Decreasing Colectomy Rates for Ulcerative Colitis: A Population-Based Time Trend Study

nature publishing group ORIGINAL CONTRIBUTIONS

INF

LAM

MAT

OR

Y B

OW

EL

DIS

EA

SE

1879

© 2012 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY

see related editorial on page x

INTRODUCTION Colectomy rates for ulcerative colitis (UC) have varied between

cohorts and across time. Langholz et al. published in 1994 that

25 % of UC patients underwent colectomy within the fi rst 10

years of diagnosis ( 1 ). However, more recent studies from Europe

following cohorts of newly diagnosed UC patients in the early

1990s have shown lower 10-year colectomy rates of 8.7 % ( 2 )

and 9.8 % ( 3 ), respectively. Similarly, in Canada, the incidence of

colectomy at 10 years was 10.4 % and this value was lower among

UC patients diagnosed in recent eras ( 4 ). Although a nation-

wide US study concluded that rates of colectomy for UC had not

changed in the last decade, the study was unable to determine the

association between immunomodulator use and surgery because

prescription data were not available ( 5 ).

In addition, the aforementioned studies on UC and colectomy

( 1 – 5 ) were primarily conducted before the publication of studies

demonstrating the effi cacy of infl iximab in inducing and maintain-

ing remission for UC ( 6 – 8 ). Randomized controlled trials have also

shown that infl iximab reduce colectomies in UC patients in both the

outpatient ( 9 ) and in-patient settings ( 6,7 ). However, population-

based studies demonstrating the uptake of infl iximab pre- and post-

2005 and the corresponding impact on colectomy rates are lacking.

Finally, prior studies evaluating UC colectomy rates have

not taken into account the indication for surgery. UC patients

Decreasing Colectomy Rates for Ulcerative Colitis: A Population-Based Time Trend Study Gilaad G. Kaplan , MD, MPH, FRCPC 1 , 2 , 3 , 4 , Cynthia H. Seow , MBBS, MSc 1 , 2 , 3 , 4 , Subrata Ghosh , MD 1 , 2 , 3 , Natalie Molodecky , MSc 1 , 2 , 3 , Ali Rezaie , MD, MSc 1 , 2 , 3 , Gordon W. Moran , MD, PhD 1 , 2 , 3 , Marie-Claude Proulx , MSc 1 , 2 , 3 , James Hubbard , MSc 1 , 2 , 3 , Anthony MacLean , MD 5 , Donald Buie , MD, MSc 5 and Remo Panaccione , MD 1 , 2 , 3

OBJECTIVES: Colectomy rates for ulcerative colitis (UC) have been inconsistently reported. We assessed temporal trends of colectomy rates for UC, stratifi ed by emergent vs. elective colectomy indication.

METHODS: From 1997 to 2009, we identifi ed adults hospitalized for a fl are of UC. Medical charts were reviewed. Temporal changes were evaluated using linear regression models to estimate the average annual percent change (AAPC) in surgical rates. Logistic regression analysis compared: (i) UC patients responding to medical management in hospital to those who underwent colectomy; (ii) UC patients who underwent an emergent vs. elective colectomy; and (iii) temporal trends of drug utilization.

RESULTS: From 1997 to 2009, colectomy rates signifi cantly dropped for elective colectomies with an AAPC of − 7.4 % (95 % confi dence interval (CI): − 10.8 % , − 3.9 % ). The rate of emergent colectomies remained stable with an AAPC of − 1.4 % (95 % CI: − 4.8 % , 2.0 % ). Azathioprine / 6-mercaptopurine prescriptions increased from 1997 to 2009 (odds ratio (OR) = 1.15; 95 % CI: 1.09 – 1.22) and infl iximab use increased after 2005 (OR = 1.68; 95 % CI: 1.25 – 2.26). A 13 % per year risk adjusted reduction in the odds of colectomy (OR = 0.87; 95 % CI: 0.83 – 0.92) was observed in UC patients responding to medical management compared with those who required colectomy. Emergent colectomy patients had a shorter duration of fl are ( < 2 weeks vs. 2 – 8 weeks, OR = 5.31; 95 % CI: 1.58 – 17.81) and underwent colectomy early after diagnosis ( < 1 year vs. 1 – 3 years, OR = 5.48; 95 % CI: 2.18 – 13.79).

CONCLUSIONS: From 1997 to 2009, use of purine anti-metabolites increased and elective colectomy rates in UC patients decreased signifi cantly. In contrast, emergent colectomy rates were stable, which may have been due to rapid progression of disease activity.

Am J Gastroenterol 2012; 107:1879–1887; doi: 10.1038/ajg.2012.333; published online 20 November 2012

1 Infl ammatory Bowel Disease Clinic, University of Calgary , Calgary , Alberta ; 2 Division of Gastroenterology, University of Calgary , Calgary , Alberta ; 3 Department of Medicine, University of Calgary , Calgary , Alberta ; 4 Department of Community Health Sciences, University of Calgary , Calgary , Alberta ; 5 Department of Surgery, University of Calgary , Calgary , Alberta . Correspondence: Gilaad G Kaplan, MD, MPH, FRCPC , Teaching Research and Wellness Center , Calgary , Alberta. E-mail: [email protected] Received 8 April 2012; accepted 23 August 2012

The American Journal of GASTROENTEROLOGY VOLUME 107 | DECEMBER 2012 www.amjgastro.com

1880 IN

FLA

MM

ATO

RY

BO

WE

L D

ISE

AS

E Kaplan et al.

typically undergo colectomy in one of two settings: (i) emer-

gent colectomy, whereby patients are operated on due to an

acute complication (e.g., toxic megacolon) or lack of response to

in-hospital medical therapies; and (ii) elective colectomy, whereby

patients are generally less sick but require an operation due to

lack of response to outpatient medical therapies or require sur-

gery for the management of colonic dysplasia or cancer ( 10,11 ).

