|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 22, No 12 (June 15), 2004: pp. 2388-2394 © 2004 American Society of Clinical Oncology. DOI: 10.1200/JCO.2004.02.027 The Risk of Cancer in Users of StatinsFrom the Academic Medical Center, Departments of Clinical Pharmacy and Oncology, Amsterdam; Department of Pharmaco-epidemiology & Pharmacotherapy, Utrecht Institute for Pharmaceutical Sciences (UIPS), Utrecht; and Department of Clinical Pharmacy and Toxicology, Leiden University Medical Centre, Leiden, the Netherlands Address reprint requests to A.C.G. Egberts, PhD, Department of Pharmaco-epidemiology & Pharmacotherapy, Utrecht Institute for Pharmaceutical Sciences (UIPS), PO Box 80082, 3508 TB, Utrecht, The Netherlands; e-mail: A.C.G.Egberts{at}pharm.uu.nl
PURPOSE: Several preclinical studies suggested a role for 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors (statins) in the treatment of cancer. The objective of this study was to compare the risk of incident cancer between users of statins and users of other cardiovascular medication. METHODS: Data were used from the PHARMO database, containing drug dispensing records from community pharmacies and linked hospital discharge records for residents of eight Dutch cities. The study base included all patients with one or more prescriptions for cardiovascular drugs in the period between January 1, 1985 and December 31, 1998. Cases were identified as patients in the study base with a diagnosis of incident cancer and matched with four to six controls on sex, year of birth, geographic region, duration of follow-up, and index date. The analysis was adjusted for diabetes mellitus; prior hospitalizations; comorbidity; and use of diuretics, angiotensin-converting enzyme inhibitors, calcium-channel blockers, nonsteroidal anti-inflammatory drugs, sex hormones, and other lipid-lowering drug therapies. RESULTS: In the study base, 3,129 patients were identified and matched to 16,976 controls. Statin use was associated with a risk reduction of cancer of 20% (adjusted odds ratio [OR], 0.80; 95% CI, 0.66 to 0.96). Our data suggest that statins are protective when used longer than 4 years (adjusted OR, 0.64; 95% CI, 0.44 to 0.93) or when more than 1,350 defined daily doses are taken (adjusted OR, 0.60; 95% CI, 0.40 to 0.91). CONCLUSION: This observational study suggests that statins may have a protective effect against cancer.
3-Hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors (statins) are efficient and widely used drugs in the treatment of lipid disorders, especially hypercholesterolemia. In large follow-up studies their efficacy on cardiovascular events has been proven irrefutably for both reduction of morbidity and mortality.1–5 Accordingly, the use of statins in the Netherlands increased four-fold in the period from 1995 to 1999.6 The mechanism of action of this class of drugs is considered to be the inhibition of cholesterol biosynthesis through inhibition of the enzyme HMG-CoA reductase. This results in depletion of mevalonate, which is a precursor of cholesterol. Mevalonate also is a precursor of farnesyl and geranylgeranyl moieties, which are essential for the activation of a variety of intracellular proteins through so-called farnesylation or geranylgeranylation (prenylation).7 Several proteins involved in signaling are dependent on prenylation for their activity, such as Ras, Rho, nuclear lamins, transducin  , Rap1, and Cdc42. The Ras protein is important in the regulation of cell differentiation and proliferation. Given that approximately 30% of the human tumors have a mutation of k-Ras oncogene, expression of this oncogene is thought to be related to aberrant cellular growth.8 Given that statins are able to inhibit farnesylation and hence activation of Ras, they might have the capacity to inhibit the expression of the malignant phenotype of a tumor cell and to restore normal cellular growth. Several in vitro studies have suggested that statins do have antitumor potential.9–15 In addition, in vivo experiments in laboratory animals indicated that pancreatic tumors and tumors of the colon and the breast are sensitive to statins.11,16,17 Furthermore, in a phase I clinical study with lovastatin, one minor response was observed in a patient with recurrent high-grade glioma.18 Recently, a small, nonblinded, randomized, controlled clinical trial was performed in patients with advanced hepatocellular carcinoma. Both a reduction in maximum tumor diameter and a prolonged survival were observed in the pravastatin group compared with the control group.19 In contrast, Newman and Hulley20 reviewed animal studies with lipid-lowering therapy and concluded that statins and fibrates might cause cancer in rodents. However, no carcinogenicity of statins has been observed in individual clinical trials and meta-analyses.1–5,21 Blais et al22 investigated the association between the use of statins and the incidence of cancer in a cohort of patients using lipid-lowering therapy. Lower risk of cancer was found in patients using statins, as compared with patients using bile acid-binding resins (relative risk, 0.72; 95% CI, 0.57 to 0.92). However, the median follow-up time (2.7 years) was relatively short.23 In addition, statin users were defined on the basis of one prescription of statins, irrespective of the duration of therapy. The aim of this study was to investigate the association between statin therapy and the risk of cancer. Therefore, we conducted a population-based, nested, case-control study exploring prospectively gathered automated pharmacy data and linked hospital morbidity data.
Setting Data were used from the PHARMO record linkage system, a database that contains drug dispensing records from community pharmacies and linked hospital discharge records of a defined population of approximately 300,000 residents of eight medium-sized cities in the Netherlands from 1985 onward. Because almost all individuals designate a single pharmacy to fill their prescriptions from general practitioners or medical specialists, dispensing histories are virtually complete.24 In the PHARMO system, the drug dispensing history of each individual is linked to hospital discharge records using a validated and reliable probabilistic algorithm.25 The computerized drug dispensing histories contain data regarding all dispensed prescriptions and include type and quantity of the dispensed drug, dosage form, strength, type of prescriber, dispensing date, and prescribed daily dose. Hospital discharge records include detailed information regarding primary and secondary diagnoses, performed medical procedures, and dates of admission and discharge. All diagnoses are coded according to the International Classification of Disease (9th revision, clinical modification; ICD9-CM). The potential study period consisted of the 14-year period between January 1, 1985 and December 31, 1998. Participants of the PHARMO population enter the database with the first prescription filled in a PHARMO community pharmacy and are observed until the last prescription.
Study Population Controls were sampled from patients in the study base that had not been discharged with a primary or secondary code for cancer, who had not filled prescriptions for anticancer medication (defined by Anatomical Therapeutic Chemical Classification codes L01, L02, L03), and who never underwent chemotherapy or radiotherapy. A potential latent period between the inhibition of malignant cell growth and the diagnosis of cancer of 6 months was assumed. Therefore, the index date was calculated by subtracting 6 months from the date of diagnosis. For every case, four to six controls were matched on sex, year of birth (± 2.5 years), geographic region, duration of follow-up (± 20%), and index date. Duration of follow-up was defined as the difference between the date of entry in PHARMO and the index date.
Drug Exposure
End Points
Potential Confounders
Data Analysis
In the study population, 3,789 patients with a primary diagnosis of cancer were identified. Of these patients, 630 had a history of cancer, chemotherapy, or radiotherapy before the date of diagnosis and were excluded because of an increased baseline risk due to reoccurrence of cancer. For 30 patients, no controls could be found. On the basis of the matching criteria, 3,129 patients could be matched to 16,976 controls. The baseline characteristics for patients and controls are listed in Table 1. Patients were present in PHARMO for a mean of 7.2 years. The chronic disease score was not statistically different for patients and controls. The median number of hospital admissions before the date of diagnosis of cancer in the patient was lower for controls than for patients. In contrast, patients used less comedication, such as diuretics, ACE inhibitors, calcium-channel blockers, and NSAIDs, compared with controls. The use of hormones was comparable in groups of patients and controls.
Approximately 6% of controls received statin therapy before the index date for at least 6 months, as compared with 5% of patients (P < .01). Participants that received prescriptions of statins for at least 6 months were prescribed mainly simvastatin (79.6%). Other prescribed statins were pravastatin (6.6%), fluvastatin (2.5%), atorvastatin (0.4%), or a combination (10.9%). Cerivastatin was not prescribed during the study period. In general, the prescribed dose was approximately 0.6 defined daily doses or approximately 1.3 defined daily doses. The use of other lipid-lowering therapy did not differ between patients and controls. Table 2 lists the effect of statin therapy on incident cancer. Statin use (at least 6 months) was associated with a risk reduction of incident cancer of 20% compared with no use of statins (adjusted OR, 0.80; 95% CI, 0.66 to 0.96). No significant difference between men and women or between different age groups was detected (data not shown). When users of statins were compared with users of other lipid-lowering therapies, an adjusted risk estimate of 0.89 (95% CI, 0.56 to 1.41) was found.
Furthermore, the effect of duration of statin use and cumulative dose is presented in Table 2. The risk decreased with duration of statin use, although the P value for trend was not statistically significant (P = .08). For cumulative dose, a significant risk reduction was found only for statin users that received more than 1,350 defined daily doses. However, an effect of the prescribed daily dose was not found (data not shown). In addition, the effect of latent period (lag time) is listed in Table 2. Patients who were past statin users and had their therapy stopped for at least 6 months had the same risk of incident cancer as nonusers of statins. In contrast, current statin users and past users that stopped for less than 6 months before the index date had a significant risk reduction of 22% to be diagnosed with cancer (OR, 0.78; 95% CI, 0.63 to 0.95). Table 3 lists the effect of statins on individual cancer sites. In the adjusted model, risk reduction was statistically significant only for renal cancer. Nonsignificant risk reductions were found for other cancer sites, except for lung, bladder, and breast cancer. Although an increased risk of lung, bladder, and breast cancer was found, numbers were small and statistical significance was not reached. To enlarge the number of patients that were diagnosed to have a cancer in the investigated sites, patients with a second diagnosis of cancer within 1 year after the primary diagnosis were included. However, this did not affect the risk estimates significantly (data not shown).
Our study suggests that the risk of incident cancer is decreased by the use of statins (OR, 0.80; 95% CI, 0.66 to 0.96). Although the observed relative risk reduction is only 20%, given the high prevalence of statin use6 and the high incidence of cancer,27 even a modest risk reduction means a considerable effect on public health. Our data suggest that statin use for a longer period and in high dosages decreases the risk of incident cancer. The protective effect was only present in current users and past users that had their therapy stopped for less than 6 months before the index date. Although statins are approved only for the treatment of lipid disorders, there is mounting evidence that they might be useful in the treatment of other diseases, such as Alzheimer's28–30 and osteoporosis.31 In addition, in vivo studies with laboratory animals11,16,17 and in vitro studies9–15 suggest a possible role for statins in the treatment of malignancies. Although the underlying biologic mechanism was originally assumed to be inhibition of Ras farnesylation, recent studies indicate that inhibition of prenylation of other proteins might be involved.32 These observations prompted us to perform a case-control study to investigate the risk of cancer among users of statins, as compared with cardiovascular patients that never used statins. One of the strengths of our study is the use of a computerized database, allowing the inclusion of 3,129 patients with 22,470 person years of follow-up. The availability of pharmacy data that were prospectively gathered before disease onset allowed us to calculate various levels of exposure of prescribed drugs. Therefore, recall bias was avoided. Using pharmacy records representing dispensing data rather than usage data might have introduced an overestimation of statin use. However, we excluded nonchronic statin users and therefore reduced the likelihood of overestimation. Prevalent cancer or the onset of cancer is not considered to exhibit symptoms that lead to a prescription of statins. In addition, we introduced a latent period of 6 months, thereby reducing protopathic bias. Furthermore, risk factors for cancer do not play a role in the prescription of statins. Therefore, it is not likely that our findings can be explained by confounding by indication. As with observational studies in general, we cannot rule out unknown biases or confounders as possible explanations for our findings. However, patients and controls were matched on sex, age, geographic region, follow-up time, and calendar time. In addition, the analysis was adjusted for diabetes mellitus and chronic use of diuretics, ACE inhibitors, calcium antagonists, NSAIDs, hormones, and other lipid-lowering therapies. Furthermore, we investigated potential confounding effects of polyposis coli and familiar hypercholesterolemia. Because of the chronic character of these diseases, we searched medical discharge records until 1 year after the index date for diagnoses of these diseases. However, neither polyposis coli nor familiar hypercholesterolemia seemed to affect the analysis significantly (data not shown). The confounding effect of medical attention could be corrected for by introducing a general score of comorbidity and the number of hospitalizations into the conditional logistic regression model. The PHARMO database does not provide information on lifestyle variables, such as body-mass index and smoking status. Nevertheless, we made an attempt to minimize the effect of these variables by confining the base population to cardiovascular patients. This group of patients is likely to receive strong recommendations regarding lifestyle from their physician. Not all of these patients will follow their doctor's orders strictly, but given that the chronic disease score was found comparable between patients and controls, there is no reason to assume that controls were more likely to improve their lifestyle habits than were patients. Moreover, coincidental differences between groups might become diluted because of the magnitude of both the patient and control group. A limitation of our study is caused by the relatively high frequency of simvastatin prescriptions: 79.6% of the statin-exposed patients were prescribed simvastatin. Because the inhibitory potency of different statins is not equal,33 the results of this study cannot be generalized to the use of other statins. Furthermore, diagnoses recorded in the PHARMO database came from hospital discharge records, but were not verified with original medical charts. This might have introduced misclassification. However, there is no reason to assume that this would be different for patients and controls. The results of our study are consistent with the assumed biologic mechanism of statins. The enzyme HMG-CoA reductase is inhibited by statins. As a consequence, statins reduce protein farnesylation and geranylgeranylation, resulting in decreased activation of proteins. Possibly, activation of oncoproteins such as k-Ras is prevented by this mechanism. However, transcription of oncogenes is not affected. Therefore, statins might prevent the expression of the oncogenic phenotype without affecting the malignant genotype. We propose that statins are able to decrease the development of existing cancer, rather than the initiation of cancer. On the basis of this hypothesis, it is expected that after patients stop statin therapy, the cancer risk returns to baseline values because of normalization of oncoprotein activation. The mean duration of observation was 7.2 years for patients and 7.3 years for controls. This period might be too short to study the influence of statins on the initiation of cancer. However, the aim of our study is to investigate the effect of statins on the development of cancer. There is no reason to assume that statins can only influence the development of the malignant phenotype when statins are used during the entire period between the initiation of cancer and the diagnosis of cancer. Therefore, the duration of follow-up is sufficient to investigate the end point of the study. Our data are consistent with the study performed by Blais et al,22 who found a risk reduction of 28% for statin users compared with users of bile acid-binding resins (risk estimate, 0.72; 95% CI, 0.57 to 0.92). Other epidemiologic studies did not find a significant effect of statins on the incident cancer risk. However, in contrast to the current study and the study performed by Blais et al,22 these epidemiologic studies were not designed to investigate the effect of statins on the occurrence of cancer. Given that these studies were aimed at investigating cardiovascular end points, patients with pre-existing cancer were not excluded.1–5 Blais et al22 compared users of statins with users of bile acid-binding resins. Their results indicate that the protective effect of statins cannot be explained by their lipid-lowering effect. We found a risk reduction when users of statins were compared with users of other lipid-lowering therapies, although the protective effect was not statistically significant. Inhibition of HMG-CoA reductase by statins has been associated with upregulation of this enzyme.34 This feedback system would theoretically indicate that cells might acquire resistance against statins by extended treatment. As a consequence, the risk of incident cancer would be expected to return to baseline values after extended statin treatment. However, this phenomenon did not appear in our study; we found a decreased risk in patients who used statins for more than 4 years, indicating that protection by statins is not temporary. In conclusion, our observational study suggests a protective effect of statins against cancer. In addition to experimental studies on the underlying mechanisms of the anticancer activity of statins, prospective randomized studies are necessary to validate the value of statins in cancer prevention.
