Response to Corticosteroids in Severe Ulcerative Colitis - Clinical ...

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2007;5:103–110

Response to Corticosteroids in Severe Ulcerative Colitis: A Systematic Review of the Literature and a Meta-Regression DAN TURNER,* CATHARINE M. WALSH,* A. HILLARY STEINHART,‡ and ANNE M. GRIFFITHS* *Division of Gastroenterology, Hepatology and Nutrition, the Hospital for Sick Children, and ‡Division of Gastroenterology, Mount Sinai Hospital, Toronto, Canada

See Colombel, J–F et al on page 52; Kamm MA et al on page 66; and Schreiber S et al on page 76 for companion articles in Gastroenterology. Background & Aims: Colectomy is a potentially lifesaving procedure for patients with severe attacks of UC who fail medical therapy. We aimed to systematically review studies that reported the short-term colectomy rate in severe UC or reported variables that could predict treatment failure. Methods: We conducted a systematic literature search for cohort studies and controlled trials published between 1974 –2006. Results: Thirty-two studies met the inclusion criteria; 16 reported short-term outcome and predictors of therapy failure, 13 only outcome, and 3 only predictors. In the pooled analysis, 581 of 1991 patients required colectomy (weighted mean 27; 95% confidence interval [CI], 26%–28%), and 22 died (1%; 95% CI, 0.7%–1.5%). In a heterogeneity-controlled meta-regression, colectomy rate did not change during the last 30 years (R2 ⴝ 0.07, P ⴝ .8). Cyclosporine was used in only 100 patients, with a 51% (95% CI, 41%– 60%) short-term success rate. A second metaregression failed to demonstrate a dose-colectomy response of methylprednisolone therapy beyond 60 mg daily (R2 < 0.01, P ⴝ .98). More than 20 variables were identified in 19 studies to predict medical therapy failure, but only a few were consistently reproduced: disease extent, stool frequency, temperature, heart rate, C-reactive protein, albumin, and radiologic assessment. Conclusions: The shortterm colectomy rate in severe UC has remained stable during the last 30 years, despite the introduction of cyclosporine, which was not used frequently. We could not find any support for administering methylprednisolone at a higher dose than 60 mg/day. Variables that predict outcome of corticosteroid therapy could aid in the development of guidelines for introduction of rescue therapies in severe UC.

I

nflammation in UC is localized to the colonic mucosa and extends from the rectum proximally in varying extents. The lifetime risk of a severe exacerbation requiring hospitalization is estimated to be 15%.1 Patients with extensive UC (ie, macroscopic disease proximal to splenic flexure) are more likely to develop acute severe colitis. Extensive colitis is, in turn, significantly more common in pediatric-onset versus adult-onset disease (50%– 80% vs 8%–20%).2,3 In 1955, Truelove et al4 published the first controlled trial of corticosteroids in the treatment of UC and in 1974 established

intravenous corticosteroids (IVCS) as the mainstay of treatment in severe exacerbations.5 Addition of calcineurin inhibitors was reported to avoid colectomy in some patients with corticosteroid-refractory UC, but their significant toxicity precludes longterm use.1,6 When used as a bridge to azathioprine maintenance therapy, it is estimated that 30%–50% of patients with steroidrefractory UC remain in clinical remission at the end of 1 year of follow-up.1,6,7 Recently, infliximab has been used in the randomized controlled trial setting to induce remission in adults with steroid-refractory severe UC,8 as well as to induce and maintain remission in ambulatory adults with steroid-dependent UC.9 Some studies in adults have attempted to identify variables that predict response to IVCS therapy. Reliable predictors of response would avoid ineffective continuation of corticosteroids and facilitate a timelier introduction of second-line therapies. As new biologic agents become available,10 it is important to develop guidelines for their use on the basis of predictive clinical and laboratory variables. The short-term colectomy rates in patients who fail intensive medical therapy are variably reported among studies, but no attempt has hitherto been made to systematically review trends in colectomy rates. We hypothesized that the introduction of cyclosporine as second-line therapy and the general advance in medical care would result in a decline in the need for colectomy. Therefore, we aimed to perform a systematic review of the literature to identify studies describing the short-term outcome of adult and pediatric patients treated with IVCS, with or without cyclosporine, for exacerbations of UC. A second aim was to identify studies designed to determine variables predictive of therapeutic response or failure. This study is reported according to the MOOSE guidelines (Meta-analysis Of Observational Studies in Epidemiology).11

Methods Eligibility Criteria Retrospective and prospective studies evaluating adult or pediatric patients with UC admitted for first or subsequent exacerbation, severe enough to require IVCS therapy, were included if the short-term outcome and/or analysis of predictors of response were reported. We searched for studies published Abbreviations used in this paper: CI, confidence interval; CRP, Creactive protein; IVCS, intravenous corticosteroids; MHC, major histocompatibility complex; PPV, positive predictive value; UK, United Kingdom. © 2007 by the AGA Institute 1542-3565/07/$32.00 doi:10.1016/j.cgh.2006.09.033

104

TURNER ET AL

since 1974, when IVCS were established as the mainstay of therapy for severe UC attacks.5 “Short-term” was defined as any time during the same admission, but the timing of colectomy after the initiation of IVCS was recorded to control for possible heterogeneity related to duration of therapy. Studies that evaluated variables predictive of response to therapy were included if a systematic evaluation of numerous parameters used in clinical practice was reported, to allow for multivariable assessment. For the predictive studies, all end points were allowed, providing that they were clearly defined (eg, colectomy, clinical response, and death). Because our aim was to describe outcome and predictive variables in clinical practice with identifiable patient populations, only full cohorts receiving IVCS were included. Studies that reported subgroups (eg, cohorts of patients receiving cyclosporine, undergoing surgery, or being diagnosed with fulminant colitis), without describing the full cohort of patients admitted for IVCS therapy, were excluded.

