Colonoscopy is an important tool for colorectal cancer screening and diagnosis, but reports of the incidence of perforation, a serious complication, vary widely, and risk factors have not been clearly identified. Using meta-analysis and a qualitative literature review, the authors systematically assessed the incidence of and risk factors for colonoscopy-associated perforation. Data on 966,172 screening and diagnostic colonoscopies from 38 original studies was combined in random-effects meta-analysis. The overall perforation incidence was 91 (95% CI: 77 to 104) per 100,000 colonoscopies. Screening colonoscopy had a lower incidence rate (41 [95% CI: 8 to 75] per 100,000 colonoscopies) than diagnostic colonoscopy and studies with mixed indications for colonoscopy (102 [95% CI: 86 to 118] per 100,000 colonoscopies). No other risk factor was identified by this meta-analysis. Published evidence suggests advanced age, polypectomy/biopsy, previous abdominal surgery, gastrointestinal comorbidities, and hemodialysis may be associated with increased risk. Despite the widespread use of colonoscopy and the seriousness of bowel perforation, information on risk factors, especially modifiable risk factors, is limited.
Perforations of the colon associated with colonoscopy were addressed in an article in the December 2006 issue of the Pennsylvania Patient Safety Advisory. At that time, the analysts noted 125 to152 perforations reported to the Pennsylvania Patient Safety Authority for a one-year period in Pennsylvania. Since then, the number of reports mentioning perforation with colonoscopy has not changed greatly. The number for the most recent year is estimated to be in the range of 140 to 170. The Authority publishes this meta-analysis to provide facilities with benchmarks to assess their perforation rates for screening and for diagnostic colonoscopies. Facilities that would like to improve their perforation rates may wish to reference a June 2008 Advisory article, “Colon Perforations Complicating Colonoscopies: What is the Best Known Evidence for Prevention?”
Colonoscopy is a standard method of screening for colorectal cancer, the second leading cause of cancer death in the United States.1 In the United States in 2009, an estimated 146,970 new cases of colon or rectal cancer were diagnosed, and 49,920 individuals died.2 Worldwide in 2008, an estimated 695,000 people died from colon or rectal cancer.3 Early detection of lesions by screening has been shown to reduce mortality from colorectal cancer.4,5
Colonoscopy has advantages over other colorectal cancer screening methods: it allows for examination of the entire length of the colon, has high diagnostic sensitivity, and enables the endoscopist to immediately biopsy or remove discovered polyps or other suspicious lesions. However, colonoscopy can result in bowel perforation, a rare but serious and potentially fatal adverse event. Because a range of perforation rates has been reported in the medical literature, the actual risk is unclear.6 Furthermore, risk factors, especially modifiable ones, have not been well defined. The authors’ research group has previously found very little published literature on the prevention of perforation or the identification of modifiable risk factors.7 These deficiencies complicate estimating risk for an individual patient, assessing whether the perforation rate of a particular provider or center is aberrant, and determining cost-effectiveness.
Although previous reviews have addressed this topic,8,9 to the authors’ knowledge, this is the first published full-text review to systematically investigate perforation risk factors and summarize perforation incidence using a meta-analytic method in patients not selected for any particular characteristics. The authors also performed a systematic qualitative literature review to report risk factors identified in primary literature.
The authors searched 12 databases for clinical studies and reference lists published from January 1, 1990, through June, 16, 2010, and hand-searched the results. Databases searched, controlled vocabulary terms, and search strategy are shown in Table 1.
|Cumulative Index to Nursing and Allied Health Literature, Cochrane Central Register of Controlled Trials, Embase, Excerpta Medica database, MEDLINE, PreMEDLINE, Cochrane Database of Systematic Reviews, Cochrane Database of Methodology, Health Technology Assessment Database, Healthcare Standards Directory, National Guideline Clearinghouse, and National Health Service Economic Evaluation Database|||||
|A combination of controlled vocabulary terms and text words, including but not limited to: adverse, colonoscop*, complication*, etiology, harm*, hazard*, iatrogenic, intestine perforation, perforat*, prevention and control, and risk|||||
Search Strategy (Ovid Format, Parallel Strategies Created For Other Databases)|||||
|1||Colonoscopy||Colonoscopy/ or colonoscope|
|2||Perforation ||Exp intestine perforation/ or intestine perforation/ or perforate|
|3||Combine sets||1 and 2|
|4||Eliminate overlap||Remove duplicates from 3|
|5||Limit by publication type||4 not (letter/ or editorial/ or news/ or comment/ or note/ or conference paper)/ or (letter or editorial or news or comment).pt.)|
|6||Risk||5 and (exp risk/ or risk$.ti. or proportional hazards models/ or proportional hazards model/)|
|7||Etiology||5 and (etiology or et.fs.)|
|8||Prevention||5 and pc.fs.|
|9||Complication||5 and (ae.fs. or co.fs.)|
|10||Iatrogenic disease||5 and (iatrogenic disease/ or iatrogenic.ti,ab.)|
The authors included in their study English-language full-length studies published between January 1, 1990, and June 16, 2010, that assessed patients undergoing conventional colonoscopy and reported original data collected since January 1, 1990, from objective records (i.e., patient record charts, databases, or prospective clinical studies, not recall-based surveys or questionnaires).
