Materials and Methods

Criteria for selection of cases—A retrospective case series was conducted. Medical records of foals aged 3 weeks to 6 months admitted to Texas A&M University's Large Animal Teaching Hospital between 1987 and 2007 were analyzed. Some data from 61 of these foals were previously published.² For purposes of this study, the diagnosis of R equi infection was defined as a foal aged 3 weeks to 6 months that met at least one of the following criteria: clinical signs consistent with pneumonia (fever, cough, or respiratory distress) or failure to thrive and supportive diagnostic imaging findings (ultrasonographic or radiographic evidence of pulmonary or intra-abdominal abscesses); cytologic identification of gram-positive coccobacilli in, or isolation of R equi by microbiologic culture from, a body fluid or tissue; or lesions identified as caused by R equi at necropsy. Radiographic evidence of R equi pneumonia included an alveolar pattern with regional consolidation or focal or multifocal nodular or cavitary lesions.²⁴ Ultrasonographic evidence of R equi pneumonia included hypoechoic lesions of the peripheral portion of the lungs consistent with abscess formation or consolidation.²⁵

Medical record review—The following information was collected for each foal included in the study: date of admission, signalment, history, clinical signs, results of diagnostic testing, treatments administered, duration of hospitalization, invoice, and outcome. Definitions were created for each EPD.

Diarrhea was defined as historical or observed passage of 3 or more episodes of liquid feces. Ulcerative enterotyphlocolitis was defined as microscopic evidence of ulcerative and pyogranulomatous lesions in the small intestine, cecum, or large colon. Intra-abdominal lymphadenitis was defined as the presence of at least 1 enlarged, reactive intra-abdominal lymph node measuring > 2 cm but ≤ 5 cm in maximal diameter. Intra-abdominal abscess formation was defined as the presence of at least 1 abscess, presumably previously associated with a lymph node, > 5 cm in diameter with a caseous center from which R equi was isolated by microbiologic culture. A diagnosis of peritonitis was based on peritoneal fluid sample findings of a nucleated cell count ≥ 30,000 cells/μL, cytologic evidence of neutrophils with degenerative changes or intracellular bacteria, or isolation of R equi on microbiologic culture.

Nonseptic or immune-mediated polysynovitis was defined as effusion of at least 1 synovial structure from a foal lacking lameness and from which no bacteria were identified either cytologically or by microbiologic culture of synovial fluid, if arthrocentesis was performed. Septic synovitis was defined as a foal with joint effusion and lameness in which synovial fluid had a nucleated cell count ≥ 30,000 cells/μL, had a total protein concentration > 3 g/dL, had cytologic evidence of intracellular bacteria, or yielded growth of R equi by microbiologic culture. Osteomyelitis was defined by use of radiographic evidence of bone lysis or by means of gross and microscopic evidence of bone lysis on necropsy with positive microbiologic culture results for R equi. Paravertebral abscesses were defined as radiographic or ultrasonographic findings consistent with abscess formation in foals with neurologic signs consistent with the lesion location.

Uveitis was defined as aqueous flare, hypopyon, hyphema, miosis, or iris discoloration. To be classified as septic uveitis, microbiologic culture of the aqueous humor had to yield R equi. Foals with positive results of immunofluorescence of the uveal tract were classified as having definitive evidence of immune-mediated uveitis.

Mediastinal lymphadenopathy was defined as enlargement of the mediastinal lymph nodes identified radiographically, ultrasonographically, or during necropsy. Rhodococcus equi bacteremia was defined as growth of the organism by microbiologic culture of blood. Peripheral lymphadenopathy was defined as a peripheral lymph node that was subjectively deemed enlarged by palpation and from which R equi was isolated by microbiologic culture. Subcutaneous abscesses were defined as subcutaneous masses that yielded R equi by microbiologic culture but that were not associated with peripheral lymph nodes. Cellulitis and lymphangitis were defined as diffuse external swelling with warmth and pain that yielded R equi by microbiologic culture. Pericarditis was defined as pericardial fluid in excess of that normally identified by use of ultrasonography or radiography or as postmortem identification of > 100 mL of turbid pericardial fluid. Pleural effusion was defined as a quantity of pleural fluid in excess of that expected; that was identified ultrasonographically, radiographically, or during necropsy; and that had abnormal cytologic findings such as neutrophils with degenerative or toxic changes or intracellular bacteria. Pyogranulomatous hepatitis, nephritis, and meningitis were all determined microscopically after death and were defined as pyogranulomatous lesions or microabscesses in the liver, kidney, or meninges, respectively. Immune-mediated hemolytic anemia was defined as a PCV < 15% in a foal with positive results of testing for surface-bound antibodies by use of an immunoassay (Coombs' test). Other EPDs were identified either by use of positive microbiologic culture results, by use of microscopic evidence of pyogranulomatous lesions of the associated tissues, or by disorders that appeared concurrently with R equi bronchopneumonia and resolved with treatment for the pneumonia. For each EPD identified, further information was collected regarding the methods of identification of the EPD, location of the EPD, treatment of the foal, and outcome (lived, died, or euthanized) of the foal.

