Diagnosis of Abdominal Tuberculosis: Sonographic Findings

Transcription

1 X/95/ American Roentgen Ray Society Rajeev Jam1 Sukhpal Sawhney1 Dinesh K. Bhargava2 Manorama Berry1 Received May 20, 1994; accepted after revision July 19, Presented at the 18th International Congress of Radiology, Singapore, January I Department of Radiodiagnosis, All India Institute of Medical Sciences, New Delhi, , India. Address correspondence to R. Jam. 2 Department of Gastroenterology, All India Institute ofmedical Sciences, New Delhi, , India. Diagnosis of Abdominal Tuberculosis: Sonographic Findings in Patients with Early Disease OBJECTIVE. The diagnosis of abdominal tuberculosis is often difficult, because clinical manifestations and results of laboratory studies are nonspecific. If sonographic findings are sufficiently characteristic for diagnosis, sonography would be useful, especially in India, where abdominal tuberculosis is common and more expensive imaging techniques are not easily available. Accordingly, we performed sonography to establish the sonographic findings in cases of early tuberculosis in 56 patients with abdominal tuberculosis who had normal barium studies of the small bowel. SUBJECTS AND METHODS. Fifty-six patients with clinical features suggestive of abdominal tuberculosis (history of fever, abdominal pain, and weight loss) with no history of intestinal obstruction and normal barium studies of the small bowel had abdominal sonography. All sonograms were independently assessed by three radiologists, and the findings were tabulated by consensus. Diagnosis of tuberculosis was confirmed by sonographicaily guided biopsy of mesenteric lymph nodes in 19 patients, analysis of aspirated ascitic fluid in 12, and response to antitubercuious chemotherapy in 25. Sonography was repeated 1, 3, 6, and 12 months after antituberculous chemotherapy was begun. Abdominal sonograms were also performed in 30 healthy volunteers, and measurements of mesenteric thickness were recorded. The mesenteric thickness was statistically compared in two groups of patients: patients at presentation with patients at the end of antituberculous chemotherapy and patients at presentation with healthy individuals. RESULTS. The mesenteric thickness in healthy Individuals ranged from 5 to 14 mm. Sonographic findings in all patients with abdominal tuberculosis included an echogenic thickened mesentery (15 mm) wfth mesenteric lymphadenopathy. Other findings were dilated small bowel loops In 38 patients, minimal ascites in 17, matted small bowel loops in five, and omental thickening with altered echogenicity in three. Regression of these changes was noted on follow-up of all patients undergoing treatment. CONCLUSION. The characteristic sonographic features of early abdominal tuberculosis are mesenteric thickness of 1 5 mm or more and an increase in the mesenteric echogenicity (due to fat deposition), combined with mesenteric lymphadenopathy. Presence of dilated small bowel loops and ascites further substantiate the diagnosis. AJR 1995;165: Abdominal tuberculosis is endemic in India. Typically, patients have one of three main types of disease, that is, intestinal, penitoneal, or mesentenic lymph node involvement, though there is considerable overlap, and mesentenic lymphadenopathy has been documented in almost all patients irrespective of the predominant type of disease [1]. By the time a clinical diagnosis is established, patients usually have a history of frank intestinal obstruction [2] or abnormal barium studies of the small bowel. Confirmation of the final diagnosis of tuberculosis ideally is established by cultuning bacilli from tissues or by demonstrating caseating granulomata at histopathology. Unfortunately in most of our patients such confirmation is not possible. Thus treatment for abdominal tuberculosis is often instituted on the basis of a high index of clinical suspicion and ancillary supportive evidences like raised erythro-

2 1392 JAIN ET AL. AJR:165, December 1995 cyte sedimentation rate, analysis of aspirated ascitic fluid, and nadiologic investigations. Further, there is a subgroup of patients who have abdominal tuberculosis with definite clinical symptoms, without a history of frank intestinal obstruction, and with normal barium studies of the small bowel. In these patients a definite diagnosis cannot be established by noninvasive means. Abnormalities can be documented on barium studies of the small bowel only after intestinal luminal involvement has occurred, so barium studies are frequently normal in many of our patients. If sonography could detect the abnormalities in the mesentery, bowel wall, and omentum, it might help support the diagnosis and may provide a target for collection of tissue by guided biopsies for a bacteriological diagnosis or histopathologic confirmation of the disease. Although a standard distinction between stages of abdominal tuberculosis does not exist in the literature, some authors consider mesentenic lymphadenopathy and small amounts of ascites as evidence of early disease, whereas intestinal disease and