Colonoscopy and its complications across a Canadian regional health authority

ORIGINAL ARTICLE: Clinical Endoscopy

Colonoscopy and its complications across a Canadian regionalhealth authority

Harminder Singh, MD, MPH, Robert B. Penfold, PhD, Carolyn DeCoster, PhD, RN, MBA, Lisa Kaita, RN, BN,Cindy Proulx, DipISS, Gerry Taylor, BSc, Charles N. Bernstein, MD, Michael Moffatt, MSc, MD

Winnipeg, Manitoba, Canada

Background: Defining the complication rate of endoscopy performed across an entire city will capture usual asopposed to referral center data.

Objective: Our purpose was to evaluate the current practice of colonoscopy and complications associated withlower GI endoscopy in usual clinical practice.

Design: All admissions within 30 days of an outpatient lower GI endoscopy at any of the 6 adult-care Winnipeghospitals were identified. This includes endoscopy for both complex and routine patients. A chart audit of allcases with potential complications was performed.

Results: A total of 24,509 outpatient lower GI endoscopies for adults were performed at the 6 hospitals over the2 study years (April 1, 2004, to March 31, 2006). There were 303 admissions with potential complications. Thecolonoscopy completion rate was 65% (72% for gastroenterologists vs 59% for general surgeons, P ! .005).Quality of bowel preparation and nature of polyps were often not documented. The overall rate of complica-tions was 2.9/1000 procedures; the perforation rate after polypectomy was 1.8/1000; and the postpolypectomybleeding rate was 6.4/1000. Most (67%) complications were recognized after discharge for the index procedure.The complication rate was highest for the endoscopists performing fewer than 200 procedures per year(5.4/1000 vs 2.7/1000 for the rest, P Z .02, relative risk 2 [95% CI, 1.1-3.7]).

Limitations: Chart audit was limited to cases requiring admission within 30 days of the index procedure.

Conclusions: The overall complication rate after lower GI endoscopy in usual clinical practice in Winnipeg iscomparable to that previously reported. A higher complication rate after endoscopy by low-volume endoscopistsneeds to be further evaluated. The reporting of endoscopy must be standardized to enhance outcomes inter-pretation. (Gastrointest Endosc 2009;69:665-71.)

Population-based colorectal cancer screening programshave recently started in 2 Canadian provinces and are inthe developmental stages in several others. Even beforethe onset of these programs, the rates of lower GI endos-

Abbreviations: ASGE, American Society for Gastrointestinal Endoscopy;

DSS, decision support system; MHHL, Manitoba Health and Healthy Liv-

ing; NCCRT, National Colorectal Cancer Roundtable; WRHA, Winnipeg

Regional Health Authority.

DISCLOSURE: The following authors disclosed financial relationships

relevant to this publication: H. Singh is supported in part by a Dr F.

W. Du Val Clinical Research Professorship Award. C. N. Bernstein is

supported in part by a Research Scientist Award of the Crohn’s and

Colitis Foundation of Canada. All other authors disclosed no

financial relationships relevant to this publication.

Copyright ª 2009 by the American Society for Gastrointestinal Endoscopy

0016-5107/$36.00

doi:10.1016/j.gie.2008.09.046

www.giejournal.org Volume 69

copy had been rising rapidly and are expected to continueto rise because of the increasing awareness of colorectalcancer in Canada.1 The 2006 Behavioral Risk Factor Sur-veillance System survey from the United States suggeststhat 56% of the U.S. population over the age of 50 yearshad at least 1 lower GI endoscopy in the 10 years preced-ing the survey.2

Recent studies suggest that there is a wide variation inthe performance of colonoscopy leading to varying out-comes among different endoscopists.3 The Quality Assur-ance Task Group of the U.S. National Colorectal CancerRoundtable (NCCRT) has recently developed a standard-ized colonoscopy reporting and data system to facilitatequality improvement.4 The NCCRTrecognized that withoutadequate, consistent, and standardized documentation it isdifficult to evaluate the quality of care or to conduct com-parative studies across different settings. Furthermore,

, No. 3 : Part 2 of 2 : 2009 GASTROINTESTINAL ENDOSCOPY 665

Colonoscopy and its complications Singh et al

complete documentation of the endoscopic episodes isnecessary to communicate between clinicians or even toprompt the memory of the clinician performing the indexprocedure at a later date. However, in much of Canada, en-doscopy is still reported on paper-based reporting forms,and there are no standardized reporting formats. Thereare no published reports of Canadian audits of the currentreporting practice or performance of lower GI endoscopyin the usual clinical practice.

