A provocative new editorial in the Journal of the National Cancer Institute (Harris 2011) has challenged the colorectal cancer screening strategy currently dominant in the US. In the US, of course, almost all major medical organizations and societies recommend colonoscopy as the primary screening strategy for colorectal cancer. Harris and Kinsinger in the editorial point out that there are randomized controlled trials that demonstrate that screening for colorectal cancer with fecal occult blood tests (FOBT) and flexibile sigmoidoscopy reduces mortality from colorectal cancer but they point out that evidence of the magnitude of additional benefit from colonoscopy is much less robust.
The editorial should not really be considered “provocative”, given the recent US Preventive Services Task Force (USPSTF) recommendations regarding screening for prostate and breast cancer. In the case of the USPSTF recommendations and the JNCI editorial, the call is really to look at the “net benefit” of the screening programs. That is, it is not enough any longer to just look at how many cancers are detected by the screening methods and how many deaths due to those cancers are prevented, but to also look at the potential harms of such screening methods and ensure that those potential harms do not outweigh the potential benefits.
They point out in the editorial that harm may occur at both the individual level and the societal level. At the individual level they note potential harms such as anxiety related to what might be found, the discomfort associated with bowel preparation for the procedure, inconvenience of the preparation and the procedure, effects of procedural sedation, loss of productivity (at work or at home), the risk of overdiagnosis, having to remember that another colonoscopy is needed in the future, in addition to the traditional potential complications of the procedure itself (eg perforation, hemorrhage, etc.). They make a case for development of tables that could compare the potential risks and benefits for individual patients so that the potential net benefit might be better gleaned.
A few years ago we tried to find out what the rate was for complications of colonoscopy. That was not as straightforward as we expected. The rate depends on whether the colonoscopies are for screening vs. diagnostic, whether polypectomy was done, and also relates to the age of the population and associated comorbidities (Warren 2009). All these factors make it extremely difficult to compare rates of colonoscopy complications across facilities. Some of the same factors likely explain why rates at free standing ambulatory sites tend to be lower than at hospital-based endoscopy units.
Last month there were a couple timely reviews of colonoscopy complications. First is a guideline from the American Society of Gastrointestinal Endoscopy (ASGE) Standards of Practice Committee (Fisher 2011). The guideline provides statistics from multiple different sources and prior studies and notes that the overall pooled rate of serious complications is on the order of 2.8 per 1000 colonoscopies. They note that over 85% of the serious complications occur in patients undergoing colonoscopy with polypectomy performed. Rates of colonic perforation are generally below 1 per 1000 procedures and are probably closer to 1 per 10,000 for purely screening colonoscopies. Hemorrhage, immediate or delayed, occurs in 1 to 6 cases per 1000 colonoscopies and is more common in diagnostic colonoscopies and those with polypectomy. Polyp size and type are risk factors and patient comorbidities increase the risk of hemorrhage. It’s not clear whether the latter increases the risk or is just a marker for increased use of anticoagulants and antithrombotic agents. They note there is a guideline for management of antithrombotic therapy in patients undergoing endoscopy (Anderson 2009). Other more serious complications such as death, infection, and colonic gas explosions are relatively rare. They also mention the postpolypectomy electrocoagulation syndrome, where a transmural burn gives rise to signs of localized peritonitis but is usually treated medically rather than surgically. Among less serious complications, they note transient GI symptoms in as many as one third of all patients. Cardiopulmonary complications occur frequently, both during the procedure and up to 30 days following the procedure. However, in the Warren study (Warren 2009) the latter events were more likely related to the comorbidities in this patient population rather than to the procedure itself. And while many of the cardiopulmonary events during the procedure may be relatively minor (transient hypoxemia or transient hypotension), we’ve cautioned about the need for careful monitoring during procedural sedation, particularly in patients who may be at risk for undiagnosed conditions like obstructive sleep apnea. A review of procedural sedation in endoscopic procedures (McQuaid 2008) notes relatively few controlled studies involving colonoscopy but does note that the common practice of adding an opiate to a sedating agent for colonoscopy increases the likelihood of deep sedation and oxygen desaturation.
While infectious complications are uncommon with colonoscopies, don’t forget the risks of mass cross contamination from inadequately reprocessed endoscopes. See the recent multisociety guideline on avoiding the latter (Petersen 2011).
A second study on complications of colonoscopy in the Medicare population (Day 2011) confirmed increased rates of complications with advancing age and especially in octogenarians. While some of the higher rates in octogenarians are due to more comorbidities (and, hence, more medications) and the likelihood that more of the colonoscopies in that group were diagnostic rather than screening, they noted some physiologic changes with aging that may make complications more likely. Those include things like more tortuous colons, higher prevalence of diverticula, greater probability of polyps, and more advanced cancers.
There is an excellent existing guideline on management of antithrombotic agents for endoscopic procdures (Anderson 2009). That guideline is quite comprehensive and, while it notes that agents may not need to be stopped in many diagnostic studies, it does provide good suggestions based upon both procedure and patient risk factors. It also provides good guidance about how to start and stop these agents, bridging therapy, etc. Obviously, those guidelines are going to need revision, given the recent introduction of multiple new oral anticoagulants.
Not surprisingly, the studies in the endoscopy literature do not mention overdiagnosis at all in their discussions of complications and harm. The recent JNCI editorial points out that overdiagnosis is a significant issue in colonoscopic screening, noting that the majority of findings at colonoscopy are small low-risk adenomas and non-adenomatous polyps, not cancers. We’ve already noted above that removal of polyps is one of the factors that increases the risk of complications from colonoscopy.
And all we’ve said is based upon “net benefit” as it applies clinically. We have not even touched upon the cost implications. Most cost effectiveness analyses take into account primarily the costs of the procedures and the individual complications compared to the costs saved by early detection of cancers. Seldom do they take into account costs related to management of the non-cancerous findings. To truly quantify the incremental benefit of colonoscopy over other screening strategies you need to take all these factors into account.
Just as I, as a neurologist, need to know the procedural complication rate of the local surgeons before I recommend carotid endarterectomy to a patient, you should know the complication rates of your local endoscopists before you recommend colonoscopy to your patients. And those rates should be risk stratified (age, comorbidities, type of procedure, etc.). We’re willing to bet you probably have no idea what those rates are at your local facility. So, in most cases, you really don’t have a good handle on the potential “net benefit” when you recommend screening colonoscopy to your patients. Of course, we all know someone whose life was “saved” by a screening colonoscopy and, conversely, someone who died because they never had a screening colonoscopy. But the time has come to look at the bigger picture and really be able to advise our patients in a truly informed way.
We’ve entered a new era in the way we look at screening tests. No longer is “earlier detection is best” the sole driving force. We need to take a balanced look at both sides of the equation for each individual patient.
Harris R, Kinsinger LS. Less is More: Not “Going the Distance” and Why.
JNCI 2011; 103(23): 1-3
Warren J.L., Klabunde C.N., Mariotto A.B., et al: Adverse events after outpatient colonoscopy in the Medicare population. Ann Intern Med 2009; 150: 849-857
Fisher DA, Maple JT, Ben-Menachem T, et al for the ASGE Standards of Practice Committee, Complications of colonoscopy. Gastrointestinal Endoscopy 2011; 74(4): 745-752
Anderson MA, Ben-Menachem T, Gan SI, et al: Management of antithrombotic agents for endoscopic procedures. Gastrointestinal Endoscopy 2009; 70: 1060-1070
McQuaid K.R., Laine L. A systematic review and meta-analysis of randomized, controlled trials of moderate sedation for routine endoscopic procedures. Gastrointestinal Endoscopy 2008; 67: 910-923
Petersen BT, Chennat J, Cohen J, et al for the ASGE Quality Assurance In Endoscopy Committee. Multisociety guideline on reprocessing flexible gastrointestinal endoscopes: 2011. Gastrointestinal Endoscopy 2011; 73(6): 1075-1084
Day LW, Kwon A, Inadomi JM, et al. Adverse events in older patients undergoing colonoscopy: a systematic review and meta-analysis. Gastrointestinal Endoscopy 2011; 74(4): 885-896