massive ascites are later manifestations and usually follow mesentenic lymphadenopathy [3, 4]. We defined early abdominal tuberculosis as a stage of disease in which there was no evidence of intestinal disease, that is, normal barium studies of the small bowel and no clinical history of intestinal obstruction. The diagnosis of abdominal tuberculosis in patients included in our study was established by biopsy of enlarged mesentenc lymph nodes, analysis of aspirated ascitic fluid, or a positive response to a therapeutic trial of antitubenculous chemotherapy. Therefore, we carried out a prospective study to define sonognaphic abnormalities in patients with early abdominal tuberculosis. Subjects and Methods Fifty-six patients (3-65 years old) who were included in this study were referred from the gastroenterology services of our hospital over an 18-month period. Although these patients had clinical suspicion of abdominal tuberculosis, they did not have any history of intestinal obstruction and had normal barium studies of the small bowel. The diagnosis of tuberculosis was subsequently established in all patients by histopathology, analysis of ascitic fluid, or a positive response to a therapeutic trial of antituberculous treatment. Only patients who met the aforementioned cnitena were included in the study. Presenting symptoms were abdominal pain in 43 patients, weight loss in 38 patients, intermittent fever in 37 patients, anorexia in 29 patients, and vomiting in 21 patients. Five patients had additional evidence of extraabdominal tuberculosis: thoracic in three, vertebral in one, and pelvic in one. A prospective study was done over a 3-year period. Abdominal sonograms were obtained in all patients with a commercially available 3.5-MHz curved linear array transducer with special attention to the appearance of the small-bowel mesentery, mesenteric thickness, mesentenic and retropenitoneal lymphadenopathy, ascites, and small bowel loops. The mesentery was first profiled in its width, radiating from the midline just above the umbilicus, and its maximum thickness was measured. Then the same segment of mesentery was profiled at right angles to the first direction, and the thickness was measured again (Fig. 1). The greater of the two measurements was recorded. All sonograms were obtained by the same radiologist. Hard copy images were independently reviewed by two other radiologists. The diagnosis of abdominal tuberculosis was made only when mesentenic Fig. 1.-Sonogram of left upper quadrant of abdomen shows normal small bowel mesentery In healthy 30-year-old man. Normal mesentery is shown as echogenic linear structure (between arrowheads) with mesenteric vessels in it. No lymph nodes can be Identified within mesentery. All around mesentery are normal loops of small intestine. thickness equal to or more than 15 mm and obvious mesentenic lymphadenopathy were observed. Additional findings, namely dilated small bowel loops, ascites, and omental abnormalities, were noted and considered as supportive evidence. The diagnosis of tuberculosis was established by sonographically guided core biopsy of enlarged mesentenic lymph nodes (19/56), analysis of aspirated ascitic fluid (12/56), and response to antituberculous chemotherapy (25/56). Follow-up sonograms were obtained at 1, 3, 6, and 12 months after starting antituberculous chemotherapy to check for any resolution or change in the initial sonographic findings. However, only 21 patients reported for regular follow-up until completion of therapy. Thirty-five patients dropped out of the sonographic follow-up protocol at various stages of chemotherapy (19 after 1 month, 11 after 3 months, and five after 6 months). Measurements of mesentenic thickness in 30 agematched healthy subjects were noted to serve as a baseline for companison with those of patients. The differences in mesenteric thickness between patients at presentation and at the end of antituberculous chemotherapy and between patients at presentation and healthy subjects were analyzed statistically. Results The predominant sonognaphic finding in all patients was thickening of the small bowel mesenteny combined with mesentenic lymphadenopathy. The mesentenic thickness ranged from 13 to 47 mm (mean ± SD, 25 ± 8) in patients with disease. Mesentenc thickening was detected most often in the left upper quadrant (47/56), paraumbilical region (43/56), night lumbar region (41/56), night lower quadrant (35/56), and suprapubic region (28/56). The mesentenc lymph nodes were 5-30 mm (mean ± SD, 12 ± 6 mm) in diameter. The lymph nodes were either discrete on seen as matted conglomerate masses (Fig. 2A). Larger lymph nodes, more than 15 mm in diameter, were often necrotic (as noted on cone biopsy specimens). Other frequent observations in 38 patients (67.8%) were dilatation and hyperpenistalsis of small bowel loops in the region of mesentenic abnormalities. Matted and fixed small bowel loops arranged radially around thickened mesentery were seen in five patients (8.9%). Wall thickening in the tenminal ileum was noted in three patients (5.3%).

