Duodenal Tuberculosis: Radiological Features on Barium Studies and their Clinical Correlation in 28 Cases
BACKGROUND: A retrospective analysis of 28 cases of duodenal tuberculosis (TB) was done to evaluate radiological findings and their value in the diagnosis of the disease. subjects AND METHODS: Upper gastrointestinal and small bowel series of 28 patients with duodenal tuberculosis were analysed for radiological findings. The diagnosis of duodenal TB was confirmed by surgery and biopsy in 18, on the basis of radiological findings and response to treatment in 9, and on the basis of findings on upper gastrointestinal scopy and biopsy in 1 patient. RESULTS: The study included 28 patients (14 males, 14 females). The mean age was 32.1 (range 5-65). Twenty-three (82.2%) patients presented with obstructive symptoms while five manifested with dyspeptic symptoms. Of the latter, 4 had ulcerations in the third and fourth parts of the duodenum. In the remaining patient, the mucosa of the duodenum could not be clearly visualised. Two patients had extrinsic impression at the D2-D3 and D3-D4 segments. In 23 patients with obstructive symptoms, 18 demonstrated luminal narrowing of varying degrees and 5 had a sharp band-like cut-off at the third part of the duodenum. Of the 18 patients with luminal narrowing, 13 had extrinsic compression, 12 had proximal dilatation and 14 had ulcerations mainly in the second and third parts of the duodenum. Biliary involvement was seen in 3 patients without any signs or symptoms directly referable to the biliary involvement. CONCLUSION: Though duodenal TB lacks specific radiological features, barium studies help to localise and define the area of narrowing and ulcerations and help to confirm the presence of lymph nodes causing compression of the duodenum.
Although gastrointestinal (GI) tuberculosis (TB) is a major health problem in India, involvement of the duodenum in the tuberculous is considered rare. In fact, the duodenum is the least commonly affected segment of the small bowel. Clinically, duodenal TB mimics various other GI pathologies. There is a lack of specific clinical, radiological and endoscopic signs of duodenal TB. This communication presents a review of the radiological findings noted in 28 cases of duodenal TB and evaluates the nature of the correlation between the radiological findings and the clinical presentation with a view to assess their value in the diagnosis of the disease.
We retrospectively reviewed the barium studies of 28 patients with duodenal TB seen in our hospital between 1988 and 2002. The diagnosis of TB was confirmed on the basis of a combination of surgical, pathological, radiological and gastroscopic findings [Table - 1]. Barium studies included either UGI series or a small bowel series. All films were reviewed for radiological features of duodenal TB. Ultrasonography (USG) and Computed Tomography (CT) were used to correlate the findings in 6 patients. Four patients were subjected to upper GI scopy. The radiological findings were correlated with clinical complaints.
The mean age of patients was 32.1 years (range:5-65 years), of these 14 were male.
The clinical presentation could broadly be divided into dyspeptic (n=5) and obstructive symptoms (n=23). Obstructive symptoms included vomiting frequently after meals and bilious, epigastric pain, and generalised abdominal pain. The dyspeptic symptoms included epigastric pain, nausea, and occasional vomiting since a year or so. All the patients had history of having received anti-ulcer treatment that usually resulted in temporary relief. All the patients who presented with dyspeptic symptoms were below 30 years of age. Only 5 of the 28 patients had history of pulmonary TB.
In patients with dyspeptic symptoms, radiological findings that were encountered included- luminal narrowing at the second and third parts of the duodenum (D2-D3, n=4) and (D3-D4, n=1) [Figure - 1], ulcerations in the second to fourth parts of the duodenum (n=4) [Figure - 2] and extrinsic compression at the third and fourth parts of the duodenum (n=2). In one patient, the mucosa was not clearly visualised. Other findings included scarred and deformed duodenal cap (n=2), widening of the C-loop of the duodenum (n=1) and peri-pancreatic lymphadenopathy on USG (n=1). One of these 5 patients with dyspeptic symptoms required surgery; the remaining 4 patients were managed conservatively.
In 23 patients with an obstructive presentation, 18 showed luminal narrowing of varying degrees and 5 showed a sharp, band-like cut-off of the duodenum at the D2-D3 junction [Figure - 3], resembling the superior mesenteric artery syndrome. Proximal dilatation was seen in all these 5 patients and they required bypass surgery for relieving the obstruction. At surgery, mesenteric and peri-pancreatic lymph node enlargement was noted. The associated findings noticed in the 18 patients with luminal narrowing were extrinsic compression (13, 72.2%), ulcerations (12, 66.7%) and lymphadenopathy as noted during surgery or at USG or CT scan (14, 77.8%). In this group 12 patients underwent bypass surgery for reliving the obstruction. D2-D3 involvement was the commonest, accounting for 20 (71.5%) patients; followed by D3-D4 (4, 14.3%), D2-D4 (3, 10.7%) and D2 alone (1, 3.6%). Concomitant tuberculous involvement of other parts of the gastrointestinal tract was noted in 9 out of 10 patients who underwent barium follow-through studies. Involvement of the following portions was noted: proximal portion of the jejunum was involved in 5 and ileum and ileo-caecal junction was involved in 4 patients.
A persistent narrow stream of barium in the bowel suggestive of the 'string sign' was seen in patients with the involvements of D2-D3 or D3-D4 segments [Figure - 1]. The biliary tract was involved in 3 subjects. The involvement was in the form of reflux of barium into the biliary tree (2 subjects) and presence of air in the common bile duct in one. None of these 3 patients had symptoms or signs related to the biliary system.
Calcified lymph nodes were noted in one patient in the region of pancreas.
Lymph nodes with necrotic centres in the peri-pancreatic region, peritoneal thickening and free fluid in the abdomen were seen on the CT and USG scans in 9 patients.
Out of the 28 cases in our series, 13 were seen prior to 1997. All these cases required surgery for the diagnosis and treatment. In the 15 cases in last 5 years, only 5 required surgery. The radiological features which helped to decide on conservative therapy with anti-tubercular drugs were: D2-D3 involvement with mass effect, narrowing and ulcerations.
The incidence of GI TB in India is approximately 1% of all general hospital admissions. Duodenal involvement constitutes 2% of all GI TB. This uncommon duodenal involvement is supposedly related to the inhibitory influence of gastric acid on the mycobacteria, rapid transit time through the duodenum allowing for a reduced contact time and the comparative paucity of the lymphoid tissue in the duodenal segment as compared to the other segments of the gastrointestinal tract. The male: female ratio in our study was 1:1. However, both male predominance and female predominance have been reported in different studies.
Takeshita et al observed that most cases before 1987 had involvement of the most distant portions of the duodenum and recent cases after 1987 had tuberculous lesions in the descending portions of the duodenum. We have not encountered such a change of location in our patients over the period of 14 years. In their series, Balikian et al reported 4 cases of TB specifically involving the pyloro-duodenal area. In our series, we encountered no such patient. However, changes of chronic duodenal ulcerations and scarred deformed duodenal cap were noted in 2 patients.
In a series reported by Kolwale and co-workers, widening of the duodenal 'C'-loop was seen in 5 out of 28 patients. We encountered this specific finding in only one patient. Our study seems to confirm the statement of Anderson et al, who stated that tuberculous involvement of the duodenum is rare and is generally seen in cases with massive involvement of the rest of the intestinal tract. In our series, in 9 out of 10 patients who underwent barium follow-through study, other portions of the intestine were seen to have been involved. As discussed by Chawla and Berry, the contributions of enlarged lymph nodes and of intrinsic involvement (through ulcerations and cicatrisation) in the causation of duodenal obstruction are difficult to estimate. Out of the18 patients with luminal narrowing and obstructive symptoms in our series, 14 had ulcerations and lymph node enlargement. Out of 14 with lymph node enlargement, 4 did not have ulcerations and out of 14 with ulcerations, 4 did not have lymph node enlargement. Hence most of the times the obstruction is the result of both the processes-extrinsic compression by enlarged lymph nodes and narrowing by intrinsic affection acting together. However, an individual process can cause obstruction in isolation.
Lymph node enlargement around the superior mesenteric artery (SMA) produces a sharp band-like cut-off of the D2-D3 portion and simulates the SMA syndrome. According to Gupta et al, the cause of the duodenal obstruction is more likely to be due to the caseating mass than the lymph nodes themselves.
Out of the 28 cases in our series, 13 were seen prior to 1997. All these cases required surgery for the diagnosis and treatment. In the 15 cases in the last 5 years, only 5 required surgery. In 10 patients the diagnosis was suspected on the basis of clinical and radiological features. The radiological features which helped to decide on conservative therapy with anti-tubercular drugs were: D2-D3 involvement with mass effect and involvement in the form of the narrowing of the tract and mucosal ulcerations. The presence of peri-pancreatic lymph nodes with necrotic centres, peritoneal thickening and free fluid in the abdomen were the main features noted on the CT and USG scans.
Duodenal TB should be considered as one of the differential diagnoses in patients presenting with features of duodenal obstruction and in patients with dyspepsia not responding to medical therapy in countries where TB is endemic.
Duodenal TB lacks specific clinical and radiological features, barium studies help to localise and define the area of narrowing and ulcerations, which cause obstruction or dyspepsia.