SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Summary

Background

Assessing the extent of ulcerative colitis determines therapeutic strategies and provides prognostic information. Colonoscopy with mucosal biopsy is considered unsafe in patients with severe disease.

Aim

To assess the correlation between proximal extent of ulcerative colitis as determined by Technitium-99-m hexamethylpropylene amine oxime labelled leucocyte scan (white cell scan) with that determined by histological assessment.

Methods

One hundred and thirty-five patients, with histologically-confirmed ulcerative colitis, who had a white cell scan and histological assessment of colonic inflammation within 6 months of each other, during the years 1991–2004, were included. Overall agreement, quadratic-weighted kappa (κ) and polychoric correlations (ρ) were calculated to estimate the inter-rater reliability.

Results

The correlation between white cell scan and histological extent was excellent (κ = 0.7 ρ = 0.8). Macroscopic appearance on colonoscopy did not correlate as well with histological extent (κ = 0.62 and ρ = 0.67). White cell scans correlated significantly better in patients with extensive disease (P = 0.02). Colonoscopy predicted disease extent more accurately in patients with limited colitis (P = 0.002).

Conclusions

Proximal extent of ulcerative colitis determined by white cell scans correlates well with histological assessment especially in patients with more extensive disease. White cell scans offer a reasonable alternative to colonoscopy with mucosal biopsies in determining the proximal extent of colitis.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

In patients with ulcerative colitis (UC), it is important to determine the proximal extent of involvement. This has implications on therapeutic strategy; in particular, first line treatment of patients with limited colitis is usually delivered topically and patients with more extensive disease would need systemic therapy.1 It also provides prognostic information, as patients with active pancolitis are more likely to need colectomy2 and are at greater long-term risk of developing colorectal cancer.3

Plain X rays of the abdomen4 and barium enema5 underestimate disease extent and contrast enemas may potentially precipitate toxic dilatation in patients with active disease.6 Colonoscopy has been considered the gold standard for determining disease extent. However, microscopic inflammation can be found in macroscopically normal colon in patients with UC.7 It has therefore been suggested that histological extent is perhaps a better indicator of the extent of colitis than macroscopic assessment alone. Histological assessment performed by serial mucosal biopsies taken during colonoscopy is invasive and carries some risk. Colonoscopy carries an increased risk of perforation1 and can potentially precipitate toxic dilatation in patients with severe colitis.8 Therefore, colonoscopy with serial mucosal biopsies to assess the proximal extent of UC is usually contraindicated in patients with severe active disease.

Prior studies utilizing white cell scans to assess the disease location involving patients with UC have evaluated small samples sizes (range: 12–52)9–17 and many studies have focused on correlation of scintigraphic scores in individual colonic segments with endoscopic or histological grading.11–13,15 Concordance between white cell scans and either histological or colonoscopic extent of inflammation in patients with active colitis has been noted to be 57% (8 of 14 patients), 67% (18 of 27 patients) and 74% (11 of 15 patients) in several small studies.9,10,12 Studies reporting only on correlation of inflammation in separate colonic segments alone have reported impressive figures of 90% in a study on 11 patients,13 and a rather poor 70% with a kappa (κ) of 0.41 in a study involving 45 patients with Crohn's disease (CD) or UC.16 A Swedish study in 25 patients with UC noted a κ of 0.19 for correlation between colonoscopy and white cell scan and white cell scans underestimated the extent of active inflammation.11 A French study on 31 patients with acute UC noted that white cell scan-based extent correlated well both with histological extent (r = 0.85) and endoscopic extent (r = 0.83) with agreement between histological and white cell scan-based extent in 73% of cases.14 In the largest study to date on 52 patients (28 UC and 24 CD), with active disease, agreement on disease extent between histology and white cells scans using Indium111 labelled granulocytes was noted in 62% with a correlation coefficient was 0.9 between the two modalities.17 We used Technitium-99-m hexamethylpropylene amine oxime instead of Indium111 for labelling leucocytes. The formed technique has superseded the latter as it reduces radiation dose, has a simpler labeling procedure, use a more readily available isotope and improves image quality.18

