Circulation Journal
Online ISSN : 1347-4820
Print ISSN : 1346-9843
ISSN-L : 1346-9843
Late Breaking Clinical Trials (JCS 2019)
Long-Term Treatment With Evolocumab Among Japanese Patients ― Final Report of the OSLER Open-Label Extension Studies ―
Atsushi HirayamaShizuya YamashitaAndrea RuzzaHyoe InomataMarcoli CyrilleChen LuAndrew W. HamerMasayuki YoshidaArihiro KiyosueTamio Teramoto
Author information

2019 Volume 83 Issue 5 Pages 971-977


Background: Treatment with evolocumab reduces mean low-density lipoprotein cholesterol (LDL-C) up to 75% and cardiovascular events by 16% in the first year and 25% thereafter.

Methods and Results: Japanese patients with hypercholesterolemia enrolled in the parent YUKAWA-1-2 studies could enroll, once eligible, in the OSLER studies (n=556). OSLER re-randomized patients 2:1 to evolocumab plus standard of care (SOC; evolocumab+SOC) or SOC alone for 1 year; after year 1, patients could enter the all-evolocumab+SOC open-label extension of OSLER. Patients received evolocumab+SOC from the 2nd year through up to 5 years. Long-term efficacy and safety, including antidrug antibodies, were evaluated. Of 556 patients, 532 continued to the all-evolocumab+SOC extension: mean (standard deviation [SD]) age 61 (10) years, 39% female. A total of 91% of 532 patients completed the studies. Mean (SD) LDL-C change from parent-study baseline with evolocumab from a mean (SD) baseline of 142.3 (21.3) and 105.0 (31.1) mg/dL in OSLER-1 and OSLER-2, respectively, was maintained through the end of the study: −58.0% (19.1%) at year 5 in OSLER-1, −62.7% (25.6%) at year 3 in OSLER-2. The overall safety profile of the evolocumab+SOC periods was similar to that of the year-1 controlled period. Antidrug antibodies were detected transiently in 3 patients. No neutralizing antibodies were detected.

Conclusions: Japanese patients who continued evolocumab+SOC for up to 5 years experienced sustained high LDL-C level reduction. Long-term evolocumab+SOC exposure showed no new safety signals.

Proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitors have emerged as a novel treatment option for people requiring further low-density lipoprotein cholesterol (LDL-C) lowering despite maximally-tolerated statin therapy.13 Studies of the PCSK9 inhibitor evolocumab added to maximally tolerated statins reported mean reduction of LDL-C by approximately 60% to 75% in the Japanese population,4,5 which is consistent with the efficacy shown in global studies.6 In the placebo-controlled, global cardiovascular outcomes study, FOURIER, evolocumab resulted in a relative risk reduction in cardiovascular death, stroke, or myocardial infarction by 20% (16% in the first year and 25% thereafter).2 The 2017 Japan Atherosclerosis Society (JAS) guidelines for atherosclerotic cardiovascular disease (ASCVD) risk prevention therefore recommend PCSK9 inhibitor use in appropriate patients.1

Recent reports have confirmed the long-term safety and efficacy of evolocumab in broad international studies7 and up to 1 year in the Japanese population.8 Longer-term use of evolocumab among Japanese patients has not yet been reported. Such data may be informative for Japanese clinicians treating patients with hypercholesterolemia or mixed dyslipidemia who require further LDL-C lowering than can be achieved with maximally-tolerated statin therapy.

We report the longest final safety and efficacy results of PCSK9 inhibition with evolocumab in the subgroup of Japanese patients of the OSLER studies.


Study Design

The design and interim results of the OSLER studies have been previously described.2,7,9 Briefly, Japanese patients enrolled in the randomized YUKAWA-1 and YUKAWA-2 studies4,5 (NCT01652703 and NCT01953328) were given the opportunity to enroll in the open-label extension studies OSLER-1 and OSLER-2 (NCT01439880 and NCT01854918),710 respectively. The YUKAWA studies4,5 enrolled patients with elevated LDL-C and high cardiovascular risk as defined by the JAS.10 Patients were re-randomized to 1 year of open-label standard of care (SOC) or evolocumab+SOC,8 after which patients received open-label evolocumab+SOC from the 2nd year through up to 5 and 3 years from the OSLER study baseline in OSLER-1 and OSLER-2, respectively. We report the results of Japanese patients enrolled in the OSLER studies. Study protocols were approved by the independent ethics committee at each study site. Amgen Inc. sponsored and designed the YUKAWA and OSLER studies and was responsible for data collection and analysis. All patients provided written informed consent before study enrollment.


