1999.07.26 Meta-analysis-Mycobacterium avium subsp. paratuberculos in Crohn's Disease
BACKGROUND: There are conflicting reports on the role of Mycobacterium avium subsp. paratuberculosis as an etiological agent of Crohn's disease in humans. An attempt has been made to resolve this conflict by reviewing available scientific data on the detection of M. paratuberculosis from patients with Crohn's disease. METHODS: The National Library of Medicine (PubMed) databases were reviewed for information on M. paratuberculosis and/or the association of M. paratuberculosis with Crohn's disease between 1990 and 1998. The papers were analyzed using meta - analytic techniques to determine the association of M. paratuberculosis with Crohn's disease. RESULTS: A total of 727 patients with Crohn's disease were listed in 27 studies between 1990 and 1998. It was observed that 59 % of the studies did not support an etiological role for M. paratuberculosis in Crohn's disease, while 33% concluded that M. paratuberculosis was an etiological agent, and the remaining 7 % of the studies could not draw any conclusion. Between 1990-1998, meta - analysis of the data on Insertion Sequence 900-based polymerase chain reaction suggested that M. paratuberculosis was more likely to be detected in patients with Crohn's disease (Odds ratio 2.35:1). However, the odds of detecting M. paratuberculosis in Crohn's patients was confounded by both geographical location and the year of the study. CONCLUSIONS: Based on meta analysis of the 20 available peer reviewed scientific papers (458 patients with Crohn's disease, 584 others) that used IS 900 PCR based assays, it is inferred that M. paratuberculosis is not an etiological agent of Crohn's disease. Jayarao BM, Shreekumar P. A meta-analytic study on the detection of Mycobacterium avium subsp. paratuberculosis from patients with Crohn's Disease. [Online] Available http://www.medjournal.com/ July 26, 1999. From The Department of Veterinary Science, Pennsylvania State University, University Park, Pennsylvania. Corresponding author: Dr Bhushan Jayarao, 111, Henning Building, The Pennsylvania State University, University Park, PA 16802 - 3500. Tel: 814 - 863 - 5939. Fax: 814 - 863 - 6140. e-mail: bmj3@psu.edu INTRODUCTION In 1932, Crohn and co-workers (1) described a chronic disease in human patient's which was characterized by fever, diarrhoea, progressive anemia and loss of weight. The pathological lesions consisting of ulcerative granulomas and fistulas were mainly found in the terminal ileum (1,2). As yet, the cause of Crohn's disease remains unknown, but a multi factorial etiology including environmental, genetic, immunologic, and microbial factors has been suggested (3,4). Mycobacterium paratuberculosis, the causative organism of Johne's disease in animals has been implicated as one of the etiologic factors of Crohn's disease (2,3). Johne's disease is a chronic enteritis of ruminants. The symptoms consist of diarrhoea, progressive anemia and weight loss. The lesions are usually restricted to the ileum. The pathological lesions, as in Crohn's disease consist of granulomatous inflammation, without caseation necrosis (5). A possible association between Crohn's and Johne's disease was first described by Chiodini and co-workers (6,7) in 1984, when they isolated a Mycobacterium from 3 out of 14 patients with Crohn's disease, which was later proved to be M. paratuberculosis by DNA-DNA hybridization techniques (8). Since then, M. paratuberculosis has been detected in few cases of Crohn's disease, but the frequency of detection (3 - 15%) has been very low (9). Polymerase Chain Reaction (PCR) studies based on IS900 (hereafter referred to as IS900 PCR), a multi-copy insertion element specific for M. paratuberculosis have also produced contrasting results (2). Despite this, the interest in the association of M. paratuberculosis with Crohn's disease remains high, especially with recent reports that M. paratuberculosis may survive pasteurization (10,11,12). In the US, this is a matter of grave concern since the reported nationwide incidence for Johne's disease is 2.9% (13), with the possibility that estimates may double with newer diagnostic techniques (14). In this review we have attempted to analyze worldwide data on the detection of M. paratuberculosis in patients with Crohn's disease over the past decade in order to try and determine whether M. paratuberculosis is one of the etiological agents of Crohn's disease. METHODS The National Library of Medicine (PubMed) databases were searched for any English language study on M. paratuberculosis and/or the association of M.paratuberculosis with Crohn's disease between 1990 and 1998. In addition, the relevant references cited in the PubMed articles were reviewed also to supplement the PubMed search. A total of 70 studies on M. paratuberculosis were available and were reviewed, out of which 27 studies (15 - 41) were included for further evaluation as only these studies examined the detection of M. paratuberculosis in Crohn's disease. A study (42) which evaluated paraffin embedded sections of tissue from patients with Crohn's disease was not included, because patient numbers were not