Cancer Treatment Reviews xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Cancer Treatment Reviews journal homepage: www.elsevierhealth.com/journals/ctrv
Systematic or Meta-analysis Studies
Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies Shanliang Zhong a,1, Xiaohui Zhang a,1, Lin Chen b, Tengfei Ma a, Jinhai Tang c, Jianhua Zhao a,⇑ a
Center of Clinical Laboratory Science, Jiangsu Cancer Hospital Affiliated to Nanjing Medical University, Nanjing 210009, China Departments of Oncology, Xuzhou Medical College, Xuzhou 221004, China c Department of General Surgery, Jiangsu Cancer Hospital Affiliated to Nanjing Medical University, Nanjing 210009, China b
a r t i c l e
i n f o
Article history: Received 24 January 2015 Received in revised form 5 April 2015 Accepted 7 April 2015 Available online xxxx Keywords: Prognosis Survival Outcome Tumor Neoplasm
a b s t r a c t Background: Previous studies have examined the effect of statin use on the mortality in cancer patients, but the results are inconsistent. A meta-analysis was performed to assess the association with all available studies. Methods: Relevant studies were identified by searching PubMed and EMBASE to April 2015. We calculated the summary hazard ratios (HRs) and 95% confidence intervals (CIs) using random-effects models. We estimated combined HRs associated with defined increments of statin use, using random-effects meta-analysis and dose–response meta-regression models. Results: Thirty-nine cohort studies and two case-control studies involving 990,649 participants were included. The results showed that patients who used statins after diagnosis had a HR of 0.81 (95% CI: 0.72–0.91) for all-cause mortality compared to non-users. Those who used statin after diagnosis (vs. non-users) had a HR of 0.77 (95% CI: 0.66–0.88) for cancer-specific mortality. Prediagnostic exposure to statin was associated with both all-cause mortality (HR = 0.79, 95% CI: 0.74–0.85) and cancer-specific mortality (HR = 0.69, 95% CI: 0.60–0.79). Stratifying by cancer type, the three largest cancer-type subgroups were colorectal, prostate and breast cancer and all showed a benefit from statin use. HRs per 365 defined daily doses increment were 0.80 (95% CI: 0.69–0.92) for all-cause mortality and 0.77 (95% CI: 0.67–0.89) for cancer-specific mortality. A 1 year increment in duration only conferred a borderline decreased risk of death. Conclusions: In conclusion, the average effect of statin use, both postdiagnosis and prediagnosis, is beneficial for overall survival and cancer-specific survival. Ó 2015 Elsevier Ltd. All rights reserved.
Introduction Cancer is a very serious health problem worldwide, and is the leading cause of death in economically developed countries and the second leading cause of death in developing countries [1]. Considering the different causes, the different tissues affected, and the different symptoms, cancer is a very complex and still incurable disease. Although much effort has been directed at comprehending carcinogenesis and a lot of progress has been achieved, there is still no effective treatment for most cancers. Recently, the potential anticancer properties of statins have attracted more interest. Statins, among the most frequently ⇑ Corresponding author at: Center of Clinical Laboratory Science, Jiangsu Cancer Hospital Affiliated to Nanjing Medical University, Baiziting 42, Nanjing 210009, China. Tel.: +86 25 83283334; fax: +86 25 83351406. E-mail addresses:
[email protected],
[email protected] (J. Zhao). 1 Equal contributors.
prescribed drugs worldwide, reduce serum cholesterol and prevent cardiovascular diseases [2]. They block 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, which inhibits the conversion of HMG-CoA to the cholesterol precursor mevalonate, the rate limiting step in cholesterol synthesis [3]. Statins may exert their anticancer effect via lowering protein prenylation [4], reducing tumor cell proliferation and migration [5,6], inhibiting of rat sarcoma (Ras) signaling [7], inducing apoptosis through phosphorylation of Akt and down-regulation of mammalian target of rapamycin (mTOR) [8], and other pleiotropic effects on the cellular level. In the last decade, a number of observational studies have tried to examine the effect of statin use on outcome in patients with several cancer types including breast [5,9–12], prostate [13–18], ovarian[19–21], lymphoma [22,23], renal cell carcinoma [8,24,25] and colorectal cancer [4,26–32] et al.; some have suggested that statin use was associated with longer survival, while others report no benefit. To date, no meta-analysis has been conducted concerning
http://dx.doi.org/10.1016/j.ctrv.2015.04.005 0305-7372/Ó 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
2
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx
the therapeutic value of statins on the survival of cancer patients. Therefore, we performed a meta-analysis with all available studies to explore the association between pre- and post-diagnosis statin use and the survival of cancer patients, for both cancer-specific mortality and all-cause mortality. Besides, we also performed a dose–response analysis to further evaluate the potential dose–response relation.
Material and methods Literature search We searched PubMed (from 1981 to present) and Embase (from 1991 to present) using the following terms (‘‘Statin’’ or ‘‘Atorvastatin’’ or ‘‘Cerivastatin’’ or ‘‘Compactin’’ or ‘‘Fluvastatin’’ or ‘‘HMG-CoA’’ or ‘‘Lovastatin’’ or ‘‘Mevastatin’’ or ‘‘Pravastatin’’ or ‘‘Rosuvastatin’’ or ‘‘Rosvastatin’’ or ‘‘Simvastatin’’) and (‘‘mortality’’ or ‘‘survival’’) and ‘‘cancer’’. The latest date of this search was April 2015. All cohort or case-control studies evaluating the association between statin use and mortality in cancer patients were eligible, without language restriction. Reference lists of every article retrieved and relevant reviews were examined manually to further identify potentially relevant studies. All searches were conducted independently by two reviewers; differences were checked by the two and resolved by discussion. When two or more studies presented possible overlap, the one with largest populations was included.
Inclusion criteria All the studies were included in this meta-analysis if they met the following criteria: (a) the exposure of interest was statin use assessed before or after diagnosis; (b) The study design was casecontrol or cohort; (c) the outcomes of interest were all-cause mortality or cancer-specific mortality; (d) the follow-up period was longer than 1 year; and (e) risk estimates of mortality and 95% confidence intervals (CIs) were reported (or information to calculate them).
Data extraction Data were extracted from the eligible articles by two independent investigators. The extracted data included the last name of first author, year of publication, origin of the study, follow-up period, sample size, study design, patient characteristics, statin use, risk estimates and corresponding 95% CIs, and covariates adjusted for in the multivariable analysis. If risk estimate and corresponding 95% CI were not available [10,33,34], the data were calculated using curve method described by Tierney [35]. For studies provided more than one risk estimate, we extracted the one that was adjusted for the greatest number of confounding factors. Discrepancies were resolved by consensus, involving a third investigator. Study quality assessment The methodological quality of the studies included in present meta-analysis was independently assessed using the nine-star Newcastle Ottawa scale (NOS) [36] by two investigators. Each study was evaluated based on eight items, categorized into three broad perspectives including selection, comparability, and outcome for cohort studies or exposure for case-control studies. We considered studies with a score of 7 or greater as high quality.
Discrepancies were resolved by discussion or through consultation with a third investigator. Statistical methods Because outcomes were relatively rare, the odds ratios (ORs) and relative risks (RRs) were considered approximations of hazard ratios (HRs). Summary estimates of HR and 95% CIs were obtained using a random effects model where the restricted maximum likelihood estimator was used to evaluate the inter-study heterogeneity [37,38]. Prediction interval (PI) of summary estimate for the random effects model was calculated to depict the uncertainty around the estimate [39]. If studies did not report a summary risk estimate for statin use, a summary risk estimate was calculated using risk estimates for each of the statin use categories [9]. For a study provided risk estimates for cancer-specific deaths and other-cause deaths, risk estimates for all-cause mortality were calculated firstly [14]. If studies provided separate risk estimates by statin type [11], tumor stage [9] or treatment [16] without a summary risk estimate, we treated them as different studies. Interstudy heterogeneity was estimated using a chisquare-based Q-test [40], with a P value of 0.21) and Egger’s tests (P > 0.44). Association of prediagnosis statin use with mortality
Fig. 1. Flow chart of the selection of publications included in the meta-analysis.
studies reported on the association between postdiagnosis statin use and all-cause and/or cancer-specific mortality and eighteen studies reported on the association between prediagnosis statin use and all-cause and/or cancer-specific mortality, with six studies having data on both postdiagnosis and prediagnosis statin use (Table 1). Table S1 summarizes the methodological quality of all the included studies. The NOS results showed that the average overall score was 6.0 (range 3–9) and there were eighteen studies with a score of 7 or more. Association of postdiagnosis statin use with mortality Postdiagnostic statin use was defined as any use of statins after cancer diagnosis. Risk estimates of the association between postdiagnosis statin use and mortality in cancer patients are shown in Figs. 2 and 3. When the association of statin use on all-cause mortality was analyzed as users vs. non-users, a HR of 0.81 (95% CI: 0.72–0.91, P < 0.01; 95% PI: 0.49–1.32) was found. In stratified by cancer type, statin use was associated with decreased risk of death from any cause for prostate cancer patients (HR = 0.59, 95% CI: 0.35–0.99, P < 0.05; 95% PI: 0.22–1.55) and ovarian cancer patients (HR = 0.39, 95% CI: 0.22–0.71, P < 0.01) (Fig. 2). In the subgroup analysis by gender, postdiagnostic statin use showed a significant benefit for overall survival in both male (HR = 0.66, 95% CI: 0.47–0.95, P = 0.02; 95% PI: 0.31–1.44) and female (HR = 0.78, 95% CI: 0.62–0.98, P = 0.03; 95% PI: 0.98–1.52) patients (Fig. S1). In terms of cancer-specific mortality, statin use was significantly associated with reduced risk of cancer-specific mortality (HR = 0.77, 95% CI: 0.66–0.88, P < 0.01; 95% PI: 0.49–1.21). When stratified by cancer type, a decreased risk of death from cancer was found in breast (HR = 0.60, 95% CI: 0.41–0.88, P < 0.01; 95% PI: 0.30–1.21), and prostate (HR = 0.77, 95% CI: 0.70–0.85, P < 0.01; 95% PI: 0.70–0.85) cancer patients who took statins after cancer diagnosis (Fig. 3). Further analysis by gender indicated that there was no association between postdiagnosis statin use and cancer-specific mortality in females (HR = 0.81, 95% CI: 0.56– 1.18, P = 0.27; 95% PI: 0.33–1.98) but among males there was a significant association (HR = 0.75, 95% CI: 0.69–0.82, P < 0.01; 95% PI: 0.69–0.82) (Fig. S1).
Prediagnosis statin use was defined as having received statins prior to cancer diagnosis. Fig. 4 presents the estimated HRs of cancer patients with prediagnosis statin use. The results revealed that patients who took statins prior to diagnosis had a HR of 0.79 (95% CI: 0.74–0.85, P < 0.01; 95% PI: 0.70–0.90) for all-cause mortality. When stratifying by cancer type, a reduced risk of death from any cause was found in breast (HR = 0.73, 95% CI: 0.62–0.86, P < 0.01; 95% PI: 0.57–0.94) and colorectal (HR = 0.77, 95% CI: 0.66–0.89, P < 0.01; 95% PI: 0.62–0.95) cancer patients (Fig. 4). In the subgroup analysis by gender, prediagnosis statin use showed a significant benefit for overall survival in both male (HR = 0.76, 95% CI: 0.65–0.87, P < 0.01; 95% PI: 0.62–0.92) and female (HR = 0.79, 95% CI: 0.73–0.85, P < 0.01; 95% PI: 0.73–0.85) patients (Fig. S1). Regarding the cancer-specific mortality, prediagnosis statin use was associated with a significant benificial effect on the risk of cancer-specific mortality (HR = 0.69, 95% CI: 0.60–0.79, P < 0.01; 95% PI: 0.45–1.06). In the subgroup analysis by cancer type, a decreased risk of death from cancer was found in breast (HR = 0.73, 95% CI: 0.61–0.89, P < 0.01; 95% PI: 0.56–0.97), colorectal (HR = 0.82, 95% CI: 0.73–0.91, P < 0.01; 95% PI: 0.70–0.95), and prostate (HR = 0.44, 95% CI: 0.20–0.93, P < 0.01; 95% PI: 0.11–1.74) cancer patients (Fig. 4). When stratified by gender, prediagnosis statin use showed a significant benefit for cancer-specific survival in both males (HR = 0.59, 95% CI: 0.40–0.87, P < 0.01; 95% PI: 0.25–1.43) and females (HR = 0.79, 95% CI: 0.72–0.87, P < 0.01; 95% PI: 0.72–0.87) (Fig. S1). There was significant heterogeneity between studies in cancerspecific mortality analysis (I2 = 88.12%, P < 0.01). When stratified by cancer type, the heterogeneity was still significant for prostate cancer (I2 = 86.34%, P < 0.01; Fig. 4). From the results of the leaveone-out sensitivity analysis, all the results above were not materially altered (data not shown). We found some evidence of publication bias in any analyses using Begg’s (P > 0.05) and Egger’s tests (P 6 0.01). After a correction for potential publication bias using the trim-and-fill method, the main findings for all-cause mortality (HR = 0.82, 95% CI: 0.77–0.88, P < 0.01) and cancer-specific mortality (HR = 0.83, 95% CI: 0.69–0.99, P = 0.04) remained significant. Dose–response meta-analysis The dose–response effects of postdiagnosis statin use on mortality were assessed with eight studies [5–7,9,13,26,28,32] including five studies [5,9,13,26,32] for cancer-specific mortality and eight studies [5–7,9,13,26,28,32] for all-cause mortality. Among these studies, exposure to statins was expressed in units of defined daily dose (DDD) [5,9,13,26] or duration of exposure [6,7,9,13,28,32]. An increment of 365 DDDs was associated with a 20% lower all-cause mortality (HR = 0.80, 95% CI: 0.69–0.92, P < 0.01) and a 23% lower cancer-specific mortality (HR = 0.77, 95% CI: 0.67–0.89, P < 0.01) (Fig. 5). A 1 year increment in duration of exposure to statins conferred a HR of 0.93 (95% CI: 0.86–1.01, P = 0.09) for all-cause mortality and 0.90 (95% CI: 0.81–1.00, P < 0.05) for cancer-specific mortality. Significant heterogeneity was found for all of the dose–response meta-analyses (Fig. 5).
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
4
First author Year
Country
Follow-up period
Study design
Patient characteristics
Statin use
Cardwell et al. (2015) [5]
UK
1998–2013 Average 5.7 years (Range 1–14)
Cohort
17,880 stage I–IV breast patients
Prediagnosis Non-user User Postdiagnosis Non-user User
Desai et al. (2015) [12]
USA
Hoffmeister et al. (2015) [32]
Germany
Shao et al. (2015) [31]
China
Viers et al. (2015) [25]
USA
Song et al. (2015) [34]
Korea
Yoon et al. (2015) [65]
USA
Caon et al. (2014) [17]
Canada
Cardwell et al. (2014) [26]
UK
1992–2010 Average 11.5 years (Standard deviation = 3.7)
Cohort
2003–2009 Median 3.4 years (IQR 2.3–5.0)
Cohort
2004–2008 Median 4.2 years
Cohort
1995–2009 Median 7.8 years (IQR 5.3–11.2)
Cohort
2007–2013 Median 38.6 months
Cohort
2007–2010
Cohort
2000–2007 Median 8.4 years
Cohort
1998–2012 Mean 5 years (Range 1–14)
Cohort
Cancer-specific mortality
All cause mortality
HR 1.00 0.81 (0.71–0.93)
HR 1.00 0.78 (0.70–0.86)
HR 1.00 0.84 (0.68–1.04)
HR 1.00 0.84 (0.72–0.97)
7883 postmenopausal women with in situ, local, regional and distant-stage breast cancer
Prediagnosis No Yes
HR 1.00 0.59 (0.32–1.06)
2697 Stage I–IV CRC patients older than 30 years
Postdiagnosis No statin use Any statin use
HR 1.00 1.11 (0.82–1.50)
HR 1.00 1.10 (0.85–1.41)
17,115 stage I, II, or III CRC patients who received curative surgery
Prediagnosis Non-user User
HR 1.00 0.77 (0.68–0.88)
HR 1.00 0.82 (0.74–0.92)
2357 patients treated with radical or partial nephrectomy for pNx/0, M0 RCC
Postdiagnosis No Yes
HR 1.00 1.02 (0.74–1.39)
HR 1.00 0.84 (0.69–1.00)
409 patients with de novo DLBCL who received RCHOP therapy
Postdiagnosis Non-user User
HR 1.00 1.23 (0.89–1.72)
2987 stage I–IV endometrial cancer patients aged 65 years or older who received a hysterectomy
Postdiagnosis Non-user User Non-user User
HR 1.00 0.92 (0.70–1.20) HR 1.00 0.92 (0.65–1.29)
3851 prostate cancer patients treated with external beam radiation therapy ± androgen deprivation therapy
Postdiagnosis No Yes
7657 stage I to III CRC patients
Prediagnosis Non-user User Postdiagnosis Non-user User
Covariate adjustment Year of diagnosis, age at diagnosis, surgery within 6 months, radiotherapy within 6 months, chemotherapy within 6 months, hormone therapy use, comorbidities, other medication usage stage and socioeconomic deprivation in women with nonmissing values Race, education, smoking, BMI, waist circumference, mammogram in the past 2 years, Gail 5-year risk, female relative with breast cancer, age at menarche, number of live births, breast biopsy, hysterectomy, hormone use, oral contraceptive, aspirin use and study component Age at diagnosis, sex, Union Internationale Contre le Cancer stage, location of CRC, surgery, neoadjuvant treatment, chemotherapy, radiotherapy, BMI, lifetime pack-years of active smoking, average lifetime physical activity, ever regular use of NSAIDs, ever use of hormone replacement therapy (women), previous large bowel endoscopy, diabetes, hyperlipidemia, myocardial infarction, stroke, heart failure, participation in general health check-ups, and for a time-dependent effect of chemotherapy Age, sex, tumor stage, adjuvant therapy, and the propensity score were adjusted Age at surgery, sex, race, type of surgery, Charlson comorbidity index, primary tumor classification, preoperative estimated glomerular filtration rate, symptoms, and year of surgery Not state
Age, acetylsalicylic acid, year of treatment, radiation dose, androgen deprivation therapy, initial pretreatment PSA, T–stage, Charlson index, Gleason score
HR 1.00 0.769 (0.548– 1.08)
HR 1.00 0.86 (0.79–0.93) HR 1.00 0.71 (0.61–0.84)
Follow-up time, age, race, neighborhood income, cancer stage, tumor grade, hysterectomy type, chemotherapy, radiation, impaired glucose tolerance, obesity, dyslipidemia and diabetes
HR 1.00 0.75 (0.66–0.84)
Year of diagnosis, age at diagnosis, sex, stage, surgery within 6 months, radiotherapy within 6 months, chemotherapy within 6 months, site, comorbidities, other medication use after diagnosis as time-varying covariates, grade, deprivation, and smoking before diagnosis in individuals without missing values
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
Table 1 Characteristic of the included studies.
First author Year
Country
Follow-up period
Study design
Patient characteristics
Statin use
Gaist et al. (2014) [58]
Denmark
2000–2011 Median 6.9 months (IQR 3.8–13.4)
Cohort
1562 GBM patients aged between 20 and 85 years
Prediagnosis Unexposed Exposed
Grytli et al. (2014) [18]
Norway
2004–2011Median 39 months
Cohort
3,699 prostate cancer patients with high-risk or metastatic disease
Prediagnosis Non-user User Postdiagnosis Non-user User
Habis et al. (2014) [19]
USA
Hamilton et al. (2014) [24]
USA
Kaffenberger et al. (2014) [8]
USA
Krens et al. (2014) [4]
The Netherlands
Livingstone et al. (2014) [6]
The Netherlands
Lipworth et al. (2014) [61]
USA
Murtola et al. (2014) [9]
Finland
1992–2013
Cohort
1995–2010 Median 3 years
Cohort
2000–2010 Median 42.5 months (IQR 19.1–67.1)
Cohort
2005–
Cohort
1998–2010 3 years
Cohort
2002–2010 Median 5.6 years
Cohort
1995–2003 3.25 years
Cohort
442 stage I–IV epithelial ovarian patients
Postdiagnosis No Yes
2,608 patients with localized RCC who were treated surgically
Postdiagnosis No Yes
916 patients who underwent radical or partial nephrectomy for RCC
Postdiagnosis Non-user User
529 KRAS mutant metastatic CRC patients treated with capecitabine, oxaliplatin bevacizumab ± cetuximab
Postdiagnosis Non-user User
791 cutaneous melanoma patients (Breslow thickness > 1 mm) aged 18 years or older
Prediagnosis Non-user User Postdiagnosis Non-user User
86,000 adult participants aged 40–79 years
Prediagnosis No Yes
31,236 breast cancer patients of all stages
Prediagnosis Localized 0 1–495 DDD P496 DDD Metastatic 0 1–495 DDD P496 DDD Postdiagnosis Localized
Cancer-specific mortality
All cause mortality
HR 1.00 0.79 (0.63–1.00)
Covariate adjustment
Year of diagnosis, age at diagnosis, gender, Charlson Comorbidity Index score, history of diabetes, hospital contact for allergy or asthma, number of years of schooling, and use of anti-asthma drugs, antihistamines, HRT, low-dose aspirin, and nonaspirin NSAIDs Age, prostate-specific antigen level, Gleason score, clinical T stage, presence and type of metastases, performance status, and androgen-deprivation therapy initiated within 6 months after diagnosis
HR 1.00 0.78 (0.67–0.90) HR 1.00 0.78 (0.67–0.90)
Age at diagnosis, race, BMI, ASA class, metformin use, residual tumor > 1 cm, primary cytoreductive surgery, histologic subtype, stage, tumor site and grade
HR 1.00 1.48 (0.68–3.22)
HR 1.00 0.89 (0.71–1.13)
HR 1.00 0.48 (0.28–0.83)
HR 1.00 0.62 (0.43–0.90)
HR 1.00 1.41 (0.95–2.10)
Demographic, clinical, preoperative, age, gender, black race, Charlson score, glomular filtration rate, surgery year, symptom presentation, and T stage Age, ASA score, pT category, Fuhrman grade, node status, metastatic status, ABO blood group, preoperative anemia, and preoperative hypercalcemia Prior adjuvant therapy, number of affected organs, age, and aspirin use
Age and sex HR 1.00 0.88 (0.58–1.34) HR 1.00 0.76 (0.50–1.61)
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx
Age, year of Southern Community Cohort Study enrollment, marital status, education, income, health insurance, BMI, cigarette smoking, alcohol consumption, history of hypertension, myocardial infarction/coronary artery bypass surgery, diabetes, stroke, race and sex
HR 1.00 0.90 (0.72–1.12)
Age, tumor morphology and treatment selection HR 1.00 0.76 (0.58–0.99) 0.44 (0.28–0.67)
HR 1.00 0.69 (0.56–0.84) 0.51 (0.39–0.68)
HR 1.00 0.79 (0.51–1.20) 0.51 (0.30–0.86)
HR 1.00 0.91 (0.63–1.31) 0.58 (0.37–0.92) Age, tumor stage and morphology, treatment selection and pre–diagnostic statin use 5
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
Table 1 (continued)
(continued on next page)
6
First author Year
Country
Nam et al. (2014) [7]
Korea
Yu et al. (2014) [13]
UK
Brewer et al. (2013) [11]
USA
Crivelli et al. (2013) [56]
3 institutions
Geybels et al. (2013) [14]
USA
Follow-up period
2006–2012
Study design
Cohort
1998–2012 Mean 4.4 years
Cohort
1995–2011 Median 2.9 years
Cohort
1996–2007 Median 62.7 months (IQR 25.0–110.7)
Cohort
2002–2011 Average 6.1 years
Cohort
Patient characteristics
Statin use
Cancer-specific mortality
All cause mortality
0 10–322 DDD 333–800 DDD P801 DDD Metastatic 0 10–322 DDD 333–800 DDD P801 DDD
HR 1.00 0.54 (0.40–0.72) 0.43 (0.31–0.61) 0.42 (0.28–0.62)
HR 1.00 0.56 (0.45–0.69) 0.48 (0.38–0.61) 0.37 (0.27–0.50)
HR 1.00 0.66 (0.42–1.01) 0.37 (0.18–0.79) 0.24 (0.03–1.74)
HR 1.00 0.73 (0.49–1.08) 0.41 (0.21–0.80) 0.38 (0.09–1.53)
241 stage II and III gastric cancer patients who underwent radical gastrectomy
Postdiagnosis Non-user User
11,772 men newly diagnosed with nonmetastatic prostate cancer
Postdiagnosis No Yes
723 patients diagnosed with primary inflammatory breast cancer
Postdiagnosis L-statin users No Yes H-statin users No Yes
HR 1.00 1.169 (0.570– 2.397)
HR 1.00 0.76 (0.66–0.88)
HR 1.00 0.86 (0.78–0.95)
HR 1.00 1.46 (0.73–2.90) HR 1.00 0.80 (0.43–1.49)
1117 patients treated with transurethral resection of the bladder for NMIBC
Postdiagnosis No Yes
HR 1.00 1.23 (0.69–2.19)
HR 1.00 1.14 (0.89–1.44)
1001 prostate cancer patients of all stages
Prediagnosis No Yes
HR 1.00 0.19 (0.06–0.56)
HR 1.00 0.45 (0.10–2.01)
Covariate adjustment
Age group, sex, Lauren’s classification, tumor invasion, lymph node invasion, tumor differentiation, adjuvant chemotherapy, Charlson comorbidity index, lymphovascular invasion, perineural invasion and tumor size Age, year of prostate cancer diagnosis, ethnicity, excessive alcohol use, smoking status, obesity, chronic kidney disease, myocardial infarction, ischemic stroke, transient ischemic attack, peripheral artery disease, previous cancers, prostate-specific antigen level, Gleason score, metformin, sulfonylureas, thiazolidinediones, insulins, other oral antihypoglycemic agents, angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, calcium channel blockers, b-blockers, diuretics, other antihypertensive drugs, aspirin, other nonsteroidal anti-inflammatory drugs, a-reductase inhibitors, prediagnostic statin use, prostate-specific antigen testing activity, prostatectomy, radiation therapy, chemotherapy, and androgen deprivation therapy Radiation therapy, hormonal receptor status and HER2 status as the stratification factors and adjusted for lymphatic/vascular invasion for progression-free survival and lymphatic/vascular invasion, nuclear grade and surgery within 1 year for overall survival and disease-specific survival None
Age at diagnosis, Gleason score, stage at diagnosis, diagnostic PSA level, primary treatment approach, race, first-degree family history of prostate cancer, BMI, smoking status, lifetime alcohol consumption, aspirin use, non-aspirin NSAID use, history of diabetes mellitus, and history of prostate cancer screening
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
Table 1 (continued)
First author Year
Country
Follow-up period
Study design
Patient characteristics
Statin use
Lavie et al. (2013) [20]
Israel
2003–>2010 9 years
Cohort
150 ovarian cancer cases
Postdiagnosis No Yes Postdiagnosis No Yes
274 endometrial cancer cases Mace et al. (2013) [30]
USA
2000–2012 5 years
Cohort
Xylinas et al. (2013) [64]
11 international institutions
1987–2007 Median 45 months (IQR 20–81)
Cohort
Eindhoven et al. (2012) [57]
The Netherlands
2000–2005 Median 5 years (Range 3–9)
Cohort
Lakha et al. (2012) [27]
UK
1999–2009
Cohort
Cancer-specific mortality
All cause mortality
Covariate adjustment Age
HR 1.00 0.24 (0.06–0.78) HR 1.00 0.35 (0.12–0.96)
407 patients with primary rectal adenocarcinoma who received neoadjuvant therapy then proctectomy
Postdiagnosis Non-user User
HR 1.00 0.62 (0.32–1.18)
HR 1.00 0.81 (0.49–1.35)
2490 patients with UTUC treated with radical nephroureterectomy
Postdiagnosis No Yes
HR 1.00 0.86 (0.72–1.03)
HR 1.00 0.83 (0.72–0.95)
5647 patients who underwent percutaneous coronary intervention
Prediagnosis Non-user User
HR 1.00 0.48 (0.34–0.67)
309 CRC patients aged 16– 79 years
Prediagnosis No Yes Postdiagnosis No Yes
Age, BMI, ASA class III/IV (relative to I/II), and pathological stage
Standard clinicopathologic factors
Age, sex, indication, prior MI, prior PCI, prior CABG, diabetes, hypertension, current smoking, family history of coronary disease, multivessel disease and the use of beta blockers, angiotensin-converting enzyme inhibitors, calcium antagonists, nitrates, diuretics, digitalis and anticoagulants
HR 1.00 0.60 (0.327–1.32)
HR 1.00 0.59 (0.28–1.24)
HR 1.00 0.54 (0.19–1.50)
HR 1.00 0.61 (0.26–1.41)
Marcella et al. (2012) [15]
USA
1997–2000
Casecontrol
380 cases aged 55–79 who died from prostate cancer; 380 matched controls
Prediagnosis No Yes
0.37 (0.23–0.60)
Nielsen et al. (2012) [62]
Denmark
1995–2009 Median 2.6 years (Range 0–15)
Cohort
295,925 patients aged 40 or older who received a diagnosis of cancer
Prediagnosis No Yes
HR 1.00 0.85 (0.82–0.87)
Haukka et al. (2012) [60]
Finland
1997–2005 Mean 4.4 years (IQR 1.8–6.6)
Cohort
336,618 pairs of individuals
Prediagnosis No statin Statin
RR 1.00 0.53 (0.43–0.64)
Chae et al. (2011) [10]
USA
1999–2008 Median 55 months; maximal 118 months
Cohort
Ng et al. (2011) [28]
USA
1999–2009 Median 6.5 (Range 4.4– 7.3)
Cohort
Age, sex, region of residence, family history of CRC, past medical history of cancer, past medical history of bowel disease, BMI, smoking, physical activity, and regular NSAID intake
Education level, body mass index, waist size, number of comorbidities and matched for race and age, and antihypertensive medication use HR 1.00 0.85 (0.83–0.87)
Age at diagnosis; cancer stage; status with regard to chemotherapy, radiotherapy, diagnosis of cardiovascular disease before cancer, and diagnosis of diabetes mellitus before cancer; year of birth; sex; race and ethnic descent; highest level of education; and size of residential area
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx
Baseline diseases, sex, age, follow-up time, statin group, and interaction term between statin and follow-up time
703 stage II/III breast cancer patients
Postdiagnosis Non-user User
HR 1.00 0.98 (0.56–1.70)
842 stage III colon cancer patients with regional lymph node metastases but no distant metastases who received surgical resection then adjuvant bolus 5fluorouracil or leucovorin
Postdiagnosis No Yes
HR 1.00 1.16 (0.77–1.72)
Age, race, menopausal status, family and smoking history, diabetes, hormonal receptor, and Her–2/ Neu status, as well as hormonal therapy Age, sex, family history of colorectal cancer, baseline performance status, depth of invasion through bowel wall, number of positive lymph nodes, perineural invasion, extravascular invasion, postoperative carcinoembryonic antigen, treatment arm, body mass index, physical activity, Western pattern diet, consistent aspirin use, and KRAS mutation status (continued on next page)
7
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
Table 1 (continued)
8
First author Year
Country
Follow-up period
Study design
Patient characteristics
Statin use
Katz et al. (2010) [16]
USA
1990–2006 Median 4 years (Range 0–16)
Cohort
7042 men who underwent radical prostatectomy or radiotherapy for prostate cancer
Postdiagnosis RP No Yes RT No Yes
Nowakowski et al. (2010) [22]
USA
Samaras et al. (2010) [23]
Switzerland
Gardette et al. (2009) [59]
France
Siddiqui et al. (2009) [29]
USA
van Gestel et al. (2009) [63]
The Netherlands
Elmore et al. (2008) [21]
USA
Graf et al. (2008) [33]
Germany
2002–2007 Median 47 months (Range 13–80)
Cohort
2001–2009 Median 24.4 months (Range 2.27–98)
Cohort
1991– Mean 9.6 years
Cohort
1997–2003 5 years
Cancer-specific mortality
All cause mortality
Age, race, BMI, comorbid illness, smoking status at diagnosis, and androgen-deprivation therapy HR 1.00 0.35 (0.21–0.58) HR 1.00 0.59 (0.37–0.94)
228 DLBCL patients aged age 18 years or older
Postdiagnosis No Yes
HR 1.00 0.89 (0.42–1.90)
145 DLBCL patients who received R-CHOP as firstline therapy
Postdiagnosis No Yes
HR 1.00 0.58 (0.07–4.55)
7855 men aged 50–59 years at baseline
Prediagnosis No Yes
Casecontrol
1309 male patients with a new diagnosis of CRC
Prediagnosis Non-user User
1990–2006 Median 5 years (IQR 2.0–9.1)
Cohort
3371 patients with peripheral arterial disease who underwent vascular surgery
Prediagnosis No Yes
1996–2001
Cohort
126 patients with advanced stage (III/IV) epithelial ovarian cancer undergoing primary cytoreductive surgery followed by at least 6 cycles of platinum-based chemotherapy
Postdiagnosis Non-user User
RR 1.00 0.45 (0.23–0.88)
183 patients with early HCC (1 nodule < 5 cm or 3 nodules < 3 cm each) who received transarterial chemoembolization
Postdiagnosis No Yes
HR 1.00 0.66 (0.45–0.99)
2003–2008
Cohort
Covariate adjustment
International prognostic index
Not state
Center, age, educational level, histories of cardiovascular and severe chronic diseases, hypertension, and diabetes, smoking habits, alcohol consumption, physical activity, waist circumference, and HDL and non-HDL cholesterol
HR 1.00 0.41 (0.16–1.06)
Age and NSAIDs use OR 1.00 0.7 (0.6–0.9) Age, gender, type of surgery, diabetes, smoking, hypercholestrolaemia, corticosteroids, aspirin and propensity score
HR 1.00 0.82 (0.57–1.20
Age, stage, grade, and suboptimal cytoreduction
Not state
IQR, interquartile range; RCC, renal cell carcinoma; GBM, glioblastoma multiforme; CRC, colorectal cancer; DLBCL, diffuse large B cell lymphoma; NMIBC, non-muscle-invasive bladder cancer; UTUC, upper tract urothelial carcinoma; HCC, hepatocellular carcinoma; R-CHOP, rituximab-cyclophosphamide, doxorubicin, vincristine, prednisone; DDD, defined daily dose; L-statin, lipophilic statin; H-statin, weakly lipophilic to hydrophilic statins; RP, radical prostatectomy; RT, radiotherapy; HR, hazard ratio; OR, odds ratio; RR, relative risk; NSAID, non-steroidal anti-inflammatory drug; BMI, body mass index.
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
Table 1 (continued)
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx
9
Fig. 2. Pooled analyses and subgroup analyses by cancer type for the association between post-diagnostic statin use and all-cause mortality. The squares and horizontal lines correspond to the study-specific hazard ratio and 95% CIs. The area of the square is proportional to the inverse of the sum of the between studies variance and the studyspecific variance. The diamond represents the pooled multivariate hazard ratio and 95% CI. The dotted line of the diamond indicates the bounds of the 95% prediction interval.
Discussion The present meta-analysis investigated the relationship between statin use and mortality in cancer involving 947,410 participants for prediagnosis statin use and 104,811 participants for postdiagnosis statin use with cancer survival outcomes. The summary results, as derived from thirty-nine cohort studies and two case-control studies, indicated that the average effect of statin use post- or prediagnosis was beneficial for overall survival and cancer-specific survival. However, statin use after diagnosis may not always be beneficial in an individual setting (95% PI: 0.49– 1.32 for all-cause mortality; and 95% PI: 0.49–1.21 for cancerspecific mortality) considering the significant heterogeneity between studies. With respect to the effect of prediagnosis statin
use on cancer-specific mortality, statin use will be beneficial in most settings (95% PI: 0.45–1.06), although there was significant heterogeneity. Five studies have explored the effect of statin use on survival of breast cancer patients [5,9–12]. The two most recent studies [5,9] reported a beneficial effect on survival of breast cancer patients who used statin before or after diagnosis. Although the other three studies showed no effect, Chae et al., found that breast cancer patients who took statins more than 6 months have a reduced risk of breast cancer recurrence [10], Brewer et al. showed that weakly lipophilic to hydrophilic statins were associated with significantly improved progression-free survival compared with no statin [11], and Desai et al. reported that prior statin use was associated with lower breast cancer stage at diagnosis [12].
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
10
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx
Fig. 3. Pooled analyses and subgroup analyses by cancer type for the association between post-diagnostic statin use and cancer-specific mortality. The squares and horizontal lines correspond to the study-specific hazard ratio and 95% CIs. The area of the square is proportional to the inverse of the sum of the between studies variance and the studyspecific variance. The diamond represents the pooled multivariate hazard ratio and 95% CI. The dotted line of the diamond indicates the bounds of the 95% prediction interval.
Fig. 4. Pooled analyses and subgroup analyses by cancer type for the association between prediagnostic statin use and all-cause and cancer-specific mortality. The squares and horizontal lines correspond to the study-specific hazard ratio and 95% CIs. The area of the square is proportional to the inverse of the sum of the between studies variance and the study-specific variance. The diamond represents the pooled multivariate hazard ratio and 95% CI. The dotted line of the diamond indicates the bounds of the 95% prediction interval.
The association between statin use and mortality in colorectal cancer patients was evaluated in eight studies [4,26–32], five of which showed no association [4,27,28,30,32]. Only the three
largest studies [26,29,31] showed a statistically significant reduction in mortality for colorectal cancer patients who used statins. In the study of Siddiqui et al. [29], the long-term statin use prior
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx
11
Fig. 5. Meta-analysis of HRs for all-cause mortality and cancer-specific mortality per increment of 365 DDDs or per increment of 1 year duration of statin use.
to diagnosis was associated with a less advanced tumor stage, a higher prevalence of right-sided tumors, a lower frequency of distant metastases, and a better survival rate. In the two largest,
population-based cohort studies, a reduced death risk was observed in colorectal cancer patients who used statin before or after diagnosis [26]. In addition, although Mace et al. failed to
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
12
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx
found an association between statin use and mortality in colorectal cancer patients, they showed that rectal cancer patients with use of statins had a better response to neoadjuvant chemoradiation (65.7% versus 48.7%, P = 0.004) and lower median regression grade (1 versus 2, P = 0.01) [16]. There are six studies on the relationship of statin use and mortality in prostate cancer patients [13–18]. The study by Katz et al. reported a reduced risk of all-cause mortality in prostate cancer patients who used statins after treatment with radical prostatectomy or radiotherapy [16]. Another study [17] also evaluated the effect of statins on the survival of men receiving prostate cancer treatment but found no effect. The investigators ascribed this nonresult to the adjustment for comorbidity. They expected that statins are used more often with increased comorbidity, which is significantly associated with worse survival. In the largest, prescription database study [13], statin use after diagnosis was associated with a better survival for prostate cancer patients, and this beneficial effect was stronger in those who also used statins before diagnosis. Grytli et al. found that prostate cancer patients who used statins both pre- and postdiagnosis had a lower risk of death from prostate cancer [18]. The other two studies [14,15] which have assessed statin use prior to diagnosis also showed a protective effect against prostate cancer death. Clinical data remain inconclusive regarding the effect of statins on human cancer, although much evidence suggested statins may exhibit anticancer properties through inducing cell cycle arrest and apoptosis, sensitizing cancer cells to specific chemotherapies, inhibiting invasion and angiogenesis, inducting tumor differentiation, and reducing cholesterol [66]. Several randomized controlled trials have assessed the statin use on mortality in cancer patients [67–72]. Strandberg et al. showed that mortality from cancer was slightly, although non-significantly, reduced in patients with coronary heart disease who used simvastatin before a diagnosis of cancer relative to those who used placebo [67]. In study by Han et al., the authors found that gefitinib plus simvastatin showed longer progression-free survival compared with gefitinib alone in non–small cell lung cancer patients with wild-type EGFR nonadenocarcinomas [68]. A study of 83 advanced hepatocellular carcinoma patients showed that pravastatin prolonged the survival of the patients [72]. However, the other three studies showed that adding statin to chemotherapy in advanced pancreatic cancer [70] or advanced gastric cancer [69,71] did not provide clinical benefit, although it also did not result in increased toxicity. As expected, there is significant heterogeneity among included studies. Several reasons may account for the heterogeneity. First, statins may play a different role in different cancer types. Even the same cancer focus has different grades and stages. Second, differences between statins may exist since the hydrophilic statins have a decreased ability to penetrate cell membranes, which may contribute to different effect on cancer death. Third, misclassification of statin use is likely to impact on the effect estimates of statin use. There are nine studies assessed the exposure to statins based on self-reported, while the others obtained exposure data from prescription database or medical record (Table S1). Self-reported data are not expected to be accurately recalled. Fourth, the included studies are varying in time periods of exposure assessment and follow-up period. Fifth, some of the included studies have excluded deaths in the year after diagnosis using a lag, while others have not. All of these may explain the differing results and significant heterogeneity among included studies. The potential limitations of our study should be considered when interpreting the results. First, we only used all-cause mortality and cancer-specific mortality as outcome, and studies which have investigated other outcomes (e.g., cancer recurrence) and
statin use were not included. Second, in the subgroup analysis, the sample size of each subgroup was relatively small not having enough statistical power to explore the real association. Third, we have not assessed the impact of prediagnosis statin use on the relationship between statin use after cancer diagnosis and patient survival with limited information from original article. So did prediagnosis statin use, without assessment the impact of postdiagnosis statin use. Fourth, although many of the studies had adjusted for important risk factors, unmeasured factors related to statin use may also have influenced results of individual studies. Fifth, there is some evidence of publication for prediagnosis statin use. We only included published studies and a number of meeting abstracts were excluded from present study. Sixth, a number of included studies [4,8,10,11,16–18,20– 23,25,27,34,56,59,61,63,64] may suffer from immortal time bias [73]. Seventh, the dose–response analyses were performed with limit studies in mixed cancer types; therefore, the results should be interpreted with caution. In conclusion, the average effect of statin use, both prediagnosis and postdiagnosis, is beneficial for overall survival and cancerspecific survival. However, statin use may not always be beneficial in an individual setting, especially for postdiagnosis statin use. The full potential effect of statin use on mortality of cancer patients should be further accessed through randomized controlled trials in the future.
Conflict of interest The authors declare no conflict of interest.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ctrv.2015.04.005.
References [1] Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin 2011;61:69–90. [2] Maron DJ, Fazio S, Linton MF. Current perspectives on statins. Circulation 2000;101:207–13. [3] Schachter M. Chemical, pharmacokinetic and pharmacodynamic properties of statins: an update. Fundam Clin Pharmacol 2005;19:117–25. [4] Krens LL, Simkens LH, Baas JM, Koomen ER, Gelderblom H, Punt CJ, et al. Statin use is not associated with improved progression free survival in cetuximab treated kras mutant metastatic colorectal cancer patients: results from the CAIRO2 study. PLoS ONE 2014;9:e112201. [5] Cardwell CR, Hicks BM, Hughes C, Murray LJ. Statin use after diagnosis of breast cancer and survival: a population-based cohort study. Epidemiology 2015;26:68–78. [6] Livingstone E, Hollestein LM, van Herk-Sukel MP, van de Poll-Franse L, Joosse A, Schilling B, et al. Statin use and its effect on all-cause mortality of melanoma patients: a population-based Dutch cohort study. Cancer Med 2014;3:1284–93. [7] Nam DH, Lee H, Park JC, Shin SK, Lee SK, Hyung WJ, et al. Long-term statin therapy improves oncological outcome after radical gastrectomy for stage II and III gastric cancer. Anticancer Res 2014;34:355–61. [8] Kaffenberger SD, Lin-Tsai O, Stratton KL, Morgan TM, Barocas DA, Chang SS, et al. Statin use is associated with improved survival in patients undergoing surgery for renal cell carcinoma. Urol Oncol 2014;33(21):e11–7. [9] Murtola TJ, Visvanathan K, Artama M, Vainio H, Pukkala E. Statin use and breast cancer survival: a nationwide cohort study from Finland. PLoS ONE 2014;9:e110231. [10] Chae YK, Valsecchi ME, Kim J, Bianchi AL, Khemasuwan D, Desai A, et al. Reduced risk of breast cancer recurrence in patients using ACE inhibitors, ARBs, and/or statins. Cancer Invest 2011;29:585–93. [11] Brewer TM, Masuda H, Liu DD, Shen Y, Liu P, Iwamoto T, et al. Statin use in primary inflammatory breast cancer: a cohort study. Br J Cancer 2013;109:318–24. [12] Desai P, Lehman A, Chlebowski RT, Kwan ML, Arun M, Manson JE, et al. Statins and breast cancer stage and mortality in the Women’s Health Initiative. Cancer Causes Control 2015.
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx [13] Yu O, Eberg M, Benayoun S, Aprikian A, Batist G, Suissa S, et al. Use of statins and the risk of death in patients with prostate cancer. J Clin Oncol 2014;32:5–11. [14] Geybels MS, Wright JL, Holt SK, Kolb S, Feng Z, Stanford JL. Statin use in relation to prostate cancer outcomes in a population-based patient cohort study. Prostate 2013;73:1214–22. [15] Marcella SW, David A, Ohman-Strickland PA, Carson J, Rhoads GG. Statin use and fatal prostate cancer: a matched case-control study. Cancer 2012;118:4046–52. [16] Katz MS, Carroll PR, Cowan JE, Chan JM, D’Amico AV. Association of statin and nonsteroidal anti-inflammatory drug use with prostate cancer outcomes: results from CaPSURE. BJU Int 2010;106:627–32. [17] Caon J, Paquette M, Hamm J, Pickles T. Does statin or ASA affect survival when prostate cancer is treated with external beam radiation therapy? Prostate Cancer 2014;2014:184297. [18] Grytli HH, Fagerland MW, Fossa SD, Tasken KA. Association between use of beta-blockers and prostate cancer-specific survival: a cohort study of 3561 prostate cancer patients with high-risk or metastatic disease. Eur Urol 2014;65:635–41. [19] Habis M, Wroblewski K, Bradaric M, Ismail N, Yamada SD, Litchfield L, et al. Statin therapy is associated with improved survival in patients with nonserous-papillary epithelial ovarian cancer: a retrospective cohort analysis. PLoS ONE 2014;9:e104521. [20] Lavie O, Pinchev M, Rennert HS, Segev Y, Rennert G. The effect of statins on risk and survival of gynecological malignancies. Gynecol Oncol 2013;130:615–9. [21] Elmore RG, Ioffe Y, Scoles DR, Karlan BY, Li AJ. Impact of statin therapy on survival in epithelial ovarian cancer. Gynecol Oncol 2008;111:102–5. [22] Nowakowski GS, Maurer MJ, Habermann TM, Ansell SM, Macon WR, Ristow KM, et al. Statin use and prognosis in patients with diffuse large B-cell lymphoma and follicular lymphoma in the rituximab era. J Clin Oncol 2010;28:412–7. [23] Samaras P, Heider H, Haile SR, Petrausch U, Schaefer NG, Siciliano RD, et al. Concomitant statin use does not impair the clinical outcome of patients with diffuse large B cell lymphoma treated with rituximab-CHOP. Ann Hematol 2010;89:783–7. [24] Hamilton RJ, Morilla D, Cabrera F, Leapman M, Chen LY, Bernstein M, et al. The association between statin medication and progression after surgery for localized renal cell carcinoma. J Urol 2014;191:914–9. [25] Viers BR, Houston Thompson R, Psutka SP, Lohse CM, Cheville JC, Leibovich BC, et al. The association of statin therapy with clinicopathologic outcomes and survival among patients with localized renal cell carcinoma undergoing nephrectomy. Urol Oncol 2015. [26] Cardwell CR, Hicks BM, Hughes C, Murray LJ. Statin use after colorectal cancer diagnosis and survival: a population-based cohort study. J Clin Oncol 2014;32:3177–83. [27] Lakha F, Theodoratou E, Farrington SM, Tenesa A, Cetnarskyj R, Din FV, et al. Statin use and association with colorectal cancer survival and risk: case control study with prescription data linkage. BMC Cancer 2012;12:487. [28] Ng K, Ogino S, Meyerhardt JA, Chan JA, Chan AT, Niedzwiecki D, et al. Relationship between statin use and colon cancer recurrence and survival: results from CALGB 89803. J Natl Cancer Inst 2011;103:1540–51. [29] Siddiqui AA, Nazario H, Mahgoub A, Patel M, Cipher D, Spechler SJ. For patients with colorectal cancer, the long-term use of statins is associated with better clinical outcomes. Dig Dis Sci 2009;54:1307–11. [30] Mace AG, Gantt GA, Skacel M, Pai R, Hammel JP, Kalady MF. Statin therapy is associated with improved pathologic response to neoadjuvant chemoradiation in rectal cancer. Dis Colon Rectum 2013;56:1217–27. [31] Shao YY, Hsu CH, Yeh KH, Chen HM, Yeh YC, Lai CL, et al. Statin use is associated with improved prognosis of colorectal cancer in Taiwan. Clin Colorectal Cancer 2015. [32] Hoffmeister M, Jansen L, Rudolph A, Toth C, Kloor M, Roth W, et al. Statin use and survival after colorectal cancer: the importance of comprehensive confounder adjustment. J Natl Cancer Inst 2015;107. [33] Graf H, Jungst C, Straub G, Dogan S, Hoffmann RT, Jakobs T, et al. Chemoembolization combined with pravastatin improves survival in patients with hepatocellular carcinoma. Digestion 2008;78:34–8. [34] Song M, Chung J, Lee G, Cho S, Hong J, Shin D, et al. Statin use has negative clinical impact on non-germinal center in patients with diffuse large B cell lymphoma in rituximab era. Leuk Res 2015;39:211–5. [35] Tierney JF, Stewart LA, Ghersi D, Burdett S, Sydes MR. Practical methods for incorporating summary time-to-event data into meta-analysis. Trials 2007;8:16. [36] Wells G, Shea B, O’Connell D, Robertson J, Peterson J, Welch V, et al. The Newcastle–Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. www.ohri.ca/programs/clinical_epidemiology/ oxford_web.ppt. 2010. [37] Viechtbauer W. Bias and efficiency of meta-analytic variance estimators in the random-effects model. J Educat Behav Statistics 2005;30:261–93. [38] Raudenbush SW. Analyzing effect sizes: Random-effects models. The Handbook of Research Synthesis and Meta-analysis 2009;2:295–316. [39] Riley RD, Higgins JP, Deeks JJ. Interpretation of random effects meta-analyses. BMJ 2011;342:d549. [40] Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539–58. [41] Hedges LV, Pigott TD. The power of statistical tests in meta-analysis. Psychol Methods 2001;6:203–17.
13
[42] Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629–34. [43] Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics 2000;56:455–63. [44] Greenland S, Longnecker MP. Methods for trend estimation from summarized dose-response data, with applications to meta-analysis. Am J Epidemiol 1992;135:1301–9. [45] Viechtbauer W. Conducting meta-analyses in R with the metal for package. J Stat Softw 2010;36:1–48. [46] Crippa A, Orsini N. Dosresmeta Performing multivariate dose-response metaanalysis. Vienna, Austria: R Foundation for Statistical Computing; 2013. [47] Kollmeier MA, Katz MS, Mak K, Yamada Y, Feder DJ, Zhang Z, et al. Improved biochemical outcomes with statin use in patients with high-risk localized prostate cancer treated with radiotherapy. Int J Radiat Oncol Biol Phys 2011;79:713–8. [48] Mehta JL, Bursac Z, Hauer-Jensen M, Fort C, Fink LM. Comparison of mortality rates in statin users versus nonstatin users in a United States veteran population. Am J Cardiol 2006;98:923–8. [49] Cuaron J, Pei X, Cohen GN, Cox BW, Yamada Y, Zelefsky MJ, et al. Statin use not associated with improved outcomes in patients treated with brachytherapy for prostate cancer. Brachytherapy 2014;14:179–84. [50] Rotta M, Storer BE, Storb R, Martin PJ, Flowers MED, Vernon MS, et al. Impact of recipient statin treatment on graft-versus-host disease after allogeneic hematopoietic cell transplantation. Biol Blood Marrow Transplant 2010;16:1463–6. [51] da Silva RD, Xylinas E, Kluth L, Crivelli JJ, Chrystal J, Chade D, et al. Impact of statin use on oncologic outcomes in patients with urothelial carcinoma of the bladder treated with radical cystectomy. J Urol 2013;190:487–92. [52] Kaffenberger SD, Lin-Tsai O, Stratton KL, Morgan TM, Barocas DA, Chang SS, et al. Statin use is associated with improved survival in patients undergoing surgery for renal cell carcinoma. Urol Oncol 2015;33:21.e11–7. [53] Stavrou EP, Buckley N, Olivier J, Pearson SA. Discontinuation of statin therapy in older people: does a cancer diagnosis make a difference An observational cohort study using data linkage. BMJ Open 2012;2. [54] Suna S, Sakata Y, Nakatani D, Okuda K, Shimizu M, Usami M, et al. Decreased mortality associated with statin treatment in patients with acute myocardial infarction and lymphotoxin-alpha C804A polymorphism. Atherosclerosis 2013;227:373–9. [55] Nickels S, Vrieling A, Seibold P, Heinz J, Obi N, Flesch-Janys D, et al. Mortality and recurrence risk in relation to the use of lipid-lowering drugs in a prospective breast cancer patient cohort. PLoS ONE 2013;8:e75088. [56] Crivelli JJ, Xylinas E, Kluth LA, da Silva RD, Chrystal J, Novara G, et al. Effect of statin use on outcomes of non-muscle-invasive bladder cancer. BJU Int 2013;112:E4–E12. [57] Eindhoven JA, Onuma Y, Oemrawsingh RM, Daemen J, Van Nierop JWI, De Jaegere PPT, et al. Long-term outcome after statin treatment in routine clinical practice: Results from a prospective PCI cohort study. EuroIntervention 2012;7:1420–7. [58] Gaist D, Hallas J, Friis S, Hansen S, Sorensen HT. Statin use and survival following glioblastoma multiforme. Cancer Epidemiol 2014;38:722–7. [59] Gardette V, Bongard V, Dallongeville J, Arveiler D, Bingham A, Ruidavets JB, et al. Ten-year all-cause mortality in presumably healthy subjects on lipidlowering drugs (from the Prospective Epidemiological Study of Myocardial Infarction [PRIME] prospective cohort). Am J Cardiol 2009;103:381–6. [60] Haukka J, Niskanen L, Partonen T, Lonnqvist J, Tiihonen J. Statin usage and allcause and disease-specific mortality in a nationwide study. Pharmacoepidemiol Drug Saf 2012;21:61–9. [61] Lipworth L, Fazio S, Kabagambe EK, Munro HM, Nwazue VC, Tarone RE, et al. A prospective study of statin use and mortality among 67,385 blacks and whites in the Southeastern United States. Clin Epidemiol 2014;6:15–25. [62] Nielsen SF, Nordestgaard BG, Bojesen SE. Statin use and reduced cancerrelated mortality. N Engl J Med 2012;367:1792–802. [63] van Gestel YR, Hoeks SE, Sin DD, Huzeir V, Stam H, Mertens FW, et al. COPD and cancer mortality: the influence of statins. Thorax 2009;64:963–7. [64] Xylinas E, Kluth LA, Crivelli JJ, Rieken M, Margulis V, Seitz C, et al. Impact of statin use on oncologic outcomes of patients with upper tract urothelial carcinoma treated with radical nephroureterectomy. Eur Urol 2013;63:1134–5. [65] Yoon LS, Goodman MT, Rimel B, Jeon CY. Statin use and survival in elderly patients with endometrial cancer. Gynecol Oncol 2015. [66] Barkas F, Rizzo M, Dinicolantonio JJ, Liberopoulos E. Beyond cholesterol reduction, the pleiotropic effects of statins: Is their use in cancer prevention hype or hope? Clin Lipidol 2013;8:273–7. [67] Strandberg TE, Pyorala K, Cook TJ, Wilhelmsen L, Faergeman O, Thorgeirsson G, et al. Mortality and incidence of cancer during 10-year follow-up of the Scandinavian Simvastatin Survival Study (4S). Lancet 2004;364:771–7. [68] Han JY, Lee SH, Yoo NJ, Lee SH, Moon YJ, Yun T, et al. A randomized phase II study of gefitinib plus simvastatin versus gefitinib alone in previously treated patients with advanced non-small cell lung cancer. Clin Cancer Res 2011;17:1553–60. [69] Konings IR, van der Gaast A, van der Wijk LJ, de Jongh FE, Eskens FA, Sleijfer S. The addition of pravastatin to chemotherapy in advanced gastric carcinoma: a randomised phase II trial. Eur J Cancer 2010;46:3200–4. [70] Hong JY, Nam EM, Lee J, Park JO, Lee SC, Song SY, et al. Randomized doubleblinded, placebo-controlled phase II trial of simvastatin and gemcitabine in
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005
14
S. Zhong et al. / Cancer Treatment Reviews xxx (2015) xxx–xxx
advanced pancreatic cancer patients. Cancer Chemother Pharmacol 2014;73:125–30. [71] Kim ST, Kang JH, Lee J, Park SH, Park JO, Park YS, et al. Simvastatin plus capecitabine-cisplatin versus placebo plus capecitabine-cisplatin in patients with previously untreated advanced gastric cancer: a double-blind randomised phase 3 study. Eur J Cancer 2014;50:2822–30.
[72] Kawata S, Yamasaki E, Nagase T, Inui Y, Ito N, Matsuda Y, et al. Effect of pravastatin on survival in patients with advanced hepatocellular carcinoma. A randomized controlled trial. Br J Cancer 2001;84:886–91. [73] Levesque LE, Hanley JA, Kezouh A, Suissa S. Problem of immortal time bias in cohort studies: example using statins for preventing progression of diabetes. BMJ 2010;340:b5087.
Please cite this article in press as: Zhong S et al. Statin use and mortality in cancer patients: Systematic review and meta-analysis of observational studies. Cancer Treat Rev (2015), http://dx.doi.org/10.1016/j.ctrv.2015.04.005