Summary
The pleiotropic effects of statins, beyond their cholesterol-lowering properties, are much debated. In primary prevention, several observational cohort and case-control studies appear to show that statins reduce the incidence of venous thromboembolism by about 30%. In a single randomized placebo-controlled clinical trial (JUPITER), which included 17,000 patients, rosuvastatin 20 mg/day reduced the risk of venous thromboembolism by 43%. However, these patients were at low risk of venous thromboembolism, and the frequency of the event was, in principle, low. In secondary prevention, several observational studies and post-hoc analyses of randomized clinical trials have suggested that statins may prevent recurrence of venous thromboembolism. However, none of these studies had enough scientific weight to form the basis of a recommendation to use statins for secondary prevention. The putative preventive effect of statins appears to be independent of plasma cholesterol concentration and could be a pharmacological property of the statin class, although a dose-effect relationship has not been demonstrated. The mechanism through which statins might prevent venous thrombosis is thought to involve their anti-inflammatory and antioxidant effects or perhaps a more specific action, by blocking the degradation of antithrombotic proteins. A mechanism involving the action of statins on interactions between risk factors for atherosclerosis and venous thromboembolism is supported by some studies, but not all. In the absence of firm evidence, statins cannot currently be recommended for primary or secondary prevention of venous thromboembolism.
Résumé
Au-delà de la réduction du taux de cholestérol plasmatique, les effets pléiotropes des statines sont discutés. En prévention primaire, plusieurs études observationnelles de type registres ou études cas-témoins semblent montrer un effet réducteur des statines d’environ 30 % sur la fréquence de la maladie veineuse thromboembolique. Un seul essai clinique contrôlé et randomisé (JUPITER), comparant la rosuvastatine 20 mg/jour au placebo et ayant inclus 17 000 patients, a montré une réduction du risque de maladie thromboembolique veineuse de 43 % dans le groupe statine. Il s’agissait, cependant, de patients à faible risque de maladie thromboembolique veineuse chez lesquels la fréquence de l’évènement était a priori basse. En prévention secondaire, quelques études observationnelles ou analyses post-hoc d’essais cliniques randomisés suggèrent un possible effet préventif des statines sur la récidive de maladie thromboembolique veineuse. Cependant, aucune de ces études n’avait le poids scientifique permettant de recommander l’utilisation des statines en prévention secondaire. L’effet préventif possible des statines paraît indépendant des taux plasmatiques du cholestérol et pourrait être relié à la classe pharmacologique elle-même sans qu’une relation dose-effet puisse être construite. Le mécanisme préventif des statines sur le thrombus veineux ferait intervenir les effets anti-inflammatoires et anti-oxydants des statines, voire une action plus spécifique bloquant la dégradation des protéines anti-thrombotiques. Une action des statines via des liens entre les facteurs de risque de l’athérosclérose et la maladie thromboembolique veineuse est discutée. Aujourd’hui, en l’absence de preuves fermement établies, les statines ne peuvent être recommandées en prévention primaire ou secondaire de la maladie thromboembolique veineuse.
Background
Statins reduce the risk of onset or recurrence of conditions and events caused by atheroma, such as coronary heart disease, angina, myocardial infarction, peripheral artery disease and stroke . This effect is observed regardless of initial cholesterol concentration in patients at both high and low risk of cardiovascular disease . In addition to their cholesterol-lowering effect, the pleiotropic effects of statins, based mainly on their anti-inflammatory and antioxidant properties, have been debated ever since they came into use. A preventive effect was thus postulated in heart failure, certain cancers, osteoporosis and dementia. A number of observational studies published in the 2000s suggested that statins might also have a preventive effect against venous thromboembolism (VTE) . More recently, the randomized placebo-controlled trial JUPITER provided more tangible evidence of the possible efficacy of statins in the primary prevention of VTE . These studies employed very different designs and methodologies, and the conclusions of meta-analyses have been inconsistent . High quality evidence for the value of statins in secondary prevention is lacking, as the current data were obtained from observational studies or post-hoc analyses of clinical trials . We will consider in turn the results of the main studies published on the use of statins in primary and secondary prevention of VTE, before addressing mechanisms that could explain their possible efficacy, a subject that remains a source of debate.
