-5. Blastomycosis. In addition to bacteria, other types of infectious agents may also involve the joint. This photomicrograph contains numerous neutrophils that are "rounded up" and almost appear like mononuclear cells owing to the thickness of the smear. In the center of the photo, broad-based budding yeast are found that are consistent with Blastomyces dermatitidis (arrow). Fungal organisms may be present infrequently and are best found on low-power examination. . 13-7A&B ). As the name implies, polyarthritis typically affects multiple joints, but on occasion may present with only a solitary affected joint. Joints affected by immune-mediated disease have increased numbers of nondegenerate neutrophils. In rare cases, increased numbers of lymphocytes and plasma cells may be found. Smaller distal joints are most commonly affected. Diagnosis of immune-mediated disease depends not only on demonstrating joint infl ammation, but also on ruling out infection via culture, serology, and/ or empiric therapy. Some cases of immune-mediated disease will have ragocytes ( Fig. 13 -8A-C) or (lupus erythematosus) LE cells ( Fig. 13-8D ). These are infrequent fi ndings and should not be relied on to make a diagnosis of immune-mediated disease.

Erosive arthritis is suggested when there are lucent cystlike areas in the subchondral bone with narrowing or widening of the joint spaces found on joint radiographs. Types of erosive arthritis described in animals include rheumatoid arthritis, polyarthritis of Greyhounds, and feline chronic progressive polyarthritis . The classic fi nding is progressive loss of subchondral bone with deformation and destruction of affected joints. Infection or neoplasia may also cause erosive joint disease. Erosive arthritis, as with other types of infl ammatory joint disease, is characterized by increased numbers of neutrophils in the synovial fl uid. Synovial fl uid analysis alone cannot distinguish erosive disease from nonerosive disease and, for this reason, radiographs should be done on animals with infl ammatory joint disease. Other clinical features of noninfectious erosive arthritis include morning stiffness, swelling of same or multiple joints within a 3-month period, symmetric swelling of joints, mononuclear infi ltrates observed microscopically in a synovial membrane biopsy, and positive RF titer.

Degenerative joint disease (osteoarthritis, osteoarthropathy) is characterized by degeneration of articular cartilage with secondary changes in associated joint structures. The disorder usually occurs secondary to conditions such as osteochondrosis, hip dysplasia, joint instability, chronic bicipital tenosynovitis, and trauma. Changes in the synovial fl uid are not as remarkable as seen with infl ammatory disease (Fig. 13-9A&B) . A mild increase in the number of mononuclear cells is the predominant fi nding (Stobie et al., 1995) . These cells are likely a mixture of macrophages, lymphocytes, and synovial lining cells ( Fig. 13-9C-E) . Occasionally osteoclasts can be observed which may suggest erosion of cartilage and exposure of underlying subchondral bone ( Fig. 13-10A&B ).

If recent trauma has occurred, joint hemorrhage may be appreciated (Fig. 13-11A&B ). True hemorrhage must be distinguished from the much more common artifact of blood contamination. This is best done at the time of sample collection. If previous hemorrhage has occurred, the withdrawn fl uid will appear xanthochromic (yellowish color due to old hemorrhage) to homogeneously red and cloudy. Besides trauma, other causes of hemorrhagic joint fl uid include coagulation defects and neoplasia. A congenital coagulation factor defi ciency should be considered in a puppy or kitten that presents with 

repeated episodes of hemarthrosis or with hemarthrosis and a history of minimal trauma. Low numbers of red blood cells can be observed in normal synovial fl uid but should not be present in high enough numbers to discolor the fl uid. Cytologically, hemorrhage can be distinguished from peripheral blood contamination by identifi cation of erythrophagia, hemosiderin-laden macrophages, and other red blood cell pigments such as hematoidin. Occasionally platelets will be observed in samples with severe peripheral blood contamination. Care must be taken not to overinterpret erythrophagia because this can occur ex vivo if the sample is not evaluated quickly.