The authors indicated no potential conflicts of interest.
Presented in part at the Meeting of the Royal Dutch Society of Clinical Pharmacology and Biopharmacy, October 10, 2002, Lunteren, the Netherlands; and the 39th Annual Meeting of the American Society of Clinical Oncology, May 31-June 3, 2003, Chicago, IL. Authors' disclosures of potential conflicts of interest are found at the end of this article.
1. Pedersen TR, Wilhelmsen L, Faergeman O, et al: Follow-up study of patients randomized in the Scandinavian simvastatin survival study (4S) of cholesterol lowering. Am J Cardiol 86:257–262, 2000[CrossRef][Medline]
2. Shepherd J, Cobbe SM, Ford I, et al: Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia. N Engl J Med 333:1301–1307, 1995
3. Downs JR, Clearfield M, Weis S, et al: Primary prevention of acute coronary events with lovastatin in men and women with average cholesterol levels. JAMA 279:1615–1622, 1998
4. Prevention of cardiovascular events and death with pravastatin in patients with coronary heart disease and a broad range of initial cholesterol levels. The Long-Term Intervention with Pravastin in Ischaemic Disease (LIPID) Study Group. N Engl J Med 339:1349–1357, 1998
5. Sacks FM, Pfeffer MA, Moye LA, et al: The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels. N Engl J Med 335:1001–1009, 1996 6. College voor zorgverzekeringen: Lipidenverlagende middelen 1995-1999. GIPeilingen 16:44–46, 2001 7. Goldstein JL, Brown MS: Regulation of the mevalonate pathway. Nature 343:425–430, 1990[CrossRef][Medline]
8. Bos JL: Ras oncogenes in human cancer: A review. Cancer Res 49:4682–4689, 1989 9. Rubins JB, Greatens T, Kratzke RA, et al: Lovastatin induces apoptosis in malignant mesothelioma cells. Am J Respir Crit Care Med 157:1616–1622, 1998 10. Maksumova L, Ohnishi K, Muratkhodjaev F, et al: Increased sensitivity of multidrug-resistant myeloid leukemia cell lines to lovastatin. Leukemia 14:1444–1450, 2000[CrossRef][Medline] 11. Sumi S, Beauchamp RD, Townsend CM Jr, et al: Inhibition of pancreatic adenocarcinoma cell growth by lovastatin. Gastroenterology 103:982–989, 1992[Medline] 12. Kawata S, Nagase T, Yamasaki E, et al: Modulation of the mevalonate pathway and cell growth by pravastatin and d-limonene in a human hepatoma cell line (Hep G2). Br J Cancer 69:1015–1020, 1994[Medline] 13. Macaulay RJ, Wang W, Dimitroulakos J, et al: Lovastatin-induced apoptosis of human medulloblastoma cell lines in vitro. J Neurooncol 42:1–11, 1999[CrossRef][Medline]
14. Dimitroulakos J, Nohynek D, Backway KL, et al: Increased sensitivity of acute myeloid leukemias to lovastatin-induced apoptosis: A potential therapeutic approach. Blood 93:1308–1318, 1999
15. Dimitroulakos J, Ye LY, Benzaquen M, et al: Differential sensitivity of various pediatric cancers and squamous cell carcinomas to lovastatin-induced apoptosis: therapeutic implications. Clin Cancer Res 7:158–167, 2001 16. Alonso DF, Farina HG, Skilton G, et al: Reduction of mouse mammary tumor formation and metastasis by lovastatin, an inhibitor of the mevalonate pathway of cholesterol synthesis. Breast Cancer Res Treat 50:83–93, 1998[CrossRef][Medline] 17. Iishi H, Tatsuta M, Baba M, et al: rasp21 Isoprenylation inhibition induces flat colon tumors in Wistar rats. Dis Colon Rectum 43:70–75, 2000[CrossRef][Medline] 18. Thibault A, Samid D, Tompkins AC, et al: Phase I study of lovastatin, an inhibitor of the mevalonate pathway, in patients with cancer. Clin Cancer Res 2:483–491, 1996[Abstract] 19. Kawata S, Yamasaki E, Nagase T, et al: Effect of pravastatin on survival in patients with advanced hepatocellular carcinoma: A randomized controlled trial. Br J Cancer 84:886–891, 2001[CrossRef][Medline]
20. Newman TB, Hulley SB: Carcinogenicity of lipid-lowering drugs. JAMA 275:55–60, 1996 21. Bjerre L, Lelorier J: Do statins cause cancer? A meta-analysis of large randomized clinical trials. Am J Med 110:716–723, 2001[CrossRef][Medline]
22. Blais L, Desgagné A, Lelorier J: 3-Hydroxy-3-methylglutaryl coenzyme A reductase inhibitors and the risk of cancer. Arch Intern Med 160:2363–2368, 2000
23. Goldstein M: Reductase inhibitors and the risk of cancer. Arch Intern Med 161:1460, 2001 24. Lau HS, de Boer A, Beuning KS, et al: Validation of pharmacy records in drug exposure assessment. J Clin Epidemiol 50:619–625, 1997[CrossRef][Medline] 25. Herings R: Pharmo, A Record Linkage System for Postmarketing Surveillance of Prescription Drugs in The Netherlands [thesis in pharmaco-epidemiology and pharmacotherapy]. The Netherlands, Utrecht University, 1993, pp 17–32 26. Von Korff M, Wagner E, Saunders K: A chronic disease score from automated pharmacy data. J Clin Epidemiol 45:197–203, 1992[CrossRef][Medline] 27. Visser O, Coebergh JWW, van Dijck JAAM, et al: Incidence of cancer in the Netherlands 1998. Utrecht, 2002. http://www.ikc.nl/vvik/kankerregistratie/img_kr_cijfersaug02/divers/VIKC-Incidence.pdf
28. Simons M, Keller P, Dichgans J, et al: Cholesterol and Alzheimer's disease: Is there a link? Neurology 57:1089–1093, 2001 29. Jick H, Zornberg GL, Jick SS, et al: Statins and the risk of dementia. Lancet 356:1627–1631, 2000[CrossRef][Medline]
30. Wolozin B, Kellman W, Ruosseau P, et al: Decreased prevalence of Alzheimer disease associated with 3-hydroxy-3-methyglutaryl coenzyme A reductase inhibitors. Arch Neurol 57:1439–1443, 2000 31. Coons JC: Hydroxymethylglutaryl-coenzyme A reductase inhibitors in osteoporosis management. Ann Pharmacother 36:326–330, 2002[Abstract]
32. Adjei A: Blocking oncogenic ras signaling for cancer therapy. J Natl Cancer Inst 93:1062–1074, 2001 33. Negre-Aminou P, van Vliet AK, van Erck M, et al: Inhibition of proliferation of human smooth muscle cells by various HMG-CoA reductase inhibitors: Comparison with other human cell types. Biochim Biochim Biophys Acta 1345:259–268, 1997[Medline] 34. Wang W, Macaulay RJ: Apoptosis of medulloblastoma cells in vitro follows inhibition of farnesylation using manumycin A. Int J Cancer 82:430–434, 1999[CrossRef][Medline] Submitted February 5, 2003; accepted March 23, 2004.
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2004 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2 statistics were used (or the Fisher's exact test whenever the expected cell counts were < 5). Conditional logistic regression was used to estimate the relative magnitude of the association between statin use and the risk of cancer, expressed as odds ratios (ORs) and 95% CIs. The analysis was adjusted for diabetes mellitus; number of hospitalizations; comorbidity; and use of diuretics, ACE inhibitors, calcium-channel blockers, NSAIDs, sex hormones, and other lipid-lowering therapies. Because the time since last exposure might be important, the effect of lag time was assessed; this was defined as the time between last statin use and the index date. All statistical tests were performed two-sided with a rejection of the null hypothesis at a P value of less than .05. All statistical analyses were performed using SPSS 10.0 (SPSS Inc, Chicago, IL). 