Statins and primary prevention of VTE
Not long after statins were first marketed, it was hypothesized that they might have pharmacological properties beyond simply lowering plasma cholesterol concentrations. The results of observational cohort and case-control studies published in the 2000s hinted at a possible preventive effect against VTE ( Table 1 ). This aspect of the pharmacological properties of statins was first studied in postmenopausal women, a population at increased risk of both arterial and venous thromboembolic events. The Heart and Estrogen/progestin Replacement Study (HERS) was a randomized clinical trial designed to study the effects of oestrogen and progesterone supplementation on cardiovascular morbidity and mortality, conducted in 2763 postmenopausal women with coronary heart disease . The authors later compared the incidence of VTE events in the statin users included in the trial with that observed in the non-users. Statin use reduced the frequency of VTE events by 50% (relative hazard 0.5, 95% confidence interval [CI] 0.2–0.9). However, as this was a post-hoc analysis, it only provided low-level evidence. A case-control study was conducted by the health insurer Group Health Cooperative in Washington State, USA, between January 1995 and December 2000, in postmenopausal women aged between 30 and 89 years . A group of 465 women who had a first VTE during the study period was matched with 1962 controls. At the time of their inclusion in the study, 4.5% of the cases and 5.6% of the controls were using a statin, mainly simvastatin or pravastatin. In this study, statin use was associated with a 16% reduction in the risk of VTE (odds ratio [OR]: 0.84, 95% CI: 0.51–1.37). This reduction was unaffected by adjustment for hormone replacement therapy use, body weight or the presence of diabetes.
| Study | Design | OR [95% CI] | Level of evidence a |
|---|---|---|---|
| HERS; Grady et al., 2000 | RCT; post-hoc analysis in 2763 postmenopausal women with CHD | 0.50 [0.2–0.9] | Low |
| GHC; Doggen et al., 2004 | Case-control in postmenopausal women (465 cases, 1962 controls) | 0.84 [0.51–1.37] | Low |
| Ontario Residents Cohort Study; Ray et al., 2001 | Retrospective cohort study in 125,862 men and women aged > 65 years; 77,993 statin users | Total 0.78 [0.69–0.87]; women 0.72 [0.63–0.82]; men 0.97 [0.78–1.22] | Low |
| EDITH; Lacut et al., 2004 | Case-control (377 patients/group); hospital-based | 0.42 [0.23–0.76] | Low |
| MEGA; Ramcharan et al., 2009 | Case-control (4538 cases, 5914 controls); population-based | 0.55 [0.46–0.67] | Low |
| Olmsted County; Ashrani et al., 2015 | Case-control (1340 cases, 1538 controls) | 0.73 [0.55–0.96] | Low |
| JUPITER; Glynn et al., 2009 | RCT (17,804 patients); rosuvastatin 20 mg/day versus placebo | 0.57 [0.37–0.86] | High |
a Levels of evidence are based on the recommendations of the French National Authority for Health (HAS) .
Using data from Ontario provincial healthcare administrative databases, Ray et al. conducted a retrospective population-based study in Canada to determine the influence of statin use on the incidence of VTE . The cohort consisted of 125,862 men and women aged > 65 years with neither cancer nor cardiovascular disease. The incidence of VTE in a group of 77,993 statin users was compared with that of a group of 11,891 patients prescribed a non-statin lipid-lowering drug, and 35,978 patients for whom thyroid hormone replacement had been prescribed. The authors postulated that patients with hypothyroidism were not at increased risk of VTE, and could therefore constitute a control group. Compared with the control group, the VTE risk was 22% lower among statin users (adjusted hazard ratio [HR]: 0.78, 95% CI: 0.69–0.87). The reduction was significant when only women were considered (HR: 0.72, 95% CI: 0.63–0.82), but not in men (HR: 0.97, 95% CI: 0.78–1.22). No risk reduction was observed in patients receiving non-statin lipid-lowering therapy (HR: 0.97, 95% CI: 0.79–1.18). Despite the large number of patients included in this study, it had certain limitations: its retrospective design, the failure to discriminate between the various statins available on the market at that time, and the potential confounding role of hormone replacement therapy among the women. Furthermore, the characteristics of these patients’ lipid abnormalities were unknown.
Lacut et al. evaluated the environmental and genetic risk factors for VTE in the case-control EDITH study, conducted in Brest, France . The authors analysed the role of statins in 377 patients hospitalized for pulmonary embolism and/or deep vein thrombosis, and the same number of controls hospitalized in cardiology or internal medicine at the same hospital. The cases and controls were matched for VTE risk factors. The higher frequency of statin use in the control group (9.5% of patients) compared with the VTE group (5%) equated to a risk reduction of 58% (OR: 0.42, 95% CI: 0.23–0.76; P = 0.002). This reduction was not observed in fibrate users (OR: 1.38, 95% CI: 0.76–2.52; P = 0.26). The risk reduction was at the limit of significance for aspirin (OR: 0.66, 95% CI: 0.42–1.05; P = 0.06), and was non-significant for thienopyridine users (OR: 1.07, 95% CI: 0.48–2.41; P = 0.85). After adjustment for aspirin use or atherosclerotic disease, statins retained a protective effect against VTE, suggesting that the effect is independent of these confounding factors. However, because this was a non-interventional case-control study, it only provides low-level evidence.