Normal synovium consists loose connective tissue containing blood vessels, fi broblasts, adipocytes, and histiocytes lined by a layer of synoviocytes. The synoviocytes involves three cell populations: macrophages, antigenpresenting dendritic cells, and glycosaminoglycansproducing cells. In a study of 35 dogs, the most common neoplasm of canine synovium (51%) was of histiocytic origin (Craig et al., 2002) . Other neoplasms in this study were (17%) synovial myxomas, (14%) synovial sarcomas, and the remaining 17% were mixed sarcomas including malignant fi brous histiocytoma, fi brosarcoma, chondrosarcoma, and undifferentiated. Immunohistochemical They are thought to represent nuclear remnants or phagocytosed immune complexes. They should be distinguished from bacteria. Observations suggest that these cells are seen more commonly in association with immune-mediated polyarthropathies but are not considered diagnostic. Serologic evaluation for immune-mediated disease and extra-articular nonbacterial infections such as ehrlichiosis and borreliosis is recommended when polyarthritis is identifi ed. (Wright; HP oil) B, Neutrophils frequently contain several variably sized, dark, cytoplasmic granules in a case of immune-mediated polyarthropathy. Stifl e fl uid had WBC 7,400/μl, protein 3.6 g/dl, good mucin clot, 21% nondegenerate neutrophils, 45% small lymphocytes, 34% large mononuclear cells, positive ANA titer, and negative rickettsial titers. (Wright; HP oil.) C, Close-up of affected neutrophils containing fragments of nuclear material. (Wright; HP oil.) D, Lupus erythematosus cell. Dog. Synovial fl uid from a dog with shifting leg lameness. The total cell number is moderately increased and composed of predominantly nondegenerate neutrophils with lesser numbers of lymphocytes and monocytes. The neutrophil in the center contains a large, round, homogeneous eosinophilic inclusion in the cytoplasm that displaces the nucleus to the periphery of the cell membrane (arrow). This is an LE cell. The phagocytized material is thought to be nuclear material that has been structurally altered by antinuclear antibody. The homogeneous, light-staining appearance of the material distinguishes it from normal nuclear material. LE cells are rare, but when found, support the diagnosis of systemic lupus erythematosus. 

staining was necessary to distinguish between the histologic types of synovial tumors as prognosis varied greatly between them. A recommended panel of antibodies is suggested to include cytokeratin (AE1/AE3) for synovial cell sarcoma, CD18 for histiocytic sarcoma, and smooth muscle actin for malignant fi brous histiocytoma. Synovial cell sarcomas may appear most commonly as the spindle cell form (Fig. 13-12A ) or alternatively as a mixed spindle and epithelioid variant. Histiocytic sarcomas arise from the antigen-presenting dendritic cells of the synovium layer. These neoplasms are frequently associated with Rottweilers, Bernese Mountain Dogs, and retrievers (Affolter and Moore, 2002) . Cytologically, histiocytic sarcoma displays a round cell appearance with anaplastic characteristics including cellular pleomorphism, multinucleation, anisokaryosis, coarse chromatin, and prominent nucleoli with abundant basophilic cytoplasm ( Fig. 13-12B&C ). Another neoplasm occasionally encountered in the joint is a metastatic form of carcinoma. Cases with metastatic neoplastic cells in synovial fl uid have been documented arising from the lung and mammary gland (Meinkoth et al., 1997) .

Aspiration of muscle, connective tissue, or bone lesions is much the same as aspiration of other lesions. Generally there is a mass, evidence of lysis, or swelling that warrants A B ■ FIGURE 13-11. Hemarthrosis. Because of the small size of canine and feline joints, it is common to get some degree of blood contamination in most joint aspirates. To help distinguish true hemorrhage from blood contamination, the smears should be routinely examined for erythrophagia, hematoidin crystals, hemosiderin, and platelet clumps. In (A), the macrophage contains a small, golden hematoidin crystal (arrow), while the smaller of two macrophages in (B) contains a phagocytosed erythrocyte in the lower left area of its cytoplasm (arrow). These fi ndings indicate that there has been previous hemorrhage in the joint. Potential causes of hemarthrosis include trauma, coagulopathy, and neoplasia. Coagulopathies may have evidence of multiple joint involvement and bleeding elsewhere. Abnormal hemostasis is documented by coagulation testing. (Wright; HP oil.)

aspiration. Fine-needle aspiration, fenestration, or impression smears of tissue taken for biopsy are common methods of obtaining a sample. Components of the musculoskeletal system that will be reviewed in this section of the chapter include skeletal muscle and bone.