3 AJR:165, December 1995 SONOGRAPHY OF EARLY ABDOMINAL TUBERCULOSIS 1393 Fig. 2.-Serial abdominal sonograms of 64-year-old patient with abdominal tuberculosis receiving antituberculous chemotherapy. A, Sonogram at presentation shows thlckened(40 mm), echogenic mesentery containing multiple enlarged hypoechoic, discrete, and conglomerate lymph nodes. Small amount of ascites is seen (arrows). Dilated, fluid-filled, thick-walled ioop of small intestine Is seen at periphery (arrowheads). This figure highlights characteristic sonographic findings in abdominal tuberculosis. Core biopsy of mesenteric lymph nodes showed caseating granulomata with giant cells. B, Sonogram after I month of antituberculous chemotherapy shows mesenteric thickness reduced to 28 mm. Significant reduction in size and number of mesenteric nodes is noted. Ascites has decreased. Note marked, early response of enlarged lymph nodes to therapy. C, Sonogram after 6 months of antituberculous chemotherapy shows echogenic mesentery has further reduced to 20 mm. Lymphadenopathy and ascites have almost completely resolved. Surrounding bowel loops are no longer dilated. Persistent thickening and echogenicity of mesentery Is due to fat deposftion and edema. D, Sonogram after completion of antftuberculous chemotherapy shows mesenteric thickness has reverted to normal level (9 mm). Small, normalsized nodes are seen within mesentory. Ascites and dilatation of bowel loops have completely regressed. All sonographic findings of abdominal tuberculosis regress completely with antituberculous therapy. Ascites was noted in 1 7 patients (30%). It was usually small in volume, with free fluid seen in the pelvis or between dilated loops of small bowel; massive ascites was present in only one patient. In five of these patients, thin, incomplete, floating septa were noted. Thickened and edematous omentum with altered echogenicity was seen in three patients (5.3%). Retropenitoneal lymphadenopathy was seen in one. An associated gnanulomatous mass in the gallbladder (histologically proven to be tuberculosis) was seen in one patient and resolved completely on antitubenculous chemotherapy. The diagnosis of abdominal tuberculosis was confirmed in 56 patients by the following methods: sonographically guided cone biopsy of mesentenic lymph nodes in 1 9 patients (33.9%), analysis of aspirated ascitic fluid in 12 (21.4%), and a positive response to antitubenculous chemotherapy in 25 (44.6%). Histopathologic examination of biopsy specimens from mesentenic lymph nodes showed granulomata with giant cells and caseous necrosis. Acid-fast bacilli were not detected by staining or cultures of ascitic fluid or in biopsy specimens of any patient. Regression of mesentenic lymphadenopathy, ascites, bowel-loop dilatation, and mesentenic thickening were consistently noted at sonognaphy (Figs. 2B and 2C). Mesentenic thickening was the last abnormality to regress completely (Fig. 2D). At the end of antituberculous therapy, the mesentenic thickness in 21 patients ranged from 6 to 13 mm (mean ± SD, 10 ± 2 mm). In 30 age-matched healthy subjects, the mesentenic thickness ranged from 5 to 14 mm (mean ± SD, 9 ± 2 mm). The difference in mesentenic thickness between patients at presentation and at the end of therapy and between patients at presentation and healthy subjects was significant (p <.001). Discussion In a study of normal sonographic appearances of the small bowel mesentery in 30 healthy subjects, the thickness of the mesentery ranged from 7 to 12 mm, irrespective of the age on body habitus of the patient, except that the mesentery is more easily identified in obese subjects [5]. We noted a similar thickness of the mesentery (5-14 mm) in our healthy volunteers. A review of the literature revealed few reports on the sonographic findings of abdominal tuberculosis [3, 4, 6-10]. In patients with abdominal tuberculosis in our study, the most common abnormalities noted were mesentenic thickening (15 mm) and mesentenic lymphadenopathy. The echogenicity of the mesentery increased markedly, presumably because of fat deposition due to lymphatic obstruction. Pathologically, this mesentenic thickening results from edema, lymphadenop-

4 1394 JAIN ET AL. AJR:165, December 1995 athy, and fat deposition [1]. In addition, multiple enlarged discrete or conglomerate lymph-node masses were identified within the thickened mesentery. Dilated fluid-filled loops of small intestine with hyperpenistalsis are often seen around the abnormal mesentery (Fig. 2A). Other sonographic findings, namely ascites, matted fixed small bowel loops, omental inflammation, and terminal ileal wall thickening, further supported the diagnosis of tuberculosis. An interesting observation was the relative lack of metropenitoneal lymphadenopathy in patients with tuberculosis, considering the extensive mesentenic and penitoneal abnormalities that are present [3]. This can be explained by the fact that mesentenic lymph nodes drain directly into the thomacic duct (via the cistemna chyli). This