Moreover, although lower GI endoscopy is rarelyassociated with the risk of potentially life-threateningcomplications, such as excessive postbiopsy and postpoly-pectomy bleeding, colon wall perforation, and cardiorespi-ratory compromise, there are no published reports of thecomplication rates outside tertiary care centers in Canada.5

Another continuing Canadian study of colonoscopy-associated complications is restricted to evaluating therisk of perforations and postpolypectomy bleeding aftercolonoscopy and is not evaluating the cardiorespiratoryor other serious complications or complications after flex-ible sigmoidoscopy.6 Most of the previous reports on com-plications associated with lower GI endoscopy from othercountries have also been limited to the risk of colon perfo-rations and bleeding.7,8

Previous reports of lower GI endoscopy-associated com-plications in the usual clinical practice have been unable tostudy the effect of differences among the endoscopistsbecause of the uniformity of the experience and volumeof procedures performed by the involved endoscopists.9

The aims of our study were to evaluate the currentpractice of lower GI endoscopy, focusing on the reportingand colonoscopy completion rates, and to determine therates of lower GI endoscopy–associated complications inthe usual clinical practice. We also report on the currentmanagement practice of lower GI endoscopy–associatedcomplications in usual clinical care.

METHODS

We performed a retrospective chart audit of lower GI en-doscopies performed on adult (aged O16 years) outpatientsat Winnipeg hospitals between April 1, 2004, and March 31,2006.

Manitoba, as elsewhere in Canada, has a universalhealth care system with a single health care provider(the provincial government). Health care in Manitoba isadministered by 11 regional health authorities. WinnipegRegional Health Authority (WRHA) oversees all hospitals(n Z 6) and nursing homes in the city of Winnipeg, alongwith several outpatient facilities. The WRHA accounts for56% of the provincial population and 47% of health careexpenditures in the province. All hospitals in Manitoba ab-stract admission and discharge information on outpatient(day surgery) endoscopies performed in the hospitals.Hospital discharge abstracts are reported to Manitoba

666 GASTROINTESTINAL ENDOSCOPY Volume 69, No. 3 : Part 2 of

Capsule Summary

What is already known on this topic

d GI endoscopy is associated with the risk of life-threatening complications, such as excessive postbiopsyand postpolypectomy bleeding, colon wall perforation,and cardiorespiratory compromise.

What this study adds to our knowledge

d A review of 24,509 adult outpatient GI endoscopiesperformed at 6 Winnipeg hospitals over 2 years resultedin an overall complication rate of 2.9 per 1000procedures.

d The complication rate was highest for endoscopistsperforming fewer than 200 procedures each year.

Health and Healthy Living (MHHL), which reports to theCanadian Institute for Health Information. MHHL is theprovincial agency, with the overall responsibility for healthcare in the entire province. In addition to submitting toMHHL, all hospitals in Winnipeg also submit hospital dis-charge abstracts to the WRHA, which maintains a decisionsupport system (DSS) to aid in the planning of the ser-vices in the city. About two thirds of the lower GI endos-copies in the province are performed in Winnipeg, withmost (87%) of the procedures performed in the 6 hospi-tals covered by this audit.

We performed an electronic search of WRHA DSS toidentify all individuals who had an admission within 30days of the initial outpatient lower GI endoscopy to oneof the hospitals in Winnipeg. The Canadian Classificationof Interventions codes 1.NM.??.BA* and 2.NM.??.BA*were used to identify the lower GI endoscopies. All admis-sions with a potential cardiovascular, GI, renal, or a pulmo-nary complication associated with a lower GI endoscopywere identified according to an exhaustive list of 129International Classification of Disease, 10th revision,CA codes (list available from authors, at request). Thislist was developed by discussion with several managersof medical information departments (medical records) inWinnipeg. The sensitivity of this list was validated by re-viewing the charts of a 20% sample (n Z 118) of 30-dayadmissions after lower GI endoscopy that did not containthese codes; none of these additional admissions wereassociated with endoscopy-associated complications.