Patients who completed either of the YUKAWA studies and did not experience a treatment-related serious adverse event (AE) that led to study drug discontinuation were eligible. Major exclusion criteria included pregnancy or breastfeeding and disorders that would interfere with study participation, in the opinion of the investigator.

Study Treatment in the All-Evolocumab+SOC Period

After completion of the SOC-controlled period, patients who continued then received subcutaneously administered evolocumab 420 mg once monthly (OSLER-1) or a choice of this dose or 140 mg once every 2 weeks (OSLER-2). Investigators, patients, and research staff were unblinded to lipids after the first 12 weeks of the OSLER studies. The SOC for the Japan OSLER cohort was defined according to JAS guidelines.1,10


The 1-year visit was at week 52 for OSLER-1 and week 48 for OSLER-2 and continued to approximately 5 and 2.7 years, respectively. Analyses of LDL-C used a reflexive testing approach, whereby the calculated LDL-C values were used in the analysis except when <40 mg/dL or triglycerides (TG) were >400 mg/dL; these values were replaced with ultracentrifugation LDL-C values for the analysis when available.


The primary safety objective was to characterize the safety and tolerability of longer-term administration of evolocumab (subject incidence of AEs).

Secondary and exploratory objectives were to characterize the effects of longer-term administration of evolocumab as assessed by LDL-C, high-density lipoprotein cholesterol (HDL-C), TG, very LDL-C (VLDL-C), non-HDL-C, total cholesterol, lipoprotein(a) (Lp[a]), apolipoprotein (Apo) B and ApoA1.

Another safety endpoint was the incidence of anti-evolocumab antibodies.

Statistical Analysis

For baseline characteristics and safety analyses, the Japanese populations from OSLER-1 and OSLER-2 were pooled. For efficacy analyses, the studies were analyzed separately because of the differing visit schedules. Baseline was defined throughout the analyses as the patient’s baseline in the parent YUKAWA-1 or YUKAWA-2 study. AEs were coded using the Medical Dictionary for Regulatory Activities version 21.0; severity assessments were made using the Common Terminology Criteria for Adverse Events version 4.03. All analyses are presented descriptively. There was no imputation for missing values. Data were analyzed using SAS software version 9.4 (SAS Institute).



YUKAWA-1 (n=307) and YUKAWA-2 (n=404) patients were able to choose to participate in the open-label extension studies OSLER -1 and OSLER-2 (Figure 1). Of 711 patients, 556 entered the open-label extension studies and were re-randomized to receive SOC or evolocumab+SOC for 1 year.

Figure 1.

Patient disposition. SOC, standard of care.

A total of 532 of 556 (96%) Japanese OSLER-1 and OSLER-2 patients continued from the year-1 randomized, controlled period to the all-evolocumab+SOC extension.

Of these 532 patients, 18 (3%) did not receive further evolocumab treatment and therefore did not continue in the study, and 514 (97%) received at least 1 dose of evolocumab. In the evolocumab+SOC extension (years 2 to up to 5), a total of 457 patients (86%) completed evolocumab treatment and 482 (91%) completed the study.

The mean (standard deviation [SD]) duration of exposure to evolocumab in the evolocumab+SOC group during year 1 was 12.7 (1.2) months in OSLER-1 and 10.6 (1.7) months in OSLER-2. The mean (SD) duration of exposure to evolocumab in the all-evolocumab+SOC period (from year 2 to end of study) was 43.7 (9.6) and 21.2 (5.6) months, respectively.

At baseline, the cohort had a mean (SD) age of 60.9 (9.9) years, and 61% of patients were male. Most patients (62%) had a high risk of coronary heart disease (CHD), as defined by the National Cholesterol Education Program (CHD, CHD-risk equivalent, or ≥2 risk factors plus a 10-year risk for CHD >20%); 15% had coronary artery disease (CAD); 13% had cerebrovascular or peripheral artery disease; 45% had type 2 diabetes mellitus; and 28% had metabolic syndrome without diabetes.

