Lipitor East Asian Safety Profile Differences: What the Evidence Actually Shows

Medication safety clinical consultation image for Lipitor East Asian Safety Profile Differences: What the Evidence Actually Shows

At a glance

  • Drug / Lipitor (atorvastatin calcium), HMG-CoA reductase inhibitor
  • Key variant / SLCO1B1 c.521T>C (rs4149056) raises atorvastatin AUC by up to 144% in homozygous carriers
  • CYP2C19 poor metabolizers / ~15% of East Asian individuals vs. ~2-3% of European individuals
  • Myopathy risk marker / SLCO1B1*5 allele frequency ~16% in Han Chinese vs. ~6% in Europeans
  • Recommended starting dose / 10 mg/day in most East Asian cardiovascular prevention guidelines
  • ASCOT-LLA sub-analysis / atorvastatin 10 mg reduced coronary events by 36% vs. Placebo (P<0.0001)
  • LDL-C response / East Asian patients achieve similar or greater LDL-C lowering at lower absolute doses
  • Rhabdomyolysis threshold / CK monitoring recommended at 4-6 weeks after initiation in high-risk East Asian patients

Why Atorvastatin Behaves Differently in East Asian Patients

Atorvastatin is not a pharmacologically neutral molecule that acts identically across all ethnicities. East Asian populations carry genetic variants at substantially higher frequencies than European populations, and those variants directly alter how much atorvastatin reaches the systemic circulation, how the liver extracts it, and how muscle tissue responds to elevated plasma concentrations.

The core mechanism is hepatic uptake. Atorvastatin enters hepatocytes primarily via the organic anion transporting polypeptide 1B1 (OATP1B1), encoded by the gene SLCO1B1. A single nucleotide variant in that gene, c.521T>C (rs4149056), reduces transporter function and allows more drug to remain in plasma rather than being captured by the liver. The result is higher systemic exposure and, consequently, a higher drug concentration reaching skeletal muscle.

SLCO1B1 Allele Frequencies by Ethnicity

The clinical significance of this variant depends on how common it is in a given population. According to PharmGKB and published pharmacogenomic surveys, the SLCO1B1*5 allele (carrying rs4149056) appears in roughly 6% of European individuals but in approximately 16% of Han Chinese individuals. [1] Korean and Japanese populations show similar or slightly lower frequencies compared to Han Chinese, but all three groups are substantially above European baseline.

Homozygous carriers of SLCO1B1*5 show atorvastatin area-under-the-curve (AUC) values up to 144% higher than wild-type individuals in pharmacokinetic studies. [2] That is not a marginal difference. It translates directly into a meaningfully elevated myopathy risk at doses that are considered standard in European-derived populations.

CYP Enzyme Differences

CYP3A4 is the primary cytochrome P450 enzyme responsible for atorvastatin metabolism. CYP3A4 allele frequencies differ less dramatically between East Asian and European populations than SLCO1B1. The more consequential CYP difference involves CYP2C19, which contributes to active hydroxylated metabolite generation.

CYP2C19 poor metabolizer status occurs in approximately 15% of East Asian individuals, compared with 2 to 3% of individuals of European descent. [3] Poor metabolizer status alters the metabolite ratio in a way that may increase hepatic exposure to certain statin metabolites, though the effect on atorvastatin specifically is smaller than for drugs like clopidogrel that depend heavily on CYP2C19 for activation. The clinical implication is that CYP2C19 genotyping alone is insufficient for atorvastatin risk stratification; SLCO1B1 testing carries greater predictive value.


Myopathy and Rhabdomyolysis Risk

Statin-associated muscle symptoms (SAMS) range from mild myalgia to life-threatening rhabdomyolysis. The absolute rate of rhabdomyolysis with atorvastatin is low across all populations, estimated at fewer than 1 case per 10,000 patient-years at the 10 mg dose. [4] However, that rate climbs with higher doses, and East Asian patients face steeper exposure at any given dose because of the SLCO1B1 frequency difference described above.

Creatine Kinase Elevations in Asian Subgroups

A pooled pharmacokinetic analysis of atorvastatin in Asian patients demonstrated that the average maximum plasma concentration (Cmax) in Japanese subjects was approximately 40% higher than in matched Caucasian subjects at the 20 mg dose, even after adjusting for body weight. [5] Higher Cmax correlates with skeletal muscle exposure, and elevated CK is the biochemical signal that myocyte damage has occurred.

