1319 Rectal Epithelial Proliferation in Persons with or without a History of Adenoma and Its Association with Diet and Lifestyle Habits Paul Rozen, M.B. B.S.1 Flora Lubin, M.Sc.1 Naomi Papo, R.N.1 Gershom Zajicek, M.D.2 1 Department of Gastroenterology, Tel Aviv Medical Center, Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel. 2 H. Humphrey Center for Experimental Medicine and Cancer Research, Hebrew University, Jerusalem, Israel. Presented as a poster at the annual meeting of the American Gastroenterological Association, Washington, DC, May 11–17, 1997; and published in abstract form in: Rozen P, Lubin F, Papo N, Farbstein H, Farbstein N, Zejicek G. Association of diet and lifestyle habits with rectal epithelial proliferation measurements in persons with or without colorectal adenomas. Gastroenterology 1997;112: A647. Drs. Rozen and Lubin were supported by a grant from the National Council for Research and Development, Israel and the German Cancer Research Center (DKFZ), Heidelberg (in collaboration with Dr. H. Boeing). The authors thank Ms. S. Zimmerman for secretarial assistance, Ms. V. Liberman and Ms. J. Knaani for data and patient management, and H. and M. Farbstein (ICRF) for statistical analysis. Address for reprints: Paul Rozen, M.B., B.S., Department of Gastroenterology, Tel Aviv Medical Center, 6 Weizmann Street, Tel Aviv 64239, Israel. Received September 15, 1997; revision received January 14, 1998; accepted March 18, 1998. © 1998 American Cancer Society BACKGROUND. Rectal epithelial proliferation (REP) measurements are used as a biomarker of risk for colorectal neoplasia and response to chemoprevention. The authors evaluated REP in screenees with and without a history of adenoma and its association with demographic and adenoma characteristics, diet, and other lifestyle habits. METHODS. Long term lifestyle habits were evaluated and proliferation assessed by in vitro bromodeoxyuridine labeling of rectal biopsies in 223 screenees, 132 of whom had adenomas removed . 3 years previously. Analyses included the total population, screenees with a previous history of adenomas and adenoma free screenees separately, and a subgroup of 55 matched adenoma cases and controls. RESULTS. Crypt proliferation measurements were not elevated in screenees with a history of adenomas compared with adenoma free screenees (mean total labeling index [LI] of 4.8% and 4.9%, respectively). This was confirmed by the case-control analysis, in which the LI of the most superficial crypt compartment was lower in the adenoma cases (P 5 0.05). Moreover, their total LI correlated negatively with the number of adenomas removed previously (P , 0.01). Proliferation was more frequent in the most superficial crypt compartments of female adenoma free screenees than in female screenees with a history of adenomas (P 5 0.02), and in men age . 65 years compared with younger men (P 5 0.06). In the total population, negative Spearman rank correlations were found between total LI and long term dietary intake of calcium (correlation coefficient [r] 5 20.15; P 5 0.02), LI of the two most superficial crypt compartments and intake of fiber (r 5 2 0.18; P 5 0.01), water (r 5 20.12; P 5 0.08), and carbohydrates (not significant). A positive correlation was found between LI of the most superficial crypt compartment and cigarette smoking (r 5 0.4; P 5 0.02). CONCLUSIONS. REP measurements did not discriminate between screenees with a history of adenomas and adenoma free screenees. Long-term lifestyle habits, gender, and age were associated with REP levels and need to be considered when evaluating human intervention studies. Cancer 1998;83:1319 –27. © 1998 American Cancer Society. KEYWORDS: adenoma, chemoprevention, colorectal, diet, family history, lifestyle habits, proliferation. M easurements of colonic epithelial proliferation are believed to be elevated in persons at risk for or who have colorectal neoplasia.1–5 In addition, they have been useful biomarkers for following the effects of chemoprevention or nutritional intervention in animal experiments, in which there is full control over confounding factors such as genetic strain, gender, age, and diet.2,6 In humans, some interventional studies have shown that short term dietary changes or 1320 CANCER October 1, 1998 / Volume 83 / Number 7 dietary supplements had an effect on the rectal epithelial proliferation (REP) and these are believed to reflect similar changes throughout the entire colon.2,7–9 However, other intervention studies in humans were not always successful in demonstrating such a response in REP.10,11 This possibly was due to the difficulty in controlling for the confounding influences on proliferation of genetic predisposition for neoplasia, age, gender, diet, and other lifestyle habits of the normal adult.12–14 There have been only a few studies in humans examining and demonstrating the effects of dietary and other lifestyle habits on REP measurements.1,2,4,5,15–17 For this reason, within the framework of an adenoma case– control nutrition and lifestyle study,18 we examined the REP measurements in cases who had adenomas in the past and in controls who consistently were adenoma free. We also examined the associations between dietary and other lifestyle habits with REP in the total group and its study populations. degree relative with colorectal neoplasia were not included in the current study. There were 196 adenoma cases and matched controls who fulfilled the inclusion criteria and agreed to participate in the case– control study.18 Of these, 132 adenoma cases (67.3%) also agreed to participate in the evaluation of their REP. None had adenomas at this time. After excluding controls with a family history of colorectal neoplasia, 91 of the adenoma free screenees (46.4%) also fulfilled the entry criteria for this part of the study and agreed to participate in the evaluation of their REP. STUDY POPULATION AND METHODS Rectal Epithelial Proliferation Study Population REP was assessed in rectal biopsies taken between 8 and 10 A.M. in the nonfasting state, and without bowel preparation. This technique and the immunohistochemical method used have been described in detail elsewhere.21 In summary, by guiding flexible endoscopic forceps under the index finger inserted into the rectum, 3 pinch biopsies were taken from each individual, 8 –10 cm from the anal verge. Under a dissecting microscope the specimens were oriented mucosal side up on a membrane filter. They then were incubated in Eagle’s medium with bromodeoxyuridine (BrdU), from the Amersham proliferation kit (Amersham, Buckinghamshire, U.K.), at 37 °C for 1 hour in a hyperbaric chamber containing 95% oxygen and 5% carbon dioxide under 1 added atmosphere pressure. The ethanol fixed specimens then were carefully reoriented, paraffin blocks prepared, and sections cut. Every consecutive five sections were excluded systematically to obtain samples of different crypts on at least four slides. Sections were dewaxed and rehydrated through graded ethanols. After DNA denaturation with hydrochloric acid (which was neutralized), and the addition of normal sheep serum, the Amersham kit monoclonal antibody was used for the immunohistochemistry to reveal nuclei that had taken up the BrdU. These were intensified by dipping in diaminobenzidine with NiCl2 and the slides were counterstained weakly with hematoxylin and mounted. The biopsies were coded and evaluated twice by a single experienced technician who was blinded to the Our population was comprised of 223 urban adults who participated in the nutrition-adenoma case– control study.18 These included all the consecutive adenoma patients identified in the Screening Program of the Gastroenterology Department at the Tel Aviv Medical Center19 who had been followed for at least 3 years by repeated colonoscopic examinations. Our follow-up protocol, after clearing the colon of polyps, included repeated colonoscopic examinations after 1 year and then every 2 years to prove that the colon was free of adenomas or to determine whether recurrence had occurred.19,20 The controls were asymptomatic adenoma free screenees, matched to cases by age (6 5 years), gender, country of origin, and duration of follow-up ($ 3 years) in the screening program. All controls had been screened by annual fecal occult blood tests and a flexible sigmoidoscopic examination that had been repeated at least twice, after 1 year and then every 2 years.19,20 No adenoma ever was detected in the controls. All pathology specimens were reviewed by one pathologist and classified by standard criteria as part of an ongoing study of colorectal polyps.19 The inclusion criteria for the nutrition-adenoma case– control study were: ages 21–75 years, adenoma size $ 0.5 cm; systematic follow-up that included the endoscopic and fecal occult blood examinations; and medical status (no history of malignancy, inflammatory bowel disease, or chronic disease requiring a strict diet or use of calcium supplements). Controls with a first- Data Collected Data included information available from the adenoma nutrition case– control study with regard to family history of neoplasia, dietary habits, physical activity, height, weight, lifetime smoking habits, alcohol intake, vitamin or mineral supplementation, and medications used.18 The site, size, number, and histology of the initial and recurrent polyps (if there were any) was noted. Adenomas, Diet, and Epithelial Proliferation/Rozen et al. diagnosis of adenoma. Under a microscopic magnification of 3 500 –750, at least 15 complete longitudinal crypts were scored with care taken not to reexamine the same crypt. Each side of the crypt was counted from the base (muscularis mucosa) to the surface, recording the number of cells, the number of labeled cells, and their position in the hemicrypt. The scoring technique had been well established and standardized in our laboratory with good reproducibility.22 The scoring data for each person were recorded into a database for statistical analysis together with the other collected data regarding nutrition and lifestyle habits. Diet and Lifestyle Habits The nutritional methodology had been developed at the Epidemiology Department, Chaim Sheba Medical Center and is described in full elsewhere.18 Trained interviewers conducted a standardized interview using a quantified dietary questionnaire that included 180 food items. Subjects were asked to describe for each item the frequency and number of portions consumed and the usual portion size by natural or standard units of package, or individual portion size as determined by life-size color photographs. The adult population is known to change nutritional habits for health and other reasons. Therefore, for each item, interviewers were trained to ask whether any changes occurred in the subject’s consumption habit during the last 15 years. If yes, first the more recent and then the long term, prechange habits (“past” consumption) were recorded in an identical set of columns that included the same questions as well as the number of years since the change occurred. For items in which no change was reported, current habits were equivalent to past habits. This method allowed evaluation of the “past” habits (before changes occurred) in the same detail as for the more “recent” ones. The results presented are based on analysis of the long term “past” dietary habits and their association with REP. Physical activity was categorized using activity criteria (resting and sitting hours and time spent in light and moderate to heavy activity) similar to those described in the Recommended Daily Allowances.23 Lifetime tobacco smoking habits were ascertained by noting the number of cigarettes currently smoked daily or in the past, and duration of smoking. Current weight and height, weight at age 18 years, and weight for the majority of the subject’s adult life were recorded. Written informed consent for the study had been obtained from all subjects and the study was approved by the hospital’s Ethics Committee in 1989. 1321 Data Analysis and Statistical Evaluation An adenoma score was made by: size( 0.5– 0.9 cm 5 1; and $ 1.0 cm 5 2) and histology (tubular 5 1; villotubular 5 2; villous 5 3; and high grade dysplasia 5 4) for each adenoma detected. Analysis of REP parameters included: mean numbers (6 standard deviation) of hemicrypts counted, cells per hemicrypt, and BrdU-labeled cells per hemicrypt. A mean total labeling index (LI) was computed of the ratio of labeled to total number of crypt cells. In addition, each hemicrypt was divided into 5 equally sized compartments (from base to surface) and the mean LI in each compartment, and then the LI of 2 most superficial compartments (upper 40%), also were evaluated. Calculations also were made of the ratio of the number of labeled cells in the upper 40% of the crypts to the total number of labeled cells (Øh4&5),24 or to the number of hemicrypts counted (Lh4&5) (M. Lipkin, personal communication), or of having any labeled cells in the most superficial compartment.2,3 Three levels of LI (low, intermediate, and high) also were determined according to the distribution of LI values for the total population. Reported food consumption was converted into average daily amounts consumed. A comprehensive computer program contained information regarding food composition from international and local sources and took into account seasonal fruits and vegetables.25,26 The program converted all frequencies and portion sizes to mean daily, long term, past consumption. For some calculations, three levels of intake (low, medium, and high) were determined based on the distribution of intake by the total population, the adenoma cases, and the adenoma-free screenees. Physical activity was converted to mean daily active hours and an “activity score” computed for each individual. This was comprised of the number of daily hours spent in moderate to heavy activity and half of the light activity hours.18,23 Weight and weight gain from age 18 years to adulthood were analyzed as such and by body mass index (BMI). Tobacco use was assessed for present and past smokers by a score based on the mean number of cigarettes smoked per day 3 365 3 number of years smoking. The number, size, and histology of each adenoma detected was recorded. Statistical analyses were performed both on the total population and on the adenoma cases and adenoma free screenees separately. Conditional logistic analysis, adjusting for physical activity and energy intake, was performed in a subgroup of 55 adenoma cases and 55 adenoma free screenees, matched by age (6 5 years), gender, country of origin, and duration of 1322 CANCER October 1, 1998 / Volume 83 / Number 7 TABLE 1 Distribution of the Total Study Population (n5223) by Diagnosis, Demographic, and Rectal Epithelial Proliferation Data No. Men (%) European origin (%) Age (yrs) (mean 6 SD) Family history of colorectal carcinoma (%) No. hemicrypts counted (mean 6 SD) No. cells/hemicrypts (mean 6 SD) No. labeled cells/hemicrypt (mean 6 SD) LI 3 100 (mean 6 SD) LI/compartment 3 100 (mean 6 SD) Compartment 1 2 3 4 5 Adenoma cases Adenoma free screenees 132 60.6a 61.4 61.3 6 8.1 91 45.1a 60.0 63.5 6 6.3 33.3 0b 29.9 6 4.8 30.3 6 4.6 53.5 6 6.7 51.9 6 6.4 2.54 6 1.10 4.8 6 1.8 2.53 6 1.08 4.9 6 1.9 7.1 6 3.2 8.5 6 3.3 5.7 6 3.2 2.2 6 1.8 0.2 6 0.4a 7.3 6 2.9 8.8 6 3.5 5.8 6 3.3 2.3 6 1.8 0.3 6 0.5a SD: standard deviation; LI: labeling index. a P , 0.05. b Excluded by inclusion criteria. follow-up in the screening program.27,28 Spearman correlation coefficients were evaluated between the various measurements of REP and each of the dietary and lifestyle habits. To examine whether having a family history of colorectal neoplasia influenced the results, the analyses were repeated excluding those pairs with adenoma cases having a positive family history. RESULTS Demographic Characteristics There were 223 subjects in the total population studied and, consistent with the screened population of Tel Aviv, the majority were of European origin entering their sixth decade of life. The demographic characteristics of the adenoma cases and adenoma free screenees in the study population are presented in Table 1. There were relatively more women in the adenoma free screenees because fewer healthy men agreed to have their REP assessed (P , 0.05). Fortyfour of the adenoma cases (33.3%) had a first-degree relative with colorectal neoplasia; because of the study inclusion criteria there were no adenoma free screenees with a positive family history. REP Characteristics of the Study Population In the 132 adenoma cases, Spearman correlation coefficients were examined between the various parameters of proliferation and the adenoma characteristics: their number in each case, maximum size, sum of sizes, and a score that included these data and histology (data not shown). No significant positive correlations were found. Unexpectedly, a significant negative correlation was found between the total number of adenomas removed and the mean total LI measurement (correlation coefficient [r] 520.23; P , 0.01). The proliferation characteristics of the adenoma cases and adenoma free screenees (Table 1) were not significantly different, other than in Compartment 5. In Compartment 5 the adenoma cases had significantly less (P , 0.05) proliferation than the adenoma free screenees. To further explore these findings, we performed a conditional logistic analysis on 55 matched adenoma cases and controls, adjusting for energy intake and physical activity. Results of this analysis showed a significantly negative association between risk for adenoma and level of proliferation (LI) in Compartment 5 (odds ratio 5 0.37; 95% confidence interval, 0.14 – 0.99; P 5 0.05). However, no risk was associated with the total crypt LI or any other measure of crypt proliferation. Analyses also were performed by gender within and between the adenoma cases and adenoma free screenees. Female adenoma patients were less likely to have any proliferation in Compartment 5 compared with female adenoma free screenees (23.1% vs. 44.0%; P 5 0.02). Analysis then was made comparing the combined male and female populations from each study population and there were no significant differences in proliferation parameters between genders. The effect of age was evaluated as well. Men age . 65 years of age were more likely to have proliferation occurring in Compartment 5 than those age , 65 years (45.8% vs. 30.0%; P 5 0.06). The differences in proliferation measurements between past adenoma cases and adenoma free screenees were not consistent with increased epithelial proliferation in adenoma cases. Differences based on gender or age were limited to Compartment 5.29 Therefore, further analyses were performed for the total population as well as for the adenoma patients and adenoma free screenees separately. Diet and Other Lifestyle Habits and REP in the Total Population The significant associations of dietary and other lifestyle habits with the REP parameters of the total Adenomas, Diet, and Epithelial Proliferation/Rozen et al. TABLE 2 Significant Spearman Correlation Coefficients between Long-Term Dietary and Other Lifestyle Habits and Rectal Epithelial Proliferation Measurements in the Total Population (n 5 223) Parameters measured Total labelling index: Calcium LI Compartments 4 and 5: Fiber LI Compartment 5: Tobacco smokers’ score (n 5 32) Correlation coefficient P value 20.15 0.02 20.18 0.01 0.40 0.02 LI: labeling index. population are presented in Table 2. Increased nutritional intake of calcium was negatively correlated with the mean total LI (P 5 0.02) and a similar finding was observed for fiber, but limited to combined Compartments 4 and 5 (P 5 0.01). Trends in the same direction were found with both water and carbohydrate intakes and LI levels in combined Compartments 4 and 5 (data not shown) (r 5 20.12, P 5 0.08; and r 5 20.11, P 5 0.11, respectively). No significant correlations were observed with mean daily energy or fat intake or physical activity and any measurements of REP, but a nonsignificant negative correlation between weight gain through adult life and LI was found in combined Compartments 4 and 5 (data not shown) (r 5 20.11, P 5 0.13). The total amount of cigarettes smoked was significantly and positively correlated with the LI of Compartment 5 (P 5 0.02). To further evaluate the quantitative effect of calcium intake on REP we studied the percentage of individuals consuming the highest and lowest amounts of calcium by their levels of total LI (Table 3). Among those in the highest tertile of calcium consumption ($ 1027 mg/day), there was a significantly lower percentage of individuals within the highest tertile of LI levels (P 5 0.01); conversely, among the lower calcium consumers, a lower percentage of individuals was observed in the lowest tertile of LI levels. There was a significantly decreasing linear trend for total LI among individuals in the higher level of calcium intake (P 5 0.01). The same results were obtained after adjusting for energy intake (P for linear trend 5 0.004). A nonsignificant trend, in the same direction, was observed with fiber (data not shown). 1323 Dietary and Other Lifestyle Habits and REP by Adenoma Status Among adenoma free screenees, the intake of fiber was significantly negatively correlated with the total LI (P 5 0.02) (Table 4). In the adenoma cases, this negative association was limited to the combined crypt Compartments 4 and 5 (P 5 0.03). Negative but nonsignificant correlations between calcium intake and total LI were observed in both the adenoma free screenees and adenoma cases (data not shown) (r 5 20.17, P 5 0.1; and r 5 20.13, P 5 0.1, respectively). No significant correlations were found between energy or fat intake or level of physical activity and measurements of proliferation in either group. There was a negative correlation between weight gain during adult life and LI in combined Compartments 4 and 5 of the adenoma cases (data not shown) (r 5 20.18, P 5 0.05), but not in the adenoma free screenees. In the adenoma cases, there was a significantly negative correlation (P , 0.01) between the LI of combined Compartments 4 and 5 and being a nontobacco smoker or past smoker (n 5 79). Family History We repeated the analyses after excluding the adenoma cases with a family history of colorectal neoplasia. The results were fully consistent with the original findings and statistical significance was maintained. DISCUSSION The main findings in this large study were the lack of major differences in REP measurements between persons having had an adenoma .3 years previously, with or without a family history of colorectal neoplasia, and adenoma free individuals without this family history. This was not consistent with expectations that persons at risk for colorectal carcinoma would have persistently increased proliferation (a “field effect”) in the normal-appearing rectal epithelium.1,3 In addition, we found associations between certain dietary and lifestyle habits with REP measurements. In the adenoma-nutritional case– control study (which included this same population), numerous dietary and lifestyle factors were evaluated.18 The main risk factors for adenoma found in that study included: lower carbohydrate, increased animal fat, reduced water and fiber, and increased energy intake; weight gain; reduced physical activity; and cigarette smoking. Therefore, their associations with REP measurements were examined specifically in addition to numerous other lifestyle habits. Although calcium intake had not been found to be associated with adenoma in our 1324 CANCER October 1, 1998 / Volume 83 / Number 7 TABLE 3 Joint Distribution of Long-Term Calcium Intake (2 Levels) and 3 Levels of Total LI in the Total Population (n 5 223) % of population by tertiles of LI% Level of calcium intake/day No. Low < 3.8 Intermediate 3.8–5.2 High > 5.2 P valuea Lowest and intermediate tertiles (, 1027 mg) Highest tertile ($ 1027 mg) 152 71 27.0 46.5 36.2 28.2 36.8 25.8 0.01 LI: labeling index. a Mantel–Haenszel chi-square test for linear trend (P 5 0.004). TABLE 4 Significant Spearman Correlation Coefficients between Long-Term Dietary and Other Lifestyle Habits and Rectal Epithelial Proliferation Measurements of Each Study Group in the Total Population Adenoma cases (n 5 132) Parameter measured Total LI: Fiber LI Compartments 4 and 5: Fiber Nonsmokers & past tobacco smokers (n 5 79) Adenoma free screenees (n 5 91) Correlation coefficient P value Correlation coefficient P value 0.07 NS 20.24 0.02 20.19 0.03 20.16 NS 20.31 , 0.01 0.06 NS LI: labeling index; NS: not significant. nutritional case– control study,18 we examined its association with REP as part of the calcium intervention trial that also was being performed.20 Limitations and Strengths of the Study We attempted to find reasons to explain the relatively low REP measurements in our population and the lack of REP differences between our past adenoma cases and the adenoma free screenees. The use of hyperbaric oxygen ensured uniform tissue labeling with BrdU which, in contrast to proliferating cell nuclear antigen, easily is recognized and scored.2,21 Measurements of proliferation are lower in the rectum than in the colon; nevertheless, they have been accepted, at least qualitatively, to reflect the epithelial proliferation throughout the colon.30,31 However, in many other studies the samples were taken during endoscopy, whereas in our study they were taken during unprepared digital rectal examination and most likely taken closer to the anus than in the other reports. Even so, a study of a U. S. popula- tion, using biopsies taken from the unprepared rectum and without using hyperbaric oxygen, demonstrated higher levels of REP than found in the current study.32 These different results may very well reflect differences in diet between the U. S. population and ours, namely that our population consumes fewer calories from animal fat, less beef and alcohol, and more calcium.6,18,33 Because the REP was measured a significant amount of time (.3 years) after the initial adenoma detection and removal, the patients may have become more health conscious and changed their diet and other lifestyle habits. We took this into account in the analysis of their diet and other lifestyle habits by our evaluation of both their past and recent lifestyle habits.18 However, we cannot exclude the possibility that the REP changed since adenoma diagnosis and removal. Another criticism is that our adenoma free screenees were not examined routinely by colonoscopy. Their repeatedly negative sigmoidoscopic examinations and annual fecal occult blood tests strongly suggested that they were free of adenomas and that misclassification was very unlikely. Based on a multicenter report of screening in Israel and a previous study of asymptomatic screenees with a normal sigmoidoscopic examination followed by colonoscopy, we estimate that 2 of our 91 adenoma free screenees might have harbored undetected proximal adenomas.34,35 The main strength of this study is its detailed analysis of long term diet and lifestyle habits and the REP measurements in the same population. We did not find increased REP in the past adenoma cases compared with the adenoma free screenees or by using a case– control analysis. This allowed us to evaluate the association between diet and lifestyle habits with the REP of all the examined population. In addition, we deliberately excluded from the study adenoma free screenees with a family history of colorectal Adenomas, Diet, and Epithelial Proliferation/Rozen et al. carcinoma to eliminate any additional risk influencing their REP measurements.2 After excluding the adenoma cases with a positive family history, the results were unchanged. History of Adenoma and REP Characteristics We did not find increased REP in the past adenoma cases compared with the adenoma free screenees. These results were not affected by the number, size, or histologic type of the adenomas. In fact, REP measurements in Compartment 5 showed the opposite: a decreased level in adenoma cases. Some other studies did find higher proliferation measurements in persons with hyperplastic polyps only.36 Our study was not designed to address the relation between having hyperplastic polyps only and REP measurements in flat rectal mucosa. We had only two adenoma free screenees recorded as having only these polyps and their REP measurements were not significantly different from the other screenees. There are descriptions of lower REP measurements in patients with small adenomas compared with those with large adenomas.30,36 –38 Like us, other authors did not find a significant difference, nor any association with numbers of adenomas or histologic type,5,39 Risio et al. found that the REP was lower in cases in whom the adenoma was removed $ 2 years previously.37 They did not address the possibility that it was due to a change in lifestyle. We could not evaluate this because the entry criteria for our study required a duration of .3 years since polypectomy and measurement of REP. We took no measurements at the time of polypectomy. The issue of whether increased proliferation is found in the normal-appearing large bowel mucosa of adenoma or colon carcinoma patients also has been addressed in recent publications.40 – 42 These authors did not detect significant differences in REP using direct microscopy or computer image analysis or in examining isolated crypts, and this is consistent with our own previous studies.3,29 In contrast, other authors did find such differences, but they were not always marked.2,5,36 –38,43,44 Many of these studies had small numbers of cases or controls and/or were not gender- or age-matched or controlled for other confounding variables such as diet and lifestyle habits. This topic also was recently reviewed by Jass et al.45 who reached a similar conclusion, namely that diffuse colonic epithelial hyperproliferation is not necessarily present in persons at risk for colorectal neoplasia. 1325 Diet, Lifestyle and REP Characteristics Our results are in concordance with the limited amount of literature showing that human diet and other lifestyle habits influence REP.4,5,15–17 In a previous study, we did not find differences in REP in healthy screenees of European versus nonEuropean origin, but did find urban-rural differences that might have been explained by differing dietary and other habits.29 In the current study the population was all urban. In North America, studies of SeventhDay Adventists showed that they had significantly less colonic epithelial proliferative activity than persons who previously had sporadic adenomas or familial colon carcinoma. This was attributed to their different dietary habits, lifestyle, and genetic risk.4,15 In a more recent study, also in Seventh-Day Adventists and other selected whites, a significantly negative association was found between their calcium intake and REP after controlling for age, gender, BMI, dietary fat, and energy intake.16 A significantly positive association was found between their BMI and REP and positive, but nonsignificant, associations were found for total energy intake and female gender. In a study of vegetarians in the U. K. the total LI of their rectal epithelial crypts was the same as that of adenoma free controls.5 A study of mitotic activity in the rectal epithelium of minimally symptomatic persons living in two distinct geographic areas of Italy showed that both their mitotic activity and dietary habits were significantly different with a higher intake of protein, fiber, starches, and nitrites found among those with the highest level of proliferation.17 However, multivariate logistic regression analysis did not confirm the occurrence of significant associations. In contrast to a defined population maintaining a specific diet and lifestyle, our study group’s habits were those of a free-living, asymptomatic population. Nevertheless, the number of cells per crypt was similar to that described in Seventh-Day Adventists,4 indicating that the rectal epithelium of our population was relatively “quiescent.”15 The most significant finding was the negative associations of REP with both calcium and fiber intake, indicating their influence in suppressing proliferative cells from reaching the crypt surface. This confirms the findings of Morgan and Singh16 and is in accord with some clinical observations and chemoprevention studies.2 Tobacco smoking also was significantly and positively associated with the degree of REP in the most superficial crypt compartment. This is consistent with epidemiologic studies showing that tobacco smoking is a risk factor for adenoma.18 The negative correlation between 1326 CANCER October 1, 1998 / Volume 83 / Number 7 weight gain and some REP measurements in the adenoma patients was unexplained. 10. REP, Diet, and Adenoma Etiology We could not show that REP measurements discriminated between past adenoma patients and adenoma free individuals. However, some demographic factors such as age and gender (but not family history of sporadic colorectal neoplasia) as well as diet and lifestyle habits were found to influence REP. Therefore, intervention studies in patients with a history of adenoma using REP as a biomarker should recognize the limitations of this biomarker in humans and take diet, lifestyle, and demographic variables into account as potential confounding factors.2,12–14 Many of these lifestyle associations that were found with REP are consistent with epidemiologic data related to adenoma etiology.18 This suggests that they may have influenced the normal-appearing rectal mucosa in the preneoplastic period. Therefore, subject to a further local influence, possibly genetic or genotoxic, these long term lifestyle habits could provide the milieu for the stepwise progression to future adenoma formation or growth. 11. 12. 13. 14. 15. 16. 17. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. Lipkin M. Biomarkers of increased susceptibility to gastrointestinal cancer: new application to studies of cancer prevention in human subjects. Cancer Res 1988;48:235– 45. Rozen P. An evaluation of rectal epithelial proliferation measurement as biomarker of risk for colorectal neoplasia and response in intervention studies. Eur J Cancer Prev 1992;1: 215–24. Rozen P, Fireman Z, Fine N, Chetrit A, Zajicek G. Rectal epithelial proliferation in persons post sporadic colorectal neoplasia. Cancer Lett 1991;59:1– 8. Richter A, Yang K, Richter F, Lynch HT, Lipkin M. Morphological and morphometric measurements in colorectal mucosa of subjects at increased risk for colonic neoplasia. Cancer Lett 1993;74:65– 8. Rooney PS, Clarke PA, Gifford KA, Hardcastle JD, Armitage NC. The identification of high and low risk groups for colorectal cancer using rectal mucosal crypt cell production rate (CCPR). Br J Cancer 1993;68:172–5. Rozen P, Liberman V, Lubin F, Angel S, Owen S, Trostler N, et al. A new dietary model to study colorectal carcinogenesis. Experimental design, food preparation and experimental findings. Nutr Cancer 1996;25:79 –100. Bartram H-P, Gostner A, Scheppach W, Reddy BS, Rao CV, Dusel G, et al. Effects of fish oil on rectal cell proliferation, mucosal fatty acids, and prostaglandin E2 release in healthy subjects. Gastroenterology 1993;105:1317–22. Steinbach G, Heymsfield S, Olansen NE, Tighe A, Holt PR. Effect of caloric restriction on colonic proliferation in obese persons: implications for colon cancer prevention. Cancer Res 1994:54:1194 –7. Anti M, Armelao F, Marra G, Percesepe A, Bartoli GM, Palozza P, et al. Effects of different doses of fish oil on rectal cell 18. 19. 20. 21. 22. 23. 24. 25. 26. proliferation in patients with sporadic colonic adenomas. Gastroenterology 1994;107:1709 –18. Gregoire R, Yeung KS, Stadler J, Stern HS, Kashtan H, Neil G, et al. Effect of high-fat and low-fiber meals on the cell proliferation activity of colorectal mucosa. Nutr Cancer 1991;15:21– 6. Lipkin M, Friedman E, Winawer SJ, Newmark H. Colonic epithelial cell proliferation in responders and nonresponders to supplemental dietary calcium. Cancer Res 1989; 49:248 –54. Lipkin M, Newmark H. Chemoprevention studies: controlling effects of initial nutrient levels. J Natl Cancer Inst 1993; 85:1870 –1. Schatskin A, Freedman LS, Dorgan J, McShane LM, Schiffman MH, Dawsey SM. Surrogate end points in cancer research: a critique. Cancer Epidemiol Biomarkers Prev 1996; 5:947–53. Einspahr JG, Alberts DS, Gapstur SM, Bostik RM, Emerson SS, Gerner EW. Surrogate end-point biomarkers as measures of colon cancer risk and their use in cancer chemoprevention trials. Cancer Epidemiol Biomarkers Prev 1997;6: 37– 48. Lipkin M, Uehara K, Winawer S, Sanchez A, Bauer C, Philips R, et al. Seventh-Day Adventist vegetarians have a quiescent proliferative activity in colonic mucosa. Cancer Lett 1985;26: 139 – 44. Morgan JW, Singh PN. Diet, body mass index and colonic epithelial cell proliferation in a healthy population. Nutr Cancer 1995;23:247–57. Caderni G, Bianchini F, Russo A, Spagnesi MT, Gabbrielli M, Ginanneschi U, et al. Mitotic activity in colorectal mucosa of healthy subjects in two Italian areas with different dietary habits. Nutr Cancer 1993;19:263– 8. Lubin F, Rozen P, Arieli B, Farbstein M, Knaani Y, Bat L, et al. Nutritional and lifestyle habits and water-fiber interaction in colorectal adenoma etiology. Cancer Epidemiol Biomarkers Prev 1997;6:79 – 85. Rozen P. Screening for colorectal neoplasia in the Tel Aviv area: cumulative data 1979 –1989 and initial conclusions. Isr J Med Sci 1992;28(Suppl):8 –20. Lubin F, Boeing H, Rozen P. Design and background of the Tel Aviv-Heidelberg dietary study of colonic adenoma patients and calcium intervention trial. In: Rozen P, Reich CH, Winawer SJ, editors. Large bowel cancer: policy, prevention, research and treatment. Basel: S. Karger, 1991:74 – 87. Darmon E, Pincu-Horstein A, Rozen P. A rapid and simple in vitro method for evaluating human colorectal epithelial proliferation. Arch Pathol Lab Med 1990;114:855–7. Fireman Z, Rozen P, Fine N, Chetrit A. Reproducibility studies and effect of bowel preparations on measurements of rectal epithelial proliferation. Cancer Lett 1989;45:59 – 64. Recommended Dietary Allowances. 10th edition. Washington, DC: National Academy Press, 1989:27. Lipkin M, Blattner WE, Fraumeni JF, Lynch HT, Deschner EE, Winawer S. Tritiated thymidine (Øp, Øh) labelling distribution as a marker for hereditary predisposition to colon cancer. Cancer Res 1983;43:1899 –904. Guggenheim YK, Kaufman NA, Reshef A. Composition of foods. 6th edition. Jerusalem: Magnes Press, The Hebrew University, 1991. Paul AA, Southgate DAT. McCance and Widowson’s the composition of foods. 4th edition. London: Elsevier, 1978. Adenomas, Diet, and Epithelial Proliferation/Rozen et al. 27. Breslow NE, Day NE. Statistical methods in cancer research. Volume 1. The analysis of case-control studies. Lyon, France: International Agency for Research on Cancer Scientific Publications, 1980. 28. Schlesselman JJ. Case control studies. Oxford: Oxford University Press, 1982. 29. Fireman Z, Rozen P, Fine N, Chetrit A. Influence of demographic parameters on rectal epithelial proliferation. Cancer Lett 1989;47:133– 40. 30. Terpstra OT, van Blankenstein M, Dees J, Eilers GAM. Abnormal pattern of cell proliferation in the entire colonic mucosa of patients with colon adenoma or cancer. Gastroenterology 1987;92:704 – 8. 31. Potten CS, Kellett M, Rew DA, Roberts SA. Proliferation in human gastrointestinal epithelium using bromodeoxyuridine in vivo: data for different sites, proximity to a tumour, and polyposis coli. Gut 1992;33:524 –9. 32. Lipkin M, Enker WE, Winawer SJ. Tritiated-thymidine labelling of rectal epithelial cells in “non-prep” biopsies of individuals at increased risk for colonic neoplasia. Cancer Lett 1987;37:153– 61. 33. Rozen P, Horwitz C, Tabenkin C, Ron E, Katz L. Dietary habits and colorectal cancer incidence in a second-defined kibbutz population. Nutr Cancer 1987;9:177– 84. 34. Odes HS, Rozen P, Ron E, Bass D, Bat L, Keren S, et al. Screening for colorectal neoplasia: a multicenter study in Israel. Isr J Med Sci 1992;28(Suppl):21– 8. 35. Rozen P, Ron E. A cost analysis of screening methodology for family members of colorectal cancer patients. Am J Gastroenterol 1989;84:1548 –51. 36. Risio M, Arrigoni A, Pennazio M, Agostinucci A, Spandre M, Rossini FP. Mucosal cell proliferation in patients with hyperplastic colorectal polyps. Scand J Gastroenterol 1995;30: 344 – 8. 37. Risio M, Lipkin M, Candelaresi GL, Bertone A, Coverlizza S, Rossini FP. Correlations between rectal mucosa cell proliferation and the clinical and pathological features of nonfa- 38. 39. 40. 41. 42. 43. 44. 45. 1327 milial neoplasia of the large intestine. Cancer Res 1991;51: 1917–21. Paganelli GM, Biasco G, Santucci R, Brandi G, Lalli AA, Miglioli M, et al. Rectal cell proliferation and colorectal cancer risk level in patients with nonfamilial adenomatous polyps of the large bowel. Cancer 1991;68:2451– 4. Roncucci L, Scalmati A, Ponz de Leon M. Pattern of cell kinetics in colorectal mucosa of patients with different types of adenomatous polyps of the large bowel. Cancer 1991;68: 873– 8. Nakamura S, Goto J, Kitayama Y, Sheffield JP, Talbot IC. Flow cytometric analysis of DNA synthetic phase fraction of the normal appearing colonic mucosa in patients with colorectal neoplasms. Gut 1995;37:398 – 401. Wong AJ, Kohn GJ, Schwartz HJ, Ruebner BH, Lawson MJ. Colorectal cancer and non-cancer patients have similar labelling indices by microscopy and computed image analysis. Hum Pathol 1995;26:1329 –32. Hardman WE, Cameron IL, Beer WH, Speeg KV, Kadakia SC, Lang KA. Negative correlations between the size of the differentiated and proliferative cell compartments in rectal mucosal crypts of humans with various types of colonic tumors [abstract]. Cancer Res 1996;37:268. Anti M, Marra G, Armelao F, Percesepe A, Ficarelli R, Ricciuto GM, et al. Rectal epithelial cell proliferation patterns as predictors of adenomatous colorectal polyp recurrence. Gut 1993;34:525–30. Ikeda N, Mizuno M, Okada H, Tomoda J, Tsuji T. Immunohistochemical analysis of epithelial cell proliferation in normal-appearing rectal mucosa of patients with colorectal adenoma and cancer using an in vitro labeling method with bromodeoxyuridine. Acta Med Okayama 1994;48:243–7. Jass JR, Ajioka Y, Radojkovic M, Allison LJ, Lane M. Failure to detect colonic mucosal hyperproliferation in mutation positive members of a family with hereditary non-polyposis colorectal cancer. Histopathology 1997;30:201–7.