Assessing Quality of Studies There is no consensus of how to assess the quality of observational studies.11 Comparative studies can be assessed by using the same criteria as for clinical trials (eg, concealment of allocation, randomization, selection bias of the comparable groups, and blinding of outcome).11,12 However, this does not apply to observational studies lacking a concurrent control group. For this review, therefore, studies were excluded on the basis of completeness of data; required data were colectomy rate (the primary end point), method of steroid administration (the major inclusion criteria), and timing of outcome evaluation (critical in assessing heterogeneity).

Search Strategy Two investigators (D.T., C.W.) independently performed a database literature search for human studies published between January 1, 1974 –January 31, 2006 without language restrictions. MEDLINE (NLM, National Library of Medicine, Bethesda, MD; 1966 –2006) and EMBASE (Elsevier, New York, NY; 1980 –2006) on OVID were searched by using the following terms: ulcerative colitis (MeSH, explode all and also as free text) AND (inpatient$ (text) OR admit$ (text) OR admission (text) OR hospitalized (text) OR predict (text)). A similar search with key words was performed by using PubMed (NLM). The search for interventional studies was supplemented by using a similar search strategy on Cochrane Central Register of Controlled Trials (The Cochrane Collaboration, United Kingdom [UK]; searched 2nd quarter 2006) and on the American College of Physicians journal club. References of all identified studies were manually searched for further relevant studies. The manual reference search was expanded to include review articles on (acute OR severe) AND (ulcerative colitis) and from personal library. Where the 2 investigators disagreed on the eligibility of an article, a senior investigator (A.M.G.) was consulted, and decision was made by mutual agreement.

Assessment of Heterogeneity The established statistical methods to explore heterogeneity (ie, Cochran Q test and I2) cannot be implemented on non-comparative data. Thus, heterogeneity among the studies was addressed in several other ways. First, clinically important factors that might explain study heterogeneity were explored. Next, sensitivity analysis was performed on the basis of these

CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 5, No. 1

factors, and finally, we controlled for the important identified factors in the meta-regression.13

Statistical Analysis Statistical analysis was performed by using SPSS for Windows V12.0 (SPSS Inc, Chicago, IL) and SAS V9.1 (SAS Institute, Inc, Cary, NC). Dichotomous variables were compared by using ␹2 or Fisher exact tests, accompanied by odds ratios (95% confidence interval [CI]), as appropriate. A meta-regression was performed to explore the relationship between colectomy rates with time and separately with IVCS dosing, while controlling for possible causes of heterogeneity. Studies were weighted by using the inverse variance method. The regression line was verified by plotting the predicted values of the full model against colectomy rate. Ninety-five percent CI was calculated for all pooled point estimates. Statistical significance was considered to have been reached if the calculated 2-tailed P value was less than .05.

Results Search Result and Study Characteristics The search strategy yielded 918 non-duplicated articles. Fifty-seven studies that were potentially eligible, on the basis of abstract review, were retrieved. Twenty-five studies were excluded; 9 included subpopulation cohorts only, 6 used predominantly oral steroids, 3 reported only the long-term outcome, and 7 were of poor quality as defined above. The remaining 32 studies were included; 16 reported both the short-term outcome and predictors of therapy failure, 13 reported only the outcome, and 3 reported only the predictors without sufficient data to report the outcome (Tables 1 and 2). Of the 29 outcome studies, 14 were retrospective cohort studies, 6 were prospective cohort studies, 5 were randomized controlled trials, and 4 did not specify. Severity was assessed by using the classification of Truelove and Witts4 in 20 studies, the classification of Lichtiger et al14 in 3 studies, the index of Seo et al15 in 1 study, novel individual definitions in 4, and was not reported in 4 studies. Among the studies that used the classification of Truelove and Witts, 12 required the fulfillment of all of the 5 items, and 8 applied a more liberal definition. Of the predictive studies, 11 used colectomy as the end point measure,16 –27 one used clinical improvement,28 and one used number of days to improvement.20 Composite end point of colectomy or death29,30 was used in 2 studies, colectomy or the need for further medical therapy (ie, calcineurin inhibitors or infliximab) was used in 2 studies,31,32 and a combined clinical and endoscopic score in 1 study.33

Assessing Heterogeneity and Subgroup Analysis The year of publication, IVCS dose, the time at which need for colectomy was evaluated, cyclosporine use, inclusion of moderately active patients versus only severe patients, and the publication type (clinical trial/observational) were identified as the clinically important factors that could have contributed to heterogeneity. Sensitivity analyses were performed on the basis of these factors. Colectomy rate was similar in the 5 interventional studies (44/149, 30% [95% CI, 23%–37%]) and the other descriptive