A nurse auditor abstracted information on demograph-ics, indication for the index procedure, extent of the colonexamined, duration of the procedure, and documentationof bowel preparation on the index procedure and detailsof complications and their management. An endoscopist(H. S.) reviewed all the abstracted information. Additionaldiscussion with a second endoscopist resolved all unclearcases.

2 : 2009 www.giejournal.org

Singh et al Colonoscopy and its complications

Acute myocardial infarctions, renal failure, dehydration,and intestinal obstruction were attributed to be related tothe index procedure if the onset of symptoms was within2 days of the index procedure. Similarly, episodes of pneu-monia and acute diverticulitis were attributed to the indexprocedure if the onset of symptoms was within 4 days ofthe index procedure. There were no cases of pneumoniaor acute diverticulitis with onset of symptoms betweendays 4 and 7 from the index procedure. Postpolypectomysyndrome was defined as in other studies as new-onset se-vere abdominal pain (indicative of transmural burn) with-out evidence of frank perforation on the imaging studies.9

Of those colonoscopies associated with an admissionwithin 30 days, the rate of colonoscopy completion, de-fined as reported intubation of the cecum, was calculatedfrom the colonoscopies in which the end point reachedwas documented, excluding individuals with prior largebowel resection.

Results were tabulated by using standard descriptiveanalysis. The c2 test was used to compare differences inproportions. The Mantel-Haenszel c2 test was used for lin-ear trends.

This study was performed for the WRHA’s MedicineStandards Committee and was approved by the Universityof Manitoba’s Health Research Ethics Board.

RESULTS

There were 24,509 outpatient lower GI endoscopiesperformed at the 6 Winnipeg hospitals between April 1,2004, and March 31, 2006. General surgeons performed13,705 (56%), gastroenterologists 9618 (39%), and familyphysicians 1180 (5%) of the procedures. The nature ofthe procedures is listed in Table 1. The mean (�SD) ageof individuals undergoing the procedures was 59 � 15years, and 56% were women.

The 303 admissions with potential complications andthe preceding 301 endoscopies (84% colonoscopies; 16%sigmoidoscopies) were reviewed. There was no mentionof the quality of the bowel preparation in 85% (n Z 256)of cases; 6% had adequate preparation (n Z 18) and9% (n Z 27) were reported to have poor preparation.

The reported completion rate was 65% (158/244) for allcolonoscopies and 55% (64/116) for colonoscopies with-out additional procedures. The completion rate was 72%(60/83) for the colonoscopies performed by gastroenterol-ogists compared with 59% (89/150) for those performedby general surgeons and 82% (9/11) for family physicians(P ! .005 for the comparison between general surgeonsand gastroenterologists). The depth of insertion was notdocumented in 4% of the colonoscopies. If it is assumedthat all cancers or strictures could not be passed by theendoscope and these procedures are removed from thecalculations, the overall colonoscopy completion ratewas 69% (123/179), for gastroenterologists 78% (50/64),

www.giejournal.org Volume 69,

general surgeons 62% (65/105), and family physicians80% (8/10) (P Z .03 for the comparison between generalsurgeons and gastroenterologists).

The median (range) duration of colonoscopies withoutany additional procedure at the 6 different hospitals was12.5 (5-24), 13.5 (2-47), 14.5 (4-37), 19 (5-36), 29 (5-117),and 40 (15-105) minutes, with the longest time at the 2teaching hospitals.

ComplicationsThere were 69 procedures associated with 71 complica-

tions in the 30 days after the index procedure (Table 2:complication rates for all lower GI endoscopies consid-ered together; Table 3: complications associated witheach category of procedures). One individual was initiallyadmitted for postpolypectomy syndrome the day after theprocedure, managed conservatively, and then readmitted2 weeks later for postpolypectomy bleeding. A second in-dividual was admitted within 6 hours of the initial proce-dure with acute myocardial infarction, was treated withthrombolytic therapy, and had postpolypectomy bleeding.There were 230 admissions for other indications, such aselective unrelated surgery. The mean (�SD) age of the in-dividuals who had procedure-related complications was65 � 15 years; 65% were men. The rate of complicationsfor individuals older than 50 years was 3.3 per 1000 proce-dures (60/17,918).