Statistical analysis—Data were summarized as median and range values (for continuous data) and by use of contingency tables (for categoric data). Data were analyzed with the χ² or Fisher exact test (for categoric data) and Wilcoxon rank sum tests (for continuous data). A significance level of P < 0.05 was used.

Results

Frequency and types of EPDs—One hundred fifty foals met the inclusion criteria for the study. Of these, 111 (74%) foals had at least 1 EPD. The median number of EPDs among all foals was 2 (range, 0 to 9); the median number of EPDs among foals that had at least 1 EPD was 2 (range, 1 to 9). A wide variety of EPDs were identified (Table 1). Some EPDs were always recognized in live foals while others required postmortem examination for identification. However, other EPDs could be identified either before or after death. Among 63 foals in which a necropsy was performed, 76% (48/63) had at least 1 EPD that was subclinical and had not been recognized before death.

Abdominal manifestations—Diarrhea was the most common EPD identified. The etiology of diarrhea in each affected foal was often undetermined. In 26 of 50 foals, diarrhea was not observed until after initiation of antimicrobial treatment with a macrolide. Of those 26 foals, 15 were treated with erythromycin (15 of 89 erythromycin-treated foals [17%]), 7 were treated with azithromycin (7 of 26 azithromycin-treated foals [27%]), and 4 were treated with clarithromycin (4 of 12 clarithromycin-treated foals [33%]). Of the 31 foals in which ulcerative enterotyphlocolitis was diagnosed at necropsy, only 12 (39%) had diarrhea before death. Nine of the 31 foals were admitted for failure to thrive or grow appropriately. Twenty-eight foals with enterotyphlocolitis also had an intra-abdominal abscess (n = 13 foals) or abdominal lymphadenitis (15).

Abdominal lymphadenitis alone was detected in 15 foals, intra-abdominal abscesses alone were detected in 15 foals, and both lymphadenitis and abscesses were detected in 10 foals. Of the 25 foals in which abdominal lymphadenitis was diagnosed, 5 had abdominal lymphadenitis detected before death via ultrasonography and 20 had abdominal lymphadenitis detected at necropsy. Abdominal ultrasonography allowed for antemortem detection of abdominal lymphadenopathy in 5 of 12 foals in which it was performed. Only 5 of 25 foals with lymphadenitis had diarrhea before death. One foal was admitted for colic, and 6 foals were admitted for failure to thrive or grow appropriately. Thoracic radiography was performed in 24 foals with abdominal lymphadenitis, and all 24 foals had radiographic evidence of pneumonia. The single foal that did not undergo thoracic radiography had clinical signs of pneumonia, and multiple R equi pulmonary abscesses were identified at necropsy.

Intra-abdominal abscesses were identified ultrasonographically in 12 of 25 foals before death, whereas 13 of 25 foals had intra-abdominal abscesses identified only at necropsy. Abdominal ultrasonography allowed for antemortem detection of intra-abdominal abscesses in 12 of 17 foals in which it was performed. The diagnosis of an intra-abdominal abscess was confirmed by exploratory celiotomy in 5 of the 12 foals prior to euthanasia. Nine of 25 foals had antemortem signs of diarrhea. One foal was admitted for an acute episode of colic, and 8 foals were admitted for failure to thrive or grow appropriately. Twenty-three of 24 foals had radiographic evidence of pneumonia. One foal did not undergo thoracic radiography, but had clinical signs of pneumonia and multiple R equi–positive abscesses in the lungs at necropsy. The other foal had no clinical signs of pneumonia, normal findings on thoracic radiography, and no gross or microscopic evidence of pneumonia at necropsy, but was admitted because of diarrhea and colic; this foal had a single large mesenteric abscess that yielded R equi by microbiologic culture at necropsy.

Of the 11 foals with peritonitis, 7 had concurrent intra-abdominal abscesses, 3 had abdominal lymphadenopathy, and 3 had ulcerative enterotyphlocolitis. Five of the foals also had diarrhea, 4 were admitted for fever of unknown origin, and 1 was admitted for evaluation of failure to thrive.

Pyogranulomatous hepatitis was detected in 16 foals at necropsy. None of those foals had clinical signs specifically referable to the liver before death, and there was no significant correlation with antemortem serum γ-glutamyltransferase or alkaline phosphatase activities in these foals. Six of these foals underwent abdominal ultrasonography, and no abnormalities were observed in images of the liver.

Pyogranulomatous nephritis was detected in 7 foals at necropsy. None of these foals underwent abdominal ultrasonographic examination before death. All 7 foals had been admitted for evaluation of respiratory tract disease.

Musculoskeletal manifestations—Nonseptic or immune-mediated polysynovitis was observed in 37 foals. Arthrocentesis was performed in 15 of those foals, and no bacteria were identified via cytologic examination or microbiologic culture of these synovial fluid samples. In the remaining 22 foals, the diagnosis was made presumptively on the basis of synovial effusion that was not associated with signs of lameness. Immunofluorescence was performed at necropsy on synovial fluid samples from 2 of these 37 foals. Results were positive in 1 foal and negative in the other. The foal with negative results had been treated with a glucocorticoid for 15 days prior to euthanasia for a concurrent EPD (immune-mediated hemolytic anemia and thrombocytopenia). The median number of synovial structures affected among foals with polysynovitis was 5 (range, 1 to 14 joints). Synovial structures affected included tarsal (n = 30 foals), carpal (18), metatarsophalangeal (18), metacarpophalangeal (15), stifle (ie, femoropatellar or femorotibial; 15), proximal interphalangeal (2), distal interphalangeal (1), scapulohumeral (1), and coxofemoral (1) joints and the extensor carpi radialis tendon sheath (2).

Septic synovitis was documented for 15 foals. Microbiologic culture of synovial fluid samples from 8 of these foals yielded R equi. Among foals with septic synovitis, the median number of joints affected was 2 (range, 1 to 10 joints). Joints affected included tibiotarsal (n = 12 foals), stifle (7), carpal (5), metacarpophalangeal (2), metatarsophalangeal (2), coxofemoral (1), and radiohumeral (1) joints. Of the 15 foals, 13 had clinical and radiographic evidence of bronchopneumonia. Two foals had no clinical evidence of pneumonia and did not undergo thoracic radiography or ultrasonography before death. Of these 2 foals, 1 had pulmonary abscesses found at necropsy while the other had no evidence of pulmonary disease identified at necropsy.

Osteomyelitis was found in 5 foals. None of those foals had clinical signs of pneumonia, although 3 of 4 that underwent thoracic radiography had evidence of pneumonia. Three of these 5 foals had concurrent septic synovitis. Two of those foals had septic tibiotarsal joints, with involvement of the medial trochlear ridge in 1 foal and the lateral trochlear ridge of the talus in the other. One of these foals also had a septic medial femorotibial joint with osteomyelitis in the medial femoral condyle. The other foal with a septic coxofemoral joint had osteomyelitis of the acetabulum. Of the 2 foals without joint involvement, 1 had septic physitis of the distal femoral physis and the other had distal radial physis. The latter foal developed a subsequent angular limb deformity that necessitated surgical correction. Two foals had cellulitis as an EPD; 1 had an affected metatarsal area with no underlying lesion, and the other had cellulitis of the ventrolateral aspect of the thorax and pectoral region associated with a mediastinal abscess.

Thoracic manifestations—Mediastinal lymphadenopathy was identified in 12 foals, 8 of which were admitted while in respiratory distress or had an abnormal respiratory noise. This EPD was detected before death in 7 foals (4 by radiography and 3 by a combination of radiography and ultrasonography). The diagnosis was made only at necropsy in 5 foals, all of which had thoracic radiography performed and 4 of which also had thoracic ultrasonography performed before death.

Pericarditis was identified in 6 foals. Pericarditis was identified in 2 of the 6 foals by antemortem examination via radiography or ultrasonography. Pericarditis was identified for the remaining 4 foals at necropsy. Likewise, pleural effusion was identified in 2 of 5 foals before death by use of either radiography or ultrasonography. Pleural effusion was identified in the 3 other foals at necropsy.

Neurologic manifestations—Vertebral body osteomyelitis was diagnosed radiographically in 2 foals and at necropsy in 1 foal. One foal was admitted with cauda equine syndrome and was found to have osteomyelitis of vertebrae S2, S3, and S4; this foal was successfully treated.¹² This foal had had a diagnosis of R equi pneumonia 6 weeks previously, but clinical and radiographic signs of pneumonia had resolved following 4 weeks of treatment with trimethoprimsulfamethoxazole. A second foal was admitted for evaluation of hind limb paralysis; it died before radiography could be performed, but was found to have osteomyelitis of T13 at necropsy as well as substantial pneumonia. The third foal was admitted for evaluation of a painful swelling over the cranial cervical portion of the vertebral column and hind limb ataxia; osteomyelitis of S2 and S3 as well as pneumonia were diagnosed radiographically. This foal was euthanized at the farm with no necropsy performed. All 3 foals had paravertebral abscesses associated with the osteomyelitis.

Granulomatous meningitis was detected in 5 foals at necropsy. None of those foals had been observed to have any specific neurologic signs, although they may have had a depressed attitude, which was attributed to their pulmonary disease. None of these 5 foals had concurrent vertebral osteomyelitis. There was no CSF collected for analysis from any of these 5 foals.

Hematologic manifestations—Rhodococcus equi bacteremia was found in 11 of 19 R equi–infected foals in which at least 1 blood sample was submitted for microbiologic culture. There was no significant association between bacteremia and any of the other EPDs.

Immune-mediated hemolytic anemia was detected in 2 foals. One was admitted for evaluation of fever of unknown origin and the other for evaluation of pneumonia. Both also had a concurrent nonseptic polysynovitis, and one had bilateral uveitis. Both had a precipitous decline in PCV 3 to 4 days after treatment with erythromycin was implemented for the treatment of pneumonia. Both foals developed profound weakness, tachycardia, tachypnea, and icterus. Hematologic evaluation revealed marked anemia (PCV < 10%), increased serum total and indirect bilirubin concentrations, and positive Coombs' test results. One foal also developed a concurrent immune-mediated thrombocytopenia. The foal with immune-mediated hemolytic anemia and immune-mediated thrombocytopenia had marked clinical and hematologic improvement after treatment with dexamethasone, but this foal was euthanized 2 months later for failure to thrive and multiple intra-abdominal abscesses were found at necropsy. The second foal was euthanized without further treatment.

Miscellaneous—Uveitis was detected in 16 foals, 4 of which were described as being unable to see. Four foals had unilateral signs, and 12 were affected bilaterally. Microbiologic culture of aqueous humor fluid obtained after death from 4 foals yielded R equi, and these foals were thus classified as having septic uveitis. Four foals were confirmed to have immune-mediated uveitis on the basis of positive results of immunofluorescent staining and a lack of microbial growth in culture. The etiology of the uveitis remained undetermined in the other 8 foals.

Peripheral lymphadenopathy was found in 11 foals, 9 of which had submandibular lymphadenopathy, 1 of which had retropharyngeal lymphadenopathy, and 1 of which had inguinal lymphadenopathy. Eight foals had subcutaneous abscesses in areas unrelated to lymph nodes. Aspirates of the abscesses all yielded growth of R equi by microbiologic culture. Three of the 8 foals had underlying septic synovitis or osteomyelitis, and 2 had a mediastinal abscess. Two of the 8 foals with subcutaneous abscesses had no clinical or radiographic evidence of pneumonia.

There were 17 other EPDs that were each documented once. Some of these (eg, sinusitis, myositis, stomatitis, and pyometra) were infectious in origin, in which specimens grew R equi in microbiologic culture. Others were deemed to be immune mediated (eg, telogen effluvium). There were also EPDs that were undetermined in origin but appeared concurrently with R equi pneumonia and resolved with treatment (eg, bone marrow erythroid hypoplasia and seizures).

Although the finding was not considered to be an EPD, hyperthermia related to macrolide administration was described in 6 foals with R equi bronchopneumonia. This was defined as a rectal temperature > 40°C (104°F) associated with exposure to hot, humid environmental conditions in a foal that was receiving a macrolide antimicrobial. Five of those foals were being treated with erythromycin, and 1 foal was being treated with azithromycin at the time that hyperthermia developed. All 6 foals improved with environmental modification, including provision of a cooled stall and avoidance of exposure to hot, humid weather conditions. Three of the 6 foals with hyperthermia had an accompanying EPD.

Outcome of foals with EPDs versus those without—The duration of hospitalization was not significantly associated with the presence of EPDs. The invoice for foals with EPDs (median, $1,316; range, $29 to $8,877) was not significantly (P = 0.405) different from that for foals without EPDs (median, $1,117; range, $29 to $11,532). Overall, 70 of the 150 (47%) foals either died (27/150 [18%]) or were euthanized (43/150 [29%]). The proportion of foals that survived was significantly (P < 0.001) higher among foals without EPDs (32/39 [82%]) than among foals with EPDs (48/111 [43%]). Similarly, the total number of EPDs per foal was significantly (P < 0.001) associated with survival status; medians (ranges) of EPDs for foals that survived, died, or were euthanized were 1 (0 to 4), 3 (0 to 5), and 4 (0 to 9), respectively.

Of the EPDs that were detected in more than 1 foal, several were significantly associated with decreased survival, compared with all foals without that EPD (Table 2). Ulcerative enterotyphlocolitis, pyogranulomatous hepatitis, pyogranulomatous nephritis, and granulomatous meningitis were all histopathologic diagnoses determined only at necropsy and were consequently all significantly associated with nonsurvival. The association between each of the remaining EPDs and survival was analyzed by examining either all foals or only those foals with EPDs that were identified before death. Uveitis, bacteremia, septic synovitis, and intra-abdominal abscesses were significantly associated with nonsurvival, whether they were identified before death or at any time (ie, before or after death; Table 3). Abdominal lymphadenitis, pericarditis, and pleural effusion were significantly associated with nonsurvival when analysis included foals with EPDs identified at any time; however, the association was no longer significant when foals in which these conditions were diagnosed only after death were excluded. Although some other EPDs appeared to be associated with decreased probability of survival, too few foals were examined to evaluate significance. These EPDs included vertebral body osteomyelitis (1/3 foals survived), paravertebral abscess (1/3 survived), peritonitis (3/11 survived), mediastinal lymphadenopathy (3/12 survived), and immune-mediated hemolytic anemia (0/2 survived).

Association of EPDs with factors determined at admission—No significant association was found between breed or sex of foal and whether or not a foal had any EPDs. Foals without an EPD were younger at admission (median, 57 days; range, 9 to 154 days) than foals with EPDs (median, 61 days; range, 21 to 183 days), but this difference was not significant (P = 0.25). The duration of clinical signs prior to admission was significantly (P = 0.024) longer for foals with EPDs (median, 7 days; range, 0 to 90 days) than for foals without EPDs (median, 3 days; range, 0 to 70 days). The proportion of foals examined by referring veterinarians prior to admission was significantly (P < 0.001) higher for foals with EPDs (96/111 [86%]) than for foals without EPDs (21/39 [54%]). The total number of EPDs was also significantly (P < 0.001) greater among foals that were referred (median, 2 EPDs; range, 0 to 9) than foals that were not referred (median, 0 EPDs; range, 0 to 7 EPDs).

No significant differences were found in rectal temperature or respiratory rate between foals with and without EPDs; however, the heart rate at admission was significantly (P = 0.031) greater among foals with EPDs (median, 92 beats/min; range, 48 to 180 beats/min) than among foals without EPDs (median, 80 beats/min; range, 52 to 130 beats/min). Clinicopathologic data (Table 4) that differed significantly between foals with and without EPDs were WBC count, neutrophil count, and BUN concentration. When WBC count was considered as a categoric variable (WBC count > 20 × 10³ cells/μL vs WBC count ≤ 20 × 10³ cells/μL), the proportion of foals with WBC count > 20 × 10³ cells/μL was significantly (P = 0.043) greater for foals with EPDs (49/101 [49%]) than for foals without EPDs (9/32 [28%]). The total number of EPDs in an individual foal was also significantly (P = 0.001) associated with WBC count. Neutrophil counts were also considered as a categoric variable (neutrophil count > 15 × 10³ cells/μL vs neutrophil count ≤ 15 × 10³ cells/μL). The proportion of foals with neutrophil counts > 15 × 10³ cells/μL was significantly (P = 0.003) greater for foals with EPDs (61/101 [60%]) than for foals without EPDs (9/32 [28%]). The total number of EPDs in an individual foal also was significantly (P < 0.001) associated with neutrophil count. Concentration of BUN was significantly (P < 0.001) different between foals with EPDs (median, 11 mg/dL; range, 4 to 91 mg/dL) and foals without EPDs (median, 6 mg/dL; range, 3 to 32 mg/dL), but the median value for both groups fell within the reference range for our laboratory (7 to 28 mg/dL). Concentration of BUN was significantly (P = 0.018) higher among foals in which pyogranulomatous nephritis was identified at necropsy (median, 54.5 mg/dL; range, 11 to 91 mg/dL) than among foals without nephritis (median, 15 mg/dL; range, 3 to 80 mg/dL). While not significant (P = 0.06), serum creatinine concentrations were higher among foals with nephritis (median, 2.3 mg/dL; range, 1.2 to 7.0 mg/dL) than among foals without nephritis (median, 1.5 mg/dL; range, 0.7 to 11.0 mg/dL). Serum activities of alkaline phosphatase and J-glutamyltransferase were not significantly different between foals with and without EPDs or between foals with and without pyogranulomatous hepatitis.

Discussion

Extrapulmonary manifestations of R equi infection are highly prevalent among infected foals, as manifested by the finding that 74% (111/150) of foals in this study had at least 1 EPD. This prevalence is higher than that found by Zink et al³ (EPD identified in 54% of 131 foals at necropsy) and Takai et al⁴ (EPD identified in 59% of 108 foals at necropsy). Extrapulmonary disorders included in those reports were primarily intestinal or orthopedic. A wide variety of EPDs were identified in our study that included aberrant sites of infection outside the lung, immune-mediated disorders, and complications resulting from antimicrobial treatment. For most foals in our study, it was impossible to determine the exact pathogenesis of the EPD. Given that 76% (48/63) of foals that underwent necropsy had at least 1 subclinical EPD identified, it is conceivable if not likely that the prevalence of EPDs was even higher among foals in our study because foals that survived (and thus did not undergo necropsy) could have had subclinical EPDs that were not identified (such as subclinical intraabdominal lymphadenitis). Previous reports by Zink et al³ and Takai et al⁴ of foals with R equi undergoing necropsy indicate intestinal tract EPDs were found in 46% (60/131) and 58% (63/108) of foals, respectively. These findings illustrate the great likelihood of our study having underestimated the prevalence of EPDs in foals surviving to discharge, given that we found at least 1 intestinal tract EPD (ulcerative enterotyphlocolitis, abdominal lymphadenitis, or intra-abdominal abscesses) in only 33% (50/150) of our foals.

The high prevalence of subclinical EPDs also underscores the potential difficulty in recognizing affected foals. In this study, only 39% (12/31) of foals with ulcerative enterotyphlocolitis had diarrhea, only 24% (6/25) of foals with abdominal lymphadenitis had diarrhea (n = 5 foals) or colic (1) before death, and only 40% (10/25) of foals with intra-abdominal abscesses had diarrhea (9) or colic (1) before death. Previously, Zink et al³ found that 38% (23/60) of foals with enterotyphlocolitis, mesenteric lymphadenopathy, or intra-abdominal abscesses at necropsy had had antemortem signs referable to the gastrointestinal tract. In many foals with abdominal manifestations of R equi infection, failure to thrive or grow appropriately was the only admitting complaint. Vertebral body osteomyelitis has also been historically difficult to detect radiographically, with only 2 of 6 foals reported in the literature^12–15 having radiographic evidence despite extensive gross lesions at necropsy. Two of our 3 foals with vertebral body osteomyelitis had a radiographic diagnosis, with the third foal dying before radiography could be performed. Given the reported discrepancy between diagnosis and radiographically detectable abnormalities, we may have misclassified some foals as normal because they did not yet have radiographic changes. If these foals did not have a necropsy performed, their lesions would have been missed in this study.

It is also interesting that EPDs may occur without evidence of concurrent pneumonia as was previously reported by Zink et al,³ who found that 5 of 60 foals with intestinal EPDs at necropsy had no evidence of pneumonia. Takai et al⁴ also found that 3 of 63 foals with enteritis or intestinal lymphadenopathy had no evidence of bronchopneumonia. Likewise, 2 of 6 foals reported to have vertebral osteomyelitis in the literature^12–15 had no historical, radiographic, or necropsyidentified evidence of pneumonia. One foal in our study that was euthanized because of a large intra-abdominal abscess and another with septic synovitis had no history, clinical signs, or radiographic or necropsy evidence of R equi bronchopneumonia. Another foal with osteomyelitis had no clinical or radiographic evidence of pneumonia; however, it survived and thus no necropsy was performed. In some instances, as for our foal with sacral vertebral osteomyelitis, the pneumonia was previously diagnosed and successfully treated weeks prior to development of the EPD.

Some EPDs, such as nonseptic polysynovitis, appear to be highly suggestive of R equi pneumonia because they are not seen commonly with other infectious causes of foal pneumonia.⁵ Our prevalence of 25% (37/150) was slightly lower than that previously reported by Sweeney et al,⁵ in which 35% (17/48) of foals with pneumonia attributed to R equi had a nonseptic synovitis characterized by joint effusion without lameness.⁵ The lower prevalence in our study may have been the result of lack of recognition or inadequate record keeping. We often recognize polysynovitis in these foals to be an incidental finding with no specific treatment indicated (other than to treat the pneumonia). Therefore, in some foals, this EPD may have not been recorded in the medical record. We also found the tibiotarsal joints to be most commonly affected, but as previously described,⁶ many other joints or synovial structures may be involved.

The etiology of each EPD was not always able to be determined. For example, 26 of 50 foals that developed diarrhea did so only after the initiation of macrolide treatment. It is possible that the diarrhea was an adverse drug reaction and not an extrapulmonary manifestation of the disease. As was previously reported,²⁶ diarrhea was seen most commonly in foals treated with clarithromycin, although there was no significant difference among treatments. Of the 26 foals that developed diarrhea after induction of macrolide treatment in the present study, 4 were found to have ulcerative enterotyphlocolitis at necropsy, which may have been the cause of the diarrhea. This raises the possibility that the temporal relationship between the development of diarrhea and treatment was circumstantial in those foals. Fifteen of the 26 foals survived and thus did not have a necropsy to definitively rule in or out the presence of enterotyphlocolitis. Therefore, the etiology of the diarrhea remained undetermined in most foals. It could also be argued that both foals with immune-mediated hemolytic anemia reported here developed anemia subsequent to erythromycin treatment and could therefore be a drug reaction. However, the 1 foal that was treated with corticosteroids remained on erythromycin as well and had an initial positive response to treatment prior to succumbing to intra-abdominal abscesses. We cannot differentiate the etiology of the immune-mediated hemolytic anemia, but we consider it likely to be attributable to the infectious process as reported with other bacterial infections. Another commonly reported clinical finding in foals with R equi infections being treated with a macrolide antimicrobial is hyperthermia. As it is believed, although not certain, that this is attributable to the drug and not the infection itself, we did not include hyperthermia as an EPD. However, of these 150 foals, hyperthermia was reported in 6 foals. This 4% prevalence in our population was significantly lower than the 25% previously reported,²⁷ a discrepancy that may be attributed to the fact that most of our foals were housed in an air-conditioned hospital.

The presence of at least 1 EPD was significantly associated with nonsurvival in foals with R equi infections in our study. When looking at the total number and specific EPDs, it is likely that the associations of some EPDs with failure to survive were a result of, rather than a cause of, poor prognosis because necropsy was required for identification of some EPDs (eg, ulcerative enterotyphlocolitis or granulomatous meningitis) and because some of these EPDs were subclinical and identified only at necropsy. The effect of postmortem diagnosis of EPDs such as pyogranulomatous hepatitis, nephritis, or meningitis on the clinical status or prognosis of a live foal cannot be ascertained from this study. Certainly foals that survived may have had subclinical hepatitis that went undiagnosed and never caused a clinical problem. An attempt was made to differentiate between the effect of antemortem and postmortem diagnosis in the statistical analysis. Unfortunately, because of small numbers of some EPDs, significance was not always maintained despite clinical impression. For example, despite their lack of significance, all foals with pericarditis, pleural effusion, or immune-mediated hemolytic anemia failed to survive. Other EPDs such as abdominal lymphadenitis were no longer significantly associated with nonsurvival when analyzing only those foals in which these conditions were identified after death. If anything, it is likely that some surviving foals had EPDs that went unrecognized, resulting in an underrepresentation of the proportion of surviving foals with EPDs. Even for foals with EPDs that were clinically apparent and identified before death, it was difficult to determine the true effect on survival of that specific EPD. For many of those foals, it was impossible to determine from the case record the exact cause for euthanasia (eg, was the foal euthanized for financial reasons, because of the clinical assumption of a poor prognosis for that specific EPD, or because of the severity of another EPD or the respiratory tract disease?). Therefore, we were unable to conclude whether that EPD alone would have resulted in death, whether the combination of that EPD with pneumonia or other EPDs would have resulted in death, or whether the EPD would have resolved with treatment.

The association between some specific EPDs and nonsurvival may have also been confounded by other concurrent EPDs. The strong association between uveitis and failure to survive was surprising to us. This EPD was readily apparent before death and did not require necropsy for diagnosis, so the question remains whether uveitis was merely an indicator of other EPDs or whether the association with nonsurvival was real and uveitis represented a more severe systemic response.

Historical factors associated with increased odds of EPDs included referral status and duration of clinical signs prior to admission. The majority of the foals in our population were referred to us by other veterinarians (117/150 [78%]), which may indicate that the study population represented more severely affected foals and therefore a higher prevalence of EPDs than one might see in a population of foals at a breeding farm. Increased disease severity might therefore account for the association between referral status and the presence of and total number of EPDs. Many of the foals that were referred to our hospital were those that were either severely ill from the onset of clinical signs or had failed to respond to initial treatment provided by the referring practitioner. Because EPDs were significantly more common among foals that were referred for evaluation (relative to foals for which our hospital provided initial evaluation), the proportion of foals with EPDs identified in general practice would be expected to be lower than that observed in this study. Although we would not expect the types of EPDs to vary substantially among geographic regions, there might be differences in the prevalence, risk factors, and prognosis for EPDs in different areas as a result of differences in management factors or diagnostic and therapeutic procedures. The magnitude of the significant difference in duration of clinical signs prior to admission between foals with EPDs and those without was small (7 days vs 3 days) and thus of questionable clinical importance.

The increased heart rate in foals with EPDs may have also been an indicator of increased disease severity. The tachycardia may have been a response to hypoxemia resulting from pulmonary disease, hypovolemia, or hypoperfusion caused by decreased cardiovascular stability attributable to EPDs such as diarrhea and pericarditis or to pain associated with EPDs such as septic synovitis or intra-abdominal abscesses causing colic. Increased WBC and neutrophil counts were significantly associated with the presence of and total number of EPDs. This finding may be clinically important when attempting to diagnose EPDs before death. Foals in which R equi bronchopneumonia is identified or those with nonspecific clinical signs warrant further diagnostic investigation (eg, abdominal ultrasonography and blood culture) when they have markedly increased blood WBC or neutrophil counts; however, the high prevalence of EPDs among foals with R equi infection indicated that all foals with suspected or confirmed R equi infection should be thoroughly evaluated for EPDs. The importance of the association between BUN concentration and EPDs was difficult to interpret because the medians for both groups were within the reference range for BUN concentrations. However, because increased BUN concentration was significantly associated with nephritis caused by R equi, foals with R equi bronchopneumonia and increased BUN and serum creatinine concentrations should be evaluated for this EPD (rather than attributing the increased BUN concentration to either prerenal azotemia or chance alone).

As with any retrospective study, this study had a number of limitations. The principal limitation is the potential for misclassification of foals without EPDs. Because not all foals were examined either by abdominal ultrasonography or necropsy, it is possible that some foals considered free of EPDs had subclinical EPDs. Moreover, additional EPDs may have existed in foals for which at least 1 EPD was identified before death. This limitation does not nullify the value of this study because it underscores the importance of evaluating foals carefully for EPDs. Further prospective studies of the frequency of EPDs, risk factors for EPDs, and their impact on survival are warranted. A thorough diagnostic evaluation should be undertaken for all foals with confirmed or suspected R equi bronchopneumonia, as there are a wide variety of extrapulmonary manifestations that may be difficult to diagnose but may affect outcome.