January 2007

CORTICOSTEROIDS IN SEVERE UC

105

Table 1. Short-term Outcome of Medical Treatment in Hospitalized UC During the Corticosteroid Era

Year (ref) Adults 1974 1975 1977 1978 1979 1985 1986 1987 1990 1993 1994 1995 1996 1998 2000 2000 2000 2001 2001 2002 2002 2004

R/P/ RC

n

Days to evaluation

Disease severity

(5) (18) (50)a (51) (52)b (53)c (54)d (55) (20)e (21)f (56) (17) (19) (16) (30)a (33)

NS R R NS P NS RC R R NS RC P P R R R

49 189 23 100 179 79 27 114 26 42 39 30 51 97 85 71

60 DA 5 5 DA 5 DA DA DA DA 10 20 DA 30 30 7–14

(57) (58) (59) (60) (23) (32)

RC RC RC P R R

13 55 15 45 127 167

60 10 8 21 DA 30

Severe NS Severe Severe NS Severe Severe Mod ⫹ severe Severe Severe Severe Severe Severe Mod ⫹ severe Severe Mod-severe ⫹ severe Mod ⫹ severe Severe Severe Severe Mod ⫹ severe Severe

P P R R

67 55 149 54

30 7 21 5

Severe Severe Severe Severe

R R

13 10

60 36

Severe Severe

R

20

21

Severe

2004 (25) 2004 (29) 2004 (31)g 2005 (22) Pediatric 1977 (61) 1995 (34)

2002 (62) Sum Adults, n (%, 95% CI) Pediatric, n (%, 95% CI)

Daily dose of IVCS

MP 60 mg NS HC 240 mg MP 60 mg NS MP 60 mg MP 60 mg NS MP 40–100mg NS HC 400 mg HC 400 mg HC 400 mg BM 8–16mg MP 0.8 mg/kg MP 1–1.5mg/kg or pulse 1 g once MP 0.75–1 mg/kg HC 400 mg MP 40 mg HC 400 mg MP 40–60 mg MP 60 mg or HC 400 mg MP 1 mg/kg HC (dose NS) MP 1 mg/kg HC 200–400 mg HC 10 mg/kg MP 1.7 mg/kg, max 67 mg (or equivalent HC) MP 2 mg/kg, max 60 mg

1948 43

N (%) colectomy

N (%) response to IVCS

N (% of total cohort) receiving cyclosporine (n responded)

N (%) death

13 (27) 34 (18) 11 (48) 25 (25) 35 (20) 32 (41) 14 (52) 28 (25) 6 (23) 21 (50) 14 (36) 12 (40) 15 (29) 33 (34) 37 (44) 13 (18)

36 (73) 144 (76) 12 (52) 75 (75) 143 (80) 45 (56) 13 (48) 86 (75) 20 (77) 21 (50) 25 (64) 18 (60) 29 (57) 64 (66) 41 (48) 58 (82)

0 0 0 0 0 0 0 0 0 NS 0 0 14 (27) (4) 0 13 (15) (6) 0

0 0 5 (2.7) 2 (2.5) 2 (7.4) 0 0 NS 0 0 0 0 1 (1) 0

1 (8) 12 (22) 3 (20) 20 (44) 33 (26) 68 (40)

11 (85) 43 (78) 8 (53) 23 (51) 88 (69) 88 (53)

2 (15) (1) 0 7 (47) (4) 0 NS 21 (13) (9)

0 0 0 2 (4.4) 0 2 (1.2)

14 (20) 8 (15) 24 (16) 39 (72)

34 (50) 45 (82) 112 (75) 15 (28)

25 (37) (19) 3 (5) (0) 15 (10) (8) 0

0 1 (0.5) 0 3 (5.6)

6 (46) 1 (10)

7 (54) 9 (90)

9 (45)

11 (55)

565 (27, 26– 29)h 16 (37, 24–52)

0 0

2 (10) (1)

1429 (67, 65–69)

100 (5, 4–6) (51)

27 (63, 48–76)

2 (5, 1–15) (1)

0 4 (2)

0 0

0

22 (1, 0.7–1.6) 0

NOTE. Days to evaluation represents the time frame by which colectomy was assessed after IVCS therapy. Only reference 31 included patients who received infliximab; 6 patients (4%) were treated with infliximab, of whom 4 responded. NS, not stated; R, retrospective; P, prospective; RC, randomized controlled study; DA, during admission (time frame not reported); MP, methylprednisolone; HC, hydrocortisone; BM, betamethasone. aData represent only the fraction of patients who were treated with IVCS. bData represent only the fraction of UC patients. cData represent only the fraction of the patients with severe colitis and the short-term follow-up period. dData represent only UC patients of the cohort. Patients who required immediate colectomy were excluded in the study. eFew patients who did not tolerate IVCS were excluded in the study. fData represent only the fraction of the patients with severe exacerbation. gConservative approach was taken. Data represent only the short-term follow-up. hWeighted mean.

cohorts (537/1842, 29% [95% CI, 27%–31%]; P ⫽ .9). Thus, the 2 groups were pooled in the analysis. Colectomy rate was slightly higher in studies in which the need for colectomy was evaluated within 2 weeks (157/451, 32% [95% CI, 28%–36%]) as compared with those in which it was evaluated after 2 weeks of IVCS (424/1500, 28% [95% CI, 26%– 30%]), but this did not reach significance (P ⫽ .13, odds ratio, 1.2 [95% CI, 0.95–1.5]). Nonetheless, because this variable seems

clinically important in assessing heterogeneity, it was forced into the final model. Colectomy rate was higher in studies that included only severe patients (404/1201, 34% [95% CI, 31%–36%]) compared with studies that included also moderately active patients (108/ 422, 26% [95% CI, 22%–30%]; P ⫽ .003, odds ratio, 1.47 (95% CI, 1.1–1.9). Thus this variable was included in the final model to control for heterogeneity.

106 TURNER ET AL

Table 2. Predictors of IVCS Failure in Severe UC: Summary of Systematic Evaluations Lindgren Gulati Lennard-Jones Travis Oshitani Chakravarty Seo Carbonnel Truelove Ho Choquet Oshitani Benazzato Kumar Elloumi Daperno Spicer Spicer Meyers (16) (17) (18) (19) (20) (21) (23) (30) (5) (32) (24) (33) (25) (29) (22) (31) (26) (27) (28) 3

1

1

3

⫹

1

1

⫹

⫹

1

1

⫹ ⫹

⫹ ⫹ ⫹

⫹ ⫹ ⫹

⫹

⫹

1–3

1

?

1

⫹ ⫹

⫹

⫹ ⫹

⫹ ⫹

⫹ ⫹

⫹

⫹

⫹

ESR, erythrocyte sedimentation rate; WBC, white blood cells.

⫹ ⫹

1

1

1

⫹

⫹

⫹

⫹ ⫹

⫹

⫹ ⫹

⫹ ⫹

⫹

⫹

⫹

⫹ ⫹

⫹

⫹ ⫹ ⫹

⫹ ⫹ ⫹

⫹ ⫹ ⫹

⫹

⫹

⫹ ⫹

⫹

1

⫹

⫹

⫹ ⫹ ⫹

3

⫹

⫹

⫹

1-2

⫹ ⫹

⫹

1

⫹

⫹ ⫹

⫹

⫹

⫹

⫹ ⫹ ⫹


Day of evaluation Disease characteristics Disease extent Disease duration Duration of present attack No. of previous attacks Previous oral steroids Patient characteristics Sex Temperature Pulse rate Stool frequency Blood in stool Stool frequency change Pedal edema Laboratory results CRP ESR Albumin Hemoglobin WBC Prothrombin Fibrinogen Total protein Cholinesterase Imaging Radiologic assessment Endoscopy

January 2007

Nine studies reported the use of cyclosporine in the cohort, and none reported the use of tacrolimus. Although the pooled number of patients included in the cyclosporine studies was 622, only 100 (16%; 95% CI, 14%–20%) received cyclosporine, of which 51 (51%; 95% CI, 41%– 60%) responded and were discharged without colectomy. The overall colectomy rate in these 9 studies combined was 29% (179/622 patients; 95% CI, 25%– 32%), identical to the combined results of the other 18 studies in which cyclosporine was not used (353/1200 [29.4%; 95% CI, 27%–32%]; P ⫽ 1, odds ratio, 1 [95% CI, 0.8 –1.3]). The dose of IVCS reported in 24 studies (Table 1) was standardized as methylprednisolone equivalent while using a mean adult weight of 70 kg. The mean daily dose was 68 ⫾ 13 mg (range, 40 –100 mg). Only 3 studies used a dose less than 60 mg/day. In a meta-regression controlled for disease severity at baseline, there was no correlation of the dose used with colectomy rate (Figure 1; R2 ⬍ 0.01, P ⫽ .98). Nonetheless, because of the potential clinical importance, this variable was also forced into the final model. Only one cohort31 reported the use of infliximab in 6 of 149 patients (4%), of whom 4 (67%) responded and were discharged without further therapy. Finally, a meta-regression of colectomy rate versus time was performed while controlling for the dose of IVCS used, disease severity at baseline, and the time to colectomy evaluation (Figure 2). Evaluation of temporal trends in colectomy rates indicates that there has been no change during the last 30 years (R2 ⫽ 0.07, P ⫽ .8). In the multivariate meta-regression, none of the factors assumed to contribute to heterogeneity were statistically significant (P ⫽ .9 for IVCS dose, P ⫽ .62 for disease severity, P ⫽ .58 for time to colectomy evaluation). Furthermore, the best fit regression line was not changed after excluding these variables from the model. It was impossible to pool the results of the predictive studies (Table 2) because of the extreme variation of the included studies. Of the studies tabulated in Table 2, three attempted to derive a valid predictive index from the measured variables. In a prospective study, Travis et al19 found that a stool frequency of ⬎8/day or 3– 8/day and C-reactive protein (CRP) ⬎45 mg/dL on the third day of therapy have a positive predictive value (PPV) of 85% for colectomy. Lindgren et al16 developed the fulminant colitis index (n ⫽ 97) by using a regression formula to predict the likelihood of medical failure, stool frequency/day ⫹ 0.14 ⫻ CRP (mg/L). The sensitivity, specificity, and PPV of this index in predicting colectomy at a cutoff score of ⬎8, during the third day of corticosteroid therapy, were 76%, 81%, and 72%, respectively. The fulminant colitis index was prospectively assessed in a controlled trial of rescue infliximab therapy in severe UC, in which the PPV for colectomy was 69% in the placebo group.8 Recently, Ho et al32 developed a different index to predict colectomy on the basis of a retrospective chart review of 167 UC patients with acute severe attacks. The score (0 –9) included mean stool frequency (⬍4, 4 – 6, 7–9, ⬎9), presence of colonic dilatation (⬎4 cm), and hypoalbuminemia (⬍30 g/L). The sensitivity and specificity for predicting medical therapy failure with a score of ⬎4 were 85% and 75%, respectively (area under a receiver operating characteristic curve of 0.88). This score has yet to be validated in a separate prospective cohort. The index of Seo et al,15 calculated after 1 week of IVCS therapy, had a PPV of 52% and negative predictive value of 97% to predict short-term colectomy at a cutoff point of 180.23 The