The authors compiled the original studies’ results into a meta-analysis in order to generate an overall summary estimate of perforation incidence, investigate reasons for differences in findings among studies, and assess the robustness of the authors’ own findings. (For further details, see online-exclusive "Methods Summary.")
The searches identified 535 citations, of which most were excluded, most frequently due to lack of relevance to colonoscopy-associated perforation or for reporting data collected prior to 1990. The authors included 38 clinical studies that addressed perforation incidence, of which 16 also reported risk factors (see Figure 1). The 38 studies were published between 1998 and June 16, 2010, and reported on a total of 966,172 colonoscopies (for a list of all included studies and a summary of their characteristics, see the online-exclusive Table 2). Fifteen studies were conducted in the United States; the rest were conducted in Israel, Canada, Europe, or Asia. Only 11 studies collected perforation data prospectively. Duration of follow-up was periprocedural.
Figure 1. Study Selection Process
Reason for Colonoscopy||
|Tian YF, Liang JT, Chang KJ, et al. The clinical values and pitfalls of colonoscopy; our five year experience.
J Surg Assoc Repub China 1998;31(3):173-9. ||1998||1992-1996||Record review||Mostly symptomatic (70%) or follow-up (17%)||Taiwan|
|Eckardt VF, Kanzler G, Schmitt T, et al. Complications and adverse effects of colonoscopy with selective sedation.
Gastrointest Endosc 1999 May;49(5):560-5.||1999||1995-1997||Prospective case series||Mostly follow-up (29.5%) or symptomatic||Germany|
|Stringer MD, Pinfield A, Revell L, et al. A prospective audit of paediatric colonoscopy under general anaesthesia.
Acta Paediatr 1999 Feb;88(2):199-202.||1999||1994-1998||Prospective case series||Symptomatic||United Kingdom|
|Imperiale TF, Wagner DR, Lin CY, et al. Risk of advanced proximal neoplasms in asymptomatic adults according to the distal colorectal findings.
N Engl J Med 2000 Jul 20;343(3):169-74.||2000||1995-1998||Retrospective data review ||Asymptomatic screening||United States|
|Tran DQ, Rosen L, Kim R, et al. Actual colonoscopy: what are the risks of perforation?
Am Surg 2001 Sep;67(9):845-7; discussion 847-8.||2001||1994-1999||Retrospective database review||65% diagnostic, 35% therapeutic||United States|
|Wexner SD, Garbus JE, Singh JJ, SAGES Colonoscopy Study Outcomes Group. A prospective analysis of 13,580 colonoscopies. Reevaluation of credentialing guidelines.
Surg Endosc 2001 Mar;15(3):251-61.||2001||1998-1999||Prospective database investigation||62.4% diagnostic, 37.6% therapeutic||United States|
|Nelson DB, McQuaid KR, Bond JH, et al. Procedural success and complications of large-scale screening colonoscopy.
Gastrointest Endosc 2002 Mar;55(3):307-14.||2002||1994-1997||Prospective clinical study||Asymptomatic screening||United States|
|Gatto NM, Frucht H, Sundararajan V, et al. Risk of perforation after colonoscopy and sigmoidoscopy: a population-based study.
J Natl Cancer Inst 2003 Feb 5;95(3):230-6.||2003||1991-1998||Retrospective multicenter database review||49% diagnostic, 51% not reported ||United States|
|Gondal G, Grotmol T, Hofstad B, et al. The Norwegian Colorectal Cancer Prevention (NORCCAP) screening study: baseline findings and implementations for clinical work-up in age groups 50-64 years.
Scand J Gastroenterol 2003 Jun;38(6):635-42.||2003||Not reported||Retrospective review||72% therapeutic, 28% diagnostic; all had other positive screening result||Norway|
|Korman LY, Overholt BF, Box T, et al. Perforation during colonoscopy in endoscopic ambulatory surgical centers.
Gastrointest Endosc 2003 Oct;58(4):554-7.||2003||1999||Retrospective record review with survey||Not reported||United States|
|Bowles CJ, Leicester R, Romaya C, et al. A prospective study of colonoscopy practice in the UK today: are we adequately prepared for national colorectal cancer screening tomorrow?
