Relationships between defecographic findings, rectal emptying, and colonic transit time in constipated patients

Please download to get full document.

View again

of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Relationships between defecographic findings, rectal emptying, and colonic transit time in constipated patients
Document Share
Document Transcript
  Gut 1995; 36: 907-912 Relationships between defecographic findings, rectal emptying, and colonic transit time in constipated patients U Karlbom,   Pahlman, S Nilsson,   Graf Abstract The relations between defecographic findings, rectal emptying, and colonic transit time were analysed in80consti- pated patients  median age 49years, range 22-87 . Patients were classified into three clinically definedgroups slow transit, outlet obstruction, and a mixed group .Rectal evacuation was evaluated bycomputer-based areacalculation. There were no differences in defeco- graphic findingsexcept that evacuation was lessefficient in the slow transit group compared with the mixed group  p<0.01 and with the outlet obstruction group  p<0.05 . Transit time was prolonged in the slow transit andmixed groups com- pared with the outlet group  p<0.001 . Prominent impression of the puborectalis muscleduring straining and the size of a rectocoele correlatedwith rectal emptying  p<0*01 . Perineal descent, anorectalangles, enterocoele, or intussusception were not significantly related to emptying. Prominent impression of the puborectalis muscle  p<0.05 and impaired rectal emptying  p<0.05 were more frequent in patients with prolonged transit time  six or more days . There wasno significant correlation between transit time and rectal evacuation in the total study popu- lation. There was, however, an inverse relation between thesevariables  r=-0-40, p<002 when all patients who claimed infrequent defecation  two or fewer/week were analysed separately. These results did not confirm a directrelation between rectal evacuation and colonic transit time in constipated patients overall. The resultsare consistent with the suggestion that impaired colonic function may develop secondary to oudet obstruction in some patients.  Gut 1995; 36: 907-912 Keywords: constipation, defecography, rectal emptying, colonic transit time. Constipation can be defined as passing two or fewer stools per week or excessive straining at stool These symptoms may be attributed to delayed colonic transit2 or impaired rectal emptying.3 The srcin of the colonic motor defect is debatable. Histopathological changes in intramural nerve plexa compatible with nerve damage have been described in patients with longstanding constipation,4 as have altered levels of neurotransmitters.5 It is not known, however, if these changes are primary or secondary to chronicuse of laxatives. Inhibition of small6 and large7 bowel transit has been observed after induced rectal stasis. These findings suggest that slow transit consti- pation may developsecondary to impaired rectal evacuation. This hypothesis is of special interest since impaired rectal emptying is sometimes associated with an inability to relax the puborectalis muscle during evacuation,8 a condition that may be treated with biofeedback training.9 10 Consequently, it may beargued that pelvic floor retraining, or indeedany therapeutic action aiming at improving rectal evacuation, shouldprecede all other measures in slow transit constipation.   closerelationship between rectal evacua- tion and colonic transit time would favour a causual relation between these variables. The inter-relationships between structural defeco- graphic findings, rectal emptying, and colonic transit time, however, have not been systemat- ically studied. Suchknowledgewould help to define the potential benefit of therapeutic measures aiming at correcting morphological abnormalities and improving rectal emptying in the management of intractableconstipation. The aim of this study was to evaluate the relations between defecographic findings,rectal emptying, and the number of distribu-tion of markerson colonic transit examinations overall and in separate clinically defined groupsof constipated patients. Patients and methods Eighty patients 71 women 9 men, median age 49 years, range 22-87 referred to the UniversityHospital, Uppsala for investigation of constipation of a median duration of10 y rs  range 0-5-45 between January 1988 and June 1993 were included in this study. Patients with neurological illness were excluded. The stated bowelfrequency ranged between one every three weeks and