We therefore carried out a population-based study to assess the

temporal trends of colectomy rates for UC, stratifi ed by emergent

vs. elective colectomy indication.

METHODS Patient data source Th e Data Integration, Measurement, and Reporting (DIMR) hos-

pital discharge abstract administrative database captures all hospi-

talizations in the Calgary Health Zone of Alberta Health Services,

Canada. Th e Calgary Health Zone is a population-based health

authority under a public, single-payer system, and provides all

levels of medical and surgical care to the residents of the city of

Calgary and over 20 nearby smaller cities, towns, villages, and

hamlets. Th e estimated population of the Calgary Health Zone in

2008 was 1,275,664 ( 12 ). Th e DIMR database contains 42 diagnos-

tic and 25 procedural coding fi elds. Th e International Classifi ca-

tion of Disease, Ninth Revision, Clinical Modifi cation (ICD-9-CM)

was used up to 2001, whereas the ICD-10-CA and the Canadian

Classifi cation of Health Intervention coding were used aft er 2001.

Study population Th e DIMR administrative discharge database was used to iden-

tify all adults ( ≥ 18 years of age) who were admitted to hospital

with a diagnosis of UC, between 1 January 1997 and 31 Decem-

ber 2009. Th is cohort was split into two populations: (i) UC

patients who underwent colectomy; and (ii) UC patients who

were admitted to hospital for a fl are, but did not undergo colec-

tomy. We identifi ed 763 discharge abstracts containing codes for

both UC (ICD-9-CM 556.X or ICD-10-CA K51.X) and colec-

tomy (ICD-9-CM: 45.7 and 45.8 or Canadian Classifi cation of

Health Intervention: 1.NM.87, 1.NM.89, 1.NM.91, 1.NQ.89, 1.

NQ. 90), as well as 969 discharge abstracts with UC coded in

the primary diagnostic position. Charts were excluded for the

following reasons: (i) UC patient resided outside of the Calgary

Health Zone on the date of their admission ( n = 404); (ii) the

hospital chart was not available to review ( n = 147); (iii) other

admissions for the same individual ( n = 173); (iv) patient did

not have a diagnosis of UC aft er chart review ( n = 103); and (v)

patients were not admitted for a fl are of UC ( n = 91); and (vi)

patients had a colectomy before 1997 ( n = 39). Th e fi nal cohorts

included 437 UC patients who underwent colectomy and 338

UC patients who were admitted to hospital for a fl are, but did

not undergo colectomy between 1997 and 2009. UC patients

who underwent colectomy were not included in the fl are cohort.

When an UC patient had more than one hospital admission,

we randomly selected the hospital admission that served as the

index date for analysis.

Data extraction Th e data extracted from the chart review included: age at colec-

tomy; year of surgery; sex; smoking status at hospital admission,

defi ned as current, ex-smoker, or never smoker; duration of dis-

ease, defi ned as time between UC diagnosis and date of surgery;

number of comorbidities before admission to hospital as previ-

ously described ( 11,13 ) (see Appendix 1 for comorbidity classi-

fi cation); disease activity at admission as defi ned by frequency of

bowel movements (above or below 5) and blood in stool (yes vs.

no); extent of disease, classifi ed as pancolitis vs. left -sided colitis

(i.e., infl ammation limited to below the splenic fl exure); duration

of fl are (defi ned as < 2 weeks, between 2 and 8 weeks, and more

than 8 weeks); length of stay in hospital; duration between admis-

sion and the operation; and dysplasia or colon cancer at colec-

tomy. For each patient the chart reviewer determined the use of

the following medications before or during admission: mesala-

mine or sulfasalazine; azathioprine (AZA) or 6-mercaptopurine

(6-MP); prednisone; cyclosporine; and infl iximab. Th e admission

was stratifi ed by colectomy indication; that is, elective vs. emer-

gent. An operation was defi ned as ‘ elective ’ if the decision to oper-

ate on the UC patient was made before admission to hospital. In

contrast, the decision for an emergent colectomy was determined

during the hospital admission (e.g., acute complication or refrac-

tory to in-hospital medical management).

Data analysis Descriptive statistics were performed using the Fisher ’ s exact test

for proportions; continuous variables were expressed as medians

with interquartile ranges and compared using the Wilcoxon rank

sum test. Th e yearly incidence of colectomies for UC was calcu-

lated by dividing the annual total number of colectomies by the

annual adult population size for the Calgary Health Zone from

1997 to 2009. In a sensitivity analysis, we divided the annual total

number of colectomies by the estimated prevalence of UC in the

Calgary Health Region from 1999 to 2005 ( 14,15 ). UC patients

were stratifi ed by colectomy indication: emergent vs. elective.

Th ree time trend analyses were performed on colectomy rates.

First, the average annual percent change (AAPC) from 1997 to

2009 was calculated using a generalized linear model that assumed

a Poisson distribution. AAPCs were not signifi cant when the 95 %

confi dence intervals (CI) crossed 0. Joinpoint regression was used to

explore whether each temporal trend would be better modeled with

two or three log-linear segments connected at ‘ joinpoints ’ instead

of just one; this method uses statistical criteria to determine if there

are signifi cant changes in trend at points in the time series, referred

to as joinpoints ( http://surveillance.cancer.gov/joinpoint/ . 2011).

Second, time trend analyses were conducted using a multivari-

ate logistic regression model comparing UC patients admitted

to hospital for a fl are of UC but not requiring colectomy vs. UC

patients who underwent colectomy. Year was modeled as a continu-

ous variable (1997 – 2009) and adjusted for the following variables

using the stepwise selection approach with 0.1 as the entry and exit

P -value: age (defi ned as 18 – 34, 35 – 64, and > 64 years), comorbidity

(defi ned as 0, 1 – 2, > 2); sex; smoking status (current, ex-smoker,

never / unknown); length of fl are ( < 2 weeks, 2 – 8 weeks, > 8 weeks,


1881

INF

LAM

MAT

OR

Y B

OW

EL

DIS

EA

SE

Colectomy Rates in Ulcerative Colitis

or undetermined); disease activity (defi ned as > 5 stool per day and

the presence of blood in stool); extent of disease (left -sided colitis,

pancolitis, or unknown); and mesalamine / sulfasalazine, prednisone,

AZA / 6-MP, and / or infl iximab prescribed before admission.

Th ird, multivariate logistic regression analysis was performed

to examine patient characteristic diff erences between UC patients

who underwent an emergent as compared with an elective colec-

tomy. Year was modeled as a continuous variable (1997 – 2009) and

adjusted for the following variables using the stepwise selection

approach with 0.1 as the entry and exit P -value: age, comorbid-

ity, sex, smoking, length of fl are, disease activity, extent of disease,

and infl ammatory bowel disease (IBD) medications. Signifi cance

was defi ned as a two-sided P -value < 0.05. Risk estimates were pre-

sented as adjusted odds ratios (ORs) with 95 % CI.

Annual drug utilization patterns between 1997 and 2009 for

hospitalized UC patients were determined by evaluating all UC

patients who underwent colectomy and UC patients admitted to

hospital for a fl are of UC but not requiring colectomy. To evaluate

changes in drug utilization practices between 1997 and 2009, we

determined the annual proportion of UC patients (colectomy and

non-colectomy) who were prescribed the following drugs before

or during admission to hospital: mesalamine or sulfasalazine; oral

corticosteroids; AZA or 6-MP; and / or infl iximab. Cyclosporine

was subsequently excluded from the analysis because the drug

was rarely prescribed ( n = 3). Th e temporal trends analysis of drug

utilization excluded hospital admissions that occured within 3

months of diagnosis for mesalamine/sulfasalazine and AZA / 6-MP.

Time trend analyses of drug utilization practices of mesalamine /

sulfasalazine, oral corticosteroids, AZA / 6-MP, and infl iximab were

evaluated using separate multivariate logistic regression models

with the drug as the outcome and year (1997 – 2009) as a continu-

ous variable with adjustment for age and sex. For infl iximab, the

analysis was restricted to aft er the year 2005.

Statistical analyses were conducted using SAS version 9.2 (SAS

Institute, Cary, NC, USA) and Joinpoint Regression Program 3.5.1

( http://surveillance.cancer.gov/joinpoint/ . 2011). Th e study was

approved by the Conjoint Health Research Ethics Board of the

University of Calgary.

RESULTS Between 1 January 1997 and 31 December 2009 we identifi ed a

population-based cohort of 437 patients with UC who underwent

a colectomy, with over half of all operations carried out on an emer-

gent basis (53.1 % ). Table 1 presents the characteristics of patients

with UC who were admitted to hospital for a fl are and responded to

medical management as compared with UC patients who under-

went colectomy, stratifi ed by emergent vs. elective colectomy.

Between 1997 and 2009 the incidence of colectomy for UC

signifi cantly decreased ( P < 0.01) from a high of 5.4 to a low of

2.3 per 100,000 person-years, respectively ( Figure 1 ). Over the

13-year study period, the calculated AAPC signifi cantly decreased

by − 4.3 % (95 % CI: − 6.6 % , − 1.8 % ). Colectomy rates signifi cantly

dropped among UC patients who underwent elective colectomies

( Figure 1 ) with an AAPC of − 7.4 % (95 % CI: − 10.8 % , − 3.9 % ).

However, colectomy rates remained stable for UC patients who

underwent an emergent colectomy ( Figure 1 ) with an AAPC

of − 1.4 % (95 % CI: − 4.8 % , 2.0 % ). Infl ection points were not statis-

tically observed in joinpoint analysis. In sensitivity analysis, using

the prevalence of UC as the denominator, the AAPC from 1999 to

2005 were: (i) − 5.2 % ( − 10.9, 0.9) in all colectomies; (ii) − 2.7 %

( − 10.9, 6.2) in emergent colectomies; and (iii) − 7.6 % ( − 15.4, 0.9)

in elective colectomies.

A 13 % per year risk adjusted reduction in the odds of colectomy

(OR = 0.87; 95 % CI: 0.83 – 0.92) was observed in hospitalized UC

patients who responded to medical management compared with

UC patients who required colectomy ( Table 2 ).

Among all UC patients who underwent a colectomy, year of

surgery was an independent predictor of emergent vs. elective

colectomies (OR = 1.12; 95 % CI: 1.03 – 1.22; Table 2 ) aft er adjusting

for covariates including duration of fl are ( < 2 weeks vs. 2 – 8 weeks,

OR = 5.31; 95 % CI: 1.58 – 17.81) and time between diagnosis and

surgery ( < 1 year vs. 1 – 3 years, OR = 5.48; 95 % CI: 2.18 – 13.79).

Between 1997 and 2009 the proportion of hospitalized UC

patients (i.e., medically responsive and those who underwent

colectomy) prescribed mesalamine / sulfasalazine (OR = 1.05; 95 %

CI: 0.99 – 1.11) and prednisone (OR = 1.00; 95 % CI: 0.94 – 1.06)

remained stable ( Figure 2 ). Between 1997 and 2009, prescription

of AZA or 6-MP steadily increased (OR = 1.20; 95 % CI: 1.14 – 1.27;

Figure 2 ). From 2005 to 2009 infl iximab utilization increased

(OR = 1.28; 95 % CI: 1.07 – 1.54; Figure 2 ).

