Conflicts of interest: The authors declare that there is no conflict of interest.
Jing Yuan FANG, Department of Gastroenterology, Renji Hospital, Shanghai Jiaotong University School of Medicine Shanghai Institute of Digestive Disease, Shanghai 200001, China. Email: firstname.lastname@example.org
OBJECTIVE: The best cure for colorectal cancer (CRC) lies on its early diagnosis and treatment. We aimed to provide the epidemiological features of advanced colorectal adenoma (A-CRA) and CRC in symptomatic patients and to determine whether the incidences of A-CRA and CRC increased simultaneously in China between 1990 and 2009.
METHODS: A total of 157 943 patients who had undergone a colonoscopy from 1990 to 2009 were enrolled, of which 6 777 patients had A-CRA and 3 503 had CRC. They were compared with controls in a stratified analysis. The detection rates of A-CRA and CRC in the 1990s and 2000s were also compared.
RESULTS: The detection rate of A-CRA increased 1.88-fold over the two decades, while that of CRC increased 0.66-fold. The percentages of patients with A-CRA and CRC who were elder than 50 years were significantly higher in the 2000s than those in the 1990s (P = 0.000). The changes of location of A-CRA and CRC during the two decades indicated a shift of lesions from the distal colon to proximal colon.
CONCLUSION: There was a significant increase in detection rate of A-CRA in the 2000s, but CRC did not increase at a similar speed. Our results indicated that the early detection and removal of colorectal adenoma in symptomatic patients might decrease the incidence of CRC.
The global incidence of colorectal cancer (CRC) is around one million cases per year. CRC causes nearly 500 000 deaths each year, making it one of the most commonly seen tumors and one of the leading causes of death from cancer worldwide.1,2 A study in Shanghai urban area showed that CRC was ranked as the third most prevalent malignancy.3 On the basis of the Hong Kong Cancer Registry, Hospital Authority, we know that the incidence of CRC in men in Hong Kong SAR, China has increased from 40 cases per 100 000 individuals in 1992 to 60 cases per 100 000 individuals in 2002.4,5
The pathological sequence that converts colorectal adenoma (CRA), especially advanced CRA (A-CRA), into CRC is well known. Multivariate analysis has revealed that the presence of villous features or a ≥11 mm diameter were the significant determinants of the malignant potential of CRA.2,6 A colonoscopy, which enables the detection and removal of CRA, is an effective method for the prevention of CRC.7,8 The National Polyp Study demonstrated that there was a reduction in risk of 76–90% in the incidence of CRC among carriers of colorectal polyps after a colonoscopic polypectomy over prolonged years of surveillance.9 Large case-control studies of screening colonoscopy have also shown a 50% reduction in the incidence of cancer in the examined segment of colorectum10.
On the basis of accumulating evidences, a colonoscopy is recommended by experts in many countries for the early detection and prevention of CRC in asymptomatic individuals with an average risk for CRC11–13. On the other hand, the Canadian Association of Gastroenterology recommended the colonoscopy as the screening modality only for patients with an increased risk for CRC. Currently, for individuals with an average risk the choice of screening method depends on the preferences of physicians and patients.14
Most CRC screening trials were performed in asymptomatic people15–17 and little is known about the screening of hospital-based symptomatic individuals. Although CRC in most patients develops from A-CRA the differential detection rates of A-CRA and CRC in symptomatic patients are not clear, especially in China. The aim of this study was to provide the epidemiological feature of colorectal neoplasia, as detected by colonoscopy in symptomatic patients, and to determine whether the incidences of A-CRA and CRC have increased simultaneously across Mainland China between 1990 and 2009.
MATERIALS AND METHODS
A multicenter, cross-sectional study was performed of Han Chinese symptomatic patients who had undergone their first colonoscopy in five medical centers in China from January 1990 to December 2009. These hospitals are located in four major cities in China: Nanfang Hospital in Guangzhou, Guangdong Province (Southern China); The Affiliated Drum Tower Hospital of Nanjing University Medical School in Nanjing, Jiangsu Province (Eastern China); Renji Hospital and Shanghai First People's Hospital in Shanghai (Eastern China) and the PLA. Military General Hospital of Beijing (Northern China).