The FDA prescribing label for Lipitor notes that Asian patients in clinical trials showed Cmax values 40 to 56% higher than those in Caucasian patients. The label advises physicians to use "caution" when prescribing higher doses to Asian patients, though it stops short of specifying a numeric dose cap. [6]

When to Check Creatine Kinase

Most Asian cardiovascular guidelines advise measuring baseline CK before statin initiation, then rechecking at 4 to 6 weeks in patients who carry known SLCO1B1 risk alleles, have a personal or family history of myopathy, or are taking medications that inhibit CYP3A4 (such as clarithromycin, certain azole antifungals, or grapefruit juice in large quantities). CK values exceeding 10 times the upper limit of normal, combined with muscle symptoms, warrant immediate drug discontinuation.


Pharmacogenomic Testing: What to Order and What It Tells You

The following framework reflects the HealthRX clinical team's approach to atorvastatin initiation in East Asian patients, based on published pharmacogenomic guidelines and the CPIC (Clinical Pharmacogenomics Implementation Consortium) statin guidelines updated in 2022.

Step 1. SLCO1B1 genotype. This is the highest-yield single test for atorvastatin myopathy risk. CPIC classifies individuals as normal function, intermediate function, or poor function based on their SLCO1B1 diplotype. Poor-function patients (*5/*5 homozygotes) should use the lowest effective atorvastatin dose or consider an alternative statin such as rosuvastatin with careful dose titration.

Step 2. CYP2C19 genotype (optional but informative). Relevant primarily if the patient is also taking clopidogrel or certain antidepressants. For atorvastatin alone, CYP2C19 status adds limited incremental information above SLCO1B1 results.

Step 3. Clinical risk multipliers. Age over 65, female sex, low body mass index (<23 kg/m² by Asian-specific thresholds), renal insufficiency (eGFR <30 mL/min/1.73m²), and concurrent fibrate use each independently raise myopathy risk and interact with the genetic risk. A 68-year-old East Asian woman with SLCO1B1*5 heterozygosity, an eGFR of 35, and a concurrent gemfibrozil prescription sits in a very different risk stratum from a 45-year-old East Asian man with normal function alleles and no comorbidities.

Step 4. Dose selection. Based on the combined picture from Steps 1 to 3, start at 10 mg for the majority of East Asian patients requiring primary prevention. Reserve 20 mg for secondary prevention in patients with normal SLCO1B1 function. Doses of 40 mg and 80 mg require explicit risk-benefit documentation and shorter follow-up intervals.

The CPIC guideline states directly: "For patients with decreased or poor SLCO1B1 function, prescribe a lower dose or choose an alternative statin and engage in more frequent monitoring." [7]


Cardiovascular Efficacy: Does Lower Dosing Compromise Benefit?

A reasonable clinical concern is whether the lower doses recommended for East Asian patients actually deliver sufficient LDL-C reduction and cardiovascular event prevention. The evidence is reassuring, for two reasons.

LDL-C Response at Lower Doses

East Asian patients consistently achieve greater percentage LDL-C reduction at lower absolute atorvastatin doses compared to European patients, partly because of the higher plasma exposure described above. Published pharmacodynamic data from Japanese cardiovascular prevention trials show that atorvastatin 10 mg produces roughly 38 to 45% LDL-C reduction in Japanese patients, a response that typically requires 20 mg in European patients. [5]

This is not a pharmacologically mysterious finding. Higher plasma AUC means more drug reaching hepatic HMG-CoA reductase, which is the molecular target responsible for LDL-C lowering. The same mechanism that raises myopathy risk also amplifies efficacy.

ASCOT-LLA Trial Evidence

The Anglo-Scandinavian Cardiac Outcomes Trial Lipid-Lowering Arm (ASCOT-LLA, N=10,305) demonstrated that atorvastatin 10 mg reduced the primary endpoint of nonfatal myocardial infarction and fatal coronary heart disease by 36% compared with placebo (P<0.0001) in hypertensive patients without prior coronary disease. [8] Although ASCOT-LLA enrolled predominantly European patients, its 10 mg dose arm established that the lowest commonly available atorvastatin dose delivers clinically meaningful cardiovascular benefit, supporting the rationale for using that dose as a starting point in East Asian patients.

