Tibolone Decreases Lipoprotein(A) Levels in Postmenopausal Women: a Systematic Review and Meta-Analysis of 12 Studies with 1009 Patients

Accepted Manuscript

Tibolone Decreases Lipoprotein(a) levels in Postmenopausal Women: A Systematic Review and Meta-Analysis of 12 studies with 1009 patients

Kazuhiko Kotani, Amirhossein Sahebkar, Corina Serban, Florina Andrica, Peter P. Toth, Steven R. Jones, Karam Kostner, Michael J. Blaha, Seth Martin, Jacek Rysz, Stephen Glasser, Kausik K. Ray, Gerald F. Watts, Dimitri P. Mikhailidis, Prof. Maciej Banach, MD, PhD, FNLA, FAHA, FESC; FASA, Head

PII: S0021-9150(15)30017-4 DOI: 10.1016/j.atherosclerosis.2015.06.056 Reference: ATH 14168

To appear in: Atherosclerosis

Received Date: 28 May 2015 Revised Date: 28 June 2015 Accepted Date: 29 June 2015

Please cite this article as: Kotani K, Sahebkar A, Serban C, Andrica F, Toth PP, Jones SR, Kostner K, Blaha MJ, Martin S, Rysz J, Glasser S, Ray KK, Watts GF, Mikhailidis DP, Banach M, Lipid and Blood Pressure Meta-analysis Collaboration (LBPMC) Group, Tibolone Decreases Lipoprotein(a) levels in Postmenopausal Women: A Systematic Review and Meta-Analysis of 12 studies with 1009 patients, Atherosclerosis (2015), doi: 10.1016/j.atherosclerosis.2015.06.056.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT

Tibolone Decreases Lipoprotein(a) levels in Postmenopausal Women: A Systematic Review and Meta-Analysis of 12 studies with 1009 patients

Kazuhiko Kotani 1, Amirhossein Sahebkar 2,3 , Corina Serban 4, Florina Andrica 5, Peter P. Toth 6,7 , Steven R. Jones 7, Karam Kostner 8, Michael J. Blaha 7, Seth Martin 7, Jacek Rysz 9, Stephen Glasser 10 , Kausik K. Ray 11, Gerald F. Watts 12, Dimitri P. Mikhailidis 13 , Maciej Banach 14 ; Lipid and Blood Pressure Meta-analysis Collaboration (LBPMC) Group

1Department of Clinical Laboratory Medicine, Jichi Medical University, Shimotsuke City, Tochigi, Japan; 2Biotechnology Research Center, Mashhad University of Medical Sciences, Mashhad, Iran; 3Metabolic Research Centre, Royal Perth Hospital, School of Medicine and Pharmacology, University of Western Australia, Perth, Australia; 4Department of Functional Sciences, Discipline of Pathophysiology, “Victor Babes” University of Medicine and Pharmacy, Timisoara, Romania; 5Faculty of Pharmacy, Discipline of Pharmaceutical Chemistry “Victor Babes” University of Medicine and Pharmacy, Timisoara, Romania; 6Preventive Cardiology, CGH Medical Center, Sterling, Illinois, USA; 7The Johns Hopkins Ciccarone Center for the Prevention of Heart Disease, Baltimore, MD, USA; 8Mater Hospital, University of Queensland, St Lucia, QLD, Australia; 9Chair of Nephrology and Hypertension, Medical University of Lodz, Poland; 10 Department of Preventive Medicine, University of Alabama at Birmingham, Birmingham, AL, USA; 11Department of Primary Care and Public Health, Imperial College London, London, UK; 12Lipid Disorders Clinic, Cardiovascular Medicine, Royal Perth Hospital, School of Medicine and Pharmacology, University of Western Australia,MANUSCRIPT Perth, WA, Australia; 13 Department of Clinical Biochemistry, Royal Free Campus, University College London Medical School, University College London (UCL), London, UK; 14Department of Hypertension, Chair of Nephrology and Hypertension, Medical University of Lodz, Poland.

*Corresponding author: Prof. Maciej Banach, MD, PhD, FNLA, FAHA, FESC; FASA, Head, Department of Hypertension, WAM University Hospital in Lodz, Medical University of Lodz, Zeromskiego 113; 90-549 Lodz, Poland. Phone: +48 42 639 37 71; Fax: +48 42 639 37 71; E-mail: [email protected]

Conflict of InterestACCEPTED Disclosures: None

Number of words: 2625

ACCEPTED MANUSCRIPT

ABSTRACT

INTRODUCTION: Circulating lipoprotein (a) (Lp(a)) is a recognized risk factor for cardiovascular disease (CVD). Tibolone, a synthetic steroid, may lower Lp(a) levels; however, evidence of the effects of tibolone on Lp(a) still remain to be defined. Therefore, we investigated the effects of tibolone treatment on circulating Lp(a) levels in postmenopausal women. METHODS: The search included PUBMED, Web of Science, Scopus, and Google Scholar (up to January 31 st , 2015) to identify controlled clinical studies investigating the effects of oral tibolone treatment on Lp(a) levels in postmenopausal women. Random-effects meta-regression was performed using unrestricted maximum likelihood method for the association between calculated weighted mean difference (WMD) and potential moderators. RESULTS: Meta-analysis of data from 12 trials (16 treatment arms) suggested a significant reduction of Lp(a) levels following tibolone treatment (WMD: -25.28%, 95% confidence interval [CI]: -36.50, -14.06; p<0.001). This result was robust in the sensitivity analysis and its significance was not influenced after omitting each of the included studies from the meta- analysis. When the studies were categorized accordiMANUSCRIPTng to the tibolone dose, there were consistent significant reductions of Lp(a) in the subsets of studies with doses <2.5 mg/day (WMD: -17.00%, 95%CI: -30.22, -3.77; p<0.012) and 2.5 mg/day (WMD: -29.18%, 95%CI: -45.02, -13.33; p<0.001). Likewise, there were similar reductions in the subsets of trials with follow-up either <24 months (WMD: -26.79%, 95%CI: -38.40, -15.17; p<0.001) or ≥24 months (WMD: - 23.10%, 95%CI: -40.17, -6.03; p=0.008). CONCLUSIONS: This meta-analysis shows that oral tibolone treatment significantly lowers circulating Lp(a) levels in postmenopausal women. Further studies are warranted to explore the mechanism of this effect and the potential value and place of tibolone or its analogues in the treatment of elevated Lp(a) in individuals at risk of CVD. ACCEPTED Keywords: cardiovascular disease, cardiovascular risk, lipoprotein(a), tibolone.

