Results
We screened 8,623 titles and abstracts and selected 80 studies for full-text screening. Of these, eight studies were included (Figure 1 ). The most common reason for excluding studies during full-text screening was due to the studies not concerning newly initiated statin treatment (n = 29). A full overview of the included studies can be found in Table 1 . Out of the eight included studies, four studies investigated the effect of a specific statin (simvastatin[13, 14], rosuvastatin[15], and atorvastatin[16]), while three studies investigated the effect of more than one statin type.[17-19] Finally, one study did not specify which statin type the patients were initiating treatment with.[20] Concerning the VKA-treatment, seven studies included patients in warfarin treatment only,[13-16],[18-20] whereas one study concerned patients in phenprocoumon and acenocoumarol treatment.[17] The size of the included population ranged from 7 to 5,637 patients in six of the studies, while one study did not report the size of the study population.[20] Although all eight studies met our inclusion criteria, only seven studies focused specifically on the effect of initiating statin treatment on coagulation activity,[13-19] while one study addressed several drug interactions with warfarin, including statins.[20] Two of the included studies were prospective controlled drug trials, comprising patients in stable warfarin treatment receiving statin treatment only due to the trial, and not due to any known medical condition.[15, 16]
The three studies that focused on simvastatin initiation in patients treated with warfarin, all reported a small increase in mean INR. The largest study, including data from 5,637 warfarin users, found an increase in mean INR from 2.43 at baseline to 2.58, four weeks after initiation of simvastatin treatment.[13] Similarly, another large register-based cohort study including 1,363 patients found an increase in mean INR from 2.40 to 2.71 also peaking approximately four weeks after initiation of simvastatin treatment.[19] Furthermore, it was shown that high-dose (>40 mg) and low-dose simvastatin (< 40 mg) led to comparable changes in INR (increase of 0.33 vs 0.29).[19] A smaller study (n=29) reported an increase in mean INR from 2.50 at baseline to 3.15 after initiation of simvastatin treatment.[14] In this study, the INR values were obtained from the patient’s visit to the anticoagulation clinic prior to initiating simvastatin treatment and from the subsequent clinic visit after simvastatin had been initiated, with no information about the number of days/weeks in between.[14] Two studies reported a mean reduction of the daily warfarin dose following statin initiation of 7% and 9%, respectively.[13, 14] An observational case-control study assessed whether initiation of fibrates or statins in chronic warfarin users increased the risk of hospitalization due to gastrointestinal (GI) bleeding.[13] The study included warfarin patients initiating treatment with fluvastatin (n = 16), simvastatin (n = 277), atorvastatin (n = 499) and pravastatin (n = 113). Each patient case with GI bleeding was matched with up to 50 controls, whom had initiated treatment with the same type of statin as the case, but had not been hospitalized due to GI bleedings. In chronic warfarin users, the risk of GI bleeding was highest 31-60 days after simvastatin initiation with an Odds Ratio (OR) of 1.60 (95% CI 1.07-2.39), and for atorvastatin the risk of GI bleeding was highest 1-30 days after initiating treatment with an OR of 1.39 (95% CI 1.07-1.81). Concerning fluvastatin treatment, there were too few cases to obtain reliable estimates. No increased risk of GI bleeding was observed, when initiating treatment with the chosen reference drug; pravastatin, as pravastatin is not metabolized by CYP enzymes and therefore was not expected to interact with warfarin.[18]
A pilot study of warfarin-associated drug interactions did not find any changes in INR related to initiation of statin treatment, when patients with INR >1.8 and <3.2 in the period of 60 days prior to initiation of statin treatment, were categorized in four groups according to the change in INR. The four groups compromised: nonevents (2.0<INR<3.0), increasing (INR≥4.0), decreasing (INR≤1.5) and residual. Most patients were in the nonevent group (67.9%), while 7.8% and 7.1% experienced increased and decreased INR levels, respectively, and 17.3% patients were in the residual group in between the other three groups.[20] No specification of the type of the included statins was stated, and neither were the size of the exposed population of warfarin users, therefore this study was considered to have lesser relevance to the evaluation of the effect of initiation of statin treatment on level of coagulation.
In another study, the immediate (on average 1 week after initiation of statin treatment) and long-term (6-12 weeks after initiation of statin treatment) effects of statins on INR in 435 patients in phenprocoumon treatment and 303 patients in acenocoumarol treatment were studied.[17] The patients treated with phenprocoumon (n = 435) and acenocoumarol (n = 303), respectively, initiated treatment with the statins simvastatin, atorvastatin, pravastatin, rosuvastatin or fluvastatin. For all statins, the immediate effect on INR was an increase of 0.1 among phenprocoumon users, while no effect was observed for acenocoumarol. VKA dose was decreased with 0.02 mg/day in phenprocoumon users and 0.04 mg/day in acenocoumarol users within the first six weeks after initiation of statin treatment, and reduced furthermore with 0.07 mg/day and 0.11 mg/day respectively after 12 weeks of statin treatment.[17]
Finally, two prospective clinical studies investigated the potential drug interaction between rosuvastatin and warfarin and atorvastatin and warfarin, respectively. One study investigated the effect of initiation and dosage increasement of rosuvastatin in seven patients in stable warfarin treatment. Initiation of treatment with 10 mg rosuvastatin per day in a period of up to 14 days resulted in an INR > 4 in two out of the seven patients. Four out of the remaining five patients experienced an INR > 4 within the next 14 days, when rosuvastatin dosage was increased to 80 mg.[15] A small study with 12 patients in warfarin treatment described the change in mean prothrombin time the first 15 days after initiation of atorvastatin treatment. The mean prothrombin time remained unchanged except for a small yet statistically significant decrease of 1.6 seconds from day 3-5 after initiation of atorvastatin treatment. However, the decrease was not considered therapeutically important.[16]