The MEGA (Multiple Environmental and Genetic Assessment of risk factors for VTE) study employed a population-based case-control study design . This observational study, conducted at six hospitals in the Netherlands between 1999 and 2004, measured the influence of various drug regimens on the incidence of VTE, by comparing 4538 patients with VTE with 5914 controls. Only 3% of patients in the VTE group were using a statin, versus 6% in the control group, corresponding to a risk reduction of 45% (OR: 0.55, 95% CI: 0.46–0.67). No risk reduction was seen with use of a non-statin lipid-lowering drug (OR: 0.88, 95% CI: 0.46–1.71). Antiplatelet therapy significantly reduced the risk of VTE by 24% (OR: 0.76, 95% CI: 0.60–0.95), and simultaneous use of aspirin and antiplatelet therapy reduced the risk by 62% (OR: 0.38, 95% CI: 0.25–0.57). The authors also compared the impact of the various statins used on VTE risk. The majority of patients were using simvastatin, atorvastatin or pravastatin, and no difference was found in the risk reduction observed with these three drugs. There were too few rosuvastatin and fluvastatin users to draw any valid conclusions about these statins. The statin doses used were not specified in this study, and it was therefore impossible to determine whether a dose-effect relationship exists. And again, as this was an observational case-control study, the level of evidence it provides is low.
Recently, Ashrani et al. performed a case-control study using the longitudinal population-based resources of the Rochester Epidemiology Project to compare all Olmsted County residents experiencing a first VTE event between 1988 and 2000 with age- and sex-matched controls (one or two controls for each case) . Lipid-lowering therapy was associated with a lower risk of VTE (OR: 0.73, 95% CI: 0.55–0.96; Table 1 ), and the association was statistically stronger for statins. However, after adjustment for common VTE risk factors, the association remained consistent but not statistically significant (OR: 0.67, 95% CI: 0.43–1.04).
JUPITER was a double-blind randomized placebo-controlled clinical trial in patients with a cholesterol concentration below the accepted threshold for cardiovascular risk (low-density lipoprotein cholesterol < 1.3 g/L) but with high-sensitivity C-reactive protein (hs-CRP) concentrations of 2 mg/L or higher . This patient group was chosen because of the observation that hs-CRP concentration might be a marker for cardiovascular risk. As statins reduce hs-CRP concentrations, the authors postulated that these drugs might reduce the incidence of cardiovascular events in a population whose cardiovascular risk was theoretically low. More than 17,000 patients were randomized to receive either rosuvastatin 20 mg/day or placebo. The primary outcome was the occurrence of a first major cardiovascular event. In JUPITER, rosuvastatin produced a reduction of 44% in the primary endpoint (HR: 0.56, 95% CI: 0.46–0.69; P < 0.00001) ( Table 2 ). Analysis of the individual components of the primary endpoint showed a 54% risk reduction for fatal or non-fatal myocardial infarction (HR: 0.46, 95% CI: 0.30–0.70; P < 0.0002) and a 48% reduction for fatal or non-fatal stroke (HR: 0.52, 95% CI: 0.34–0.79; P < 0.002). Hypothesising that rosuvastatin may also act on venous thrombus formation, the authors prespecified that the frequency of VTE would also be measured in each randomization arm of this trial . A significant result was observed for this endpoint too, with a 43% reduction in VTE events in the rosuvastatin arm (HR: 0.57, 95% CI: 0.37–0.86; P = 0.007). The reduction was significant for provoked VTE (HR: 0.52, 95% CI: 0.28–0.96; P = 0.03), but non-significant for unprovoked events (HR: 0.61, 95% CI: 0.35–1.09; P = 0.09). Rosuvastatin reduced the risk of deep vein thrombosis by 55% (HR: 0.45, 95% CI: 0.25–0.79; P = 0.004), but the risk of pulmonary embolism was not significantly reduced (HR: 0.77, 95% CI: 0.41–1.45; P = 0.42). Subgroup analysis showed a 50% risk reduction in men (HR: 0.50, 95% CI: 0.30–0.84; P < 0.05), but a non-significant reduction in women (HR: 0.74, 95% CI: 0.35–1.56). This difference could be related to the fact that the trial contained more men than women (62% men). A 45% reduction in the risk of VTE was demonstrated in patients aged between 50 and 69 years (HR: 0.55, 95% CI: 0.31–0.96; P < 0.05), while the reduction was not significant among participants aged 70–97 years (HR: 0.59, 95% CI: 0.31–1.11). The authors concluded that rosuvastatin reduces the risk of VTE to the same degree as it reduces the incidence of arterial events, and that these two effects are independent. Furthermore, the total number of venous events (94 cases) was similar to the total number of cases of stroke (97 cases) and myocardial infarction (99 cases), showing that, in this population, VTE was as common as these major life-threatening cardiac and vascular events. This study also confirmed that the effect of statins on the incidence of VTE was independent of low-density lipoprotein cholesterol, high-density lipoprotein cholesterol and triglyceride concentrations, a finding consistent with earlier observations that non-statin lipid-lowering agents have no effect on VTE risk. The relatively high dose of rosuvastatin used in this study could argue in favour of a dose-dependent effect, consistent with earlier data suggesting a preventive effect for high doses of statins on VTE risk . However, because the effect on VTE is modest, many patients (714 per year) need to be treated in order to prevent one event.