Use of white cell scans has been extensively evaluated and found to be useful in assessment of inflammatory bowel disease activity and response to therapy,18, but its value in accurately determining the proximal extent of involvement of UC is less clear with mixed results noted in prior small studies.9–17 The aim of this study was to determine, using a retrospective cohort, whether white cell scans can be used as alternative to histological evaluation in assessing proximal extent of UC in routine clinical practice.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Subjects

All adult patients with a diagnosis of UC made on histology between 1991 and 2004 were identified from the histopathology database at St George's Hospital, London. Patients from this group who had undergone a white cell scan and either colonoscopy with serial mucosal biopsies or total colectomy within 6 months of each other were then identified from the electronic patient records. Age, sex, C reactive protein (CRP) within 2 weeks of white cell scan, proximal extent of UC as defined by histology, colonoscopy and white cell scan were collated from the patient records.

Definition of disease extent

Disease extent was divided into (1) proctitis: colitis confined to the rectum (2) proctosigmoiditis: involving rectum and sigmoid colon, (3) left-sided colitis: up to the splenic flexure, (4) extensive colitis: up to the hepatic flexure, and (5) pancolitis: affecting the whole colon. This is referred to as ‘expanded extent’ throughout this paper. We also assessed a broader and clinically useful division of extent1 into (a) ‘distal colitis’ defined as disease up to the sigmoid descending junction (b) ‘left sided’ defined as disease extending proximal to the sigmoid/descending colon junction up to the splenic flexure and (c) ‘extensive’ UC as extending proximal to the splenic flexure. The definition is referred to as ‘clinically useful extent’ throughout this paper.

Statistical analysis

Overall agreement, quadratic-weighted κ19 and polychoric correlation (ρ)20 were calculated to estimate the inter-rater (extent based on histology, white cell scan and colonoscopy) reliability of the ordered categorical rating of proximal extent of UC. Weighted κ statistics were calculated as extent assessment was considered ordinal.19

We also calculated the polychoric correlation (ρ) by using R for windows version 2.2.1. We assumed that the distribution of extent in the population studied was normal on a continuous scale.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Of the 751 patients with histologically proven UC screened, 135 patients [78 (58%) male and 57 (42%) female] with a mean age of 41 (±16) years met the inclusion criteria. Histological assessment was based on the colectomy in 40 (30%) patients and by serial mucosal biopsies taken during colonoscopy in the rest. One hundred and one patients had a colonoscopic assessment of disease extent and 34 had total colectomy without having a colonoscopy.

The inter-rater agreement between white scan and histology, colonoscopy and histology and white cell scan and colonoscopy is given Table 1. Correlation with histological assessment was better for white cell scans rather than colonoscopy for both the expanded and clinically useful definitions of disease extent. White cell scans and colonoscopy correlated poorly in assessing proximal extent of disease.

Table 1.   Overall agreement, quadratic-weighted kappa and polychoric correlation among white cell scans, histology and colonoscopy for assessment of proximal extent of disease in patients with ulcerative colitis
Correlation between groupsNo. of patientsExpanded definition of extentClinically useful definition of extent
Overall agreement %Quadratic- weighted kappa (S.E.)Polychoric correlation (S.E.)Overall agreementQuadratic- weighted kappa (S.E.)Polychoric correlation (S.E.)
White cell scan and histology135650.72 (0.1)0.82 (0.03)760.7 (0.1)0.84 (0.03)
Colonoscopy and histology101740.62 (0.17)0.68 (0.06)780.56 (0.17)0.67 (0.06)
White cell scan and colonoscopy101490.54 (0.13)0.6 (0.07)560.47 (0.12)0.57 (0.08)

The proximal extent based on the white cell scans and histology did not correlate in 47 (35%) patients. Of these 14 had normal white cell scans. The CRP in 10 of these negative white cell scans was below the upper limit of the reference range (8 mg/dL) and was not available in the other 4. The CRP was available in 25 of the 47 uncorrelated cases of which 14 had normal values and in 53 of the 88 correlated cases of which 22 had normal values. Raised CRP values (more than 8 mg/dL) were not significantly associated with better correlation between histological and white cell scan-based proximal extent of disease (P = 0.34). There was no significant difference between the mean CRP values in cases that correlated (24 mg/dL, S.E.M. 4) and for cases that did not correlate (16 mg/dL, S.E.M. 4) (P = 0.23).