specified. Meta-analytic techniques were used to analyze the papers based on the guidelines suggested by Hunter and co-workers (43). The articles were categorized based on the year of study, geographic location (Europe, USA and Other countries including Japan, New Zealand, Kuwait and Argentina), technique used to detect M. paratuberculosis, and the number of Crohn's and other patients included in the study (Table 1). Descriptive statistical analysis was conducted using data compiled from 20 papers that employed IS900 PCR- based methods for detection of M.paratuberculosis using Epi Info version 6.0, a word processing, database, and statistical program for epidemiology on microcomputers (44). To reduce bias, confounders were detected, and the data then stratified based on year of study and geographic location. The stratified data set was analyzed to determine the association between M. paratuberculosis and Crohn's disease. RESULTS Between 1990 and 1998, there were 70 reported studies on Crohn's disease worldwide, of which 27 studies specifically explored the association of M. paratuberculosis with Crohn's disease. It was observed that 16 of 27 (59 %) of the studies did not support an etiological role for M. paratuberculosis in Crohn's disease, while 9 of 27 (33%) concluded that M. paratuberculosis was an etiological agent, and the remaining 2 (7 %) of the studies did not draw any conclusion (Table 1). Interestingly, 6 of 12 (50%) of the studies reported between 1990 - 1994 supported an etiological role for M. paratuberculosis, whereas only 2 of 15 (13%) of the studies reported between 1995 - 1998 supported an etiological role for M. paratuberculosis in Crohn's disease. The techniques used to detect M. paratuberculosis were either DNA based techniques (IS900 PCR) or immunological techniques, primarily Enzyme Linked Immuno Sorbent Assay (ELISA). Between 1990 - 1998, 20 of 27 (75 %) of the studies utilized DNA based techniques, whereas 7 of 27 (25%) utilized immunological assays for the detection of M. paratuberculosis. Between 1990 - 1994, 7 of 12 (59%) studies utilized DNA based techniques as opposed to 5 of 12 (41%) which used immunological techniques. In comparison most of the studies between 1995 - 1998 (13 of 15 studies, 87%) used DNA based techniques whereas only 2 of 15 (13%) utilized immunological techniques to detect M. paratuberculosis. Between 1990 - 1998, M. paratuberculosis was detected in 115 of 727 (16%) patients with Crohn's disease. During the same period M. paratuberculosis was detected also in 21 of 335 (6%) cases of ulcerative colitis, 16 of 155 (10%) cases of other inflammatory bowel diseases, and 32 of 603 (5 %) controls (Table 1). Between 1990 - 1994, M. paratuberculosis was detected in 95 of 373 (25%) of patients with Crohn's disease, 6 of 166 (4%) patients with ulcerative colitis, 16 of 54 (30%) patients with other inflammatory bowel diseases, and 18 of 394 (5%) control subjects. As opposed to this, between 1995 - 1998, M. paratuberculosis was detected in 20 of 354 (6%) of patients with Crohn's disease, 15 of 169 (9%) patients with ulcerative colitis, and 14 of 294 (5%) control subjects, but could not be detected in 101 patients with other inflammatory bowel diseases (Table 1). Between 1990 - 1998, 20 of 27 (74%) of the studies on Crohn's disease were reported from Europe, of which 13 (356 patients with Crohn's disease) were reported between 1990 - 1994, and the remaining 7 (259 patients with Crohn's disease) were reported between 1995 - 1998. It was observed that 7 of 14 (50%) of the studies from Europe between 1990 - 1994 supported an etiological role for M. paratuberculosis, whereas only 1 of 7 (14 %) of the studies between 1995 - 1998 supported an etiological role for M. paratuberculosis in Crohn's disease. There were 3 studies on Crohn's disease from the USA (32 patients with Crohn's disease), all of which were reported between 1995 - 1998. Only 1 of 3 (33%) of the studies supported an etiological role for M. paratuberculosis in Crohn's disease. Between 1990 - 1998, there were 4 studies (80 patients with Crohn's disease) on Crohn's disease from other countries including New Zealand, Kuwait, Japan and Argentina. Except for the study from Argentina (17 patients with Crohn's disease) which was reported in 1994, the other 3 studies were reported between 1995 - 1998. One study from New Zealand was inconclusive, but the remaining 3 (75%) studies failed to support an etiological role for M. paratuberculosis in Crohn's disease. Sub-analysis of IS900 PCR studies to detect Mycobacterium paratuberculosis in patients with Crohn's disease Studies have shown that IS 900 PCR-based detection methods have a higher sensitivity and 100% specificity for detection of the presence of M. paratuberculosis DNA in tissue samples. 2, 15, 30 With regard to this review, it was evident that nearly 75% of the investigators had employed IS 900 PCR-based assays for detection of M. paratuberculosis in Crohn's patients. These two important factors led to the selection of the data set that used IS 900 PCR-based detection method for conducting descriptive statistical analysis of the subjects included in these investigations. Crohn's disease Between 1990 and 1998, M. paratuberculosis was detected from a significantly higher number of patients with Crohn's disease studied worldwide (odds ratio 2.35:1) or from Europe (odds ratio 4.52:1) (Table 2). During the same period, there were no significant differences in the detection of M. paratuberculosis from patients with Crohn's disease from the United States or other countries (New Zealand, Kuwait, Japan). Between 1990 - 1994, it was evident that M. paratuberculosis was detected from a significantly higher number of patients with Crohn's disease worldwide (odds ratio 5.77:1) or from Europe (odds ratio 5.77:1), but during the same period no studies were reported from USA or other countries. Between 1995-1998, there were no significant differences in the detection of M. paratuberculosis from patients with Crohn's disease as compared to those with ulcerative colitis, other inflammatory bowel diseases or controls, either worldwide, or from any specific geographic location (Table 2). A plot of the annual percentage of patients with Crohn's disease detected with M. paratuberculosis showed that the worldwide pattern of detection of M. paratuberculosis closely paralleled the pattern of detection of M. paratuberculosis in patients with Crohn's disease from Europe (Fig 1). Ulcerative colitisBetween 1990 and 1998, Mycobacterium paratuberculosis was detected from a significantly lower number of patients with ulcerative colitis studied worldwide or from Europe (odds ratio 0.17). During the same period there were no significant differences in the detection of M. paratuberculosis from patients with ulcerative colitis from other countries (New Zealand, Kuwait, Japan), while the odds ratio from USA could not be calculated because there were too few patients. Between 1990 - 1994, M. paratuberculosis was detected from a significantly lower number of patients with ulcerative colitis worldwide (odds ratio 0.17), or from Europe (odds ratio 0.17), while no studies were reported from USA or other countries. Between 1995 - 1998, there were no significant differences in the detection of M. paratuberculosis from patients with ulcerative colitis worldwide, or from the group of other countries (Table 2). During the same period, M. paratuberculosis was not detected from any patient with ulcerative colitis studied from Europe, while the odds ratio for USA could not be calculated since there were too few subjects. Other inflammatory bowel diseasesBetween 1990 - 1998, M. paratuberculosis was detected from a significantly lower number of patients with other inflammatory bowel diseases (diseases other than Crohn's disease or ulcerative colitis) from the group of other countries (New Zealand, Kuwait, Japan; odds ratio 0.0). During the same period there were no significant differences in the detection of M. paratuberculosis from patients with other inflammatory bowel diseases worldwide, or from Europe, while the odds ratio for USA could not be calculated since there were too few subjects (Table 2). Between 1990 - 1994, there were no significant differences in the detection of M. paratuberculosis from patients with other inflammatory bowel diseases either worldwide, or from Europe, while no studies were reported from USA or other countries. Between 1995 - 1998, the pattern of detection of M. paratuberculosis from patients with other inflammatory bowel diseases was similar to the pattern seen between 1990 - 1998, except that during this period M. paratuberculosis was not detected from any patient with other inflammatory bowel diseases from Europe. ControlsBetween 1990 and 1998, Mycobacterium paratuberculosis was detected from a significantly lower number of healthy controls worldwide (odds ratio 0.59), or from Europe (odds ratio 0.44) or USA. During the same period, there were no significant differences in the detection of M. paratuberculosis from control subjects from other countries (New Zealand, Kuwait, Japan). Between 1990 - 1994, M. paratuberculosis was detected from a significantly lower number of healthy controls worldwide (odds ratio 0.3), or from Europe (odds ratio 0.3), while no studies were reported from USA or other countries (Table 2). Between 1995 - 1998, M. paratuberculosis was detected from a significantly lower number of healthy controls from USA. During the same period, there were no significant differences in the detection of M. paratuberculosis from control subjects worldwide or from other countries, while M. paratuberculosis was not detected from any control subject from Europe. DISCUSSIONThere have been several studies conducted over the past decade to evaluate the role of M. paratuberculosis as an etiological factor in Crohn's disease (15-42). However, conflicting results from different laboratories have confounded the issue (2,3) and the role of M. paratuberculosis in Crohn's disease continues to remain controversial. Although several excellent reviews (2,3,45) on M. paratuberculosis and Crohn's disease are available, to our knowledge, a specific attempt to collate and statistically evaluate the data over the past decade has not been made. In this review, we have gathered and analyzed all the available data from studies