Outpatient colonoscopy complications in the CDC's Colorectal Cancer Screening Demonstration Program
A prospective analysis
The articles in this supplement were commissioned based on participation in evaluating the Centers for Disease Control and Prevention-funded Colorectal Cancer Screening Demonstration Program.
The opinions or views expressed in this supplement are those of the authors and do not necessarily reflect the opinions or recommendations of the journal editors, the American Cancer Society, John Wiley & Sons Inc, or the Centers for Disease Control and Prevention.
To the authors's knowledge, there are few published prospective cohort studies of colonoscopy complications in patients at average risk for colorectal cancer who received colorectal cancer screening from a community-based program. In this article, the authors report the rate of colonoscopy complications in the Centers for Disease Control and Prevention (CDC)'s Colorectal Cancer Screening Demonstration Program (CRCSDP), which provided colorectal cancer screening to a medically underserved population aged 50 years to 64 years for screening, diagnostic follow-up after positive stool blood tests, and surveillance purposes.
Clinical data were collected prospectively from 5 community-based colorectal cancer screening programs. Complications were identified by reviewing the standardized clinical data and medical complication reporting forms submitted by the programs to the CDC. Serious complications were defined as conditions or symptoms that resulted in hospital admission within 30 days after the procedure, including perforation, gastrointestinal bleeding requiring or not requiring blood transfusion, cardiopulmonary events, postpolypectomy syndrome, excessive abdominal pain, or death.
A total of 3215 individuals underwent 3355 colonoscopies. Of these, 89% of the colonoscopies were conducted for screening, 9% were conducted for diagnostic follow-up, and 2% were conducted for surveillance purposes. The mean age of the individuals was 55.9 years. Eight individuals experienced serious complications, for an incidence of 2.38 per 1000 colonoscopies. Three patients experienced bowel perforations that required surgery, 1 patient was hospitalized for postpolypectomy bleeding, 3 patients experienced cardiopulmonary events, and 1 patient visited the emergency room for excessive abdominal pain and underwent surgery for an identified colorectal mass. No deaths were reported.
In the CDC's CRCSDP, in which a total of 3215 individuals underwent 3355 colonoscopies, the overall incidence of serious complications from colonoscopy was found to be low. Cancer 2013;119(15 suppl):2849-54. © 2013 American Cancer Society.
Colonoscopy is one of the primary screening modalities for colorectal cancer, and is also used as a follow-up procedure to evaluate abnormal findings detected by other screening tests.[1, 2] Although colonoscopy is generally considered to be a safe procedure, it can result in serious medical complications. An understanding of the complications associated with colonoscopy is important when planning for a colorectal cancer screening program. Although medical complications from colonoscopy have been reported in many published studies, these studies tend to be retrospective and are not exclusive to settings in which colonoscopy was used primarily for screening purposes.[3-9] To the best of our knowledge, few studies to date have been prospective studies of a screening cohort, making it difficult to assess the likelihood of complications occurring among individuals at average risk of colorectal cancer who undergo colonoscopy for colorectal cancer screening. In what to our knowledge is the largest prospective study in the United States, Ko et al reported 2.01 adverse events per 1000 colonoscopies. In this article, we report the complications that occurred in the first federally funded colorectal cancer screening program, which was established in 2005 by the Centers for Disease Control and Prevention (CDC). The 4-year Colorectal Cancer Screening Demonstration Program (CRCSDP) funded 5 sites across the country to implement screening programs in community-based settings.
MATERIALS AND METHODS
This report involves a prospective cohort of men and women who underwent outpatient colonoscopy for colorectal cancer screening or as a follow-up test after receiving a positive result from a fecal occult blood test. The cohort also includes patients who underwent colonoscopy for surveillance purposes because of a personal history of colorectal cancer or adenomatous polyps.
The goal of the CRCSDP was to screen individuals aged 50 years to 64 years who were at average risk of colorectal cancer. Those with a personal or family history of colorectal cancer or adenomatous polyps were eligible for the program. Approximately 83% of patients in the CRCSDP were at average risk of colorectal cancer, and 17% reported a personal or family history of colorectal cancer or adenomatous polyps. Individuals excluded from the program included those with symptoms of colorectal cancer, a history of inflammatory bowel disease (ulcerative colitis or Crohn disease), or a history of genetic syndromes that predispose a person to colorectal cancer such as familial adenomatous polyposis or hereditary nonpolyposis colorectal cancer. These individuals generally require genetic counseling or intensive clinical and surveillance services that were beyond the scope of this program. Income level and health insurance status also determined program eligibility. An individual's household income could not have exceeded 250% of the federal poverty level, and they must have had either no health insurance or inadequate coverage for colorectal cancer screening.
Screening Sites and Tests
The 5 CRCSDP-funded programs included the Maryland Department of Health and Mental Hygiene, which subcontracted for services provided in 5 Baltimore City hospitals; Stony Brook University Medical Cancer, which serves Suffolk County, New York; the Nebraska Department of Health and Human Services' statewide program, which provided services through the state health department and selected physician practices; the Public Health–Seattle and King County program, located in Washington State, which delivered services in primary care clinics and physician practices; and the St. Louis screening site of the Missouri Department of Health and Human Services, in which services were delivered at a university hospital or federally qualified health centers, and within a specialty care health provider network. The 5 screening programs were able to select the colorectal cancer screening modality that would be used by their program, provided the selected test or tests were recommended by the U.S. Preventive Services Task Force and the site had the capacity to offer the selected test or tests. Three of the 5 programs used colonoscopy as the primary screening test. Two sites used colonoscopy to evaluate positive fecal occult blood test results, to screen individuals at an increased risk due to family history, or for the surveillance of patients with a personal history of colorectal cancer or adenomatous polyps.
Data Source and Identification of Medical Complications
The CRCSDP sites used 2 methods to report medical complications to the CDC. First, the programs used a medical complications reporting form that was intended to be used in real time to capture any immediate and delayed complications within 30 days of the procedure. If a patient was hospitalized as a result of the complication, the screening site notified the CDC within 72 hours of the hospitalization and submitted the form within 5 days. For patients who experienced complications within 30 days of receiving a program test, program sites submitted the medical complications reporting form to the CDC within 30 days of the reported complication. The medical complications reporting form included 18 questions to elicit information regarding the procedure and the types of complications. The complications included on the form were bleeding, cardiopulmonary events, complications related to anesthesia, perforation, excessive abdominal pain, emergency room visit, death, and “other, specify.” Second, the screening sites submitted standardized clinical data to the CDC on a quarterly basis for the purposes of program evaluation and quality assurance. These Colorectal Cancer Clinical Data Elements (CCDEs) captured demographic and clinical information on each patient, including information concerning the screening and diagnostic tests performed by the program, and are described elsewhere in this supplement. The occurrence of a complication was reported in the CCDEs as a “yes” or “no” under the variable name “complications of endoscopy or double-contrast barium enema.”
Potential complications were identified by reviewing both the submitted medical complications reporting forms and the CCDEs. The CDC followed up with programs to resolve discrepancies between the data reported on the forms and in the CCDEs. Intensive tracking and record review were required to reconcile the data and capture the major complications. Medical and scientific staff at the CDC, including a collaborating gastroenterologist, reviewed the clinical information provided in the form to understand the details of the complication. In this report, major complications were defined as any conditions or symptoms that resulted in hospital admission within 30 days of the procedure and included perforation, gastrointestinal bleeding requiring or not requiring blood transfusion, cardiopulmonary events, postpolypectomy syndrome, excessive abdominal pain, and death. All complications occurring within 30 days of any program test were included, regardless of whether the medical records associated the complication with the procedure. Minor complications such as self-limited bleeding were not reported consistently from all 5 programs and therefore the numbers and rates of these complications are not reported herein.
Of the 5233 individuals who were screened in the CRCSDP, 3215 (61%) underwent colonoscopy for all indications. A total of 3355 colonoscopies were performed. Of these, 89% were conducted for screening, 9% for diagnostic follow-up after positive stool blood tests, and 2% were performed for surveillance purposes. The mean age of the patients at the time of colonoscopy was 55.9 years; 26% of the patients were men and 74% were women (Table 1).
Table 1. Characteristics of Patients and Complications Associated With Colonoscopies Performed in the CDC's Colorectal Cancer Screening Demonstration Program
|No. of patients||3215|
|No. of colonoscopies||3355|
|Mean age (SD)||55.9 (5.7)|
|Indication, %|| |
|Diagnostic follow-up to FOBT||9|
|Major complications, no. (rate/1000 colonoscopies)||8 (2.38)|
|Gastrointestinal bleeding||1 (0.29)|
|Cardiopulmonary events||3 (0.89)|
|Postpolypectomy syndrome||0 (0)|
|Excessive abdominal pain||1 (0.29)|
A total of 8 patients experienced major complications (incidence of 2.38 per 1000 colonoscopies) that required hospitalization based on the definition used in this report. Three patients experienced bowel perforations (for an incidence of 0.89 per 1000 colonoscopies) that required surgery to repair the bowel. Two of the 3 perforations occurred in patients who received colonoscopy for screening purposes and no interventions, such as biopsy or polypectomy, were performed during the procedure. The third perforation occurred in a patient who underwent a colonoscopy for screening; however, the patient received 2 additional colonoscopies for diagnostic and treatment purposes. The perforation occurred during the piecemeal removal of a sessile polyp during follow-up colonoscopy.
One patient was hospitalized briefly with postpolypectomy bleeding that did not require a blood transfusion (incidence of 0.29 per 1000 colonoscopies). Three patients experienced cardiopulmonary events (incidence of 0.89 per 1000 colonoscopies). Two patients were admitted to the hospital within 30 days of the colonoscopy with symptoms of chest pain and shortness of breath. A medical record review suggested that both patients had preexisting cardiac conditions and the chest pain and shortness of breath occurring after the colonoscopy were believed to be unrelated to the colonoscopy in either patient. One additional person developed a self-limited supraventricular tachycardia that was believed to be related to the bowel preparation. One patient visited the emergency room for excessive abdominal pain 2 weeks after undergoing a screening colonoscopy and surgery was performed to remove a large cancerous colorectal mass identified during the colonoscopy 2 weeks prior (incidence of 0.29 per 1000 colonoscopies). No deaths were reported. No complications occurred within 30 days of any other screening tests used in this program, including sigmoidoscopy and double-contrast barium enema.
Colonoscopy is the most invasive of all colorectal cancer screening tests, and its potential complications can be serious. The majority of complications are typically minor, such as bloating, nausea, abdominal pain, pain at the intravenous catheter site, or gastrointestinal bleeding that does not require medical attention. Serious complications, such as perforation and hemorrhage, can lead to significant morbidity and mortality, and they are more likely to result from colonoscopy compared with other screening tests.[6, 14] To our knowledge, only a few studies published to date have captured the immediate complications associated with colonoscopy and anesthesia.
The incidence rates of medical complications vary across studies, which may be attributed to differences in study populations and inclusion and exclusion criteria. Many studies have not been limited to a screening population.[4, 7-9, 15, 16] Among prospective cohort studies of asymptomatic adults who underwent colonoscopy for screening or surveillance because of a history of cancer or adenomatous polyps, complication rates ranged from 0.7917 to 8.4 per 1000 colonoscopies (Table 2).[10, 17-20] Johnson et al reported a complication rate of 0.79 per 1000 colonoscopies in a study in which the primary objective was to assess the accuracy of computed tomography colonography in the detection of large adenomas and cancers. In their 15-site study, colonoscopy was performed after the computed tomography colonographic examination. The study cites the incidence of adverse events, but does not detail the methods used to capture these events. Leffler et al reported a complication rate of 8.4 per 1000 colonoscopies in a study performed at a single medical center in Boston, Massachusetts. An electronic medical record-based system to automatically record admissions to the emergency department within 14 days after colonoscopy was developed. This approach was used to overcome the limitations of more commonly used practices such as physician reporting, medical record review, and telephone follow-up interviews. The study also compared the number of hospital visits captured by the automated tracking system with the medical center's standard voluntary physical reporting system. During the 9-month study period, 31 complications resulted in an emergency department visit or hospitalization according to the data captured in the standard physician reporting system, compared with 134 complications captured using the automated system.
Table 2. Prospective Cohort Studies on Serious Complications Associated With Screening and Surveillance Colonoscopy in Asymptomatic Individuals
|Current study||CDC Colorectal Cancer Screening Demonstration Program (practices from 5 states)||2005-2009||3355/55.9 y||All||2.38|
| || || || ||Perforation||0.89|
| || || || ||Bleedinga||0.29|
| || || || ||Cardiopulmonary||0.89|
| || || || ||Postpolypectomy syndrome||0|
| || || || ||Excessive abdominal pain||0.29|
|Leffler 2010||Beth Israel Deaconess Medical Center, Boston, MA||2007||7392/60.6 y||All||8.4|
|Ko 2010||18 Clinical Outcomes Research Initiative sites|| ||21,375||All||2.01|
| || || || ||Bleeding||1.59|
| || || || ||Perforation||0.19|
| || || || ||Diverticulitis||0.23|
| || || || ||Postpolypectomy||0.09|
| || || || ||syndrome|| |
|Johnson 2008||15 US medical centers||2005-2006||2531/58.3 y||All||0.79|
| || || || ||Bleeding||0.39|
| || || || ||Bacteremia||0.39|
|Regula 2006||40 centers in Poland||2000-2004||50,148/55.2 y||All||1.01|
| || || || ||Perforation||0.09|
| || || || ||Bleeding||0.25|
| || || || ||Cardiovascular||0.43|
| || || || ||Others||0.21|
| || || || ||Death||0|
|Lieberman 2000||13 US Veterans Affairs hospitals||1994-1997||3196/62.9 y||All||3.12|
| || || || ||Perforation||0|
| || || || ||Bleeding||1.87|
| || || || ||Myocardial infarction||0.31|
| || || || ||Cerebrovascular||0.31|
| || || || ||Fornier gangrene||0.31|
| || || || ||Thrombophlebitis||0.31|
The study by Ko et al, which to our knowledge is the largest prospective cohort study published in the United States, showed a complication rate of 2 per 1000 colonoscopies across 18 practices. Similar to the CRCSDP cohort, this study used a 30-day follow-up and the majority of patients were enrolled from community-based practices. Patients were contacted at 7 days and 30 days after colonoscopy. A questionnaire was administered to identify the occurrence of complications, and medical records review was used to validate the information collected in the questionnaires.
In the CRCSDP, the complication rate was 2.38 per 1000 colonoscopies, which is consistent with other prospective cohort studies. As demonstrated in the study by Leffler et al, tracking complications based on patient and physician reporting limits the potential to document a complication rate with 100% accuracy.
The overall low rate of complications noted herein is likely attributed to characteristics of the CRCSDP cohort, in which the mean age was 55.9 years among individuals who underwent colonoscopy. It has been found that patients aged ≥ 60 years have higher rates of complications. In addition, the CRCSDP focused on screening those individuals considered to be at average risk.
In the CRCSDP, 2 of the 3 colonic perforations requiring surgery occurred in patients who underwent screening colonoscopies that did not require biopsy or polypectomy. Although several studies have shown that complications are more likely to occur when an intervention is performed, a recent, large, US prospective study demonstrated that 2 of the 4 reported perforations occurred in patients who did not have any intervention during the procedure. Major bleeding is an adverse event that is more frequently associated with colonoscopy than perforation. In our cohort, 1 patient developed postpolypectomy bleeding that required inpatient admission, but did not require a blood transfusion.
The main limitation of the complications reported in this article is the small number of individuals screened. In addition, we learned that it is difficult to track any but the major complications in decentralized clinical settings because of inconsistent reporting of real-time clinical information from the subcontracted screening sites to the grantee site and back to the CDC. Tracking was particularly difficult if the presentation of complications was delayed. We learned that the tracking system designed for this program was not consistently able to prospectively capture all complications. Extensive follow-up was required with programs to reconcile the data in the CCDEs and medical complications reporting form, and to capture the major complications experienced in the 4-year program. The process to track complications in the CDC's newer Colorectal Cancer Control Program was simplified. Programs are now required to report medical complications that occur during or within 30 days of the procedure to their CDC Program Consultant by telephone. Programs are required to report complications that result in an emergency room visit, hospitalization, or death in the CCDEs and are no longer required to submit information via a medical complications reporting form.
The findings of the current study regarding complication rates were consistent with those of other prospective studies of screening in asymptomatic populations. The strength of the current report is that it prospectively examined major complications from colonoscopy in individuals who underwent screening in multiple community practices across the country. The low colonoscopy complication rate and reporting processes described herein will be useful for informing the planning efforts of the growing number of colorectal cancer screening programs in the United States, in which screening tests are coordinated across multiple health care provider systems.
The Colorectal Cancer Screening Demonstration Program evaluated in this supplement was funded by the Centers for Disease Control and Prevention Funding Opportunity Number RFA AA030.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.