Table 2 lists the agreement, overestimation and underestimation of disease extent between white cell scans or colonoscopy compared with histological extent. When assessing patients with extensive colitis white cell scans did significantly (P = 0.02) better than colonoscopy. Conversely, for patients with left sided colitis and proctosigmoiditis, colonoscopy-based extent performed significantly (P = 0.002) better than white cell scans.

Table 2.   Agreement, overestimation and underestimation of disease extent between white cell scans and colonoscopy compared with histological extent
 Total numberAgreement (%)Overestimation (%)Underestimation (%)
  1. N/A, not applicable.

White cell scan-based extent
 Proctosigmoiditis2211 (50)2 (9)9 (41)
 Left sided3119 (61)5 (16)7 (23)
 Extensive8373 (88)N/A10 (12)
Colonoscopic extent
 Proctosigmoiditis2219 (86)3 (14)0
 Left sided2924 (83)1 (3)4 (14)
 Extensive5036 (72)N/A14 (28)

Additionally of the 40 patients who had a colectomy, 37 (93%) had extensive disease and overall agreement between white cell scan and histological extent was 88%.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

In this large retrospective study, we have demonstrated very good correlation between white cell scans and histological assessment of the proximal extent of involvement in patients with UC. Both the κ statistic and the polychoric correlation (ρ) suggested substantial agreement between the two modalities. White cell scans performed better in assessing the patients with more extensive colitis and conversely colonoscopy fared better in patients with more limited disease.

Our results are similar to previous reports in selected patients with acute colitis14,17 This study is unique in having chosen unselected patients with UC and not patients with active disease colitis and is the largest cohort.

Although overall agreement with histological assessment was marginally better for colonoscopy compared with white cell scans, the κ statistics showed better correlation for white cell scans. This was probably because of higher prevalence of widely varying results in assessment of extent between colonoscopy and histology. As disease extent assessment was considered as an ordinal parameter, we used a quadratic-weighted κ to assess the degree of disagreement. Quadratic weighting emphasizes the size of disagreements as differences are squared. There are several criticisms of the κ statistic including the fact that it is influenced by trait prevalence (distribution) and base-rates and hence it is seldom comparable across studies, procedures or populations.20,21 We therefore also calculated the polychoric correlation (ρ), which addresses some of these concerns.20,22 Polychoric correlation confirmed the good agreement between white cell scans and histological assessment.

As this was a retrospective study, there are several potential biases in the design. By limiting our subjects to those who had either colonoscopy with serial mucosal biopsies or colectomy to determine histological extent, within 6 months of the leucocyte scan, the proportion of patients with more extensive disease is higher in our group. Patients with more limited colitis are less likely to have a full colonoscopy and also have a lower risk for surgical intervention2 and hence were a smaller proportion in our study population. Secondly, colonoscopy, histological assessment and interpretation of the white cell scans were done by several different observers over this time period which may have led to inter-observer variation. Moreover, by including patients with relatively quiescent disease and normal CRP values and avoiding a second look interpretation of the white cell scans, we have aimed to parallel actual clinical practice.

Disease extent in UC is not static and having an interval of up to 6 months between the colonoscopy and white cell scanning could have affected the results. A Scandinavian study showed that in 1 year, there was histological progression in 20%, regression in 24% and normal histological findings in 24%.23 More recently, it has been shown that histological distribution between consecutive colonoscopies at least 1 year apart remained the same in 61% of patients.24 Thus, changes in disease distribution, in the period intervening the two modalities used for assessing disease extent in our study, may potentially underestimate correlation. This is corroborated by the finding that among the 40 patients with histological assessment of disease extent made after colectomy, in whom white cell scans were performed between 1 and 14 days prior to surgery, there was agreement in disease extent based on the histology and white cell scans in 35 (88%) patients. We used CRP levels as a surrogate marker of disease activity. CRP levels have not been shown to correlate well with histological inflammation in patients with UC.25 Moreover, a third of patients with chronic quiescent UC over 12 months still have an acute inflammatory cell infiltrate identified on rectal biopsy.26