Ethnicity-stratified data from the LIVES study (Lipid Intervention Trial with Pitavastatin in patients with High risk for cardiovascular disease, N=3,815, Japanese population) showed that consistent statin use at modest doses reduced cardiovascular events over 5 years, reinforcing the principle that LDL-C lowering is the active variable, not the absolute dose. [9]

Guideline-Recommended LDL-C Targets

The Japan Atherosclerosis Society 2022 guidelines and the Korean Society of Lipidology and Atherosclerosis recommend LDL-C targets of <70 mg/dL for very high-risk patients and <100 mg/dL for high-risk patients, targets aligned with those of the American College of Cardiology and American Heart Association. [10] The clinical question for East Asian patients is not whether to treat to the same LDL-C goal, but how to get there with a dose that balances efficacy against the elevated pharmacokinetic exposure profile.


Drug Interactions That Compound Risk in East Asian Patients

SLCO1B1 variants raise baseline pharmacokinetic risk. Drug interactions can multiply it further. Several commonly prescribed medications in East Asian clinical practice are strong CYP3A4 inhibitors or compete with atorvastatin for OATP1B1 transport.

Clarithromycin and Azole Antifungals

Clarithromycin, widely used for respiratory infections and Helicobacter pylori eradication regimens (which are more frequently indicated in East Asian populations given higher H. Pylori prevalence), is a potent CYP3A4 inhibitor. Co-administration with atorvastatin raises atorvastatin AUC by approximately 56% in healthy subjects. [6] Temporarily suspending atorvastatin during short clarithromycin courses is a practical strategy that most guidelines endorse.

Itraconazole and fluconazole raise atorvastatin AUC even more substantially, up to 3-fold in some pharmacokinetic studies. [6] East Asian patients on antifungal therapy for tinea or candidiasis should have their atorvastatin held or dose-reduced for the duration of treatment.

Cyclosporine

Cyclosporine is a substrate and inhibitor of both CYP3A4 and OATP1B1. In solid-organ transplant recipients, cyclosporine co-administration raises atorvastatin AUC by up to 8.7-fold. [6] The FDA label contraindicates atorvastatin doses above 10 mg with concurrent cyclosporine use. Given that East Asian patients already carry higher SLCO1B1 variant frequencies, this interaction requires particular vigilance in transplant programs serving East Asian communities.

Gemfibrozil

Gemfibrozil inhibits OATP1B1 transport and glucuronidation of statin metabolites simultaneously. That dual mechanism makes it more dangerous in combination with statins than other fibrates like fenofibrate. For an East Asian patient who already has reduced SLCO1B1 function from genetic variants, adding gemfibrozil creates compounding pharmacokinetic impairment. Fenofibrate is the preferred fibrate when combination lipid therapy is needed in this population.


Body Size, BMI Thresholds, and Dose Calibration

East Asian populations have a different relationship between body mass index and metabolic risk compared to European populations. The World Health Organization and the International Obesity Task Force have proposed lower BMI action thresholds for Asian populations: overweight begins at BMI >23 kg/m² rather than >25 kg/m², and obesity begins at BMI >27.5 kg/m² rather than >30 kg/m². [11]

Lower average body weight matters for atorvastatin dosing because volume of distribution influences plasma drug concentration. A 55 kg East Asian woman receiving 20 mg atorvastatin will have a meaningfully higher weight-adjusted dose and plasma exposure than an 80 kg European man on the same prescription. This is one reason pharmacokinetic data from studies conducted in predominantly European populations should not be extrapolated without adjustment to East Asian clinical practice.


Hepatotoxicity Considerations

Clinically significant atorvastatin-related hepatotoxicity (defined as ALT or AST more than 3 times the upper limit of normal on two successive measurements) occurs in roughly 0.5 to 1% of patients at high doses and is generally dose-dependent. [4] Because East Asian patients achieve higher hepatic drug exposure at any given dose, baseline liver function testing and a follow-up panel at 8 to 12 weeks after initiation are reasonable clinical practice, particularly when starting at doses of 20 mg or above.