107

classification of Truelove and Witts and the index of Lichtiger et al have also shown some predictive value, but these have not been tested systematically.29

Discussion The introduction of IVCS by Truelove and Jewell5 greatly reduced the mortality and morbidity previously associated with severe exacerbations of UC. Our review of the literature suggests that further reduction in short-term colectomy rates has not occurred during the subsequent 3 decades, despite the introduction of calcineurin inhibitors to the therapeutic arsenal. Mortality rates also differed across studies, but no significant trend was observed, and it remains low at 1% (95% CI, 0.7%–1.5%). The wide range of colectomy rates in the published studies could reflect heterogeneity of inclusion criteria and time of follow-up, as well as different approaches to management. The study that reported a 72% colectomy rate was from a center in which colectomy was routinely performed after failure of only 5 days of IVCS therapy,22 whereas the 10% colectomy rate was observed after a very conservative approach.34 Because of the imprecision of disease severity classification, it is possible that patients included in hospitalized cohorts 30 years ago are now being treated in an ambulatory fashion with oral prednisone. As a result, patients who are now hospitalized might have, in some ways, more severe disease, and this might account for the lack of apparent decline in colectomy rates over time. Therefore, we controlled the final meta-regression model for variables that could have explained heterogeneity, but this did not change the consistent colectomy rate of 27%. Optimization of therapy in severe UC is hindered by the surprising lack of dose-ranging studies with corticosteroids. Only one clinical trial in ambulatory patients compared oral prednisolone dosing and concluded that 60 mg/day was as effective as 40 mg/day but with more adverse events.35 The only comparative study in severe UC used ACTH versus hydrocortisone.28 Our review revealed the common use of higher doses than with oral corticosteroid therapy, but meta-regression failed to demonstrate a dose-colectomy association. It is possible, however, that more severe patients received higher doses, and therefore no reduction in colectomy rate was shown, but the vast majority of the studies provided the doses as standardized local guidelines, rather than individualized dosing. Clearly, a randomized, dose-ranging study of IVCS is needed to definitively determine optimal first-line management of acute severe colitis. Studies of cyclosporine or tacrolimus therapy in corticosteroid-refractory patients showed that colectomy could be avoided in the short-term in roughly 70%– 80% of cases.7 In contrast, we found a pooled 51% (95% CI, 41%– 60%) short-term success rate, which suggests that the real life success rate is lower than that in standardized clinical trials and small specialized subcohorts that could be affected by publication bias of positive findings. Moreover, it seems that cyclosporine has not gained popularity in the treatment of patients with IVCS failure, because the actual proportion of patients receiving cyclosporine was low, even among studies that did report the use of cyclosporine in clinical practice. It is, therefore, not surprising that the relative benefit of cyclosporine has not translated into reduced overall colectomy rate. The reasons for the limited use of cyclosporine could be related to the potential for significant

108

TURNER ET AL

Figure 1. Colectomy rates in hospitalized UC patients treated with various doses of IVCS during the last 30 years, controlled for disease severity at baseline. All doses were standardized as methylprednisolone equivalent for an adult weighing 70 kg. Sum of squares regression line (continuous line) and 95% CI (dotted lines) were calculated after weighting each observation by the inverse variance method (size of the diamonds represents the relative weight of each study to the analysis).

toxicity and the need for colectomy in the long-term in a substantial portion of the initial responders.7 Publication bias could have affected the results of this systematic review, but the large number of studies included likely provided an accurate reflection of the actual use of cyclosporine in the overall population. More recently, there has been the option of prescribing infliximab for rescue or maintenance therapy in UC.8,9 Only one cohort in our review reported the use of infliximab in a small number of patients, and the implication of infliximab therapy in the outcome of severe colitis is yet to be reflected in the literature. UC develops during childhood or adolescence in up to 25% of patients and is extensive in 50%– 80%.3 Remarkably, however, only 3 small retrospective studies (43 patients in total) have assessed the short-term response rate of severe colitis occurring specifically in children. One of these studies reported lower rate of colectomy than in adults (10%) and the other 2 reported higher rates (45%), and no clear conclusions could be drawn. Management guidelines in severe colitis could aid physicians in optimizing the medical therapy available and should be based on variables that could predict response to therapy. More than 20 variables were found to predict response to IVCS in the adult literature (Table 2). Pooling these studies for meta-analysis was impossible, but a few variables were consistently reproduced: disease extent, stool frequency, temperature, heart rate, CRP, albumin, and radiologic assessment (mucosal tags and bowel dilatation). Other variables could be important but might have been understudied. Although clinicians might consider these variables independently, 3 major indices were developed to predict short-term colectomy in hospitalized adults with UC.16,19,32 It will be necessary to validate these indices and compare their accuracy in one prospective cohort study. The development of these indices did not incorporate second-line medical therapies, and they should be tested in a cohort that allows for infliximab or calcineurin inhibitors therapy.


It might be possible in the future to predict a patient’s steroid responsiveness from genetic, serologic, or fecal markers, but, to date, no study has systematically evaluated these markers with respect to short-term outcome in a cohort of hospitalized UC patients. As the genetic basis of susceptibility to IBD has been explored, some studies have suggested an association between specific genes and the severity of UC. The IBD3 region contains the major histocompatibility complex (MHC) genes (also referred to as the HLA complex).36 HLA DRB1*0103 is consistently associated with both extensive and severe disease requiring colectomy and is also associated with shorter time to surgery.37– 42 Unfortunately, the low frequency of this allele suggests that this association is unlikely to be clinically useful in predicting disease course. Biallelic structural polymorphism in the inhibitor of ␬B-like (IKBL) gene, in the central MHC region, has been also associated with severe course,41 and multidrug resistance 1 (MDR-1) has been shown to predict steroid resistance in UC.43 A polymorphism in the hMLH1 gene, a mismatch repair gene involved in hereditary non-polyposis colorectal cancer, has been shown to infer susceptibility to refractory UC.44 Serologic testing might play a role in predicting disease behavior in Crohn’s disease,45 but there are no such data regarding UC. Fecal calprotectin, an important granulocyte cytosolic protein, is closely correlated with disease activity and mucosal healing in patients with UC46,47 and can predict 1-year relapse rate.46,48 This review has many of the limitations common to systematic reviews in general and observational studies in particular. The heterogeneity among the cohorts makes it difficult to draw definite conclusion, despite the efforts to control for various variables in the final model (all of which were statistically not significant). Nonetheless, the large number of studies included in this review is likely to have diluted the impact of outlying studies, and the results probably reflect clinical outcomes in real life. Less heterogeneous support to the premise of this article

Figure 2. Colectomy rates in hospitalized UC patients requiring IVCS during the last 30 years, controlled for dose of treatment, disease severity at baseline, and time to evaluation of colectomy. Sum of squares regression line (continuous line) and 95% CI (dotted lines) were calculated after weighting each observation by the inverse variance method (size of the diamonds represents the relative weight of each study to the analysis).

January 2007

comes from the UK. Colectomy rate at a single center (Oxford) has not changed not only between 1974 –1996 (Table 1) but also since 1955.4 Furthermore, a prospective study of 116 patients admitted during a 3-month period to 29 UK centers also found 28% colectomy rate in severe UC, and only 15% were treated with cyclosporine.49 This was not included in the meta-regression because it was published in an abstract form only. In conclusion, colectomy rates in UC attacks significant enough to require IVCS therapy have been stable during the last 3 decades, despite the improvement in medical therapy and the introduction of calcineurin inhibitors. This review does not provide any evidence to support increasing IVCS dose beyond 60 mg/day of methylprednisolone, whereas lower doses were not evaluated because of the small sample size. Various clinical variables at the third day of IVCS (ie, CRP, albumin, stool frequency, temperature, heart rate, and radiology assessment) might aid the clinicians to decide early on second-line therapy, but the optimal treatment sequence of calcineurin inhibitors, infliximab, and colectomy is yet to be determined. The current review and further cohort studies should facilitate the development of accepted guidelines for the management of severe colitis in the attempt to minimize morbidity and mortality. Well-validated predictive indices could play an important role in the guideline development.50-62 References 1. Daperno M, Sostegni R, Rocca R, et al. Review article: medical treatment of severe ulcerative colitis. Aliment Pharmacol Ther 2002;16(Suppl 4):7–12. 2. Antonioli DA. Pediatric inflammatory bowel disease. Pediatr Dev Pathol 2005;8:2–19. 3. Griffiths AM. Specificities of inflammatory bowel disease in childhood. Best Pract Res Clin Gastroenterol 2004;18:509 –523. 4. Truelove SC, Witts LJ. Cortisone in ulcerative colitis: final report on a therapeutic trial. Br Med J 1955;4947:1041–1048. 5. Truelove SC, Jewell DP. Intensive intravenous regimen for severe attacks of ulcerative colitis. Lancet 1974;1:1067–1070. 6. Kugathasan S, Dubinsky MC, Keljo D, et al. Severe colitis in children. J Pediatr Gastroenterol Nutr 2005;41:375–385. 7. Durai D, Hawthorne AB. Review article: how and when to use cyclosporin in ulcerative colitis. Aliment Pharmacol Ther 2005; 22:907–916. 8. Jarnerot G, Hertervig E, Friis-Liby I, et al. Infliximab as rescue therapy in severe to moderately severe ulcerative colitis: a randomized, placebo-controlled study. Gastroenterology 2005;128: 1805–1811. 9. Rutgeerts P, Sandborn WJ, Feagan BG, et al. Infliximab for induction and maintenance therapy for ulcerative colitis. N Engl J Med 2005;353:2462–2476. 10. Ardizzone S, Bianchi Porro G. Biologic therapy for inflammatory bowel disease. Drugs 2005;65:2253–2286. 11. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting: Metaanalysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000;283:2008 –2012. 12. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17:1–12. 13. Higgins JPT, Green S, eds. Cochrane handbook for systematic reviews of interventions 4.2.5: the Cochrane Library. 3rd ed. Chichester, UK, John Wiley & Sons, Ltd; 2005. 14. Lichtiger S, Present DH. Preliminary report: cyclosporin in treatment of severe active ulcerative colitis. Lancet 1990;336: 16 –19.


109

15. Seo M, Okada M, Yao T, et al. An index of disease activity in patients with ulcerative colitis. Am J Gastroenterol 1992;87: 971–976. 16. Lindgren SC, Flood LM, Kilander AF, et al. Early predictors of glucocorticosteroid treatment failure in severe and moderately severe attacks of ulcerative colitis. Eur J Gastroenterol Hepatol 1998;10:831– 835. 17. Gulati R, Rawal KK, Kumar N, et al. Course of severe ulcerative colitis in northern India. Trop Gastroenterol 1995;16:19 –23. 18. Lennard-Jones JE, Ritchie JK, Hilder W, et al. Assessment of severity in colitis: a preliminary study. Gut 1975;16:579 –584. 19. Travis SP, Farrant JM, Ricketts C, et al. Predicting outcome in severe ulcerative colitis. Gut 1996;38:905–910. 20. Oshitani N, Kitano A, Fukushima R, et al. Predictive factors for the response of ulcerative colitis patients during the acute-phase treatment. Digestion 1990;46:107–113. 21. Chakravarty BJ. Predictors and the rate of medical treatment failure in ulcerative colitis. Am J Gastroenterol 1993;88:852– 855. 22. Elloumi H, Ben Abdelaziz A, Derbel F, et al. Predictive factors of glucocorticosteroid treatment failure in severe acute idiopathic colitis. Acta Gastroenterol Belg 2005;68:226 –229. 23. Seo M, Okada M, Yao T, et al. Evaluation of the clinical course of acute attacks in patients with ulcerative colitis through the use of an activity index. J Gastroenterol 2002;37:29 –34. 24. Choquet A, Yamamoto-Furusho JK, Reyes E, et al. [Predictors of colectomy in patients with ulcerative colitis: a cohort analysis of 184 cases]. Rev Invest Clin 2004;56:11–15. 25. Benazzato L, D’Inca R, Grigoletto F, et al. Prognosis of severe attacks in ulcerative colitis: effect of intensive medical treatment. Dig Liver Dis 2004;36:461– 466. 26. Spicer CC. The prediction of success or failure of medical treatment for acute colitis. J R Coll Physicians Lond 1975;9:252– 256. 27. Spicer CC, Buckell NA, Jones JE. Prognosis in acute colitis. J R Coll Physicians Lond 1979;13:201–202. 28. Meyers S, Lerer PK, Feuer EJ, et al. Predicting the outcome of corticoid therapy for acute ulcerative colitis: results of a prospective, randomized, double-blind clinical trial. J Clin Gastroenterol 1987;9:50 –54. 29. Kumar S, Ghoshal UC, Aggarwal R, et al. Severe ulcerative colitis: prospective study of parameters determining outcome. J Gastroenterol Hepatol 2004;19:1247–1252. 30. Carbonnel F, Gargouri D, Lemann M, et al. Predictive factors of outcome of intensive intravenous treatment for attacks of ulcerative colitis. Aliment Pharmacol Ther 2000;14:273–279. 31. Daperno M, Sostegni R, Scaglione N, et al. Outcome of a conservative approach in severe ulcerative colitis. Dig Liver Dis 2004;36:21–28. 32. Ho GT, Mowat C, Goddard CJ, et al. Predicting the outcome of severe ulcerative colitis: development of a novel risk score to aid early selection of patients for second-line medical therapy or surgery. Aliment Pharmacol Ther 2004;19:1079 –1087. 33. Oshitani N, Matsumoto T, Jinno Y, et al. Prediction of short-term outcome for patients with active ulcerative colitis. Dig Dis Sci 2000;45:982–986. 34. Gold DM, Levine JJ, Weinstein TA, et al. Prolonged medical therapy for severe pediatric ulcerative colitis. Am J Gastroenterol 1995;90:732–735. 35. Lennard-Jones JE. Toward optimal use of corticosteroids in ulcerative colitis and Crohn’s disease. Gut 1983;24:177–181. 36. Walters TD, Silverberg MS. Genetics of inflammatory bowel disease: current status and future directions. Can J Gastroenterol 2006;20:633– 639. 37. Ahmad T, Armuzzi A, Neville M, et al. The contribution of human leucocyte antigen complex genes to disease phenotype in ulcerative colitis. Tissue Antigens 2003;62:527–535.

110

TURNER ET AL

38. Bouma G, Crusius JB, Garcia-Gonzalez MA, et al. Genetic markers in clinically well defined patients with ulcerative colitis (UC). Clin Exp Immunol 1999;115:294 –300. 39. Trachtenberg EA, Yang H, Hayes E, et al. HLA class II haplotype associations with inflammatory bowel disease in Jewish (Ashkenazi) and non-Jewish Caucasian populations. Hum Immunol 2000;61:326 –333. 40. Roussomoustakaki M, Satsangi J, Welsh K, et al. Genetic markers may predict disease behavior in patients with ulcerative colitis. Gastroenterology 1997;112:1845–1853. 41. de la Concha EG, Fernandez-Arquero M, Lopez-Nava G, et al. Susceptibility to severe ulcerative colitis is associated with polymorphism in the central MHC gene IKBL. Gastroenterology 2000; 119:1491–1495. 42. Yap LM, Ahmad T, Jewell DP. The contribution of HLA genes to IBD susceptibility and phenotype. Best Pract Res Clin Gastroenterol 2004;18:577–596. 43. Farrell RJ, Murphy A, Long A, et al. High multidrug resistance (P-glycoprotein 170) expression in inflammatory bowel disease patients who fail medical therapy. Gastroenterology 2000;118: 279 –288. 44. Bagnoli S, Putignano AL, Melean G, et al. Susceptibility to refractory ulcerative colitis is associated with polymorphism in the hMLH1 mismatch repair gene. Inflamm Bowel Dis 2004;10:705– 708. 45. Mow WS, Vasiliauskas EA, Lin YC, et al. Association of antibody responses to microbial antigens and complications of small bowel Crohn’s disease. Gastroenterology 2004;126:414 – 424. 46. Costa F, Mumolo MG, Bellini M, et al. Role of faecal calprotectin as non-invasive marker of intestinal inflammation. Dig Liver Dis 2003;35:642– 647. 47. Roseth AG, Aadland E, Grzyb K. Normalization of faecal calprotectin: a predictor of mucosal healing in patients with inflammatory bowel disease. Scand J Gastroenterol 2004;39:1017– 1020. 48. Costa F, Mumolo MG, Ceccarelli L, et al. Calprotectin is a stronger predictive marker of relapse in ulcerative colitis than in Crohn’s disease. Gut 2005;54:364 –368. 49. Hawthorne AB, Travis SP. Outcome of inpatient management of severe ulcerative colitis: a BSG IBD Clinical Trials Network survey. Gut 2002;50:A16. 50. Schjonsby H, Heimann P, Kremer D, et al. Intensive treatment in severe acute attacks of ulcerative colitis. Acta Med Scand Suppl 1977;603:43– 46.


51. Truelove SC, Willoughby CP, Lee EG, et al. Further experience in the treatment of severe attacks of ulcerative colitis. Lancet 1978;2:1086 –1088. 52. Buckell NA, Lennard-Jones JE. How district hospitals see acute colitis. Lancet. 1979;1:1226 –1229. 53. Jarnerot G, Rolny P, Sandberg-Gertzen H. Intensive intravenous treatment of ulcerative colitis. Gastroenterology 1985;89:1005– 1013. 54. McIntyre PB, Powell-Tuck J, Wood SR, et al. Controlled trial of bowel rest in the treatment of severe acute colitis. Gut 1986; 27:481– 485. 55. Jones HW, Grogono J, Hoare AM. Acute colitis in a district general hospital. Br Med J (Clin Res Ed) 1987;294:683– 684. 56. Mantzaris GJ, Hatzis A, Kontogiannis P, et al. Intravenous tobramycin and metronidazole as an adjunct to corticosteroids in acute, severe ulcerative colitis. Am J Gastroenterol 1994;89: 43– 46. 57. Panes J, Esteve M, Cabre E, et al. Comparison of heparin and steroids in the treatment of moderate and severe ulcerative colitis. Gastroenterology 2000;119:903–908. 58. Mantzaris GJ, Petraki K, Archavlis E, et al. A prospective randomized controlled trial of intravenous ciprofloxacin as an adjunct to corticosteroids in acute, severe ulcerative colitis. Scand J Gastroenterol 2001;36:971–974. 59. D’Haens G, Lemmens L, Geboes K, et al. Intravenous cyclosporine versus intravenous corticosteroids as single therapy for severe attacks of ulcerative colitis. Gastroenterology 2001;120: 1323–1329. 60. Latella G, Vernia P, Viscido A, et al. GI distension in severe ulcerative colitis. Am J Gastroenterol 2002;97:1169 –1175. 61. Werlin SL, Grand RJ. Severe colitis in children and adolescents: diagnosis, course, and treatment. Gastroenterology 1977;73(Pt 1):828 – 832. 62. Barabino A, Tegaldo L, Castellano E, et al. Severe attack of ulcerative colitis in children: retrospective clinical survey. Dig Liver Dis 2002;34:44 – 49.

Address requests for reprints to: Anne Griffiths, MD, Division of Pediatric Gastroenterology, Hepatology, and Nutrition, Hospital for Sick Children, 555 University Ave, Toronto M5G-1X8, Canada. e-mail: Anne.griffi[email protected]; fax: (416) 813-6531.