Gut 2004 Feb;53(2):277-83.||2004||2000-2001||Prospective multicenter study||61.2% diagnostic, 6.8% therapeutic, 40.3% surveillance ||United Kingdom|
|Cobb WS, Heniford BT, Sigmon LB, et al. Colonoscopic perforations: incidence, management, and outcomes.
Am Surg 2004 Sep;70(9):750-7; discussion 757-8.||2004||1997-2003||Retrospective database review||Not reported||United States|
|Misra T, Lalor E, Fedorak RN. Endoscopic perforation rates at a Canadian university teaching hospital.
Can J Gastroenterol 2004 Apr;18(4):221-6.||2004||1998-2001||Retrospective care center database||60% diagnostic, 40% therapeutic||Canada|
|Vokurka J. Iatrogenic perforation during an endoscopic examination of the gastrointestinal tract.
Bratisl Lek Listy 2004;105(10-11):387-9.||2004||1998-2002||Retrospective record review||Diagnostic and therapeutic ||Czech Republic|
|Iqbal CW, Chun YS, Farley DR. Colonoscopic perforations: a retrospective review. J Gastrointest Surg 2005 Dec;9(9):1229-35: discussion 1236.||2005||1994-2000||Retrospective database review||Not reported||United States|
|Friedland S, Soetikno R. Colonoscopy with polypectomy in anticoagulated patients.
Gastrointest Endosc 2006 Jul;64(1):98-100.||2006||2004-2005||Retrospective review||Not reported||United States|
|Lee YC, Wang HP, Chiu HM, et al. Factors determining post-colonoscopy abdominal pain: prospective study of screening colonoscopy in 1000 subjects.
J Gastroenterol Hepatol 2006 Oct;21(10):1575-80.||2006||2004-2004||Prospective case series||Asymptomatic screening||Taiwan|
|Levin TR, Zhao W, Conell C, et al. Complications of colonoscopy in an integrated health care delivery system.
Ann Intern Med 2006 Dec 19;145(12):880-6.||2006||1994-2002||Retrospective database review||Screening or follow-up of other screening method||United States|
|Rathgaber SW, Wick TM. Colonoscopy completion and complication rates in a community gastroenterology practice.
Gastrointest Endosc 2006 Oct;64(4):556-62.||2006||2002-2004||Retrospective record review||Not reported||United States|
|Regula J, Rupinski M, Kraszewska E, et al. Colonoscopy in colorectal-cancer screening for detection of advanced neoplasia.
N Engl J Med 2006 Nov 2;355(18):1863-72.||2006||Since 2000||Prospective database analysis||Asymptomatic screening||Poland|
|Tulchinsky H, Madhala-Givon O, Wasserberg N, et al. Incidence and management of colonoscopic perforations: 8 years' experience.
World J Gastroenterol 2006 Jul 14;12(26):4211-3.||2006||1994-2001||Database review||Not reported||Israel|
|Choe JW, Chang HS, Yang SK, et al. Screening colonoscopy in asymptomatic average-risk Koreans: analysis in relation to age and sex.
J Gastroenterol Hepatol 2007 Jul;22(7):1003-8.||2007||1998-2004||Retrospective review||Asymptomatic screening||South Korea|
|Eisele R, Vogelsang E, Kraft K, et al. Screening for colorectal lesions with high-resolution video colonoscopes in a German male average-risk population at 40 to 59 years of age.
Z Gastroenterol 2007 Sep;45(9):952-7.||2007||1998-2003||Prospective case series||Asymptomatic screening||Germany|
|Kahi CJ, Azzouz F, Juliar BE, et al. Survival of elderly persons undergoing colonoscopy: implications for colorectal cancer screening and surveillance.
Gastrointest Endosc 2007 Sep;66(3):544-50.||2007||1999-2000||Retrospective case series||Not reported||United States|
|Lahat A, Yanai H, Menachem Y, et al. The feasibility and risk of early colonoscopy in acute diverticulitis: a prospective controlled study.
Endoscopy 2007 Jun;39(6):521-4.||2007||2004-2006||Prospective trial||Follow-up of acute diverticulitis||Israel|
|Schmilovitz-Weiss H, Weiss A, Boaz M, et al. Predictors of failed colonoscopy in nonagenarians: a single-center experience.
J Clin Gastroenterol 2007 Apr;41(4):388-93.||2007||2000-2005||Retrospective review ||48% symptomatic, 46% screening||Israel|
|Magdeburg R, Collet P, Post S, et al. Endoclipping of iatrogenic colonic perforation to avoid surgery.