Normal cytology of skeletal muscle has a characteristic appearance. Usually a tissue fragment is aspirated and the cytoplasm of the cells stains deeply basophilic (see Figure 2 -1). Often, striations can be visualized by focusing up and down on the cell aggregate. The nuclei are round with a condensed chromatin pattern. Myositis is diffi cult to diagnose with cytology because it is difficult to associate the infl ammatory cells with the myocytes (Fig. 13-13) . Therefore, cytology is of limited use in diagnosing myositis. Diagnosis of myositis typically requires consideration of history, signalment, and chemistry fi ndings (increased creatine kinase and aspartate transaminase), as well as electromyographic (EMG), immunologic, and serologic tests. Histopathology is necessary as well for defi nitive characterization of infl ammatory and degenerative muscle lesions. Tumors arising from skeletal muscle include rhabdomyoma and rhabdomyosarcoma. These neoplasms are uncommon. Cytologically these tumors may appear similar to other mesenchymal tumors. Rhabdomyomas in particular often exfoliate poorly; however, rhabdomyosarcomas can be cellular enough to diagnose as a sarcoma. Usually these cells are round to spindle-shaped with abundant amounts of basophilic cytoplasm and oval nuclei. Occasionally, multinucleated cells are observed, with the nuclei arranged in a row consistent with a straplike A A A 20m B 20m C ■ FIGURE 13-12. Neoplasia. A, Synovial sarcoma. Synovial fl uid aspirated from a dog with lameness localized to a solitary joint. The sample is predominated by large sheets of pleomorphic spindle cells that are sometimes separated by a fi ne, pink, streaming stroma. The cells display moderate pleomorphism. This joint had an associated soft tissue mass that was ultimately diagnosed as synovial sarcoma. The cells in this photograph display some cytologic features of malignancy and may be neoplastic, but could potentially be reactive synovial cells. As with many mesenchymal tumors, it is diffi cult to defi nitely diagnose malignancy based solely on cytologic detail. (Wright; HP oil.) Same case B-C. Histiocytic sarcoma. Dog. B, An aspirate from a localized soft tissue mass that arose around the joint and infi ltrated the muscle revealed a pleomorphic population of round cells. These cells display malignant features of anisokaryosis, variable nuclear-to-cytoplasmic ratio, coarse chromatin, and prominent nucleoli. The abundant basophilic cytoplasm suggests histiocytic origin, which was confi rmed by immunohistochemistry. (Wright; HP oil.) C, Pleomorphic round cell neoplasm that was negative for CD3, CD79, and MUM1 (lymphoid) antigens and positive for CD45, CD18, E-cadherin (leukocyte, histiocytic, and dendritic cell antigens, respectively). (H&E; IP.) (B and C, Courtesy of Dr. Rose Raskin, Purdue University.) cell ( Fig. 13-14) . This feature has been reported in cytology of rhabdomyosarcomas (Fallin et al., 1995) . Rarely, striations are visible within the cytoplasm. Specifi c diagnosis of a rhabdomyoma or rhabdomyosarcoma on cytology is diffi cult; however, presence of striations and strap cells can assist with the diagnosis. Histopathology with immunohistochemistry is often necessary for a defi nitive diagnosis (refer to Chapter 17).

Fine-needle aspiration of bone is becoming a more commonly used technique (Britt et al., 2007) . Aspiration of bone is indicated if an osteolytic or osteoproliferative lesion is observed, which may involve cortical lysis or periosteal bone proliferation. Healthy bone does not exfoliate well; however, infl amed or neoplastic bone exfoliates much more readily. Bone aspiration can be performed with an 18-gauge needle, or if there is considerable lysis a smaller gauge needle may be used. Both aspiration and fenestration techniques can be used to obtain a specimen for cytology. Additionally, imprints of tissue obtained for a biopsy can be used for cytology. Care must be taken to blot as much blood off the biopsy specimen as possible before making the imprints. The bone should be sampled in the center of the lesion rather than on the periphery of the lesion, where there may be a transition between normal and abnormal bone.