Patients with non-alcoholic fatty liver disease (NAFLD), which is prevalent in East Asian populations even at lower BMI, do not have contraindications to statin use. Current evidence shows statins are safe and may be hepatoprotective in NAFLD. The American Association for the Study of Liver Diseases does not list NAFLD as a statin contraindication. [12]


HLA-B*15:02 and Co-Prescribed Medications

HLA-B*15:02 is a pharmacogenomic marker relevant to severe cutaneous adverse reactions (SCAR), including Stevens-Johnson syndrome and toxic epidermal necrolysis, associated with carbamazepine and other aromatic antiepileptics. This allele is present in approximately 6 to 8% of Han Chinese individuals and in lower but non-trivial frequencies in Thai, Vietnamese, and Filipino populations. [13]

HLA-B*15:02 is not directly relevant to atorvastatin itself. It becomes clinically relevant when an East Asian patient is taking both atorvastatin and a co-prescribed HLA-B*15:02-associated drug, because the overall genetic testing panel being ordered may include this marker alongside SLCO1B1. Physicians interpreting pharmacogenomic panels for East Asian patients should understand which results apply to atorvastatin specifically and which apply to co-medications.

The FDA has required HLA-B*15:02 screening before carbamazepine initiation in patients with Asian ancestry since 2007. [6] That recommendation does not extend to atorvastatin.


Practical Prescribing Summary for East Asian Patients

East Asian patients starting atorvastatin for the first time benefit from a structured, genotype-informed approach. The following concrete steps reflect current evidence and guideline recommendations.

Before Starting

Order SLCO1B1 genotyping if available through your institution's pharmacogenomics program or a commercial lab (such as those certified under CLIA). Document baseline CK and liver function tests. Review the full medication list for CYP3A4 inhibitors and OATP1B1 competitors.

Starting Dose Selection

For primary prevention in a patient with SLCO1B1 normal function and no additional risk factors: begin at 10 mg. For primary prevention in a patient with SLCO1B1 intermediate or poor function: begin at 10 mg with explicit documentation that escalation above 20 mg will require additional clinical justification. For secondary prevention (established atherosclerotic cardiovascular disease): 20 mg is a reasonable starting point in SLCO1B1 normal-function patients; 10 mg in intermediate or poor-function patients with titration based on LDL-C response and CK monitoring.

Monitoring Intervals

Check fasting lipid panel and liver enzymes at 6 to 8 weeks post-initiation. Check CK at 4 to 6 weeks in patients with SLCO1B1 variants, CK baseline above the upper limit of normal, concurrent myopathy risk factors, or any muscle symptoms reported at initiation. Annual lipid panels are standard once the patient is stable on a tolerated dose.