ACCEPTED MANUSCRIPT

INTRODUCTION

Lipoprotein(a) (Lp(a)) is a unique lipoprotein particle, which consists of an apolipoprotein B containing lipoprotein moiety very similar to low-density lipoprotein (LDL) covalently linked to the glycoprotein component apoprotein(a) [apo(a)] 1. Apo(a) is exclusively produced in the liver, and its secretions highly correlated with circulating Lp(a) levels 2. Clinical and epidemiological evidence reveals an increased level of Lp(a) to be a causal, independent risk factor for cardiovascular disease (CVD) 3-6. The atherogenic properties of Lp(a) are associated with several mechanisms, including the inhibition of the fibrinolitic system by homologous structure of apo(a) with plasminogen 7, the interaction with extracellular matrix conjugates such as glycoproteins 8, the binding to scavenger receptors on macrophages 9, 10 , and the induction of inflammatory molecules 11 .

Given the above, there has been an interest in Lp(a) as a target for residual risk therapy 12-14 . In general, circulating Lp(a) levels are genetically MANUSCRIPTdetermined 15 , while the levels can change in certain circumstances, such as under acute vascular and inflammatory pathologies 16-18 . The

European Atherosclerosis Society (EAS) Consensus Panel recommends screening for elevated

Lp(a) in those at intermediate or high CVD/coronary heart disease (CHD) risk, a desirable level

<50 mg/dL as a function of global CV risk, and use of niacin for Lp(a) and CVD/CHD risk reduction 19 . Since currently available LDL cholesterol-lowering drugs, such as statins, have little effects of Lp(a), and other, such as niacin is poorly tolerated and is not available in many countries, there has been a continuous search for effective agents for lowering circulating Lp(a) levels 20 . ACCEPTED

Tibolone (Livial, Org OD 14) is a synthetic steroid with estrogenic, androgenic and progestogenic properties, and used orally for the prevention of osteoporotic bone loss and

ACCEPTED MANUSCRIPT

hormonal replacement treatment on climacteric symptoms in postmenopausal women 21, 22 .

Tibolone itself has no activities, but two estrogenic metabolites (3α- and 3 β-hydroxy [OH] tibolone) and the third metabolite (4-tibolone) exert the effects on climacteric symptoms, and the third metabolite also exerts additional progestogenic effects in endometrium 23, 24 . Tibolone treatment can modulate the lipid profile and its possible reduction of Lp(a) has been observed 25 , although the effects of tibolone on Lp(a) were not previously analysed within systematic review and meta-analysis 22 . Therefore, we performed a meta-analysis to evaluate the efficacy of tibolone treatment on circulating Lp(a) levels in postmenopausal women.

MATERIAL AND METHODS

Search Strategy

This study was designed according to the guidelines of the 2009 preferred reporting items for systematic reviews and meta-analysis MANUSCRIPT (PRISMA) statement 26 . PubMed (http://www.ncbi.nlm.nih.gov/pubmed), Web of Science (http://apps.webofknowledge.com),

SCOPUS (http://www.scopus.com), and Google Scholar (http://www.scholar.google.com) databases were searcher and the search was limited to the controlled studies (mainly randomized controlled trials [RCTs], as well as controlled clinical trials, perspective and open-label studies) carried out up to January 31, 2014, investigating the potential effects of tibolone treatment on

Lp(a) concentrations in the group of postmenopausal women. The databases were searched using the following search terms in titles and abstracts (also in the combination with MESH terms):

(tibolone OR OrgOD14ACCEPTED OR “Org OD14” OR livial OR livial ®) AND (lipoprotein(a) OR

“lipoprotein (a)” OR Lp(a) OR “Lp(a)”). The wild-card term ‘‘*’’ was used to increase the sensitivity of the search strategy. No language restriction was used in the literature search. The

ACCEPTED MANUSCRIPT

search was limited to studies in humans. Selected articles were searched to identify further relevant studies. Two reviewers (KK and AS) evaluated each article separately. Disagreements were resolved by agreement and discussion with a third party (MB).

Study Selection

Original studies (all in postmenopausal women) were included if they met the following inclusion criteria: (i) being a controlled clinical study (RCT, controlled perspective or open-label study), (ii) investigating the impact of tibolone on plasma/serum concentrations of Lp(a), (iii) presentation of sufficient information on Lp(a) concentrations at baseline and at the end of follow-up in each group or providing the net change values.

Exclusion criteria were: (i) lack of a an appropriate control group in the study design, (ii) non-clinical observational studies with case-control, cross-sectional or cohort design, (iii) lack of sufficient information on baseline or follow-up Lp(MANUSCRIPTa) concentrations, (iv) inability to obtain adequate details of study methodology or results from the article or the investigators, and, (v) the study was ongoing.

Data extraction

Eligible studies were reviewed and the following data were abstracted: 1) first author's name; 2) year of publication; 3) study location; 4) study design; 5) tibolone dose and the route of administration; 6) number of participants in the tibolone and control groups; 7) inclusion and exclusion criteria;ACCEPTED 8) age, gender and body mass index (BMI) of study participants; 9) prevalence of diabetes mellitus; and 10) baseline and follow-up plasma concentrations of Lp(a).

ACCEPTED MANUSCRIPT

Quality assessment

A systematic assessment of bias in the included studies was performed using the Cochrane criteria 27 . The items used for the assessment of each study were as follows: adequacy of sequence generation, allocation concealment, blinding, addressing of dropouts (incomplete outcome data), selective outcome reporting, and other potential sources of bias. According to the recommendations of the Cochrane Handbook, a judgment of “yes” indicated low risk of bias, while “no” indicated high risk of bias. Labeling an item as “unclear” indicated an unclear or unknown risk of bias.