There were 14 colon perforations after colonoscopywithout additional procedures (1.0/1000), 6 after polypec-tomy (1.8/1000), 3 after stricture dilation (58.8/1000), 2after colonoscopy with biopsy (0.5/1000), 2 after flexiblesigmoidoscopy without additional procedures (0.8/1000),and 2 after flexible sigmoidoscopy with biopsy (3.1/1000). Most perforations occurred in the rectum or sig-moid colon (66%, 19/29). All complications after stricture

TABLE 1. Breakdown of the procedures performed as

outpatient procedures at Winnipeg hospitals in the 2

study years

Scope category No. (%)

Colonoscopy 13,775 (56)

Colonoscopy with biopsy 4148 (17)

Lower GI endoscopy with polypectomy* 3268 (13)

Flexible sigmoidoscopy 2608 (11)

Flexible sigmoidoscopy with biopsy 655 (3)

Lower GI endoscopy with balloon dilation* 51 (0.2)

Lower GI endoscopy with destruction

of tissue with laser*

4 (0)

*Canadian Classification of Interventions does not distinguish

between colonoscopies with polypectomy (or balloon dilatation/

laser treatment) from sigmoidsocopies with polypectomy (or

balloon dilatation/laser treatment).

No. 3 : Part 2 of 2 : 2009 GASTROINTESTINAL ENDOSCOPY 667


dilatation or sigmoidsocopy were perforations. Therewere 21 episodes of postpolypectomy bleeding (6.4/1000polypectomies).

There was one death related to the procedure in anindividual who had colon perforation after snare polypec-tomy of a small polyp in the cecum; the patient died asa result of postoperative complications of the subsequentsurgery. Most complications (70%, 50/71) occurred afterprocedures with biopsies, polypectomies, or balloon dila-tation; no additional diagnostic or therapeutic interven-tion was performed during 21 procedures. Sixteen (23%)of the 69 patients were receiving aspirin, nonsteroidalanti-inflammatory medications, clopidogrel, or warfarinbefore the index procedure.

There was no difference in the complication rateamong procedures performed at teaching hospitals com-pared with the community hospitals (19/6186 vs 50/18,323, P Z .6) or among the endoscopists of differentspecialties (general surgeons 41/13,705, gastroenterolo-gists 23/9618, family physicians 5/1180; P Z .5). Therewas a slightly increased perforation rate for colonoscopieswithout additional procedures performed by general sur-geons (11/8543) and family physicians (2/587) comparedwith the gastroenterologists (1/4635) (P Z .03).

There was a linear trend for a decreasing rate of compli-cations for endoscopists performing a higher number ofprocedures (P Z .02 for trends) (Table 4); for this analysisthe endoscopists were grouped into 3 groups with aboutequal number of endoscopists in each group. The compli-cation rate for those performing less than 200 per yearwas twice that for the rest (13/2400 or 5.4/1000 vs 55/

TABLE 2. Description of the 30-day complications after

lower GI endoscopies*

Complication No.

Rate (per 1000

procedures) (95% CI)

Perforations 29 1.18 (0.79-1.7)

Postpolypectomy bleeding 21 0.86 (0.53-1.31)

Postpolypectomy syndrome 9 0.37 (0.17-0.70)

Intestinal obstruction 3 0.12 (0.03-0.36)

Acute myocardial infarction 3 0.12 (0.03-0.36)

Acute diverticulitis 2 0.08 (0.01-0.30)

Bleeding after biopsy 1 0.04 (0-0.23)

Persistent vomiting 1 0.04 (0-0.23)

Pneumonia 1 0.04 (0-0.23)

Acute renal failure 1 0.04 (0-0.23)

Total 71 2.9 (2.26-3.65)

*The complication rates described in this table are for all lower GI

endoscopies considered together. For the complications associated

with each particular category of procedures, see Table 3 and text.


