RSS Feeds
-
The Authors Respond
Submit responseOur recent study investigating the association between colonoscopy and colorectal cancer (CRC) mortality has certainly generated substantial interest. The media attention in response to publication of our study has at least in part been due to a misinterpretation of our methodology and primary outcome. We did not evaluate the rate of missed cancers at colonoscopy and the degree to which missed cancers (or missed precursor lesions) may have contributed to our findings - that colonoscopy is significantly associated with a reduction in mortality from left-sided CRC but not right-sided CRC - is unknown. There has been considerable speculation with respect to the influence of procedural quality in Ontario on our findings, however our understanding of the quality of colonoscopy in the general population in any jurisdiction is limited and is almost certainly lower than that published in series from expert centers. Additionally, this line of argument reveals an unwavering belief in the potential of colonoscopy for the prevention of the vast majority of CRC cases and CRC deaths. However, our understanding of the molecular basis of CRC carcinogenesis is evolving and it seems likely that a considerable proportion of cancers do not originate from easily identified adenomas that have a slow rate of progression (1). Although it is possible that improvements in the quality of colonoscopy and / or the use of other screening methods may increase our ability to detect CRC or precursor lesions (such as the serrated adenoma), what is really needed is a rigorous evaluation of the relative effectiveness of CRC screening methods in a randomized trial. Specifically, we need to know the marginal benefit of colorectal cancer screening with colonoscopy compared with flexible sigmoidoscopy in terms of reducing deaths from this disease.
Case-control studies are a challenging methodology and prone to numerous biases. We agree with Romagnuolo et al that case and control selection is important and can be difficult particularly when evaluating maneuvers that may have a role in screening, investigation of symptoms and surveillance of patients after treatment. We did not include controls diagnosed with CRC prior to the date of diagnosis of their matched case, as the appropriate exposure interval in such individuals, given that the vast majority would be undergoing regular surveillance colonoscopy after diagnosis, is unclear. As exposure in our study was based on administrative data, our information on exposure to colonoscopy was complete for all cases and controls and accurate with respect to timing, necessary prerequisites to minimize bias when only those without disease are selected as controls (2). The length of follow up after diagnosis might be considered too short to ensure that enough time had elapsed after diagnosis to identify all cases (i.e. all CRC deaths in those diagnosed with CRC in our time period); however this would tend to bias the study towards an overestimate of the strength of the association between colonoscopy and mortality from CRC. Far more important in terms of a threat to the validity of our study results, we were unable to determine the indication for colonoscopy necessitating the use of an exclusion window, where colonoscopy performed close to the date of diagnosis (whether for screening or investigation of symptoms) was not considered an exposure. We agree with Weiss et al that because of this it would be ill advised to consider our odds ratio estimates of the association between colonoscopy and CRC death as precise, or generalizable to other populations. Although many assume that the limitations of our study have resulted in an underestimate of the strength of the association between colonoscopy and CRC mortality, Weiss correctly highlights that, because of the necessity of the 6 month exclusion window, our study may have overestimated the strength of the association. However, despite the limitations there is no reason to conclude that potential biases due to the study design would have influenced the major findings of this study; that colonoscopy is strongly associated with a reduction in CRC mortality, but that the association is not uniform throughout the colon. Although our study does not provide an explanation for the lack of association between colonoscopy and prevention of right-sided CRC deaths, our findings do appear to be consistent with an emerging literature (3-5).
References
1.Jass JR. Classification of colorectal cancer based on correlation of clinical, morphological and molecular features. Histopathology 2007, 50, 113–130.
2.Weiss NS, Lazovich D. Case-control studies of screening efficacy: the use of persons newly diagnosed with cancer who later sustain an unfavorable outcome. Am J Epidemiol. 1996 Feb 15;143(4):319-22.
3.Gupta AK, Melton LJ, Petersen GM, Timmons LJ, Vege SS, Harmsen WS, Diehl, NN, Zinsmeister AR, Ahlquist DA. Changing trends in the incidence, stage, survival, and screen-detection of colorectal cancer: a population- based study. Clin Gastroenterol Hepatol. 2005 Feb; 3(2):150-8.
4.Brenner H, Chang-Claude J, Seiler CM, Sturmer T, Hoffmeister M. Does a negative screening colonoscopy ever need to be repeated? Gut. 2006;55:1145- 50.
5.Singh G, Mannalithara A, Wang HJ, Graham DJ, Gerson LB, Triadafilopoulos G. Is protection against colorectal cancer good enough: a comparison between sigmoidoscopy and colonoscopy in the general population. Gastroenterology. 2007;132(Suppl 2):A81.
Conflict of Interest:
None declared
-
Definition of controls underestimates effect of colonoscopy on colorectal cancer mortality
Submit responseDear editor:
Baxter et al (1) concluded colonoscopy’s effect on colon cancer (particularly right-sided) mortality was weaker than previously thought. However, we feel the study design introduced significant bias by the definition of controls. Had the outcome been development of colon cancer (rather than death from colon cancer), elimination of patients with non-fatal colon cancer from the control group would have been appropriate. However, since the study outcome was “death”, only patients who died of colon cancer should have been excluded from the control group, irrespective of whether prior cancers were excluded from being cases (2-4). Patients diagnosed with colonoscopy early enough to have prevented death should have been allowed to be “non-cases”, since they did not die from colon cancer (2-4). Exclusion of even a small percent of these patients, almost all of whom would have undergone colonoscopy, likely dropped the control colonoscopy exposure by a few percent. Each percent change corresponds to a 0.1 difference in the odds ratio, since the baseline exposure rate was only 9.8% (1). Furthermore, should the rate of detecting non-fatal colon cancer with colonoscopy be different for right-sided cancers, this could differentially bias the right-sided odds ratio, explaining in part the surprising lack of apparent benefit in that group.
Although the above is by far the most important (and clinically significant) potential bias, lack of adjustment for flexible sigmoidoscopy (known confounder, associated with reduced mortality and with exposure to colonoscopy), family history (associated with cancer and colonoscopy) and prep quality (particularly for the right colon) may have introduced bias.
Other issues include the age range chosen. Indeed, deaths at age 50 are not expected to be impacted by screening colonoscopy, which is recommended for most after age 50, and may not affect mortality for years afterwards. Similarly, the upper age of 90 included deaths 15-20 years after screening generally stopped. This dilutes estimates of screening effectiveness. Lastly, in-patient colonoscopy is generally performed for symptoms, rather than for screening; lack of adjustment for inpatient status reduces the apparent benefit of screening outpatient colonoscopy. Inpatient colonoscopies may have poorer visualization (especially right-sided), such as in the context of bleeding, which may again bias the results towards the null (lower apparent benefit).
These methodologic concerns have likely contributed to an underestimation (by a clinically significant magnitude) of colonoscopy’s benefit in general, and in the right colon in particular. Further data are required to support the claimed overall and site-specific conclusions.
REFERENCES
1. Baxter NN, Goldwasser MA, Paszat LF, Saskin R, Urbach DR, Rabeneck L. Association of colonoscopy and death from colorectal cancer. Ann Intern Med. 2009;150(1):1-8.
2. Weiss NS. Application of the case-control method in the evaluation of screening. Epidemiol Rev. 1994;16(1):102-8.
3. Knox G. Case-control studies of screening procedures. Public Health. 1991;105(1):55-61.
4. Cronin KA, Weed DL, Connor RJ, Prorok PC. Case-control studies of cancer screening: theory and practice. J Natl Cancer Inst. 1998;90(7):498-504.
Conflict of Interest:
Drs. Romagnuolo and Barkun have consultancy relationships with Olympus Corp.
-
Rethinking Colon Cancer Screening Strategies
Submit responseTo the Editor:
Baxter et al have provided further data which conflicts with the conclusions drawn from the National Polyp Study (1) and draws into question commonplace use of colonoscopy as a colon cancer screening test in the US for low risk, asymptomatic patients. Colon cancer incidence in patients under close colonoscopic surveillance has been shown to approach SEER estimates for the general population (2) and now Baxter’s data raises the possibility right colon cancer deaths may be unchanged by a test which carries with it significant expense and risks.
Because cecal intubation rates in Baxter’s study were lower than expected, defending colonoscopy based on “quality” issues has become the popular counter-argument. However, enthusiasm for “withdrawal time” and “polyp identification rates” is largely based on a study reporting more polyps were found by a colonoscopist with longer withdrawal times compared to his colleagues (3), but the study failed to include that endoscopist’s unexpected three year interval cancer incidence was the highest in his group. More diligence searching for polyps to meet “benchmark” standards may result in more polyps being removed, but not necessarily fewer colon cancer deaths. As Baxter points out, tumor biology variability may prevent some colon cancers from being detected by colonoscopy. To date, no new colonoscope technologic advances have proven superior to what is already in wide-spread use for identifying flat polyps or lesions hidden behind folds.
The “Katie Couric era” of colon cancer screening by the most costly and invasive test available has yet to be objectively justified. As concluded by the US Preventive Task Force report (4), no study has demonstrated a superiority of colonoscopy over annual fecal occult blood testing to decrease colon cancer deaths or incidence. If the low end estimate that colonoscopy decreases colon cancer deaths by 50% proves accurate, the 30% reduction already published for fecal occult blood testing (5) becomes increasingly impressive. Similarly, if only left colon cancer deaths are significantly impacted by endoscopy, the flexible sigmoidoscope’s value might be more appreciated.
Baxter et al have raised another warning that colonoscopy may not deliver the expected protection from colon cancer. For the average risk population with no symptoms, colonoscopy may not prove to be cost effective in these economically challenging times. Perhaps an initiative to make Fecal Immunochemical Testing kits available in the neighborhood grocery store (akin to home pregnancy tests) would be a better use of the shrinking US health care dollar.
References
1) Winawer SJ et al. Prevention of colorectal cancer by colonoscopic polypectomy. N Engl J Med 1993; 329;1977-81
2) Robertson DJ et al. Colorectal Cancer in Patients Under Close Colonoscopic Surveillance. Gastro 2005; 129;34-41
3) Barclay RL et al. Colonoscopic Withdrawal Times and Adenoma Detection during Screening Colonoscopy. N Engl J Med 2006;355:2533-41
4) Screening for Colorectal Cancer: US Preventive Task Force Recommendation Statement. Annals Int Med 2008 149;9; 627-637
5) Jorgensen OD et al. A randomized study of screening for colorectal cancer using faecal occult blood testing: results after 13 years and seven biennial screening rounds. Gut 2002;50:29-32
Conflict of Interest:
None declared
-
Association of Colonoscopy and Death From Colorectal Cancer
Submit responseThe impressive study by Baxter et al (1) approached the question of colonoscopic effectiveness from a unique direction, but their findings of insignificant screening benefit in the proximal colon did not match our clinical experience or that of other studies since the 1993 National Polyp Study (2). We sought to explore the discrepancy between Baxter's conclusions and our clinical experience by assessment of our practice.
After IRB approval, we reviewed 5 years of stage II-IV colorectal cancer (CRC) cases at our institution. For comparison we also reviewed a month of colonoscopies for completeness to cecum/TI determined by photo documentation, description, and diagram.
Of 111 CRC cases found, 93 (84%) had no prior history of colonoscopy within 10 years of diagnosis. Of the other 18 cases, 12 were performed at outside institutions, mostly by non-gastroenterologists. In the 6 patients with CRC who had colonoscopies at our institution, 1 had inflammatory bowel disease and 5 did not return for colonoscopic follow up as directed (3 poor preps, 2 adenomatous polyps). Of 191 colonoscopies performed at our institution, 179 were complete; 169 with retrievable photo documentation of the cecum/TI; 12 were aborted (8 poor prep; 3 inadequate sedation; 1 bleeding).
In contrast to Baxter's conclusions, our data suggests that limited right colonoscopic effectiveness is best explained by inadequate colonoscopy. Specifically, Baxter's conclusions cannot be extrapolated to our center, as standards of care can vary significantly between institutions. If 30% of Baxter’s colonoscopies are billed as incomplete, and "completeness" for CPT-4 billing is “proximal to the splenic flexure”, perhaps only half actually reached the cecum (3).
We recommend that: (1) informed consent for colonoscopy include the potential of an incomplete procedure; (2) the CPT 4 code for complete colonoscopy be amended to “colonoscopy proximal to the ileocecal valve”; (3) photo documentation of cecal/TI intubation be routinely performed; (4) emphasis be placed on completing follow up; (5) individual centers be responsible for their colonoscopic quality even to the point of publishing their cecal intubation rates; (6) if cecal intubation rates are inadequate other methods of right colon visualization should be considered (e.g. CT Colonography).
References:
1. Baxter NN, Goldwasser MA, Paszat LF, Saskin R, Urbach DR, Rabeneck L. Association of Colonoscopy and Death From Colorectal Cancer. Ann Intern Med 2009; 150: 1-8.
2. Winawer SJ, Zauber AG, Ho MN, O'Brien MJ, Gottlieb LS, Sternberg SS, Waye JD, Schapiro M, Bond JH, Panish JF, 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.
3. Rodney WM, Dabov G, Cronin C. Evolving colonoscopy skills in a rural family practice: the first 293 cases. Fam Pract Res J. 1993 Mar;13(1):43-52.
Conflict of Interest:
None declared
-
Let's not miss the Race
Submit responseTo the Editor: We read with great interest the article by Baxter et al (1) regarding the association between colonoscopy and death from colorectal cancer. The results of this study are likely to have a far-reaching impact on the day- to-day practice of colonoscopy. Expertise of the colonoscopist in conjunction with the difficulties encountered in getting to the right side of colon are the factors likely to be responsible for the observed discrepancy between the left and right colon cancer related mortality.
One thing strikes our minds while going through the article. In this high quality study, although cases were matched to controls for factors like sex, socioeconomic status, age and even geographical location; race and ethnicity as an important matching variable was missing in the data (2). In our view this is an important factor in the pathogenesis, progression and survival in colorectal carcinoma as has been shown in various previous studies (3-5).
References
1) Baxter NN, Goldwasser MA, Paszat LF, Saskin R, Urbach DR, Rabeneck L. Association of colonoscopy and death from colorectal cancer. Ann Intern Med. 2009; 150(1):1-8. [PMID: 19075198]
2) Comparing colorectal cancer by race and ethnicity. Available at http://www.cdc.gov/cancer/Colorectal/statistics/race.htm. Accessed January 2009.
3) Koo JH, Kin S, Wong C, Jalaludin B, Kneebone A, Connor SJ, et al. Clinical and pathologic outcomes of colorectal cancer in a multi-ethnic population. Clin Gastroenterol Hepatol. 2008 (9):1016-21.[PMID: 18558515]
4) Chien C, Morimoto LM, Tom J, Li CI. Differences in colorectal carcinoma stage and survival by race and ethnicity. Cancer. 2005; 104(3):629-39. [PMID: 15983985]
5) Brim H, Mokarram P, Naghibalhossaini F, Saberi-Firoozi M, Al-Mandhari M, Al-Mawaly K, et al. Impact of BRAF, MLH1 on the incidence of microsatellite instability high colorectal cancer in populations based study. Mol Cancer. 2008; 7:68. [PMID: 18718023]
Conflict of Interest:
None declared
-
Colorectal cancer screening in the US - implications from the study by Baxter et al
Submit responseThe findings from the Canadian study by Baxter and colleagues cannot be readily generalized to the United States. In their study, nearly 70% of the endoscopists who participated in the study were GP's, internists, etc. and not fellowship trained gastroenterolgists. Unfortunately the media, in disseminating the conclusion of this study, failed to pick up on this important detail. The editors of the Annals of Internal Medicine have done a disservice to the American public and colorectal cancer screening.
Conflict of Interest:
None declared
-
Association of Colonoscopy and Death From Colorectal Cancer
Submit responseTO THE EDITOR: We read with great interest the article by Nancy and colleagues (1) which showed that colonoscopy was associated with fewer deaths from CRC developing in the left side of the colon. The authors found that the patients which got a colonoscopy and died of CRC varied by site of primary CRC, 4.1%, 10.6%, and 9.1% for left-sided cancer, right-sided cancer and unknown site of cancer, respectively. Generally, the conclusion of this study was solid. However, there are still some points which need to be clarified.
First, the reason why colonoscopy was less effective in preventing death from right-sided CRC should be discussed. In addition to what the author has already explained, we should also keep in mind that the left colon is narrower than the right colon. Cancers of the left colon are more likely to cause partial or complete bowel obstruction and hence present symptoms that mandate seeking medical attentation. Therefore, possibly, it is detected much earlier for the patients to do a colonoscopy if the tumors localized at the left colon than the right side.
Secondly, the incidence of right-sided colorectal carcinoma was much higher than the left in older individuals (2). The age was an important factor to evaluate the CRC incidence location. Finally, healthy eating habits, especially low fat intake and high fiber diet is helpful to prevent CRC. Different individuals have different habits. So, we believe that the quality of the paper will be greatly improved after taking these potential factors into consideration.
References
1. Baxter NN, Goldwasser MA, Paszat LF, Saskin R, , et al. Association of colonoscopy and death from colorectal cancer. Ann Intern Med. 2009;150(1):1-8. [PMID: 19075198]
2. Rabeneck L, Davila JA, El-Serag HB. Is there a true "shift" to the right colon in the incidence of colorectal cancer? Am J Gastroenterol. 2003;98:1400-9. [PMID: 12818288]
Conflict of Interest:
None declared
-
The Evidence and Efficacy of Colonoscopy for Colorectal Cancer Screening
Submit responseBeing a Chief Resident at our University based Internal Medicine residency program, I got the privilege to choose this article for our January Journal Club meet. I would like to comment Baxter and colleagues for their extensive work on this extremely important screening test (1). This landmark paper demonstrated a clear benefit from colonoscopy for left -sided colorectal cancer (CRC) (OR 0.39 (0.34-0.45)) but failed to show any survival benefit for right-sided CRC (OR 1.07 (0.94-1.21))(1). In today's world of Evidence based medicine, most of my residents were surprised at the evidence available behind the recommendation for screening colonoscopy as there are no randomized controlled trials.
One problem that we had, and would like the authors of the study to clarify is the methodology, although very extensive, was very confusing. To my surprise, many of my residents could not confidently conclude whether all the control arm patients had CRC or not. This confusion would be even more prevalent among our primary care readers and is already evident in some rapid responses posted in response to the study. A more clear explanation as to whether all the patients in the control arm had CRC would be extremely helpful in understanding the methodology and the results.
We would like to congratulate all the authors for their extensive work and explaining in great details each limitation for their study. We would eagerly wait to see the results of the ongoing randomized control trial (2) as highlighted in the editorial by Dr Ransohoff (3).
References
1. Baxter NN, Goldwasser MA, Paszat LF, et al. Association of colonoscopy and death from colorectal cancer: a population-based, case-control study. Ann Intern Med 2009;150:1-8.
2. Bretthauer M, Ekbom A, Malila N, Stefansson T, Fischer A, Hoff G; et al. NordICC-gruppen (Nordic Initiative on Colorectal Cancer). [Politics and science in colorectal cancer screening]. Tidsskr Nor Laegeforen. 2006;126:1766-7.
3. Ransohoff D. F. How Much Does Colonoscopy Reduce Colon Cancer Mortality? Ann Intern Med, 2009; 150(1): 50 - 52.
Conflict of Interest:
None declared
-
A smoking gun to explain observed differences between colonoscopy and CT colonography?
Submit responseTo the Editor:
We read with interest the recent article by Baxter and colleagues (1). This landmark paper demonstrated a clear benefit from colonoscopy for left -sided colorectal cancer (CRC) but failed to show any survival benefit for right-sided CRC. Plausible explanations for this important observation include: 1) a unique tumor biology for a subset of right-sided CRC and 2) technical factors resulting in missed right-sided CRC by colonoscopy. In fact, an alternate cancer pathway distinct from the typical adenoma-carcinoma sequence – the serrated polyp pathway – has been identified, which is a predominately right-sided process and may account for 15% of sporadic CRC. However, the precursor serrated lesion is typically large and appears to have a very long dwell time, suggesting that neither cancer prevention nor detection should be diminished. There is no convincing evidence for a rapidly progressive or de novo sporadic CRC that favors the proximal colon. This leaves missed right-sided cancers at colonoscopy as the most likely explanation.
It stands to reason that proximal colonic tumors at a greater distance from the anus would pose a greater diagnostic challenge for colonoscopy. Several inter-related factors likely limit colonoscopic detection of right -sided pathology: 1) the proximal aspect of right-sided colonic folds represents a recognized blind spot for physical endoscopy (2, 3); 2) suboptimal bowel preparation tends to obscure the right colon to a greater degree; 3) incomplete examination (whether recognized or not); and 4) the presence of flat lesions, which are particularly vulnerable to limitations 1 and 2 above. In contrast, CT colonography (CTC) has distinct advantages for right-sided evaluation, including the absence of physical directional constraints allowing for complete evaluation, relative ease of distention of the proximal colon, the ability to tag residual stool with oral contrast, and combined 3D-2D assessment for flat lesions. This may help explain the striking differences observed in CRC detection between primary CTC and colonoscopy.
In our comparison trial (4), eight right-sided cancers were found in the CTC screening cohort, versus just one in the closely-matched colonoscopy arm. In a Mayo clinic validation trial (5), 4 of 5 cancers (80%) were missed at initial colonoscopy, despite serving as the reference standard for CTC, which prospectively detected all five cancers. With a priori knowledge of CTC-detected lesions, the miss rate at colonoscopy is considerably lower. Perhaps a primary screening strategy that alternates between flexible sigmoidoscopy and CTC would optimize cancer detection and minimize invasiveness.
References
1. Baxter NN, Goldwasser MA, Paszat LF, Saskin R, Urbach DR, Rabeneck L. Association of Colonoscopy and Death From Colorectal Cancer: A Population-Based, Case-Control Study. Ann Intern Med. 2008.
2. Pickhardt PJ, Nugent PA, Mysliwiec PA, Choi JR, Schindler WR. Location of adenomas missed by optical colonoscopy. Annals of Internal Medicine. 2004;141(5):352-359.
3. Pickhardt PJ, Choi JR, Hwang I, et al. Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults. New England Journal of Medicine. 2003;349(23):2191-2200.
4. Kim DH, Pickhardt PJ, Taylor AJ, et al. CT colonography versus colonoscopy for the detection of advanced neoplasia. New England Journal of Medicine. 2007;357(14):1403-1412.
5. Johnson CD, Fletcher JG, MacCarty RL, et al. Effect of slice thickness and primary 2D versus 3D virtual dissection on colorectal lesion detection at CT colonography in 452 asymptomatic adults. AJR Am J Roentgenol. 2007;189(3):672-80.
Conflict of Interest:
None declared
-
Inadequate Sedation May Have Influenced Findings
Submit responseTo the Editor: In the recent study by Baxter, et al (1), there is no discussion of what sedation, if any, that the patients received. Although the study took place in Canada, in the United States it has become increasingly common for a separate anesthesia care provider to be present during colonoscopies both to ensure the safety of the patient and to provide a deeper level of sedation. This deeper level of sedation presumably provides the endoscopist a better opportunity for conducting a thorough exam, especially the right side of the colon. The fact that about 70% of the procedures were performed by surgeons and internists leads one to suspect that a separate anesthesia provider was not present for many of these procedures and that optimal sedation was not achieved. The poor tolerance of colonoscopies in inadequately sedated patients may be reflected in the relatively low cecal intubation rates and the inability to examine the right side of the colon sufficiently well to impact mortality rates.
References:
1. Baxter NN, Goldwasser MA, Paszat LF, Saskin R, Urbach DR, Rabeneck L. Association of colonoscopy and death from colorectal cancer: a population-based, case-control study. Ann Intern Med 2009; 150:1-8.
Conflict of Interest:
None declared
-
Efficacy of colonoscopy screening
Submit responseIn order to provide information regarding the efficacy of screening colonoscopy in preventing death from colorectal cancer, Baxter et al. (1) analyzed administrative claims data from the province of Ontario. While this choice had the advantage of yielding a very large sample (10,292 deaths from colorectal cancer), it had some drawbacks that we believe seriously threaten the validity of the results obtained:
1. Colonoscopy, like most screening tests for cancer, can be used as well to evaluate persons in whom, based on the presence of symptoms or signs, there is reason to believe that the cancer is present. Because the cases in this case-control study all had colorectal cancer, many of them would have received colonoscopy in response to symptoms or signs. The data available to Baxter et al. could not distinguish between screening and diagnostic colonoscopies. To the extent that tests not screening in nature were labeled as “screening”, the proportion of “screened” cases would be spuriously high, leading to a spuriously high odds ratio and thus a spuriously low estimate of screening efficacy.
2. Cognizant of the above limitation, Baxter et al. omitted from their analysis those colonoscopies most likely not to represent screening, i.e. tests done within six months of diagnosis of the person who went on to die of colorectal cancer (and the corresponding date in their matched controls). However, this analytic choice has the potential to produce a strong bias in the opposite direction. The problem stems from the fact that screening tests among cancer cases and controls during the period prior to the date of the cases’ diagnoses are not distributed in time in the same way. Especially for a sensitive test such as colonoscopy, almost all screening tests performed in cases (during the period in which the cancer is detectable by means of this test) will have been done close to the time of diagnosis: had a screening colonoscopy been done earlier during the preclinical phase of the cancer, the tumor would have been found then. In controls, in whom nearly all do not have colorectal cancer, the distribution of screening colonoscopy during the corresponding period of time would have been far more uniform.
To gauge the impact of a strategy of deleting from consideration all tests done within six months of diagnosis, assume there is a perfectly sensitive screening test for a cancer that, in the absence of screening, would be present for several years prior to its diagnosis on the basis of symptoms or signs. Assume as well that the receipt of screening by study subjects can be accurately ascertained, and also that there is no effective treatment for the condition being screened for. A valid case- control study of the efficacy of the screening test, a test which failed to yield any mortality reduction, ought to observe an odds ratio of one, i.e. a similar proportion of cases and controls with a history of screening. However, a far larger proportion of screening tests done in cases than in controls will be excluded in an analysis that ignores tests performed in the six months prior to the date of diagnosis, leading to a spuriously low odds ratio and a spuriously high estimate of screening efficacy (2).
For these reasons, as well as other (probably smaller) sources of potential bias – failure to include all deaths from colorectal cancer that ultimately will occur among cases diagnosed in the study population during 1992-2001 (3), confounding by family history and other risk factors – the odds ratios obtained in the study of Baxter et al. may well not be indicative of the actual impact of screening colonoscopy in Ontario on mortality from colorectal cancer. We agree with Baxter et al., and with the opinion expressed in the accompanying editorial (4), that it is plausible that colonoscopy has greater efficacy in the prevention of death from distal than proximal colorectal cancer. However, we believe that it would be fortuitous if the efficacy of colonoscopy estimated by the study of Baxter et al. against mortality from either distal or proximal colorectal cancer corresponded to the true value.
References
1. Baxter NN, Goldwasser MA, Paszat LF, et al. Association of colonoscopy and death from colorectal cancer: a population-based, case-control study. Ann Intern Med 2009;150:1-8.
2. Weiss NS. Analysis of case-control studies of the efficacy of screening for cancer: how should we deal with tests done in persons with symptoms? Am J Epidemiol 1998;147:1099-102.
3. Weiss NS, Lazovich DA. Case-control studies of screening efficacy: the use of persons newly diagnosed with cancer who later sustain an unfavorable outcome. Am J Epidemiol 1996;143:319-22.
4. Ransohoff DF. How much does colonoscopy reduce colon cancer mortality. Ann Intern Med 2009;150:50-2.
Conflict of Interest:
None declared
-
A Need to Get Screening Right for Right-sided Colorectal Cancer
Submit responseTo the Editor:
In the nice study by Baxter, et al (1), the authors observed that colonoscopy was associated with a significant decline in left-sided but not right-sided colorectal cancer (CRC) mortality. If data are available, it would be instructive to report differences in polypectomy rates between cases and controls. It is ultimately removal of premaligant polyps, rather than colonoscopy per se, that accounts for CRC prevention.
We had earlier reported our findings from a population-based study of all CRC cases in Olmsted County assessing concurrent trends in CRC incidence, screen-detection, and polypectomy rates (2). Over the study period 1980-1999, there was a four-fold increase in polypectomies and a 23% reduction in CRC incidence. The decline in CRC incidence was accounted for by a 40% reduction in left-sided CRCs (p< 0.001), as the incidence of right-sided CRC remained essentially unchanged.
We speculated in our study that the limited impact of polypectomies on right-sided CRC incidence may have been due to a distal bias in screening, to changing host-environmental factors, or to a different precursor lesion and natural history.
While both of these retrospective population-based studies have inherent limitations, results are corroborating and raise questions about the effectiveness of currently-practiced colonoscopy for right-sided CRC prevention. In face of the well-documented left-to-right shift in CRC incidence across many countries, a concerted effort is needed to better understand the biology of right-sided CRC and, accordingly, to develop rational screening interventions.
References:
1. Baxter NN, Goldwasser MA, Paszat LF, Saskin R, Urbach DR, Rabeneck L. Association of colonscopy and death from colorectal cancer: a population-based, case-control study. Ann Intern Med 2009; 150:1-8.
2. Gupta AK, Melton LJ 3rd, Petersen GM, Timmons LJ, Vege SS, Harmsen WS, Diehl NN, Zinsmeister AR, Ahlquist DA. Changing trends in the incidence, stage, survival, and screen-detection of colorectal cancer: a population-based study. Clin Gastroenterol Hepatol. 2005 Feb;3(2):150-8.
Conflict of Interest:
None declared
-
Chromoendoscopy enhances rate detections in CRC
Submit responseAfter reading the article by Baxter N et al in your journal, I had the chance to read the press release of the American Society for Gastrointestinal Endoscopy (ASGE), reminding us the importance of a proper bowel preparation before colonoscopy in order to reach adecuately the cecum, and also the importance of the withdrawal time, which should be of 6 minutes or more(1).
I would like to add to this statement, the importance of chromoendoscopy nowadays in the detection of small popyps and flat lesions. We do not necessarily require Narrow Band Imaging (NBI), and/or magnification endoscopes. Using indigocarmine solution through the biopsy channel of a standard colonoscope we can improve the detection rate of both, small polyps and flat lesions, throughout the colon, and particularly at the cecum and ascending colon (2-3).
References
1) ASGE Urges Patients to Seek A Qualified ndoscopist Before Undergoing a Colonoscopy for Colorectal Cancer. At www.asge.org/Pressroomindex.aspx? id=6024
2) Kiesslich R, von Bergh M, Hahn M et al. Chromoendoscopy with indigocarmine improves the detection of adenomatous and non adenomatous lesions in the colon. Endoscopy 2001; 33(12):1001-6.
3) Park SY, Lee SK, Kim BC et al. Efficacy of chromoendoscopy with indigocarmine for the detection of ascending colon and cecum lesions. Scand J Gastroenterol. 2008;43(7):878-85.
Conflict of Interest:
None declared
-
Improving Colon Cancer Screening with sDNA
Submit responseTo the Editor:
The recent study by Baxter and colleagues highlights that while colonoscopic screening decreases death rates from colon cancer, it has limitations, particularly with respect to preventing deaths from cancers of the right colon (1). The authors suggest that these limitations might reflect the technical difficulty of routinely reaching and identifying right-sided colonic lesions , which may be flat and harder to visualize, and/or might have a different biology than more distal cancers. Indeed, it is well established, for example in the context of Lynch Syndrome, that colon cancers with microsatellite instability, which predominantly arise in the right colon, develop at an accelerated rate, and can appear sometimes within 12-24 months of previously normal colonoscopies. The more common non-familial MSI colon cancers are similarly also likely to easily give rise to “interval” colon cancers arising in-between screening colonoscopies.
What then can be done to enhance the efficacy of colon cancer screening programs? We suggest that the addition of stool DNA (sDNA) based tests at intervals between colonoscopies as one attractive strategy. sDNA based tests directly detect DNA molecules that are shed from colonic neoplasms into the stool, and that bear signature molecular alterations of colonic neoplasia. sDNA, with high sensitivity for cancer detection, has been endorsed as an acceptable option for colorectal cancer screening by the American Cancer Society / Multi-GI Society Task Force (American Gastroenterological Association, American College of Gastroenterology, American Society of GI Endoscopy) and American College of Radiology and included in their joint national colorectal screening guideline (2). The test is noninvasive, inexpensive, and appears to be equally sensitive for detecting lesions in the right and left colon (3-5). Thus, sDNA testing is an attractive technology for identifying individuals in whom intensive examination of the right colon could be of value during an initial colonoscopy, or routinely for identifying individuals in whom a right sided lesion develops following a normal colonoscopy. Broadening routine screening to use, sequentially, a visual method (colonoscopy) complemented by an interval molecular method (sDNA) could more effectively identify patients harboring active colorectal neoplasia than either approach alone.
References
1. Baxter NN, Goldwasser MA, Paszat LF, Saskin R, Urbach DR, Rabeneck L. Association of Colonoscopy and Death From Colorectal Cancer: A Population-Based, Case-Control Study. Ann Intern Med. 2008.
2. Levin B, Lieberman DA, McFarland B, et al. Screening and surveillance for the early detection of colorectal cancer and adenomatous Polyps, 2008: a joint guideline from the American Cancer Society, the US Multi-Society Task Force on Colorectal Cancer, and the American College of Radiology. CA Cancer J Clin. 2008;58(3):130-60.
3. Itzkowitz S, Brand R, Jandorf L, et al. A simplified, noninvasive stool DNA test for colorectal cancer detection. Am J Gastroenterol. 2008;103(11):2862- 70.
4. Ahlquist DA, Sargent DJ, Loprinzi CL, et al. Stool DNA and occult blood testing for screen detection of colorectal neoplasia. Ann Intern Med. 2008;149(7):441-50, W81.
5. Chen WD, Han ZJ, Skoletsky J, et al. Detection in fecal DNA of colon cancer-specific methylation of the nonexpressed vimentin gene. J Natl Cancer Inst. 2005;97:1124-1132.
Conflict of Interest:
Sanford Markowitz holds patents on technologies Exact Sciences has licensed, that are employed in currently available sDNA tests.
Steven Itzkowitz, has conducted clinical trials of sDNA tests and has received research support and consulting fees from Exact Sciences.
Barry M. Berger is a developer of DNA based technologies for early cancer detection.
-
Do not be disheartened by study's demonstration of high colon cancer miss rate
Submit responseThere are several issues in the recent article entitled “Association of Colonoscopy and Death From Colorectal Cancer: A Population-Based, Case-Control Study” that raise concern about the adequacy of the exams performed and hence the clinical significance of the findings . First, the cecal intubation rates ranged from 79-83% amongst all colonoscopists. For gastroenterologists specifically, the cecal intubation rate was only 83% which is well below current accepted rates. In subgroup analysis, the cecal intubation rate was even lower (73%) in the case patients. The authors appropriately performed additional subgroup analysis to compare complete versus incomplete colonoscopy in terms of cancer detection rates. However, it can certainly be hypothesized, that even if a complete colonoscopy was performed, a colonoscopist with a low cecal intubation rate may lack the experience and technical skill to adequately identify polyps, particularly flat lesions, during the exam. Supporting this is the fact that in the study, the polyp detection rate was 26% in cases and 21% of controls. A significant proportion of these polyps may have been hyperplastic and hence the overall adenoma detection rate (ADR) may have been significantly lower than the current standards in the gastroenterology literature which suggest that adenomas be detected in 15% of women and 25% in men (1). In addition, one might expect that in this older, predominantly male case population (median age of 73, 54% male) in which cancer is already present, the adenoma detection rate in other parts of the colon would be even higher. In a recent study in the American Journal of Gastroenterology (2), fellow participation in colonoscopy yielded an adenoma detection rate of 37%, raising concern that our accepted ADR may be too low.
Based on above, it is not surprising that the rate of missed cancers, particularly in the right colon, was high especially given the fact that many of these lesions may have arisen from flat, subtle lesions that would be much more difficult to detect than a well defined polyp. This study underscores the need to have all gastroenterologists strive to meet accepted rates of cecal intubation and polyp detection. The aggressive and evasive characteristics of right sided lesions that this paper seems to support also highlights the idea that a careful optical colonoscopy to identify subtle mucosal changes in the right colon may be superior to "virtual" colonoscopy which can miss flat lesions.
References
1) Rex DK, Bond JH, Winawer S, et al. Quality in the technical performance of colonoscopy and the continuous quality improvement process for colonoscopy: Recommendations of the U.S. Multi-Society Task Force on Colorectal Cancer. Am J Gastroenterol 2002;97:1296-308
2) Rogart, JN, Siddiqui UD, Jamidar PA, Aslanian HR. Fellow Involvement May Increase Adenoma Detection Rates During Colonoscopy. Am J Gastroenterol 2008;103:2841-2846
Conflict of Interest:
None declared