In conclusion, white cell scans offer a reliable alternative to histological assessment of disease extent in patients with UC and is most accurate in assessing the extensive disease. Previous studies indicate that white cell scans also provide a semi-quantitative estimate of disease activity in UC. 18 Thus, useful information about extent and activity of disease can be obtained from a single investigation. White cell scans are therefore particularly valuable in assessing the disease extent in patients with acute severe colitis in whom colonoscopy is usually contraindicated. In tandem with white cell scans, a limited flexible sigmoidoscopy with biopsies to confirm the histological diagnosis of UC may also be prudent. A recent population-based study from Denmark suggests that only 27% of patients with UC have extensive disease. 27 White cell scans seem to be most accurate in those who have histological inflammation extending beyond the splenic flexure. As our study shows, extensive disease forms the majority of patients in a hospital setting and white cell scans would be effective in this group. In the out-patient setting with smaller proportion of patients with extensive disease, a cost-effective approach would be to perform initially a flexible sigmoidoscopy and follow that up with a white cell scan in those with histological inflammation up to and including the most proximal point of biopsy. Additionally, this technique also offers a reasonable alternative in assessment of disease extent, in patients with relatively quiescent disease whether or not CRP levels are raised and for patients who do not tolerate or decline colonoscopy.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Carter MJ, Lobo AJ, Travis SP. IBD Section, British Society of Gastroenterology Guidelines for the management of inflammatory bowel disease in adults. Gut 2004; 53: V116.
  • 2
    Moum B, Ekbom A, Vatn MH, et al. Clinical course during the 1st year after diagnosis in ulcerative colitis and Crohn's disease. Results of a large, prospective population-based study in southeastern Norway, 1990–93. Scand J Gastroenterol 1997; 32: 100512.
  • 3
    Venkataraman S, Mohan V, Ramakrishna BS, et al. Risk of colorectal cancer in ulcerative colitis in India. J Gastroenterol Hepatol 2005; 20: 7059.
  • 4
    Langmead L, Rampton DS. Plain abdominal radiographic features are not reliable markers of disease extent in active ulcerative colitis. Am J Gastroenterol 2002; 97: 3549.
    Direct Link:
  • 5
    Elliott PR, Williams CB, Lennard-Jones JE, et al. Colonoscopic diagnosis of minimal change colitis in patients with a normal sigmoidoscopy and normal air-contrast barium enema. Lancet 1982; 1: 6501.
  • 6
    Koudahl G., Kristensen M. Toxic megacolon in ulcerative colitis. Scand J Gastroenterol 1975; 10: 41721.
  • 7
    Mutinga ML, Odze RD, Wang HH, et al. The clinical significance of right-sided colonic inflammation in patients with left-sided chronic ulcerative colitis. Inflamm Bowel Dis 2004; 10: 2159.
  • 8
    Norland CC, Kirsner JB. Toxic dilatation of colon (toxic megacolon): etiology, treatment and prognosis in 42 patients. Medicine 1969; 48: 22950.
  • 9
    Stahlberg D, Veress B, Mare K, et al. Leukocyte migration in acute colonic inflammatory bowel disease: comparison of histological assessment and Tc-99 m-HMPAO labeled leukocyte scan. Am J Gastroenterol 1997; 92: 2838.
  • 10
    Stahlberg D, Mare K, Granqvist S, et al. Instant technetium 99 m hexamethyl propylenamine oxime-labeled leukocyte scan compared with colonoscopy in early assessment of disease extent and activity in acute colitis. Dis Colon Rectum 1996; 39: 114652.