fourper day  median two perweek . The patients were classified into three groups dependingon stool frequency, difficulties inrectal emptying, and general symptoms. The slow transit group  n= 10 consisted of patients with infrequentdefecation  two or fewer/week and abdominalpain or distension, orboth. The group with outlet obstruction  n=45 had normal stool frequency  more than two/week butclaimed evacuation difficulties. Patients with both infrequentdefecation and evacuation difficul- ties were classified as mixed  n=25 . Clinical Departments of Surgery and Radiology,University Hospital, Uppsala, Sweden U Karlbom LPhlman S Nilsson   Graf Correspondence to: Dr U Karlbom,Department of Surgery, Akademiska Sjukhuset, S-751 85 Uppsala,Sweden. Accepted for publication 7 October 1994 907  group.bmj.comon July 14, 2011 - Published by gut.bmj.comDownloaded from   Karlbom, Pdhlman Nilsson, Graf TABLE   Clinicalcharacteristics of thethree patient groups Slow transit Mixed Outletobstruction  n= 10)  n=25)  n =45) Age  y) 43  23-66) 46  22-82)52 31-87) Male:female 0:10 1:24 8:37 Symptoms  y) 15  2-30) 12  05-45) 9  0 5-36) Stoolslwk 1  0.35-2)ti-t 1  0 5-2)t-tt 7  2-1-20) Abdominal pain 7 22tt 26 Abdominal distension 5 18t 20 Digitation 0*t 1317 Motor stimulants/enemas 8* 21tt 19 Abdominal surgery 7 9 14Anal/transanalsurgery 1 5 9 Values are median  range) or number of patients. Abdominal surgery=boweland/or gynaecological surgery. *p<0.05, ***p<0.01  slow transit v outlet obstruction); tp<0 05, ttp<0-01, jttp<0-001  mixedv outlet obstruction); *p<005  slow transit v mixed).Kruskal Wallis and Mann-Whitney U test or Fischer s exact test. characteristics of the patients are shown in Table I Each patient underwent a colonic transit timeexamination and a defecography during the same investigation. No treatment was given between these examinations. The colonic width at the pelvic brim ranged between 2-8 and 5.9 cm in 34 patients whose double contrast enema was available forreview, and in the remaining46 patients the corresponding range was 2-7 to 5-5 cm on plain abdominal images thus ruling our mega- colon in all patients. 11 DEFECOGRAPHY Oral contrast  200 ml liquid MixobarColon 0 5 g/ml, Astra Laboratories, Sweden) was given and one hour later 250 ml semisolid barium paste  MixobarOesophagus 1 g/ml, Astra Laboratories) was instilled in the rectum. The patient was then seated on a commode, and after a short periodof rest was instructed to squeeze, and then to strain and evacuate as completely as possible. Radiographs  lateral view)weretaken during rest, squeezing, and straining. The whole investigation was docu- mented on a videotape. In measuring the anorectal angle, the axis of the anal canal and the posterior rectal wall was used. Perineal descent was defined as the posi- tion of the anorectal junction in relation to Figure 1: Distal 8 cm of rectum at rest. The boundaries of homogenous contrast is outlined. tuber os ischi at straining as compared with rest. The size of a rectocoele was calculated during straining  s thedistance from the apexof theanterior rectal wall t line drawn in the axis of theanal canal. The presenceofanenterocoele and its relation to the rectum was noted.Circumferentialintussusception was graded in four groups: -6 mm >6 mm butnotreaching theanal canal, intra-anal, and externalprolapse.   prominent impressionof the puborectalis muscleduring straining was judged as a sign of a paradoxical puborectalis contraction. All figures presented in the results section are corrected for a magnification factor of 1-7 noted by incorporating a lead marker at the centre of the commode. Rectal evacuation on coded videotapes was evaluated by computerbased area calculation. An image from the videotape was transferred to a computer withan analogue/digital con- verter  Matrox, type VIP 1024) and processedwith the Epsilonsystem  IMTEC Uppsala, Sweden) consisting of an arrayprocessor, a display processor, and 4  b imagememory. The host computer was a Motorola68030 with a 300  b data and program disc. Images from the resting phase, after first strain and the postevacuationphasewas selected for analysis. The boundariesof homogenous contrast was outlined manually - that is a thin layer of contrast covering the mucosa was not included in the regionof interest. The area of the distal 8 cm of the rectum was calculated atrest  Fig 1), after first strain, and after the total evacuation period  Fig 2). The time used for the first strain  that is maximum evacuation) as well as the total evacuation time was noted.Rectal evacuation was expressed as the per- centage evacuated area, percentage evacuated area per second during maximum evacuation, and percentage evacuated area per secondduring the total straining period  Fig 3). This was a modification of the method described by Tumbull et al 12 The three evacuation parameters were repro- ducible as evaluated by the method described by Bland and Altman 3 in 20 unselected Figure 2: The same region as shown in Figure 1 after repeated attempts of evacuation. Note that the areawith homogenous contrast has diminished. 