20,365 or 2.7/1000. P Z .02, relative risk 2 [95% CI, 1.1-3.7]).

Forty of the 69 patients with complications had polyps.The polyp size was not documented in 17 (43%), and thetype of polypectomy was not documented in 6 (15%). Thepolyps were removed by hot biopsy in 5 (13%), regular bi-opsy in 2, snare polypectomy in 26 (including 1 casewhere both snare polypectomy and hot biopsy wasused), and not removed in 2 cases.

Presentation and management ofcomplications

Two thirds (n Z 46) of the individuals with complica-tions were readmitted after discharge for the day surgeryprocedure. Eight (12%) individuals were recognized tohave had complications during the procedure and another15 (22%) before discharge from the day surgery proce-dure. However, 54% of individuals with complicationswere admitted within a day of the index procedure andthe rest within an additional 15 days.

Postpolypectomy bleeding was the only complicationthat led to admission beyond 5 days from the index proce-dure. The median (and range) of time interval betweenthe index procedure and readmission was 1 (range 0-5)day for colon perforations and 6 (range 0-16) days forpostpolypectomy bleeding.

Twenty-four of 29 with colon perforations required sur-gery. Three patients had a simple repair of the defect and8 resection and repair without a stoma; in 13 a stoma wascreated. Of the 21 cases with postpolypectomy bleeding, 7required blood transfusions, and 15 required repeat en-doscopy; hemostasis for control of continuing bleedingwas necessary in 7, and 2 required laparotomy.

The median length of hospital stay for complicationswas 7 days, with a range of 1 to 146 days.

DISCUSSION

These results suggest that the overall rate of differentcomplications associated with lower GI endoscopy per-formed as outpatient procedures in Winnipeg hospitalsis similar to that reported by others.9-12 The unadjustedcolonoscopy completion rate among the reviewed casesis lower than that recommended or reported in largeseries. The reporting and documentation of the endo-scopic examination findings such as the quality of bowelpreparation and characteristics of the colon polyps were of-ten missing. The importance of these data is that they arenot selective but representative of city-wide practice of anurban center that has a population of 700,000. Althoughprospective studies are typically held in higher regardthan retrospective ones, in terms of the ability to ensurethat all data are collected, in the case of recording endo-scopic outcomes a retrospective review provides a ‘‘realworld’’ experience avoiding the bias of practitioners



TABLE 4. Complication rate among endoscopists performing varying volume of procedures*

No. of procedures

performed per year

No. of

endoscopists

No. of procedures

associated with

complications

Total no. of

procedures

performed

Rate of complications

(per 1000 procedures)

(95% CI)

P value

for trends

!200 13 13 2400 5.4 (3.0-9.0) .02

200-300 14 22 6967 3.2 (2.0-4.7)

O300 15 33 13,398 2.5 (1.7-3.4)

*This analysis was limited to endoscopists who performed lower GI endoscopies in the entire study period.

TABLE 3. Complications associated with the different categories of lower GI endoscopy

Procedure

No. of procedures

associated with complications

Total no. of

procedures

Rate (per 1000

procedures) (95% CI)

Colonoscopy 19 13,775 1.4 (0.9-2.1)

Colonoscopy with biopsy 6 4148 1.4 (0.6-3.2)

Lower GI endoscopy with polypectomy 37 3268 11.3 (8.1-15.4)

Flexible sigmoidoscopy 2 2608 0.8 (0.1-2.5)

Flexible sigmoidoscopy with biopsy 2 655 3.1 (0.5-10.5)

Lower GI endoscopy with stricture dilation 3 51 58.8 (15.2-151.2)

Total 69 24,509 2.8 (2.2-3.5)

changing their practice because of awareness of data re-cording. Further, there is no financial advantage for report-ing complete colonoscopy in Manitoba. The results aredisappointing in terms of what practitioners are recording.

The recommended colonoscopy completion rate is90% to 95%.12 The completion rate in this study was calcu-lated from colonoscopies performed for individuals whowere admitted within 30 days of the index procedure.These findings need to be further explored. A recent anal-ysis of physicians’ billing claims data from Ontario sug-gested an 87% colonoscopy completion rate13; however,the reimbursement rates in Ontario are higher for com-plete colonoscopy, which may potentially provide someimpetus for completing or reporting complete proce-dures. Most of the other case series are based on prospec-tively collected data, where endoscopists may have beenaware of the data collection, or are from teaching hospi-tals.14 It is possible that colonoscopies that are associatedwith an admission are more difficult or complicated, andhence they may be more likely to be incomplete than nor-mal or uncomplicated colonoscopies. Hence the overallrate of incompletion of colonoscopies in the WRHA is un-known. Furthermore, this rate of colonoscopy completioncannot be reliably extrapolated to a population undergo-ing screening, who presumably would be on average un-complicated patients. However the completion rates of72% for even the gastroenterologists is concerning and

www.giejournal.org Volume 69

needs urgent attention. The colonoscopy completionrate was low even after individuals with colon cancer orstrictures were excluded. Even in individuals with knowncolorectal cancer, completion of colonoscopy to the ce-cum can be helpful to ensure that no synchronous lesionswere missed on the prior examination. Moreover, the co-lonoscopy completion rates in our study were calculatedfrom physician reports. We did not systematically assessphoto documentation. It is possible that colonoscopycompletion rates with proper photo documentation ofthe cecal landmarks would have been even lower. It is in-teresting to speculate whether colonoscopy completionrates would be similarly as low or at least less than the tar-get rate of 90% to 95% in other jurisdictions if audits (suchas our study), as opposed to prospective studies (or retro-spective studies involving few select endoscopists) areundertaken.

The poor reporting likely reflects the lack of require-ment for standardized reporting in Manitoba and mostof the rest of Canada. The quality of the bowel preparationand the characteristics of the colon polyps visualizedduring colonoscopy guide the recommendations forfollow-up colonoscopy and further management. A secondphysician (or even the same physician on a day distant tothe date of the performance of the index procedure) willhave difficulty providing optimal care without completeinformation of the initial procedure. In such situations,

, No. 3 : Part 2 of 2 : 2009 GASTROINTESTINAL ENDOSCOPY 669


a decision may be made to repeat the procedure at a shortinterval, adding to health care costs, unnecessary inconve-nience, and risk of adverse events to the patients. Perhapsthe most reliable way to ensure complete reporting wouldbe to institute electronic endoscopy reporting with man-datory data entry fields. One of the impacts of the currentstudy is that WRHA is considering transitioning to sucha system.

The total procedure times varied across the hospitals.Many colonoscopies were completed within 10 minutes,suggesting that withdrawal time may be less than that rec-ommended during many procedures.12 The longer proce-dure time in the teaching hospitals may reflect the impactof trainee participation, closer adherence to the necessityfor sufficient withdrawal times, or the fact that patientswith complex medical histories or longer surveillance co-lonoscopies may be more likely to occur in teaching hos-pitals. Finally, the durations may reflect differences indocumenting starting and ending points of procedures be-tween facilities.

Although the overall complication rate associated withlower GI endoscopies in Winnipeg is similar to that re-ported by others, the rate of complications (1.4/1000 pro-cedures) is high after colonoscopy without additionalprocedures.9,12 This needs to be investigated further.The complication rate was highest after procedures per-formed by low-volume endoscopists; if this finding is con-firmed in other population-based studies, there will beclear implications for policy makers and guidelines. Thesefindings do support the recommendations of Cancer CareOntario’s Colonoscopy Standards Expert Panel, on the ba-sis of which only the endoscopists performing more than200 colonoscopies per year can participate in Ontario’scolorectal cancer screening program.14

Deaths resulting from endoscopy-related complicationsare rare but can occur in spite of all the advances in med-ical care. Patients undergoing lower GI endoscopy need tobe advised about this small but definite risk. Some of thecomplications, such as postpolypectomy bleeding, can bedelayed and still be life threatening. Patients and referringphysicians need to be aware of the risk of delayedcomplications.

In spite of recent recommendations to limit their use,the use of hot biopsy forceps appears to be common inWinnipeg. This may be one explanation for the slightlyhigher rate of postpolypectomy complications reportedin our study compared with previous studies.5,9,15

General surgeons performed a large proportion of thelower GI endoscopies in our study. This is reflective of thepractice in Canada.16 All gastroenterology fellowshiptrainees in Canada over the course of the 2- to 3-year fel-lowship training program perform more colonoscopiesthan the minimum recommended by the American Soci-ety for Gastrointestinal Endoscopy (ASGE) (140 colonos-copies) or the Canadian Association of Gastroenterology(150 colonoscopies).17 Most general surgical trainees


acquire much of their endoscopy training in a 3-monthblock rotation, and on completion of their training,many may not meet the ASGE or Canadian guidelines.17,18

Additional experience may be gained during their residen-cies, but this is not uniform. Since 1996 the 2 main teach-ing hospitals in Winnipeg have required endoscopists tohave previously performed at least 150 supervised colo-noscopies and be certified by their training director asto their endoscopic competence to obtain privileges toperform colonoscopy in the 2 hospitals. Similar guidelineswere also developed for the rest of Manitoba but to thebest of our knowledge were never implemented.

There are several limitations to our study. We did notcontrol for comorbidities and prior surgery. We did nothave information on procedures performed at outpatientfacilities. About 15% of elective lower GI endoscopies areperformed at outpatient/ambulatory care facilities in Win-nipeg. We were unable to separate colonoscopy with poly-pectomy from sigmoidoscopy with polypectomy in thehospital abstract database, and hence we are reportingon complications after all polypectomies. Our databasesdo not allow us to conclusively determine the indicationsfor the performance of all procedures; many were likelyperformed on symptomatic patients, and our resultsmay not apply to pure screening populations. However,our data are applicable to an unselected group of patientsundergoing outpatient lower GI endoscopy. Most colonos-copies in the United States and elsewhere are still per-formed for diagnostic indications.19,20 Similar to many ofthe previous reports, we evaluated the more seriouscomplications, defined as those requiring admission toa hospital9,11; we did not assess visits to the emergencydepartment with minor complications that did not leadto hospitalization. For the incomplete colonoscopies, wedid not abstract information regarding the plans for repeatprocedures or radiologic investigations.

One of the strengths of the current analysis is evaluatingfor the risk after sigmoidoscopy and colonoscopy in thesame study. Previous studies, which restricted analysis to co-lonoscopies alone, could have missed complications associ-ated with failed colonoscopies, which could have beencoded as flexible sigmoidoscopies. It has been suggestedthat 1 of the reasons for the increased rates of complicationsreported to be associated with flexible sigmoidoscopy ina population-based study was due to inclusion of failed co-lonoscopies as flexible sigmoidoscopies.21 Similarly, exclu-sion of failed colonoscopy-associated complications fromthe calculation of colonoscopy-associated complicationsmay lead to determination of a lower risk of colonoscopy-associated complications. There were at least 2 such casesin our analyses, where the planned colonoscopy wasstopped as a result of recognition of perforation duringthe procedure; although these procedures were coded inthe hospital abstracts as flexible sigmoidoscopies, we con-sidered them as complications associated with colono-scopy. An important advantage of our study was that we



did review and confirm all cases of complications and en-doscopy reports by careful and systematic chart review.

In conclusion, the overall complication rates after lowerGI endoscopy in Winnipeg are comparable to those re-ported in other studies. The reporting of endoscopieshas several deficiencies and needs the generation and im-plementation of a more standardized reporting format,ideally a standardized electronic reporting system. The co-lonoscopy completion rate was poor among the casesaudited and needs to be further evaluated in an unse-lected group of patients. The higher complication rateassociated with procedures performed by low-volume en-doscopists needs to be further evaluated and may haveimplications on the policy for recredentialing for endos-copy privileges.

REFERENCES

1. Singh H, Demers AA, Xue L, et al. Time trends in colon cancer inci-

dence and distribution and lower gastrointestinal endoscopy utiliza-

tion in Manitoba. Am J Gastroenterol 2008;103:1249-56.

2. Use of colorectal cancer testsdUnited States, 2002, 2004, and 2006.

MMWR Morb Mortal Wkly Rep 2008;57:253-8.

3. Rex DK. Maximizing detection of adenomas and cancers during colo-

noscopy. Am J Gastroenterol 2006;101:2866-77.

4. Lieberman D, Nadel M, Smith RA, et al. Standardized colonoscopy report-

ing and data system: report of the Quality Assurance Task Group of the Na-

tional Colorectal Cancer Roundtable. Gastrointest Endosc 2007;65:757-66.

5. Misra T, Lalor E, Fedorak RN. Endoscopic perforation rates at a Cana-

dian university teaching hospital. Can J Gastroenterol 2004;18:221-6.

6. Rabeneck L, Paszat L, Hilsden RJ, et al. Factors associated with risk of

serious complications of colonoscopy: a population-based study

[abstract]. Gastroenterology 2008;134(1 Suppl):A88.

7. Farley DR, Bannon MP, Zietlow SP, et al. Management of colonoscopic

perforations. Mayo Clin Proc 1997;72:729-33.

8. Luning TH, Keemers-Gels ME, Barendregt WB, et al. Colonoscopic per-

forations: a review of 30,366 patients. Surg Endosc 2007;21:994-7.

9. Levin TR, Zhao W, Conell C, et al. Complications of colonoscopy in an

integrated health care delivery system. Ann Intern Med 2006;145:880-6.

www.giejournal.org Volume 69,

10. Gatto NM, Frucht H, Sundararajan V, et al. Risk of perforation after

colonoscopy and sigmoidoscopy: a population-based study. J Natl

Cancer Inst 2003;95:230-6.

11. Levin TR, Conell C, Shapiro JA, et al. Complications of screening flex-

ible sigmoidoscopy. Gastroenterology 2002;123:1786-92.

12. Rex DK, Petrini JL, Baron TH, et al. Quality indicators for colonoscopy.

Am J Gastroenterol 2006;101:873-85.

13. Shah HA, Paszat LF, Saskin R, et al. Factors associated with incomplete

colonoscopy: a population-based study. Gastroenterology 2007;132:

2297-303.

14. Rabeneck L, Rumble RB, Axler J, et al. Cancer care Ontario colonoscopy

standards: standards and evidentiary base. Can J Gastroenterol

2007;21(D Suppl):5-24D.

15. Gibbs DH, Opelka FG, Beck DE, et al. Postpolypectomy colonic hemor-

rhage. Dis Colon Rectum 1996;39:806-10.

16. Hilsden RJ, Tepper J, Moayyedi P, et al. Who provides gastrointestinal

endoscopy in Canada? Can J Gastroenterol 2007;21:843-6.

17. Romagnuolo J, Enns R, Ponich T, et al. Canadian credentialing guide-

lines for colonoscopy. Can J Gastroenterol 2008;22:17-22.

18. Asfaha S, Alqahtani S, Hilsden RJ, et al. Assessment of endoscopic

training of general surgery residents in a North American health

region. Gastrointest Endosc 2008;68:1056-62.

19. Lieberman DA, Holub J, Eisen G, et al. Utilization of colonoscopy in the

United States: results from a national consortium. Gastrointest Endosc

2005;62:875-83.

20. Harris JK, Vader JP, Wietlisbach V, et al. Variations in colonoscopy prac-

tice in Europe: a multicentre descriptive study (EPAGE). Scand J Gas-

troenterol 2007;42:126-34.

21. Schoen RE, Levin TR. Re: risk of perforation after colonoscopy and sig-

moidoscopy: a population-based study. J Natl Cancer Inst 2003;95:830-1.

Received July 17, 2008. Accepted September 18, 2008.

Current affiliations: Departments of Internal Medicine (H.S, C.N.B.)

Community Health Sciences (H.S., R.B.P., C.D., M.M.), University of

Manitoba, Winnipeg Regional Health Authority (R.B.P., L.K., C.P., G.T.,

M.M.), Winnipeg, Manitoba, Canada.

Reprint requests: Harminder Singh, MD, Section of Gastroenterology, 804-

715 McDermot Ave, Winnipeg, Manitoba, Canada R3E3P4.

If you want to chat with an author of this article, you may contact him at

[email protected].

No. 3 : Part 2 of 2 : 2009 GASTROINTESTINAL ENDOSCOPY 671

mailto:[email protected]

Documents

Colonoscopy and its complications across a Canadian regional health authority