Increasing experimental evidence suggests that the renin angiotensin

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2007;5: Angiotensin-Converting Enzyme Inhibitors and Risk of Esophageal and Gastric Cancer: A Nested Case-Control Study TOMAS SJÖBERG,* LUIS A. GARCÍA RODRÍGUEZ, and MATS LINDBLAD* *Unit of Esophageal and Gastric Research, Karolinska Institutet, Stockholm, Sweden; and Centro Español de Investigación Farmacoepidemiologica, Madrid, Spain Background & Aims: There are contradictory results regarding the association between angiotensin-converting enzyme (ACE) inhibitors and cancer. We aimed to investigate whether ACE inhibitors protect against esophageal and gastric cancer. Methods: We conducted a population-based case-control study nested within the General Practitioners Research Database in the United Kingdom. All individuals in the General Practitioners Research Database aged years between 1994 and 2001 were followed up until detection of an esophageal or gastric cancer (cases), other cancer, age of 85 years, death, or end of study period. Results: Among 4.34 million person-years, 909 cases of esophageal and 1023 cases of gastric cancer were identified, and 10,000 matched controls were selected at random. Adjustments included smoking, body mass index, concurrent medication, and gastrointestinal disorders. Current use of ACE inhibitors decreased the risk of esophageal by 29% (odds ratio [OR], 0.71; 95% confidence interval [CI], ), but not of squamous-cell carcinoma (OR, 1.27; 95% CI, ) compared with nonusers. A high daily dose of ACE inhibitors decreased the risk of both and squamous-cell cancer of the esophagus and rendered a 45% decrease of total esophageal cancer (OR, 0.55; 95% CI, ). Our data showed no clear association between the use of ACE inhibitors and risk of gastric cancer (OR, 1.07; 95% CI, ). Conclusions: The use of ACE inhibitors may decrease the risk of developing esophageal cancer, particularly among users with a high daily dose. No association was found between gastric cancer and ACE inhibitors. Increasing experimental evidence suggests that the renin angiotensin system is involved in tumor growth, angiogenesis, and metastasis. The key enzyme of the renin angiotensin system, the angiotensin-converting enzyme (ACE), produces the renin angiotensin system s active end product: angiotensin II (Ang II). Experimentally, Ang II is a growth factor that stimulates cell replication and tumor growth via paracrine actions: some tumors express renin or Ang II receptors. As a consequence, an antineoplastic potential of ACE inhibitors has been suggested. 1,2 ACE inhibitors inhibit expression of vascular endothelial growth factor, a well-known promoter of angiogenesis, and a novel drug target for cancer treatment. 1,3 5 Clinically, ACE inhibitors are used most commonly for treating hypertension and cardiac heart failure. Results from clinical randomized trials and observational studies have been conflicting on the association between use of ACE inhibitors and cancer; there are some findings in favor of an anticarcinogenic effect, 6 and others against. 7,8 Recently, a case-control study presented at Digestive Disease Week 2006 pointed out a 55% reduction in the incidence of esophageal cancer, a 48% reduced risk of pancreatic cancer, and a 47% reduced risk of colon cancer among users of ACE inhibitors (Presented at Digestive Disease Week, 2006 in Los Angeles, CA). Globally, esophageal and gastric cancers are common tumors, both carrying a bad prognosis, being the sixth and second leading causes of cancer mortality, respectively. 9 Therefore, there is an urgent need to investigate and potentially unveil new preventive factors, and new treatment modalities of these tumors. We aimed at testing the hypothesis of a protective effect of ACE inhibitors on the development of esophageal and gastric cancer by conducting a large case-control study within the comprehensive General Practitioners Research Database (GPRD). Material and Methods General Practice Research Database The study used data generated from the GPRD in the United Kingdom, a database that has been described in detail elsewhere. 