Baseline patient characteristics are displayed in Table 1.

Table 1. Baseline Demographic Characteristics of the Japanese Population From the OSLER Studies
Characteristic Pooled OSLER
Japan cohort (n=556)
Age, years, mean (SD) 60.9 (9.9)
Male sex, n (%) 337 (60.6)
NCEP CHD high risk,a n (%) 343 (61.7)
Coronary artery disease, n (%) 84 (15.1)
Cerebrovascular or peripheral artery disease, n (%) 71 (12.8)
Cardiovascular risk factors, n (%)
 Current cigarette use 142 (25.5)
 Type 2 diabetes mellitus 248 (44.6)
 Hypertension 412 (74.1)
 Family history of premature CHD 53 (9.5)
 Low HDL-C 80 (14.4)
 ≥2 cardiovascular risk factors 317 (57.0)
 Metabolic syndrome without diabetesb 156 (28.1)

When the calculated LDL-C was <40 mg/dL or triglycerides were >400 mg/dL, calculated LDL-C was replaced with ultracentrifugation LDL-C from the same blood sample, if available. Baseline was defined as the parent-study baseline. aCHD, CHD-risk equivalent, or ≥2 risk factors plus 10-year risk for CHD >20%. bDefined as ≥3 of the following and without diabetes mellitus: increased waist circumference, triglycerides ≥150 mg/dL, low HDL-C, systolic blood pressure ≥130 mmHg or diastolic blood pressure ≥85 mmHg or hypertension, or fasting glucose ≥100 mg/dL. Low HDL-C was defined as baseline HDL-C <40 mg/dL in men and <50 mg/dL in women. Increased waist circumference was defined as ≥90 cm for Asian men and ≥80 cm for Asian women. CHD, coronary heart disease; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; NCEP; National Cholesterol Education Program; SD, standard deviation.

Efficacy Results

As previously reported,8 the mean (SE) percent change treatment difference in LDL-C from baseline to year 1 with evolocumab+SOC vs. SOC alone was −69.1% (1.2%) in OSLER-1 and −65.1% (2.2%) in OSLER-2. The mean (SD) baseline LDL-C was 142.3 (21.3) mg/dL in OSLER-1 and 105.0 (31.1) mg/dL in OSLER-2. From year 1 through to the end of the studies, the mean (SD) LDL-C changed by approximately −60% from baseline (ranging between −58% and −68%) at all visits, and by −58.0% (19.1%) and −62.7% (25.6%) at the end of OSLER-1 and OSLER-2, respectively (Figure 2, Table 2).

Figure 2.

Plot of mean percent change from baseline in LDL-C by scheduled visit and treatment group for Japanese patients in OSLER-1 (A) and OSLER-2 (B). When the calculated LDL-C was <40 mg/dL or triglycerides were >400 mg/dL, calculated LDL-C was replaced with ultracentrifugation LDL-C from the same blood sample, if available. Baseline is defined as the parent-study baseline. Vertical lines represent the standard error around the mean. Plot is based on observed data and no imputation is used for missing values. EOS, end of study; IP, investigational product; LDL-C, low-density lipoprotein cholesterol; OLE, open-label extension; SOC, standard of care; w, week.