Histology of normal bone consists of osteocytes housed in lacunae with low numbers of osteoblasts and osteoclasts. Osteoblasts make osteoid, which appears on cytology and histology as a pink, amorphous, proteinaceous material. On the outer surface of the bone is the periosteum, which consists of fi brous connective tissue. Cytology of normal bone is usually of very low cellularity and may consist of 1 to 2 cells per slide or less. Usually only the spindle-shaped mesenchymal cells of the periosteum will exfoliate. However, when bone remodeling occurs secondary to trauma, infl ammation, or neoplasia, reactive osteoblasts can be observed on cytology. These cells are round with an eccentrically placed nucleus with prominent nucleoli and occasionally prominent Golgi apparatus (Fig. 13-15 ). It is important not to mistake reactive osteoblasts for neoplastic osteoblasts and sometimes this is quite diffi cult. The presence of osteoblasts in the absence of infl ammation 

and minimal criteria of malignancy should be interpreted with caution.

Cytology of lytic bone exfoliates more readily. Processes associated with bone lysis include infl ammation, neoplasia, hypertrophic osteopathy, and aneurismal bone cyst. Osteomyelitis usually consists of suppurative to pyogranulomatous infl ammation with varying numbers of neutrophils, macrophages, and multinucleated giant cells depending on the cause of infl ammation. Reactive osteoblasts and other mesenchymal cells may also be observed. Osteomyelitis can be caused by bacteria or fungus. Bacterial osteomyelitis can occur uncommonly via hematogenous spread but more commonly secondary to bite wounds, trauma, postsurgical infections, or foreign body. There are many causes of bacterial osteomyelitis; however, organisms commonly associated with osteomyelitis include Actinomyces and Nocardia. The infl ammatory process associated with bacterial osteomyelitis is suppurative rather that pyogranulomatous. It is important to remember when aspirating bone that there is often peripheral blood contamination, and some white blood cells will be observed secondary to the hemodilution. It may be necessary to evaluate a CBC or peripheral blood smear on the patient to determine if there are truly increased numbers of neutrophils within the sample. Observation of intracellular bacteria is diagnostic for bacterial osteomyelitis; however, culture is recommended for all infl ammatory bone aspirates.

Fungal osteomyelitis consists of a pyogranulomatous to suppurative infl ammatory process and often consists of neutrophils, macrophages, and multinucleated giant cells. Organisms are not always observed in the aspirate so it is important to aspirate more than once and examine all of the slides. Fungal organisms known to cause osteomyelitis include Blastomyces dermatitidis (Fig. 13-16A&B) , Cryptococcus spp., Coccidioides immitis ( Fig. 13-17A&B) , Histoplasma capsulatum, and, less commonly, Candida spp., Aspergillus spp., Geomyces spp. (Erne et al., 2007) , and Sporothrix spp. Blastomyces is a round yeast organism with a double-contoured wall and broad-based bud. Coccidioides organisms are large (10 to 100μm) blue or clear spheres with fi nely granular protoplasm. Histoplasma organisms by comparison are quite small (2 to 4 μm) and are easily phagocytized by macrophages and can be observed within the cytoplasm of macrophages. The organisms are round with a thin capsule and crescentshaped, eccentrically placed, eosinophilic nuclei. Cryptococcal organisms are round with a narrow-based bud and thick, nonstaining with Wright stain, mucoid capsule.

Bone tumors often cause boney lysis or proliferation. They can be categorized as primary bone tumors, tumors of bone marrow, tumors that invade bone, or tumors that are metastatic to bone (Rosol et al., 2003) . Primary bone tumors include osteosarcoma, chondrosarcoma, fi brosarcoma, hemangiosarcoma, and synovial cell sarcoma (Chun, 2005) . Cytologically these tumors can be diffi cult to differentiate from one another. 