10 In brief, the GPRD includes prospective recordings of diagnoses, prescriptions of drugs, demographics, details of every general practitioner s (GP s) consultation, results from laboratory tests, hospital letters, and a free text section. All prescriptions are entered automatically into the GPRD because they are generated directly from the participating GP s computer, thus ensuring a complete recording. Validation studies have shown that the GPRD data are of high quality. 11 More than 90% of all referrals are entered into the GP s computer with the specialist s diagnosis correctly coded. 11 Design Our design has been described in detail previously. 12 In short, we identified all individuals registered in the GPRD, aged years during the period from January 1, 1994, to December 31, To be included in the study cohort patients had to be enrolled with a GP for at least 2 years, and have had at least 1 year of prescription history recorded in the GPRD. A history of cancer (other than nonmelanoma skin cancer) recorded in the database rendered exclusion. All subjects in the Abbreviations used in this paper: ACE, angiotensin-converting enzyme; Ang II, angiotensin II; -blockers, -receptor inhibitors; BMI, body mass index; Ca-blockers, calcium channel blockers; CI, confidence interval; GP, general practitioner; GPRD, General Practitioners Research Database; OR, odds ratio by the AGA Institute /07/$32.00 doi: /j.cgh

2 October 2007 ACE INHIBITORS AND RISK OF ESOPHAGEAL AND GASTRIC CANCER 1161 cohort were followed up and considered at risk until one of the following events occurred, whichever came first: (1) detection of an esophageal or gastric cancer (selected as cases), (2) detection of any other cancer (except nonmelanoma skin cancer), (3) age of 85 years, (4) death, or (5) end of study period (December 31, 2001). The Scientific and Ethical Advisory Group in the United Kingdom approved the study. Identification of Case Patients The follow-up of the study cohort rendered 2128 patients with a new diagnosis code indicating esophageal or gastric cancer. We obtained information on the site and histology to verify and further classify the tumors. We also requested additional paper-based information (eg, surgical and pathology reports), and correspondence from specialists from the GPs for a consecutive sample of 1280 patients. The reviewer of these records and additional information (M.L.) was blinded to the exposure data. Patients were excluded if any of the following occurred: (1) the tumor was benign, (2) the origin of the cancer was unknown, (3) the tumor was a metastasis, (4) the patient had another concurrent cancer, (5) the cancer was first diagnosed before start date, or (6) the histologic type was not or squamous-cell carcinoma. The effect of the review of the additional information provided by the GPs was negligible with respect to the overall final case status compared with the information already present in the GPRD. Therefore, we did not request complementary information from the GPs for the remaining computer-detected case patients, who also all were reviewed individually by M.L. The index date among the case patients was set to be the date when the cancer was first recorded or when the manual review revealed an earlier date of clinical diagnosis. Selection of Control Participants All members of the study cohort were assigned a random date within the study period. If the random date was within that individual s eligible person-time, that person was marked as an eligible control subject and we used his or her random date as the index date. Thereafter, we randomly frequency-matched 10,000 controls on sex, age (within 1 year), and calendar year (year at index date). Definition of Drug Exposure Drug exposure was recorded or classified as not recorded for the use of the drug. If the drug had been used within 1 year before the index date it was further classified as current use, whereas past use represented use before that. The duration of treatment was calculated adding the periods of consecutive prescriptions, defined as an interval of less than 6 months between 2 prescriptions of the same drug. Treatment duration was grouped into less than 3 years, or more than 3 years. The daily dose of ACE inhibitors was categorized as low/medium or high dose for each individual ACE inhibitor. Limits of these doses of ACE inhibitors are presented in the Appendix (see supplementary material online at Statistical Analyses We evaluated the association between the use of ACE inhibitors and the risk of esophageal and gastric cancers compared with nonuse. For comparison we also evaluated the association between the use of 3 other common antihypertensive agents and these tumors, that is, calcium channel blockers (Ca-blockers), -receptor inhibitors ( -blockers), and diuretics. Unconditional logistic regression was conducted to calculate odds ratios (ORs) with 95% confidence intervals (CIs). In multivariable analyses, all estimates of risk were adjusted for potential confounding factors in 2 models. A basic model included age (in 10-year intervals), sex, calendar year, tobacco smoking (categorized into 4 groups: nonsmoker, current smoker, ex-smoker, or unknown), alcohol consumption (categorized into 5 groups: 0 2, 3 15, 16 34, 34 U/day 1, or unknown, in which 1 unit corresponded to 10 ml or 7.9 g of pure ethanol), and body mass index (BMI) (categorized into 5 groups: 20, , , 30 kg/m 2, or unknown). A full model also included upper gastrointestinal disorders (grouped into the following: gastroesophageal reflux, esophagitis, dyspepsia, and peptic ulcer disease) and concurrent drug use (ie, use of nonsteroidal inflammatory drugs, aspirin, ACE inhibitors, Cablockers, -blockers, and diuretics). To further evaluate the influence of protopathic bias (ie, reversed causality) all basic analyses also were performed using 1 year of lag-time (ie, excluding all exposure and comorbidity data recorded 1 year before index date both among cases and controls). Results Patients and Basic Characteristics A total of 2128 patients were identified with a newly diagnosed cancer in the esophagus or the stomach among 4,340,270 person-years at risk. Our manual review excluded 178 patients because of at least one of the reasons stated in the Materials and Methods section, and another 18 patients because the review could not distinguish whether the tumor was of esophageal or gastric origin. A total of 909 cases of esophageal cancer was gathered, among which 287 had, 140 had squamous-cell cancer, and the remaining 482 patients had an unknown histologic type of esophageal cancer (Table 1). All in all, 1023 cases of gastric cancer of the different subsites as shown in Table 2, and 10,000 matched random controls were selected for analysis. Tables 1 and 2 show the distribution of some risk factors (eg, high BMI was a risk factor for esophageal ). Smokers were more common among cases of esophageal carcinoma, and squamous-cell cancer. High alcohol consumption ( 34 U/day) was a risk factor for squamous-cell cancer. High BMI was more common among cases of gastric cardia than the controls. Smoking was a risk factor for noncardia gastric, but not for the cardia subsite. Other basic characteristics among the study participants are shown in Tables 1 and 2. Angiotensin-Converting Enzyme Inhibitors and Risk of Esophageal Cancer The association between the use of antihypertension drugs and esophageal cancer is presented in Table 3. Current users of ACE inhibitors had a statistically nonsignificant lowered risk of esophageal as compared with nonusers (OR, 0.71; 95% CI, ), although no such effect could be observed for esophageal squamous-cell carcinoma (OR, 1.27; 95% CI, ). Individuals on a high daily dose of ACE inhibitors were at a decreased risk of both and squamous-cell carcinoma of the esophagus, the OR for all esophageal cancer was 0.55 (95% CI, ) (Table 3).

3 1162 SJÖBERG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 5, No. 10 Table 1. Distribution of Main Characteristics Among Cases With Esophageal Cancer and Controls Controls Esophageal Esophageal squamous-cell carcinoma Esophageal cancer of unknown histology Total esophageal cancer Exposure N(%) N (%) N (%) N (%) N (%) Total 10, BMI, kg/m (2.6) 8 (2.8) 9 (6.4) 20 (4.2) 37 (4.1) (22.2) 49 (17.1) 34 (24.3) 88 (18.3) 171 (18.8) (24.5) 94 (32.8) 39 (27.9) 97 (20.1) 230 (25.3) (8.7) 36 (12.5) 4 (2.9) 48 (10.0) 88 (9.7) Unknown BMI 4210 (42.1) 100 (34.8) 54 (38.6) 229 (47.5) 383 (42.1) Tobacco smoking status Nonsmoker 4625 (46.2) 127 (44.2) 60 (42.9) 176 (36.5) 363 (39.9) Ex-smoker 825 (8.2) 27 (9.4) 6 (4.3) 44 (9.1) 77 (8.5) Current smoker 1571 (15.7) 65 (22.6) 33 (23.6) 115 (23.9) 213 (23.4) Unknown status 2979 (29.8) 68 (23.7) 41 (29.3) 147 (30.5) 256 (28.2) Use of alcohol, U/day (33.7) 95 (33.1) 49 (35.0) 150 (31.1) 294 (32.3) (16.6) 59 (5.2) 20 (14.3) 77 (16.0) 156 (17.2) (5.6) 15 (5.2) 13 (9.3) 26 (5.4) 54 (5.9) (1.8) 9 (3.1) 5 (3.6) 16 (3.3) 30 (3.3) Unknown use 4219 (42.2) 109 (38.0) 53 (37.9) 213 (44.2) 375 (41.2) More than 3 years of current use of ACE inhibitors did not translate into a decreased risk of esophageal (OR, 0.91; 95% CI, ) or squamous-cell carcinoma (OR, 1.03; 95% CI, ), whereas less than 3 years of use lowered the risk statistically nonsignificantly by 39% (OR, 0.61; 95% CI, ) for esophageal, but did not decrease the risk of squamous-cell carcinoma (OR, 1.43; 95% CI, ). No clear association was found between current use of Cablockers, -blockers, or diuretics and esophageal cancer, respectively (Table 3). Antihypertensive Agents and Risk of Gastric Cancer Our data showed no clear association between the use of any of the assessed antihypertensive agents and gastric cancer Table 2. Distribution of Main Characteristics Among Cases With Gastric Cancer and Controls Controls Gastric cardia Noncardia gastric Unknown subsite of gastric Total gastric Exposure N(%) N (%) N (%) N (%) N (%) Total 10, BMI, kg/m (2.6) 2 (1.03) 16 (4.9) 11 (2.2) 29 (2.8) (22.2) 36 (18.5) 70 (21.4) 111 (22.2) 217 (21.2) (24.5) 55 (28.2) 83 (25.4) 116 (23.2) 254 (24.8) (8.7) 20 (10.3) 23 (7.0) 55 (11.0) 98 (9.6) Unknown BMI 4210 (42.1) 82 (42.0) 135 (41.3) 208 (41.5) 425 (41.5) Tobacco smoking status Nonsmoker 4625 (46.2) 75 (38.5) 134 (41.0) 196 (39.1) 405 (39.6) Ex-smoker 825 (8.2) 21 (10.8) 25 (7.6) 48 (9.6) 94 (9.2) Current smoker 1571 (15.7) 36 (18.5) 75 (22.9) 111 (22.2) 222 (21.7) Unknown status 2979 (29.8) 63 (32.3) 93 (28.4) 146 (29.1) 302 (29.5) Use of alcohol, U/day (33.7) 55 (28.2) 124 (37.9) 172 (34.3) 351 (34.3) (16.6) 33 (16.9) 61 (18.6) 72 (14.4) 166 (16.2) (5.6) 14 (7.2) 19 (5.8) 25 (5.0) 58 (5.7) (1.8) 4 (2.0) 2 (0.6) 10 (2.0) 16 (1.6) Unknown use 4219 (42.2) 89 (45.6) 121 (37.0) 222 (44.3) 432 (42.2)

4 Table 3. Associations Between Use of ACE Inhibitors, Ca-Blockers, -Blockers, and Diuretics and Risk of Esophageal Cancer Compared With Nonusers, Presented as ORs and 95% CIs Exposure Controls Esophageal Esophageal squamous-cell carcinoma Esophageal cancer of unknown histology Total esophageal cancer N(%) N (%) OR (95% CI) N (%) OR (95% CI) N (%) OR (95% CI) N (%) OR (95% CI) Total 10,000 (100) 287 (100) 140 (100) 482 (100) 909 (100) Use of ACE inhibitors Never 8955 (89.6) 262 (91.3) 1.00 (reference) 124 (88.6) 1.00 (reference) 444 (92.1) 1.00 (reference) 830 (91.3) 1.00 (reference) Past 215 (2.2) 7 (2.4) 1.08 ( ) a 1.02 ( ) b 2 (1.4) 0.70 ( ) a 0.62 ( ) b 8 (1.7) 0.74 ( ) a 0.71 ( ) b 17 (1.9) 0.84 ( ) a 0.78 ( ) b Current use 830 (8.3) 18 (6.3) 0.67 ( ) a 0.71 ( ) b 14 (10.0) 1.44 ( ) a 1.27 ( ) b 30 (6.2) 0.73 ( ) a 0.73 ( ) b 62 (6.8) 0.80 ( ) a 0.79 ( ) b Current use lowmedium dose 546 (5.5) 11 (3.8) 0.64 ( ) a 0.68 ( ) b 11 (7.9) 1.72 ( ) a 1.56 ( ) b 24 (5.0) 0.89 ( ) a 0.78 ( ) b 46 (5.1) 0.91 ( ) a 0.85 ( ) b Current use high dose 284 (2.8) 7 (2.4) 0.73 ( ) a 0.78 ( ) b 3 (2.1) 0.89 ( ) a 0.84 ( ) b 6 (1.2) 0.43 ( ) a 0.36 ( ) b 16 (1.8) 0.60 ( ) a 0.55 ( ) b Current use 3 y 570 (5.7) 10 (3.5) 0.55 ( ) a 0.61 ( ) b 11 (7.9) 1.66 ( ) a 1.43 ( ) b 20 (4.1) 0.71 ( ) a 0.59 ( ) b 41 (4.5) 0.78 ( ) a 0.70 ( ) b Current use 3 y 260 (2.6) 8 (2.8) 0.93 ( ) a 0.91 ( ) b 3 (2.1) 0.95 ( ) a 1.03 ( ) b 10 (2.1) 0.79 ( ) a 0.76 ( ) b 21 (2.3) 0.86 ( ) a 0.84 ( ) b Use of Ca-blockers Never 8302 (83.0) 239 (83.3) 1.00 (reference) 120 (85.7) 1.00 (reference) 400 (83.0) 1.00 (reference) 759 (83.5) 1.00 (reference) Past 421 (4.2) 7 (2.4) 0.56 ( ) a 0.58 ( ) b 4 (2.9) 0.69 ( ) a 0.61 ( ) b 25 (5.2) 1.22 ( ) a 1.19 ( ) b 36 (4.0) 0.93 ( ) a 0.91 ( ) b Current use 1277 (12.8) 41 (14.3) 1.08 ( ) a 1.10 ( ) b 16 (11.4) 1.01 ( ) a 0.90 ( ) b 57 (11.8) 0.94 ( ) a 0.94 ( ) b 114 (12.5) 0.99 ( ) a 0.98 ( ) b Current use 3 y 743 (7.4) 20 (7.0) 0.90 ( ) a 10 (7.1) 1.06 ( ) a 29 (6.0) 0.82 ( ) a 59 (6.5) 0.88 ( ) a Current use 3 y 534 (5.3) 21 (7.3) 1.34 ( ) a 6 (4.3) 0.94 ( ) a 28 (5.8) 1.13 ( ) a 55 (6.0) 1.17 ( ) a Use of -blockers Never 7934 (79.3) 229 (79.8) 1.00 (reference) 110 (78.6) 1.00 (reference) 403 (83.6) 1.00 (reference) 742 (81.6) 1.00 (reference) Past 784 (7.8) 19 (6.6) 0.80 ( ) a 0.76 ( ) b 11 (7.9) 0.93 ( ) a 0.89 ( ) b 27 (5.6) 0.69 ( ) a 0.68 ( ) b 57 (6.3) 0.77 ( ) a 0.74 ( ) b Current use 1282 (12.8) 39 (13.6) 1.07 ( ) a 1.05 ( ) b 19 (13.6) 1.17 ( ) a 1.15 ( ) b 52 (10.8) 0.85 ( ) a 0.84 ( ) b 110 (12.1) 0.96 ( ) a 0.96 ( ) b Current use 3 y 655 (6.6) 21 (7.3) 1.12 ( ) a 10 (7.1) 1.14 ( ) a 22 (4.6) 0.69 ( ) a 53 (5.8) 0.89 ( ) a Current use 3 y 627 (6.3) 18 (6.3) 1.02 ( ) a 9 (6.4) 1.20 ( ) a 30 (6.2) 1.01 ( ) a 57 (6.3) 1.04 ( ) a Use of diuretics Never 6881 (68.8) 202 (70.4) 1.00 (reference) 93 (66.4) 1.00 (reference) 313 (64.9) 1.00 (reference) 608 (66.9) 1.00 (reference) Past 826 (8.3) 22 (7.7) 0.95 ( ) a 0.90 ( ) b 17 (12.1) 1.55 ( ) a 1.55 ( ) b 34 (7.0) 0.87 ( ) a 0.86 ( ) b 73 (8.0) 1.28 ( ) a 0.97 ( ) b Current use 2293 (22.9) 63 (22.0) 1.06 ( ) a 1.03 ( ) b 30 (21.4) 1.07 ( ) a 1.06 ( ) b 135 (28.0) 1.27 ( ) a 1.26 ( ) b 228 (25.1) 1.17 ( ) a 1.15 ( ) b Current use 3 y 1443 (14.4) 33 (11.5) 0.88 ( ) a 23 (16.4) 1.28 ( ) a 98 (20.3) 1.44 ( ) a 154 (16.9) 1.24 ( ) a Current use 3 y 850 (8.5) 30 (10.4) 1.39 ( ) a 7 (5.0) 0.69 ( ) a 37 (7.7) 0.96 ( ) a 74 (8.1) 1.05 ( ) a a OR adjusted for age, sex, calendar year, smoking, alcohol consumption, and BMI. b OR adjusted for age, sex, calendar year, smoking, alcohol consumption, BMI, gastroesophageal reflux, esophagitis, dyspepsia, peptic ulcer, use of nonsteroidal anti-inflammatory drugs, aspirin, Ca-blockers, -blockers, and diuretics. October 2007 ACE INHIBITORS AND RISK OF ESOPHAGEAL AND GASTRIC CANCER 1163

5 Table 4. Associations Between Use of ACE Inhibitors, Ca-Blockers, -Blockers, and Diuretics and Risk of Gastric Cancer, Compared With Nonusers, Presented as ORs and 95% CIs Exposure Controls Gastric cardia Gastric noncardia Gastric of unknown subsite Total gastric N(%) N (%) OR (95% CI) N (%) OR (95% CI) N (%) OR (95% CI) N (%) OR (95% CI) Total 10,000 (100) 195 (100) 327 (100) 501 (100) 1023 (100) Use of ACE inhibitors Never 8955 (89.6) 175 (89.7) 1.00 (reference) 288 (88.1) 1.00 (reference) 439 (87.6) 1.00 (reference) 902 (88.2) 1.00 (reference) Past 215 (2.2) 3 (1.5) 0.70 ( ) a 0.60 ( ) b 5 (1.5) 0.72 ( ) a 0.62 ( ) b 12 (2.4) 1.10 ( ) a 0.97 ( ) b 20 (2.0) 0.90 ( ) a 0.80 ( ) b Current use 830 (8.3) 17 (8.7) 1.00 ( ) a 0.97 ( ) b 34 (10.4) 1.32 ( ) a 1.27 ( ) b 50 (10.0) 1.18 ( ) a 1.09 ( ) b 101 (9.9) 1.18 ( ) a 1.11 ( ) b Current use lowmedium dose 546 (5.5) 10 (5.1) 0.91 ( ) a 0.76 ( ) b 25 (7.7) 1.47 ( ) a 1.39 ( ) b 36 (7.2) 1.29 ( ) a 1.16 ( ) b 71 (6.9) 1.26 ( ) a 1.14 ( ) b Current use high dose 284 (2.8) 7 (3.6) 1.17 ( ) a 1.02 ( ) b 9 (2.8) 1.04 ( ) a 0.97 ( ) b 14 (2.8) 0.98 ( ) a 0.90 ( ) b 30 (2.9) 1.04 ( ) a 0.93 ( ) b Current use 3 y 570 (5.7) 10 (5.1) 0.86 ( ) a 0.73 ( ) b 27 (8.3) 1.52 ( ) a 1.42 ( ) b 32 (6.4) 1.10 ( ) a 0.98 ( ) b 69 (6.7) 1.18 ( ) a 1.05 ( ) b Current use 3 y 260 (2.6) 7 (3.6) 1.30 ( ) a 1.10 ( ) b 7 (2.1) 0.88 ( ) a 0.86 ( ) b 18 (3.6) 1.37 ( ) a 1.28 ( ) b 32 (3.1) 1.20 ( ) a 1.11 ( ) b Use of Ca-blockers Never 8302 (83.0) 152 (78.0) 1.00 (reference) 265 (81.0) 1.00 (reference) 417 (83.2) 1.00 (reference) 834 (81.5) 1.00 (reference) Past 421 (4.2) 9 (4.6) 1.16 ( ) a 1.13 ( ) b 16 (4.9) ( ) a 1.11 ( ) b 27 (5.4) 1.16 ( ) a 1.03 ( ) b 52 (5.1) 1.16 ( ) a 1.07 ( ) b Current use 1277 (12.8) 34 (17.4) 1.16 ( ) a 1.30 ( ) b 46 (14.1) 1.16 ( ) a 1.08 ( ) b 57 (11.4) 0.84 ( ) a 0.74 ( ) b 137 (13.4) 1.16 ( ) a 0.94 ( ) b Current use 3 y 743 (7.4) 23 (5.3) 1.67 ( ) a 28 (8.6) 1.22 ( ) a 39 (7.8) 0.99 ( ) a 90 (8.8) 1.19 ( ) a Current use 3 y 534 (5.3) 11 (5.6) 1.15 ( ) a 18 (5.5) 1.10 ( ) a 18 (3.6) 0.64 ( ) a 47 (4.6) 0.87 ( ) a Use of -blockers Never 7934 (79.3) 147 (75.4) 1.00 (reference) 255 (78.0) 1.00 (reference) 403 (80.4) 1.00 (reference) 805 (78.7) 1.00 (reference) Past 784 (7.8) 16 (8.2) 1.09 ( ) a 0.98 ( ) b 28 (8.6) 1.09 ( ) a 0.98 ( ) b 29 (5.8) 0.73 ( ) a 0.69 ( ) b 73 (7.1) 0.91 ( ) a 0.84 ( ) b Current use 1282 (12.8) 32 (16.4) 1.37 ( ) a 1.32 ( ) b 44 (13.5) 1.10 ( ) a 1.06 ( ) b 69 (13.8) 1.07 ( ) a 1.04 ( ) b 145 (14.2) 1.14 ( ) a 1.10 ( ) b Current use 3 y 655 (6.6) 15 (7.7) 1.25 ( ) a 22 (6.7) 1.06 ( ) a 33 (6.6) 1.00 ( ) a 70 (6.8) 1.06 ( ) a Current use 3 y 627 (6.3) 17 (8.7) 1.51 ( ) a 22 (6.7) 1.16 ( ) a 36 (7.2) 1.14 ( ) a 75 (7.3) 1.21 ( ) a Use of diuretics Never 6881 (68.8) 126 (64.6) 1.00 (reference) 220 (67.3) 1.00 (reference) 313 (62.5) 1.00 (reference) 659 (64.4) 1.00 (reference) Past 826 (8.3) 19 (9.7) 1.36 ( ) a 1.21 ( ) b 28 (8.6) 1.07 ( ) a 0.98 ( ) b 58 (11.6) 1.46 ( ) a 1.38 ( ) b 105 (10.3) 1.32 ( ) a 1.23 ( ) b Current use 2293 (22.9) 50 (25.6) 1.33 ( ) a 1.24 ( ) b 79 (24.2) 1.13 ( ) a 1.05 ( ) b 130 (26.0) 1.17 ( ) a 1.13 ( ) b 259 (25.3) 1.18 ( ) a 1.12 ( ) b Current use 3 y 1443 (14.4) 27 (13.8) 1.13 ( ) a 52 (15.9) 1.17 ( ) a 86 (17.2) 1.22 ( ) a 165 (16.1) 1.19 ( ) a Current use 3 y 850 (8.5) 23 (11.8) 1.70 ( ) a 27 (8.3) 1.06 ( ) a 44 (8.8) 1.09 ( ) a 94 (9.2) 1.18 ( ) a a OR adjusted for age, sex, calendar year, smoking, alcohol consumption, and BMI. b OR adjusted for age, sex, calendar year, smoking, alcohol consumption, BMI, gastroesophageal reflux, esophagitis, dyspepsia, peptic ulcer, use of nonteroidal anti-inflammatory drugs, aspirin, Ca-blockers, -blockers, and diuretics SJÖBERG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 5, No. 10

6 October 2007 ACE INHIBITORS AND RISK OF ESOPHAGEAL AND GASTRIC CANCER 1165 (Table 4). Current use of ACE inhibitors was not associated with gastric cardia (OR, 0.97; 95% CI, ). Neither was current use related to risk of gastric noncardia (OR, 1.27; 95% CI, ) nor with gastric cancer of unknown subsite (OR, 1.09; 95% CI, ). No further association was found when assessing dose, duration (Table 4), and risk of gastric cancer. When we used a 1-year lag-time analysis the results remained virtually identical (data not shown). No obvious correlation was found between current use of Ca-blockers, -blockers, or diuretics and gastric cancer, respectively (Table 4). Discussion Our large population-based study suggests that ACE inhibitors could lower the risk of developing esophageal, but no clear effect could be found for esophageal squamous-cell carcinoma or gastric cancer. The risk of esophageal and gastric cancer was not altered among patients taking Ca-blockers, -blockers, or diuretics compared with nonusers. The main strengths of our study include its populationbased design, its prospectively recorded exposure data, especially details on drug use, and its thorough manual review of cases. Our study was a large pharmacoepidemiologic study conducted on esophageal and gastric cancers, however, our analyses were hampered by limited statistical power, which could be an explanation for the wide CIs and inability to reach statistical significance. Despite this, we have analyzed the cases according to type of esophageal carcinoma and subsite of gastric cancer. Other potential limitations of the study are the lack of complete information on some exposures (eg, tobacco smoking, alcohol consumption, and BMI) and missing information on some other risk factors for the studied tumors (eg, Helicobacter pylori, dietary habits, socioeconomic status, hereditary antecedents, and gastroesophageal reflux disease) None of these exposures, however, should be associated strongly with the use of ACE inhibitors, and therefore should not act as confounders of the studied associations. The epidemiologic evidence for a protective effect on cancer development of ACE inhibitors is not clear-cut. The first published report, a retrospective cohort study from Scotland, showed a significantly lower overall cancer incidence as well as overall cancer mortality among patients prescribed ACE inhibitors. 6 No clear protective effect on overall cancer incidence was found in 2 Scandinavian studies, 7,8 however, the investigators concluded that a long-term effect on some digestive cancers, such as esophageal, gastric, and liver cancer, could not be ruled out. 7,8 These studies were hampered by the inability to adjust for potential confounders such as smoking and alcohol and of limited statistical power to study gastrointestinal tumors. 6 8 Recent, yet unpublished, results presented at Digestive Disease Week 2006, from a case-control study within the Veterans Integrated Service Network database lend further support to our findings, reporting a 55% significant risk reduction of esophageal cancer among ACE inhibitor users) Presented at Digestive Disease Week, 2006 in Los Angeles, CA). However, no data on esophageal cancer histologic subtypes, dosage, or duration of ACE inhibitor use was presented. In our results, exposure of less than 3 years conferred a statistically significant risk reduction whereas exposure of more than 3 years did not. This finding should be interpreted cautiously because our number of cases was small, but we have no readily available explanation for these results. The important finding in our study was the consistent nonsignificant statistically but potentially clinically significant reduced risk of esophageal cancer in association with ACE inhibitors. A possible biological mechanism through which ACE inhibitors prevent tumor growth has been suggested by in vitro studies in which Ang II has been shown to have many physiologic effects, including strong pro-angiogenic activity, by inducing the potent angiogenic factor vascular endothelial growth factor. Recent studies have revealed that angiogenesis is an essential process in many pathologic events, including a crucial step for a tumor to exceed 2 mm 3 in size. 17 In addition, it has been shown in vitro and in animal studies that Ang II could promote cell proliferation and growth. 