Table 2. Percent Change From Baseline in Lipids of the Japan Population From the OSLER Studies
 Baseline, mg/dL, mean (SD) 142.3 (21.3) 105.0 (31.1)
 Percent change at ∼2 years, mean (SD)a −63.2 (15.9) −64.5 (21.4)
 Percent change at EOS, %, mean (SD)b −58.0 (19.1) −62.7 (25.6)
 Baseline, mg/dL, mean (SD) 54.8 (12.9) 56.8 (13.7)
 Percent change at ∼2 years, mean (SD)a 12.8 (16.5) 10.1 (16.8)
 Percent change at EOS, mean (SD)b 11.9 (19.5) 9.4 (17.3)
 Baseline, mg/dL, median (IQR) 125.3 (98.0 to 160.0) 121.5 (92.0 to 161.5)
 Percent change at ∼2 years, median (IQR)a −15.8 (−32.0 to 2.6) −14.1 (−32.2 to 8.0)
 Percent change at EOS, median (IQR)b −15.7 (−35.3 to 13.9) −16.8 (−35.0 to 15.5)
 Baseline, mg/dL, mean (SD) 26.2 (11.4) 26.5 (11.5)
 Percent change at ∼2 years, mean (SD)a −4.1 (42.9) −5.8 (40.9)
 Percent change at EOS, mean (SD)b 2.0 (52.7) −6.6 (37.7)
 Baseline, mg/dL, mean (SD) 168.4 (25.5) 132.3 (35.3)
 Percent change at ∼2 years, mean (SD)a −56.4 (15.8) −57.4 (20.9)
 Percent change at EOS, mean (SD)b −52.4 (18.9) −55.8 (24.3)
Total cholesterol
 Baseline, mg/dL, mean (SD) 223.2 (26.3) 189.1 (35.9)
 Percent change at ∼2 years, mean (SD)a −39.7 (12.3) −37.1 (15.0)
 Percent change at EOS, mean (SD)b −37.0 (14.9) −36.4 (17.2)
 Baseline, nmol, median (IQR) 33.0 (13.0 to 63.0) 34.0 (13.0 to 58.0)
 Percent change at ∼2 years, median (IQR)a −39.5 (−56.4 to −14.6) −48.7 (−63.8 to −21.7)
 Percent change at EOS, median (IQR)b −28.6 (−46.7 to 0) NAc
 Baseline, mg/dL, mean (SD) 111.5 (16.9) 93.3 (22.1)
 Percent change at ∼2 years, mean (SD)a −54.4 (13.9) −55.7 (18.8)
 Percent change at EOS, mean (SD)b −48.4 (16.9) NAc
 Baseline, mg/dL, mean (SD) 157.1 (22.9) 161.8 (25.6)
 Percent change at ∼2 years, mean (SD)a 6.7 (12.1) 5.5 (13.0)
 Percent change at EOS, mean (SD)b 8.3 (12.9) NAc

When the calculated LDL-C was <40 mg/dL or triglycerides were >400 mg/dL, calculated LDL-C was replaced with ultracentrifugation LDL-C from the same blood sample, if available. Baseline is defined as the parent-study baseline. aWeek 100 in OSLER-1, week 104 in OSLER-2. bWeek 260 in OSLER-1, week 152 in OSLER-2. cData available through week 104. Apo, apolipoprotein; EOS, end of study; IQR, interquartile range; Lp(a), lipoprotein(a); NA, not applicable; VLDL-C, very-low-density lipoprotein cholesterol. Other abbreviations as in Table 1.

Mean (SD) percent changes from baseline to the end of the study for other lipids were as follows: HDL-C, 11.9% (19.5%) and 9.4% (17.3%); VLDL-C, 2.0% (52.7%) and −6.6% (37.7%); non-HDL-C, −52.4 (18.9%) and −55.8 (24.3%); total cholesterol, −37.0% (14.9%) and −36.4% (17.2%); ApoB, −48.4 (16.9) and −55.7 (18.8); and ApoA1, 8.3% (12.9%) and 5.5% (13.0%) in OSLER-1 and OSLER-2, respectively. Median (interquartile range) percent change from baseline to end of study for TG was −15.7% (−35.3% to 13.9%) and −16.8% (−35.0% to 15.5%); and for Lp(a), −28.6% (−46.7% to 0%) and −48.7 (−63.8% to −21.7%) in OSLER-1 and OSLER-2, respectively (Table 2).

Efficacy results are displayed in Table 2.

Safety Results

As previously reported,8 the proportion of patients with AEs was 72.0% with SOC alone and 74.9% with evolocumab+SOC during year 1. The proportion of patients experiencing at least 1 AE in all years of evolocumab exposure was 90.8%; serious AE, 18.6%; and leading to evolocumab discontinuation, 3.3%. These rates were similar to the year-1 SOC alone group when analyzed by year of evolocumab exposure. Rates of neurocognitive-related, potential hypersensitivity and injection-site reactions, and muscle-related AEs were similar by year of evolocumab exposure to that of the year-1 SOC alone group (Table 3).