The study protocol was approved by the institutional ethical review boards at each participating center and was in agreement with the revised Helsinki Declaration of 1983.
Colonoscopy in symptomatic individuals
A total of 157 943 symptomatic patients who had undergone a colonoscopy were enrolled. Lower gastrointestinal symptoms were defined as a change of bowel habit, abdominal pain, bloody stool or fecal occult blood. In instances in which the participant had undergone a colonoscopy repeatedly within 5 years, only the results of the first examination were taken for analysis. Individuals who had had inadequate bowel preparation, a colonoscopy that did not reach the cecum (incomplete), or aged under 16 years were excluded.
A colonoscopy with adequate withdrawal time and excellent patient preparation were performed by well-trained gastroenterologists using standard colonoscopic equipment. The extent of the colorectal segment examined and the quality of the bowel preparation were recorded. The polyp size was measured with a ruler after the removal of the polyp or by placing opened biopsy forceps against the lesion. Polyps up to 10 mm in diameter were removed immediately and larger polyps were removed during a separate procedure. The location of polyp was defined as distal (from the rectum to the descending colon), proximal (from the splenic flexure to the cecum) or both (multiple polyps located in different places). When available, the distance in cm from the anus to the lesion was documented.
Uniform diagnostic criteria for both A-CRA and CRC were adhered to, and a similar miss rate of colorectal neoplasia was found in all these five centers. A-CRA was defined as the presence of adenomatous polyp sized ≥10 mm in diameter, villous features (at least, those that were 25% villous) or high-grade dysplasia.7,10,17 In instances in which an individual had multiple adenomas, the data of the largest one were taken.18 In instances in which two or more adenomas were equal in size, then the location and histological features of all relevant adenomas were taken.19 CRC was defined by its pathological features in accordance with the World Health Organization guidelines.20 If patients met the criteria of A-CRA or CRC they were categorized into Group A-CRA or Group CRC for further analysis.
Biopsy specimens were examined by pathologists in accordance with the standard protocol issued by the World Health Organization. If available, the histopathology reports of surgical resection specimens were also used to confirm the abovementioned data.
Patients diagnosed with familial adenomatous polyposis or hereditary nonpolyposis colorectal cancer were excluded. Adherence to the standard criteria of colonoscopy was assessed by analyzing the endoscopy report along with the histopathology report, as well as the histopathology report of the surgical resection specimen, if available.
In order to make the data from 2000s period comparable to those of 1990s we chose the same time period (one decade) for comparison and analysis: 1990−1999 (the 1990s) and 2000–2009 (the 2000s). Individuals were stratified by their gender and age data. As most current US and European guidelines set the cut-off age threshold for an endoscopy at 50 years and Asian guidelines recommend an endoscopy for those aged ≥50 years. The age groupings in this study were stratified into ≤49 years, 50–74 years and ≥75 years. Pathological variables, including size, location and number of lesions were also stratified for analysis.
Moreover, a logistic regression analysis was performed for the age, gender and time period of the colonoscopies in patients with and without A-CRA and CRC. The increase of detection rates of A-CRA and CRC between the 1990s and 2000s was calculated and analysed.
The statistical analysis was performed using the SPSS 15.0.1 (SPSS Inc., Chicago, IL, USA). Descriptive statistical analysis included rates and proportions for categorical data and mean ± standard deviation (SD) for continuous data. Continuous variables were tested by Student's t-test. Categorical variables were tested by a χ2 test, Fisher's exact test or Wilcoxon rank–sum test. A standard logistic regression analysis was used to calculate the relative risk as odds ratios (OR) with 95% confidence intervals (CI).21 A P-value of less than 0.05 was considered significant.
Altogether 157 943 patients who had undergone a colonoscopy between January 1990 and December 2009 at these five medical centers were enrolled and categorized into two groups: 1900s (1990–1999) and 2000s (2000–2009), according to the date of their colonoscopy. The study flow chart is summarized in Figure 1. The characteristics of these individuals is shown in Table 1.
Table 1. Characteristics of individuals (N= 157 943) who had undergone a colonoscopy including detection rates of advanced colorectal adenoma (A-CRA) and colorectal cancer (CRC)
1990–1999 (N = 26 026)
2000–2009 (N = 131 917)
Age in years
15 230 (58.52)
59 296 (44.95)
10 251 (39.39)
64 859 (49.17)
7 762 (5.88)
14 982 (57.57)
70 861 (53.72)
11 044 (42.43)
61 056 (46.28)
Detection rate of A-CRA
6 343 (4.81)
Detection rate of CRC
3 130 (2.37)
Of all patients studied 6 777 were detected as having A-CRA and 3 503 patients had CRC. The remaining participants (147 663 individuals) served as controls. Logistic regression analysis showed that their age (≤49, 50–74 or ≥75 years) and the time when they underwent the colonoscopy (1990s or 2000s) were the significant factors influencing the detection rates of A-CRA and CRC (P = 0.000) (Table 2).
Table 2. Logistic regression analysis in the cohort study of advanced colorectal adenoma (A-CRA) and colorectal cancer (CRC) (N = 157 943)
Clinicopathological characteristics of A-CRA and CRC (Table 3)
Table 3. Clinicopathological characteristics of advanced colorectal adenoma (A-CRA) (n = 6777) group and colorectal cancer (CRC) (n = 3503) group over two decades
SD, standard deviation.
Detection rate of A-CRA (%, n)
Age (years, mean ± SD)
57.75 ± 13.87
56.98 ± 11.11
≤49 years (n, %)
1 625 (25.62)
50–74 years (n, %)
4 205 (66.29)
≥75 (n, %)
Male (n, %)
4 090 (64.48)
Female (n, %)
2 253 (35.52)
Proximal only (n, %)
1 287 (20.29)
Distal only (n, %)
4 485 (70.71)
Both (n, %)
Number = 1 (n, %)
4 839 (76.29)
Number ≥2 (n, %)
1 504 (23.71)
Detection rate of CRC (%, n)
Age (years, mean ± SD)
53.12 ± 13.29
52.59 ± 12.12
≤49 years (n, %)
50–74 years (n, %)
2 033 (64.95)
≥75 years (n, %)
Male (n, %)
1 869 (59.71)
Female (n, %)
1 261 (40.29)
Proximal (n, %)
Distal (n, %)
2 428 (77.57)
Both (n, %)
Overall the male : female ratio was 6:4, and most of the patients with A-CRA were elder than 50 years. The percentage of patients elder than 50 years in the 2000s was significantly higher than that in the 1990s (P = 0.000). The location of the A-CRA lesion in the 2000s was different from that in the 1990s (P = 0.002), with a trend towards multiple lesions seen in both the proximal and distal colon. The detection rate of A-CRA located in both the distal and proximal colon increased from 4.15% in the 1990s to 9.00% in the 2000s. As for the number of A-CRA, the detection rate of a single polyp was higher than that of many polyps (≥2) over the full course of the study (3.27% vs 1.02%). However, there was no significant difference in the percentages of single to multiple polyps during both periods (P > 0.05).
For CRC, the overall male : female ratio was 6:4 and most of the patients with CRC were elder than 50 years. The percentage of patients elder than 50 years in the 2000s was significantly higher than that in 1990s (80.13% in the 2000s vs 69.71% in the 1990s, P < 0.05), which was similar to that of A-CRA. The percentage of CRC located in the proximal colon in the 2000s was significantly higher than that in the 1990s (P = 0.005), indicating a trend towards a shift of location from the distal colon to the proximal colon.
Overall, patients in the age group 50–74 years remained the largest percentage of symptomatic patients with A-CRA or CRC. During the 2000s the constitution ratios of A-CRA and CRC in this age group were 66.29% and 64.95%, respectively; and during the 1990s they were 62.67% and 64.88%, respectively.
Increase of detection rates of A-CRA and CRC
The detection rates of A-CRA in the 1990s and the 2000s were 1.67% and 4.81%, respectively, and the detection rates of CRC were 1.43% and 2.37%, respectively (Table 4). From these data we calculated that the increase in detection rate of A-CRA was approximately 1.88-fold over the past two decades, while the increase of CRC was 0.66-fold, significantly lower than that of A-CRA.
Table 4. Comparison of detection rate between advanced colorectal adenoma (A-CRA) group and colorectal cancer (CRC) group over two decades
Detection rate of A-CRA (%, n)
≤49 years (%)
50–74 years (%)
Proximal only (%)
Distal only (%)
Number = 1 (%)
Number =≥2 (%)
Detection rate of CRC (%, n)
≤49 years (%)
50–74 years (%)
≥75 years (%)
Increase of detection rates of A-CRA and CRC in different age groups
The detection rates of A-CRA increased from 0.86% to 2.74%, from 2.65% to 6.48%, and from 5.69% to 6.61% in the 1990s to the 2000s for the groups aged ≤49 years, 50–74 years and ≥75 years, respectively.
Although patients with A-CRA or CRC were most likely to be 50–74 years, the increase of detection rate of A-CRA in age groups ≤49 years (a 2.19-fold increase) was higher than that of 50–74 years group (1.45-fold increase). The CRC detection rate increased more in the ≥ 75 years age group (a 0.85-fold increase).
Increase of detection rate of CRC in different locations
The increase of the rate of detecting CRC in the proximal colon was higher than that in distal colon (a 1.27-fold vs a 0.52-fold increase). This might indicate a proximal shift of the CRC lesion over the past two decades. A sigmoidoscopy alone is not a reliable test for excluding the proximal cancer that occured in 20.99% of patients in the 2000s.
The incidence of colorectal neoplasia is increasing rapidly in Asia, but differences in the detection rates of CRA and CRC are not clear, particularly in China. In order to address this issue we carried out this multicenter, hospital-based study to analyze the clinical and pathological features of A-CRA and CRC detected by colonoscopy in symptomatic patients during the last two decades in China.
The results of our study represent the changes in colonoscopy results in symptomatic patients over the last two decades. We believe that the results are valid for at least two following reasons. First, the universal access to medical procedures is a feature of the health-care insurance system in China. Since the urban medical insurance scheme in China accounts for approximately 95% of the population, we make the assumption that over 90% of the symptomatic individuals in urban areas would choose to undergo a colonoscopy. Second, the patient data were collected from five major class A tertiary hospitals located in Beijing (Northern China), Nanjing and Shanghai (both in Eastern China) and Guangzhou (Southern China), which denotes an excellent level of medical service and qualified colonoscopy.
The data presented demonstrate that a significant increase in detection rate of A-CRA was observed over the recent 10 years; however, the detection rate of CRC did not increase with similar speed. This might partly be explained by the increasing use of colonoscopies, as well as the improvement in the recognition of adenomas, which affords an opportunity to remove benign adenomatous polyps before they progress and become malignant.11 The detection rate of CRC would rise if A-CRA were not identified and removed. This suggests the importance of screening colorectal neoplasms with the removal of the A-CRA thus detected for reducing the incidence of CRC.
Current guidelines recommend screening colonoscopies in asymptomatic average-risk individuals who aged over 50 years.22–24 Few studies have been performed on the detection rate of premalignant A-CRA in symptomatic patients. Rundle et al.15 reported that the detection rate of A-CRA in screening colonoscopies of average-risk individuals aged 40–49 years was similar to that in individuals aged 50–59 years. In our study the detection rate of all these lesions increased when patients were elder than 50 years, which agreed with previously published reports.25 However, although patients with A-CRA or CRC were most likely to be 50–74 years in our study, the increase in the detection rate of A-CRA in symptomatic patients aged ≤49 years and the detection rate of CRC in symptomatic patients aged ≥75 years were higher than that in the 50–74 year-old group.
The correlation between the location of A-CRA and the ultimate outcome has been a focus of study.11 Levin et al. demonstrated that advanced proximal neoplasia were not uncommon in patients with or without distal tubular adenoma.26 However, other scientists have suggested that right-sided lesions were more likely to be fast growing.27 Recently, Schoepfer et al. analyzed the results of a screening study performed in asymptomatic individuals underlining the significance of findings in the rectosigmoid for the detection of neoplastic lesions in the proximal colon.11 Data from the Prince of Wales Hospital, Hong Kong SAR, China, demonstrated that the incidence of distal A-CRA was negatively associated with the time period, whereas the incidence of proximal and synchronous A-CRC was positively correlated with it.5 Our data in symptomatic patients show a significant difference in the percentages of lesions (A-CRA or CRC) distributed in the proximal colon versus the distal colon when comparing the 2000s with the 1990s. The increase of detection rate of CRC in the proximal colon was higher than that in the distal colon. This might indicate a proximal shift of CRC lesions over the last two decades. A sigmoidoscopy alone is not a reliable test for excluding the proximal cancer that occurred in 20.99% of patients during the 2000s. Although the presence of distal advanced neoplasia was a significant risk factor for a proximal advanced neoplasm,27 in our study, 22.12% and 20.29% of patients with A-CRA in the proximal colon during the 1990s and the 2000s had no evidence of CRA in the distal colon, respectively. Therefore, many of these neoplasms could not been detected by a sigmoidoscopy. Many studies have shown that a significant number of patients with advanced proximal neoplasms did not have their adenomas detected by sigmoidoscopy28 and up to 70% of those with proximal CRC lack a distal lesion.27 This is a challenge to clinicians because CRA in the proximal colon is more difficult to treat.
The recommendations for CRC screening are currently based solely on age and a family history of cancer.10 Our study demonstrated there were higher percentages of A-CRA or CRC in men than in women, which may result from the different lifestyles and dietary habits, as well as the different work environments between men and women in urban areas of China. Further studies will be needed to find out whether gender might be a factor related to risk of CRA.
Sung et al. have reported the first multicenter study of colonoscopies performed in 860 patients to examine the characteristics of colorectal neoplasms in Asia.4 However, our study is a long-term cohort study of a large sample in China and our results could be used as a reference for colonoscopy practice in China.
There were several important limitations to our study. The major limitation was that we failed to collect information about dietary habits and use of medication, which might affect the development of adenomas.7,18,29 Another drawback was the cross-sectional nature of this study, which prevented the including of the development of colonoscopy techniques in the analysis and might lead to an underestimate of detection rates obtained during the 1990s compared with the 2000s.
In conclusion, this is the first large multicenter colonoscopy study comparing the detection rates and characteristics of A-CRA and CRC in symptomatic patients in Mainland China. Our data demonstrated that the detection rate of A-CRA increased 1.88-fold while that of CRC increased 0.66-fold over the last two decades. The increase of CRC would be higher if the adenomas had not been detected and removed. The changes of location of A-CRA and CRC during the last two decades indicated a shift of lesions from the distal colon to the proximal colon. These results will have important implications for planning screening programs for colorectal neoplasia, emphasizing the removal of A-CRA to reduce the incidence of CRC.
We thank Dr Xiao Yu CHEN, Zhi Zheng GE, Yao Hao CUI, Ji Lin WANG, Qi MIAO, Yun CUI and Yan Wei LIN for their help during the preparation of this manuscript. The authors also thank the pathologists at each of the five participating centers who made a significant contribution to the performance and successful completion of the study.
We also thank the Ministry of Public Health, China for Grant No. 200802094, Shanghai Jiaotong University School of Medicine Foundation for Science and Technology for Grant No. BXJ0914 to Hui Min CHEN, and the Ministry of Education for the Scholarship Award for an Excellent Doctoral Student to Hui Min CHEN.