1 4 However, to establish a causal pathway for ACE inhibition and reduced risk of esophageal cancer more research is needed. In conclusion, our study adds further support to the hypothesis that ACE inhibitors may protect against the development of esophageal cancer, and particularly of. We found a suggestion of a dose-response relationship with greater decreased risk of esophageal cancer at higher doses. No association was found between ACE inhibitors and gastric cancer. There is a need for more basic and clinical research to determine the possible role of ACE inhibitors and cancer prevention before any firm conclusions can be drawn. Supplementary Data Note: To access the supplementary material accompanying this article, visit the online version of Clinical Gastroenterology and Hepatology at References 1. Ino K, Shibata K, Kajiyama H, et al. Manipulating the angiotensin system new approaches to the treatment of solid tumours. Expert Opin Biol Ther 2006;6: Achard JM, Pruna A, Fernandez LA, et al. Prevention of stroke and cancer: could angiotensin II type 1 receptor antagonists do better than angiotensin II converting enzyme inhibitors? Am J Hypertens 1999;12: Smith GR, Missailidis S. Cancer, inflammation and the AT1 and AT2 receptors. J Inflamm (Lond) 2004;1:3. 4. Egami K, Murohara T, Shimada T, et al. Role of host angiotensin II type 1 receptor in tumor angiogenesis and growth. J Clin Invest 2003;112: Herbst RS, Hidalgo M, Pierson AS, et al. Angiogenesis inhibitors in clinical development for lung cancer. Semin Oncol 2002;29: Lever AF, Hole DJ, Gillis CR, et al. Do inhibitors of angiotensin-iconverting enzyme protect against risk of cancer? Lancet 1998; 352: Lindholm LH, Anderson H, Ekbom T, et al. Relation between drug treatment and cancer in hypertensives in the Swedish Trial in Old Patients with Hypertension 2: a 5-year, prospective, randomised, controlled trial. Lancet 2001;358: Friis S, Sorensen HT, Mellemkjaer L, et al. Angiotensin-converting enzyme inhibitors and the risk of cancer: a population-based cohort study in Denmark. Cancer 2001;92: Parkin DM, Bray FI, Devesa SS. Cancer burden in the year The global picture. Eur J Cancer 2001;37(Suppl 8):S4 S66.

7 1166 SJÖBERG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 5, No Garcia Rodriguez G. Use of the UK general practice research data base for pharmacoepidemiology. Br J Clin Pharmacol 1998;45: Jick SS, Kaye JA, Vasilakis-Scaramozza C, et al. Validity of the general practice research database. Pharmacotherapy 2003;23: Lindblad M, Lagergren J, Garcia Rodriguez LA. Nonsteroidal anti-inflammatory drugs and risk of esophageal and gastric cancer. Cancer Epidemiol Biomarkers Prev 2005;14: Jansson C, Johansson AL, Nyren O, et al. Socioeconomic factors and risk of esophageal : a nationwide Swedish case-control study. Cancer Epidemiol Biomarkers Prev 2005;14: Terry PD, Lagergren J, Wolk A, et al. Dietary intake of heterocyclic amines and cancers of the esophagus and gastric cardia. Cancer Epidemiol Biomarkers Prev 2003;12: Lagergren J, Ye W, Lindgren A, et al. Heredity and risk of cancer of the esophagus and gastric cardia. Cancer Epidemiol Biomarkers Prev 2000;9: Lagergren J, Bergstrom R, Lindgren A, et al. Symptomatic gastroesophageal reflux as a risk factor for esophageal. N Engl J Med 1999;18: Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000; 100: Address requests for reprints to: Tomas Sjöberg, Unit of Esophageal and Gastric Research, P9:03, Karolinska Institutet, SE Stockholm, Sweden. tomas.sjoberg@ki.se; fax: (46) Supported by AstraZeneca R&D, Sweden. AstraZeneca did not have any influence on study design or conduct of the study, nor collection, management, analysis, interpretation of data. They did not participate in preparation, review, approval of the manuscript, or submission.

8 1166.e1 SJÖBERG ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 5, No. 10 Appendix Daily dose limits of ACE inhibitors for categorization into low-medium or high daily dose Generic drug name Limit for low-medium daily dose, mg Limit for high daily dose, mg Captopril Cilazapril Generic drug name Limit for low-medium daily dose, mg Limit for high daily dose, mg Enalapril Fosinopril Lisinopril Perindopril Quinapril Ramipril