specifically reporting the detection of Mycobacterium paratuberculosis from patients with Crohn's disease from all over the world, between 1990 and 1998. The data suggested that a larger number of studies failed to support an etiological role for M. paratuberculosis in patients with Crohn's disease between 1990 - 1998. However, it was observed that during this period M. paratuberculosis was detected in a larger number of patients with Crohn's disease than those patients with ulcerative colitis, other inflammatory diseases or control subjects. The data showed also that M. paratuberculosis was detected from patients with Crohn's disease mainly from Europe between 1990 -1994. To further investigate the apparent geographical bias confounded by the period of study, meta-analysis was conducted on the data from the 20 available IS900 PCR based studies. The IS900 PCR based studies were selected for 2 reasons. Firstly, IS900 PCR based studies are sensitive and highly specific for the detection of M. paratuberculosis (2,15,30). Secondly, a majority (75%) of the studies between 1990 -1998 used IS900 PCR based assays for the detection of M. paratuberculosis in patients with Crohn's disease. Meta-analysis revealed that between 1990 - 1998, M. paratuberculosis was detected from a significantly higher number patients with Crohn's disease worldwide by IS900 PCR. It was evident that this was mainly because M. paratuberculosis was detected from a significantly higher number of patients with Crohn's disease from Europe, especially between 1990 - 1994. Between 1990 - 1994 there were no studies reported from the USA or other countries using IS900 PCR- based assays for comparison, but a study reported from Argentina 33 during this period that used immunological techniques failed to detect M. paratuberculosis in patients with Crohn's disease. It is not clear why M. paratuberculosis was detected in a larger number of patients with Crohn's disease from Europe. Researchers have suggested that the variability in detection of M. paratuberculosis in patients with Crohn's disease could be due to environmental or genetic factors, or an interaction between these 2 factors (3,46). Alternatively, there could be a true association of Crohn's disease with areas of high prevalence of Johne's disease, since some studies have suggested that people living in the vicinity of dairy cattle and sheep herds are at higher risk of contracting Crohn's disease (47-49). However, two (47,48) of the three studies provided circumstantial evidence only for an association between Johne's disease and Crohn's disease. The other hypothesis by Tamboli (49) which suggested that there was a similarity in the distribution of Crohn's disease and Johne's disease was based on a review of patients with Crohn's disease who were 65 or older. However, Crohn's disease affects both the very young as well as elderly people (3,45). Although survival rates of patients with Crohn's disease have increased (4), it is possible that the study by Tamboli (49) may not have included a significant proportion of the younger people with Crohn's disease. Further, the National Animal Health Monitoring system (50) Dairy' 96 study has reported that there were no significant differences in Johne's disease prevalence in the United States of America related to Geographic region. Therefore if there was an association between M.paratuberculosis and patients with Crohn's disease, there would not be any difference related to geographic location. The author (49) has failed to explain also why Sweden has a high prevalence of patients with Crohn's disease, even though Johne's disease has not been reported from Sweden. In addition, the author has suggested that heavy rains may wash off M. paratuberculosis from the environment into rivers contributing to the dissemination of M. paratuberculosis. If this were true, then regions with high rainfall and Johne's disease should have higher prevalence of Crohn's disease, which has not been reported. All these observations suggest that there is no true overlap of areas with Johne's disease and Crohn's disease. One of the major findings in our study was that although M. paratuberculosis was detected in a significantly larger number of patients with Crohn's disease between 1990 - 1994, there were no significant differences in the detection of M. paratuberculosis from patients with Crohn's disease between 1995 - 1998, either worldwide, or from Europe, USA or other countries. In fact, M. paratuberculosis was not detected from even a single patient with Crohn's disease from Europe between 1995 and 1998. If there was a true association between Johne's disease and Crohn's disease, M. paratuberculosis should have been detected in a higher number of patients with Crohn's disease during both periods. Therefore, it appears more likely that M. paratuberculosis does not play an etiological role in Crohn's disease. Investigators who support a role for M. paratuberculosis in Crohn's disease have argued that culturing the organism from tissues of Crohn's patients is extremely difficult, as is the ability to histologically demonstrate the presence of M. paratuberculosis in the affected tissues.9 Some have suggested that detection is difficult because M. paratuberculosis may exist as cell wall deficient spheroplasts in tissues. 