908  group.bmj.comon July 14, 2011 - Published by gut.bmj.comDownloaded from   Defecography, rectal emptying, and colonic transit in constipation 0. Time  s) Figure 3: Graphical representation of the percentage decrease in rectal area C/A, the rate of emptying during maximum evacuation B/A(t) and the rate of emptying during the total emptying period C/Aml-. subjects. The reproducibility of the percentageevacuated area per second during the total evacuationperiod is shown in Figure 4. The reproducibility of the otherevacuationpara- meters was of   similar magnitude  date notshown). Inorder to validatefurther the method, a subjective grading wasdone inde- pendently by three investigators  1l=complete evacuation, 2=intermediate, and 3=poor). There were good correlations between the subjective grading and thearea calculation methods (r=0-51-O 66, p<0-0Ol). All defeco-graphic evaluations were performed without knowledge of the clinical characteristics or the results of the transit examination. (upper95th centile 4-7 days). For analysis of the segmental distribution of markers, the colon was divided into four main segments by lines drawn from thecentre of the promontory. Markers situated between lines through the right pelvic outlet and through the middle of the vertebral column were considered toreside in the right colon. Markersbetween lines through the vertebral column and to the left iliac crest wereregarded as located in the left colon. Those situated between lines through the leftiliac crest and the centre of the left femoral head werejudged as belonging to the sigmoid colon. Finally, markers between lines through the right pelvic outlet and the centre of left femoral head were considered as being in the rectum. This method has pr viously been used for assessingregional colonic transit.15 STATISTICAL ANALYSES Values are expressed as median  range) unless otherwise stated. Correlations were studied with the Spearman rank correlation test and expressed as an r coefficient and a p value. Kruskal Wallis and Mann-Whitney U test wereused for comparisonsbetween groups. Fisher s exact test was used to compare proportions. All p values are two tailed. Results COLONIC TRANSIT TIME The patients ingested 10 radiopaque ring- markers(diameter 3 mm daily at noon for six days. On day seven, 24 hours after the last intake of markers, a plain abdominal x ray was taken. Except for the fact that no rod markers were used, this was the same method as described by Abrahamsson et al.14 No stool softeners, laxatives, or enemas were allowed during the investigation. The x ray films were analysed with regard to the total number ofretained markers and their distribution. The gastrointestinal transit time (=colonic transit time) was calculated as previously described. 4 With this method, the median transit time for healthy men is 1-8 days (upper 95th centile 2.8 days) and for healthy women 2.8 days CO, CN 2- C1 W a E cn   CO E   ci 4 3 2 1   1 -21   K-0~0cC 0   o 3 4~~~~~~~~~~~-4 0 1 23 45 6 7 89 lo 11 12 Average measurement 1-2  /61s) Figure 4: Reproducibility of rectal evacuation measurements  percentage evacuated/second during the total evacuation period) in 20 subjects. The difference between the measurements is plotted against their mean The horizontal lines indicatethe mean  2 SD). The filled circles represent two paired measurements. RADIOLOGICALFINDINGS Some degree of intussusception was observed in 51 of 71 women (72 ) and in four of nine men (44 , p>0.20). The infolding was 6 mm or less in 11 patients, >6 mm but not reaching theanalcanal in 39, intra-anal in four, and one patient had external prolapse. An enterocoele wasnoted in 37of 71 women (52 ),and twoof nine men (22 , p>0 10). The enterocoele was situated adjacent to the rectumduring straining in nine cases. A rectocoele exceeding 2 cm was observed in 43 of the women (61 ) but in none of the men  p<0.01). In 38 of the rectocoeles, entrapment of contrast at the end of straining was noted. The rectocoele