Proctocolectomy is performed on patients with ulcerative colitis (UC) who fail medical management, develop complications (eg, toxic megacolon), or develop colonic dysplasia (1). The risk of colectomy may be higher in children with UC compared with adults (2–4), with 1 study reporting a cumulative colectomy rate of 8% at 1 year, 15% at 3 years, and 20% at 5 years (4). Despite higher rates of colectomy in children, the majority of studies evaluating outcomes following colectomy in UC have focused on adult populations (5,6).
Consequently, data on postoperative complication following colectomy in children are limited. The incidence of in-hospital postoperative complications in pediatric populations was reported to be between 30% and 41% (7,8). Postoperative complications have included wound infection, abscess formation, and bowel obstruction (7–9). In contrast to adult studies, pediatric studies have not reported mortality following colectomy for UC (6–8,10). Severity of disease and preoperative corticosteroid use have been shown to increase early postoperative complications in patients with UC who underwent colectomy (11). Similarly, preoperative corticosteroid use and anemia were associated with postoperative infectious complications in children (12); however, population-based studies of in-hospital complications following colectomy for UC in children are lacking. Consequently, outcomes from prior studies may have been skewed by selection and referral biases.
Population-based administrative health care databases are widely used in the study of epidemiology, health services use, and outcomes of diseases. Administrative databases have been used to study postoperative complications following colectomy in adults with UC (13). Validation studies have shown that administrative data correctly identified patients with UC who underwent colectomy in 86% of cases (14); however, the diagnostic coding was not validated in the pediatric population.
Hence, the objectives of our study were to evaluate the validity of administrative coding in identifying children with UC who underwent colectomy in an administrative database, assess the in-hospital postcolectomy complications and their clinical predictors in a population-based cohort of children with UC, and determine the annual incidence of colectomy in pediatric patients with UC.
The Data Integration, Measurement, and Reporting (DIMR) hospital discharge abstract administrative databases captured all of the hospitalizations in the Calgary Health Zone of Alberta Health Services, Alberta, Canada. The Calgary Health Zone is a public, single-payer system and provides all levels of medical and surgical care to the residents of Calgary and >20 nearby smaller cities, towns, villages, and hamlets, a population of 1,275,664 (∼130,000 children) in 2008 (15). The DIMR database contains 42 diagnostic and 25 procedural coding fields. The International Classification of Disease (ICD), Ninth Revision, Clinical Modification (ICD-9-CM) was used up to 2001, whereas the ICD-10-CA and the Canadian Classification of Health Intervention coding were used after 2001.
Alberta Children's Hospital (ACH) is the only referral center for pediatric gastroenterology and surgery in southern Alberta, serving a population of 1.7 million. All of the cases of pediatric inflammatory bowel disease (IBD) have been prospectively identified since 1993 and retrospectively collected from 1983 to 1998 at ACH to form the Pediatric IBD Database.
The DIMR administrative discharge database was used to identify all of the children (18 years old or older) who were admitted to hospital with a diagnosis of UC (ICD-9-CM 556.X or ICD-10-CA K51.X) and were coded as having a colectomy (ICD-9-CM: 45.7, 45.8 or Canadian Classification of Health Intervention: 1.NM.87, 1.NM.89, 1.NM.91, 1.NQ.89, 1.NQ.90) between January 1, 1996 and December 31, 2009. The ACH IBD database was separately searched to identify all of the patients with UC who underwent colectomy from 1983 to 2007.
Hospital medical charts were reviewed for all of the patients identified to confirm the diagnosis of UC. The study population was restricted to patients diagnosed with UC and who underwent colectomy before the age of 19 years from 1983 to 2009. A total of 36 patients were identified from the DIMR and ACH databases; however, 3 were excluded because they were subsequently diagnosed as having Crohn disease (CD) and 3 were excluded because hospital charts were not available for review.
Among the 30 patients with UC who underwent colectomy, we developed a separate cohort of patients UC who were 16 years old or older when they underwent colectomy, resided in the Calgary Health Zone, and were operated on between 1983 and 2009. This subpopulation (n = 18) was used to calculate the incidence rate of colectomy for pediatric patients with UC in the Calgary Health Zone. Patients 16 years old or older were used because those older than 16 years may be evaluated by adult gastroenterologists and never seen at the ACH.
The primary outcome was postoperative complications, defined as experiencing at least 1 postoperative complication that was graded II or higher according to the Clavien-Dindo Classification of Surgical Complications (online-only Appendix [http://links.lww.com/MPG/A85]) (16). Postoperative complications were defined as unexpected medical events that occurred between the time of colectomy and discharge from the hospital. Patients who experienced >1 postoperative complication were assigned the most severe postoperative complication. Complications were grouped as follows: death, gastrointestinal, cardiovascular, infections, wound, pulmonary, renal or endocrine, and neurological.
Demographic and clinical data extracted from chart review included age at colectomy, age at UC diagnosis, duration of disease (defined as time between UC diagnosis and date of surgery), sex, city of residence (stratified as living within or outside the Calgary Health Zone), length of flare before hospital admission (categorized as <2 weeks or ≥2 weeks), extent of disease (classified as pancolitis vs left-sided colitis), and length of hospital stay before colectomy (stratified as ≤2 weeks and >2 weeks). The use of mesalamine or sulfasalazine, azathioprine or 6-mercaptopurine, methotrexate, prednisone or intravenous corticosteroids, cyclosporine, tacrolimus, and infliximab before admission and during hospitalization was recorded.
The type of colectomy was stratified as elective versus emergent colectomy. An operation was defined as “elective” if the decision to operate on the patient with UC was made before admission to hospital. An “emergent” colectomy was decided during the admission (eg, acute complication or refractory to in-hospital medical management). For patients with emergent colectomy, the duration between admission and colectomy and the duration of bowel rest (defined as days of nothing by mouth before colectomy) were also recorded.
Validation of Administrative Data
Using data obtained from the chart review as the criterion standard, we calculated the sensitivity and positive predictive value (PPV) with 95% confidence interval of administrative codes identifying pediatric patients with UC who underwent colectomy. Sensitivity was defined as the proportion of true cases identified in the administrative database, and PPV was the proportion of cases identified in the administrative database that were true cases.
In the assessment of postoperative complications following colectomy, the Fisher exact test was used to compare proportions and the Wilcoxon rank sum test was used for continuous variables, expressed as medians with first quartile (Q1) and third quartile (Q3). Univariate logistic regression was performed to examine the association between study variables and postoperative complication defined as Clavien-Dindo Classification II or higher. Multivariate logistic regression was not performed because of the small number of patients.
A subgroup analysis was performed on the cohort of patients with UC who had an emergent colectomy. In this cohort, the duration of flare before admission, the time from admission to colectomy, and the duration of bowel rest before colectomy were assessed as potential predictors of postoperative complications.
Time Trend Analysis
The yearly incidence of colectomies for pediatric UC was calculated by dividing the annual total number of colectomies by the annual population size for the Calgary Health Zone. The average annual percentage change from 1983 to 2009 was calculated using a generalized linear model that assumed a Poisson distribution. Poisson regression analysis was performed to evaluate the change in incidence of colectomy in patients with UC over time.
All of the statistical analyses were conducted using STATA version 11.0 (StataCorp, College Station, TX). P values <0.05 were considered to be statistically significant. This was approved by the University of Calgary's conjoint health research ethics board.
Validation of Administrative Data
Twenty-four patients coded for both UC and colectomy were identified in the administrative database from 1996 to 2007. All of these patients had UC and colectomy confirmed by chart review, yielding a PPV of 100%. One additional true case was identified in the ACH IBD database, yielding a sensitivity of 96% (95% confidence interval 80%–99.9%).
A total of 30 children with UC underwent colectomy from 1983 to 2009 for medically refractory disease (29 patients) and toxic megacolon (1 patient). The baseline characteristics of the cohort are presented in Table 1. The median age at diagnosis of UC was 10.9 years (Q1, Q3: 7.0, 14.4), and the median age at colectomy was 12.1 (Q1, Q3: 9.4, 16.2). All of the patients had documented pancolitis and 63% underwent emergent colectomy.
Postoperatively, 33% experienced at least 1 complication. No in-hospital mortality occurred in this cohort. A total of 11 Clavien-Dindo grade II or higher complications were recorded in this cohort (Table 2). The most frequent postoperative complications were infections (17%), with the urinary tract being the most common site (10%). The majority of postoperative complications were Clavien-Dindo grade II (55%) followed by grade IV (36%). Table 3 provides a detailed overview of the 10 patients with UC who experienced a postoperative complication. Patients who had an emergent colectomy were more likely to have postoperative complications than those who had elective colectomy (90% vs 50%; P = 0.03). IBD medication use before hospitalization was not associated with postoperative complications.
In the subgroup analyses performed on the cohort of patients with UC who had emergent colectomy, a disease flare of ≥2 weeks before hospital admission was associated with development of a postoperative complication (60% vs 0%; P = 0.03). Among patients with UC who underwent emergent colectomy, prolonged time to colectomy following admission of >2 weeks was also an independent predictor of postoperative complication (80% vs 20%; P = 0.04).
Time Trend Analysis
The average annual incidence of colectomy for pediatric UC per 100,000 person-years in the Calgary Health Zone from 1983 to 2009 was 0.059, and it ranged from 0.021 during the period 1983–1989 to 0.095 for 1995–1999 (Fig. 1). Poisson regression analysis did not reveal a statistically significant temporal trend in the incidence of colectomy (P > 0.05).
This is the first population-based study to assess in-hospital postoperative complications in children with UC who underwent colectomy. The population-based study design minimized the risk of selection and referral bias that may have influenced prior studies (7,8). During the last 2 decades, we identified a total of 30 children with UC who underwent colectomy. Misclassification of cases was unlikely in our study because we searched 2 independent databases and all of the cases were confirmed by medical chart review, including surgical and pathological reports.
Medically refractory disease was the main reason for colectomy in children with UC. Postoperative complications, defined as Clavien-Dindo grade II or higher, were seen in one-third of these children, similar to previously published pediatric studies (7,9) and adult patients with UC undergoing colectomy (27.5%) (6). The Clavien-Dindo Classification of Surgical Complications is a standardized classification and has been validated in general surgery populations (16,17). This classification system focuses on the therapeutic consequences of a complication, emphasizing the level of intervention required. The Clavien-Dindo Classification has been used in the assessment of complications after percutaneous nephrolithotomy in children (18), and our study is the first to our knowledge to use this classification in the assessment of in-hospital complications postcolectomy in children with UC.
The most common complications were infections, especially urinary tract infections. Because patients undergoing surgical procedures frequently receive urinary catheterizations, clinicians should be judicious about the duration of catheter placement to minimize the frequency of urinary tract infections. We did not observe an association between immunosuppressive therapies (eg, azathioprine) and postcolectomy infections; however, this finding should be interpreted cautiously because of our low sample size. In the era of widespread immunosuppressive therapies for the management of UC, clinicians should be vigilant about the risk of infections in children with UC facing colectomy. In contrast to the risk of mortality among adult patients with UC undergoing colectomy (1.5%) (6), inpatient mortality was not observed in our cohort, which is compatible with previously published pediatric studies (7–9).
Postoperative complications were higher in patients with emergent colectomy compared with patients with elective colectomy. Prolonged disease flare before admission predicted postoperative complications among patients with emergent colectomy. Furthermore, delay to colectomy following admission to hospital was associated with postoperative complications. Patients with UC who experience delay to surgery may experience worse postoperative complications because of greater disease severity, prolonged preoperative exposure to corticosteroids, and malnutrition (12). A large cohort of >600 adult patients with UC who underwent colectomy also demonstrated that delay to surgery was independently associated with a >3-fold increased risk of postoperative complications in adults (6).
The association between prolonged time to colectomy and postoperative complications emphasizes the need to initiate salvage medical therapy (ie, cyclosporine or infliximab) and/or colectomy in a timely fashion. The 2011 consensus for managing acute severe UC in children recommended salvage therapy on day 5 to 7 of admission for severe acute UC who are refractory to intravenous corticosteroids (19). Discussion with families of therapeutic options for acute severe UC in children needs to occur before or at the time of admission to hospital to facilitate the appropriate initiation of salvage therapy in a timely manner.
In our study, chart review was used as the criterion standard to assess the validity of ICD-9 and -10 codes in administrative database for the coding of UC and colectomy. A combination of UC and colectomy codes yielded high sensitivity (96%) and PPV (100%), indicating high diagnostic accuracy. The diagnosis of UC was made based on the available clinical, radiological, and endoscopic data at the time of colectomy; however, 9% of patients in our cohort had a subsequent change in diagnosis to CD. A change in diagnosis as an outpatient would not reflect an error in coding UC during hospital admission for colectomy. The 9% diagnosis change is in keeping with the reported diagnosis of CD after colectomy for children with UC ranging from 3% to 16% (20–22). Because we did not have an unselected control group, specificity and negative predictive values were not determined; however, owing to the rare incidence of colectomy for UC in children, the specificity and negative predictive values are likely high. Overall, the high sensitivity and PPV of ICD codes enable the identification of a cohort of patients with UC who had colectomy in an administrative database for research purposes.
The annual incidence of colectomy in pediatric UC in the Calgary Health Zone has remained relatively stable during the last 25 years. The increased use of immunomodulators for UC has not resulted in a corresponding decrease in colectomy rates in children. Approval of infliximab in adults with UC occurred in 2006. A study in the Calgary Health Zone demonstrated that infliximab use in adults for UC increased considerably after 2006 and that the rate of colectomy from 2006 to 2009 was nearly half the rate observed from 1997 to 2005 (23). The effect of infliximab in reducing colectomy rates in pediatric patients with UC is not known, but it should be evaluated in future studies.
Several limitations to the present study should be pointed out. The chart review was retrospective and hence depended on the accuracy and completeness of the records by clinical providers. Also, quantification of disease activity such as the Pediatric Ulcerative Colitis Activity Index, which predicts corticosteroid response and need for salvage therapy in severe acute UC (24,25), was not possible because recording clinical symptoms was not standardized in the chart. We indirectly addressed disease activity by stratifying our analysis by elective (ie, typically less sick) and emergent colectomies. Our study is also limited by the small sample size, which reduced the identification of rare complications and prevented us from evaluating the effect of clinical predictors on postoperative complications using multivariable logistic regression models; however, because we validated ICD-9/-10 coding of UC and colectomy in the pediatric population, these limitations can be overcome in future studies by identifying large sample sizes derived from administrative databases. Another limitation of our study is the absence of long-term complications following colectomy in children with UC. This would be best addressed in a future prospective study in which long-term complications can be captured accurately or by using administrative databases to evaluate health service use following discharge postcolectomy.
In conclusion, colectomy for pediatric UC was found to be rare and surgical rates have been stable. Postoperative complications following colectomy in children with UC was common (approximately 1 in 3) and comparable to adults undergoing colectomy (6). The predominant risk factor for postoperative complication was emergent colectomy compared to elective. Furthermore, delays to surgery enhanced the risk of postoperative complications among patients with emergent colectomy. Consequently, optimal management of an acute flare of UC likely requires the close monitoring of disease severity and the timely introduction of immunomodulators and/or biologicals among children failing management with corticosteroids. Although our small sample size precluded more comprehensive analysis of preoperative risk factors for postoperative complications, we validated ICD codes in pediatric patients with UC. Thus, future studies using administrative databases should be designed to evaluate postoperative complications in a large cohort of children with UC.
1. Nicholls RJ. Review article: ulcerative colitis—surgical indications and treatment. Aliment Pharmacol Ther 2002; 16 (suppl 4):25–28.
2. Cosnes J, Gower-Rousseau C, Seksik P, et al. Epidemiology and natural history of inflammatory bowel diseases. Gastroenterology 2011;140:1785–94.
3. Falcone RA Jr, Lewis LG, Warner BW. Predicting the need for colectomy in pediatric patients with ulcerative colitis. J Gastrointest Surg 2000; 4:201–206.
4. Gower-Rousseau C, Dauchet L, Vernier-Massouille G, et al. The natural history of pediatric ulcerative colitis: a population-based cohort study. Am J Gastroenterol 2009; 104:2080–2088.
5. 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; 134:680–687.
6. 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.
7. El-Baba M, Lin CH, Klein M, et al. Outcome after surgical intervention in children with chronic inflammatory bowel disease. Am Surg 1996; 62:1014–1017.
8. Rintala RJ, Lindahl HG. Proctocolectomy and J-pouch ileo-anal anastomosis in children. J Pediat Surg 2002; 37:66–70.
9. Sako M, Kimura H, Arai K, et al. Restorative proctocolectomy for pediatric patients with ulcerative colitis. Surg Today 2006; 36:162–165.
10. Pakarinen MP, Natunen J, Ashorn M, et al. Long-term outcomes of restorative proctocolectomy in children with ulcerative colitis. Pediatrics 2009; 123:1377–1382.
11. Mahadevan U, Loftus EV Jr, Tremaine WJ, et al. Azathioprine or 6-mercaptopurine before colectomy for ulcerative colitis is not associated with increased postoperative complications. Inflamm Bowel Dis 2002; 8:311–316.
12. Markel TA, Lou DC, Pfefferkorn M, et al. Steroids and poor nutrition are associated with infectious wound complications in children undergoing first stage procedures for ulcerative colitis. Surgery 2008; 144:540–545.
13. 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.e1–687.e1.
14. Ma C, Crespin M, Proulx M, et al. Accuracy of administrative data in identifying ulcerative colitis patients presenting with acute flare and undergoing colectomy. Am J Gastroenterol 2010; 105 (suppl 467):A1270.
16. Clavien PA, Barkun J, de Oliveira ML, et al. The Clavien-Dindo classification of surgical complications: five-year experience. Ann Surg 2009;250:187–96.
17. Dindo D, Demartines N, Clavien P-A. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004; 240:205–213.
18. Ozden E, Mercimek MN, Yakupoglu YK, et al. Modified Clavien classification in percutaneous nephrolithotomy: assessment of complications in children. J Urol 2011; 185:264–268.
19. Turner D, Travis SPL, Griffiths AM, et al. Consensus for managing acute severe ulcerative colitis in children: a systematic review and joint statement from ECCO, ESPGHAN, and the Porto IBD Working Group of ESPGHAN. Am J Gastroenterol 2011; 106:574–588.
20. Mortellaro VE, Green J, Islam S, et al. Occurrence of Crohn's disease in children after total colectomy for ulcerative colitis. J Surg Res 2011;170:38–40.
21. Evers EA, Pfefferkorn MD, Steiner SJ. Factors predictive of Crohn disease following colectomy in medically refractory pediatric colitis. J Pediatr Gastroenterol Nutr 2009; 48:283–286.
22. Patton D, Gupta N, Wojcicki JM, et al. Postoperative outcome of colectomy for pediatric patients with ulcerative colitis. J Pediatr Gastroenterol Nutr 2010; 51:151–154.
23. Kaplan GG, Seow CH, Ghosh S, et al. Decreasing colectomy rates for ulcerative colitis patients following increased utilization of immunomodulators and biologics: a population-based time trend study. Gastroenterology 2011;140(5 suppl 1):S267.
24. 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.
25. Turner D, Walsh CM, Benchimol EI, et al. Severe paediatric ulcerative colitis: incidence, outcomes and optimal timing for second-line therapy. Gut 2008; 57:331–338.