DISCUSSION Th is study identifi ed a large population-based cohort of UC

patients who underwent a colectomy over a 13-year period in

both academic and community hospitals in the Calgary Health

Zone. Among hospitalized UC patients, an increase in immuno-

suppressants and biologic utilization was observed over the study

period. Elective colectomy rates fell by 7.5 % per year from over

2 colectomies per 100,000 person-years in the mid 1990s to nearly

1 per 100,000 person-years in 2009. Th e reduction in elective

colectomy rates was a consistent fi nding across three diff erent

analytical approaches indicating that these reductions were mean-

ingful. Rates of emergent colectomy remained stable throughout

the study period. Th e median time from diagnosis to emergent

colectomy was < 3 years, which was signifi cantly shorter than that

for elective colectomy. Th is data may suggest that the window of

opportunity in which to identify and intervene in patients with

an aggressive UC phenotype is narrow and future studies should

focus on identifying predictors of poor prognosis.

Historically, 10-year colectomy rates have been high ( ~ 25 % )

( 1 ). However, recent cohort studies indicate that only 10 % of UC

patients undergo colectomy aft er 10 years of diagnosis ( 2 – 4 ). It

is important to consider what may be leading to these observed

decreases in colectomy rates. In the present study, we observed

an increased use of immunomodulators and infl iximab. Immuno-

modulators nearly doubled from the mid 1990s to the later part

of the 2000s. Similarly, infl iximab prescriptions increased consid-

erably aft er 2005 with nearly one in fi ve hospitalized UC patients


1882 IN

FLA

MM

ATO

RY

BO

WE

L D

ISE

AS

E Kaplan et al.

Table 1 . Characteristics of patients with UC who were admitted to hospital with a fl are and did not undergo colectomy as compared with UC patients who underwent colectomy stratifi ed by emergent vs. elective indication

Characteristics Total ( n =775) UC fl are ( n =338)

Colectomy ( n =437)

P -value (fl are vs.

colectomy)

Elective colectomy ( n =205)

Emergent colectomy ( n =232)

P -value (elective vs. emergent)

Age at operation (year)

25th Percentile 28.0 27.0 30.0 < 0.001 31.0 30.0 0.558

Median 40.0 36.5 43.0 44.0 43.0

75th Percentile 54.0 49.0 55.0 57.0 55.0

Gender, % (n)

Male 56.4 (437) 52.7 (178) 59.3 (259) 0.068 59.5 (122) 59.1 (137) 1.000

Female 43.6 (338) 47.3 (160) 40.7 (178) 40.5 (83) 40.9 (95)

Smoking status, % ( n )

Current 10.9 (79) 13.6 (41) 9.0 (38) 0.132 9.1 (18) 8.8 (20) 0.114

Ex-smokers 31.9 (231) 31.8 (96) 31.9 (135) 26.9 (53) 36.3 (82)

Never 57.2 (415) 54.6 (165) 59.1 (250) 64.0 (126) 54.9 (124)

Missing ( n ) 50 36 14 8 6

Comorbidity, % (n)

0 44.5 (345) 48.2 (163) 41.6 (182) 0.027 45.9 (94) 37.9 (88) 0.175

1 – 2 43.6 (338) 42.3 (143) 44.6 (195) 41.0 (84) 47.8 (111)

> 2 11.9 (92) 9.5 (32) 13.7 (60) 13.2 (27) 14.2 (33)

Length of UC fl are, % (n)

< 2 Weeks 42.8 (303) 21.0 (70) 18.9 (71) < 0.001 2.7 (4) 29.3 (67) < 0.001

2.8 Weeks 19.9 (141) 49.5 (165) 36.8 (138) 19.9 (29) 47.6 (109)

> 8 Weeks 37.3 (264) 29.4 (98) 44.3 (166) 77.4 (113) 23.1 (53)

Missing ( n ) 67 5 62 59 3

Extent of disease, % (n)

Left-sided 30.7 (210) 44.5 (117) 22.0 (93) < 0.001 25.0 (49) 19.5 (44) 0.196

Pancolitis 69.3 (475) 55.5 (146) 78.0 (329) 75.0 (147) 80.5 (182)

Missing ( n ) 90 75 15 9 6

Blood in stool and stool frequency > 5 / day, % (n)

No 24.5 (151) 22.0 (63) 26.6 (88) 0.222 43.1 (50) 17.7 (38) < 0.001

Yes 75.5 (466) 78.0 (223) 73.4 (243) 56.9 (66) 82.3 (177)

Missing ( n ) 158 52 106 89 17

Cancer or dysplasia, % (n)

No NA NA 91.5 (400) NA 82.9 (170) 99.1 (230) < 0.001

Yes 8.5 (37) 17.1 (35) 0.9 (2)

Mesalamine/sulfasalazine b , % (n)

No 21.7 (167) 19.9 (67) 23.1 (100) 0.292 29.9 (60) 17.3 (40) 0.003

Yes 78.3 (601) 80.1 (269) 76.9 (332) 70.1 (141) 82.7 (191)

Missing ( n ) 7 2 5 4 1

Prednisone b , % ( n )

No 11.8 (91) 17.6 (59) 7.4 (32) < 0.001 12.9 (26) 2.6 (6) < 0.001

Yes 88.2 (677) 82.4 (276) 92.6 (401) 87.1 (176) 97.4 (225)

Missing ( n ) 7 3 4 3 1

Table continued on the following page


1883

INF

LAM

MAT

OR

Y B

OW

EL

DIS

EA

SE


became apparent during the 2000s ( 19,20 ). Th e elective colectomy

rate may also have decreased due to a greater appreciation for

long-term complications of pouch surgery (e.g., reduced fertility,

pouchitis). ( 21 ) In addition, education of UC patients, adherence

to medications, and less reluctance from patients to seek medical

care at time of fl are may have changed with time ( 22,23 ). Future

studies should be designed to account for the impact of non-

therapeutic interventions on UC outcomes.

Cohort studies have consistently reported the highest rates of

colectomy soon aft er UC diagnosis ( 1,24,25 ). However, previous

studies have not stratifi ed the colectomy rate by indication for sur-

gery (i.e., elective vs. emergent colectomy). We identifi ed a subset

of UC patients who rapidly progressed to colectomy within 3 years

of diagnosis. Th ese patients accounted for > 50 % of UC patients

who underwent emergent colectomy. Furthermore, three quar-

ters of patients undergoing emergent colectomy had a duration of

fl are < 8 weeks. In addition, emergent colectomy patients were less

likely to be prescribed AZA or 6-MP before colectomy. Th ese UC

patients may represent a distinct ‘ rapidly progressive ’ phenotype of

UC. Consequently, there is a limited ‘ window to intervene ’ medi-

cally in this subset of UC patients. Identifi cation of ‘ rapid progres-

sion ’ UC patients at diagnosis with clinical – genetic – serological

predictive models is needed ( 26 ). Randomized controlled studies

assessing an aggressive ‘ top-down ’ approach (i.e., infl iximab and

receiving the drug before admission or in-hospital in 2009.

In contrast, mesalamine and corticosteroid utilization remained

stable from 1997 to 2009. Th ese results should be interpreted with

caution because correlating the increased use of immunosuppres-

sants and infl iximab and the decreased colectomy rate is ecologi-

cal. Further, we did not observe a clear infl ection point correlating

dropping colectomy rates to increased drug utilization. Neverthe-

less, these results raise the question whether earlier and increase

use of these therapies could further reduce colectomy rates in the

UC population.

However, non-therapy-related factors also contributed to

reduce elective colectomies, including diff ering rates of disease

severity and evolving attitudes toward surgery of both physi-

cians and patients. Also, during the 2000s several guidelines were

published on medical and non-medical management of UC that

may have improved physician education and practices ( 16 – 18 ).

For example, the British Guidelines highlighted the importance

of rapid access and action for UC patients who develop a fl are

of UC ( 16 ). During our study period, the IBD clinic in Calgary

expanded allowing for greater interaction of UC patients with gas-

troenterologists trained in IBD management and access to nurse

practitioners dedicated to IBD care. Furthermore, recognition

and management of important triggers of an UC fl are (e.g., bacte-

rial gastrointestinal infections, in particular Clostridium diffi cile )

Table 1 . Continued

Characteristics Total ( n =775) UC fl are ( n =338)

Colectomy ( n =437)

P -value (fl are vs.

colectomy)

Elective colectomy ( n =205)

Emergent colectomy ( n =232)

P -value (elective vs. emergent)

Azathioprine / 6-mercaptopurine b , % (n)

No 73.3 (562) 76.4 (256) 70.8 (306) 0.085 67.2 (135) 74.0 (171) 0.137

Yes 26.7 (205) 23.6 (79) 29.2 (126) 32.8 (66) 26.0 (60)

Missing ( n ) 8 3 5 4 1

Infl iximab a , % ( n )

No 90.5 (692) 89.0 (298) 91.6 (394) 0.217 93.5 (186) 90.0 (208) 0.225

Yes 9.5 (73) 11.0 (37) 8.4 (36) 6.5 (13) 10.0 (23)

Missing ( n ) 10 3 7 6 1

Duration between diagnosis and surgery (years)

25th Percentile NA NA 1.0 NA 2.0 0.0 < 0.001

Median 3.0 6.0 2.0

75th Percentile 10.0 15.0 7.0

Missing ( n ) 19 12 7

Preoperative stay (days)

25th Percentile NA NA NA NA NA 7.0 NA

Median 10.0

75th Percentile 15.0

NA, not applicable; UC, ulcerative colitis. a Defi ned as medication used before, at the time of, or during admission to hospital. ‘ No ’ refers to no record of drug use in the medical chart.


1884 IN

FLA

MM

ATO

RY

BO

WE

L D

ISE

AS

E Kaplan et al.

AZA / 6-MP) early aft er diagnosis in these UC patients may also

be warranted.

Several limitations of this study should be considered ( 27 ).

First, the study was dependent on available data in the medical

charts. Th e chart review was performed by trained personnel

using standardized data extraction forms with random auditing

for accuracy ( 11,28 ). Th is approach is superior to relying solely

on administrative databases, which have the inherent risk of

misclassifi cation errors. Despite the comprehensiveness of the

chart review, a second limitation of such retrospective studies is

that several important factors that may have infl uenced colec-

tomy rates (e.g., role of nurse practitioner and patient education)

could not be accurately assessed. Th ird, disease severity is diffi -

cult to defi ne in a retrospective chart review and changes in dis-

ease severity overtime may have infl uenced colectomy temporal

trends. We performed three distinct analyses to account for dis-

ease severity: (i) a linear time trend model stratifi ed by emer-

gent (i.e., sicker patients) vs. elective indication; (ii) a logistic

6

5

4

3

2

Inci

denc

e pe

r 10

0,00

0 po

pula

tion

1

0

Year

Total

Elective

Emergent

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

Figure 1 . Temporal trend analyses from 1997 to 2009 of colectomy rates stratifi ed by all colectomy patients, elective colectomies, and emergent colectomies.

Table 2 . Independent predictors of: (1) UC fl are without colectomy compared with UC colectomy; and (2) emergent vs. elective colectomy for UC

Variable

UC fl are vs. colectomy

OR (95 % CI; n =774) a

Emergent vs. elective colectomy

OR (95 % CI; n =436) a

Age (years)

18 – 34 1.00 1.00

35 – 64 2.05 (1.37 – 3.06) 1.26 (0.67 – 2.39)

> 64 1.42 (0.76 – 2.66) 2.26 (0.93 – 5.51)

Comorbidity

0 1.00 Not selected

1 – 2 1.55 (1.05 – 2.28)

> 2 2.26 (1.14 – 4.45)

Length of UC fl are b

< 2 Weeks 1.37 (0.83 – 2.26) 5.31 (1.58 – 17.81)

2 – 8 Weeks 1.00 1.00

> 8 Weeks 1.29 (0.85 – 1.95) 0.13 (0.07 – 0.24)

Blood in stool and > 5 bowel movements per day b

No 1.00 1.00

Yes 0.96 (0.60 – 1.55) 3.83 (1.84 – 7.97)

Disease extent b

Left-sided colitis 1.00 Not selected

Pancolitis 3.33 (2.22 – 4.98)

Previous prednisone

No 1.00 Not selected

Yes 4.68 (3.11 – 7.05)

Previous azathiopurine or 6-mercaptopurine

No 1.00 1.00

Yes 3.37 (1.98 – 5.74) 0.32 (0.16 – 0.63)

Year of admission 0.87 (0.83 – 0.92) 1.12 (1.03 – 1.22)

Duration from diagnosis to surgery b

< 1 Year Not applicable 5.48 (2.18 – 13.79)

1 – 3 Years 1.00

4 – 10 Years 1.18 (0.57 – 2.44)

> 10 Years 0.53 (0.24 – 1.17)

CI, confi dence interval; OR, odds ratio; UC, ulcerative colitis. a Observation was deleted due to missing values for the explanatory variables. b Adjusted for missing indicator variable.

100

90

80

70

60

Pro

port

ion

(%)

50

40

20

30

10

0

1997

1998

1999

2000

2001

2002

2003

2004

Year

2005

2006

2007

2008

2009

Mesalamine or sulfalazinePrednisoneAzathioprime or 6-mercaptopurineInfliximab

Figure 2 . Temporal trend analyses from 1997 to 2009 of drug utilization for UC patients admitted to hospital. Includes all three groups — medically responsive plus those who underwent an elective or emergent colectomy. Drug utilization (prescription at, during, or before admission to hospital) for the following drugs: mesalamine or sulfasalazine, prednisone, azathioprine or 6-mercaptopurine, and infl iximab.


1885

INF

LAM

MAT

OR

Y B

OW

EL

DIS

EA

SE


CONFLICT OF INTEREST Guarantor of the article: Gilaad G. Kaplan, MD, MPH, FRCPC.

Specifi c author contributions : Gilaad G. Kaplan participated in

conceiving the study idea, developing the study design, preparing

the data, performing the analysis, interpreting results, and writing

of the manuscript. Gilaad G. Kaplan confi rms that he has had full

access to all the data in the study and had fi nal responsibility for the

decision to submit for publication. Marie-Claude Proulx partici-

pated in preparing the data, interpreting results, and editing of the

manuscript for intellectual content. James Hubbard participated

in preparing the data, conducting the data analysis, and editing of

the manuscript for intellectual content. Anthony MacLean, Don

Buie, Ali Rezaie, and Gordon Moran participated in interpreting

results and editing of the manuscript for intellectual content. Remo

Panaccione participated in conceiving the study idea, developing

the study design, interpreting results, and writing of the manuscript.

Subrata Ghosh participated in interpreting results and editing of

the manuscript for intellectual content. Cynthia Seow participated

in developing the study design, interpreting results, and writing of

the manuscript. Natalie Molodecky participated in data analysis,

interpreting results, and editing of the manuscript for intellectual

content. All the authors have seen and approved the fi nal version of

the manuscript.

Financial support : Th e study was funded by the M.S.I. Foundation

and supported in part by the Alberta IBD Consortium, which is

funded by an Alberta Innovates — Health Solutions Interdisciplinary

Team Grant.

Potential competing interests : Dr Kaplan has served as a

speaker for Merck, Schering-Plough, Abbott, and UCB Pharma.

He has participated in advisory board meetings for Abbott,

Merck, Schering-Plough, Shire, and UCB Pharma. Dr Kaplan has

received research support from Abbott and Shire. Dr Panac-

cione has served as a speaker, a consultant and an advisory board

member for Abbott Laboratories, Merck, Schering-Plough, Shire,

Centocor, Elan Pharmaceuticals, and Procter and Gamble. He

has served as a consultant and speaker for Astra Zeneca. He has

served as a consultant and an advisory board member for Ferring

and UCB. He has served as a consultant for Glaxo-Smith Kline

and Bristol Meyers Squibb. He has served as a speaker for Byk

Solvay, Axcan, Jansen, and Prometheus. He has received research

funding from Merck, Schering-Plough, Abbott Laboratories,

Elan Pharmaceuticals, Procter and Gamble, Bristol Meyers

Squibb, and Millennium Pharmaceuticals. He has received

educational support from Merck, Schering-Plough, Ferring,

Axcan, and Jansen. Dr Ghosh has served as a speaker for Merck,

Schering-Plough, Centocor, Abbott, UCB Pharma, Pfi zer,

Ferring, and Procter and Gamble. He has participated in ad-hoc

advisory board meetings for Centocor, Abbott, Merck,

Schering-Plough, Proctor and Gamble, Shire, UCB Pharma,

Pfi zer, and Millennium. He has received research funding from

Procter and Gamble, Merck, and Schering-Plough. Cynthia Seow

has served as a speaker for Merck and Schering-Plough. She has

participated in advisory board meetings for Abbott, Merck, and

Schering-Plough. Th e remaining authors declare no confl ict of

interest.

regression model comparing hospitalized medically responsive

UC patients against those who required colectomy with year as a

dependent variable, adjusting for factors related to disease sever-

ity; and (iii) a logistic regression model comparing emergent

vs. elective colectomy patients with year as an adjusted depend-

ent variable. All three analyses supported the observation that

colectomy rates declined over time with a greater correspond-

ing reduction in elective vs. emergent colectomies. Fourth, as the

study population was restricted to severe UC patients admitted

to hospital for medical or surgical management, the results of the

study are only generalizable to this hospitalized population. We

did not study UC patients who were treated only with mesala-

mine and moderate – severe outpatients who were successfully

managed with corticosteroids, immunomodulators, and / or bio-

logics without ever being hospitalized. Nor may our fi ndings be

generalizable to centers with a greater utilization of cyclosporine.

Fift h, although our total study period spanned 13 years, infl ixi-

mab was only introduced in this cohort in 2005, hence with

4 years of longitudinal follow-up until 2009, we cannot rule out

the possibility that infl iximab may delay rather than prevent sur-

gery. Ongoing studies are required to further evaluate the long-

term eff ects of infl iximab on colectomy outcomes. Sixth, we used

the annual population size of the Calgary Health Region to calcu-

late the denominator for colectomy rates, which may not account

for a change in prevalence of UC. However, in a sensitivity analy-

sis using the estimated prevalence of UC in the Calgary Health

Region from 1999 to 2005 ( 14,15 ), the reduction in colectomy

rates were similar. Seventh, we restricted our analysis to adult

UC patients because drug utilization patterns diff er in children.

A separate study has shown that colectomy rates were stable in

pediatric UC patients in our health region during the study time

period ( 29 ). Eighth, correlations between drug utilization pat-

terns and colectomy rates may merely be ecological rather than

causal associations. Finally, other centers will need to replicate

this study to determine whether our fi ndings can be generalized

outside of our large Canadian health region.

In summary, this large population-based study of colectomy

rates over more than a decade, including the pre- and post-

biologic era, has important clinical implications. From 1997 to

2009, we observed a reduction in the number of elective, but

not emergent, colectomies. Although there was a correspond-

ing greater utilization of immunomodulators and infl iximab

during this time period, other non-medication factors may have

contributed to the change in colectomy rates. Identifi cation of

an aggressive phenotype of UC at diagnosis with early introduc-

tion of immunomodulators and / or infl iximab may be required to

alter the natural history of the disease and to reduce the need for

emergent colectomy.

ACKNOWLEDGMENTS We acknowledge the DIMR department for providing data

from the Calgary Health Zone. Dr Kaplan is supported through

a New Investigator Award from the Canadian Institute of Health

Research and a Population Health Investigator Award from Alberta

Innovates — Health Solutions.


1886 IN

FLA

MM

ATO

RY

BO

WE

L D

ISE

AS

E Kaplan et al.

Study Highlights

WHAT IS CURRENT KNOWLEDGE 3 Colectomy for ulcerative colitis occurs in ~ 10 % of patients

within 10 years of diagnosis.

3 A high proportion of ulcerative colitis patients who require colectomy, undergo a colectomy within the fi rst 3 years of diagnosis.

3 Approximately half of all colectomies occur emergently and postoperative outcomes are worse in emergent as compared with elective operations.

WHAT IS NEW HERE 3 From 1997 to 2009 rates of elective colectomy for ulcera-

tive colitis have steadily decreased; however, the rates of emergent colectomies have remained stable.

3 During this time period, the use of immunomodulators and infl iximab for the ulcerative colitis has increased.

3 Ulcerative colitis patients undergoing emergent colectomies have a more rapidly progressive disease course as compared with those requiring elective colectomies.

REFERENCES 1 . Langholz E , Munkholm P , Davidsen M et al. Course of ulcerative colitis:

analysis of changes in disease activity over years . Gastroenterology 1994 ; 107 : 3 – 11 .

2 . Hoie O , Wolters FL , Riis L et al. Low colectomy rates in ulcerative colitis in an unselected European cohort followed for 10 years . Gastroenterology 2007 ; 132 : 507 – 15 .

3 . Solberg IC , Lygren I , Jahnsen J et al. Clinical course during the fi rst 10 years of ulcerative colitis: results from a population-based inception cohort (IBSEN Study) . Scand J Gastroenterol 2009 ; 44 : 431 – 40 .

4 . Targownik LE , Singh H , Nugent Z et al. Th e epidemiology of colectomy in ulcerative colitis: results from a population-based cohort . Am J Gastro enterol 2012 ; 107 : 1228 – 35 .

5 . Cannom RR , Kaiser AM , Ault GT , Beart RW Jr , Etzioni DA . Infl amma-tory bowel disease in the United States from 1998 to 2005: has infl iximab aff ected surgical rates? Am Surg 2009 ; 75 : 976 – 80 .

6 . Jarnerot G , Hertervig E , Friis-liby I et al. Infl iximab as rescue therapy in severe to moderately severe ulcerative colitis: a randomized, placebo controlled study . Gastroenterology 2005 ; 128 : 1805 – 11 .

7 . Gustavsson A , Jarnerot G , Hertervig E et al. Clinical trial: colectomy aft er rescue therapy in ulcerative colitis - 3-year follow-up of the Swedish-Danish controlled infl iximab study . Aliment Pharmacol Th er 2010 ; 32 : 984 – 9 .

8 . Rutgeerts P , Sandborn WJ , Feagan BG et al. Infl iximab for induction and maintenance therapy for ulcerative colitis . N Engl J Med 2005 ; 353 : 2462 – 76 .

9 . Sandborn WJ , Rutgeerts P , Feagan BG et al. Colectomy rate comparison aft er treatment of ulcerative colitis with placebo or infl iximab . Gastro-enterology 2009 ; 137 : 1250 – 60 ; quiz 1520 .

10 . Kaplan GG , McCarthy EP , Ayanian JZ et al. Impact of hospital volume on postoperative morbidity and mortality following a colectomy for ulcerative colitis . Gastroenterology 2008 ; 134 : 680 – 7 .

11 . de Silva S , Ma C , Proulx MC et al. Postoperative complications and mortality following colectomy for ulcerative colitis . Clin Gastroenterol Hepatol 2011 ; 9 : 972 – 80 .

12 . Alberta Health Services Zone Map . http://www.albertahealthservices.ca/745.asp , accessed 26 October 2012 .

13 . Kaplan GG , Hubbard J , Panaccione R . et al. Risk of comorbidities on post-operative outcomes in patients with infl ammatory bowel disease . Arch Surg 2011 ; 146 : 959 – 64 .

14 . Rezaie A , Panaccione R , Fedorak R et al. Time trends in prevalence and incidence of infl ammatory bowel diseases in Alberta: a population-based study . Can J Gastroenterol 2012 ; 26 : A008 .

15 . Molodecky NA , Soon IS , Rabi DM et al. Increasing incidence and preva-lence of the infl ammatory bowel diseases with time, based on systematic review . Gastroenterology 2012 ; 142 : 46 – 54 e42; quiz e30 .

16 . Carter MJ , Lobo AJ , Travis SP . Guidelines for the management of infl amma-tory bowel disease in adults . Gut 2004 ; 53 (Suppl 5) : V1 – 16 .

17 . Cohen JL , Strong SA , Hyman NH et al. Practice parameters for the surgical treatment of ulcerative colitis . Dis Colon Rectum 2005 ; 48 : 1997 – 2009 .

18 . Kornbluth A , Sachar DB . Ulcerative colitis practice guidelines in adults (update): American College of Gastroenterology, Practice Parameters Committee . Am J Gastroenterol 2004 ; 99 : 1371 – 85 .

19 . Ananthakrishnan AN , McGinley EL , Binion DG . Excess hospitalisation burden associated with Clostridium diffi cile in patients with infl ammatory bowel disease . Gut 2008 ; 57 : 205 – 10 .

20 . Nguyen GC , Kaplan GG , Harris ML et al. A national survey of the prevalence and impact of Clostridium diffi cile infection among hospitalized infl ammatory bowel disease patients . Am J Gastroenterol 2008 ; 103 : 1443 – 50 .

21 . Hueting WE , Buskens E , van der Tweel I et al. Results and complications aft er ileal pouch anal anastomosis: a meta-analysis of 43 observational studies comprising 9,317 patients . Dig Surg 2005 ; 22 : 69 – 79 .

22 . Higgins PD , Rubin DT , Kaulback K et al. Systematic review: impact of non-adherence to 5-aminosalicylic acid products on the frequency and cost of ulcerative colitis fl ares . Aliment Pharmacol Th er 2009 ; 29 : 247 – 57 .

23 . Kennedy A , Nelson E , Reeves D et al. A randomised controlled trial to assess the impact of a package comprising a patient-orientated, evidence-based self-help guidebook and patient-centred consultations on disease management and satisfaction in infl ammatory bowel disease . Health Technol Assess 2003 ; 7 iii, 1 – 113 .

24 . Allison J , Herrinton LJ , Liu L , Yu J , Lowder J . Natural history of severe ulcerative colitis in a community-based health plan . Clin Gastroenterol Hepatol 2008 ; 6 : 999 – 1003 .

25 . Gustavsson A , Halfvarson J , Magnuson A et al. Long-term colectomy rate aft er intensive intravenous corticosteroid therapy for ulcerative colitis prior to the immunosuppressive treatment era . Am J Gastroenterol 2007 ; 102 : 2513 – 9 .

26 . Henriksen M , Jahnsen J , Lygren I et al. C-reactive protein: a predictive factor and marker of infl ammation in infl ammatory bowel disease. Results from a prospective population-based study . Gut 2008 ; 57 : 1518 – 23 .

27 . Molodecky NA , Panaccione R , Ghosh S et al. Challenges associated with identifying the environmental determinants of the infl ammatory bowel diseases . Infl amm Bowel Dis 2011 ; 17 : 1792 – 9 .

28 . Ma C , Crespin M , Proulx MC et al. Postoperative complications following colectomy for ulcerative colitis: A validation study . BMC Gastroenterol 2012 ; 12 : 39 .

29 . Soon IS , Wrobel I , deBruyn JC et al. Postoperative complications following colectomy for ulcerative colitis in children . J Pediatr Gastroenterol Nutr 2012 ; 54 : 763 – 8 .


1887

INF

LAM

MAT

OR

Y B

OW

EL

DIS

EA

SE


APPENDIX 1

Comorbidity classifi cation

Category Comorbidity

Coronary artery disease Coronary artery disease; ischemic heart disease; myocardial infarction; peripheral vascular disease

Other cardiovascular Cardiac arrhythmia; valvular disorder; pericardial effusion; pericarditis; congenital heart disease; myocarditis

CHF Congestive heart failure

Cancer Lymphoma; metastatic tumor; solid tumor without metastases

Diabetes Diabetes with complications; diabetes without complications

Venous thromboembolism Deep vein thrombosis; pulmonary embolism; intra-abdominal venous thrombosis

GI C. diffi cile infection; CMV infection; pancreatitis; peptic ulcer disease

Hematological Blood loss anaemia; coagulopathy; cyclical neutropenia; iron defi ciency anaemia; von Willebrand disease

HTN Hypertensions

Liver Fatty liver; primary sclerosing cholangitis; hepatitis (chronic or acute); liver transplant; hepatic failure; portal hypertension (with or without varices)

Neurological Cerebrovascular disease; hemiplegia and paraplegia; carotid artery disease; other neurological disorders (epilepsy, encephalopathy, convulsions), multiple sclerosis, dysphagia and aphasia; Parkinson’s disease

Pulmonary Acute bronchitis; asthma; COPD; sarcoidosis

Renal Renal failure (acute or chronic); nephrectomy

Rheumatoid Ankylosing spondylitis; episcleritis, uveitis and iritis; gout; sacroiliitis; rheumatoid arthritis; systemic lupus erythematosus; vasculitis; Sjogren’s syndrome

Thyroid and adrenal Hyperthyroidism; hypothyroidism; Addison’s disease; adrenal insuffi ciency

GI, gastrointestinal; CMV, cytomegalovirus; COPD, chronic obstructive pulmonary disease.

Decreasing Colectomy Rates for Ulcerative Colitis: A Population-Based Time Trend Study

Documents

Transcript of Decreasing Colectomy Rates for Ulcerative Colitis: A Population-Based Time Trend Study