was larger in those with entrapment (median 2-7 cm, range 1-5-4-4) than in thosewith to those without (median 1.2 cm, range 0-2-7, p<0-001). Prominent imnpression of the pubo- rectalis muscle during straining was seen in seven women (10 ) and four men (44 , p<0 05). Six patients had a decreasedano- rectal angle during straining compared with at rest. The median evacuated area during the total emptying periodwas 88 range 0-100). The median rate of evacuation was 2-1 persecond(0-20) during the total period and 4.0 persecond (0-20)during maximum evacuation. For each emptying variable, men had slightly more efficient evacuation than women, but this difference was not statistically significant  data not shown). The colonic transit time was faster in men (median 2-1 days, range 1-4-5.9) than in women (median 39 days, range 0-7-6-5, p<0-05). Fiftypatients had a normal transit time - that is, withinthe 95th centilefor each 909  group.bmj.comon July 14, 2011 - Published by gut.bmj.comDownloaded from   Karlbom, Pdhlman Nilsson, Graf TABLE II Results of colonic transit and defecographic examinations inrelation to clinical classification. Values are median (range) or number of patients. Slow transit Mixed Outletobstruction (n=10) (n=25)(n=45) Colonic transit time: Transit time (d) 6-0 (3.9-6.4)*** 592 (0-7-6.5)tff 2-5(0.8-6.0)   Rightcolon 32 (22-67) 40 (0-92) 27 (0-75)   Left colon29 (14-53) 33 (0-63)* 21 (0-94)   Rectosigmoid 28 (9-55) 23 (0-58)tt 45 (0-100) Defecography: PIP 1 4 6  R strain-rest 28 (0-48) 28(-8-59)25(-17-62) Perineal descent (cm) 2-6 (18-3-5) 2-6 (0 7-4.6) 2-2 (-0 4-5.5)Rectocoele (cm) 2-4(0.5-3.3)2-4(0-4.4) 1-8 (0-3.6) Intussusception 7 16 32 Entercoele 511 23   Evacuated 70 (26-100) 81 (27-100) 95 (0-100)   Evacuated/stot 1-6 (0 5-3 6)2-1 (0.6-8 3)2-2 (0-20.0)   Evacuated/smax 1-6 (0-5.2)*tt 4-3 (0-6-8 3) 4-1 (0-20-0) ARA=anorectal angle, PIP=prominent impression of the puborectalis muscleduring straining,   evacuated/stot= evacuated per second during the total evacuation period,   evacuated/sm x= evacuated persecond during maximum evacuation. *p<0.05, ***p<0-001 (slow transit v outlet obstruction); tp<0-05, jf-p<0-01, tttp<0-001 (mixed v outletobstruction);  4p<001 (slow transit v mixed). Kruskal Wallis and Mann-Whitney U test or Fischer s exact test. TABLE III Defecographic findings in patients according to transit time Transit >6 d Transit <6 d(n=14) (n=66) PIP 5 (36)* 6 (9)  R strain-rest 18 (-8-39)* 26 (-17-62) Perineal descent (cm) 2-4 (0.7-3.8) 2-5 (-04-5.5) Rectocoele(cm) 2-5(0-4.4) 2-1 (0-3.8) Intussusception 9 (64) 46 (70) Enterocoele 7 (50) 32 (49)   Evacuated 51 (26-100)* 94 (0-100)   Evacuated/stot 1-4 (0-10)* 2-3 (0-20)   Evacuated/smax 1-7 (0-10)* 4-3 (0-20)Values are medianand range or number of patients and percentages. PIP=prominent impressionof the puborectalis muscleduring straining. ARA=anorectal angle.   evacuated/stot=-/o evacuated per second during the total evacuation period,   evacuated/sm.= evacuated per second during maximum evacuation. *p<0.05, Mann-Whitney U test or Fischer s exact test. sex. Among the 30 patients (28 women, two men) with transit time above the 95th centile, were 14 women with a transit time   6days (more than 90/O of the markers retained). RADIOLOGICAL FINDINGS IN RELATION TO CLINICAL CLASSIFICATION The frequency of intussusception, enterocoele, and paradoxical puborectalis contraction didnot differ between thethree patient groups.Neitherdid perineal descentnor change in anorectal angle (strain-rest)differ between the groups. Therewas a tendencytowards larger rectocoeles in the slow transit andmixed group compared with the outlet group (p=0.09, andp=0-06, respectively). Evacuation was less efficientin the slow transit group and there was TABLE IV Correlations between colonic transit time and rectal emptying in the three separate patient groups (Spearman rank correlation test Slow transit Mixed Outlet obstruction (n=10) (n =25) (n=45) Transit time:   Evacuated r= -0.45, p=0- 19 r=-0 *13, p=0 53 r=-O- 14 p=0 37   Evacuated/stot r=-0-20, p=0-58 r=-0 34, p=009 r=-0-15, p=0-31   Evacuatedlsmax r=-0.49, p=0-14 r=-0  12, p=056 r=-0-15, p=0-32 0 In rectosigmoid:   Evacuated r=-0.55, p=010 r=-0 39, p=0.05 r=-0.07, p=066   Evacuated/stot r= -004, p=0.90 r=-0 54, p=0.005 r=-0-13,p=0-38   Evacuatedlsmax r= -0-36, p=0-30 r= -0  45, p=0-02 r=-0-24, p=0l10 / evacuated/st,t= evacuated per second during the total evacuation period, / evacuated/smax= evacuated per second during maximum evacuation. a significant difference regarding maximum