Table 3. Summary of Patient Incidence of AEs by Years of Exposure in the Evolocumab Open Period of the Japan Population From the OSLER Studies
  No. (%) of patients with ≥1 event
Year(s) of evolocumab exposure
SOC patients
at year
1 (n=186)
All years of
Any AE 132 (71.0) 415 (76.6) 356 (71.1) 263 (63.5) 147 (75.0) 91 (72.8) 492 (90.8)
 Grade ≥2 96 (51.6) 296 (54.6) 252 (50.3) 185 (44.7) 93 (47.4) 60 (48.0) 430 (79.3)
 Grade ≥3 22 (11.8) 50 (9.2) 42 (8.4) 26 (6.3) 17 (8.7) 8 (6.4) 119 (22.0)
 Grade ≥4 1 (0.5) 3 (0.6) 1 (0.2) 1 (0.2) 0 0 7 (1.3)
 Serious 17 (9.1) 40 (7.4) 35 (7.0) 20 (4.8) 10 (5.1) 8 (6.4) 101 (18.6)
 Leading to evolocumab
NA 6 (1.1) 7 (1.4) 1 (0.2) 2 (1.0) 0 18 (3.3)
Most commonly occurring
 Nasopharyngitis 53 (28.5) 148 (27.3) 140 (27.9) 104 (25.1) 72 (36.7) 35 (28.0) 282 (52.0)
 Diabetes mellitus 7 (3.8) 44 (8.1) 19 (3.8) 10 (2.4) 11 (5.6) 4 (3.2) 76 (14.0)
 Back pain 2 (1.1) 26 (4.8) 17 (3.4) 25 (6.0) 7 (3.6) 10 (8.0) 71 (13.1)
 Bronchitis 6 (3.2) 26 (4.8) 25 (5.0) 8 (1.9) 5 (2.6) 4 (3.2) 52 (9.6)
 Diarrhea 2 (1.1) 23 (4.2) 14 (2.8) 11 (2.7) 9 (4.6) 2 (1.6) 52 (9.6)
AEs of interest
 Potential hypersensitivitya 19 (10.2) 67 (12.4) 54 (10.8) 35 (8.5) 16 (8.2) 12 (9.6) 137 (25.3)
 Potential injection-site reactionsa 0 11 (2.0) 6 (1.2) 2 (0.5) 2 (1.0) 1 (0.8) 21 (3.9)
 Muscle relateda 5 (2.7) 27 (5.0) 17 (3.4) 15 (3.6) 7 (3.6) 0 57 (10.5)
 Neurocognitive relatedb 0 1 (0.2) 2 (0.4) 1 (0.2) 1 (0.5) 0 5 (0.9)

aBased on US Food and Drug Administration search terms. bBased on Medical Dictionary for Regulatory Activities search terms. AE, adverse event; NA, not applicable; SOC, standard of care.

Antidrug antibodies were detected transiently in 3 patients: at week 12 of YUKAWA-1 and on day 1 of OSLER-1 (randomized to placebo in YUKAWA-1 and evolocumab+SOC in year 1 of OSLER-1); weeks 4 and 48 of OSLER-1 randomized period (randomized to placebo in YUKAWA-1 and evolocumab+SOC in year 1 of OSLER-1); and on day 1 of YUKAWA-2 (randomized to evolocumab in parent study and evolocumab+SOC in year 1 of OSLER-2). No patients tested positive for neutralizing antidrug antibodies; no patients tested positive for antidrug antibodies in years 2 through 3 or 5 (OSLER-2 and -1, respectively).


In the final report of the OSLER studies’ Japanese cohort, patients who continued evolocumab+SOC up to 5 years experienced profound and consistent LDL-C level reduction. Long-term evolocumab+SOC exposure showed no new safety signals, and the safety profile was similar to that of the OSLER year-1 control SOC alone group. No neutralizing antibodies were found.