Surg Endosc 2008 Jun;22(6):1500-4.||2008||2004-2006||Retrospective database review||71% diagnostic, 29% therapeutic ||Germany|
|Morse JW, Fowler SA, Morse AL. Endoscopist-administered propofol: a retrospective safety study.
Can J Gastroenterol 2008 Jul;22(7):617-20.||2008||1996-2007||Retrospective chart review||Not reported||Canada|
|Paspatis GA, Vardas E, Theodoropoulou A, et al. Complications of colonoscopy in a large public county hospital in Greece. A 10-year study.
Dig Liver Dis 2008 Dec;40(12):951-7.||2008||1996-2006||Retrospective database review||75% symptomatic, 10% surveillance, 3% screening||Greece|
|Rabeneck L, Paszat LF, Hilsden RJ, et al. Bleeding and perforation after outpatient colonoscopy and their risk factors in usual clinical practice.
Gastroenterology 2008 Dec;135(6):1899-1906, 1906.e1.||2008||2002-2003||Retrospective database review||Screening||Canada|
|Thakkar K, El-Serag HB, Mattek N, et al. Complications of pediatric colonoscopy: a five-year multicenter experience.
Clin Gastroenterol Hepatol 2008 May;6(5):515-20.||2008||2000-2005||Retrospective database review||100% symptomatic||United States|
|Arora G, Mannalithara A, Singh G, et al. Risk of perforation from a colonoscopy in adults: a large population-based study.
Gastrointest Endosc 2009 Mar;69(3 Suppl):654-64.||2009||1995-2005||Retrospective database review||78% symptomatic, 22% screening||United States|
|Bair D, Pham J, Seaton MB, et al. The quality of screening colonoscopies in an office-based endoscopy clinic.
Can J Gastroenterol 2009 Jan;23(1):41-7.||2009||2006-2007||Prospective case series||95% screening, mostly average-risk, 3.5% surveillance, 1% test follow up||Canada|
|Singh H, Penfold RB, DeCoster C, et al. Colonoscopy and its complications across a Canadian regional health authority.
Gastrointest Endosc 2009 Mar;69(3 Suppl):665-71.||2009||2004-2006||Retrospective records and chart review||Not reported||Canada|
|Warren JL, Klabunde CN, Mariotto AB, et al. Adverse events after outpatient colonoscopy in the Medicare population.
Ann Intern Med 2009 Jun 16;150(12):849-57.||2009||2001-2005||Retrospective electronic database review ||Indication not reported prospectively||United States|
|Imai N, Takeda K, Kuzuya T, et al. High incidence of colonic perforation during colonoscopy in hemodialysis patients with end-stage renal disease.
Clin Gastroenterol Hepatol 2010 Jan;8(1):55-9.||2010||2001-2008||Retrospective review||Not reported||Japan|
|Lorenzo-Zuniga V, Moreno de Vega V, Domenech E, et al. Endoscopist experience as a risk factor for colonoscopic complications.
Colorectal Dis 2010 Oct;12:273-7.||2010||1995-2008||Retrospective database review||81% diagnostic, 19% therapeutic||Spain|
|Rotholtz NA, Laporte M, Lencinas S, et al. Laparoscopic approach to colonic perforation due to colonoscopy.
World J Surg 2010 Aug;34(8):1949-53.||2010||1997-2008||Retrospective record review||73% diagnostic, 27% therapeutic||Argentina|
Many studies did not report basic demographic characteristics or clinical information. Where reported, most studies examined perforation rates in asymptomatic patients undergoing screening colonoscopy or colonoscopy for diagnostic or therapeutic reasons, but two studies limited enrollment to either screening or follow-up of another screening method (e.g., fecal occult blood test),10,11 and one only examined patients with acute diverticulitis.12 The authors of this meta-analysis did not limit inclusion by patient age, but only two studies focused on pediatric patients.13,14
Few characteristics of the clinicians providing colonoscopy services, such as board certification or experience, were reported. For the most part, gastroenterologists provided colonoscopy services, though sometimes colorectal surgeons, general surgeons, coloproctologists, hepatologists, or medical or surgical trainees performed colonoscopies. Providers were affiliated with a variety of healthcare settings, including university hospitals, general hospitals, and ambulatory centers.