observation can help differentiate abdominal tuberculosis from other systemic diseases such as lymphomas. Predominantly on the basis of sonographic findings (and results of lymph node biopsy and analysis of aspirated ascitic fluid when available), patients were classified as having probable tuberculosis and were started on antituberculous chemotherapy. Symptomatic response to antitubenculous chemotherapy was dramatic, with decrease in abdominal pain within 1 week and considerable improvement in the patient s general condition within 2 weeks. On the other hand, bacterial cultures require up to 6 weeks; therefore, therapeutic trials of antituberculous chemotherapy are justified in patients with strong presumptive evidence of tuberculosis, provided close clinical supervision is maintained [4, 10, 11]. Twenty-one patients on antituberculous chemotherapy were followed up with abdominal sonograms. Measurable reduction in enlarged mesentenic lymph nodes was the earliest observed change and in some patients is seen as early as 2 weeks after antitubenculous chemotherapy is started. Progressive reduction of ascites, dilatation and hypenpenistalsis of small intestine, and omental inflammation are also noted with antituberculous chemotherapy (Figs. 2B and 2C). Mesentenic thickening and echogenicity are the last abnormalities to regress completely, despite rapid resolution of the lymphadenopathy. This indicates that in addition to lymphadenopathy, fat deposition and edema also contribute significantly to total mesentenic thickness. As can be seen from Figure 2D, on completion of antitubenculous chemotherapy, mesentenic thickness returns to the same level as in healthy subjects. This correlates with the observations that specific pathologic features of abdominal tuberculosis regress completely after adequate antituberculous chemotherapy [1]. Analysis of sonographic data on mesentenic thickening indicates that the difference in mesentenic thickening between patients at presentation and at the end of antituberculous chemotherapy is significant (p <.001). The difference in mesentenic thickness between patients at presentation and healthy subjects is also significant (p <.001) (Fig. 3). Also, the mesentenic thickness in healthy subjects (maximum, 14 mm) and patients at the end of antitubenculous chemotherapy (maximum, 13 mm) is comparable. Thus, these data indicate that an arbitrary value of 1 5 mm (based on the maximum observed thickness in healthy subjects and 5o 3o 20 b-- : #{247}: : *;: - -. PRE TREATMENT POST TREATMENT * NORMAL Fig. 3.-Graph shows mesenteric thickness (y-axis In millimeters) in all patients (x-axis) at presentation (.), In 21 patients at completion of antituberculous chemotherapy (+), and in healthy subjects (*). Mesenteric thickness In patients after completion of therapy (mean, 10 mm) Is of same magnitude as that seen in healthy subjects at presentation (mean, 8 mm). patients at the end of antituberculous chemotherapy) can be used as a reasonable threshold for disease. In addition to tuberculosis, we have come across only two other conditions that result in mesentenic thickening, namely portal hypertension and lymphoma. Portal hypertension results in mesentenic congestion and thickening (>17 mm) [12]. Dilated and tortuous mesentenic venous channels are seen, but enlarged mesentenic lymph nodes are not (Fig. 4). Lymphoma results in mesentenic thickening with enlarged, usually discrete lymph nodes, but retropenitoneal adenopathy is more common than in tuberculosis. Multicompartmental involvement and lesions not confined to one anatomic area of drainage are strongly suggestive of lymphoma rather than tuberculosis. In lymphoma, lymph nodes in one group genenally show uniform echogenicity, whereas in tuberculosis, different nodes in the same group exhibit different echogenicities. Ascites and omental abnormalities are name in lymphoma [13]. Target lesions may be seen as a result of intestinal wall involvement. Focal lesions in the spleen and liven may be seen in both conditions. We have observed that in lymphoma, the mesentenic thickening is caused predominantly by enlarged lymph nodes. The increased echogenicity of the mesentery (as seen in tuberculosis) is not seen in lymphoma, possibly because of a lesser degree of lymphatic obstruction and consequently less or no fat deposition in the mesentery (Fig. 5). Mesentenc thickening oftubenculosis also needs to be diffementiated from that in pentoneal cancinomatosis, pseudomyxoma pentonei, and mesothelioma. In these conditions, plaques or nodules of metastases on tumor are adherent to all pentoneal sunfaces. Moderate to massive ascites is the hallmark of these conditions. Mesentenc and metropentoneal lymphadenopathy may be present but are unusual. Finally, the gross pathologic abnormalities in Crohn s disease are similar to those in tubenculosis (mesentenc thickening and adenopathy, bowel wall changes), and differentiating between the two on sonography alone in the absence of other relevant clinical data may be difficult.

5 AJR:165, December 1995 SONOGRAPHY OF EARLY ABDOMINAL TUBERCULOSIS 1395 Fig. 4.-Sonogram of mesentery in 31-year-old man with portal hypertension. Sonogram shows thickened (1 7 mm), echogenic mesentery with dliated mesenteric vascular channels. The Important difference from tuberculosis is absence of enlarged lymph nodes within mesentery. Fig. 5.-Sonogram of mesentery in 41-year-old woman with non-hodgkin s lymphoma. Sonogram shows markedly thickened mesentery with enlarged, discrete lymph nodes. Significantly, there is hardly any mesenteric fat surrounding enlarged lymph nodes. To conclude, the most important and characteristic sonographic finding in patients with early abdominal tuberculosis is the combination of mesentenic thickening of 1 5 mm or more with associated mesentenic lymphadenopathy. This combination of findings highly suggests the diagnosis of abdominal tuberculosis, especially in the proper clinical setting. Rapid regression of sonognaphic abnormalities occurs on therapeutic trial of antitubenculous chemotherapy and is a reliable confirmatory sign. It follows that sonognaphic findings can be used to monitor the adequacy of antituberculous themapy. Sonogmaphy thus promises to be the ideal technique for noninvasive diagnosis and subsequent follow-up of patients with abdominal tuberculosis who do not have other objective evidence of disease, for example, abnormal barium studies of the small bowel. REFERENCES 1. Tandon HD, Prakash A. Pathology of intestinal tuberculosis and its distinction from Crohn s disease. Gut 1 972;13: Haddad FS, Ghossain A, Sawaya E, Nelson AR. Abdominal tuberculosis. Dis Colon Rectum 1 987;30: Denton T, Hossain J. A radiological study of abdominal tuberculosis in Saudi population, with special reference to ultrasound and computed tomography. C/in Radioll993;47: Kedar RP, Shah PP, Shivde RS, Malde HM. Sonographic findings in gastrointestinal and penitoneal tuberculosis. C/in Radiol 1994;49: Derchi LE, Solbiati L, Rizzatto G, Pra LD. Normal anatomy and pathological changes of the small bowel mesentery: US appearance. Radiology 1987:164: Akhan 0, Demirkazik FB, Demirkazik A, et al. Tuberculous peritonitis: ultrasonic diagnosis. J C/in Ultrasound 1990;18: Ramaiya LI, Walter DF. Sonographic features of tuberculous peritonitis. Abdom Imaging 1993;1 8: Mitchell RS, Bristol U. Intestinal tuberculosis: an analysis of 346 cases diagnosed by routine intestinal radiography on 5529 admissions for pulmonary tuberculosis: Am J Med Sd 1954;227: Hanson RD, Hunter TB. Tuberculous peritonitis: CT appearance. AJR 1985;144: Lee DH, Lim JH, Ko VT, Yoon Y. Sonographic findings in tuberculous pentonitis of wet-ascitic type. C/in Radioll99l 44: Bhargava DK, Shniniwas, Chopra P, Nijhawan 5, Dasarathy 5, Kushwaha AKS. Penitoneal tuberculosis: laparoscopic patterns and its diagnostic accuracy. Am J Gastroenterol 1 992;87: Patnquin H, Tessier G, Gngnon A, Boisvert J. Lesser omental thickness in normal children: baselinefor detection ofportal hypertension. AJR1985;145: Lynch MA, Cho KC, Jeffrey RB Jr, Alterman DD, Federle MP. CT of penitoneal lymphomatosis. AJR 1 988; 151: