Ovarian cancer accounts for approximately 3% of all cancers in women,1 but it is a common cause of cancer death in women with an extremely high age-adjusted mortality rate of 8.2 per 100,000 women per year.2 Ovarian cancer is also ranked as 1 of top 10 deadliest cancers in women. It was ranked as the seventh deadliest cancer in women worldwide.3
The risk of ovarian cancer increases with age, family history and genetic predisposition, nulliparity, use of fertility drugs, and long-term hormone replacement therapy. On the contrary, the risk of ovarian cancer decreases with bilateral tubal ligation, hysterectomy, and history of oral contraceptive usage.4 Controversially, some studies showed that ovarian cancer in patients with type 2 diabetes decreased survival,5 whereas another large population-based study opposed such association.6
In patients with type 2 diabetes, metformin is an antidiabetic drug commonly prescribed. A number of studies have shown that tumorigenesis is promoted by high insulin and glucose levels induced by type 2 diabetes.7 Metformin is known to reduce both insulin and glucose levels, and it was postulated that metformin may exert anticancer effects. Recently, observational studies also demonstrated that diabetic patients receiving metformin have lower cancer incidence and higher survival.8,9 For ovarian cancer, preclinical studies have shown that metformin can inhibit the proliferation of cancer cell lines.10,11
However, evidence from large clinical trials and real-world studies of the effects of metformin on ovarian cancer outcomes remains limited and inconclusive.12 Therefore, we systematically searched and meta-analyzed the existing evidence to determine the effects of metformin in diabetic ovarian cancer patients in terms of ovarian cancer risk and survival outcomes.
Data Sources, Search Strategy, and Selection Criteria
Relevant studies were identified by searching the following data sources: PubMed and EMBASE with relevant key words that included all spellings of “metformin” AND (“ovarian cancer” OR “ovary tumor”) until December 2012. Reference lists from identified studies were manually scanned to identify any other relevant studies. The literature search, data extraction, and quality assessment were undertaken independently by 2 authors (P.D. and N.C.) with a standardized approach. All clinical studies assessing the effects of metformin on ovarian cancer were eligible for inclusion.
Data Extraction and Quality Assessment
Published reports were obtained, and standardized information was extracted. The data sought included purpose of studies, study design, baseline characteristics, metformin used, dose of metformin, number of ovarian cancer patients with/without metformin used, and outcomes. We used Newcastle-Ottawa quality assessment scale (NOS)13 for assessing the quality of observational studies, whereas randomized controlled trials (RCTs) were assessed using risk of bias.14 Any disagreement in abstracted data was adjudicated by a third reviewer (R.S.).
The overall odds ratios (ORs) with corresponding 95% confidence intervals (CIs) were calculated to determine the effects of metformin on ovarian cancer. All analyses were performed using Der Simonian and Laird method15 under a random-effects model. The percentage of variability across studies attributable to heterogeneity beyond chance was estimated using the Q-statistic and I2 tests.16 A Q value of 0.10 indicated statistically significant between-study heterogeneity, and I2 values less than 25% denoted no/minimal heterogeneity across studies. Begg test was used to evaluate the publication bias.17 All analyses were performed with STATA.
A total of 190 articles were identified, but only 6 studies met the inclusion criteria. Of the 184 excluded articles, 152 articles were not related to metformin or ovarian cancer, and 32 articles were not clinical studies. Of 6 articles that met the inclusion criteria, 3 articles were duplicative studies. One additional article was added after reviewing the bibliographies of the included articles. However, the additional study reported observational events for 2 RCTs. In total, 4 articles18–21 were included in our final analysis.
Characteristics of Included Studies
Among 4 included studies,18–21 2 studies determined the effects of metformin on survival in diabetic ovarian cancer patients,20,21 whereas 2 other studies determined the preventive effects of metformin on ovarian cancer occurrence in diabetic patients.18,19 Two studies were undertaken using retrospective cohort design,20,21 1 study was conducted using retrospective case-control design,18 and another study19 was an observational report of 2 RCTs (ADOPT [A Diabetes Outcome Progression Trial]22 and RECORD [Rosiglitazone Evaluated for Cardiovascular Outcomes and Regulation of Glycaemia in Diabetes]23 trials). The study characteristics are shown in Table 1.
Quality of Included Studies
All studies were undertaken using various methods to control confounders. Multivariable regression analysis was used to control confounders in 3 studies,18,20,21 whereas randomization was used in another study.19 Among 3 observational studies,18,20,21 information bias was controlled in a study21 using blinded investigators, whereas the other 2 studies18,20 did not report how information bias was minimized. An included study19 reported results from 2 RCTs.22,23 Therefore, we assessed the quality of study based on methods of original articles. ADOPT22 controlled for the information bias using double-blind approach, whereas RECORD23 did not control for information bias because it was conducted with an open-label approach. According to risk of bias, ADOPT22 has low risk of bias, whereas RECORD23 has high risk of bias. The NOS was also used to assess quality of observational studies.13 Briefly, there are separated scales of NOS for cohort study design and case-control study design. Summary scores of cohort studies were 9 of 9 score for the studies of both Kumar et al20 and Romero et al,21 whereas the summary score for an included case-control study18 was 7 out of 9 score.
Effects of Metformin on Clinical Outcomes
Studies were classified into 2 groups: (1) studies conducted to determine the effects of metformin on survival outcomes of ovarian cancer20,21 and (2) studies conducted to determine the effects of metformin on ovarian cancer occurrence in diabetic patients.18,19 Because of the different outcome measures among studies conducted to determine survival outcomes of ovarian cancer, meta-analysis of those studies could not be performed. We qualitatively summarized the results of those studies. On the other hand, studies conducted to investigate the preventive effects on ovarian cancer reported similar outcome measures. Thus, we performed meta-analysis of these studies.
Effects of Metformin on Survival Outcomes of Ovarian Cancer
Two studies20,21 evaluated the effects of metformin on treating ovarian cancer. Kumar et al20 conducted a retrospective cohort study with 72 ovarian cancer patients receiving metformin and 143 random matched ovarian cancer patients not receiving metformin. In the ovarian cancer cohort, cases were matched with control subjects for age (±5 years), International Federation of Gynecology and Obstetrics (FIGO) stage, and residual disease. The exposure of this study was metformin 500 to 1000 mg twice daily. The mean age of patients receiving metformin was 60.6 ± 11 years, whereas that of patients not receiving metformin was 60.3 ± 11 years. Severity of patients receiving metformin by FIGO stage was lower than those of patients not receiving metformin. Whereas 62.5% of patients receiving metformin were in late stage, 83.2% of patients not receiving metformin was in late stage. The 5-year disease-specific survival (DSS) rate for patients receiving metformin was 73%, whereas that of patients who did not receive metformin was 44% with statistical significance (P = 0.002). Using Cox proportional hazard model, the adjusted hazard ratio (HR) was 0.37 with 95% CI of 0.19 to 0.71 for overall DSS. The study also reported the subgroup analysis of patients with only epithelium ovarian cancer compared with ovarian cancer patients with and without diabetes. For the subgroup analysis, 61 epithelium ovarian cancer patients receiving metformin were included. For control subjects, 178 matched patients with epithelium ovarian cancer without diabetes and 103 patients with epithelium ovarian cancer with diabetes were included. The 5-year DSS rate of diabetic epithelium ovarian cancer patients receiving metformin was 67%, whereas that of nondiabetic epithelium ovarian cancer patients not receiving metformin was 47% with significant difference compared with diabetic epithelium ovarian cancer patients receiving metformin (P = 0.006). The 5-year DSS rate of diabetic epithelium ovarian cancer patients not receiving metformin was 40% with significant difference compared with diabetic epithelium ovarian cancer patients receiving metformin (log-rank P = 0.003). The adjusted HR of diabetic epithelium ovarian cancer patients receiving metformin and nondiabetic epithelium ovarian cancer patients not receiving metformin was 0.46 (95% CI, 0.26-0.83). The adjusted HRs of diabetic epithelium ovarian cancer patients receiving metformin and diabetic epithelium ovarian cancer patients not receiving metformin were not reported (Table 2). This study indicated significant benefits in survival outcomes after controlling for disease stage, grade, histology, chemotherapy, body mass index, and surgical cytoreduction. However, this study was limited with a small sample size. Moreover, it was a retrospective chart review. Confounder may remain, even though authors attempted to minimize confounders using matching method and multivariate regression model.
Romero et al21 conducted a retrospective cohort study. A total of 341 ovarian cancer patients were included in the study. Among the included patients, 297 patients (87%) were nondiabetic patients not using metformin, whereas 28 patients (8%) and 16 patients (5%) were diabetic patients not using metformin and diabetic patients using metformin, respectively. The mean age of nondiabetic patients not using metformin was 59 ± 12 years, whereas the mean ages of diabetic patients not using metformin and diabetic patients using metformin were 67 ± 9 and 60 ± 10 years, respectively. A total of 253 nondiabetic patients (82%) not using metformin was in late stage of ovarian cancer as defined by FIGO, whereas 24 (85%) and 11 (69%) of diabetic patients not using metformin and diabetic patients using metformin were in late stage of ovarian cancer. The 5-year progression-free survival of nondiabetic patients not using metformin was 23%, whereas those of diabetic patients not using metformin and diabetic patients using metformin were 8% and 51%, respectively. The overall 5-year survival rates were 37%, 23%, and 63% for nondiabetic patients not using metformin, diabetic patients not using metformin, and diabetic patients using metformin, respectively. The adjusted HR of diabetic patients using metformin and nondiabetic patients not using metformin for progression-free survival was 0.53 (95% CI, 0.27–1.15), whereas the adjusted HR of diabetic patients using metformin and diabetic patients not using metformin was 0.38 (95% CI, 0.16–0.90). For overall survival, the adjusted HR of diabetic patients using metformin and nondiabetic patients using metformin and diabetic patients not using metformin was 0.43 (95% CI, 0.16–1.19) (Table 2). This study also indicated significant benefits in survival outcomes. It was important to note that the study was a retrospective chart review. Some confounders might occur, although the authors used multivariable regression to minimize the effects of confounders.
According to the previously mentioned 2 studies, metformin was likely to be effective for treating ovarian cancer. There were trends for survival prolongation in diabetic patients using metformin compared with diabetic patients and nondiabetic patients not using metformin.
Effects of Metformin on Ovarian Cancer Prevention
Two studies18,19 were conducted to determine the effects of metformin on ovarian cancer prevention. Bodmer et al18 conducted a study using retrospective electronic database case-control study design, whereas the study by Home et al19 was conducted using observational reports of 2 RCT including ADOPT22 and RECORD.23 The study by Bodmer et al18 was conducted in 1611 ovarian cancer patients and 9170 matched control subjects. The ADOPT22 was conducted in 1454 diabetic patients using metformin and 2897 diabetic patients not using metformin. The median follow-up time of the ADOPT trial was 4.0 years. The RECORD23 was conducted in 3344 diabetic patients using metformin and 1103 diabetic patients not using metformin with 5.5-year median follow-up time. The mean age of patients in the study by Bodmer et al18 was 61.2 ± 13.1 years, whereas the mean age of those in ADOPT22 and RECORD23 ranged from 56.3 ± 10.0 to 57.0 ± 9.9 years and 57.0 ± 8.0 to 59.8 ± 8.3 years, respectively.
In the study of Bodmer et al,18 among patients who developed ovarian cancer, 41 patients (2.5%) used metformin, whereas 1570 patients (97.5%) did not use metformin. In patients who did not develop ovarian cancer, 269 patients (2.9%) used metformin, whereas 8901 patients (97.1%) did not use metformin. The unadjusted OR was 0.89 (95% CI, 0.62–1.23). In the analysis restricted to diabetic patients, among those patients who developed ovarian cancer, 38 patients (44.7%) used metformin, whereas 47 patients (55.3%) did not. In patients who did not develop ovarian cancer, 269 patients (53.9%) used metformin, whereas 221 (46.1%) did not. The unadjusted OR was 0.68 (95% CI, 0.43–1.07) (Table 3). It is important to note that the study was a retrospective case-control study; patients in both using and not using metformin groups may have also used other diabetic medications such as sulfonylurea or insulin, which might have affected the association between metformin and ovarian cancer. However, the study did not report the number of patients who took metformin, sulfonylurea, or insulin, but it was reported that there was no association between sulfonylurea or insulin and ovarian cancer.
The ADOPT22 and RECORD23 also reported the number of diabetic patients developing ovarian cancer. The number of metformin users who developed ovarian cancer were 0 (0%) of 1454 patients, whereas the number of non–metformin users who developed ovarian cancer was 3 (0.10%) of 2897 patients for the ADOPT. The OR was 0.28 (95% CI, 0.01–5.51). In the RECORD trial, the number of metformin users who developed ovarian cancer was 6 (0.18%) of 3344 patients, whereas the number of non–metformin users who developed ovarian cancer was 3 (0.27%) of 1103 patients. The OR was 0.66 (95% CI, 0.16–2.14) (Table 3).
The meta-analysis of those studies18,22,23 among diabetic patients was performed. When pooling the findings from RCTs together, we found that the pooled OR was 0.57 (95% CI, 0.16–1.99). In addition, we did not find heterogeneity among these RCTs (I2 statistic = 0.0%). When pooling all studies including 2 RCTs and 1 case-control study, we found that the pooled OR was 0.67 (95% CI, 0.43–1.04) with I2 statistic = 0.0% (Fig. 1). This was similar to the pooled OR of the meta-analysis of the RCTs. Begg test indicated no publication bias (P = 0.317). The findings indicated that there is a trend of protective effect of metformin on the occurrence of ovarian cancer in diabetic population.
Our findings revealed that metformin tended to prolong survival in diabetic ovarian cancer patients and also tended to decrease the risk of ovarian cancer in patients with diabetes mellitus. We found that approximately 50% of diabetic patients receiving metformin survived for 5 years longer than did nondiabetic patients who did not receive metformin.20 In addition, 5-year progression-free survival of diabetic patients receiving metformin was 50% more than that of nondiabetic patients who were not receiving metformin.21 Comparing with diabetic patients with ovarian cancer who did not receive metformin, 5-year progression-free survival of patients receiving metformin was 60% more than the survival of patients not receiving metformin. Furthermore, we found that metformin tended to decrease the occurrence of ovarian cancer of approximately 40% compared with nonmetformin in diabetic patients; however, this was not of statistical significance.
Our findings of survival outcomes in patients with ovarian cancer were similar to previous studies that were conducted in different types of cancer.24–28 The previous studies showed the benefit of metformin in decreasing cancer-related mortality25,28 in both overall cancer-related mortality25,27,28 and specific types of cancer-related mortality such as hepatocellular cancer26 and breast cancer.24 The previous studies indicated that the relative risk of cancer-related death in patients receiving metformin ranged from 0.3 to 0.7 compared with patients not receiving metformin. Moreover, in terms of preventive effects, our results were also similar to previous studies28,29 that showed the effects of metformin on preventing occurrence of any cancer such as hepatocellular cancer, colorectal cancer, and lung cancer.28,29
The anticancer effect of metformin has been attributed to several mechanisms. It is likely to be the result of the activation of adenosine monophosphate–activated protein kinase through liver kinase B1, a tumor suppressor protein kinase. Adenosine monophosphate–activated protein kinase suppresses the activity of mammalian target of rapamycin.30,31 Specific to ovarian cancer, metformin showed the apoptotic effects on ovarian cancer cell and decreased the growth of ovarian cancer cell lines (OVCAR-3 and OVCAR-4) by an up-regulation of adenosine monophosphate–activated protein kinase activity.32–34 Moreover, the effect of metformin on ovarian cancer is likely due to antiangiogenic effects of its ovarian cancer cells.34
Metformin is one of the potential medications to be adjuvant therapy with taxane/carboplatin, which are first-line therapy for ovarian cancer based on the National Comprehensive Cancer Network.35 An in vitro study36 indicated that metformin in conjunction with carboplatin produces a synergistic enhancement in cytotoxicity of ovarian cancer cells. However, there is no clinical evidence to support this hypothesis. Clinical studies to evaluate the synergistic effects of metformin in conjunction with carboplatin are still needed.
To the extent of our knowledge, this is the first systematic review summarizing all evidence of the therapeutic effects of metformin specifically on ovarian cancer. We sought to provide relevant empirical information of metformin in both treatment and prevention of ovarian cancer to clinical practitioners, especially gynecologists. Moreover, our findings may encourage interested researchers to conduct further RCTs to investigate the effects of metformin in ovarian cancer patients.
This study aimed to strengthen evidence of the effects of metformin on ovarian cancer; however, all but 2 studies22,23 were observational studies. The findings of this systematic review and meta-analysis should be interpreted with caution. There were differences in baseline characteristics between patients using and not using metformin among the included observational studies. This may have affected the outcome estimates. However, all observational studies included in this current review were adjusted for potential confounders, thus minimizing potential bias.
Among the studies determining the effects of metformin on ovarian cancer prevention, we also performed meta-analysis of 3 studies across study design as an additional analysis. The results of 2 RCTs and 1 case-control study were pooled together. The findings might be biased because of heterogeneity between study designs. However, we also tested the heterogeneity across studies using I2 statistic. The result of heterogeneity indicated that there was no observed statistical heterogeneity across the pooled studies in this meta-analysis. Moreover, the analysis was similar to findings from our meta-analysis of only RCTs. Thus, we believe that the result of our meta-analysis is reliable.
In conclusion, our systematic review and meta-analysis showed the potential effects of metformin in prolonging the survival of diabetic ovarian cancer patients and decreasing the occurrence of ovarian cancer in diabetic patients. However, most of the evidence was from observational studies. There is a call for further well-conducted controlled clinical trials to confirm the effects of metformin on ovarian cancer survival and ovarian cancer prevention.
The authors thank the Office of the Higher Education Commission, Ministry of Education, Thailand, and Thailand Research Fund through the Royal Golden Jubilee PhD Program (grant PHD/ 0356/2550 to P.D. and PHD/0127/2552 to R.S.) for supporting the grant fund for the PhD program.
3. American Cancer Society. Global Cancer Facts & Figures. 2nd ed. Atlanta, GA: American Cancer Society; 2011.
4. Roett MA, Evans P. Ovarian cancer: an overview. Am Fam Physician. 2009; 80: 609–616.
5. Bakhru A, Buckanovich RJ, Griggs JJ. The impact of diabetes on survival in women with ovarian cancer. Gynecol Oncol. 2011; 121: 106–111.
6. Weiderpass E, Ye W, Vainio H, et al. Diabetes mellitus and ovarian cancer (Sweden). Cancer Causes Control. 2002; 13: 759–764.
7. Ben Sahra I, Le Marchand-Brustel Y, Tanti JF, et al. Metformin in cancer therapy: a new perspective for an old antidiabetic drug? Mol Cancer Ther. 2010; 9: 1092–1099.
8. Baur DM, Klotsche J, Hamnvik OP, et al. Type 2 diabetes mellitus and medications for type 2 diabetes mellitus are associated with risk for and mortality from cancer in a German primary care cohort. Metabolism. 2010; 60: 1363–1371.
9. Bowker SL, Yasui Y, Veugelers P, et al. Glucose-lowering agents and cancer mortality rates in type 2 diabetes: assessing effects of time-varying exposure. Diabetologia. 2010; 53: 1631–1637.
10. Gotlieb WH, Saumet J, Beauchamp MC, et al. In vitro metformin anti-neoplastic activity in epithelial ovarian cancer. Gynecol Oncol. 2008; 110: 246–250.
11. Rattan R, Giri S, Hartmann LC, et al. Metformin attenuates ovarian cancer cell growth in an AMP-kinase dispensable manner. J Cell Mol Med. 2009; 15: 166–178.
12. Berstein LM. Metformin in obesity, cancer and aging: addressing controversies. Aging (Albany NY). 2012; 4: 320–329.
14. Higgins JPT, Altman DG, Sterne JAC. Assessing risk of bias in included studies. In: Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions Version 5.0.2: The Cochrane Collaboration. 2009. Updated September 2009. Available at www.cochrane-handbook.org
15. Der Simonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986; 7: 177–188.
16. Higgins J, Thompson S. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002; 21: 1539–1558.
17. Begg C, Berlin J. Publication bias and dissemination of clinical research. J Natl Cancer Inst. 1989; 81: 107–115.
18. Bodmer M, Becker C, Meier C, et al. Use of metformin and the risk of ovarian cancer: a case-control analysis. Gynecol Oncol. 2011; 123: 200–204.
19. Home PD, Kahn SE, Jones NP, et al. Experience of malignancies with oral glucose-lowering drugs in the randomised controlled ADOPT (A Diabetes Outcome Progression Trial) and RECORD (Rosiglitazone Evaluated for Cardiovascular Outcomes and Regulation of Glycaemia in Diabetes) clinical trials. Diabetologia. 2010; 53: 1838–1845.
20. Kumar S, Meuter A, Thapa P, et al. Metformin intake is associated with better survival in ovarian cancer: a case-control study. Cancer. 2013; 119: 555–562.
21. Romero IL, McCormick A, McEwen KA, et al. Relationship of type II diabetes and metformin use to ovarian cancer progression, survival, and chemosensitivity. Obstet Gynecol. 2012; 119: 61–67.
22. Kahn SE, Haffner SM, Heise MA, et al. Glycemic durability of rosiglitazone, metformin, or glyburide monotherapy. N Engl J Med. 2006; 355: 2427–2443.
23. Home PD, Pocock SJ, Beck-Nielsen H, et al. Rosiglitazone evaluated for cardiovascular outcomes in oral agent combination therapy for type 2 diabetes (RECORD): a multicentre, randomised, open-label trial. Lancet. 2009; 373: 2125–2135.
24. Bodmer M, Meier C, Krahenbuhl S, et al. Long-term metformin use is associated with decreased risk of breast cancer. Diabetes Care. 2010; 33: 1304–1308.
25. DeCensi A, Puntoni M, Goodwin M, et al. Metformin and cancer risk in diabetic patients: a systematic review and meta-analysis. Cancer Prev Res. 2010; 3: OF1–OF11.
26. Hassan MM, Curley SA, Li D, et al. Association of diabetes duration and diabetes treatment with the risk of hepatocellular carcinoma. Cancer. 2010; 116: 1938–1946.
27. Landman GW, Kleefstra N, van Hateren KJ, et al. Metformin associated with lower cancer mortality in type 2 diabetes: ZODIAC-16. Diabetes Care. 2010; 33: 322–326.
28. Noto H, Goto A, Tsujimoto T, et al. Cancer risk in diabetic patients treated with metformin: a systematic review and meta-analysis. PLoS ONE. 2012; 7: e33411.
29. Lee MS, Hsu CC, Wahlqvist ML, et al. Type 2 diabetes increases and metformin reduces total, colorectal, liver and pancreatic cancer incidences in Taiwanese: a representative population prospective cohort study of 800,000 individuals. BMC Cancer. 2011; 11 20.
30. Kourelis TV, Siegel RD. Metformin and cancer: new applications for an old drug. Med Oncol. 2012; 29: 1314–1327.
31. Shackelford DB, Shaw RJ. The LKB1-AMPK pathway: metabolism and growth control in tumour suppression. Nat Rev Cancer. 2009; 9: 563–575.
32. Yasmeen A, Beauchamp MC, Piura E, et al. Induction of apoptosis by metformin in epithelial ovarian cancer: involvement of the Bcl-2 family proteins. Gynecol Oncol. 2011; 121: 492–498.
33. Rattan R, Giri S, Hartmann LC, et al. Metformin attenuates ovarian cancer cell growth in an AMP-kinase dispensable manner. J Cell Mol Med. 2011; 15: 166–178.
34. Shank JJ, Yang K, Ghannam J, et al. Metformin targets ovarian cancer stem cells in vitro and in vivo. Gynecol Oncol. 2012; 127: 390–397.
36. Erices R, Bravo ML, Gonzalez P, et al. Reprod Sci. 2013; Metformin, at concentrations corresponding to the treatment of diabetes, potentiates the cytotoxic effects of carboplatin in cultures of ovarian cancer cells. Reprod Sci. 2013; published online May 7, 2013.