There was consistent LDL-C lowering among Japanese patients who continued/started evolocumab+SOC in OSLER after year 1 through up to 5 years. Indeed, the LDL-C lowering profile has been consistently between 60% and 70% regardless of the monthly subcutaneous dose from the 12-week YUKAWA studies,4,5 through the 1-year controlled OSLER studies’ randomized periods,8 and ≥1- to 5-year OSLER evolocumab+SOC period. Likewise, persistent increases in HDL-C and ApoA1 and decreases in non-HDL-C, TG, and Lp(a) were similar to those observed in the YUKAWA studies and year 1 of the OSLER studies.4,5,8 As was found in short-term studies of evolocumab, there were more substantial LDL-C reductions among the OSLER Japanese patients compared with those of the global, similarly long-term OSLER-1 and FOURIER studies (60–70% vs. 50–60%).2,7

After long-term evolocumab use, most AEs were mild, and few led to evolocumab discontinuation. The most commonly occurring AEs were nasopharyngitis, diabetes, back pain, bronchitis, and diarrhea. As was reported in year 1 of OSLER,8 diabetes-related AEs were reported as complications in patients with diabetes at baseline. The rates of serious AEs, AEs leading to evolocumab discontinuation, and events of interest (new-onset diabetes, cognitive-related events, injection-site reactions) were similar to those observed in the long-term, global OSLER-1 and FOURIER studies.2,7

The pooled analysis of the open-label extension OSLER studies’ year-1 randomized vs. SOC alone-controlled period provided the first indication of significant reductions in the incidence of adverse cardiovascular events with evolocumab added to SOC.11 This result was confirmed among patients with clinically evident ASCVD in the event-driven, double-blind, placebo-controlled FOURIER study.2 The FOURIER study found that evolocumab treatment over a median of 2.2 years resulted in a relative risk reduction of the composite endpoint of cardiovascular death, myocardial infarction, or stroke by 20%. Among the 3835 patients enrolled in the Asia-Pacific region (including Japan), there was not a significant difference in risk reduction compared with the overall population.2 The overall safety profile observed in FOURIER verified the tolerability of evolocumab, with most AE rates being similar to those with placebo.2 The present pooled, open-label extension analysis complements FOURIER by providing the longest-term follow-up of PCSK9 inhibition in a Japanese cohort, verifying the continued safety, efficacy, and tolerability of evolocumab up to 5 years.

This analysis was of extended follow-up in an open-label extension study. Patients and investigators were all unblinded to treatment and lipid outcomes. These limitations are mitigated by the length of the follow-up (up to 5 years) and the consistency in safety and efficacy with other Japanese and international long-term studies.4,5,7,8


This analysis of the OSLER-1 and OSLER-2 from year 1 through up to 5 years demonstrated that evolocumab treatment persistently lowers LDL-C with a reassuring safety profile in high-risk Japanese patients with hyperlipidemia or mixed dyslipidemia. This report on Japanese patients with the longest evolocumab exposure did not lead to any new safety findings.


We thank Tim Peoples, MA, ELS, of Amgen Inc. for medical writing assistance.

Data Access Policy

Qualified researchers may request data from Amgen clinical studies. Complete details are available at the following:


A.H. received research grants from Daiichi-Sankyo Pharma, Bayer, Bristol-Myers Squibb, Pfizer Japan, Boston Scientific Japan, Astellas, AstraZeneca, MSD, Otsuka, Kyowa Hakko Kirin, Sanofi, Takeda, Mitsubishi Tanabe, Sumitomo Dainippon, Nippon Boehringer Ingelheim, and Nihon Medi-Physics; was in speakers’ bureaus for Daiichi-Sankyo, Bayer, Bristol-Myers Squibb, and AstraZeneca. S.Y. received consulting fees and was in a speakers’ bureau for Amgen. A.R., M.C., C.L., and A.W.H. are employees and stockholders of Amgen. H.I. is a former employee of Amgen Astellas and stockholder of Amgen. M.Y. and A.K. declare no conflicts of interest. T.T. received research grants from Daiichi-Sankyo, Kowa, Eli Lilly, Takeda, and Shionogi; was on speakers’ bureaus for Bayer, Pfizer, Daiichi-Sankyo, Kowa, Takeda, Astellas, Kissei, Sanofi, and Amgen Astellas Biopharma Partnership.

Source of Funding

These studies were sponsored by Amgen Inc.