Quantitative Data Synthesis

Meta-analysis was conducted using Comprehensive Meta-Analysis (CMA) V2 software

(Biostat, NJ) 28 . Net changes in measurements (change scores) were calculated as follows: measure at end of follow-up − measure at baseline. MANUSCRIPT For cross-over trials, net change in plasma concentrations of Lp(a) were calculated by subtracting the value after control intervention from that reported after treatment. All values were collated in percentage changes from baseline levels. Standard deviations (SDs) of the mean difference were calculated using the following

2 2 formula: SD = square root [(SD pre-treatment ) + (SD post-treatment ) – (2R × SD pre-treatment × SD post- treatment )], assuming a correlation coefficient (R) = 0.5. If the outcome measures were reported in median and inter-quartile range, mean and standard SD values were estimated using the method described by Hozo et al. 29 . Where standard error of the mean (SEM) was only reported, standard deviation (SD) wasACCEPTED estimated using the following formula: SD = SEM × square root ( n), where n is the number of subjects. When the results were presented in multiple time points, only data relating to the longest duration of treatment were considered.

ACCEPTED MANUSCRIPT

A random-effects model (using DerSimonian-Laird method) and the generic inverse variance method were used to compensate for the heterogeneity of studies in terms of demographic characteristics of populations being studied and also differences in study design.

Heterogeneity was quantitatively assessed using I 2 index. Effect sizes were expressed as weighted mean difference (WMD) and 95% confidence interval (CI). In order to evaluate the influence of each study on the overall effect size, sensitivity analysis was conducted using leave- one-out method, i.e. removing one study each time and repeating the analysis 30, 31 .

Meta-regression

Random-effects meta-regression was performed using unrestricted maximum likelihood method to evaluate the association between calculated WMD and potential moderators including duration of treatment with tibolone. MANUSCRIPT Publication bias

Potential publication bias was explored using visual inspection of Begg’s funnel plot asymmetry, fail-safe N test, and Begg’s rank correlation and Egger’s weighted regression tests.

Duval & Tweedie “trim and fill” method was used to adjust the analysis for the effects of publication bias 32 .

RESULTS

The search providedACCEPTED 82 articles excluding duplicates. Of those, 67 were excluded after initial screening and the remaining 15 full text papers were reviewed. Finally 12 were scrutinized as full texts and were selected to be included in the analysis 33-44 . The reasons for excluding the

ACCEPTED MANUSCRIPT

remaining 3 articles were: not controlled for tibolone (n=2) and no tibolone treatment arm (n=1)

(Figure 1 ). In total, 1009 participants were randomized, of whom 634 were allocated to tibolone supplementation group and 375 to control group in the selected studies. The number of participants in these studies ranged from 6 to 136. They were published between 1996 and 2009, and were conducted in USA, Greece, Ireland, the Netherlands (3 trials), Italy (2 trials), UK,

Serbia and Turkey. A range of tibolone doses from 0.3 to 2.5 mg/day was administered in the included trials. Duration of supplementation with tibolone ranged between 3 and 60 months.

Tibolone was safe and well-tolerated in all included studies with no report of any drug-related adverse events. Baseline parameters of the included studies are shown in Table 1.

Risk of bias assessment

According to the Cochrane Collaboration 45 , a specific tool for assessing risk of bias in each involved study comprises judgment of specific featu MANUSCRIPTres of the study. This involves evaluating the risk of bias as ‘low risk’, ‘high risk or ‘unclear risk’. The last category shows either lack of information or uncertainty over the potential for bias. There are seven analyzed domains comprising: sequence generation (selection bias), allocation sequence concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment

(detection bias), incomplete outcome data (attrition bias), selective outcome reporting (reporting bias) and other potential sources of bias ( Table 2 ).

Effect of tiboloneACCEPTED on plasma Lp(a) concentrations

Overall, the impact of tibolone on plasma Lp(a) concentrations were reported in 12 trials comprising 16 treatment arms with 1009 women-patients. The results suggested a significant

ACCEPTED MANUSCRIPT

reduction of Lp(a) levels following treatment with tibolone (WMD: -25.28%, 95%CI: -36.50, -

14.06; p < 0.001), which are robust in the sensitivity analysis, and their significance was not influenced after omitting each of the included studies from meta-analysis (Figure 2 ). When the studies were categorized according to the dose of treatment, there was consistent significant reduction in Lp(a) concentrations in the subsets of trials with doses <2.5 mg/day (WMD: -

17.00%, 95%CI: -30.22, -3.77; p < 0.012) and =2.5 mg/day (WMD: -29.18%, 95%CI: -45.02, -

13.33; p < 0.001) (Figure 3 ). Likewise, there were similar reductions in the subsets of trials lasting either <24 months (WMD: -26.79%, 95%CI: -38.40, -15.17; p<0.001) or ≥24 months

(WMD: -23.10%, 95%CI: -40.17, -6.03; p=0.008) (Figure 4).

Meta-regression

Random-effects meta-regression was performed to assess if the Lp(a) response to tibolone is associated with administered dose and duration ofMANUSCRIPT treatment. The results did not suggest any significant association between the changes in plasma concentrations of Lp(a) with dose (slope:

-6.07; 95%CI: -20.16, 8.02; p = 0.399) and duration of treatment (slope: 0.27; 95%CI: -0.53,

1.08; p = 0.505).

Publication bias

The funnel plot of the study standard error by effect size (WMD) was symmetric, suggesting lack of publication bias in the analysis of the effect tibolone on plasma Lp(a) concentrations.

This finding wasACCEPTED confirmed by the results of Egger’s linear regression (intercept = -3.04, standard error = 1.72; 95%CI: -6.73, 0.65, t = 1.77, df = 14, two-tailed p = 0.099) and Begg’s rank correlation tests (Kendall’s Tau with continuity correction = -0.24, z = 1.31, two-tailed

ACCEPTED MANUSCRIPT

p-value = 0.192) (Figure 5 ). The “fail-safe N” test suggested that 523 studies would be needed to bring the effect size down to a non-significant ( p > 0.05) value.

DISCUSSION

The present meta-analysis of data from 12 trials comprising 16 treatment arms revealed that oral tibolone treatment significantly reduced circulating Lp(a) levels in postmenopausal women.

These results are of high importance because Lp(a) is an independent risk factor for CVD, there are no effective and safe treatments for high Lp(a) levels, and postmenopausal women are at higher CVD risk 3-5.

Tibolone treatment can reduce circulating Lp(a) levels by 25% (42-44). More recently, significant reductions of Lp(a) have been shown with a range of new therapies, including cholesteryl ester transfer protein (CETP) inhibitors, sequence-specific antisense oligonucleotides against apoB, inhibitory monoclonal antibodies to MANUSCRIPTproprotein convertase subtilisin/kexin type-9 (PCSK9), thyroid hormone analogues, and synthetic inhibitors of microsomal triglyceride transfer protein 46-51 . The magnitude of reductions of Lp(a) by these drugs appears to be similar to the ones with tibolone, while the overall magnitude of these recent drugs is still not established 46-48, 52 . The magnitude of reduction of Lp(a) with tibolone is similar to that of niacin.

The influences of the dose and duration of used drugs on Lp(a) is also of concern 25 . Niacin can reduce Lp(a) in a dose-dependent manner 53 . Whereas, the present meta-analysis did not find a dose- or duration-dependent manner of tibolone treatment on Lp(a). This may be an advantage in using tibolone toACCEPTED decrease Lp(a) levels.

The reduction of circulating Lp(a) by tibolone is still incompletely explained. The estrogenic property of tibolone is a possible explanation of the reduction of Lp(a) 25 ; however, the details in

ACCEPTED MANUSCRIPT

the involvement of estrogens in the Lp(a) regulation still remain to be fully clarified 54 . One of the mechanism potentially responsible might be the fact that Lp(a) gene has an estrogen response element and tibolone may active this and reduce hepatic output of Apo(a) 55 . Lp(a) is also decreased by testosterone and the androgenic component of tibolone may have a dual effect 56 .

However, further studies of the mechanism of action of tibolone on Lp(a) metabolism are still required.

The present meta-analysis has limitations. First, the studies included in the present meta- analysis were not always interventional studies focused on the Lp(a) levels, as a main endpoint, therefore well-designed interventional trials specific for Lp(a) are still necessary. Next, the included studies did not evaluate the long-term CVD outcomes associated with Lp(a) reduction.

The reduction of Lp(a) is theoretically thought to be favorable for the prevention of the events, but whether the reduction of Lp(a) with Lp(a)-targeted therapies can lead to the favorable outcomes still needs to be proven, especially that MANUSCRIPTavailable long-term controlled studies with a Lp(a)-lowering drug (nicotinic acid plus laropiprant) have failed to find a positive impact of the

Lp(a) reduction on CV events 19 . Some studies have also indicated that very low Lp(a) levels might be harmful 57, 58 , however Mendelian randomization studies did not confirm that low Lp(a) increases the risk of diabetes 59 . However, tibolone can improve not only Lp(a) level but also vascular functions (e.g. vasodilation) with its estrogenic property 21 . Additionally, tibolone is reported to modulate other lipoprotein factors, i.e. the drug with its androgenic properties can reduce high-density lipoprotein cholesterol (HDL-C) as an unfavorable change 25 . However, there are findingsACCEPTED that the change of HDL-C is usually transient 33 ; on the other hand there are more and more data that we should not focus on HDL-C as it is not main target of the lipid- lowering therapy, also due to the lasting debate on HDL functionality in different patients’

ACCEPTED MANUSCRIPT

groups 42, 60, 61 . It should be emphasized that data on circulating Lp(a) levels were obtained from studies using different assays for Lp(a). It would be also necessary to look at the tibolone effects in general, including the influence on inflammatory markers, and other risk factors of CVD (e.g. blood pressure) 62 , in order to finally confirm the possible benefits of its therapy. Finally, the included studies did not particularly focus on tibolone therapy-related side effects. One should notice that the side effects after tibolone therapy with doses ≤2.5 mg/day are uncommon 63 . In the Long-Term Intervention on Fractures with Tibolone (LIFT) trial 64 in older postmenopausal women (aged 60-85 years) tibolone therapy (in comparison to placebo) significantly reduced the risk of vertebral and nonvertebral fractures, invasive breast cancer and colon cancer, however, increased of the stroke risk. There were also no significant differences in the risk of either CHD or venous thromboembolism 64 . However the above mentioned stroke risk was not observed in other studies with younger postmenopausal women 65, 66 . In conclusion, the present meta-analysis revealed MANUSCRIPT that oral tibolone treatment significantly reduced circulating Lp(a) levels in postmenopausal women. The significance of this metabolic effect for the prevention of CVD needs to be weighed against other effects of tibolone and requires further investigations. Large-scale, well-designed studies may then be justified to explore the value of tibolone treatment in high risk subjects with elevated plasma Lp(a) levels 19 .

ACCEPTED

ACCEPTED MANUSCRIPT

DECLARATION OF INTEREST

This meta-analysis was written independently; no company or institution supported it financially. No authors have any conflict of interest concerning the preparation of this analysis.

Some of the authors have given talks, attended conferences and participated in trials and advisory boards sponsored by various pharmaceutical companies. No professional writer was involved in the preparation of this meta-analysis.

MANUSCRIPT

ACCEPTED

ACCEPTED MANUSCRIPT

REFERENCES:

1. Kronenberg F, Steinmetz A, Kostner GM, Dieplinger H. Lipoprotein (a) in health and disease. Critical reviews in clinical laboratory sciences 1996;33(6):495-543. 2. Lobentanz EM, Dieplinger H. Biogenesis of lipoprotein (a) in human and animal hepatocytes. Electrophoresis 1997;18(14):2677-2681. 3. Collaboration ERF. Lipoprotein (a) concentration and the risk of coronary heart disease, stroke, and nonvascular mortality. JAMA: the journal of the American Medical Association 2009;302(4):412. 4. Dubé JB, Boffa MB, Hegele RA, Koschinsky ML. Lipoprotein (a): more interesting than ever after 50 years. Current opinion in lipidology 2012;23(2):133-140. 5. Kostner KM, März W, Kostner GM. When should we measure lipoprotein (a)? European heart journal 2013;34(42):3268-3276. 6. Banach M, Aronow WS, Serban C, Sahabkar A, Rysz J, Voroneanu L, Covic A. Lipids, blood pressure and kidney update 2014. Pharmacological Research 2015;95-96C:111-125. 7. Angles ‐Cano E, Díaz ADLP, Loyau S. Inhibition of fibrinolysis by lipoprotein (a). Annals of the New York Academy of Sciences 2001;936(1):261-275. 8. Siegel G, Malmsten M, Ermilov E. Anionic biopolyelectrolytes of the syndecan/perlecan superfamily: Physicochemical properties and medical significance. Advances in colloid and interface science 2014;205:275-318. 9. Naruszewicz M, Giroux L-M, Davignon J. Oxidative modification of Lp (a) causes changes in the structure and biological properties of apo (a). Chemistry and physics of lipids 1994;67:167-174. 10. Tsimikas S, Hall JL. Lipoprotein (a) as a potential causal genetic risk factor of cardiovascular disease: a rationale for increased efforts to understand its pathophysiology and develop targeted therapies. Journal of the American College of Cardiology 2012;60(8):716-721. 11. Klezovitch O, Edelstein C, Scanu AM. Stimulation of interleukin-8 production in human THP-1 macrophages by apolipoprotein (a) Evidence forMANUSCRIPT a critical involvement of elements in its C- terminal domain. Journal of Biological Chemistry 2001;276(50):46864-46869. 12. Kolski B, Tsimikas S. Emerging therapeutic agents to lower lipoprotein (a) levels. Current opinion in lipidology 2012;23(6):560-568. 13. Jacobson TA. Lipoprotein (a), cardiovascular disease, and contemporary management. In: Mayo Clinic proceedings; 2013: Elsevier; 2013. p. 1294-1311. 14. Bos S, Yayha R, van Lennep JER. Latest developments in the treatment of lipoprotein (a). Current opinion in lipidology 2014;25(6):452-460. 15. Enkhmaa B, Anuurad E, Zhang W, Tran T, Berglund L. Lipoprotein (a): genotype–phenotype relationship and impact on atherogenic risk. Metabolic syndrome and related disorders 2011;9(6):411-418. 16. Hegele RA, Freeman M, Langer A, Connelly P, Armstrong P. Acute reduction of lipoprotein (a) by tissue-type plasminogen activator. Circulation 1992;85(6):2034-2038. 17. Hong SJ, Seo HS, Park CG, Rha SW, Oh DJ, Ro YM. Serially increasing change in lipoprotein (a) concentration has predictive value in acute vascular events. Annals of clinical biochemistry 2005;42(4):285-291. 18. Ramharack R, Barkalow D, Spahr MA. Dominant negative effect of TGF-β1 and TNF-α on basal and IL-6–inducedACCEPTED lipoprotein (a) and apolipoprotein (a) mRNA expression in primary monkey hepatocyte cultures. Arteriosclerosis, thrombosis, and vascular biology 1998;18(6):984-990. 19. Nordestgaard BG, Chapman MJ, Ray K, Borén J, Andreotti F, Watts GF, et al. Lipoprotein (a) as a cardiovascular risk factor: current status. European heart journal 2010;31(23):2844-2853. 20. Parhofer K. Lipoprotein (a): medical treatment options for an elusive molecule. Current pharmaceutical design 2011;17(9):871. 21. Albertazzi P, Di Micco R, Zanardi E. Tibolone: a review. Maturitas 1998;30(3):295-305.

ACCEPTED MANUSCRIPT

22. Modelska K, Cummings S. Tibolone for postmenopausal women: systematic review of randomized trials. The Journal of Clinical Endocrinology & Metabolism 2002;87(1):16-23. 23. Markiewicz L, Gurpide E. In vitro evaluation of estrogenic, estrogen antagonistic and progestagenic effects of a steroidal drug (Org OD-14) and its metabolites on human endometrium. Journal of steroid biochemistry 1990;35(5):535-541. 24. Kloosterboer H. Tibolone: a steroid with a tissue-specific mode of action. The Journal of steroid biochemistry and molecular biology 2001;76(1):231-238. 25. Gaeta G, Lanero S, Barra S, Silvestri N, Cuomo V, Materazzi C, et al. Sex hormones and lipoprotein (a) concentration. Expert opinion on investigational drugs 2011;20(2):221-238. 26. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 2009;339:b2535. 27. Higgins J. Green S. Cochrane Handbook for systematic reviews of interventions version 5.0. 2 [updated September 2009]. The Cochrane Collaboration, 2009. www cochrane-handbook org/(accessed 18 May 2009) 2013. 28. Borenstein M, Hedges L, Higgins J, Rothstein H. Comprehensive meta-analysis version 2. Englewood, NJ: Biostat 2005;104. 29. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC medical research methodology 2005;5(1):13. 30. Sahebkar A. Does PPAR γ2 gene Pro12Ala polymorphism affect nonalcoholic fatty liver disease risk? Evidence from a meta-analysis. DNA and cell biology 2013;32(4):188-198. 31. Sahebkar A. Are Curcuminoids Effective C ‐Reactive Protein ‐Lowering Agents in Clinical Practice? Evidence from a Meta ‐Analysis. Phytotherapy research 2014;28(5):633-642. 32. 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(2):455-463. 33. Bjarnason NH, Bjarnason K, Haarbo J, Bennink HJC, Christiansen C. Tibolone: influence on markers of cardiovascular disease. The Journal of Clinical Endocrinology & Metabolism 1997;82(6):1752-1756. MANUSCRIPT 34. Gallagher J, Baylink DJ, Freeman R, McClung M. Prevention of bone loss with tibolone in postmenopausal women: results of two randomized, double-blind, placebo-controlled, dose- finding studies. The Journal of Clinical Endocrinology & Metabolism 2001;86(10):4717-4726. 35. Kalogeropoulos S, Petrogiannopoulos C, Gagos S, Kampas N, Kalogeropoulos G. The influence of 5-year therapy with tibolone on the lipid profile in postmenopausal women with mild hypercholesterolemia. Gynecological endocrinology 2004;18(4):227-232. 36. Kroiss R, Fentiman I, Helmond F, Rymer J, Foidart J-M, Bundred N, et al. The effect of tibolone in postmenopausal women receiving tamoxifen after surgery for breast cancer: a randomised, double ‐blind, placebo ‐controlled trial. BJOG: An International Journal of Obstetrics & Gynaecology 2005;112(2):228-233. 37. Lloyd G, McGing E, Cooper A, Patel N, Lumb P, Wierzbicki A, et al. A randomised placebo controlled trial of the effects of tibolone on blood pressure and lipids in hypertensive women. Journal of human hypertension 2000;14(2):99-104. 38. Milner MH, Sinnott MM, Cooke TM, Kelly A, McGill T, Harrison RF. A 2-year study of lipid and lipoprotein changes in postmenopausal women with tibolone and estrogen-progestin. Obstetrics & Gynecology 1996;87(4):593-599. 39. Ostberg JE,ACCEPTED Damjanovic T, Dimkovic N, Byrne D, Mikhailidis DP, Prelevic GM. Effect of tibolone on markers of cardiovascular disease risk in postmenopausal women undergoing hemodialysis: a pilot study. Fertility and sterility 2004;81(6):1624-31. 40. Perrone G, Capri O, Galoppi P, Brunelli R, Bevilacqua E, Ceci F, et al. Effects of either tibolone or continuous combined transdermal estradiol with medroxyprogesterone acetate on coagulatory factors and lipoprotein (a) in menopause. Gynecol Obstet Invest 2009;68(1):33-9.