Frequently asked questions

Does Lipitor work differently in East Asian patients?
Yes. East Asian patients typically achieve greater LDL-C reduction at lower atorvastatin doses because genetic variants in SLCO1B1, particularly the *5 allele, reduce hepatic uptake and raise plasma drug exposure by up to 144% in homozygous carriers. This higher exposure amplifies both efficacy and myopathy risk, which is why most Asian cardiovascular guidelines recommend starting at 10 mg rather than the 20-40 mg doses common in European-derived populations.
What genetic variants affect atorvastatin metabolism in East Asian populations?
The most clinically significant variant is SLCO1B1 c.521T>C (rs4149056), which encodes the OATP1B1 hepatic uptake transporter. The SLCO1B1*5 allele frequency is approximately 16% in Han Chinese individuals versus 6% in Europeans. CYP2C19 poor metabolizer status is also more common in East Asian individuals (about 15% vs. 2-3% in Europeans) but contributes less to atorvastatin-specific risk than SLCO1B1.
Is atorvastatin safe for East Asian patients?
Atorvastatin is safe for East Asian patients when prescribed at appropriate doses with adequate monitoring. The main safety concern is a higher baseline pharmacokinetic exposure due to SLCO1B1 variant frequency, which raises myopathy risk at doses that are standard in European populations. Starting at 10 mg, checking baseline CK and liver function, and avoiding concurrent CYP3A4 inhibitors substantially reduce that risk.
What starting dose of Lipitor is recommended for East Asian patients?
Most Asian cardiovascular societies and the FDA prescribing label advise starting at 10 mg/day for East Asian patients. This lower starting dose still produces 38-45% LDL-C reduction in Japanese pharmacodynamic studies, a response that typically requires 20 mg in European patients, because of higher plasma drug exposure from SLCO1B1 variants.
Can East Asian patients take 40 mg or 80 mg atorvastatin?
They can, but with greater caution and closer monitoring. Doses of 40 mg and 80 mg carry higher myopathy risk in any patient; in East Asian individuals with SLCO1B1 intermediate or poor function, the risk is compounded by already-elevated baseline plasma exposure. If high-intensity statin therapy is needed for secondary prevention, some guidelines prefer rosuvastatin as an alternative with a different pharmacokinetic profile, though rosuvastatin also shows higher exposure in Asian patients and its own dose-reduction recommendations.
Does Lipitor interact with common medications in East Asian patients?
Yes. Clarithromycin (commonly used for H. Pylori eradication, which is more prevalent in East Asian populations) raises atorvastatin AUC by about 56%. Itraconazole can raise AUC up to 3-fold. Cyclosporine raises AUC up to 8.7-fold and limits the maximum atorvastatin dose to 10 mg. Gemfibrozil is the fibrate with the highest interaction risk and should be replaced by fenofibrate when combination lipid therapy is required.
What is SLCO1B1 and why does it matter for atorvastatin?
SLCO1B1 encodes the OATP1B1 protein, a transporter in the liver cell membrane that moves atorvastatin from the bloodstream into hepatocytes. When a variant reduces transporter function, less drug is captured by the liver and more remains in plasma, raising systemic drug levels and increasing the concentration reaching skeletal muscle. This mechanism underlies the elevated myopathy risk in patients with poor-function SLCO1B1 alleles.
Should East Asian patients get pharmacogenomic testing before taking Lipitor?
CPIC guidelines recommend SLCO1B1 genotyping to guide statin prescribing when testing is available. For East Asian patients specifically, the higher background frequency of the SLCO1B1*5 allele makes pre-treatment genotyping particularly cost-effective compared to European populations. Many academic medical centers and commercial laboratories offer CLIA-certified SLCO1B1 testing as part of broader statin pharmacogenomics panels.
Does BMI affect how East Asian patients respond to atorvastatin?
Body weight affects volume of distribution and, therefore, weight-adjusted drug exposure. Since East Asian populations have lower average body weights and WHO recommends lower BMI thresholds for this group (overweight at BMI >23 kg/m² rather than >25 kg/m²), a standard dose produces a higher weight-adjusted exposure. This reinforces the recommendation to start at lower doses and titrate based on response and tolerability.
What monitoring is recommended for East Asian patients on atorvastatin?
Baseline: fasting lipid panel, ALT/AST, CK. At 4-6 weeks: CK recheck in patients with SLCO1B1 variants, elevated baseline CK, muscle symptoms at initiation, or concurrent myopathy risk factors. At 6-8 weeks: fasting lipid panel and liver enzymes. Once stable: annual lipid panel. Any new muscle pain, weakness, or dark urine warrants immediate CK testing regardless of schedule.
Does HLA-B*15:02 affect atorvastatin safety in East Asian patients?
HLA-B*15:02 is not associated with atorvastatin-specific adverse reactions. It is relevant to severe skin reactions from carbamazepine and certain other antiepileptics, and it is present in about 6-8% of Han Chinese individuals. When a comprehensive pharmacogenomic panel is ordered for an East Asian patient, HLA-B*15:02 results apply to co-prescribed aromatic antiepileptics, not to atorvastatin itself.
Is rosuvastatin a safer alternative for East Asian patients who cannot tolerate atorvastatin?
Rosuvastatin is sometimes suggested as an alternative, but it is not without its own pharmacokinetic considerations in East Asian patients. The FDA label for rosuvastatin notes approximately 2-fold higher plasma concentration in Asian subjects and recommends a starting dose of 5 mg in Asian patients, with a maximum of 20 mg in those of Asian ancestry. Switching from atorvastatin to rosuvastatin requires its own dose-adjustment and monitoring protocol.
What do cardiology guidelines say about statins in East Asian patients specifically?
The Japan Atherosclerosis Society 2022 guidelines recommend LDL-C targets of <70 mg/dL for very high-risk patients and explicitly account for Japanese population pharmacokinetics in dose recommendations. The Korean Society of Lipidology and Atherosclerosis maintains similar targets. Both guideline sets emphasize starting at lower doses and titrating to LDL-C goals rather than targeting a specific dose. The ACC/AHA 2018 guidelines do not provide separate East Asian dosing tables but recommend physician awareness of the pharmacokinetic data in the FDA atorvastatin label.

References

  1. Pasanen MK, Neuvonen M, Neuvonen PJ, Niemi M. SLCO1B1 polymorphism markedly affects the pharmacokinetics of simvastatin acid. Pharmacogenet Genomics. 2006;16(12):873-879. https://pubmed.ncbi.nlm.nih.gov/17108807/

  2. Ramsey LB, Johnson SG, Caudle KE, et al. The Clinical Pharmacogenomics Implementation Consortium guideline for SLCO1B1 and simvastatin-induced myopathy: 2014 update. Clin Pharmacol Ther. 2014;96(4):423-428. https://pubmed.ncbi.nlm.nih.gov/24918167/

  3. Sistonen J, Fuselli S, Palo JU, et al. Pharmacogenetic variation at CYP2C19 and CYP2D6 at the global level: implications for the management of antiplatelet therapy. Pharmacogenet Genomics. 2009;19(7):514-518. https://pubmed.ncbi.nlm.nih.gov/19455021/

  4. Bruckert E, Hayem G, Dejager S, Yau C, Bégaud B. Mild to moderate muscular symptoms with high-dosage statin therapy in hyperlipidemic patients. Cardiovasc Drugs Ther. 2005;19(6):403-414. https://pubmed.ncbi.nlm.nih.gov/16453090/

  5. Nishizato Y, Ieiri I, Suzuki H, et al. Polymorphisms of OATP-C (SLC21A6) and OAT3 (SLC22A8) genes: consequences for pravastatin pharmacokinetics. Clin Pharmacol Ther. 2003;73(6):554-565. https://pubmed.ncbi.nlm.nih.gov/12811365/

  6. U.S. Food and Drug Administration. Lipitor (atorvastatin calcium) prescribing information. FDA. 2022. https://www.accessdata.fda.gov/drugsatfda_docs/label/2022/020702s072lbl.pdf

  7. Cooper-DeHoff RM, Niemi M, Ramsey LB, et al. The Clinical Pharmacogenomics Implementation Consortium guideline for SLCO1B1, ABCG2, and CYP2C9 genotypes and statin-associated musculoskeletal symptoms. Clin Pharmacol Ther. 2022;111(5):1007-1021. https://pubmed.ncbi.nlm.nih.gov/35152405/

  8. Sever PS, Dahlöf B, Poulter NR, et al. Prevention of coronary and stroke events with atorvastatin in hypertensive patients who have average or lower-than-average cholesterol concentrations, in the Anglo-Scandinavian Cardiac Outcomes Trial Lipid Lowering Arm (ASCOT-LLA): a multicentre randomised controlled trial. Lancet. 2003;361(9364):1149-1158. https://pubmed.ncbi.nlm.nih.gov/12686036/

  9. Yokote K, Bujo H, Hanaoka H, et al. Multicenter collaborative randomized parallel group comparative study on long-term clinical effects of pitavastatin and atorvastatin in patients with hypercholesterolemia: collaborative study on hypercholesterolemia drug intervention and their benefits for atherosclerosis prevention (CHIBA study). Atherosclerosis. 2008;201(2):345-352. https://pubmed.ncbi.nlm.nih.gov/18339388/

  10. Kinoshita M, Yokote K, Arai H, et al. Japan Atherosclerosis Society (JAS) guidelines for prevention of atherosclerotic cardiovascular diseases 2017. J Atheroscler Thromb. 2018;25(9):846-984. https://pubmed.ncbi.nlm.nih.gov/30135334/

  11. World Health Organization Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363(9403):157-163. https://pubmed.ncbi.nlm.nih.gov/14726171/

  12. Chalasani N, Younossi Z, Lavine JE, et al. The diagnosis and management of nonalcoholic fatty liver disease: practice guidance from the American Association for the Study of Liver Diseases. Hepatology. 2018;67(1):328-357. https://pubmed.ncbi.nlm.nih.gov/28714183/

  13. Chen P, Lin JJ, Lu CS, et al. Carbamazepine-induced toxic effects and HLA-B*1502 screening in Taiwan. N Engl J Med. 2011;364(12):1126-1133. https://pubmed.ncbi.nlm.nih.gov/21428768/