35, 39 This argument could be used to justify the difficulty in isolating and culturing M. paratuberculosis from patients with Crohn's disease, but modern methods of PCR using sensitive and specific probes easily overcome these problems (2,15,30). The reasons for the variability in PCR results between the earlier and latter half of the decade are not fully clear, but the differences in experimental approaches, tissues used and DNA extraction procedures all could contribute to the variability in PCR results (29). Therefore, it appears that controlled longitudinal studies from different geographical locations using standardized DNA isolation and IS900 PCR protocols are needed to investigate the etiological role of M. paratuberculosis in Crohn's disease. Although the primary aim of the study was to investigate the role of M. paratuberculosis in Crohn's disease, data from other disease categories were evaluated to see whether there was some association of M. paratuberculosis with inflammatory bowel diseases. The analysis revealed that M. paratuberculosis was detected either in a significantly lower number of patients with ulcerative colitis, other inflammatory bowel diseases and control subjects, or that there were no differences between the various groups. Further, M. paratuberculosis was not reported from any of the patients with inflammatory bowel diseases other than Crohn's disease or ulcerative colitis between 1995 - 1998. Therefore, it appears that M. paratuberculosis does not play a significant role in the inflammatory bowel diseases of human subjects. The sporadic detection of M. paratuberculosis from all categories of intestinal diseases studied as well as from controls supports the concept that M. paratuberculosis may be a commensal or opportunistic pathogen along with other Mycobacteria in human beings (3). In one study (31), all 4 Crohn's patients who were positive for M. paratuberculosis, were positive by PCR for non-specific Mycobacterial DNA as well. Other studies (15,16,22) using PCR probes to non-specific Mycobacterial DNA, e.g the 16S rRNA gene have shown also the presence of Mycobacterial DNA in patients with Crohn's disease, ulcerative colitis and controls. If M. paratuberculosis is an etiological factor of Crohn's disease, then specific chemotherapy should help in reducing the symptoms. However, studies on the treatment of Crohn's patients with anti-tuberculous chemotherapy have also revealed conflicting results. Some studies have shown a beneficial effect of a combination of rifampicin, ethambutol, isoniazid, and pyrazinamide or clofazamine (51,52). On the other hand a controlled longitudinal study (53) over 5 years using rifampicin, ethambutol, and isoniazid did not show any clinical benefits at all. Even in those studies with a positive response to the anti-tuberculous therapy, there is some confusion as to whether the response was due to a specific therapeutic effect, or the result of anti-inflammatory effects characteristic of several antibiotics (2). A recent in vitro study 54 has suggested that clarithromycin and rifabutin may be the best chemotherapeutic choices for treatment of M. paratuberculosis , but in vivo tests of their efficacy in Crohn's patients need to be evaluated. CONCLUSION Our review suggests that M. paratuberculosis does not play an etiological role in Crohn's disease. Although M. paratuberculosis may be detected in patients with Crohn's disease, it appears to be a commensal or opportunistic pathogen. Further controlled longitudinal epidemiological studies of patients with Crohn's disease from different geographical locations are required to clarify the exact role of M. paratuberculosis in Crohn's disease. This review may help clinicians to advise patients with Crohn's disease regarding the probability of detection of M. paratuberculosis, and the potential benefits of undergoing anti-tuberculous chemotherapy. REFERENCES
TABLE 1 Synopsis of reviewed articles exploring the relationship between Crohn's disease and Mycobacterium paratuberculosis (1990 - 1998).
*Figures in parentheses represent the number of cases in which Mycobacterium paratuberculosis was detected; PCR denotes Polymerase chain reaction; ELISA, Enzyme linked Immunosorbent assay; Cro, Crohn's disease; UC, Ulcerative colitis; Oth, Others including Inflammatory bowel diseases other than Crohn's disease or Ulcerative colitis; Con, Controls; LB, Lymphocyte blastogenesis assay TABLE 2 The effect of Period of Isolation and Geographical location on the detection of M paratuberculosis using IS900 PCR based investigations < 1990 - 1998 > < 1990 - 1994 > < 1995 - 1998 >
IS900 PCR denotes Insertion Sequence 900 based polymerase chain reaction; CI = Confidence Interval; Other IBD = other inflammatory nflammatory bowel diseases other than Crohn's disease or Ulcerative colitis; M.Para = Mycobacterium paratuberculosis; Other countries include New Zealand, Kuwait and Japan. |
All times are GMT -8. The time now is 08:16. |
Powered by: vBulletin Version 3.0.3
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.