ACCEPTED MANUSCRIPT

41. Von Eckardstein A, Schmiddem K, Hövels A, Gülbahçe E, Schuler-Lüttmann S, Elbers J, et al. Lowering of HDL cholesterol in post-menopausal women by tibolone is not associated with changes in cholesterol efflux capacity or paraoxonase activity. Atherosclerosis 2001;159(2):433- 439. 42. Von Eckardstein A, Crook D, Elbers J, Ragoobir J, Ezeh B, Helmond F, et al. Tibolone lowers high density lipoprotein cholesterol by increasing hepatic lipase activity but does not impair cholesterol efflux. Clinical endocrinology 2003;58(1):49-58. 43. Anedda FM, Velati A, Lello S, Orrù M, Paoletti AM, Melis GB, et al. Observational study on the efficacy of tibolone in counteracting early carotid atherosclerotic lesions in postmenopausal women. Hormone research 2003;61(1):47-52. 44. Demirol A, Guven S, Guven ESG, Kirazli S, Gurgan T, Ayhan A. Comparison of the effects of tibolone and estrogen therapy on hemostasis in surgical menopause: a randomized, double-blind, placebo-controlled study. Fertility and sterility 2007;87(4):842-848. 45. Collaboration C. Cochrane handbook for systematic reviews of interventions version 5.1. 0. 2011. URL: http://handbook cochrane org/[accessed 2014-03-12][WebCite Cache] 2013. 46. Bochem A, Kuivenhoven J, Stroes E. The promise of cholesteryl ester transfer protein (CETP) inhibition in the treatment of cardiovascular disease. Curr Pharm Design 2013;19(17):3143-3149. 47. Norata GD, Ballantyne CM, Catapano AL. New therapeutic principles in dyslipidaemia: focus on LDL and Lp (a) lowering drugs. European heart journal 2013;34(24):1783-1789. 48. Sahebkar A, Watts GF. New therapies targeting apoB metabolism for high-risk patients with inherited dyslipidaemias: what can the clinician expect? Cardiovascular drugs and therapy 2013;27(6):559-567. 49. Banach M, Rizzo M, Obradovic M, Montalto G, Rysz J, P Mikhailidis D, et al. PCSK9 inhibition-a novel mechanism to treat lipid disorders? Current pharmaceutical design 2013;19(21):3869-3877. 50. Dragan S, Serban MC, Banach M. Proprotein convertase subtilisin/kexin 9 inhibitors: an emerging lipid-lowering therapy? J Cardiovasc PharmMANUSCRIPTacol Ther 2015;20(2):157-68. 51. Taub R, Chiang E, Chabot-Blanchet M, Kelly MJ, Reeves RA, Guertin M-C, et al. Lipid lowering in healthy volunteers treated with multiple doses of MGL-3196, a liver-targeted thyroid hormone receptor-β agonist. Atherosclerosis 2013;230(2):373-380. 52. Banach M, Serban C, Aronow WS, Rysz J, Dragan S, Lerma EV, et al. Lipid, blood pressure and kidney update 2013. International urology and nephrology 2014;46(5):947-961. 53. G Parhofer K. Lipoprotein (a): medical treatment options for an elusive molecule. Curr Pharm Design 2011;17(9):871-876. 54. McCarty MF. Estrogen agonists/antagonists may down-regulate growth hormone signaling in hepatocytes–An explanation for their impact on IGF-I, IGFBP-1, and lipoprotein (a). Medical hypotheses 2003;61(3):335-339. 55. Hoover-Plow J, Huang M. Lipoprotein (a) metabolism: potential sites for therapeutic targets. Metabolism: clinical and experimental 2013;62(4):479-491. 56. Zmuda JM, Thompson PD, Dickenson R, Bausserman LL. Testosterone decreases lipoprotein (a) in men. The American journal of cardiology 1996;77(14):1244-1247. 57. Kotani K, Sakane N. Carotid intima-media thickness in asymptomatic subjects with low lipoprotein (a) levels. Journal of clinical medicine research 2012;4(2):130. 58. Ishikawa S, Kotani K, Kario K, Kayaba K, Gotoh T, Nakamura Y, et al. Inverse association between serumACCEPTED lipoprotein (a) and cerebral hemorrhage in the Japanese population. Thrombosis research 2013;131(2):e54-e58. 59. Kamstrup PR, Nordestgaard BG. Lipoprotein (a) concentrations, isoform size, and risk of type 2 diabetes: a Mendelian randomisation study. The Lancet Diabetes & Endocrinology 2013;1(3):220-227.