  • 11
    Almer S, Franzen L, Peters AM, et al. Do technetium-99 m hexamethylpropylene amine oxime-labeled leukocytes truly reflect the mucosal inflammation in patients with ulcerative colitis? Scand J Gastroenterol 1992; 27: 10318.
  • 12
    Middleton SJ, Li D, Wharton S, et al. Validation of 99Tcm-HMPAO leucocyte scintigraphy in ulcerative colitis by comparison with histology. Br J Radiol 1995; 68: 10616.
  • 13
    Vilien M, Nielsen SL, Jorgensen M, et al. Leucocyte scintigraphy to localize inflammatory activity in ulcerative colitis and Crohn's disease. Scand J Gastroenterol 1992; 27: 5826.
  • 14
    Guimbaud R, Beades E, Chauvelot-Moachon L, et al. Technetium Tc 99 m hexamethyl propylene amine oxine leukocyte scintigraphy in patients with ulcerative colitis: correlation with clinical, biologic, endoscopic, and pathologic intensity, and local release of interleukin 8. Gastrointest Endosc 1998; 48: 4916.
  • 15
    Spinelli F, Milella M, Sara R, et al. The 99 mTc-HMPAO leukocyte scan: an alternative to radiology and endoscopy in evaluating the extent and the activity of inflammatory bowel disease. J Nucl Biol Med 1991; 35: 827.
  • 16
    Madsen SM, Thomsen HS, Munkholm P, et al. Inflammatory bowel disease evaluated by low-field magnetic resonance imaging. Comparison with endoscopy, 99 mTc-HMPAO leucocyte scintigraphy, conventional radiography and surgery. Scand J Gastroenterol 2002; 37: 30716.
  • 17
    Saverymuttu SH, Camilleri M, Rees H, Lavender JP, Hodgson HJ, Chadwick VS. Indium 111-granulocyte scanning in the assessment of disease extent and disease activity in inflammatory bowel disease. A comparison with colonoscopy, histology, and fecal indium 111-granulocyte excretion. Gastroenterology 1986; 90: 11218.
  • 18
    Weldon MJ, Lowe C, Joseph AE, et al. Review article: quantitative leucocyte scanning in the assessment of inflammatory bowel disease activity and its response to therapy. Aliment Pharmacol Ther 1996; 10: 12332.
  • 19
    Fleiss JL, Cohen, J. The equivalence of weighted kappa and the intraclass correlation coefficient as measures of reliability. Educ Psychol Meas 1973; 33: 6139
  • 20
    Hutchinson TP. Focus on psychometrics. Kappa muddles together two sources of disagreement: tetrachoric correlation is preferable. Res Nurs Health 1993; 16: 3136.
  • 21
    Feinstein AR. Cicchetti DV. High agreement but low kappa: I. The problems of two paradoxes. J Clin Epidemiol 1990; 43: 5439.
  • 22
    Hutchinson TP. Agreement between experts: an answer, but is it the answer for you? J Clin Neurosci 2003; 10: 2135.
  • 23
    Moum B, Ekbom A, Vatn MH, Elgjo K. Change in the extent of colonoscopic and histological involvement in ulcerative colitis over time. Am J Gastroenterol 1999; 94: 15649.
    Direct Link:
  • 24
    Picco MF, Krishna M, Cangemi JR, Shelton D. Oral mesalamine and clinical remission are associated with a decrease in the extent of long-standing ulcerative colitis. Inflamm Bowel Dis 2006; 12: 53742.
  • 25
    Riley SA, Mani V, Goodman MJ, et al. Microscopic activity in ulcerative colitis: what does it mean? Gut 1991; 32: 1748.
  • 26
    Solem CA, Loftus Jr EV, Tremaine WJ, et al. Correlation of C-reactive protein with clinical, endoscopic, histologic, and radiographic activity in inflammatory bowel disease. Inflamm Bowel Dis 2005; 11: 70712.
  • 27
    Vind I, Riis L, Jess T, et al. Increasing incidences of inflammatory bowel disease and decreasing surgery rates in Copenhagen City and County, 2003–2005: a population-based study from the Danish Crohn colitis database. Am J Gastroenterol 2006; 101: 127482.
    Direct Link: