Clinical, Endoscopic And Histological Differentiations Among Crohn’s Disease, Intestinal Tuberculosis And Primary Colorectal Lymphoma

Background and AimsCrohn’s disease (CD) and intestinal tuberculosis (ITB) are granulomatous diseases of the intestine. The principal cause of intestinal tuberculosis is mycobacterium tuberculosis. According to the World Health Organization global tuberculosis control 2010 report,1.7 million people died from tuberculosis in 2009, equal to 4700 deaths a day. There were 9.4 million new TB cases in 2009. In the developed world, an increase in the number of cases with TB has been observed mainly attributable to immigration, human immunodeficiency virus. The aetiology of CD is not clear. The incidence of CD is increasing in China. According to a report from Hong Kong, the age-adjusted incidence of CD was 3.0 per 100,000 and has increased by 3 fold during the past decade.Crohn’s disease and intestinal tuberculosis have very similar clinical, radiological, endoscopic and histological characteristics. However, the therapy is different for the two diseases. If CD was misdiagnosed as ITB, unnecessary anti-tuberculosis therapy would pose a risk of drug toxicity and delays the treatment of CD. Conversely, treatment of ITB with steroids alone would lead to deterioration of condition. So it is of great importance to distinguish intestinal tuberculosis from Crohn’s disease. Because of their similarity, sometimes there is difficulty in differentiation between these two entities and misdiagnosis would happen. Differentiating CD from ITB challenges clinician in the areas where tuberculosis and CD coexist, such as China.Some studies reported that epithelioid granuloma is one of the best histological criterions for distinguishing CD from intestinal tuberculosis. CD68 is a transmembrane glycoprotein highly expressed by human monocytes and tissue macrophages. Immunohistochemisty for CD68 can help identify the epithelioid granuloma.Primary colorectal lymphoma is uncommon and simulates Crohn’s disease. Sometimes it is difficult to distinguish primary colorectal lymphoma from Crohn’s disease. This study was designed to evaluate clinical, endoscopic and histological features of patients with CD, ITB and primary colorectal lymphoma in order to identify features to differentiate Crohn’s disease from the other two diseases.Patients and MethodsPatientsPatients with a definite diagnosis of intestinal tuberculosis, Crohn’s disease or primary colorectal lymphoma from Nan Fang hospital, Guangdong Province during the period from February 2000 to June 2010 were selected for the study. Forty-three patients with intestinal tuberculosis were included in the study. The diagnosis of intestinal tuberculosis was established based on one or more of the following evidences:(1) demonstration of acid-fast bacilli or caseating granuloma in histological sections (n=6); (2) radiological, and/or operative evidence of active tuberculosis elsewhere and dramatic response to antituberculous treatment (n=26); (3) strong suspicion of tuberculosis on both clinical and histological grounds, together with sensitive response to antituberculous treatment (n=11). All the patients with intestinal tuberculosis were confirmed to complete anti-tuberculous treatment with resolution of symptoms and no recurrence.Fifty-three patients with a final definite diagnosis of Crohn’s disease were selected as the reference groups. The diagnosis of Crohn’s disease was confirmed by clinical evaluation and a combination of endoscopic, histological, radiological, and/or biochemical investigations. The conditions of intestinal infections (particularly tuberculosis), lymphoma and carcinoma were carefully excluded. Thirty-seven patients with Crohn’s disease have been followed up for at least one year in our centre.Thirty one patients with primary colorectal lymphoma confirmed histologically were recruited for the study. All patients meet the diagnostic criteria established by Dawson:(i) no superficial enlarged lymph nodes when the patient is first seen; (ii) chest radiographs show no obvious enlargement of the mediastinal nodes; (iii) the white blood cell counts, total and differential, are within normal limits; (iv) at laparotomy, only regional lymph nodes are affected by disease; and (v) the liver and spleen seem free of tumour.Patients who had undergone definitive treatment previously and patients with incomplete data set were excluded.Clinical data collectionClinical data of the patients regarding demography, the duration of disease, clinical manifests, past history of patients, laboratory examination, radiological, endoscopic and histological features, and the treatment (corticosteroid, immunosuppressive drugs, and anti-tuberculosis therapy, surgical intervention) were reviewed and noted by a senior gastroenterologist. Histological AssessmentsIn patients with Crohn’s disease and intestinal tuberculosis, biopsy specimens were taken from abnormal sites in the colon and terminal ileum. Two serial sections of 5μm thickness from the paraffin blocks were stained with HE and by immunohistochemisty for CD68 respectively. The sections were examined by a pathologist using coded slides to avoid observer bias. The granuloma was further confirmed by immunohistochemistry of the serial section. In each section, granulomas were characterised with respect to number, size (maximum diameter), location and the presence or absence of confluence and caseous necrosis. Disproportionate submucosal inflammation was also recorded.Statistical AnalysisThe T test or Mann-Whitney U test was used to compare mean values of appropriate parameters. The Chi-square test or Fisher’s exact probability test was used to test the significance of differences in frequency of the various parameters. P value (two sides) less than 0.05 was considered statistically significant. The analyses were implemented on SPSS 13.0. Each parameter with statistical significance was further analyzed for sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy.Results1. The initial diagnosis and medical interventionOf the 43 patients with intestinal tuberculosis,2 patients (5%) were initially misdiagnosed as Crohn’s disease and received prednisone treatment for three weeks, which caused the deterioration of condition, and then the therapy changed to anti-tuberculosis medication with dramatic clinical response. One patient (2%) undergone exploratory laparotomy and caseating granuloma was found in operative histological sections. The remaining 40 patients (93%) were initially treated as tuberculosis with resolution of symptoms.Of the 53 patients with Crohn’s disease, five patients (9%) had initially received anti-tuberculosis therapy trial for four weeks with no obvious clinical response. After reevaluation, the therapy of these five patients changed to prednisone and 5-aminosalicylic acid medication with dramatic resolution of symptoms. Seven patients (13%) had received laparotomy for the diagnosis was not clear, other 41 patients (78%) with Crohn’s disease had initially received 5-aminosalicylic acid, glucocorticoids, immunosuppressives or infliximab with good clinical response.Of the 31 patients with primary colorectal lymphoma, one (3%) patient had initially received anti-tuberculosis therapy for one year with symptoms relapse. Three (10%) patients were initially misdiagnosed as Crohn’s disease and received prednisone treatment. Ten (32%) patients underwent exploratory laparotomy. The other 17(55%) patients were confirmed by histology of endoscopic biopsy specimens.2. clinical featuresIntestinal tuberculosis, Crohn’s disease and primary colorectal lymphoma predominantly affect males. The median duration of symptoms of CD was significantly longer than that of ITB and primary colorectal lymphoma. The mean age of patients with primary colorectal lymphoma at the time of diagnosis was older than Crohn’s disease (P=0.000). Abdominal pain was the most common symptom in each group. Fistula and extra-intestinal manifestations were more common in patients with CD.Night sweat was more common in patients with intestinal tuberculosis.3. Laboratory and radiological featuresPositive tuberculin test and positive tuberculosis antibody were more common in patients with intestinal tuberculosis.62.8% patients with intestinal tuberculosis had pulmonary tuberculosis infiltration in their chest radiography. Ascites was significantly more common in patients with intestinal tuberculosis than in patients with CD (20.5% vs.0.0%, P=0.001).Abdominal lymphadenopathy was less common in patients with Crohn’s disease.4. Endoscopic featuresLongitudinal ulcers, cobblestone appearance and pseudopolyps were seen more often in CD compared with intestinal tuberculosis (P<0.05). A patulous ileocecal valve and transverse ulcers were significantly more common in patients with intestinal tuberculosis than in patients with CD (P<0.05). The diagnose accordance rate of colonoscopy was 25.9% in primary colorectal lymphoma. Based on colonoscopic features, primary colorectal lymphomas were divided into three types: ulcerative type, polypoid type and massive type. Compared with B-cell lymphoma, the lesions of ulcerative type were more frequently seen in patients with primary colorectal T-cell lymphoma (92.9% vs.27.3%, P=0.002).5. Histological featuresImmunohistochemical staining can clearly show epithelioid cell granulomas, hence reduce missed diagnosis and misdiagnosis. Epithelioid cell granulomas were seen more often in patients with intestinal tuberculosis than in patients with Crohn’s disease (70.0% vs.21.6%, P=0.000). The granulomas in patients with ITB were larger compared with CD (P=0.000).In all,43.3% of patients with intestinal tuberculosis had granulomas exceeding 300um in maximum diameter compared with none of patients with Crohn’s disease (P=0.000). The number of granulomas per section was more in patients with intestinal tuberculosis compared with Crohn’s disease.20.0% of patients with intestinal tuberculosis had sections containing more than five granulomas compared with none of patients with Crohn’s disease (P=0.006). Confluent granulomas were seen in 40.0% of patients with intestinal tuberculosis and in 2.7% of patients with CD (P=0.000). Caseous necrosis was identified in 13.3% of patients with intestinal tuberculosis. There was no significant difference in the feature of disproportionate submucosal chronic inflammation between Crohn’s disease and intestinal tuberculosis (P>0.05).The histological subtypes of primary colorectal lymphoma were:T-cell lymphoma (53.6%), diffuse large B-cell lymphoma (35.7%), MALT lymphoma (7.1%) and mantle-cell lymphoma (3.6%). Only 33.3% of patients with primary colorectal lymphoma were confirmed by histopathology of initial biopsy, and a part of the false negative patients were confirmed as lymphoma by histopathology of second biopsy.6. Endoscopic ultrasonography in primary colorectal lymphomaTen patients with primary colorectal lymphoma underwent endoscopic ultrasonography because the diagnosis was unclear, and nine of these ten patients had undergone coloscopy and biopsy prior to endoscopic ultrasonography. One patient underwent endosonography twice, so a total of eleven endosonographies were performed among the ten patients. The diagnosis of colorectal lymphoma was made ten times during endosonography, and the diagnosis of Crohn’s disease was made one time, therefore the diagnose accordance rate of endosonography was 90.9%. The endosonography features of primary colorectal lymphoma were:thickening of intestinal wall, disappearance of normal colorectal layers, and diffuse hypoechoic lesion instead. Regional lymphadenopathy was found in six patients with primary colorectal lymphoma during endosonography.Conclusions1. Crohn’s disease, intestinal tuberculosis and primary colorectal lymphoma have similar clinical, radiological, endoscopic characteristics, only a part of patients with primary colorectal lymphoma are confirmed by histopathology of initial biopsy, at the same time epithelioid cell granulomas are identified only in a proportion of patients with Crohn’s disease and intestinal tuberculosis. All these factors cause misdiagnosis between Crohn’s disease and intestinal tuberculosis, and primary colorectal lymphoma is apt to be misdiagnosed as Crohn’s disease or intestinal tuberculosis.Therefore intestinal tuberculosis and primary colorectal lymphoma should be considered in the differential diagnosis of all suspected cases of Crohn’s disease.2. The following features support the diagnosis of intestinal tuberculosis:(i)night sweat; (ii)positive tuberculin skin test; (iii)positive antibody to mycobacterium tuberculosis; (iv)ascites; (v)extraintestinal tuberculosis such as pulmonary tuberculosis; (vi)transverse ulcers, a patulous ileocecal valve in full length colonoscopy; (vii)granulomas exceeding 300um in maximum diameter, more than five granulomas per section, caseous necrosis, confluent granulomas.3. The following features support the diagnosis of Crohn’s disease:(i) fistulas, extraintestinal manifests; (ii) longitudinal ulcers, cobblestone appearance, and pseudopolyps in full length colonoscopy;(iii)non-caseating granulomas, maximum diameter of granulomas less than 300um, the number of granulomas per section less than five, and discrete granulomas.4. Based on colonoscopic features, primary colorectal lymphoma can be divided into three types:ulcerative type, polypoid type and massive type. Primary colorectal T-cell lymphoma often presents as ulcerative type, primary colorectal B-cell lymphoma often presents as polypoid type and massive type. Endoscopic ultrasonography is superior to colonoscopy in diagnosing primary colorectal lymphoma. The endosonography features of primary colorectal lymphoma include:thickening of intestinal wall, disappearance of normal colorectal layers, and diffuse hypoechoic lesion instead. Repeated biopsy can increase the positive rate of pathological diagnosis in primary colorectal lymphoma. 5. A comprehensive analysis of clinical, laboratory, radiological, endoscopic, and histological features may help distinguish Crohn’s disease from intestinal tuberculosis and primary colorectal lymphoma. Repeated coloscopy and biopsy during follow-up contribute to the correct diagnosis.