Colonoscopy procedures were also poorly described. Most studies did not report methods of bowel preparation, type of colonoscope used, or polypectomy methods or frequency. Most healthcare providers administered conscious sedation for anesthesia to most or all patients; however, one study reported only 10% of patients requested anesthesia,11 and one study reported administration of anesthesia to less than 5% of patients.15
Databases used in some retrospective assessments included single-hospital databases, third-party payer reimbursement databases, and the Medicare Surveillance Epidemiology and End Results database. Two studies collected data on randomized samples of patients;16,17 the others collected data from a consecutive series. The number of patients enrolled or records examined varied widely, from 2118 to 277,434.19 Some studies had so few patients that they were probably underpowered to detect a rare event (e.g., perforation). Since the authors planned to conduct a meta-analysis, studies meeting inclusion criteria were analyzed regardless of the number of patients enrolled. After running the analysis again to include only studies with at least 10,000 records, the authors did not find a significant difference in summary rate. Most studies did not report how they diagnosed or searched databases for perforation. Of those that did, several database reviews used International Classification of Diseases codes to identify perforations,10,16,17,19,20,21 and two studies reviewed only cases of perforation requiring operative intervention.22,23
Following the combination of all data, statistics indicated large differences among the study findings (I2 = 99.7%). The all-studies summary estimate of perforation incidence was 91 (95% CI: 77 to 104) per 100,000 colonoscopies (Figure 2). The incidence did not substantially change when additional analyses (i.e., cumulative and influence analyses) were conducted, suggesting that no single study has undue influence over the summary effect size, and that the overall perforation rate is not very likely to change appreciably with the publication of additional studies.
Figure 2. Incidence of Perforation per 100,000 Colonoscopies
Statistical investigation (i.e., meta-regression) showed that perforation rates were lower in asymptomatic patients undergoing screening than in symptomatic patients or groups of mixed-indication patients. However, these factors did not explain a significant proportion of the differences in results among studies, including when other factors were taken into consideration. The perforation incidence among asymptomatic individuals seeking screening was 41 (95% CI: 8 to 75) per 100,000 colonoscopies. The remaining studies, which enrolled symptomatic and mixed populations (i.e., both symptomatic and asymptomatic patients) had an incidence of 102 (95% CI: 86 to 118) perforations per 100,000 colonoscopies. Even with the clinical groups considered separately, large differences in perforation incidence rates remained. Additional statistical tests (e.g., sensitivity analyses) suggest the rate is stable among studies enrolling symptomatic and mixed-indication groups of patients, but appears less stable (and therefore more likely to change with the publication of additional studies) in studies with asymptomatic patients.
Additional analyses did not indicate that perforation incidence was significantly related to the following factors: country where the study occurred, number of patients, publication date or first or median year of data collection, or whether data was collected prospectively or retrospectively. No other potential factors could be investigated due to sparse reporting.
Literature Review of Published Risk Factors
Reviewing findings from original studies is useful because it provides findings from patient-level investigations of the data, whereas data was only available for study-level analysis in the above meta-analysis. All factors in the original studies identified as risk factors for perforation (or identified as not being risk factors) are reported in the following text.
Older age was investigated as a risk factor and found to be associated with perforation in six studies, Tian et al.,24 Gatto et al.,16 Levin et al.10, Korman et al.,25 Rabeneck et al.,21 and Arora et al.,19 but not in the two most recently published studies, Imai et al.26 and Rotholz et al.23 Findings regarding sex were inconsistent in five studies: Gatto et al.16 and Arora et al.19 reported that sex was not associated with perforation; Korman et al.25 and Paspatis et al.27 found that women were more likely to suffer perforation; and Rabeneck et al.21 found women had lower rates. Gatto et al.16 and Arora et al.19 found that race was not associated with perforation.
Number of comorbidities was identified as a risk factor by Gatto et al.,16 and higher comorbidity indexes were associated with perforation in Rabeneck et al.21 and Arora et al.19 Diverticulitis or diverticular disease was associated with perforation in two studies, Tian et al.24 and Korman et al.25 Previous abdominal surgery was identified as a risk factor by both Tian et al.24 and Korman et al.25 Arora et al.19 found an increased risk in the presence of obstruction; however, Misra et al.20 did not find an association between the reason for colonoscopy and perforation. Imai et al. found that patients on hemodialysis regimens had a statistically significant increase in occurrence of perforation.26
Wexner et al.28 found that practitioner experience was not associated with perforation, but Rabeneck et al.21 and Lorenzo-Zuniga et al.29 found that greater annual caseload was associated with lower rates of perforation. Cobb et al. reported that the rate of perforation was higher for general surgeons than gastro-enterologists, although only 1 of the 14 perforations in the study occurred when a surgeon performed the colonoscopy, and that surgeon was a resident.30 This difference was not found to be statistically significant (P = 0.353). Arora et al. found that surgeons had higher perforation rates than gastroenterologists but did not find this was true once other factors were taken into account.19 Primary care practitioners had lower rates than gastroenterologists in that study.19 Rabeneck et al.21 did not find any association between provider characteristics and perforation.
Bowles et al.,31 Vokura,32 and Rabeneck et al.21 found that polypectomy/biopsy was associated with an increased perforation rate, but Arora et al.19 did not. Cobb et al. found that four out of six colonoscopies in adults for whom colonoscopy was attempted with a pediatric colonoscope resulted in perforation, and that both cases of perforation that appeared to be due to overinsufflation in the entire series were associated with pediatric colonoscope use.30 However, the difficulties that led to the selection of the pediatric colonoscope may have been the underlying risk factor for perforation rather than the colonoscope itself. Rabeneck et al. did not find a difference in rate between hospital and clinics.21
Gatto et al.16 and Arora et al.19 did not find a significant association between year of data collection and perforation rate, though both observed a trend toward decreased rates over time.
The perforation incidence among asymptomatic patients undergoing colonoscopy for screening was 41 (95% CI: 8 to 75) per 100,000 colonoscopies, which is significantly lower than the rate in the rest of the studies combined, which was 102 (95% CI: 86 to 118) per 100,000 colonoscopies. Summary estimates subgrouped by indication should be more useful for modeling decisions or cost analyses; however, since sensitivity analyses suggest the incidence of perforation among screening studies may be subject to change with the publication of additional data, literature for the application should be monitored.
Several potential sources of bias may affect this data. The authors included data from any country but limited inclusion to English-language studies. Most of the data for the meta-analysis was from studies from the United States, Western Europe, and Canada. Only one study was from Central or South America and none were from Africa. As for any meta-analysis, the possibility that publication bias has affected these numbers cannot be ruled out. Since about 90% of the data came from retrospective studies, it is possible that these could be underestimations due to the mode of data collection, although meta-regression does not suggest prospectively and retrospectively collected data significantly differs. None of the other methodological factors investigated as potential quality indicators were associated with perforation rate either.
Although reporting was too limited to enable thorough investigation, no modifiable risk factors for perforation were identified, limiting the authors’ ability to inform clinical practice. Findings from the individual studies using patient-level data were also few and sometimes conflicting.
Identifying factors associated with an increased risk of perforation, especially modifiable risk factors, is desirable for improving the overall safety of colonoscopy. Although the risk of perforation to any individual is not high, minimizing the perforation rate is important as a greater number of individuals can be expected to seek colonoscopy for screening, diagnostic, and therapeutic purposes. Improving the safety of colonoscopy might improve patient compliance with colorectal cancer screening, reportedly only about 60% in 2006,33 and would improve the cost-effectiveness of colonoscopy by reducing iatrogenic morbidity. The need for reliable information on the identification of risk factors is particularly pressing as the number of individuals undergoing colorectal cancer screening can be expected to increase. To identify means for reducing colonoscopy-associated perforation, prospective study of risk factors is warranted. Such studies could be integrated with clinical care.
Nancy Bacci, RN; R. Bradley Hayward, MD; Ann Ouyang, MD; Robert J. Sinnott, MD; and Joel L. Weissfeld, MD, MPH, contributed to the development of this project. Jonathan R. Treadwell, PhD, contributed to the review process and statistical analysis. Eileen Erinoff, MLIS, contributed to literature searches. Helen Dunn and Tracey Monasterno-Stem organized the retrieval of articles, and Helen Dunn generated the reference list.
- Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005.
CA Cancer J Clin 2005 Jan-Feb;55(1):10-30.
- American Cancer Society (ACS). Cancer facts and figures 2009. Atlanta (GA): ACS; 2009. Also available:
- Projections of mortality and burden of disease, 2002-2030: Mortality—baseline scenario 2008 [website]. [cited 2009 Sep 1]. Geneva: World Health Organization (WHO). Available from Internet:
- Mandel JS, Church TR, Bond JH, et al. The effect of fecal occult-blood screening on the incidence of colorectal cancer.
N Engl J Med 2000 Nov 30;343(22):1603-7.
- Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup.
N Engl J Med 1993 Dec 30;329(27):1977-81.
- Perforations of the colon during colonoscopy. Pa PSRS Patient Saf Advis [online] 2006 Dec [cited 2009 Apr 16]. Available from Internet:
- Colon perforations complicating colonoscopies: what is the best known evidence for prevention? Pa Patient Saf Advis [online] 2008 Jun [2009 Apr 16]. Available from Internet: http://patientsafety.pa.gov/ADVISORIES/Pages/200806_57.aspx.
- Lohsiriwat V. Colonoscopic perforation: incidence, risk factors, management and outcome.
World J Gastroenterol 2010 Jan 28;16(4):425-30.
Panteris V, Haringsma J, Kuipers EJ. Colonoscopy perforation rate, mechanisms and outcome: from diagnostic to therapeutic colonoscopy.
Endoscopy 2009 Nov;41(11):941-51.
- Levin TR, Zhao W, Conell C, et al. Complications of colonoscopy in an integrated health care delivery system.
Ann Intern Med 2006 Dec 19;145(12):880-6.
- Gondal G, Grotmol T, Hofstad B, et al.The Norwegian Colorectal Cancer Prevention (NORCCAP) screening study: baseline findings and implementations for clinical work-up in age groups 50-64 years.
Scand J Gastroenterol 2003 Jun;38(6):635-42.
- Lahat A, Yanai H, Menachem Y, et al.The feasibility and risk of early colonoscopy in acute diverticulitis: a prospective controlled study.
Endoscopy 2007 Jun;39(6):521-4.
- Thakkar K, El-Serag HB, Mattek N, et al. Complications of pediatric colonoscopy: a five-year multicenter experience.
Clin Gastroenterol Hepatol 2008 May;6(5):515-20.
- Stringer MD, Pinfield A, Revell L, et al. A prospective audit of paediatric colonoscopy under general anaesthesia.
Acta Paediatr 1999 Feb;88(2):199-202.
- Eckardt VF, Kanzler G, Schmitt T, et al. Complications and adverse effects of colonoscopy with selective sedation.
Gastrointest Endosc 1999 May;49(5):560-5.
- Gatto NM, Frucht H, Sundararajan V, et al. Risk of perforation after colonoscopy and sigmoidoscopy: a population-based study.
J Natl Cancer Inst 2003 Feb 5;95(3):230-6.
- Warren JL, Klabunde CN, Mariotto AB, et al. Adverse events after outpatient colonoscopy in the Medicare population.
Ann Intern Med 2009 Jun 16;150(12):849-57.
- Friedland S, Soetikno R. Colonoscopy with polypectomy in anticoagulated patients.
Gastrointest Endosc 2006 Jul;
- Arora G, Mannalithara A, Singh G, et al. Risk of perforation from a colonoscopy in adults: a large population-based study.
Gastrointest Endosc 2009 Mar;69(3 Suppl):654-64.
- Misra T, Lalor E, Fedorak RN. Endoscopic perforation rates at a Canadian university teaching hospital.
Can J Gastroenterol 2004 Apr;18(4):221-6.
- Rabeneck L, Paszat LF, Hilsden RJ, et al. Bleeding and perforation after outpatient colonoscopy and their risk factors in usual clinical practice.
Gastroenterology 2008 Dec;135(6):1899-1906, 1906.e1.
- Tran DQ, Rosen L, Kim R, et al. Actual colonoscopy: what are the risks of perforation?
Am Surg 2001 Sep;67(9):845-7; discussion 847-8.
- Rotholtz NA, Laporte M, Lencinas S, et al. Laparoscopic approach to colonic perforation due to colonoscopy.
World J Surg 2010 Aug;34(8):1949-53.
- Tian YF, Liang JT, Chang KJ, et al. The clinical values and pitfalls of colonoscopy; our five year experience.
J Surg Assoc Repub China 1998;31(3):173-9.
- Korman LY, Overholt BF, Box T, et al. Perforation during colonoscopy in endoscopic ambulatory surgical centers.
Gastrointest Endosc 2003 Oct;58(4):554-7.
- Imai N, Takeda K, Kuzuya T, et al. High incidence of colonic perforation during colonoscopy in hemodialysis patients with end-stage renal disease.
Clin Gastroenterol Hepatol 2010 Jan;8(1):55-9.
- Paspatis GA, Vardas E, Theodoropoulou A, et al. Complications of colonoscopy in a large public county hospital in Greece. A 10-year study.
Dig Liver Dis 2008 Dec;
- Wexner SD, Garbus JE, Singh JJ, SAGES Colonoscopy Study Outcomes Group. A prospective analysis of 13,580 colonoscopies. Reevaluation of credentialing guidelines.
Surg Endosc 2001 Mar;15(3):251-61.
- Lorenzo-Zuniga V, Moreno de Vega V, Domenech E, et al. Endoscopist
experience as a risk factor for colonoscopic complications.
Colorectal Dis 2010 Oct;12:273-7.
- Cobb WS, Heniford BT, Sigmon LB, Hasan R, Simms C, Kercher KW, Matthews BDet al. Colonoscopic perforations: incidence, management, and outcomes.
Am Surg 2004 Sep;70(9):750-7; discussion 757-8.
- Bowles CJ, Leicester R, Romaya C, et al.
A prospective study of colonoscopy practice in the UK today: are we adequately prepared for national colorectal cancer screening tomorrow?
Gut 2004 Feb;53(2):277-83.
- Vokurka J. Iatrogenic perforation during an endoscopic examination of the gastrointestinal tract.
Bratisl Lek Listy 2004;105(10-11):387-9.
Centers for Disease Control and Prevention (CDC). Use of colorectal cancer tests—United States, 2002, 2004, and 2006.
MMWR Morb Mortal Wkly Rep 2008 Mar 14;57(10):253-8.
Methods Summary (Online Exclusive)
Data was independently extracted onto standardized forms that were distributed to all co-authors for approval. The authors combined the reported proportions of patients who underwent colonoscopy and experienced colonic perforation in random-effects meta-analysis. Standard errors were calculated using the exact binomial method. The authors explored the substantial heterogeneity (which was defined as I2 ≥50.0%1,2) to attempt to identify factors associated with perforation using meta-regression,3 which reported data was sufficient. Subgroup analyses were conducted for factors with positive meta-regression findings. To avoid an ecologic fallacy (an incorrect conclusion that can occur when patient- or provider-level characteristics are used for investigation of heterogeneity), the authors did not investigate mean patient or provider characteristics (e.g., mean age) using meta-regression. The authors recognize that experts in the field caution against conducting a “fishing expedition” out of concern for Type I errors,4 but since the analysis was exploratory, all analyses were run in which data was sufficient. Although sufficient data was not available to investigate the following factors, these are the types of information the authors sought to explore for potential risk factors:
- Procedural characteristics: Type of colonoscope used, bowel preparation, anesthesia, polypectomy/biopsy fre-quency and methods, setting (e.g., hospital, ambulatory surgical facility), conventional or enterostomal introduction of colonoscope
- Methodological and study characteristics: Number of patients in study, method of patient selection, median year of data collection, first year data was collected, diagnostic criteria used for perforation, retrospective or prospective study design, patient enrollment methods
- Patient characteristics: Age, body mass index (BMI), sex, race, comorbidities, number of co-morbidities, and reason for colonoscopy. If no reason for colonoscopy was provided but the study enrolled all, consecutive, or random patients, then the reason was considered “mixed” (i.e., comprised of both symptomatic and asymptomatic patients).
- Provider characteristics: Experience, case load, board certification, specialty (e.g., gastroenterology, surgery, internal medicine)
Risk of bias (e.g., quality) was not gauged with an instrument. Instead, methodological factors that are potential quality indicators that could affect the reported incidence of perforation, such as retrospective study design, were investigated using meta-regression.
To assess whether the incidence of colonic perforation has been stable over time, the authors conducted a cumulative meta-analysis.5,6 The authors added studies to the evidence base one at a time in increasing chronological order to gauge stability and whether the publication of additional data should be expected to strongly influence the summary effect estimate.
To assess whether any one study exerted a strong impact on the summary statistic, the authors conducted an influence meta-analysis, in which one study is removed at a time and the summary statistic is recalculated each time. The overall summary statistic is considered unstable if removing any of the single studies substantially affects the summary statistic (in this report, a difference of ≥15 perforations per 100,000 colonoscopies was considered unstable). This is a particularly useful analysis when the evidence base includes a study that differs somewhat from the others (e.g., patient characteristics, study protocols).
All analyses and figures were generated using Stata (StataCorp LP, Bryan/College Station, Texas), including a permutation test using the module written by Harbord and Higgins to conduct the meta-regression.7
- Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat
Med 2002 Jun 15;21(11):1539-58.
- Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses.
BMJ 2003 Sep 6;327(7414):557-60.
- Higgins JP, Thompson SG. Controlling the risk of spurious findings from meta-regression.
Stat Med 2004 Jun 15;23(11):1663-82.
- Lau J, Ioannidis JP, Schmid CH. Quantitative synthesis in systematic reviews.
Ann Intern Med 1997 Nov 1;127(9):820-6.
- Ioannidis JP, Contopoulos-Ioannidis DG, Lau J. Recursive cumulative meta-analysis: a diagnostic for the evolution of total randomized evidence from group and individual patient data.
J Clin Epidemiol 1999 Apr;52(4):281-91.
- Lau J, Schmid CH, Chalmers TC. Cumulative meta-analy-sis of clinical trials builds evidence for exemplary medical care.
J Clin Epidemiol 1995 Jan;48(1):45-57; 59-60.
- Harbord R, Higgins JP. Statistical Software Components, Boston College Department of Economics. METAREG: Stata module to perform meta-analysis regression [soft-ware download]. [accessed 2004 Nov 3]. Chestnut Hill (MA): EconPapers. Available from Internet: