Disparities in risk of and survival from oropharyngeal squamous cell carcinoma

Disparities in risk of and survival from oropharyngeal squamous cell carcinoma Craig S. Miller, DMD, MS,a Robert G. Henry, DMD, MPH,b and Mary Kay Rayens, PhD,c Lexington, Ky UNIVERSITY OF KENTUCKY COLLEGE OF DENTISTRY, COLLEGE OF MEDICINE, AND COLLEGE OF NURSING

Background. Oropharyngeal cancer has a low 5-year survival rate in the United States, indicating that much remains to be learned about risk and survival factors. The purpose of this research was to identify factors that contribute to risk of and survival from this disease in a region associated with high tobacco use. Methods. A retrospective chart review of 221 subjects with diagnosis of head and neck cancer at a regional University Medical Center during a 10-year period (1986 to 1995) was performed to identify relationships between risk and survival factors of oropharyngeal squamous cell carcinoma (OSCC). Data from 92 records met the inclusion criteria and were analyzed with the ␹2 test of association, the Fisher exact test, and the log-rank test for comparison of survival distributions. The software used was SAS for Windows, release 8.2. Results. Most study subjects were white men between the ages of 50 and 75 years who chronically used tobacco and alcohol. The overall median survival period of the group was 18.1 months. The percentage of subjects surviving to 3 years was 35.6%. Disparities were identified in risk and survival by sex, race, age, dwelling, anatomic site, and exposure to carcinogenic substances. In particular, the disease more adversely affected minorities and those who chronically used tobacco or alcohol. Conclusion. Our results show that, in the examined regional population, a high prevalence of tobacco use correlated with alcohol use and a lower rate of survival from OSCC than the national rate. Also, risk of and survival from OSCC are affected by demographic, geographic, and behavioral factors. This information is useful for the design and implementation of more effective preventive and intervention strategies. (Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2003;95:570-5)

Oral and pharyngeal cancer affects 40,000 people per year and has a 5-year survival rate of 52% in the United States.1-3 Chronic tobacco use and alcohol consumption are major risk factors for the most common type of oropharyngeal cancer, oral squamous cell carcinoma (OSCC).1,4,5 Advancing age, immunodeficiency, infection with human papillomavirus types 16 and 18, and chronic exposure to sunlight, particularly ultraviolet light, also increase the risk for OSCC development.1,6,7 The mortality rate from OSCC has not changed sigData presented at the American Association of Dental Research Meeting, Chicago, Ill, March 2001. a Professor of Oral Medicine, Department of Oral Health Practice and Department of Microbiology, Immunology, and Molecular Genetics, University of Kentucky College of Dentistry and College of Medicine, Lexington. b Assistant Chief Dental Service, Veterans Administration Medical Center, and Clinical Associate Professor, Department of Oral Health Science, University of Kentucky College of Dentistry, Lexington. c Associate Director, Biostatistics Consulting Unit, University of Kentucky College of Medicine, and Research Associate Professor, University of Kentucky Colleges of Nursing and Medicine, Lexington. Received for publication May 30, 2002; returned for revision Sep 23, 2002; accepted for publication Nov 11, 2002. © 2003, Mosby, Inc. All rights reserved. 1079-2104/2003/$30.00 ⫹ 0 doi:10.1067/moe.2003.108

570

nificantly during the past 30 years. This is attributed largely to the fact that most oral cancers are not diagnosed until they are in the advanced stage.3 Improved survival will likely result from scientific advances and a better understanding of risk factors and behaviors associated with the disease. Two factors that reduce the risk and enhance survival are cessation in the use of tobacco and alcohol. Early diagnosis and early entry into therapy are also crucial for improved survival.8 Other factors that are less well defined include immunologic mechanisms and susceptibility of the tumor to therapy. Previous studies have shown that significant differences in race and sex exist in the incidence rates of oropharyngeal cancer.3,9 These variations have been attributed mainly to specific risk factors, such as alcohol and tobacco use. However, dietary factors, genetic predispositions, education, cultural norms, and access to care may play a role. The release of “Oral health in America: A report of the surgeon general”10 indicates that for improved survival to be realized for all oral cancer population groups, identification of the factors associated with risk and survival is important. The purpose of this research was to examine determinants of risk and survival in a population of patients with OSCC in a state with a high prevalence of tobacco use.

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 95, Number 5

MATERIAL AND METHODS A retrospective review was performed of the medical or dental charts of 221 consecutive subjects with diagnosis of head and neck cancer during a 10-year period (1986 to 1995) as listed by the University of Kentucky Hospital Tumor Registry. The Registry contained information derived from pathology and autopsy reports and used the International Classification of Diseases, 9th Revision (ICD-9) diagnostic codes for patients with cancer who underwent treatment at the University of Kentucky Medical Center and Veterans Administration Medical Center in Lexington, Kentucky. The investigators recorded the following information from the registry and medical and dental records: age at onset, sex, race, tobacco and alcohol use, date of diagnosis and survival status, site, histologic type, treatment rendered, survival status, medical diagnoses, and duration and amount of aspirin use. Tobacco use was quantified as the number of packs of cigarettes smoked per day times the number of years involved. Chronic use was defined as at least 10 pack-years. Alcohol use was quantified as whiskey equivalents per day times the number of years involved to obtain whiskey equivalent years (WEY). Chronic use was defined as at least 10 years of use. Only records having the ICD-9 code 8070 (OSCC) or 8071 (keratinizing OSCC) at intraoral sites that included the oropharynx (ie, tonsil) were included for study. Excluded were subjects whose records lacked sufficient detail and those whose oral cancer was of nonsquamous cell epithelium origin or arose from the nasopharynx, larynx, or salivary glands. Subjects, their relatives, or significant others were contacted by telephone when necessary to complete data collection and to confirm survival status. Data analyses were performed with SAS for Windows, version 8.2 (SAS Institute, Inc, Cary, NC). The Institutional Review Board of the University of Kentucky approved the study. RESULTS Characteristics of study group One hundred twenty-nine records identified with the ICD-9 code for OSCC had sufficient detail for inclusion in the study. Of these, 37 were omitted for cancer sites outside the oral cavity, namely cancers of the nasopharynx or larynx. The remaining 92 patients formed the basis for this analysis. The study group consisted of 20.0% women and 80.0% men. The ethnicity of the group was 80.6% white, 18.1% African American, and 1.3% Hispanic/Asian. The mean age at the time of diagnosis was 65.5 years (SD, 10.6 years; range, 41 to 87 years). Chronic tobacco use was documented in 92.2% of the sample. Data collection procedures garnered sufficient information to calculate pack-years for 79% of the smokers. These 56 smokers had an average

Miller, Henry, and Rayens 571

history of 53 pack-years and a range in pack-years of 7 to 180. The comparison between heavy and light smokers was accomplished by dividing the smokers with pack-year information into 2 groups with the median of 45 as the split point. Only 1 patient had a documented history of chewing-tobacco use. Chronic alcohol use was noted in 81.2% of patients. Details on alcohol use patterns were less complete. Of the 17 subjects (30% of all alcohol users in the sample) for whom WEY could be assessed, the average WEY value was 65, with a range of 4 to 280. These patients with complete WEY information were divided into high and low alcohol users according to the median WEY value of 40. The most common site for OSCC was the tongue (27.2%), followed by the oropharynx (22.7%), floor of the mouth (19.6%), lip (13.0%), gingiva/hard palate (7.6%), buccal/labial mucosa (4.4%), posterior mandible (3.3%), and anterior mandible (2.2%). Most patients underwent surgical excision (73.9%) or radiation therapy (50.0%). Chemotherapy was provided to 10.9%, and 4.4% did not undergo treatment. Risk factors and disease characteristics Bivariate tests of association were used to identify the risk factors significantly related to both disease characteristics and survival. Relationships identified are presented categorically in the Table. Interestingly, of the risk factors identified, aspirin use was documented infrequently (17.4%), with only 3.3% of subjects using aspirin on a regular basis. Survival status and treatment Survival status was ascertained for 89 of the 92 subjects. At the end of the study period, 33.7% of the subjects were identified as alive. Of the 59 deceased patients (66.3%), death from cancer or complications of cancer occurred in 42 subjects. The median survival of the group was 18.1 months (Figure). The percentage of subjects surviving to 3 years was 35.6% (95% CI for 3-year survival rate, 22.7% to 48.5%). The 5-year survival estimate was 29.6% (95% CI, 14.5% to 44.7%). The comparison of whites with minorities on survival distribution was significant (Table), with white patients surviving longer (Figure). The median survival for minorities was 9.1 months, and the median survival for white patients was not estimable because relatively few deaths occurred in this racial group. The median survival for tobacco users was 18.1 months, and the median survival for nonusers was not estimable because nearly all nonusers were alive at the end of the study. The median survival among subjects with OSCC at nonlip sites was 16.1 months. In contrast, survival of subjects with lip cancer was significantly longer, with the median survival not estimable because of no deaths

572 Miller, Henry, and Rayens

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY May 2003

Table. Significant bivariate relationships among study variables (n ⫽ 92) Category Age

Race Behavior

Anatomic site

Stage

Survival

Treatment

* †

Finding Subjects 70 years and older were less likely to have used tobacco and alcohol compared with those diagnosed at earlier age Significant association between age and sex, with all of those diagnosed at 85 years or more being female and all those diagnosed before age 45 male Higher proportion of women than men were 75 years or older at diagnosis Compared with minorities, higher proportion of whites were rural dwellers Men were more likely than women to use tobacco and alcohol Men were also more likely than women to be heavy smokers Urban dwellers were more likely to use alcohol compared with rural dwellers Urban dwellers were more likely to use tobacco compared with rural counterparts Compared with rural dwellers, urban dwellers were more likely to use both alcohol and tobacco Tobacco users were more likely than not to use alcohol Heavy smokers were more likely than light smokers to use alcohol Lip cancer was more likely to occur among those 50 years old or more compared with rate among those younger than 50 years Lip cancer was associated more often with stage 1 disease than with later stages Decreased likelihood of cancers of pharynx with increasing age Compared with those subjects less than 75 years old, patients 75 years or older were more likely to have cancers of gingiva and hard palate Subjects with anterior site intraoral cancer were more likely to have later stage (3-4) at diagnosis, compared with subjects with other intraoral cancer sites Distribution of stage was not uniform, with most of these cancers diagnosed at stage 4 Compared with minority subjects, whites were more likely to have stages 1 to 2 at diagnosis Higher proportion of rural dwellers had stage 1 to 2, compared with proportion of urban dwellers with these lower stages 25th percentile of survival distribution for whites was 10.7 months compared with 6.2 months for minorities Comparison of survival distributions between tobacco users and nonusers was significant, with users having shorter survival Comparison of survival between patients with lip cancer and rest of subjects was significant; lip cancer patients lived significantly longer Compared with subjects who did not use tobacco, a smaller percentage of tobacco users survived to 5 years Alcohol users were significantly less likely than nonusers to survive 5 years Subjects who used tobacco or alcohol were less likely than those who used neither substance to survive 5 years after diagnosis Heavy smokers were less likely than light smokers to survive 3 years Compared with urban dwellers, higher percentage of rural dwellers lived at least 5 years after diagnosis Decreased likelihood of surgical treatment with increasing stage Increased likelihood of radiation treatment with stages 3 to 4, compared with stages 1 to 2 Compared with light smokers, heavy smokers were more likely to undergo chemotherapy

␹2 P value .02* .01* .007* .003* .001* .001* .02 .04 .03 .0001* .03* .05* .005* .05* .02* .04* .0001 .05* .01 .002† .05† .05† .03* .003* .03* .006* .04* .006* .007 .006*

Reported P value is for Fisher exact test, appropriate alternative to ␹2 test given small expected cell counts. reported P value is for log-rank test.

in this group by the end of the study. Several anatomic sites (eg, floor of mouth, tongue, and oropharynx) were associated with relatively poor survival, but these comparisons did not attain statistical significance at the .05 level. Age and type of treatment were not factors identified in the survival from oral cancer. Lack of information on aspirin use in most patients prevented the definitive exploration of whether aspirin use was related to survival or disease characteristics. Treatment type was associated with stage and tobacco use. DISCUSSION This retrospective study examined risk and survival factors of OSCC in a small group of patients diagnosed

with oral cancer in a state with a high prevalence of tobacco use. Data from 3 sources (tumor registry, medical or dental record, and select interviews of subjects, their relatives, or significant others) were analyzed. Interviews were performed only when needed to complete information specific to the data collection worksheet. Our analyses showed the following: 1. Most OSCCs were late stage (53% were stage 4), and survival of this cohort was less than the national rate. 2. Tobacco or alcohol use was implicated in most OSCCs, and greater and combined use of these carcinogenic substances was significantly related with shorter survival.

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 95, Number 5

Miller, Henry, and Rayens 573

Figure. Kaplan-Meier survival curve of patients with diagnosis of OSCC.

3. Women typically were diagnosed at later ages than men and were less likely to use tobacco or alcohol. 4. Minority subjects were disproportionately affected, more likely to live in urban areas, and more likely to have later stage disease at diagnosis and shorter overall survival. 5. Urban dwellers were more likely than their rural counterparts to use tobacco or alcohol and to have later stage disease. 6. Compared with cancer at other sites, lip cancer was more likely to be diagnosed at an earlier stage and the survival of these patients was significantly longer. 7. Pharyngeal OSCC was more likely to occur in younger patients, and gingival and hard palate OSCC was more likely to occur in older subjects. 8. Regular aspirin use was documented infrequently in this cohort. The data are demographically similar to those reported from the Oral Medicine Clinic at the University of California at San Francisco (mean, 59 years)1 and are consistent with the findings of the Surveillance, Epidemiology and End Results Program data covering similar years from 1973 to 1995.2 The site predilection for OSCC in patients of Kentucky was similar to those sites reported elsewhere.11,12 Despite a preponderance of cases occurring in white men, the racial distribution of cases reflected a higher prevalence in minorities than

the population of the Commonwealth of Kentucky would predict.13 These data are consistent with those of other studies that have shown a higher rate of OSCC in African Americans than in other racial groups.1-3,14 Possible reasons for this discrepancy may involve educational and socioeconomic issues or the fact that African Americans live more often in urban areas13 and urban dwellers are greater users of both alcohol and tobacco than rural dwellers. The finding that chronic smoking tobacco and alcohol use are the primary risk factors for OSCC has been well documented.11,15-18 Interestingly, only 1% of these OSCC cases had a history of chewing tobacco. This may seem paradoxical in a commonwealth documented to have a high prevalence of tobacco chewers.19,20 However, this finding may reflect that tobacco chewing is a habit associated with a younger age group or that the historic information obtained was incomplete. On the other hand, this may provide evidence that tobacco chewing is not a high risk factor for the development of OSCC as purported by Vigneswaran et al21 and Bouquot and Meckstroth.22 Careful prospective studies are needed to fully address this question. Our findings suggest a strong association between OSCC and the use of tobacco or alcohol, particularly among men aged 50 to 75 years. In contrast, older women in whom OSCC developed had limited tobacco and alcohol use. This finding, although previously re-

574 Miller, Henry, and Rayens

ported,23 has not been well recognized or studied. It could suggest that tobacco and alcohol perform their deleterious effects often by age 75 years, whereas after that age, OSCC develops for different reasons. Risk factors at this later age may be associated with human papillomavirus infection, change in nutritional status, declining immune surveillance, or other factors not assessed in this work. The data on anatomic site suggest that readily visible lesions (lip cancer) were diagnosed at an earlier stage than less visible cancers. Also, patients with less keratinized sites (lateral tongue, floor of mouth) were affected at earlier ages than patients with well-keratinized sites (ie, gingiva and hard palate). The latter finding is in contrast to that reported by Silverman1 but supports the hypothesis that carcinogens from tobacco and alcohol have more frequent contact with or penetrate nonkeratinized epithelium more readily than keratinized epithelium. The possibility that keratinization may limit cell susceptibility to the synergistic effects of carcinogens should be investigated in well-designed, prospective studies. Survival status of OSCC in Kentucky remains lower than the national rate.3 The percentage of subjects in this study surviving to 5 years was 29.6%. This is lower than the nationally reported rate of 52.8% and much lower than the predicted 55% in white patients who comprised the majority of patients studied.1,3 The 10year survival rate was not estimable because none of the patients had survived that long by the conclusion of the study, but the surviving percentage at 9.5 years was 14.8%, which is considerably lower than national estimates of the 10-year survival rate, which range from 39% to 45%.3,24 Survival was significantly associated with tobacco and alcohol use, with users living significantly less time after diagnosis than nonusers. The association of tobacco smoking and decreased survival in OSCC is no surprise. The Centers for Disease Control reports that 88% of deaths associated with lip, oral cavity, and the pharynx neoplasms are attributable to smoking.25 In this study, tobacco smokers were more likely to be chronic alcohol users (81.2% of the cohort used alcohol), and more persons who used alcohol had later stage disease than nonusers (P ⫽ .07). Minorities and urban dwellers also had poor survival and higher incidence rates of alcohol use. It has been reported that users of alcohol do not perceive alcohol consumption as a factor that increases their risk of OSCC.26 Although not directly assessed in this study, it is possible that education, socioeconomic status, and health care service use influenced the relationships found in this small cohort. Low educational level has been shown to have a positive association with oral cancer.17

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY May 2003

The role of aspirin in the development and progression of OSCC has not been defined. Evidence suggests that inhibitors of cyclooxygenase 2 may be useful in prevention of certain types of proliferative growths and cancer.27-31 Moreover, aspirin reduces the risk of colorectal32-33 and esophageal cancer.36,37 Although aspirin’s relationship with risk of and survival from OSCC was not clearly defined in this study, there was some indication of a beneficial effect. Analysis of the data with the 25th percentile of the survival distribution showed that regular aspirin users lived on average 8.8 months longer than patients who were not regular aspirin users (data not shown). Also, regular aspirin use was documented in only 3.3% of patients with OSCC. This could suggest that an inverse relationship exists between aspirin use and the risk of development of OSCC. The prevalence of aspirin use among adults aged 45 years and older in Wisconsin and Michigan (n ⫽ 1685; 548 in Wisconsin and 1137 in Michigan) has been reported to be 23.3%.38 Although aspirin users were underrepresented in the studied population, the beneficial effect of aspirin against gastrointestinal cancer appears to require regular usage.34 It is interesting to speculate that aspirin use may have a protective role against OSCC as in other gastrointestinal cancers. However, additional research (ie, prospective clinical trials) is necessary to more thoroughly examine the effect of aspirin on OSCC. In summary, a region with high tobacco smoking prevalence was associated with a population of patients with OSCC who had shorter survival rates than the national rate. Disparities were identified in risk and survival by race, age, sex, dwelling, anatomic site, and behavior. Specifically, OSCC more adversely affected minorities and patients who used tobacco and alcohol in combination. These findings along with other OSCC studies should be useful in the design of new preventive and intervention strategies that involve public and professional education and awareness, development of clinical guidelines for early diagnosis and screening, and strategic planning with public health service organizations. REFERENCES 1. Silverman S Jr. Oral cancer. 4th ed. Hamilton, Ontario: American Cancer Society, B.C. Decker; 1998. 2. Ries LAG, Kosary CL, Hankey BR, Miller BA, Edwards BK, eds. Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute. Cancer Statistics Review, 1973-1995. Bethesda (MD): National Cancer Institute; 1998. 3. SEER Cancer Statistics Review. Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute. Cancer Statistics Review, 1973-1999. Bethesda (MD): National Cancer Institute. Available from: URL:http://seer.cancer.gov/csr/1973_1999/sections.html#sections 4. Blot WJ, McLaughlin JK, Winn DM, Austin DF, Greenberg RS, Preston-Martin S, et al. Smoking and drinking in relation to oral and pharyngeal cancer. Cancer Res 1988;48:3282-7.

Miller, Henry, and Rayens 575

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 95, Number 5 5. Mashberg A, Garfinkel I, Harris S. Alcohol as a primary risk factor in oral squamous cell carcinoma. CA Cancer J Clin 1981; 31:146-55. 6. Maden C, Beckmann A, Thomas DB, McKnight B, Sherman KJ, Ashley RL, et al. Human papillomaviruses, herpes simplex viruses, and the risk of oral cancer in men. Am J Epidemiol 1992;135:1093-102. 7. Miller CS, White DK. Human papillomavirus expression in oral mucosa, oral premalignant conditions and squamous cell carcinoma. A retrospective review of the literature. Oral Surg Oral Med Oral Pathol 1996;82:57-68. 8. Silverman S. Early diagnosis of oral cancer. Cancer 1988;62: 1796-9. 9. Scully C, Bedi R. Ethnicity and oral cancer. Lancet Oncol 2000;1:37-42. 10. Satcher D. Surgeon General’s report on oral health. Public Health Rep 2000;115:115. 11. Jovanovic A, Schulten E, Kostense PJ, Snow GB, van der Waal I. Tobacco and alcohol related to the anatomical site of oral squamous cell carcinoma. J Oral Pathol Med 1993;22:459-62. 12. Mashberg A, Meyers H. Anatomical site and size of 222 early asymptomatic oral squamous cell carcinomas: a continuing prospective study of oral cancer. II. Cancer 1976;37:2149-57. 13. US Census Bureau. State and county quickfacts, latest revision, Feb 7, 2002. Accessed May 27, 2002. Available from: URL: http://www.census.gov/ 14. Day GL, Blot WJ, Austin DR, Bernstein L, Greenberg RS, Preston-Martin S, et al. Racial differences in risk of oral and pharyngeal cancer: alcohol, tobacco and other determinants. J Natl Cancer Inst 1993;85:465-73. 15. Mashberg A, Boffetta P, Winkelman R, Garfinkel L. Tobacco smoking, alcohol drinking, and cancer of the oral cavity and oropharynx among U.S. veterans. Cancer 1993;72:1369-75. 16. Boffetta P, Mashberg A, Winkelmann R, Garfinkel L. Carcinogenic effect of tobacco smoking and alcohol drinking on anatomic sites of the oral cavity and oropharynx. Int J Cancer 1992;52:530-3. 17. Merletti F, Boffetta P, Ciccone G, Mashberg A, Terracini B. Role of tobacco and alcoholic beverages in the etiology of cancer of the oral cavity/oropharynx in Torino, Italy. Cancer Res 1989;49: 4919-24. 18. Silverman S Jr. Demographics and occurrence of oral and pharyngeal cancers: the outcomes, the trends, the challenge. J Am Dent Assoc 2001;132:7S-11S. 19. Marcus AC, Crane LA, Shopland DR, Lynn WR. Use of smokeless tobacco in the United States: recent estimates from the current population survey. NCI Monogr 1989;8:17-23. 20. Noland MP, Kryscio RJ, Riggs RS, Linville LH, Perritt LJ, Tucker TC. Use of snuff, chewing tobacco, and cigarettes among adolescents in a tobacco-producing area. Addict Behav 1990;15: 517-30. 21. Vigneswaran N, Tilashalski K, Rodu B, Cole P. Tobacco use and cancer. A reappraisal. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 1995;80:178-82. 22. Bouquot JE, Meckstroth RL. Oral cancer in a tobacco-chewing US population: no apparent increased incidence or mortality. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 1998;86: 697-706. 23. Zavras AI, Douglass CW, Joshipura K, Wu T, Laskaris G, Petridou E, et al. Smoking and alcohol in the etiology of oral

24.

25. 26.

27.

28.

29.

30. 31. 32.

33. 34. 35.

36.

37. 38.

cancer: gender-specific risk profiles in the south of Greece. Oral Oncol 2001;37:28-35. Perez CA, Patel MM, Chao KS, Simpson JR, Session D, Spector GJ, et al. Carcinoma of the tonsillar fossa: prognostic factors and long-term therapy outcome. Int J Radiat Oncol Biol Phys 1998; 42:1077-84. Smoking-attributable mortality and years of potential life lost: United States, 1984. MMWR Morb Mortal Wkly Rep 1997;46: 444-51. Hay JL, Ostroff JS, Gustavo D, LeGeros RZ, Kenigsberg H, Franklin DM. Oral cancer risk perception among participants in an oral cancer screening program. Cancer Epidemiol Biomarkers Prev 2002;11:155-8. Eberhart CE, Coffey RJ, Radhika A, Giardiello FM, Ferrenbach S, DuBois RN. Up-regulation of cyclooxygenase 2 gene expression in human colorectal adenomas and adenocarcinomas. Gastroenterology 1994;107:11183-8. Sheehan KM, Sheahan K, O’Donoghue DP, MacSweeney F, Conroy RM, Fitzgeral DJ, et al. The relationship between cyclooxygenase-2 expression and colorectal cancer. J Am Med Assoc 1999;282:1254-7. Oshima M, Dinchuk JE, Karman SL, Oshima H, Hancock B, Kwong E, et al. Suppression of intestinal polyposis in Apc delta716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell 1996;87:803-9. Kawamori T, Rro CV, Seibert K, Reddy BS. Chemopreventive activity of celecoxib, a specific cyclooxygenase-2 inhibitor, against colon carcinogenesis. Cancer Res 1998;58:409-12. Morgan G. Deleterious effects of prostaglandin E2 in oesophageal carcinogenesis. Med Hypotheses 1997;48:177-81. Kune GA, Kune S, Watson LF. Colorectal cancer risk, chronic illnesses, operations, and medications: case control results from the Melbourne Colorectal Cancer Study. Cancer Res 1988;48: 4399-40. Marnett LJ. Aspirin and the potential role of prostaglandins in colon cancer. Cancer Res 1992;52:5575-89. Thun MJ, Callee EE, Namboodiri MM, Flanders WD, Coates RJ, Byers T, et al. Risk factors for fatal colon cancer in a large prospective. J Natl Cancer Inst 1992;84:1491-500. Heath CW Jr, Thun MJ, Greenberg ER, Levin B, Marnett LJ. Nonsteroidal antiinflammatory drugs and human cancer. Report of an interdisciplinary research workshop. Cancer 1994;74: 2885-8. Farrow DC, Vaughan TL, Hansten PD, Stanford JL, Risch HA, Gammon MD, et al. Use of aspirin and other nonsteroidal antiinflammatory drugs and risk of esophageal and gastric cancer. Cancer Epidemiol Biomarkers Prev 1998;7:97-102. Funkhouser EM, Sharp GB. Aspirin and reduced risk of esophageal carcinoma. Cancer 1995;76:1116-9. Prevalence of aspirin use to prevent heart disease: Wisconsin, 1991, and Michigan, 1994. MMWR Morb Mortal Wkly Rep 1997;46:498-502.

Reprint requests: Dr Craig S. Miller Oral Medicine Section MN 118 Department of Oral Health Practice University of Kentucky College of Dentistry 800 Rose Street Lexington, KY 40536-0297