General cytologic features include round to spin dleshaped cells with basophilic cytoplasm and an eccentrically placed nucleus, with prominent nucleoli (Fig. 13-18A ) (Reinhardt et al., 2005) . Fibrosarcoma and hemangiosarcoma are less likely to have round cells and the majority of the cells are spindle-shaped. Within the background, osteosarcoma, chondrosarcoma, fi brosarcoma, and synovial cell sarcomas can have varying amounts of eosinophilic proteinaceous material ( Fig. 13-18B ). This material can also be observed within the cytoplasm of the cells. Chondrosarcomas can have a large amount of matrix in the background, which results in understaining of the cells (Fig. 13-19A-D) . In spite of these subtle differences, these tumors can be diffi cult to differentiate cytologically and biopsy with histopathology is an important diagnostic tool. Additional cytochemical testing can be done to improve the sensitivity of cytologic diagnosis of osteosarcoma. Staining of the neoplastic cells for alkaline phosphatase activity with nitroblue tetrazolium chloride/ 5-bromo-4-chloro-3-indolyl phosphate toluidine salt (NBT/BCIP) increases the sensitivity and specifi city of differentiating osteosarcoma from other mesenchymal tumors (Barger et al., 2005) . One limitation of this staining technique is that reactive osteoblasts will stain positive as well, so obvious criteria of malignancy must be observed before this test is performed (Fig. 13-20A ). Positive staining is indicated by grayish-black staining of the cytoplasm (Fig. 13-20B ). Previously unstained slides must be used. After staining for alkaline phosphatase activity, cells can be lightly counterstained with a Romanowsky stain to examine the positive cells for the appropriate criteria of malignancy. Lymphoma and plasma cell tumors are considered tumors of bone marrow that can result in bone lysis. The morphology of these cells appears similar to that in other tissues (Fig. 13-21) . Plasma cell tumors produce a characteristic punched-out radiographic appearance. A combination of diagnostic tests is necessary to diagnose the plasma cell tumor as multiple myeloma. In addition to radiographs and cytology, protein electrophoresis of serum and urine are recommended.

Squamous cell carcinoma is the most common tumor that can invade bone. Usually the cytology reveals Cytologic features of this tumor are similar to those in other locations. Many tumors can metastasize to bone but the common tumors that metastasize to bone include prostatic, lung, and mammary carcinomas. Identification of metastatic neoplasms can be diffi cult because cytology is often accompanied by reactive osteoblasts and osteoclasts. However, a second population of cells can often be differentiated from the reactive population. Epithelial neoplasms are usually clustered, but when they metastasize they may appear more poorly differentiated ( Fig. 13-22A&B ). Additional stains are very helpful for diagnosis. 

Localized and disseminated histiocytic sarcoma of dendritic cell origin in dogs

Utilization of alkaline phosphatase staining to differentiate osteosarcoma from other vimentin positive tumors

Diagnosing appendicular osteosarcoma with ultrasound-guided fi ne-needle aspiration: 36 cases

Chun R: Common malignant musculoskeletal neoplasms of dogs and cats

The diagnosis and prognosis of synovial tumors in dogs: 35 cases

Systemic infection with Geomyces organisms in a dog with lytic bone lesions

What is your diagnosis? A 12-month-old dog with multiple soft tissue masses

Possible pseudogout in two dogs

Value of direct smears of synovial fl uid in the diagnosis of canine joint disease

Molecular identifi cation of Ehrlichia ewingii infection in dogs: 15 cases

Polyarthritis in a dog

Laboratory evaluation and interpretation of synovial fl uid

Bacterial septic arthritis in 19 dogs

Metastatic carcinoma presenting as hind-limb lameness: diagnosis by synovial fl uid cytology

Noninfectious nonerosive arthritis in dogs

Absolute and relative cell counts for synovial fl uid from clinically normal shoulder and stifl e joints in cats

Assessment of cytological criteria for diagnosing osteosarcoma in dogs

Animal models of bone metastasis

Polyarthritis associated with visceral leishmaniasis in a juvenile dog

Chronic bicipital tenosynovitis in dogs: 29 cases