ACCEPTED MANUSCRIPT

60. Dragan S, Serban C, Banach M. Can We Change the Functionality of HDL Cholesterol with Nonpharmacological and Pharmacological Agents? Current medicinal chemistry 2014;21(25):2927-2946. 61. Otocka-Kmiecik A, Mikhailidis DP, Nicholls SJ, Davidson M, Rysz J, Banach M. Dysfunctional HDL: a novel important diagnostic and therapeutic target in cardiovascular disease? Progress in lipid research 2012;51(4):314-324. 62. Prelevic GM, Kwong P, Byrne DJ, Jagroop IA, Ginsburg J, Mikhailidis DP. A cross-sectional study of the effects of hormon replacement therapy on the cardiovascular disease risk profile in healthy postmenopausal women. Fertility and sterility 2002;77(5):945-951. 63. Biglia N, Maffei S, Lello S, Nappi RE. Tibolone in postmenopausal women: a review based on recent randomised controlled clinical trials. Gynecological Endocrinology 2010;26(11):804-814. 64. Cummings SR, Ettinger B, Delmas PD, Kenemans P, Stathopoulos V, Verweij P, et al. The effects of tibolone in older postmenopausal women. New England Journal of Medicine 2008;359(7):697-708. 65. Hudita D, Posea C, Ceausu I, Rusu M. Efficacy and safety of oral tibolone 1.25 or 2.5 mg/day vs. placebo in postmenopausal women. European review for medical and pharmacological sciences 2003;7:117-126. 66. Morais-Socorro M, Cavalcanti MA, Martins R, Neto Francisco P, Rezende A, Azevedo G, Almeida M. et al. Safety and efficacy of tibolone and menopausal transition: a randomized, double-blind placebo-controlled trial. Gynecological Endocrinology 2012;28(6):483-487.

MANUSCRIPT

ACCEPTED

ACCEPTED MANUSCRIPT

TABLES

Table 1. Demographic characteristics of the included studies

Study Bjarnaso Gallagher Kalogeropou- Kroiss et Lloyd et al. Milner et al. Ostberg et Perrone Von Von Anedda et Demirol n et al. 33 et al. 34 los et al. 35 al. 36 37 38 al. 39 et al. 40 Eckardstein Eckardstein al. 43 et al. 44 et al. 41 et al. 42 Year 1997 2001 2004 2005 2000 1996 2004 2009 2001 2003 2003 2007 Location Netherlan USA Greece Netherlands UK Ireland Serbia Italy Netherlands Germany Italy Turkey ds Design Randomi Randomiz Clinical Randomized Randomized Randomized One-year Prospecti Randomized Randomized Randomized Randomiz zed ed, controlled trial , double- double-blind, placebo- open-label ve , double- , double- double-blind, ed double- double- double- blind, placebo- controlled, controlled controlled blind and blind, placebo blind, blind, blind, placebo- controlled pilot study study study placebo- placebo- controlled placebo- placebo placebo controlled, study controlled controlled study controlled controlled controlled pilot study study study study study , dose- finding studies Duration of 24 month 24 month 60 month 12 months 6 month 24 month 12 month 12 month 3 month 3 month 6 month 6 month study MANUSCRIPT (84 days) (84 days) Inclusion Healthy Healthy Non-smoking Postmenopa Hypertensive Patients at Postmenopa No Postmenopa Healthy Healthy Healthy criteria women Caucasian women, usal women patients on or least 45 usal women natural usal women postmenopa women (mean menopaus not being or Asian undergoing (≤75 years off treatment years of age, undergoing menstruat (absence of usal women age 52.0 ± al women treated women menopause with old; BMI with postmenopa chronic ion for at any with 0.34 years) 45–50 with any medicatio 45years of normal BP, 18–30 antihypertens usal for 12 hemodialysi least 12 menstrual previous who were at years of 2 n known age or without kg/m ) with ive, with 1 or month or s, with at months, bleeding for oophorecto least 6 months age who to affect older, cardiovascular newly more years of more, within least 6 FSH 12 months my or 12 beyond a 12 had coagulati within problems and diagnosed postmenopau 15% of ideal months of >60 and FSH months month period undergone on 80–130% osteopenia, but and sal on body weight amenorrhea, pg/ml, > 50 IU/l passed after of hysterecto fibrinolys of ideal with mild histologicall clinical or height age less than estradiol and estradiol the last hypergonadotr my and is lipid or body hypercholesterol grounds and had 65 years, <20 < 20 pg/ml), menstrual opic bilateral bone y confirmed metabolis weight,1y emia (TC 241+7 invasive or intact uteri and serum g/ml and 45 to 60 bleed, FSH amenorrhea salpingo- m and ear or mg/dl; LDL-C non-invasive and ovaries FSH the years of age > 50 IU/l and had never oophorec- with more more but 153+9 mg/dl) ACCEPTEDearly stage (>20 mIU/L) presence with a BMI and serum begun HRT tomy as a than 10 4 years or breast of ranging levels of result of years of less past cancer, with menopaus between 20 estradiol < benign spontaneo natural surgical al and 28 70 pmol, diseases us menopaus treatment symptoms kg/m 2 aged 45–60

ACCEPTED MANUSCRIPT

menopaus e and followed by years, BMI e without tamoxifen 20–28 kg/m 2 osteoporo sis Route of oral oral oral oral oral oral oral oral oral oral oral oral administrat ion Tibolone 1.25 0.3 2.5 mg/day 2.5 mg/ day 2.5 mg/day 2.5 mg/day 2.5 mg three 2.5 2.5 mg /day 2.5 mg /day 2.5 mg/day 2.5 dose mg/day or mg/day times per mg/day mg/day 2.5 or week mg/day 0.625 mg/day or 1.25 mg/day or 2.5 mg/day

Partici- Case 29 a 132 c 53 35 e 19 31 f 13 20 g 12 34 11 28 h pants 136 d 28 b 127 a MANUSCRIPT32 b 21 b 28 b 131 b Contr 13 130 29 35 11 50 13 19 6 31 11 27 ol Age (years) Case 66.4 52.3 (2.9) c 49.6±5 58 (6.1) e 60.11 (1.28) 52.4±0.74 f 54.5 (4.6) 51.1 ± 45 - 60 54.5 ± 3.7 51.6±0.49 48±0.6 h (7.0) a 52.8 (3.5) d 4.1 g 65.5 52.4 (3.2) a 53.6±0.77 b 53.4 ± 46±3 b (7.5) b 52.1 (3.3) b 3.6 b Contr 68.4 (6.1) 52.6 (3.4) 48.4±4 59 (5.5) 62.73 (1.79) 55.6±0.61 56.0 (7.4) 52.7 ± 45 - 60 53·5 ± 3·7 52.1±0.44 47± 0.6 ol 3.3 Male (%) Case 0a 0c 0 0e 0 0f 0 0g 0 0 0 0h 0d 0b 0a 0b 0b 0b 0b ACCEPTED Contr 0 0 0 0 0 0 0 0 0 0 0 0 ol BMI Case 24.5 24.8 (3.8) c 22–26 24.8 (3.2) e* 33.90 (5.99) 41.1± 0.6f 23.9 (4.0) 24 ± 2.7 g 20-28 24.5 ± 2.5 21.2±0.32 25.4± 0.3 h (kg/m 2) (3.4) a 24.8 (3.6) d

ACCEPTED MANUSCRIPT

23.1 25.5 (3.5) a 39.6± 0.7b 25.2 ± 24.4± 0.7 b (2.8) b 25.2 (3.7) b 2.6 b Contr 24.7 (3.2) 25.4 (3.8) 22–25 25.4 (3.7) 29.60 (1.30) 39.2± 0.5 21.2 (3.0) 23.2 8 20-28 23.7 ± 2.7 20.9±0.22 25.1± 0.1 ol 2.6 prevalence Case NS a 0c NS NS e NS NS f 0 NS g NS 0 NS 0h of diabetes 0d mellitus NS b 0a NS b NS b 0b (%) 0b Contr NS 0 NS NS NS NS 0 NS NS 0 NS 0 ol Lp(a) Case 23.1 NS c 25±4 20.4 10 (7.5–23) 14.1 (0.9- 19 (8–117) 28.0 45 (2-229) 27 ± 32 37.5±3.3 35.9 ± (mg/dL) (12.6– (25.1)e** 106.1) f (24.4) g 20.8 h NS d 44.9) a 25.8 NS a 39.4 (1.6- 26.7 38.5± 17.8 (8.95– NS b 122.4) b (17.1) b b 36.9) b Contr 24.3 NS 26±5 17.8 14 (5–66) 12.6 (0.6- 16 (8–84) 28.6 17 (3-84) 27 ± 34 35.8±5.6 39.8 ± ol (16.6– (23.9)** 87.5) (22.0) 18.2 51.2) Values are expressed as mean ± SD or median (25–75 percentiles). MANUSCRIPT ABBREVIATIONS: BMI: body mass index; BP: blood pressure; NS: not stated; TC: total cholesterol; LDL-C: low-density lipoprotein cholesterol; DM: diabetes mellitus; FSH: serum levels of follicle stimulating hormone; HRT: hormonal replacement therapy; adenotes 1.25 mg/day tibolone; bdenotes 2.5 mg/day tibolone; cdenotes 0.3 mg/day tibolone; ddenotes 0.625 mg/day tibolone; edenotes 20 mg/day oral tamoxifen (Nolvadex-D; Astra-Zeneca) plus either 2.5 mg/ day oral tibolone (Livial; NVOrganon), fdenotes a combined estrogen (0.625 mg/day) – progestin (0.15 mg for 12 days in 28) preparation; g denotes oral medroxyprogesterone acetate (MPA; 10 mg/day for 12 days); hdenotes conjugated ET (Premarin, 0.625 mg/d oral tablets; Wyeth, Istanbul, Turkey); *denotes patients belonging to tibolone group n = 34; ** denotes patients belonging to tibolone group or placebo group n = 29.

ACCEPTED

ACCEPTED MANUSCRIPT

Table 2. Risk of bias assessment in the studies considered for meta-analysis. Study Ref SEQUENCE ALLOCATION BLINDING OF BLINDING OF INCOMPLETE SELECTIVE OTHER GENERATION CONCEALMENT PARTICIPANTS OUTCOME OUTCOME OUTCOME POTENTIAL AND PERSONNEL ASSESSMENT DATA REPORTING THREATS TO VALIDITY

Bjarnason et al. 33 U U L L L L L Gallagher et al. 34 U H L L L L L Kalogeropoulos et al. 35 U U U U L L L Kroiss et al. 36 L L L L L L L Lloyd et al. 37 U U L L L L L Milner et al. 38 L U U U L L L Ostberg et al. 39 U U U U L L L Perrone et al. 40 U U U U L L L von Eckardstein et al. 41 U U L L L L L von Eckardstein et al. 42 U U L L L L L Anedda et al. 43 U U L L L L L Demirol et al. 44 U U L L L L L L: low risk of bias; H: high risk of bias; U: unclear risk of bias MANUSCRIPT

ACCEPTED

ACCEPTED MANUSCRIPT

FIGURE LEGENDS

Figure 1. Flow chart of the number of studies included to the meta-analysis.

Figure 2. Forest plot displaying weighted mean difference and 95% confidence intervals for the impact of tibolone on circulating Lp(a) levels. Lower plot shows leave-one-out sensitivity analysis.

Figure 3. Forest plot displaying weighted mean difference and 95% confidence intervals for the impact of tibolone on circulating Lp(a) levels in trials administering doses < 2.5 mg/day (upper plot) and ≥ 2.5 mg/day (lower plot).

Figure 4. Forest plot displaying weighted mean difference and 95% confidence intervals for the impact of tibolone on circulating Lp(a) levels in trials lasting < 24 months (upper plot) and ≥ 24 months (lower plot).

Figure 5. Meta-regression plots of the association between mean changes in circulating Lp(a) levels s and dose and duration of treatment with tibolone. The size of each circle is inversely proportional to the variance of change. MANUSCRIPT Figure 6. Funnel plot detailing publication bias in the studies reporting the impact of tibolone on circulating Lp(a) levels. Open diamond represents observed effect size; closed diamond represents imputed effect size.

ACCEPTED

ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT HIGHLIGHTS:

• Circulating Lp(a) is a recognized risk factor for cardiovascular disease (CVD). • Tibolone may lower Lp(a), however, evidence of this effect remains to be confirmed. • Meta-analysis suggests a significant Lp(a) reduction following tibolone (by 25.28%). • Further studies are warranted to explore the mechanism of this effect. • A place of tibolone in individuals at CVD risk needs to be further investigated.

MANUSCRIPT

ACCEPTED