This study demonstrated that statin prescriptions for cancer patients were inversely associated with the risk of all-cause mortality in a dose-response manner. This inverse association between statin prescription and all-cause mortality is similarly observed in both men and women. These findings support the results from our previous study based on the Korean NHIS-HEALS cohort17.
Cancer is the leading cause of death in Korea. With the development of diagnostic and treatment technologies, many cancers are now detected and treated in the early stages. The prevalence and number of patients diagnosed with cancer continue to increase in Korea 1. The Korean NHIS has provided a high-risk candidate national cancer screening program (NCSP) for five types of cancer since 1999: stomach, liver, colorectum, breast, and uterine cervix 18. Beyond the NCSP, many Korean adults also undergo opportunistic check-ups in private sectors. These intensive screenings for cancers increase the probability of early detection and improve survival rate. The 5-year relative survival rates for all cancers diagnosed between 2016 and 2020 were 71.5% in both sexes (65.5% in men and 77.8% in women) and have increased1. As time passes after cancer diagnosis, the attribution of primary cancer to mortality decreases, while deaths attributable to chronic diseases such as CCVDs or second primary cancer increase for long-term cancer survivors19. Thus, statins can play an important role in improving survival rates for cancer patients by preventing CCVDs, inhibiting subclinical inflammation and oxidative stress, and protecting against cancer development and progression7,8,10,13,14.
The exact mechanism by which statins improve survival rates for cancer survivors has not been clearly understood. Statins, acting as inhibitors of HMG-CoA reductase in the mevalonate pathway, block cholesterol biosynthesis and reduce blood cholesterol levels. Through inhibition of the mevalonate pathway, the biosynthesis of isoprenoids is also blocked. Isoprenoids are precursor molecules of farnesyl pyrophosphate and geranylgeranyl pyrophosphate in addition to cholesterol, dolichol, and coenzyme Q1020. Since farnesyl pyrophosphate and geranylgeranyl pyrophosphate are crucial for the membrane anchoring of Ras and Rho proteins, statins directly or indirectly prevent carcinogenesis and invasiveness of cancer. They also inhibit angiogenesis and promote apoptosis of cancer cells by modifying isoprenylation and inhibiting downstream pathways of the mevalonate pathway21. Many epidemiologic studies reported that statins were associated with a low risk of cancer development, recurrence, and metastasis10,12,14,15,22. Statins help to sensitize cancer cells to chemotherapy or radiation therapy and improve treatment response23. Beyond their effects on the mevalonate pathway, statins have pleiotropic functions such as anti-inflammatory, anti-oxidative, and immunomodulatory activities8. Thus, statins play a beneficial role in preventing atherosclerotic cardiovascular diseases, dementia, and renal impairment in addition to anti-cancer effects7,9,11. These various effects of statins may reduce all-cause mortality in cancer patients. However, nearly half of cancer survivors in Korea were not properly prescribed statins despite their eligibility for statin therapy24. Proper prescription of statins and compliance with the medication may improve the survival rate of cancer patients.
Several studies have reported an inverse association between statin use and mortality in cancer survivors25–28, although a few studies demonstrated that statins were not significantly associated with mortality29,30. Our previous study reported that statin use by cancer survivors decreased with mortality17. The presents study further supports the inverse relationship between statin prescription and mortality in cancer patients. The previous study by Kim et al. was designed based on the Korean NHIS-health screening (HEALS) cohort and included only 1,302 patients, while the present study included data of 154,916 patients. In addition, the Korean NHIS-HEALS cohort does not provide code V193 and V194, which indicate confirmed cancer diagnoses. Thus, false positive cancer patients who received a diagnostic code to rule out malignancy based on ICD-10 codes could be included, even though inclusion criteria were strictly applied. However, in the present study, cancer was confirmed histologically or radiologically because cancer was defined based on the combination of special disease codes (V193 and V194) and C codes of ICD-10. Korean authorities monitor and manage cancer patients thoroughly after enrollment using the special disease codes because the Korean NHIS, as a service provider, reimburses almost all medical expenditures (approximately 95% of all hospital bills in case of treatment covered by national insurance) instead of patients. Thus, it is unlikely that false-positive cancer patients will be included. The probability of recall bias or misclassification are extremely low. This is a significant difference from our previous study population10. In addition, overall cancer patients were further stratified into patients with the most common cancers ranked in 2007 by sex according to the Korean Statistical Information Service31. Even after stratifying, similar results were observed in each cancer type, indicating that statins may be useful to improve survival rates in many cancer types.
The present study has several advantages that distinguish it from previous studies. First, the nationwide health insurance claim database from NHID was used to investigate the association between statin prescription and all-cause mortality in cancer patients. The national health insurance corporation (NHIC) is a non-profit organization and the single insurer in Korea that operates the NHIS. Almost all Koreans (approximately 97% of the entire population) are subscribed to the NHIS and others (approximately 3%) in extreme poverty are covered by the Medical Aid 32,33. The Korean NHID includes almost all Korean medical records (medical utilization, medical expenditures, personal information, health check-up results, and death information) from the Medical Aid program as well as the NHIS. Thus, we believe that the study population is representative of the Korean population and can be generalized. Second, the study duration (median follow-up 14.6 years) was relatively long. A longer study duration aids in assessing the long-term effects of statin on all-cause mortality, given that the 5-year relative survival rate in Korea is over 70%. Third, various confounders that may affect all-cause mortality were controlled. Conditions that may affect the mortality were excluded or controlled. Individuals who died within 1 year after enrollment or had cardio- or cerebrovascular diseases before enrollment were excluded in order to minimize bias due to unrelated conditions in this investigation of the long-term effect of statin use. In addition, socioeconomic factors (household income and residential area), which are related to hospital accessibility and health disparity, as well as comorbidities (based on CCI), which correlate with patients’ death or medical resource utilization, were considered. Additionally, laboratory or anthropometric results (BMI, SBP, fasting glucose, total cholesterol, and ALT levels), closely associated with chronic diseases, were intensively adjusted in Cox proportional hazards regression models, in addition to cancer types. These models aim to offset the effects of unexpected confounding factors on mortality.
However, some limitations should be considered when interpreting this study. First, several biases such as immortal time bias or health patient bias could not be completely controlled. In addition, there is a potential for reverse causality; patients who were regularly prescribed statins are more likely to be healthy or concerned about their health than those who are not. In addition, patients who survived longer might have been prescribed more statins. Second, information on histopathology, cancer stage, and anti-cancer treatment details (surgery, chemotherapy, or radiation therapy) was not available. This information is collected by the National Cancer Center (NCC) and provided to only a small number of researches affiliated with the NCC. In addition, it cannot be merged with the NHID because all data have been deidentified. If detailed information were provided, more potential confounders could be controlled. Third, we did not distinguish hydrophilic and lipophilic statins. Lipophilic statins had a more beneficial impact on reducing the incidence or mortality of cancer than hydrophilic statins22,34,35. However, in Korea, changes in statin prescriptions by medical doctors occur very frequently, making it difficult to design a study that takes advantage of statins’ characteristics. Fourth, there is the possibility of selection bias. Of the initial 820,078 candidates who were diagnosed with any malignant neoplasm between 2007 and 2008, only 154,709 participants were analyzed. To improve sensitivity, further statistical analyses were performed after excluding individuals who were diagnosed with any cancer before enrollment or who did not participate in a national health check-up within one year after enrollment. The results were similar to those of the original analyses. This indicates that selection bias had little effect on the main results.