evacuation per second(Table II . The colonic transit time wasprolonged in the slow transit andmixed group compared with the outlet obstruction group (p<0001). Therewas alsoa difference in the distribution of markers between the slow transit, mixed, and outlet obstructiongroups. Inthe former two groups most markerswere retained in the right colon whereas in theoutlet obstruction group the majorityof markerswere in the rectosigmoid segment (Table II). When separate symptoms such as abdomi- nal pain and distension were analysed inrela- tion to radiologicalfindings, no difference wasfoundbetweenasymptomaticand sympto- matic patients, except thatdigitation was asso-ciated with large rectocoeles (p<005) and absence of the paradoxicalpuborectalis con- traction (p<0 01). CORRELATIONS The size of a rectocoelecorrelated with the percentageevacuated area (r= -0.45, p<0001), the rate of emptying during the total evacuation period (r=-0.32, p<001), and the rate of emptying during maximum evacuation  r=-0 33, p<001). Patients with prominent puborectalis sign evacuated less contrast than those without this sign (median 46 , range 0-100 v 79 , range17-100, p<0 004). Furthermore, the rate of emptying was lower in this groupboth during the total evacuation period (median 0-9 persecond,range 0-5 v median 2.3 per second,range 0-3-20, p<001), and during maximum evacuation (median 1.9 persecond,range0-6.5v 4.3 persecond,range0-20, p<0 01). Neither intussusception, entero- coele, perineal descent, norchange in ano- rectal angle were significantly correlated to any of the emptying variables  data notshown). There were no significant associations between transit time and intussusception, enterocoele, or rectocoele. The colonic transit time did not correlate with rectal emptying in the total population either when measured as the percentageevacuated area (r=-0.06, p>060), or when measured as the percentageper second during the whole evacuationperiod  r=-0 09, p>040) or during maximum evacuation (r=-0d12, p>0 20). The propor- tion ofmarkers in the rectosigmoid segment correlated to the rate of emptying during maximum evacuation (r=-0.30, p<0 01). When patients with a transit timeof six days or more were compared with those with a transit time less than six days, it was found that defecographic abnormalities consistent with paradoxicalpuborectaliscontraction and pooremptying were more frequent in the former group (Table III). Moreover, 10 of 14 (71 ) patients with a colonic transit time of six days or more exhibited poor rectal emptying asso- ciated with either a rectocoelewith entrapment of contrastor a prominent impression of the puborectalis muscle during straining. The above mentioned correlations between transit time, rectal emptying, and 910  group.bmj.comon July 14, 2011 - Published by gut.bmj.comDownloaded from   Defecography, rectal emptying, and colonic transitin constipation 911 defecographic findings were also analysed separately in each patient group. In the mixed group there was a correlation between thepro- portion of transit markers in the rectosigmoid segmentand all evacuationparameters Table IV). The association between the size of a rec- tocoele, prominent puborectalis sign, and evacuation was also seen in the groups. No other significant correlation was found. However, when patients who stated a defeca- tion frequencyof twice or less per week  slow transit andmixed group wereanalysed together, an inverse relation wasobserved between colonic transit time and the rate of emptying during the total evacuationperiod  r=-0.40, p<002 . Discussion The relationships betweenmorphological abnormalities, rectal emptying, and colonic transit are of theoretical interestin the inter- pretation of the pathophysiology of severe con- stipation. In addition, this question has practical consequences in the clinical manage- ment of patients. If delayed colonic transit is secondary to outlet obstruction, the logical treatment would be measures aiming at improving rectal emptying. If on the other hand, no such relation exists, these measures would be useless, and just a waste of resources and time. It has previously been claimed that paradoxical puborectaliscontraction plays an important role in the genesis of slow transit constipation.16 17 The former condition has been successfully treated with biofeedback.9 1 There was a good agreement between the clinicalclassification and the colonic transit time. In contrast, there was a poor correlation between a sense of obstruction and rectal evacuation as evaluated by defecography. Patients who claimed emptying difficulties actually had the most efficient evacuation. In theoutlet obstruction group,41 of 45 had intussusception or enterocoele, or both, and theoretically,these findings could give asensa- tion of incompleteemptying. These findings were equally common in the slow transit group, but withoutcausing a sense of obstruc- tion. This may be due to an impaired rectalsensibility in the slow transit group.18 The present study showed that a rectocoele is associated with impaired rectal emptying. This relation is not necessarily causual sinceintensestraining may resultin anenlarged ven- tral rectal  outpocketing . Another finding was that a prominent impression of the pubo- rectalis muscleduring straining was related toless efficient rectal emptying. The influence of a rectocoele and paradoxical puborectalis con- traction on emptying is controversial.   recent study found such a relationship,19 whereas it was notobserved in another.20 Our results support the formler study and suggest that these variables affect emptying. On the other hand, rectal intussusceptiondid not seem to hamper emptying. This finding is in agreement with the disappointing symptomatic results ofsurgery forrectal intussusception.2 22  e couldnotconfirm the results of Pezim et al who recently reported that a perineal descentof 1 cm or less was associated with impaired rectal emptying.23 Although no direct relationship between rectal emptying and colonic transit time  over- all or in the separate patient groups was veri- fiedin this study, we did find such arelation when theanalysis was restrictedto patients with infrequent defecation. A causual relation- ship between rectal emptyingand colonic transit is impossible to establish from the present study. Nevertheless, our findingsare consistent with that impaired rectal emptying may eventually lead to colonic dysfunction in some patients. This is alsoin agreement withthe frequent finding of paradoxical contraction of the puborectalis muscle and impairedemptying in patients with transit timeof six days or more. Actually, 10 of 14  7 1   in this group had signs of outlet obstruction. The pre- dominating abnormality was a rectocoele with entrapment of contrast at attempted defecation  five cases). An isolated impressionof the puborectalis muscle during straining was seen in two cases whereas a combination of these findings was observed in three cases. It has previously been claimed thatpelvic floor retraining is indicated in many patients with slow colonic transit.16 17 Our results suggest that some of these patients also have evacua- tion problemsdue toa rectocoele with incom- plete emptying. We observedan interesting difference according to gender. It was hardly surprising that rectocoeles were only observed in women. The predominance of men with paradoxical puborectalis contraction in this population suggests that this disorder may be propor- tionally more frequent in men with constipa- tion. In summary, the stated defecation frequencycorresponded to the colonic transit time whereas a sensation of obstructed defecation didnot relatetorectal evacuation as evaluated by defecography. The size of a rectocoele andprominent impression of the puborectalis muscle was correlated to rectal emptying. Rectal emptyingwas not directly relatedto colonic transit time in the total population.   separate analysis of patients with infrequent defecation revealed an inverse relation between the rate of rectal emptyingand colonic transit time.   paradoxical puborectalis contraction andpoor rectal emptying were more common in those with prolonged as compared withthose with normal colonic transit. These results indicate apossibilityfor the develop- ment of slow colonic transit secondary to outlet obstruction in some patients. Part of this  tu y has been presented at the International Society of University Colonand Rectal SurgeonsCongress in Singapore July 1994. 1 Drossman DA, SandlerRS, McKee DC, Lovitz AJ. Bowel patterns among subjects notseeking health care. Gastroenterology 1982; 83: 529-34. 2Preston DM, Lennard-Jones JE. Severe chronic constipa-tion of young women:  idiopathic slow transit constipa- tion . Gut 1986; 27: 41-8.3 Barnes PRH, Lennard-Jones JE. Balloon expulsion from the rectum in constipationof differenttypes. Gut 1985; 26: 1049-52.4 Krishnamurthy 5, Schuffler MD, Rohrmann CA, PopeCE.  group.bmj.comon July 14, 2011 - Published by gut.bmj.comDownloaded from 
Similar documents
View more...
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks