If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Division of Clinical Epidemiology and Aging Research, German Cancer Research Center (DKFZ), Heidelberg, GermanyMedical Faculty Heidelberg, University of Heidelberg, Heidelberg, Germany
Reprint requests: Hermann Brenner, MD, MPH, Division of Clinical Epidemiology and Aging Research, German Cancer Research Center, Im Neuenheimer Feld 581, D-69120 Heidelberg, Germany.
Affiliations
Division of Clinical Epidemiology and Aging Research, German Cancer Research Center (DKFZ), Heidelberg, GermanyDivision of Preventive Oncology, German Cancer Research Center (DKFZ) and National Center for Tumor Diseases (NCT), Heidelberg, GermanyGerman Cancer Consortium (DKTK), German Cancer Research Center (DKFZ), Heidelberg, Germany
Recent guidelines on colorectal cancer (CRC) screening recommend starting screening earlier than before. We performed a simulation study to examine and compare the optimal ages to have once-only screening colonoscopy and repeated colonoscopies.
Methods
A Markov model was set up using data from the German national screening colonoscopy registry to simulate the natural history of the adenoma-carcinoma process. CRC deaths and years of potential life lost (YPLL) for a hypothetical unscreened 50-year-old German population were estimated for a single screening colonoscopy or 2 or 3 screening colonoscopies with 10-year intervals at various ages.
Results
One single screening colonoscopy performed between 50 and 65 years of age was expected to reduce CRC death by 49% to 69% and YPLL by 51% to 68%. An inverted U-shaped association was found between screening age and proportion of CRC deaths or YPLL prevented. The optimal age for once-only colonoscopy that yielded the highest reductions in YPLL was around 54 years for men and 56 years for women. Estimates were approximately 6 to 8 years higher when proportions of CRC deaths prevented were examined. For 2 or 3 screening colonoscopies, the optimal starting age fell to around 50 years or even younger for both genders.
Conclusions
Based on the YPLL estimates, in a high CRC incidence and high life expectancy country like Germany, the optimal age for once-only screening colonoscopy is around 55 years and possibly slightly younger for men than for women. When 2 or more screening colonoscopies are offered with 10-year intervals, screening should start at age 50 at the latest or possibly even younger for both genders.
With approximately 1.4 million new cases and nearly 700,000 deaths every year, colorectal cancer (CRC) is the third most commonly diagnosed malignancy and the fourth leading cause of cancer death worldwide.
Because of the slow progression from adenomatous polyps to invasive cancers, a substantial proportion of CRC cases and deaths could be prevented by screening. Colonoscopy is one of the effective screening modalities and has been shown to provide long-term protection against CRC occurrence and death.
Effect of screening sigmoidoscopy and screening colonoscopy on colorectal cancer incidence and mortality: systematic review and meta-analysis of randomised controlled trials and observational studies.
When offered as a primary screening test, colonoscopy screening typically starts at age 50 or 55 and is repeated every 10 years thereafter up to age 75 or 79, with 2 or 3 examinations in total for 1 individual.
The repeated testing of the entire eligible population, together with the requirement for frequent surveillance after positive findings imposed by current guidelines,
generates great demands for colonoscopy resources and can be a challenge in settings where colonoscopy capacity is limited. For these settings, 1 single colonoscopy at an age that yields the best protection would be a practical and feasible alternative.
Age of screening initiation is crucial to the effectiveness and cost-effectiveness of screening programs. Previous simulation analyses for the U.S. Preventive Services Task Force and the American Cancer Society CRC screening recommendations have suggested that starting screening at age 45 instead of 50 yielded a more-favorable balance between life-years gained and screening burden.
The impact of the rising colorectal cancer incidence in young adults on the optimal age to start screening: microsimulation analysis I to inform the American Cancer Society colorectal cancer screening guideline.
Based on the modeling results and the increasing CRC incidence in the younger population, the recently released American Cancer Society guideline recommended average-risk populations to start screening at age 45.
Other countries have also recently adjusted their screening initiation age. In Germany, the stakeholders of the screening colonoscopy program decided to lower the starting age from 55 to 50 years for men only.
Gemeinsamer Bundesausschuss. Federal Joint Committee]. Beschluss des Gemeinsamen Bundesausschusses über eine Richtlinie für organisierte Krebsfrüherkennungsprogramme und eine Änderung der Krebsfrüherkennungs-Richtlinie. Decision of the Federal Joint Committee on a directive of organized cancer screening programs and an amendment to the cancer screening directive [In German]. Berlin.
The impact of the rising colorectal cancer incidence in young adults on the optimal age to start screening: microsimulation analysis I to inform the American Cancer Society colorectal cancer screening guideline.
evidence is limited for a single colonoscopy when capacity is inadequate and once-only test is allowed. Therefore, we performed a simulation study to examine the best ages to have once-only screening colonoscopy and compared these estimates with the optimal ages for 2 and 3 screening colonoscopies.
Methods
A multistate Markov model was set up to simulate the natural history of CRC based on the adenoma-carcinoma process. Simulation was performed on a hypothetical previously unscreened 50-year-old German population.
Model structure
The model structure is outlined in Figure 1 and has been described elsewhere.
Expected long-term impact of the German screening colonoscopy programme on colorectal cancer prevention: analyses based on 4,407,971 screening colonoscopies.
The simulated population enters the model with different proportions of no neoplasm, nonadvanced adenoma, advanced adenoma, and preclinical CRC and each year they have a certain probability to progress to the next state. Diagnosed patients may die from the disease, and at each state (including after diagnosis), people may die from other causes. Screening can alter the progression between states. People with asymptomatic CRC detected at screening colonoscopy will be moved forward to the state of diagnosed CRC because of the detection of their cancer, and people with adenoma will be moved backward to the state of no neoplasm, assuming removal of their adenoma at screening colonoscopy. Although these people might be under higher risks of developing adenomas or cancers than the general population in the following years,
Figure 1Schematic illustration of the Markov model used for the analysis. Solid lines represent the progression of colorectal disease through the adenoma-carcinoma sequence in the absence of screening; dashed lines show the movement between states because of the detection and removal of adenomas and the detection of asymptomatic CRC at screening. CRC, Colorectal cancer.
Parameters used in this analysis were based on data from several sources. Proportions of adenoma and cancer before entering the model (before screening) were previously extracted from the data of the German national screening colonoscopy registry.
Briefly, findings at screening colonoscopies, including number, size, and histologic characteristics of polyps, as well as sociodemographic information were reported anonymously to the national screening colonoscopy registry. According to the most advanced finding at screening colonoscopy, participants were then categorized into no neoplasm, nonadvanced adenoma, advanced adenoma, or cancer. Advanced adenoma is defined as at least 1 adenoma ≥1 cm or at least 1 adenoma with villous components or high-grade dysplasia. Although the offer of the German screening colonoscopy program starts at age 55 on the national level, selected regional programs offer screening colonoscopy from age 50 on, and these programs found similar prevalence of colorectal neoplasms in age groups 50 to 54 and 55 to 59.
Therefore, for this analysis, we used sex-specific proportions of no neoplasm, nonadvanced adenoma, advanced adenoma, or CRC at screening colonoscopy among 344,658 participants who had their first screening colonoscopy during 2003 to 2012 at the age of 55 years as the best available estimates for the simulated 50-year-old population (Supplementary Table 1, available online at www.giejournal.org). Transition rates between states were extracted from previous publications,
in which data from the German screening colonoscopy registry were used to derive these estimates. In this analysis, age- and sex-specific annual transition rates among the 4 states (no neoplasm, nonadvanced adenoma, advanced adenoma, and preclinical CRC) for age range 55 to 79 years and annual transition rates between preclinical and clinical CRC for age range 55 to 84 were used. Estimates for age 50 to 54 and ≥80 (or ≥85) were assumed to be the same as those for age group 55 to 59 and 75 to 79 (or 80 to 84), respectively (Supplementary Table 2, available online at www.giejournal.org).
Overall CRC-specific mortality rates by year after diagnosis were drawn from the CRC-specific survival proportions by year after diagnosis in Germany in 2011 to 2012, which were reported by the German Centre for Cancer Registry Data and cover all CRC deaths in Germany (Supplementary Table 3, available online at www.giejournal.org).
From the overall mortality rates, we then estimated the mortality rates specifically for cases detected by screening colonoscopy or symptoms. The proportion of screening colonoscopy–detected cases among all CRC cases and the hazard ratio for CRC mortality among screening colonoscopy–detected versus symptom-detected cases were used to estimate the mortality rates by mode of detection. The proportion of screening colonoscopy–detected cases was calculated as the number of CRC cases detected at screening colonoscopy divided by the total number of CRC cases in Germany during 2003 to 2012 among people aged 55 to 79 years.
Hazard ratios for CRC-specific mortality were obtained from the DACHS (Darmkrebs: Chancen der Verhütung durch Screening) study, a population-based case-control study on CRC screening with additional follow-up of cases conducted in southwest Germany.
The study was approved by the ethics committees of the University of Heidelberg and the state medical boards of Baden-Wuerttemberg and Rhineland-Palatinate. Written informed consent was obtained from each study participant. For this analysis, we used data of 2258 patients aged 55 to 79 years with a first diagnosis of CRC (C18-20 by International Classification of Diseases, 10th revision), recruited during 2003 to 2010 and followed with respect to mortality until 2013 to estimate the hazard ratio. The hazard ratios (95% confidence interval [CI]) for patients detected by screening colonoscopy versus symptoms were .24 (95% CI, .13-.43) and .18 (95% CI, .07-.48) for men and women, respectively.
The estimated CRC-specific mortality rates by mode of detection used in the analyses are displayed in Supplementary Table 3. Sex- and age-specific general mortality rates and average life expectancy of the population were extracted from German population life tables 2010/2012 (Supplementary Table 4, available online at www.giejournal.org).
We modeled disease progression for populations with or without screening from the states at age 50 to a maximum of 100 years. Progression between states was simulated in discrete time steps of 1 year based on the annual transition rates and general mortality rates. In the scenario of screening, people with asymptomatic CRC detected at screening would be moved to the state of diagnosed CRC, and people with adenoma would be moved to the state of no neoplasm. Further surveillance colonoscopy after positive findings at screening colonoscopy was not specifically examined in our analyses but was assumed to offset the higher risk of development of colorectal neoplasms in this group. Once cancer was detected, we used the annual CRC-specific mortality rates (screening colonoscopy–detected or symptom-detected CRC mortality for populations with or without screening, respectively) to obtain the CRC mortality at each age (ie, in each follow-up year) from 50 to a maximum of 100 years. CRC mortality at each age was then applied to the hypothetical cohort of 100,000 men and 100,000 women aged 50 years to estimate the number of CRC deaths at each age, which was used together with the average life expectancy at each age to eventually compute the years of potential life lost (YPLL) until age 100.
We performed simulations for different screening initiation ages to identify the scenario that yielded the highest proportion of YPLL prevented compared with no screening. Similar analyses were performed with proportion of CRC deaths prevented as comparison. Simulations were run for a single screening colonoscopy performed between ages 50 and 65, 2 screening colonoscopies with a 10-year interval with the first examination starting between ages 50 and 65, and 3 screening colonoscopies with 10-year intervals with the first examination starting between ages 50 and 59. All simulations were separately run for men and women.
Sensitivity analyses
To explore the impact of uncertainty regarding annual transition rates between states, sensitivity analyses were performed in which all point estimates of the transition rates were replaced by either the upper or lower limits of the 95% CIs. The impact of lower rates of compliance with screening colonoscopy (25%, 50%, and 75%) rather than full compliance was examined in additional sensitivity analyses. To further evaluate the effect of starting screening below age 50, all simulations were repeated with a new starting age of 45 years. Proportions of adenoma and cancer at age 45 were obtained by fitting separate regression models to the age-specific proportions of nonadvanced adenoma, advanced adenoma, and CRC among 3,593,420 participants of the German national screening colonoscopy program, aged 55 to 80 years. Log-linear models with linear and quadratic effects of age were used to fit the proportions of nonadvanced and advanced adenoma, and a log-log model was used to fit the relationship between age and the proportion of CRC. Predicted proportions at age 45 were then calculated from the regression models (Supplementary Table 5A, available online at www.giejournal.org). Transition rates between states were assumed to be the same as those for age group 50 to 54, and the overall CRC-specific mortality rates by mode of detection were used for the age group 45 to 49. General mortality rates and average life expectancy for this specific age group were also extracted from the German population life tables 2010/2012 (Supplementary Table 5B, available online at www.giejournal.org).
Results
In the absence of screening, CRC was estimated to cause a total of 5186 and 3983 deaths among 100,000 men and 100,000 women, respectively (Table 1). One single screening colonoscopy performed between 50 and 65 years of age was expected to reduce this estimate by approximately 55% to 69% for men and 49% to 69% for women (Fig. 2). An inverted U-shaped association was found between age at screening colonoscopy and the estimated proportion of CRC deaths prevented among men, and the optimal age that yielded the highest proportion was 60 years. A single screening colonoscopy among women was expected to achieve the same maximum mortality reduction but with a higher optimal screening age of 64 years.
Table 1Estimated numbers of colorectal cancer death for no screening and screening colonoscopy at various ages
Figure 2Proportions of CRC death prevented when a single screening colonoscopy was performed at various ages. A, Men. B, Women. Arrow marks the highest value. CRC, Colorectal cancer.
The estimated CRC deaths in the absence of screening were expected to yield a total of 58,429 and 46,682 YPLL for men and women, respectively (Table 2). The proportion of YPLL prevented from having a single screening colonoscopy between 50 and 65 years ranged from 51% to 68% among men and from 57% to 67% among women (Fig. 3). Similar inverted U-shaped patterns between screening age and the proportion of YPLL prevented were identified; however, the optimal age associated with the highest proportion was 54 years for men and 56 years for women, that is, 6 to 8 years younger compared with those based on examination of proportions of CRC deaths prevented.
Table 2Estimated YPLL because of CRC death for no screening and screening colonoscopy at various ages
Figure 3Proportions of YPLL prevented when a single screening colonoscopy was performed at various ages. A, Men. B, Women. Arrow marks the highest value. YPLL, Years of potential life lost.
Two screening colonoscopies with a 10-year interval were expected to reduce the YPLL from CRC death by 53% to 87% for men and 62% to 85% for women (Fig. 4). The proportion of YPLL prevented was estimated to decline substantially with delayed screening initiation, and the optimal age that yielded the highest proportion was 50 years for men and 51 years for women. Larger proportions of YPLL prevented would be expected when 3 screening colonoscopies were performed, but consistently decreasing trends in the proportion with delayed screening initiation were identified and the optimal age (within the examined range) was 50 years for both genders. Although the maximum YPLL reductions were similar between genders in the scenarios of repeated colonoscopy screening, men had slightly lower YPLL reductions than women when they started screening at a later age.
Figure 4Proportions of YPLL prevented when 2 or 3 screening colonoscopies were performed at various ages. A, Two screening colonoscopies. B, Three screening colonoscopies. Arrow marks the highest value. YPLL, Years of potential life lost.
Sensitivity analyses using upper and lower limits of 95% CI of annual transition rates yielded similar associations between screening age and proportion of CRC deaths or YPLL prevented as those in the base case analyses. In each scenario, the estimated CRC deaths or YPLL were correspondingly higher or lower, but the optimal ages associated with the highest proportions fell into ±1 year of the ages identified in the main analysis (Supplementary Fig. 1, available online at www.giejournal.org). The results were also robust to sensitivity analyses considering a screening colonoscopy compliance rate of 75%, 50%, and 25%, although the absolute proportion of CRC deaths or YPLL prevented and the differences in screening effect across various ages became smaller with lower compliance rates (Supplementary Fig. 2, available online at www.giejournal.org). Consistent optimal ages for once-only screening colonoscopy were identified when simulation started at age 45 (Supplementary Fig. 3, available online at www.giejournal.org). In the scenario of repeated colonoscopies, the proportion of YPLL prevented still declined with delayed screening initiation as in the main analysis after age 50. The extension of examined age range provided additional information in the younger population, and an even earlier screening starting age was identified, except for women having 2 colonoscopies, for which the optimal age remained as 51 years as in the base case. The optimal starting age for men having 2 or 3 colonoscopies was 49 or 45 years, respectively, and the optimal initiation age for women having 3 tests was 47 years.
Discussion
In this study, we provide model estimates of the optimal ages for screening colonoscopy. Our results indicate that for a general unscreened 50-year-old population, 1 single screening colonoscopy performed between 50 and 65 years of age is expected to reduce CRC deaths by 49% to 69% and YPLL because of CRC death by 51% to 68%. An inverted U-shaped association was identified between screening age and the proportion of CRC deaths or YPLL prevented. The optimal age that yielded the highest reduction in YPLL was around 54 years for men and 56 years for women. Estimates were approximately 6 to 8 years higher when the proportions of CRC deaths prevented were considered. The optimal age of screening initiation was predicted to fall to around 50 or lower for both genders when 2 or 3 screening colonoscopies with 10-year intervals were performed.
The earlier optimal screening ages identified in the analyses of YPLL reflects the nature of this metric in measuring premature deaths. Preventing a younger person from CRC occurrence and death could save more future productive years than saving an older person and is thus more heavily weighted in the analyses of YPLL. On the contrary, when CRC death numbers were examined, every death was weighted equally irrespective of age. The optimal ages for once-only screening colonoscopy obtained from our analyses are generally consistent with the results of another modeling study, in which the CRC mortality reduction and life-years gained from having a single colonoscopy were found to reach the maximum when colonoscopy was performed at around 60 to 61 and 53 to 54 years, respectively.
An increasing number of countries have recently implemented CRC screening programs. Although there are some variations in screening initiation age across countries, which typically ranges from 50 to 60,
the same age is recommended for men and women within each country. Our estimates suggest that for once-only colonoscopy screening, gender differentiation of screening age might yield better protection against CRC death. The earlier screening age for men obtained from our analysis is supported by the gender differences in the risk of CRC and important precursor lesions.
Age-specific CRC incidence and mortality are much lower in women than in men. Although the lifetime risk of the disease is comparable for the 2 genders because of women’s longer life expectancy,
the onset of CRC is later for women. In a study based on data from 11 countries, women were found to reach equivalent level of cumulative 10-year CRC incidence and mortality to that of men approximately 4 to 8 years later.
For repeated colonoscopy screening, our estimates suggest initiating screening at around age 50 years or even younger based on the increasing YPLL with age and the estimates from the sensitivity analyses. These results are in line with the findings from previous simulation analyses for the United States that starting screening 5 years earlier than the commonly recommended age of 50 yielded increases in life-years gained.
The impact of the rising colorectal cancer incidence in young adults on the optimal age to start screening: microsimulation analysis I to inform the American Cancer Society colorectal cancer screening guideline.
Another simulation study for the United States on individualized colonoscopy screening also found that an earlier initiation age of 47 years with more frequent repeating for blacks generated improvements in benefits.
In addition, starting screening earlier might be further supported by the rising CRC incidence rates in the younger, prescreening age groups over the past decades in several high-income countries.
Declining bowel cancer incidence and mortality in Germany: an analysis of time trends in the first ten years after the introduction of screening colonoscopy.
The impact of the rising colorectal cancer incidence in young adults on the optimal age to start screening: microsimulation analysis I to inform the American Cancer Society colorectal cancer screening guideline.
The 2 other authoritative guidelines from the U.S. Preventive Services Task Force and the U.S. Multi-Society Task Force on Colorectal Cancer, on the contrary, still suggested beginning screening at age 50 for most average-risk persons.
More analyses on the benefits, costs, and harms as well as the underlying reason for the increasing CRC incidence among young population are warranted to further inform the decision-making of screening below 50.
In Germany, colonoscopy has been offered as a primary CRC screening test since 2002. People aged over 55 years are offered up to 2 screening colonoscopies with a 10-year interval. In view of the high disease burden among men, the German screening colonoscopy program recently announced lowering the screening starting age from 55 to 50 years for men, whereas the offer for women would remain unchanged.
Gemeinsamer Bundesausschuss. Federal Joint Committee]. Beschluss des Gemeinsamen Bundesausschusses über eine Richtlinie für organisierte Krebsfrüherkennungsprogramme und eine Änderung der Krebsfrüherkennungs-Richtlinie. Decision of the Federal Joint Committee on a directive of organized cancer screening programs and an amendment to the cancer screening directive [In German]. Berlin.
the idea of once-only colonoscopy might also be relevant for some settings in light of the increasing demand for colonoscopy resources generated from the implementation of a nationwide CRC screening program, population growth, and demographic aging. For instance, in a few European countries with fecal immunochemical test–based screening programs, the cut-off points for referral to colonoscopy had to be raised to match their colonoscopy capacity.
The challenge might be even bigger for resource-constrained countries. For these settings, once-only colonoscopy could potentially enable a more rational utilization of colonoscopy resources than repeating periodically and in the meanwhile provide sufficient degree of protection with higher level of sensitivity in detecting precancerous lesions than fecal immunochemical test.
Randomized controlled trials have shown that even a single flexible sigmoidoscopy could provide long-term protection up to at least 17 years, the maximum period of follow-up completed so far.
Long term effects of once-only flexible sigmoidoscopy screening after 17 years of follow-up: the UK Flexible Sigmoidoscopy Screening randomised controlled trial.
Major strengths of this simulation analysis include the use of data from one of the world’s largest screening colonoscopy registries and a large population-based cohort study of CRC patients for deriving model parameters. Some limitations, including model simplifying assumptions, should also be considered when interpreting our results. First, because of the lack of relevant data in the German screening colonoscopy program, the proportions of adenomas/cancers and transition rates between states among people aged 50 years were assumed to be the same as those among people aged 55. However, potential bias from violation of this assumption would likely be small, given that very similar prevalence of neoplasms was observed in age groups 50 to 54 and 55 to 59 in regional programs offering screening colonoscopy from age 50 on
The uncertainty in these parameters, however, is further captured to some degree by the sensitivity analyses using the confidence limits of the transition rates and the analyses with predicted proportions at age 45, which yielded similar results. Additionally, like in several other modeling studies,
a slightly higher overall prevalence of adenomas, a slightly lower transition rate from adenoma to cancer, and slightly smaller effects of screening colonoscopy in reducing CRC deaths or YPLL might be expected when the serrated pathway is incorporated. Third, in this analysis, people with adenoma removed were moved backward to the no-neoplasm state and were considered in the same way as those without colonoscopy findings in subsequent years based on the assumption that their excess risk could be compensated through surveillance colonoscopies. If this assumption is violated in real practice, for example by noncompliance with surveillance screening guidelines, the estimated CRC deaths or YPLL prevented might be smaller than our current estimates. Another limitation relates to the potential harms of screening colonoscopy, which were not included in our study; however, the impact of this on the outcome of CRC deaths or YPLL because of CRC death prevented shall be marginal, considering that fatal adverse events are extremely rare (2 deaths were reported among >2.8 million screening colonoscopies in the German national screening colonoscopy program 2003-2008)
and nonfatal adverse events do not contribute to death number or YPLL. Although nonfatal adverse events might have an influence on the quality of life, many of these adverse events are minor and the influence is transient.
In addition, our analyses only considered removal of adenomas detected at screening colonoscopy. However, given that the proportions of neoplasms at baseline and transition rates between states were exclusively derived from neoplasms detected in the German screening colonoscopy program, estimates of prevented CRC deaths should refer to these detected neoplasms only and can thus accurately reflect effects expected in real-life screening practice. Finally, our estimates of the lifetime risk of CRC death in the absence of screening are higher than the most recent national estimates in 2013 to 2014
Robert Koch-Institut und die Gesellschaft der epidemiologischen Krebsregister in Deutschland e.V. Krebs in Deutschland für 2013/2014. 11. Ausgabe [Cancer in Germany for 2013/2014. 11th edition]. Berlin.
; however, the observed lower risk is very likely to reflect the widespread use of colonoscopy screening in Germany, where more than half of the eligible population had undergone colonoscopy within the past 10 years.
In summary, our analyses provide model estimates of the optimal ages for once-only and repeated screening colonoscopies. Our results suggest that when colonoscopy capacity is limited and only a single examination can be offered, gender differentiation of screening age is expected to yield more protection against CRC death. Compared with women, men could potentially benefit more from having a screening colonoscopy at an earlier age. When repeated screening colonoscopies are performed, the optimal initiation age falls to around 50 years or even younger. Although the decision to lower the screening starting age to below 50 years might require more analyses in terms of screening benefits, costs, and harms among these younger adults, our estimates suggest that the recent policy change of lowering colonoscopy screening initiation from 55 to 50 years among men in Germany should also apply to women.
Appendix
Supplementary Figure 1Proportions of CRC deaths or YPLL prevented when a single, 2, or 3 screening colonoscopies were performed at various ages; base case and sensitivity analyses with upper and lower limits of 95% confidence interval of annual transition rates. A, Proportions of CRC death prevented. B, Proportions of YPLL prevented. CRC, Colorectal cancer; YPLL, years of potential life lost.
Supplementary Figure 1Proportions of CRC deaths or YPLL prevented when a single, 2, or 3 screening colonoscopies were performed at various ages; base case and sensitivity analyses with upper and lower limits of 95% confidence interval of annual transition rates. A, Proportions of CRC death prevented. B, Proportions of YPLL prevented. CRC, Colorectal cancer; YPLL, years of potential life lost.
Supplementary Figure 2Proportions of CRC deaths or YPLL prevented when a single, 2, or 3 screening colonoscopies were performed at various ages; base case; and sensitivity analyses with 25%, 50%, and 75% compliance rates. A, Proportions of CRC death prevented. B, Proportions of YPLL prevented. CRC, Colorectal cancer; YPLL, years of potential life lost.
Supplementary Figure 2Proportions of CRC deaths or YPLL prevented when a single, 2, or 3 screening colonoscopies were performed at various ages; base case; and sensitivity analyses with 25%, 50%, and 75% compliance rates. A, Proportions of CRC death prevented. B, Proportions of YPLL prevented. CRC, Colorectal cancer; YPLL, years of potential life lost.
Supplementary Figure 3Proportions of CRC deaths or YPLL prevented when 1, 2, or 3 screening colonoscopies were performed at various ages and sensitivity analyses with simulation starting at age 45. A, Proportions of CRC death prevented after a single screening colonoscopy. B, Proportions of YPLL prevented after a single screening colonoscopy. C, Proportions of YPLL prevented after two or three screening colonoscopies. CRC, Colorectal cancer; YPLL, years of potential life lost.
Supplementary Figure 3Proportions of CRC deaths or YPLL prevented when 1, 2, or 3 screening colonoscopies were performed at various ages and sensitivity analyses with simulation starting at age 45. A, Proportions of CRC death prevented after a single screening colonoscopy. B, Proportions of YPLL prevented after a single screening colonoscopy. C, Proportions of YPLL prevented after two or three screening colonoscopies. CRC, Colorectal cancer; YPLL, years of potential life lost.
Supplementary Table 1Most advanced findings at screening colonoscopy among first-time participants of screening colonoscopy program aged 55 years, 2003-2012
Transition rates from no neoplasm to nonadvanced adenoma, from nonadvanced adenoma to advanced adenoma, and from advanced adenoma to preclinical colorectal cancer for age ≥80 were assumed to be the same as those for ages 75-79. Transition rates from preclinical to clinical colorectal cancer were derived for ages 80-84 and were assumed to be the same for age ≥85.
Transition rates from no neoplasm to nonadvanced adenoma, from nonadvanced adenoma to advanced adenoma, and from advanced adenoma to preclinical colorectal cancer for age ≥80 were assumed to be the same as those for ages 75-79. Transition rates from preclinical to clinical colorectal cancer were derived for ages 80-84 and were assumed to be the same for age ≥85.
1.2 (.8-1.6)
3.7 (2.8-4.7)
5.6 (4.9-6.3)
17.3 (16.0-18.8)
∗ Estimates were extracted from references 23, 24, and 25.
† Transition rates for ages 50-54 were assumed to be the same as for ages 55-59.
‡ Transition rates from no neoplasm to nonadvanced adenoma, from nonadvanced adenoma to advanced adenoma, and from advanced adenoma to preclinical colorectal cancer for age ≥80 were assumed to be the same as those for ages 75-79. Transition rates from preclinical to clinical colorectal cancer were derived for ages 80-84 and were assumed to be the same for age ≥85.
Supplementary Table 5AParameters for sensitivity analyses with simulation starting at age 45: proportions of no neoplasm, nonadvanced adenoma, advanced adenoma, and colorectal cancer at ages 55-80 among participants of the German national screening colonoscopy program during 2003-2010 and the predicted proportions at age 45
Supplementary Table 5BParameters for sensitivity analyses with simulation starting at age 45: sex- and age-specific general mortality rates and average life expectancy for ages 45-49
Effect of screening sigmoidoscopy and screening colonoscopy on colorectal cancer incidence and mortality: systematic review and meta-analysis of randomised controlled trials and observational studies.
The impact of the rising colorectal cancer incidence in young adults on the optimal age to start screening: microsimulation analysis I to inform the American Cancer Society colorectal cancer screening guideline.
Gemeinsamer Bundesausschuss. Federal Joint Committee]. Beschluss des Gemeinsamen Bundesausschusses über eine Richtlinie für organisierte Krebsfrüherkennungsprogramme und eine Änderung der Krebsfrüherkennungs-Richtlinie. Decision of the Federal Joint Committee on a directive of organized cancer screening programs and an amendment to the cancer screening directive [In German]. Berlin.
Expected long-term impact of the German screening colonoscopy programme on colorectal cancer prevention: analyses based on 4,407,971 screening colonoscopies.
Declining bowel cancer incidence and mortality in Germany: an analysis of time trends in the first ten years after the introduction of screening colonoscopy.
Long term effects of once-only flexible sigmoidoscopy screening after 17 years of follow-up: the UK Flexible Sigmoidoscopy Screening randomised controlled trial.
Robert Koch-Institut und die Gesellschaft der epidemiologischen Krebsregister in Deutschland e.V. Krebs in Deutschland für 2013/2014. 11. Ausgabe [Cancer in Germany for 2013/2014. 11th edition]. Berlin.
DISCLOSURE: This project was supported in part by the German Federal Ministry of Education and Research (grant no. 01GL 1712). The following author received research support for this study from the Helmholtz International Graduate School for Cancer Research at the German Cancer Research Center (DKFZ): C. Chen. All authors disclosed no financial relationships relevant to this publication.
If you would like to chat with an author of this article, you may contact Dr Brenner at [email protected]
30002-1&id=gr1.jpg){kind=link}
30002-1&id=gr2.jpg){kind=link}
30002-1&id=gr3.jpg){kind=link}
30002-1&id=gr4.jpg){kind=link}
30002-1&id=fx1a.jpg){kind=link}
30002-1&id=fx1b.jpg){kind=link}
30002-1&id=fx2a.jpg){kind=link}
30002-1&id=fx2b.jpg){kind=link}
30002-1&id=fx3a.jpg){kind=link}
30002-1&id=fx3c.jpg){kind=link}